Polymer Surface Modification: Relevance to Adhesion Volume 4
Polymer Surface Modification: Relevance to Adhesion Volume 4 Edited by
K.L. Mittal
LEIDEN y BOSTON 2007
VSP (an imprint of Brill Academic Publishers) P.O. Box 9000 2300 PA Leiden The Netherlands
Tel: +31 71 535 3500 Fax: +31 71 531 7532
[email protected] www.brill.nl
© VSP 2007 First published in 2007 ISBN 978-90-6764-453-2 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. PRINTED IN THE NETHERLANDS BY RIDDERPRINT BV, RIDDERKERK
Contents
vii
Preface Part 1. Surface Modification Techniques Importance of process conditions in polymer surface modification: A critical assessment J. Grace, H. K. Zhuang and L. Gerenser
3
Pretreatment and surface modification of polymers via atmospheric-pressure plasma jet treatment U. Lommatzsch, M. Noeske, J. Degenhardt, T. Wübben, S. Strudthoff, G. Ellinghorst and O.-D. Hennemann
25
The effects of excimer laser irradiation on surface morphology development in stretched poly(ethylene terephthalate), poly(butylene terephthalate) and polystyrene films J. Kim, D. U. Ahn and E. Sancaktar
33
®
XeCl excimer laser treatment of Vectran fibers in diethylenetriamine (DETA) environment J. Zeng and A. N. Netravali
87
Surface modification of polymers by ozone. Comparison of polyethylene and polystyrene treated at different temperatures T. Kobayashi and H. Kumagai
113
Atomic force microscopy based studies of photochemically-modified poly(ethylene terephthalate) surfaces T. Bahners, K. Opwis, E. Schollmeyer, S.-L. Gao and E. Mäder
127
Wool surface modification and its influence on related functional properties P. Jovancic, R. Molina, E. Bertran, D. Jocic, M. R. Julia and P. Erra
139
Surface modification of polyethylene by photosulfonation S. Temmel, W. Kern and T. Luxbacher
157
vi
Contents
Covalent coupling of fluorophores to polymer surface-bonded functional groups R. Mix, K. Hoffmann, U. Resch-Genger, R. Decker and J. F. Friedrich
171
Functionalization of fiber surfaces by thin layers of chitosan and related carbohydrate biopolymers and their antimicrobial activity D. Knittel and E. Schollmeyer
193
Dendrons for surface modification of polymeric materials H.-J. Buschmann and E. Schollmeyer
209
Surface modification of textile materials by dip-coating with magnetic nanoparticles J. Zorjanović, R. Zimehl, E. Schollmeyer, O. Petracic, W. Kleemann, D. Knittel, T. Textor and U. Schloßer
219
Part 2. Adhesion Improvement to Polymer Surfaces Adhesion improvement of epoxy resist to a benzocyclobutene layer: Application to nanoimprint lithography B. Viallet and E. Daran
231
Improvement of adhesion between poly(tetrafluoroethylene) and poly(ethylene terephthalate) films G. Bayram, G. Ozkoc and P. Kurkcu
241
Novel approaches to enhance adhesion of cellulose E. Delgado, J. A. Velásquez, G. G. Allan, A. Andrade, H. Contreras, H. Regla, L. R. Bravo and G. Toriz
251
An in-mold application of adhesion promoters to polyolefin substrates T. Schuman, M. Singh and J. Stoffer
263
Design and control of surface properties of UV-curable acrylic systems R. Bongiovanni and A. Priola
285
Detection of contaminants on polymer surfaces using laser-induced breakdown spectroscopy (LIBS) S. Markus, U. Meyer, R. Wilken, S. Dieckhoff and O.-D. Hennemann
297
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. vii–viii Ed. K.L. Mittal © VSP 2007
Preface This book chronicles the proceedings of the Fifth International Symposium on Polymer Surface Modification: Relevance to Adhesion held under the auspices of MST Conferences, LLC in Toronto, Canada, June 20–22, 2005. The premier symposium on this topic was held in Las Vegas, Nevada, November 3–5, 1993, the proceedings of which were properly chronicled [1]. The second symposium in this series was held under the aegis of MST Conferences, LLC in Newark, NJ, May 24–26, 1999, which was also documented in a proceedings book [2]. Apropos, it should be recorded that the third symposium in this vein was organized also by MST Conferences, LLC in Newark, NJ, May 21–23, 2001 but, for a variety of reasons, the proceedings of this event were not documented in the form of a hard-bound book, The fourth symposium on this topic was also organized by MST Conferences, LLC in Orlando, FL, June 9–11, 2003 the proceedings of which were documented in a hard-bound book [3]. The topic of polymer surface modification is of tremendous contemporary interest and even a casual look at the literature will attest that there is a brisk R&D activity in this arena. This high tempo of activity and interest emanates from the applications of polymeric materials for a legion of purposes in many and diverse technologies and industries. And the surface behavior (e.g., adhesion, wettability, tribological characteristics, etc.) of polymeric materials is of crucial importance. By suitably modifying polymer surfaces one can obtain the desired surface characteristics without tempering with the bulk properties. Concomitantly, there is much current interest in devising new ways or ameliorating the existing techniques. The techniques for polymer surface modification range from dry to wet, vacuum to non-vacuum, sumptuous to inexpensive, and sophisticated to simple. Apropos, recently much interest has been evinced in the atmospheric pressure plasma treatment as it offers certain advantages vis-a-vis the conventional lowpressure plasmas. The technical program for this event comprised 46 papers reflecting both overviews as well as original research contributions. The presenters hailed from academia, industry and other research organizations from many corners of the globe. The presentations focussed on various surface treatment methods, analysis and characterization of modified surfaces, understanding the life and durability of treatment methods, and relevance of surface modification in adhesion aspects of polymers. Now turning to this volume, it contains a total of 18 papers, others are not included for a variety of reasons, which were rigorously peer reviewed, revised
viii
Preface
(some twice or thrice) and edited. So it should be recorded that this book is not a mere collection of papers – which is normally the case with many proceedings volumes – rather it represents the highest standard of publication. The book is divided into two parts: Part 1. Surface Modification Techniques; and Part 2. Adhesion Improvement to Polymer Surfaces. The topics covered include: critical assessment of process conditions in polymer surface modification; various dry techniques (e.g., laser, ozone, low-pressure plasma, and atmospheric pressure plasma) to modify polymer surfaces; polymer surface modification by wet chemical techniques (e.g., photosulfonation, grafting, use of chitosan, and use of dendrons); wool surface modification, antimicrobial activity of modified fiber surfaces; AFM study of modified surfaces; relevance of adhesion in nanoimprint lithography; adhesion between polymer films; adhesion of cellulose; adhesion promoters for polyolefin substrates; surface properties of acrylic systems; and detection of contaminants on polymer surfaces by laser induced breakdown spectroscopy (LIBS). This volume and its predecessors [1–3] contain bountiful information and reflect the latest R&D activity relative to this fascinating and tremendously technologically important arena. Also it is hoped that the information contained here will serve as a fountainhead for new ideas in this field. Anyone with current interest or anticipated need to learn about polymer surface modification should find this book very relevant and of much value. Acknowledgements First, as always, it is a pleasure to express my thanks to my colleague and friend, Dr. Robert H. Lacombe, for taking care of the necessary details during the organizational phase of this symposium. Second, thanks are extended to all the contributors to this book for their interest, enthusiasm, patience and cooperation without which the book would not have seen the light of day. The reviewers provided much valuable comments which definitely improved the quality of manuscripts, and they should be thanked for their time and efforts. In closing, my appreciation goes to the staff of VSP/Brill (publisher) for transforming the raw material (manuscripts) into this book form. K. L. Mittal P.O. Box 1280 Hopewell Jct., NY 12533 1. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion. VSP, Utrecht, The Netherlands (1996). 2. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 2. VSP, Utrecht, The Netherlands (2000). 3. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 3. VSP, Utrecht, The Netherlands (2004).
Part 1 Surface Modification Techniques
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 3–24 Ed. K.L. Mittal © VSP 2007
Importance of process conditions in polymer surface modification: A critical assessment JEREMY GRACE,∗ H. KENT ZHUANG and LOUIS GERENSER Eastman Kodak Company, 1999 Lake Avenue, Rochester, NY 14650-2022, USA
Abstract—Plasma web treatment is a common practice for promoting adhesion, wettability and other surface or interfacial properties in the conversion industry. While the objective of creating new surface functional groups is conceptually simple, it can be difficult to choose the most appropriate kind and configuration of plasma source, the most appropriate feed gas composition and the most appropriate operating pressure for a given application. Such difficulties arise from the variety of species that can be formed in the plasma and the variety of possible plasma-surface interactions that can occur. A brief review of the importance of various plasma parameters (e.g., specific energy, species concentrations, and energy distributions) and an example relating nitrogen uptake in poly(ethylene-2, 6-naphthalate) to plasma diagnostic data in a low-radiofrequency capacitivelycoupled nitrogen plasma are presented. The importance of driving frequency and treatment configuration is discussed in detail. Uptake kinetics for samples treated at floating potential at low radiofrequency is compared with that for samples treated in the cathode sheath. Analysis of the treatment kinetics is based on a simple model of surface saturation. This approach can be used not only to compare practical treatment results as a function of process conditions, but also to compare different treatment techniques in a practical manner. Keywords: Polymer surface modification; plasma; capacitively-coupled discharge; process conditions; frequency effects; kinetics.
1. INTRODUCTION
Plasmas are used in a variety of polymer surface modification applications, including adhesion promotion in metallized plastics, wettability control in printing, priming of plastics and elastomers for painting or bonding, and treating catheters and other biomedical devices. Plasma polymerization, plasma treatment and plasma etching of polymers have been the subject of research for several decades, providing review articles and collections of work devoted to plasmas and polymers [1–9]. While the references cited here are by no means exhaustive, they provide a good sense of the broad range of applications, the variety of plasma chemi∗
To whom correspondence should be addressed. Tel.: (1-585) 722-0064; Fax: (1-585) 588-0710; e-mail:
[email protected]
4
J. Grace et al.
cal processes that can be employed, and the complexity of the plasma–polymer interaction. Yasuda [1] focuses on the effects of plasmas on the surface chemistry of polymers and discusses the variety of mechanisms of surface chemical modifcation by plasmas. Liston et al. [2] review plasma surface modification of polymers and plasma treatment techniques with a focus on adhesion promotion. Grace and Gerenser [4] review industrial applications and experimental studies of plasma modification of polymers, and then focus on the attempt to connect plasma characteristics with the chemical modifications they produce on polymers. Biederman and Osada [5] provide background on plasma technology and plasma chemistry and discuss applications of plasma polymerized films. The reference edited by d’Agostino [3] provides background on plasma source technology, plasma diagnostic techniques and applications of plasmas to deposition, surface modification and etching of polymers. A broad variety of modification techniques, applications and analytical techniques is presented in the collections edited by Mittal [6–9]. For even a narrow range of applications, one can find a variety of plasma sources used in practice. For example, in plasma treatment of polymer webs, industrialists and academic researchers have employed capacitively coupled lowpressure discharges driven at frequencies ranging from approx. 10 kHz to approx. 20 MHz, microwave discharges, dual frequency discharges and dielectric barrier discharges at atmospheric pressure. Furthermore, a variety of approaches can be found for generating the same chemical functionality on a polymer surface. The diversity of applications, plasma source technologies and gas chemistries employed arises largely because of the rich variety of physical and chemical processes that can occur in even the simplest plasma. In general, the polymer surfaces to be treated are placed for a specified time in contact with a plasma formed in a particular working gas at a particular pressure and flow rate. The practical dose is considered to be the applied power multiplied by the treatment time. In batch processes, one may consider the power per unit volume in the treatment zone, or one may project the volume upon the surfaces to be treated and consider the power per unit area. In these cases, energy per volume or energy per area is used as a practical treatment dose. In the case of plasma web treatment, the treatment device has some length along the direction of web motion and is at least as wide as the desired width of web to be treated. The treatment dose is found by dividing the power delivered to the treatment device by the device width and the web speed. In none of these cases does the treatment dose represent the actual energy (for example, in the form of ion kinetic energy or chemical potential energy) delivered to the surface of the article to be treated. Nonetheless, the degree of modification or practical effect of modification as a function of dose (i.e., the dose response) can be used to compare a variety of plasma chemistries and plasma sources, given a particular configuration of treatment device and sample (e.g., elongated sources applied across moving webs). Furthermore, comparing dose responses by varying the treatment time at selected settings of pressure and power can be instrumental in finding the best treatment
Importance of process conditions in polymer surface modification
5
conditions, and it can also be helpful in improving one’s understanding of the underlying mechanisms. The effects of the plasma modification are generally confined to the surface region of the treated articles. The ions, electrons and neutral species do not penetrate very far (typically approx. 10 nm), while the vacuum ultraviolet photons may penetrate micrometers deep, depending on the optical absorption properties of the article being treated. Therefore, the progression of surface chemical changes with treatment time follows a saturation curve. After the near-surface region (i.e., approx. 1–2 nm) becomes heavily modified, modification of the subsurface region can occur by diffusion of reactive species from the treated region above. Depending on the depth of analysis, one may observe simple saturation of a surface response with exposure time, or one may observe a region of rapid change with treatment time (reaction-limited regime), followed by a region of significantly slower change (diffusion-limited regime). While the general description of plasma treatment and the resultant saturation of surface modification with treatment dose are extremely simplified, they provide a useful framework in which to compare different polymer substrates, different sources, different plasma compositions, or different plasma-substrate configurations. At the core of all of these comparisons are the interrelationships among plasma process parameters, treatment configuration parameters and the species distributions that ultimately produce the chemical modification of the substrate surface. A brief discussion of these interrelationships is presented below. In Section 2, as an illustration of the importance of process conditions and treatment configuration, data are presented for treatments of polyester webs using a capacitively coupled radiofrequency nitrogen plasma. The effects of driving frequency, substrate location, power, and pressure are described and discussed. 1.1. Plasma parameters Plasma surface modification is a consequence of a variety of plasma species distributions impinging on a substrate surface. The process conditions determine the species concentrations and energy distributions. Applied power, as mentioned above, is a critical parameter with respect to the practical treatment dose. Process pressure and gas composition determine reactant concentrations. In addition, total pressure and gas partial pressures influence the energy distributions and concentrations of ions and electrons. The placement of the sample with respect to electrodes influences the relative species concentrations and their energy distributions experienced by the substrate. It is important to note that the external process parameters (i.e., applied power, pressure, gas flow and geometry) do not uniquely influence a particular species concentration and energy distribution. Of more fundamental significance than the applied power is the power divided by the gas flow through the treatment zone, as it is related to the specific energy (i.e., energy per molecule) deposited in the plasma. For example, a power of 1 W dissipated in a mass flow of 1 sccm amounts to 15.5 eV per molecule. In plasma
J. Grace et al.
Deposited Thickness in 10 min (nm)
6
60 50 40 30 20 10 0 0
0.5
1 Off -Time (ms)
1.5
2
Figure 1. Thickness of deposited fluoropolymer in a pulsed 13.56 MHz capacitively coupled CHF3 discharge as a function of time between pulses (“off-time”). The discharge pulses were 0.1 ms in duration.
polymerization processes, this parameter has been shown to be of considerable importance [10]. Additional parameters of importance in plasma polymerization using pulsed radiofrequency plasmas are the pulse period and duty cycle. Pulsing the plasma effectively reduces the time-averaged specific energy deposited relative to constant wave (CW) operation at the same driving voltage amplitude. In addition, it has important transient effects. Pulsing the power to the plasma in sufficiently short bursts makes use of higher electron temperatures during the ignition phase of the plasma (measurements and models for pulsed argon plasmas have been presented by Ashida et al. for an inductively coupled plasma source [11] and by Booth et al. for a capacitively coupled plasma source [12]). Furthermore, with sufficient delay between pulses, relatively long-lived excited neutral species may interact with each other and the substrate in the absence of ions during the period after the plasma is extinguished [13, 14]. An example of the effect of duty cycle on plasma polymerization rate in a capacitively coupled 13.56 MHz discharge in CHF3 is shown in Fig. 1. In this example, the sample (a silicon wafer) was placed on a 7.6-cm diameter electrode facing a second electrode of equal area on the opposite side of a Teflon® cylinder enclosure of 7.6 cm inner diameter and 7.6 cm height. Gas was admitted through a series of holes in the wall of the Teflon cylinder and exited through additional holes in the cylinder wall at a flow of 3 sccm and a pressure of roughly 13 Pa. A 13.56 MHz power generator was used in pulsed mode. The pulse duration (“on-
Importance of process conditions in polymer surface modification
7
time”) was 0.1 ms, and the time between pulses (“off-time”) was varied as indicated in the graph. As shown in Fig. 1, the maximum deposition rate does not correspond to the maximum power delivered (i.e., zero “off-time”, 90 W constant wave, as measured at the output of the tuning network). As the “off-time” is increased, the deposition rate increases until a maximum point beyond which it then decreases. In this latter regime, presumably the neutral species responsible for the condensation and polymerization are present in diminishing concentrations as the “off-time” is increased. Similar results have been reported by others [13, 14]. While the ratio of power to gas flow is clearly an important parameter and one of far more fundamental significance than applied power, it gives no indication of how the applied power is partitioned among the various processes occurring within the plasma. In a typical low-temperature nonequilibrium plasma, the applied power couples more effectively to the electrons in the plasma. This energized population of electrons generates ions and excited neutrals by electron impact processes. The types of species formed are determined by the electron impact cross-sections for excitation. Typical electron energy distribution functions have average energies in the range 1–5 eV, with high-energy tails extending into the range of several tens of eV. The electrons in the high-energy tail are important for driving ionization and other excitation processes having thresholds of tens of eV. In discharges sustained by secondary electron emission from the cathodes, a significant population of energetic electrons is generated by acceleration of the secondary electrons in the cathode sheath [15, 16]. These electrons can gain much of the applied voltage amplitude upon traversing the sheath. Ions traveling from the bulk plasma to surfaces acquire kinetic energy as they traverse the sheaths associated with these surfaces. For a surface floating electrically in the plasma, the ions gain the difference between the plasma potential and the floating potential. This difference scales with the electron temperature and is typically in the range 10–20 V. In contrast, ions entering a cathode sheath gain the difference between plasma potential and the cathode voltage. Depending on the relative areas of cathode and anode, the maximum energy gained by the ions as they approach the cathode can be 1–2 times the driving voltage amplitude, with time-averaged energies of 0.5–1 times the driving voltage amplitude. At low process pressures, the ions do not experience a significant number of collisions in transit to the cathode. At high process pressures, the energy with which the ions arrive is reduced significantly by collisions with neutral gas species. Neutral reactive species can be extremely important in polymer surface modification processes. For example, monatomic species formed by dissociation of molecular gases can react with polymer surfaces to form new functional groups comprising the atomic species from the plasma and atomic constituents of the polymer repeat unit. In addition, electronically excited atoms and molecules can be important participants in surface reactions leading to the formation of new chemical functionalities. The total pressure and partial pressure of feedstock gases
J. Grace et al.
(#/cm3)
8
Figure 2. (A) Plasma density and (B) density of “hot” electrons in the high-energy tail of the electron energy distribution, as determined from ion flux probe measurements in a nitrogen plasma driven at 40 kHz using the coplanar electrode configuration (see Fig. 6B). The bulk electron density (i.e., those having the lower temperature population in the bi-Maxwellian distribution) is roughly equivalent to the ion density (i.e., the plasma density), as the electrons in the high-energy tail are of much lower concentration.
influence the types and concentrations of atomic species formed by electron impact dissociation and excited atomic and molecular species formed by electron impact excitation. Furthermore, metastable species formed by electron impact
9
Relative N atom flux (arb)
Importance of process conditions in polymer surface modification
Figure 3. Relative N atom flux as a function of applied power at nitrogen pressures of 10 Pa ({), 20 Pa () and 30 Pa (z). This quantity is found from optical emission measurements by multiplying the ratio of emission intensity from N to that from N2 by the nitrogen pressure and normalizing this product of pressure and emission intensity ratio to the maximum value for all experimental runs.
(e.g., metastable states of argon or helium) can excite neutral atoms and molecules upon collision to form ions and excited neutral species, and the total system pressure and the partial pressures of the feedstock gases influence the population of products formed by these processes. A consequence of the presence of excited molecular and atomic species is the emission of photons as the excited states decay. Many of the decay processes include transitions involving the emission of vacuum ultraviolet (VUV) photons. VUV emissions can play a significant role in photochemical reactions during polymer surface modification, as well as in cross-linking reactions at the surface and in the subsurface region of the polymer. As an example of the importance of ions and reactive neutral species in the polymer surface modification process, ion densities and relative N atom fluxes (respectively from ion flux probe and optical emission data) from a 40 kHz capacitively coupled nitrogen plasma are presented in Figs 2 and 3, and their correlation with surface nitrogen uptake in poly(ethylene-2, 6-naphthalate) (PEN) is shown in Fig. 4. The plasma source, having a coplanar electrode configuration, is described elsewhere [16, 17]. The optical emission and ion flux data are discussed
J. Grace et al.
10
in Ref. [4]. Relative N atom concentrations were found by taking the ratio of Natom emission to molecular nitrogen emission and multiplying by the nitrogen pressure (i.e., the total discharge pressure), assuming that the nitrogen was only weakly dissociated. The relative N atom flux (which is taken as equivalent to the relative N atom concentration, as the two quantities are proportional, flux ∝ concentration ¥ velocity, and the effects of temperature on velocity are neglected) was normalized to a value of 1 at the maximum value of the product of emission intensity ratio and nitrogen pressure obtained in the experiment. The relative ion flux was found by taking the relative ion saturation current (at a given probe voltage significantly below floating potential) or by fitting the probe current in the ion saturation regime to a simple bi-Maxwellian model for the electrons and using the resultant ion concentration (i.e., plasma density shown in Fig. 2) and the Bohm velocity from the lower electron temperature and normalizing to a maximum value of 1. Nitrogen uptake was determined by XPS analysis of PEN samples at a series of web speeds for each of five combinations of treatment pressure and applied power [17]. A simple model of nitrogen uptake was employed (see Ref. [4] for details): dN dt
= ( AΓ i + BΓ N + C Γ N Γ i ) N max ,
(1)
t =0
where N is the surface nitrogen content, Γi and ΓN are the respective relative fluxes of ions and N atoms, Nmax is the maximum number density of available surface sites for nitrogen incorporation (estimated to correspond to 30 at%), and A, B and C are respective fitting parameters related to the cross-sections for nitrogen incorporation by molecular nitrogen ion impact, atomic neutral nitrogen impact, and incorporation by an interactive process involving ions and neutrals. The value for Nmax is a rough estimate and is expected to be similar for both PEN and PET. It is obtained by considering the nonvolatile species produced by adding N atoms to the repeat unit in the polyester. In Fig. 4, the modeled nitrogen uptake rate at t = 0 is plotted against the experimentally determined nitrogen uptake rate at t = 0 (as determined from analysis of nitrogen uptake curves (%N vs. t) for each of the five conditions). Using all three processes (i.e., coefficients A, B and C), a good fit is obtained. The quality of fit is degraded significantly by omitting the coefficient C or by using only A, B, or C alone (see Ref. [4]). The interaction term CΓNΓi may represent a two-step process, such as formation of surface radicals by ion impact, followed by reaction with atomic neutral nitrogen, or it may represent formation of atomic ions of nitrogen by electron impact processes (the ion flux is directly related to the electron concentration in the plasma). External plasma parameters, such as power, pressure and gas flow, can be measured and controlled with little difficulty. Unfortunately, these parameters are seldom related in a simple fashion to the fluxes and energy distributions of the species most important for surface modification. The relevant species fluxes and
Importance of process conditions in polymer surface modification
11
Figure 4. Modeled initial uptake rate vs. experimentally determined uptake rate for surface nitrogen in plasma-treated PEN using a 40 kHz capacitively coupled nitrogen discharge (shown schematically in Fig. 6B). The model used is given by equation (1).
energy distributions, however, can be difficult to measure. Nonetheless, using simple kinetic models, species fluxes and energy distributions can be related to surface modification effects, such as initial uptake rates of chemical species, saturation values for such species, or even distributions of chemical functionalities formed on a treated surface. An excellent example of work relating plasma physical characteristics, gas-phase chemistry and polymer surface chemistry to resultant surface modification is the investigation of the modification of polypropylene in dielectric barrier discharges in air by Dorai and Kushner [18]. 2. AN EXAMPLE OF THE IMPORTANCE OF DRIVING FREQUENCY AND TREATMENT CONFIGURATION: CAPACITIVELY-COUPLED RADIOFREQUENCY NITROGEN PLASMA TREATMENT OF POLYESTER WEBS
The example presented below illustrates how considerably different results can be obtained using the same type of treatment apparatus and varying the driving frequency and position of the article to be treated. Other researchers have noted and demonstrated the importance of such differences for steady-state treatments (relatively long exposure times) [19]. In this example, we examine, in addition, the up-
12
J. Grace et al.
take kinetics in different treatment configurations and observe significant differences in the distribution of chemical species formed for samples floating electrically in the plasma, as compared to samples located in the cathode sheath of a capacitively coupled low-radiofrequency nitrogen discharge. 2.1. Experimental details: plasma treatment The experimental configuration for treatments of stationary samples is shown in Fig. 5. A pair of water-cooled aluminum electrodes (16.5 cm ¥ 5.1 cm ¥ 1.27 cm thick) were housed in a grounded aluminum enclosure (labeled “Support and Side Walls”) and spaced 0.32 cm from the walls and each other. The pair of electrodes was placed on 0.32-cm-thick ceramic spacers (not shown), which, in turn, rested on the aluminum backing plate (approx. 17.1 cm ¥ 11.1 cm ¥ 1.27 cm). The enclosure sidewalls were 1.27 cm thick, providing support for the electrodes and backing plate and extending roughly 2.5 cm above the front surface of the coplanar electrodes. Poly(ethylene-2, 6-naphthalate) samples (100 µm thick) from Teijin were cut and placed on the electrode assembly and on the upper edge of the grounded enclosure. The grounded enclosure was installed in a cryopumped chamber (volume about 250 l) and was pumped to a base pressure below 3 ¥ 10-5 Torr. Nitrogen gas was admitted into the chamber at a flow between 80 and 100 sccm. A gate valve
Figure 5. Schematic of apparatus used for treatments of stationary samples.
Importance of process conditions in polymer surface modification
13
between the chamber and the cryopump was throttled until the steady-state chamber pressure reached 0.1 Torr. Note that the gas flow in the enclosure was unknown and considerably less than that in the chamber. Comparison with results in a pilot-scale treater with gas flow admitted directly into the electrode gap suggests that the effect of flow is negligible for the 40 kHz nitrogen treatments. After purging the chamber at steady-state flow for 3 min, power was applied to the electrodes at a specified level for a specified treatment time. Samples in positions A, B, and C were evaluated for three driving frequencies – 40 kHz, 450 kHz and 13.6 MHz. The output of the 40 kHz supply was floating and was applied directly across the two electrodes. The two higher frequency supplies had output referenced to ground; and thus the ground connection indicated in Fig. 5 was
A
B
Figure 6. Schematic of pilot-scale apparatuses used for treatments of moving webs: (A) rotating electrode configuration for treatments in the cathode sheath, (B) co-planar electrode configuration for treatments at floating potential.
14
J. Grace et al.
made. For each power supply configuration, treatment power and time were selected so that the sample in position B (electrically floating) had roughly the same nitrogen content (approx. 10 at%). For initial treatments of moving webs in the cathode sheath, the electrode assembly shown in Fig. 5 was mounted on an enclosure that was bolted to an aluminum frame with a simple web drive having a friction clutch on the unwind spindle. Web was introduced through slits between the electrode fixture and the enclosure so that the rear surface of the web was in direct contact with the electrode pair. Treatment gas was admitted to the enclosure through a series of orifices in the enclosure walls. For pilot-scale web treatments, plasma treatment devices were built and installed into an existing vacuum web coater. Schematics for the plasma source for treatments in the cathode sheath and at floating potential are respectively shown in Fig. 6A and 6B. The treatments were carried out at 40 kHz. For the treatments at floating potential, the web passed through slits in the sidewalls of the enclosure at a distance of roughly 3.3 cm from a pair of 7.6 cm ¥ 35.6 cm ¥ 1.2 cm co-planar electrodes. Treatment gas was admitted through a series of orifices fed by a manifold along a sidewall of the enclosure, producing treatment pressures from 10 to 30 Pa at nitrogen flows of 230 to 915 sccm. For treatments in the cathode sheath, the web was admitted to the treatment zone through rectangular ducts in an enclosure and conveyed over a rotating stainless steel electrode having a diameter of 12.7 cm and a length of 33 cm. A grounded counter electrode having a radius of curvature of 8.9 cm was positioned concentric with the rotating electrode to form a gap of 2.54 cm. Treatment gas was admitted through a series of orifices fed by a manifold in the ground electrode assembly, producing nitrogen pressures of 13 to 82 Pa at nitrogen flows of roughly 200–1200 sccm. An aluminum dark-space shield was machined to have a slightly larger radius of curvature than the rotating electrode and was placed roughly 0.3 cm above it. The sides of the shield along the web path were spaced from the enclosure walls to form ducts in series with the entrance and exit ducts. 2.2. Surface analysis Surface chemical changes in the PEN samples were assessed using X-ray photoelectron spectroscopy (XPS). Analyses were performed using a Physical Electronics 5601 photoelectron spectrometer with monochromatic Al Kα X-rays (1486.6 eV). The X-ray source was operated with a 2-mm filament at 350 W. Charge neutralization for these insulating polymers was accomplished by flooding the sample surface with low-energy electrons from an electron gun mounted nearly perpendicular to the sample surface (an emission current ≤25 mA and a bias voltage ≤0.5 eV were used). The pressure in the spectrometer during analysis was typically 3 ¥ 10-9 Torr. For the high-resolution spectra, the analyzer was operated at a pass energy of 11.75 eV. Under these conditions, the full width at half maximum (FWHM) for the individual components of the C1s, peak in an untreated
Importance of process conditions in polymer surface modification
15
PEN sample varied from 0.8 to 0.9 eV. All spectra were referenced to the C1s peak for the aromatic carbon atoms in the polyester repeat unit, which was assigned a value of 284.6 eV. Spectra were taken at a 45° electron take-off angle, which corresponds to an analysis depth of approx. 5 nm. XPS metrics found to be helpful in assessing changes for various nitrogen treatments are the incorporated nitrogen and the oxygen loss. In addition, the degree of change in the O1s ester doublet and changes in the chemical environment of the incorporated nitrogen were assessed. The details of the C1s spectrum provided additional information concerning the nature of the surface groups formed by plasma treatment. Detailed surface analyses of nitrogen-plasma-treated PEN with the web at floating potential are presented in Ref. [20]. For the purposes of comparing treatment configurations, quantities derived from XPS core-level spectra are tabulated and plotted below.
% Change (driven-floating)/ average
2.2.1. Results: treatments of stationary samples at various driving frequencies The run conditions and results for nitrogen plasma treatments at various driving frequencies for stationary PEN samples are listed in Table 1. U denotes untreated PEN, and the letters in the numbered runs correspond to the sample locations in Fig. 5. In the column headings, ν denotes driving frequency, %N denotes the nitrogen content from XPS, N1s centroid denotes the binding energy at the center of
60 40 20 0 -20 -40 -60 - 80 -100 10
100
1,000
10,000 100,000
Frequency (kHz) Figure 7. Differences (as a percentage of average value) in nitrogen content (•), oxygen content (□) and ester rearrangement (∆) as a function of driving frequency for nitrogen plasma-treated PEN samples on the driven (smaller) electrode and at floating potential.
J. Grace et al.
16
the N1s peak, %O denotes the oxygen content from XPS and Ester rearrangement denotes the degree of rearrangement of the ester portion of the PEN repeat unit (ester rearrangement, as judged from changes in the oxygen doublet of the O1s spectrum is described in Ref. [20]). From the results shown in Table 1, it appears that the samples in position A (i.e., located on the electrode driven by the power supply) have more incorporated nitrogen, lower N1s binding energy centroids and more oxygen loss than samples located in the other two positions. In general, the samples in position C (i.e., on the grounded coplanar electrode) have the least incorporated nitrogen and least oxygen loss. The differences seen between samples located in position A and position B are plotted in Fig. 7 as a function of driving frequency (except for the binding energy centroid). The graph illustrates that the differences diminish with increasing driving frequency. Table 1. Run conditions and XPS results for nitrogen-plasma treatments of stationary PEN samples Run*
ν (kHz)
Power (W)
Time (s)
U
N/A
N/A
N/A
%N
N1s centroid (eV)
%O
Ester rearrangement
0
N/A
22.2
0
1B 1A
40 40
100 100
10 10
10.4 16.8
399.9 399.0
22.9 9.7
5 8
2B 2A 2C
450 450 450
100 100 100
15 15 15
9.3 15.0 6.4
399.8 399.2 399.5
21.1 11.8 21.5
5 8 4
3B 3A 3C
13600 13600 13600
10 10 10
10 10 10
8.4 11.2 8.6
399.7 399.5 399.8
19.9 16.5 20.3
4 5 4
4B 4A 4C
13600 13600 13600
40 40 40
10 10 10
15.1 16.2 13.2
399.8 399.5 399.7
18.2 13.8 19.3
5 5 5
5B 5A 5C
13600 13600 13600
100 100 100
10 10 10
17.5 18.6 16.3
399.6 399.5 399.7
16.6 13.8 17.5
6 6 6
*
U denotes untreated sample; A, B and C refer to positions indicated in Fig. 5.
Importance of process conditions in polymer surface modification
17
2.2.2. Results: treatments of moving webs The simple web treatment device using the co-planar electrode assembly (as described above) was used to obtain dose–response curves at various pressures for samples treated in the cathode sheath. These results could be compared with existing data from pilot-scale treatments of PEN at floating potential. In both cases, various combinations of power and web speed were run at three different pressures.
Figure 8. Nitrogen uptake for PEN webs as a function of treatment dose: (A) treatments in the cathode sheath (using a simple web treatment device) and (B) treatments at floating potential (using the pilot-scale apparatus shown in Fig. 6B). Nitrogen pressures in the treatment devices were 6.6 Pa (○), 13 Pa (●) and 20 Pa (●).
18
J. Grace et al.
Figure 9. N1s binding energy centroid as a function of treatment dose. Circles: treatments in the cathode sheath (using simple web treatment device). Squares: treatments at floating potential (using the pilot-scale apparatus shown in Fig. 6B). Nitrogen pressures in the treatment devices were 6.6 Pa (white symbols), 13 Pa (gray symbols) and 20 Pa (black symbols).
The treatment dose was taken to be the power per unit width of the treatment device (i.e., in the direction along the web width), divided by the web speed. The data for nitrogen uptake and N1s binding energy centroid as a function of dose are respectively shown in Figs 8 and 9. The nitrogen uptake for samples treated in the cathode sheath (Fig. 8A) is enhanced relative to that observed for samples treated at floating potential (Fig. 8B). Furthermore, the N1s binding energy centroid shifts to lower values with increasing dose for samples treated in the cathode sheath, in contrast to shifting to higher values with dose for samples treated at floating potential (see Fig. 9). Poly(ethylene terephthalate) (PET) and PEN have similar nitrogen uptake characteristics when treated at floating potential at 40 kHz using the treater depicted in Fig. 6B. In addition, PET shows the same effects as PEN when treated in the cathode sheath (Fig. 6A); nitrogen uptake is comparably enhanced and the N1s binding energy centroid shifts to lower energy with increased treatment dose. Nitrogen uptake curves as a function of treatment time for PET at fixed treatment power were obtained using the pilot-scale web treater depicted in Fig. 6A. Similar uptake curves had been previously obtained for PEN [17] using the pilotscale web treater depicted in Fig. 6B. Uptake curves obtained in this fashion can be fitted to simple kinetic models and allow one to assess the effects of treatment power and treatment pressure on the basic kinetic parameters. The nitrogen uptake curves for PET treated in the cathode sheath, and the previously obtained data for PEN treated at floating potential, are shown in Fig. 10.
Importance of process conditions in polymer surface modification
19
Figure 10. Nitrogen uptake curves obtained at fixed power and varying web speeds. (A) PET treated in the cathode sheath. Treatment conditions: 60 W/13 Pa (○), 600 W/20 Pa (▲), 330 W/48 Pa (♦), 60 W/82 Pa (□), 600 W/82 Pa (■). (B) PEN treated at floating potential. Treatment conditions: 60 W/10 Pa (○), 600 W/10 Pa (●), 330 W/20 Pa (▲), 60 W/30 Pa (□), 600 W/30 Pa (■).
2.3. Analysis Comparison of PEN treated in the cathode sheath of a capacitively-coupled 40 kHz nitrogen discharge with PEN treated at floating potential in the same discharge reveals enhanced nitrogen uptake, more nitrogen-containing species that have lower N1s binding energies, and increased oxygen loss for PEN treated in the cathode sheath. Furthermore, these differences diminish with increasing driving frequency.
J. Grace et al.
20
The binding energy shifts suggest a higher degree of amine formation for samples treated in the cathode sheath. N1s binding energies for amine functionalities are typically near 399.1 eV, whereas N1s binding energies for amide functionalities are near 399.9 eV. The enhancement in amine content is consistent with the enhanced ester rearrangement and oxygen loss. See the data in Fig. 7 comparing oxygen content for treatments at 40 kHz in floating and cathode sheath configurations. More detailed studies of the oxygen content as a function of nitrogen plasma treatment dose for moving webs of PEN show that oxygen loss increases with dose in the floating potential configuration [17]. Similar studies for 40 kHz treatments in the cathode sheath configuration, however, reveal a far more dramatic loss of oxygen, consistent with the data presented in Fig. 7. Loss of oxygen lowers the probability of incorporating nitrogen in the form of amide groups. Because of the relationship between sample position and sheath voltage, and because the driving voltage decreases with increasing driving frequency for comparable applied power (approx. 1000 V at 40 kHz and approx. 100 V at 13.56 MHz for the apparatus used in this work), these results strongly suggest that high sheath voltages are responsible for the differences. These high sheath voltages increase the energy with which ions strike the polymer surface. In addition, they may produce significant shifts in the local species distributions through electron impact processes (secondary electrons are produced at the cathode, in this case the polymer surface, and are accelerated in the high sheath fields). Energetic electrons are capable of producing neutral as well as ionized atomic nitrogen. Furthermore, molecular ions falling in the high-voltage, low-frequency sheaths can produce neutral and ionized atomic nitrogen by dissociative charge exchange collisions. The data shown in Fig. 10 can be fitted to a simple model of surface saturation [4], where the nitrogen content N of the polyester surface is given by:
N=
αΓ I N max − αΓ (1 − e ( αΓ I + yΓ L
I + yΓ L
)t
).
(2)
Here, ΓΙ and ΓL are respective fluxes of species that result in nitrogen incorporation and nitrogen loss, α is the effective incorporation probability, y is the effective loss yield, Nmax is the maximum possible nitrogen incorporation (based on the number of available sites for incorporation) and t is the treatment time. The terms αΓΙ and yΓL are effectively lumped kinetic terms that represent products of interaction probabilities and species fluxes integrated over the species distributions. Because the sampling depth of the XPS measurements is greater than the modification depth, the nitrogen uptake curves exhibit a diffusion-limited regime at long times. Nonetheless, the initial linear region of the uptake curves and average saturation values in the diffusion-limited regime can be used to approximate the respective t = 0 and long-time behavior of equation (2). From Fig. 10, it is apparent that the initial uptake rates tend to be somewhat higher, and the saturation values are considerably higher for treatments in the
Importance of process conditions in polymer surface modification
21
cathode sheath than for treatments at floating potential. Using a value of 30 at% for Nmax, the ratio αΓΙ / yΓL (i.e., the ratio of nitrogen uptake to nitrogen loss) can be determined from the initial slopes of the uptake curves (Fig. 10) and their average saturation values. The ratio αΓΙ / yΓL is plotted in Fig. 11 for the two treatment configurations. For the treatments in the floating configuration (Fig. 11A), the ratio αΓΙ / yΓL is in the range 0.5–0.8, with the highest value at the highest treatment pressure and power (30 Pa/600 W). In contrast, the incorporation/loss ratio for the treatments in the cathode sheath (Fig. 11B) increases from 2 to 3 with increasing power, except for the lowest pressures (13–20 Pa), where the ratio
Figure 11. Ratio of incorporation term to loss term as calculated from experimentally observed initial uptake rates and saturation values. (A) Nitrogen plasma treatments of PEN at floating potential at pressures of 10 Pa (◊), 20 Pa (■) and 30 Pa (▲). (B) Nitrogen plasma treatments of PET in the cathode sheath at pressures of 13 Pa (◊), 48 Pa (■) and 82 Pa (▲).
22
J. Grace et al.
drops from 2 to 1.6 as the power increases from 60 to 600 W. Hence, for both treatment configurations, higher pressures and higher powers shift the uptake kinetics in favor of incorporation processes over loss processes for the polyesters studied. The treatments in the cathode sheath, however, generally have a higher incorporation/loss ratio than those at floating potential. This result is somewhat surprising, as one might expect the increased ion energies to enhance the loss mechanisms considerably. The diagnostic data in Figs 2 and 3 (applicable to the treatments at floating potential) suggest that the role of applied power is to increase the plasma density and the concentration of dissociated nitrogen. These effects lead to higher fluxes of molecular ions and atomic nitrogen to the polymer surface. In contrast, increasing the pressure has only a small effect on plasma density, while it increases the concentration of neutral atomic nitrogen. The enhanced nitrogen uptake kinetics at high powers and pressures may then be understood as resulting from an increase in species fluxes and an increase in the ratio of the flux of atomic neutrals (and perhaps atomic ions) to the flux of molecular ions. As discussed in Section 1.1, the cause for the importance of the combination of ions and atomic neutrals may be two-step surface reactions (i.e., formation of a surface radical, followed by reaction with atomic nitrogen) or formation of atomic nitrogen ions by electron impact in the plasma. In the case of the treatments in the cathode sheath, additional processes occurring in the sheath itself may generate important species (for example, dissociative charge exchange between molecular nitrogen ions and neutrals to produce atomic nitrogen species). Implicit in this discussion is the assumption that atomic nitrogen (neutral, ionized, or some combination) will produce higher saturation values of incorporated nitrogen than will molecular ions of nitrogen alone. This assumption is consistent with observations of nitrogen incorporation in poly(methyl methacrylate) surfaces exposed to low-energy (< 10 eV) N+ and N2+ [21]. In that work, Gröning et al. found that atomic ions produced considerably more nitrogen incorporation than did molecular ions. It is also consistent with observations that polyester surfaces exposed to energetic (approx. 400–1000 eV) N2+ (from ion sources) exhibit low saturation values (approx. 2 at% or less) of incorporated nitrogen, as compared to saturation values obtained by exposure to nitrogen plasmas [17]. As molecular nitrogen ions must dissociate upon impact to produce nitrogen incorporation, their loss yields may be significantly larger than their incorporation probabilities. Atomic neutrals and atomic ions of nitrogen, by contrast, may both have incorporation probabilities comparable to their loss yields. It should be noted that in the case of the treatments at floating potential, increasing the nitrogen pressure significantly above 30 Pa results in a reduced degree of nitrogen uptake. In contrast, treatments in the cathode sheath exhibit this reduced uptake near pressures of 90 Pa. These results are consistent with a decrease in the ratio of the flux of atomic neutrals to the flux of molecular ions with increasing pressure, thereby lowering the effective ratio of incorporation to loss.
Importance of process conditions in polymer surface modification
23
The results are also consistent with reduced concentrations of atomic nitrogen from collision processes at higher pressures. The wider useful pressure range for treatments in the cathode sheath is consistent with the notion that the species responsible for surface modification are generated within or near the cathode sheath and cross the sheath to reach the polymer surface, whereas for treatments at floating potential, they must cross the bulk plasma to reach the polymer surface. The relative pressure ranges are consistent with the relative dimensions of the highvoltage cathode sheath and the plasma zone between the cathode and the web at floating potential. 3. CONCLUSIONS
Plasma process parameters and treatment configuration determine the concentrations and energy distributions of species impinging a polymer surface during the plasma-modification process. Geometrical aspects of the treatment configuration (such as placement of sample with respect to the dominant cathode) can have a considerable effect on the nature of polymer surface modification in capacitively coupled radiofrequency plasmas. For example, significant differences are observed between polyester samples treated in the cathode sheath and those treated at floating potential in low-radiofrequency capacitively-coupled nitrogen plasmas. In particular, the incorporation of nitrogen, loss of oxygen, and the degree of amine content relative to amide content are enhanced in the cathode sheath. These differences diminish with increased driving frequency, suggesting that sheath potential is an important factor. The relationships between the external plasma parameters and the species concentrations and distributions are generally quite complex. Nonetheless, simple analysis of the time dependence of surface modification can provide a basis for comparing different treatment configurations or different processes using the same treatment device. Simple analysis of nitrogen uptake using lumped kinetic terms for nitrogen uptake and loss suggests that surface reactions involving species distributions from the plasma and chemical groups or molecules in the polyester surface favor incorporation mechanisms over loss mechanisms in the cathode sheath of low-radiofrequency nitrogen plasmas. In contrast, for samples at floating potential, the loss mechanisms are favored over the incorporation mechanisms. Acknowledgements The authors gratefully acknowledge Dennis Freeman, David Hawke and Michael Heinsler for their assistance with plasma treatment and plasma diagnostic experiments.
24
J. Grace et al.
REFERENCES 1. H. Yasuda, Radiat. Phys. Chem. 9, 805–817 (1977). 2. E. M. Liston, L. Martinu and M. R. Wertheimer, J. Adhesion Sci. Technol. 7, 1091–1127 (1993). 3. R. d’Agostino (Ed.), Plasma Deposition, Treatment, and Etching of Polymers, Academic Press, New York, NY (1990). 4. J. M. Grace and L. J. Gerenser, J. Dispers. Sci. Technol. 24, 305–341 (2003). 5. H. Biederman and Y. Osada, Adv. Polym. Sci. 95, 57–109 (1990). 6. M. Strobel, C. S. Lyons and K. L. Mittal (Eds.), Plasma Surface Modification of Polymers: Relevance to Adhesion. VSP, Utrecht (1994). 7. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion. VSP, Utrecht (1996). 8. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 2. VSP, Utrecht (2000). 9. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 3. VSP, Utrecht (2004). 10. M. S. Silverstein, R. Chen and O. Kesler, Polym. Eng. Sci. 36, 2542–2548 (1996). 11. S. Ashida, M. R. Shim and M. A. Lieberman, J. Vac. Sci. Technol. A14, 391–397 (1996). 12. J. P. Booth, G. Cunge and N. Sadeghi, J. Appl. Phys. 82, 552–560 (1997). 13. S. Fraser and R. D. Short, J. Phys. Chem. B106, 5596–5603 (2002). 14. C. L. Rinsch, X. Chen, V. Panchalingam, R. C. Eberhart, J.-H. Wang and R. B. Timmons, Langmuir 12, 2995–3002 (1996). 15. M. Surendra, D. B. Graves and G. M. Jellum, Phys. Rev. A41, 1112–1125 (1990). 16. S. Conti, P. I. Porshnev, A. Fridman, L. A. Kennedy, J. M. Grace, K. D. Sieber, D. R. Freeman and K. S. Robinson, Exp. Thermal Fluid Sci. 24, 79–91 (2001). 17. J. M. Grace, H. K. Zhuang, L. J. Gerenser and D. R Freeman, J. Vac. Sci. Technol. A21, 37–46 (2003). 18. R. Dorai and M. J. Kushner, J. Phys. D: Appl. Phys. 36, 666–685 (2003). 19. Y. Khairallah, F. Arefi, J. Amouroux, D. Leonard and P. Bertrand, J. Adhesion Sci. Technol. 8, 363–381 (1994). 20. L. J. Gerenser, J. M. Grace, G. Apai and P. M. Thompson, Surface Interface Anal. 29, 12–22 (2000). 21. P. Gröning, O. M. Küttel, M. Collaud-Coen, G. Dietler and L. Schlapbach, Appl. Surface Sci. 89, 83–91 (1995).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 25–32 Ed. K.L. Mittal © VSP 2007
Pretreatment and surface modification of polymers via atmospheric-pressure plasma jet treatment U. LOMMATZSCH,* M. NOESKE, J. DEGENHARDT, T. WÜBBEN, S. STRUDTHOFF, G. ELLINGHORST and O.-D. HENNEMANN Fraunhofer Institute for Manufacturing Technology and Applied Materials Research, Wiener Str. 12, D-28359 Bremen, Germany
Abstract—A novel atmospheric pressure plasma jet, that is operated with air, is used for the pretreatmet of different polymers. The resulting adhesive bond strengths and the corresponding changes of the polymer substrate surface are studied. The plasma treatment induces chemical and topographical changes on the polymer surface. It is likely that both types of surface modification contribute to the adhesion improvement. Results for poly(ethylene terephthalate) indicate that surface chemical composition is more influential in adhesion enhancement than surface roughness. Keywords: Atmospheric pressure plasma; pretreatment; polymers; adhesion.
1. INTRODUCTION
Adhesive bonding of most polymers requires a surface pretreatment (activation) to achieve sufficient adhesion between the polymer substrate surface and the adhesive. A variety of pretreatment methods exist [1, 2]. They can be classified according to their main operating principle as mechanical (e.g., blasting), physical (e.g., plasma, laser) or chemical (e.g., pickling) [3]. When choosing the optimal pretreatment method for industrial applications one has to consider a number of aspects. Some of these aspects are outlined in Table 1. In terms of high pretreatment speeds and compatibility with continuous production lines, atmospheric pressure plasmas (APPs) are especially advantageous. APPs that are used for the activation of polymer surfaces can be classified as coronas, plasma jets, or dielectric barrier discharges [4]. In this study a novel APP jet was used for the activation of six different polymers. The effectiveness of the pretreatment was studied by measuring the bond strength of pretreated samples. The surface modifications were investigated by surface analysis techniques. Studies on the use of a very similar plasma jet for the *
To whom correspondence should be addressed. Tel.: (49-421) 224-6456; Fax: (49-421) 224-6430; e-mail:
[email protected]
26
U. Lommatzsch et al.
Table 1. Some aspects to consider when choosing a pretreatment method Compatibility with production line (e.g. treatment speed and available space) Required level of activation Reliability of activation process Time stability of activation effect Health & safety Initial investment and running cost Ease of maintenance
surface treatment of polypropylene and metals have been reported by Green et al. [5] and Kim et al. [6], respectively. 2. EXPERIMENTAL
The polymers polyamide 6 (PA6), poly(ethylene terephthalate) (PET), highdensity polyethylene (HD-PE), poly(vinylidene fluoride) (PVDF), polypropylene (PP) and poly(phenylene sulfide) (PPS) were commercial-grade obtained from Rocholl (Aglasterhausen, Germany). A plasma jet system from Plasmatreat (Steinhagen, Germany) was used for the pretreatment. The plasma is generated inside the nozzle and expelled through an orifice onto the substrate surface (Fig. 1) [7]. The jet was operated with dry air at an input pressure of 5.9 bar and a flow rate of 17 l/min. The excitation frequency for the plasma was between 17 and 22 kHz, with pulse peak heights of approximately 5 kV. The samples were moved on a translation stage through the plasma jet at a distance from the nozzle exit of 10 mm (5 mm for PA6 and PPS, and 3 mm for HD-PE and PP) at a speed of 50 m/min (100 m/min for PET and PVDF). Only one treatment cycle was applied (except for PVDF, where 10 treatment cycles were applied). The gas temperature of the plasma jet outside the nozzle is approximately between 1300 K and 350 K, depending on the distance to the nozzle exit. Still, polymers can be treated without significant melting/degradation processes (see below) because the treatment time is on the order of a few milliseconds only. Substrates used were standard lap-shear samples with dimensions of 100 ¥ 24.8 ¥ 4 mm. An overlap of 12.5 mm along the long axis was used and the thickness of the adhesive layer was 0.1 mm. The adhesive layer thickness was controlled by using calibrated spacers. The samples were cleaned by ultrasonic rinsing in iso-propanol for 30 s before use. Lap shear strengths were measured at a pull rate of 5 mm/min at room temperature according to DIN EN 1465. Failure modes were determined visually. A commercially available two-component polyurethane adhesive was used and applied manually. The bonded samples were allowed to cure for 7 days at room temperature. Surface energies were determined from contact angle measurements
Modification of polymers with an atmospheric-pressure plasma jet
27
Figure 1. Schematic of the plasma jet used. The diameter of the discharge chamber is approx. 20 mm with a length of approx. 100 mm.
with water, glycerol, and diiodomethane as probe liquids (dynamic mode, advancing angle) using the approach of Owens and Wendt, and Kaelble [8, 9]. Surface topography and surface composition were determined by atomic force microscopy (AFM) and X-ray photoelectron spectroscopy (XPS), respectively. The AFM was operated in the tapping mode with a resonance frequency of the sensor of around 250 kHz. XPS measurements were performed with a Kratos Ultra system with a monochromatic Al Ka X-ray source and a pass energy of 20 eV for detailed spectra. 3. RESULTS
3.1. Adhesion The polymers PA6, PET, HD-PE, PVDF, PP and PPS were activated by the APP jet, and after adhesive bonding the lap shear strengths were measured. Figure 2 shows the lap-shear strengths of the samples with and without APP jet treatment. The plasma treatment improves the bond strength for all polymers substantially. Noteworthy is the different failure mode for the plasma activated samples. Only substrate failure (SF) or cohesive failure within the adhesive (CF) are observed for the activated samples in contrast to an interfacial failure dominant for the untreated samples.
28
U. Lommatzsch et al.
Figure 2. Lap-shear strength of polymer samples with and without APP jet treatment bonded with a polyurethane adhesive. The non-treated samples show an interfacial failure mode, while treated samples show substrate failure (SF) or cohesive failure within the adhesive (CF).
3.2. Surface modification The plasma treatment alters both the chemical and topographical states of the substrate surface. The APP jet treatment increases the surface energies of all polymers considerably (Fig. 3). Surface energies range from 35 to 62 mJ/m2 after activation. The increase is mainly caused by the increase of the polar component of the surface energy, e.g., for the commercial-grade PP containing additives the polar component increases from 12 to 34 mJ/m2. A high substrate surface energy is a prerequisite for wetting of the substrate with the adhesive. The increase in surface energy is caused by the incorporation of oxygen-containing groups in the substrate surface by the plasma treatment. Figure 4 shows high-resolution XPS spectra of the C1s region of HD-PE before and after APP jet treatment. The amount of oxygen increases from 2.0 at% in the untreated sample (probably due to contaminants or additives) to 24.4 at%. The spectral features between 286 and 290 eV indicate that different oxygen-containing functional groups are formed, such as alkoxy, carbonyl and ester groups. For the other polymers similar changes in chemical composition were observed (e.g., for PET the surface oxygen concentration increased from 15.2 to 32.4 at%). It is likely that such chemical modifications of the substrate surface are the most important for adhesion enhancement. The new functional groups on the surface give rise to good adhesion (high bond strength) because covalent bonds can be formed between the adhesive and the
Modification of polymers with an atmospheric-pressure plasma jet
29
Figure 3. Surface energy of polymer samples with and without APP jet treatment.
Figure 4. High-resolution XPS spectra of the C1s region of HD-PE before and after APP jet treatment. Peak fitting and area integration in the survey spectra (not shown) shows a relative atomic concentration (hydrogen atoms neglected) of oxygen of 2.0 at% before treatment and 24.4 at% after treatment.
30
U. Lommatzsch et al.
Figure 5. AFM images of PET surface before (left) and after (right) APP jet treatment. The surface roughness (RMS) decreases from 81 to 26 nm due to the treatment.
substrate. The APP jet treatment alters also the topography of the surface as shown for PET in Fig. 5. The AFM images show that the relatively rough surface is smoothened by the plasma treatment. Note that, in spite of a smoother surface, the bond strength is enhanced considerably by the plasma treatment. The topographical change can arise from thermal effects of the plasma or from chemical reactions and sputtering. It should be noted that smoothening of the surface was not observed for PVDF where the surface roughness remained nearly unchanged. The other polymers were not investigated by AFM. The improvement in adhesion can be related to individual surface modifications (such as change in surface energy, chemical composition, or surface roughness) or a combination of these effects [10, 11]. Also, cleaning or cross-linking effects of the plasma treatment cannot be neglected. Since different surface modifications occur simultaneously in the plasma and cannot be separated from one another, it is difficult to discriminate which effects are the most important for good pretreatment results. Therefore, it is not possible to evaluate if the change in surface roughness/topography contributes to or interferes with the adhesion enhancement. However, the observations for PET indicate that the chemical modifications of the surface are more influential than surface topography for high bond strengths. 3.3. Hydrophobic recovery The maximum time between activation and the subsequent bonding step with sufficient adhesion is of great technical importance. Due to molecular processes
Modification of polymers with an atmospheric-pressure plasma jet
5
2
Surface Energy [mJ/m ]
50
40
4
30
3
20
2
10
1
0
Lap Shear Strength [MPa]
Surface Energy Lap Shear Strength
31
0 untreated
1h
1d
7d
30d
Time after Treatment Figure 6. Surface energy and lap shear strength of PP as a function of time after APP jet treatment. A slightly different set of treatment parameters was used than for Figs 2 and 3 (nozzle–substrate distance: 10 mm; sample speed: 100 m/min). Between activation and bonding the samples were stored under ambient conditions.
occuring at the surface, pretreated samples lose their activation effect with time. This effect is called ageing or hydrophobic recovery [12]. Figure 6 shows the surface energies and bond strengths of PP at different times after plasma treatment. Both surface energy and bond strength decline only slowly with increasing time after pretreatment. After 30 days, the bond strength is about 68% of the initial level of 3.7 MPa of the sample bonded directly after pretreatment. For the other polymers studied here, similar observations were made. This long stability of the activation effect is an important advantage of APP jet treatment over corona treatment. 4. CONCLUSIONS
The polymers PP, HD-PE, PVDF, PA6, PPS and PET were activated by a commercially available plasma jet at atmospheric pressure. The plasma activation increases substantially the adhesive bond strength for all polymers studied. The improvements in adhesion are correlated to the plasma-induced changes on the surface. Surface analysis shows incorporation of 10–24 at% of oxygen into the surface as well as topographical changes of the surface but without an increase of surface roughness.
32
U. Lommatzsch et al.
REFERENCES 1. K.L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 2. VSP, Utrecht (2000). 2. K.L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 3. VSP, Utrecht (2004). 3. A. Kruse, G. Krueger, A. Baalmann and O.-D. Hennemann, in: Polymer Surface Modification: Relevance to Adhesion, K.L. Mittal (Ed.), pp. 291–302. VSP, Utrecht (1996). 4. A. Schütze, J.Y. Jeong, S.E. Babayan, J. Park, G.S. Selwyn and R.F. Hicks, IEEE Trans. Plasma Sci. 26, 1685 (1998). 5. M.D. Green, F.J. Guild and R.D. Adams, Int. J. Adhesion Adhesives 22, 81 (2002). 6. M.C. Kim, S.H. Yang, J.-H. Boo and J.G. Han, Surface Coatings Technol. 174–175, 839 (2003). 7. Agrodyn Hochspannungstechnik, German Patent No. DE 2,9919,142 (2001). 8. D.K. Owens and R.C. Wendt, J. Appl. Polym. Sci. 13, 1741 (1969). 9. D.H. Kaelble, J. Adhesion 2, 50 (1970). 10. M. Noeske, J. Degenhardt, S. Strudthoff and U. Lommatzsch, Int. J. Adhesion Adhesives 24, 171 (2004). 11. M.R. Wertheimer, L. Martinu, J.E. Klemberg-Sapieha and G. Czerenuszkin, in: Adhesion Promotion Techniques: Technological Applications, K.L. Mittal and A. Pizzi (Eds.), pp. 139–173. Marcel Dekker, New York, NY (1999). 12. F. Garbassi, M. Morra and E. Occhiello, Polymer Surfaces: from Physics to Technology. Wiley, Chichester (1998).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 33–86 Ed. K.L. Mittal © VSP 2007
The effects of excimer laser irradiation on surface morphology development in stretched poly(ethylene terephthalate), poly(butylene terephthalate) and polystyrene films JONGDAE KIM, D. U. AHN and EROL SANCAKTAR* Department of Polymer Engineering, The University of Akron, Akron, OH 44325-0301, USA
Abstract—Polystyrene (PS), poly(ethylene terephthalate) (PET) and poly(butylene terephthalate) (PBT) surfaces were irradiated using an excimer laser (LPX 240i, Lambda Physik). Polymer films were made first by cast film processing and then stretched with a biaxial stretching machine. With excimer laser treatment of polymer surfaces, it was found that 1–2-µm size structures could be produced depending on material properties and film processing conditions. Materials with lower UV absorption coefficient produced double-digit micrometer-size structures, while those with higher UV absorption coefficients produced single-digit micrometer-size structures. In all these cases the structures formed only on stretched films. In addition to these microstructure developments, the determination of ablation threshold fluence was of interest mainly for understanding the fundamentals of ablation behavior and technological applications. To understand the ablation phenomenon, and how microstructures could be developed during ablation, different material processing and excimer laser conditions were chosen for experimentation. From the observation of microstructures, we found that the initial microstructure development was a key for the following microstructures. We proposed a mechanism for this initial structure development based on polymer chain or melt movement caused by non-homogeneous ablation on the surface. Keywords: Excimer laser; non-homogeneous ablation; polystyrene, poly(ethylene terephthalate); poly(butylene terephthalate); stretched film; surface morphology; microstructure.
1. INTRODUCTION
During the last 50 years, plastics industries have grown at an enormous rate. Today, polymeric materials are very versatile and used in many fields including food, electronics, automotive, construction and other manufacturing industries. As polymer industries have grown, surface modification methods have attracted increased attention. The purpose of surface treatment usually is to modify the surface layer of polymer by incorporating certain functional group and/or inducing *
To whom correspondence should be addressed. Tel.: (1-330) 972-5508; e-mail:
[email protected]
34
J. Kim et al.
roughness on the surface to improve its wettability, printability, and its adhesion to other materials. One of the new technologies developed in this area during the last few decades is surface treatment using excimer lasers. This technology has many advantages compared to conventional surface treatment methods such as mechanical and/or chemical modifications of surfaces. Probably the most competitive advantage is that excimer laser treatments can achieve both chemical and physical modifications of surfaces at the same time. Excimer laser was first developed in 1975 and it is the most powerful of practical ultraviolet lasers. In 1982, Srinivasan and his colleagues discovered that the ultraviolet radiation induced removal of thin layers of organic polymers by using intense radiation from an excimer laser [1, 2]. This discovery stimulated scientific interest in the physics and chemistry of excimer laser ablation process on polymer surfaces, and experiments have been extended to many polymers using different excimer lasers by many researchers during the last 20 years [3–12]. Other reports on the application of the excimer laser irradiation on polymer surfaces relate to adhesion [13–15], annealing of silicon for thin-film transistor (TFT) manufacturing [16, 17], fiber dyeing [18–21], microlens fabrication [22–25], electroless metallization of polymers [26–28], etc. In addition to the studies of excimer laser ablation on polymer materials there were some studies on metal surfaces [29–31], especially on the surfaces of aluminum, titanium and copper to enhance their adhesion to other materials. It has been reported that excimer laser surface treatment of these metals could replace the hazardous process of chromic acid anodizing performed as a pretreatment for bonding or painting in aerospace industries [29]. One of the great advantages of excimer lasers treatment is its ability to treat surfaces only within the micrometer-scale without any damage to the bulk of the material. This is possible with most polymers due to the fact that they have high absorption coefficients for UV light and have low heat-conduction properties, thus extending excimer laser application to medical fields such as eye surgery, tattoo removal, etc. The process of excimer laser ablation on polymer surfaces is accompanied by plumes followed by explosive sound coming from the surface, while modifying the surface chemically and physically. The purpose of this research was to investigate the ablation characteristics and structure modification due to excimer laser irradiation, especially microstructure formation as a result of laser ablation, on polymer surfaces. This was accomplished by a fundamental study of ablation characteristics, polymer characteristics and their relationship. In order to investigate the ablation characteristics depending on different states of materials, we have applied different laser parameters, energy fluence, repetition rate and pulse number for excimer laser ablation of different polymer films. For this purpose PS, PET and PBT were chosen. These materials are known to be amorphous, slow-crystallizing and fast-crystallizing materials, respectively. With different processing conditions, we expect different states of orientation or crys-
Effects of excimer laser irradiation on surface morphology
35
tallinity for different materials, which will show different behaviors in excimer laser ablation. One of the important ablation characteristics in excimer laser irradiation process is the existence of ablation threshold. In this study, we have determined this value. To do this, different excimer laser fluences were applied to films and ablation characteristics were investigated. Some researchers have observed structures formation on stretched films or fibers [7, 32–34]. In order to determine the effect of stretching conditions on structure formation, we have produced films with different stretching conditions and have investigated structure formation after excimer laser irradiation. Some studies have reported that the origin of microstructure formation was due to the different etching ratios of crystalline and amorphous portions in polymers [9, 35]. Some others stated that the origins of microstructures were in chain or melt movements driven by internal/or external stress relaxations, made possible by heating and melting effects during laser irradiation [32, 36, 37]. Through this study we wished to understand the formation of microstructures due to laser ablation and, finally, to propose a mechanism of microstructure formation based on experimental data. 2. BACKGROUND
2.1. Excimer laser ablation of polymers Ever since the first discovery of the laser in 1960, powerful beams of light have been directed at solid materials for a variety of purposes. After developing the rare-gas–halide excimer lasers in the late 1970s, researchers discovered that these intense, pulsed UV-lasers could be used to machine precise patterns in organic or inorganic solid materials without assistance of a liquid or gaseous chemical ambient [38]. Even though it is not precisely defined, the removal of material from a solid without liquid or gaseous chemical ambient by intense laser radiation is commonly referred to as ablation. Ablation can lead to clean and precise material removal at an irradiated polymer surface. Dry etching of solid-state polymers has been important for lithographic [39] and micro-machining processes [40]. 2.1.1. Ablation threshold Ablation threshold can be defined as a level of energy required for breaking apart of solid materials by UV photon absorption. The ablation thresholds for variety of materials are shown in Table 1. According to the Beer’s law, the laser intensity transmitted at a distance x can be described as:
I ( x ) = I 0e −α x ,
(1)
where I 0 is the intensity of incident beam and α is the absorption coefficient (cm–1) at the laser wavelength used.
J. Kim et al.
36
Table 1. Ablation threshold and UV absorption coefficient for various polymers at 248 nm Material
Ablation threshold (mJ/cm2)
UV absorption coefficient, α (cm-1)
PET
36 [42], 22 [43], 44 ± 3 [44]
1.6E5 [43]
PBT
70 [26], 38 ± 12 [45]
6.7E4 [26]
PS
33 [43], 140 [26]
6.3E3 [43], 1.94E4 [26]
A general assumption made is that ablation can occur only when the intensity is higher than a threshold intensity level ( I th ) . If xth is the distance through which the threshold intensity can be transmitted, Beer’s law can be rewritten as:
I th = I 0e −α xth
(2)
From equation (2), the etch depth per pulse, xth is
xth =
I ln 0 α I th 1
(3)
Furzikov suggested the following relationship between threshold intensity and UV absorption coefficient [41]:
I th [α eff ]1/ 2 ≈ const,
(4)
where αeff is an effective UV absorption coefficient. 2.1.2. Characteristics of UV-laser ablation of polymers In case of excimer laser irradiation, the ablation phenomenon is related to two important polymer properties. One is the ability of the polymer material to absorb UV photons. This causes electronic excitation of polymer molecules in a very thin depth followed by ejection of materials or relaxation of electrons to the ground state. The second related property is thermal conductivity. In general, polymers have very poor ability to transfer heat. The correct combination of these two properties results in the ablation process to be confined only to the surface of a polymer and not cause any damage within the bulk. Most polymers strongly absorb a certain range of UV radiation. With this absorption electronic transition occurs from a ground to the excited state. The reactions that the electronically excited organic molecules undergo subsequently are varied and depend on the chemical structure of the molecule, the wavelength of UV light and the medium in which the reaction is carried out. In 1982, Srinivasan [2] reported that when a polymer was exposed to pulsed-UV laser radiation, the material at the surface was spontaneously etched away to a depth of several micrometers. This phenomenon was termed ablative photodecomposition (APD). It has been reported that the absorption of laser light in polymers is governed by
Effects of excimer laser irradiation on surface morphology
37
Table 2. Absorption coefficient and penetration depth for some polymers [49] PET
PS
PC
PI
16 ¥ 104
6.3 ¥ 103
10 ¥ 104
2.5 ¥ 105
62
1587
1000
40
Absorption coefficient (cm-1)
30 ¥ 104
8 ¥ 105
5.5 ¥ 105
4.25 ¥ 105
Penetration depth (nm)
34
12.5
18.2
23.5
KrF (248 nm) laser Absorption coefficient (cm-1) Penetration depth (nm) ArF (193 nm) laser
PC, polycarbonate; PI, polyimide.
Beer’s law and that the etch depth exhibits a logarithmic dependence on the incident laser fluence above a threshold value [46–48]. The most important parameter in excimer laser processing of polymers is their UV absorption coefficient, which depends on the polymer material and the wavelength of UV beams. For the same polymer material, absorption coefficient varies inversely with the wavelength. Consequently, the transmission intensity of UV beam into a polymer also depends on the nature of the polymer material, and the beam wavelength. Penetration depth is defined as the distance at which 65% of incident laser energy is absorbed. It is reciprocal of absorption coefficient, i.e,
xpenetration =
1
α
,
(5)
where α is the UV absorption coefficient. Table 2 provides some examples of penetration depth for different polymers at different wavelengths of excimer lasers. 2.2. Surface modification by excimer laser irradiation By using excimer lasers, one can incorporate functional groups, as well as induce roughness on polymer surfaces to improve their wettability, printability and adhesion to other materials. 2.2.1. Chemical modification The nature of surface functional groups generated on PET surfaces has been studied using X-ray photoelectron spectroscopy (XPS) subsequent to irradiation by excimer laser [50–53]. When the PET film was irradiated with ArF excimer laser (193 nm), the oxygen to carbon atomic ratio at the surfaces was found to decrease significantly. Lazare and Srinivasan [53] explained that the low oxygen content in irradiated surfaces was due to the loss of volatile molecules such as CO and CO2. They explained the formation of carboxylic acids, olefins, and alcohol groups
38
J. Kim et al.
Figure 1. UV degradation schemes for PET material.
from XPS study as shown in Fig. 1. In Fig. 1, schemes (a) and (c) were proposed by Marcotte et al. in 1967 [54] and scheme (b) was proposed by Day and Wiles in 1972 [55]. In 1989, Kokai et al. [51] characterized the PET surface using Infrared Spectroscopy subsequent to KrF laser ablation. Using the multiple internal reflection (MIR) technique, they found that the relative intensity of the absorption band at 725 cm-1 (C–H out-of plane bending vibration of aromatic ring) of irradiated PET film was greater than that of the original PET film. With this finding, they suggested that the surface after ablation was mainly composed of aromatic rings. They also argued that the surface structure of PET after excimer laser ablation was completely different from that after continuous irradiation by weak ultraviolet light, which results in the formation of carboxylic acid end-groups [56]. In 1999, Watanabe and Yamamoto [57] suggested the formation of both OH and CHO groups on PET surfaces after 248 nm KrF excimer laser irradiation. They explained that formation of the CHO group seems to be caused by photoreaction, whereas formation of the OH group appears to be due to a decomposition reaction from momentary heat at the time of laser irradiation. Chemical modification of PBT surface due to KrF excimer laser irradiation was also investigated by Wesner et al. in 1997 [45]. They found oxygen depletion and bond scission on surfaces near the threshold intensity, and decrease of O/C ratio at higher fluences, as were earlier reported for PET surfaces by Watanabe and Yamamoto [57], and Srinivasan and Lazare [58].
Effects of excimer laser irradiation on surface morphology
39
2.2.2. Physical modification In addition to chemical changes there have been reports of physical modifications, such as microstructure development on surfaces of oriented polymer films and fibers [33, 58–68] due to excimer laser irradiation. Such physical modification is important in the fields of micro-machining and micro-fabrication. Combined chemical and physical modification improves surface wettability and, consequently, adhesion. 2.2.2.1. Microstructure development. UV-laser-induced ablation of polymers is often accompanied by a modification of the surface morphology in the irradiated regions. The nature of the surface structures developed by irradiation depends on the polymer material and excimer laser variables. Srinivasan and Lazare [58] reported that the height of the microstructure developed on the polymer surfaces increased with the number of laser pulses, and a steady value occurred at roughly 10 pulses. Andrew et al. [9] suggested that such microstructures were due to spherulites created by the preferential etching of amorphous regions in polymers. They used XeCl excimer laser to ablate the PET but a more detailed investigation was not performed. The mechanism of microstructure development subsequent to excimer laser irradiation is not clear, but it is thought that several effects are simultaneously operative. Niino et al. [60] also reported a few micrometer-sized periodic patterns on PET surfaces irradiated by an excimer laser above the ablation threshold. 2.2.2.2. Laser-induced periodic (LIP) surface structures. In addition to the microstructures developed beyond the threshold as described above, sub-micrometer periodic structures were also observed below the threshold fluence [69–72]. For this purpose, Bolle and co-workers [69, 70] used polarized excimer laser beam with repetitive exposures up to about 1000 pulses, and produced periodic structures in the region of fluence far below the photoablation threshold. This structure is believed to be due to the interference of the incident wave with a surface scattered wave. Bolle and Lazare reported that polymers with an absorption coefficient smaller than 104 cm-1 could not produce sub-micrometer periodic structures [70]. 2.2.2.3. Microstructure development mechanism. The exact mechanism for microstructure development on polymer surfaces due to excimer laser irradiation is not clear. Andrew et al. [9] suggested that the origin of microstructure was in the different etch rates of crystalline and amorphous regions of polymers during irradiation. In 1988, Novis et al. [35] conducted a systematic SEM study of the structures on PET films subsequent to excimer laser irradiation at 193 nm. They also argued that the laser-induced surface structure was due to a difference in the etch rates of the amorphous and crystalline regions of the material rather than differences in their ablative photodecomposition (APD) threshold or thermally-induced stress relaxation of the material. They observed that just above the APD threshold, the etch rate of the amorphous polymer was about 60% larger than that for
40
J. Kim et al.
the semicrystalline PET. This difference in etch rates can be due to a difference in the reflectivity or absorbance of amorphous and crystalline PET. They suggested that UV laser etching was a quick and easy method for observing the crystalline subsurface structure of aromatic semicrystalline polymers at low fluence and low pulse repetition rates. Lazare et al. [62] also observed microstructures on PET films, and reported that the development of the structure was due to a combination of several factors. These factors include uneven etching due to sample inhomogeneity, Marangoni convection turbulence in the transient fluidized layer formed during the laser pulse, stress field in the polymer surface and modulation of the absorbed energy by surface scattered electromagnetic waves. They also reported that during excimer laser irradiation the surface temperature probably exceeds the melting point of the polymer or degraded polymer. In 1996, Hopp et al. [36] reported microstructure development on PET films as a result of ArF excimer laser irradiation. They suggested that some layers of polymers would have different temperatures due to varying laser beam intensities. The first layer of the polymer, where the absorbed fluence exceeds Fth, is ablated, i.e., explodes from the surface, while the second layer, where the absorbed fluence is below Fth, but more than the melting fluence Fm, is melted. Thus, in different depths the layers of the molten polymer have different temperatures at a given time. They also suggested that the molten layer in which the liquid begins to move and a material stream is developed could cause special structures. The surface structure may be the frozen remnants of these convective patterns. They pointed out that there might be other processes which could affect the formation of surface structures. For example, increasing internal friction due to cooling of the molten layer hinders mass convection. Moreover, if the liquid layer is very thin, then the surface tension may play an important role in the formation of the structures. The features of the surface structure also depend on the temperature gradient. The unit cells are re-formed after each laser shot until the height of the unit cells is smaller than the melting depth. After this, the developed pattern changes only slowly, shot by shot, due to ablation. They also reported that the average dimension and shape of microstructure unit cell depend on the laser fluence, the incident angle of the ablating laser beam and the number of the laser pulses. Watanabe and Yamamoto [73] observed small ripples on PET surfaces as the irradiation energy increased above the threshold level. They argued that in the fusion layer next to the ablation layer, the temperature is above the melting point of PET (320oC). After laser irradiation, this layer loses energy and is stabilized. At this moment, heat shrinkage of the polymer film occurs and the mode of stabilization differs at each point, depending on the local molecular orientation and thermal conditions, causing the appearance of ripples on the polymer surface. Bahners and Schollmeyer [32] observed roll-type structures on PET fiber surfaces, and these structures are rather irregular or random when one considers a single roll, but are highly regular on a larger scale, e.g., the mean roll-to-roll distances. They regarded this as typical for unstable chaotic systems. In their ex-
Effects of excimer laser irradiation on surface morphology
41
periments they treated several samples with the same draw rate but different laser fluences. After irradiation, the microstructures had almost the same dimensions indicating that the energy density deposited by the laser beam was not the driving force behind structure growth. They explained that the reason for microstructure development was the Marangoni convection due to the stress fields existing in the polymer intrinsically or induced externally. These stress fields are created because rather high drift velocities occur, especially at the free surface, thus leading to dramatic density variations along the surface. The corresponding density gradients would act as the origin for material convection. According to Knittel et al. [37], there are no requirements for the phenomenon of UV laser-induced surface structure, other than very high absorption characteristics for the polymer, resulting in a very small penetration depth for the energy deposited, together with a simultaneous existence of a tension field within the polymer. They also argued that there were no influences of fiber crystallinity and polymer chain alignment on the microstructure. They explained that surface structure formation was purely due to the thermal character of the laser irradiation. In addition to the absorption properties and temperature gradients, the simultaneous effect of stress release from the intrinsic (or external) tension plays an important role in surface structuring of fibers. The release of frozen-in tension within fibers is coupled with high temperature gradients, thus leading to selfarranged behavior of the movement of polymer chains. Using excimer laser wavelengths for which the materials show low absorption may cause fiber damage due to heat accumulation and thermal bulk degradation. Development of cone structures on polyimide (PI) surface has been reported [11, 74–76]. Krajnovich and Vazquez [77] argued that enrichment of the sample surface by carbon initiates cone formation by locally shifting the ablation threshold to higher values. In effect, the polymer surface becomes radiation hardened. It appears that radiation hardening, in the form of enrichment by carbon, plays an integral role in the ablation process, affecting not only the rate of ablation but also the dynamics and morphology of surfaces. Silvain et al. [78] reported that cone-like structure formation during KrF excimer laser ablation was associated with the diffusion and aggregation of micrometer and nanometer carbon particles. Growth of larger cone-like structure can then proceed to aggregates of smaller cone-like microstructures in thick molten layer upon repeated laser irradiation. They noted that carbon particles formed within the laser plume had a shielding effect towards laser ablation. Hopp et al. [75] suggested a simple etching model based on the diffraction and interference theories to explain the formation of conical structures. 2.3. Ablation mechanism The mechanism of excimer laser ablation on polymer surfaces is still not clear and difficult to define. In the field of polymer ablation, researchers [79–84] have argued for years whether mechanisms of photochemical or photothermal effects
42
J. Kim et al.
were mainly responsible for material removal, even though there are no exact definitions for these two terms. Krajnovich and Vazquez [77] distinguished these terms depending on how fast the electronic excitation by laser irradiation transformed into the form of vibration. Loosely speaking, advocates of photochemical ablation argue that electronic excitation leads to direct bond breakage and material ejection before the electronic to vibrational (E→V) transformation occurs. The assumption behind the photothermal mechanism, on the other hand, is that E→V transformation is fast compared to the rate of material removal, and that ablation proceeds the same way as it would if the same amount of energy were to be deposited in the sample in the form of vibrational rather than electronic energy. Many investigations have been performed in order to understand the mechanisms of photo-etching of polymers with excimer lasers. It was concluded that [85] the process of ablation was mainly photochemical in nature if ArF (193 nm) laser pulses were used, but that both photochemical and thermal mechanisms were operative at longer laser wavelengths (KrF, 248 nm; XeCl, 308 nm; XeF, 351 nm). There is a general agreement between the researchers that modification of the chemical structure occurs during laser ablation at an intensity lower than the threshold intensity. Lasers are employed at such intensity levels ( I < I th ) for heat treatments that do not alter the geometry of the workpiece. However, modification of chemical structure can occur. At higher intensities ( I > I th ), ablative processing such as drilling and cutting can be accomplished and laser-induced surface structures develop (Fig. 2).
Figure 2. Explanation of threshold intensity.
Effects of excimer laser irradiation on surface morphology
43
In 1994, Srinivasan [86] proposed hypothetical steps in the interaction of a UVlaser pulse with a polymer. During the interaction of a laser pulse with a polymer surface, a stream of photons from a single laser pulse impinging on a polymer surface is absorbed in a depth that is believed to be of the order of 10–30 µm for a weak absorber, 1–5 µm for a moderate absorber and <1 µm for a strong absorber. While the absorption coefficient that applies to such absorption process is not well established, it is accepted that the penetration of photons into the material will fall off as a logarithmic function. The absorption of the photons causes breakup of the polymer chains. The exact mechanism by which this happens is still under vigorous debate. The products of the decomposition of the polymer are ejected at the polymer surface at supersonic velocities. The ablation process is believed to be a volume explosion [87], which is likely to produce small molecules. In 1993, Srinivasan [88] showed a series of photographs of the ablation of poly(methyl methacrylate) (PMMA) by a KrF excimer laser. The speed of the ejected plume could be calculated approximately as above Mach 8. In 1998, Horiuchi et al. [89] photographed the plumes ejected from the surface of Fe using CCD photography. They reported that the plume velocity exceeded 4.5 km/s, corresponding to Mach 12. The plume expands so rapidly that it can shield the target surface from irradiation by a laser pulse. 3. MATERIALS AND EXPERIMENTAL PROCEDURES
3.1. Materials Poly(butylene terephthalate) (PBT, BASF Ultradur B 4520), poly(ethylene terephthalate) (PET, Eastpack 7352) and polystyrene (PS, Dow Styron 685D) were used in this study. PET and PBT pellets were dried in a vacuum oven for 24 h before being fed into an extruder for cast film processing. 3.2. Film casting Films were prepared using film-casting process equipment including a single screw extruder, a coat hanger die and a take-up device, as shown in Fig. 3. The take-up device consists of a chilled roll with a temperature controller and winding rolls. Materials were extruded above the melting temperature for PET and PBT (above 50oC), and above the glass transition temperature for PS (above 100oC). The polymer melts were transparent when they emerged from the die, but the PET and PBT melts turned hazy after cooling in air. By controlling the cooling rate of the cast film, optical properties could be controlled. Two methods were used for controlling the cooling rate of films. One was controlling the chilled roll temperature and the other was controlling the windup speed. A high windup speed produces a thin film which can be cooled more easily than a thick one. In this experiment, chilled roll temperature was maintained at 10oC during the process and
44
J. Kim et al.
Figure 3. Schematic diagram of cast film fabrication using an extruder with a coat-hanger die.
chilled roll was kept as close as possible to coat hanger die to achieve rapid quenching of materials. Screw and windup speeds were optimized at 28 and 150 rpm for PBT, and at 35 and 80 rpm for PET, respectively, to produce optically homogeneous transparent films. 3.3. Film stretching Cast film fabricated using an extruder with a “coat hanger die” was cut to dimensions of 12 cm by 12 cm and stretched using a biaxial stretching machine (model BIX-702, Iwamoto Seisakusho, Tokyo, Japan), which consists of a heating chamber and a stretching device. When films were stretched at the glass transition temperature or 5oC higher than that temperature, necking was observed for PET and PBT. To stretch the films without necking, temperatures were set to about 10oC higher than the glass-transition temperature of each film and after reaching the desired set temperature, samples were held in the heating chamber of the machine for 5 min to achieve temperature uniformity in the film. Films were stretched 100%, 200% and 300% at 0.3 mm/s. Stretched films were immediately cooled down to below the glass transition temperature by a cooling fan while the stretch levels were maintained with pneumatic clips in both directions. 3.4. Film characterization To investigate the relationship between material properties and ablation characteristics, mechanical, thermal and physical properties of stretched films were measured before performing excimer laser irradiation on film surfaces.
Effects of excimer laser irradiation on surface morphology
45
3.4.1. Orientation measurements Refractive indices of the films in the machine, transverse and thickness (normal) directions were measured using a Bellingham Stanley Abbé refractometer with a polarizing eyepiece. The light source used for these measurements was a monochromatic sodium lamp, which produced 589.6 nm wavelength. More than five measurements were recorded and averaged for measurement accuracy. Temperature was also recorded at each measurement for temperature correction. Refractive indices were determined using a conversion table and temperature correction factor provided by the manufacturer. Birefringence for stretched polymer films was determined using the refractive indices measured from Abbé refractometer. The birefringence value contains information on orientation and crystallinity for semi-crystalline polymers, and it directly indicates the orientation for amorphous polymers. For PBT, wide-angle X-ray scattering (WAXS) patterns were also obtained to determine the orientation of crystalline region. X-ray diffraction is the most important method for the determination of the crystallographic structure of crystalline polymers. WAXS flat film photographs were taken using a General Electric X-ray generator (GE-XRD6) equipped with a copper target tube. This equipment produced CuKα radiation at λ=1.54178 Å. Operating conditions were set at 30 kV and 30 mA. 1.2 mm ¥ 15 mm films were cut and stacked with a glue to prepare 1.2-mm-thick samples. Samples were mounted perpendicular to the X-ray beam and the sample to film distance was kept at 32.19 mm. For the construction of pole figures, a GE copper target X-ray generator with a quarter circle goniometer was used. Samples were mounted on a single crystal orienter with the machine direction as spindle axis. The azimuthal angle, χ was varied from 0o to 90o in 5o increments, and the meridional angle, φ, was varied from 0o to 360o using 10o intervals. The diffraction intensities collected by this method represent the densities of crystallographic poles of certain planes in all directions. 3.4.2. Thermal properties Differential scanning calorimetry (DSC, TA Instruments model 2920 MDSC V2.6A) was used to characterize thermal properties of materials. Crystallinity, glass transition temperature, and melting temperature were measured. A heating rate of 20oC/min and about 7 mg of sample were used. In order to calculate the percent crystallinity, Φc (%), of stretched films, the following equation was used:
φc (%) =
∆H exp ∆H 0
× 100,
(6)
where ∆H exp = ∆H melt − ∆H cryst and ∆H 0 is the heat of fusion of 100% crystalline polymer, reported to be 119.8 J/g for PET [90] and 142.0 J/g for PBT [91]. To determine the thermal stability of the materials, TA 2920 TGA instrument was used. In this experiment, an approx. 10 mg sample was used for each scan, and the temperature was raised up to 600oC at 20oC/min.
J. Kim et al.
46
Table 3. The gas pressures for the KrF (248 nm) excimer lasers
Halogen Rare Buffer Inert Total
Gas
Pressure (mbar)
5% F2/He Kr Ne He
80 130 2990 0 3200
3.5. Surface treatment For this research, an excimer laser (LPX 2401) by Lambda Physics was used. With this excimer laser, the energy, frequency and total pulse numbers can be controlled. By varying the gas in the laser tube, different laser wavelengths are also available. For the current research, Kr2 and F2 gases were filled in the laser tube to produce laser beam of 248 nm wavelength. The laser beam area produced was about 1 cm2 (2 cm ¥ 0.5 cm), and central regions showed more uniform energy distribution than edges. When the laser is operating with KrF (248 nm), the input voltage available is up to 26 kV, with which up to about 300 mJ output energy can be obtained. Up to 22 kV, the maximum frequency of the laser pulse is 400 Hz, and after 22 kV the maximum value is 10 Hz. The pulse duration is 25 ns. In the case of ArF (193 nm), the laser has 180 mJ maximum pulse energy, 300 Hz maximum frequency and 20 ns pulse duration. The laser was operated at different energies, total pulse numbers and frequency levels to obtain the relationship between laser operating variables and ablation characteristics for different polymers. The typical gas pressures for the KrF excimer laser are shown in Table 3. There are two methods to control the laser energy. One is constant high voltage mode and the other is constant energy mode. With constant high voltage, pulse energy decreases with time because the gases have limited life times. If the constant energy mode is selected, the controller adjusts the high voltage to achieve laser operation at the preset energy level. 3.6. Irradiated surface properties In order to investigate the effects of excimer laser irradiation on polymer surfaces, we observed surface property changes before and after laser irradiation, including surface morphology. 3.6.1. Surface morphology Surface morphology was observed using SEM (Scanning Electron Microscope, Hitachi S-2150) after polymer films were treated by excimer laser. In order to relate morphology development to excimer laser variables, films were treated with different total pulse numbers and fluence levels. Different repetition rates were
Effects of excimer laser irradiation on surface morphology
47
Figure 4. Experimental apparatus for ablation depth measurement.
also used to investigate the dependence of morphology on heat accumulation in the surface region. The samples were coated with a thin layer of conducting silver to facilitate this procedure. Surface morphology pictures were analyzed by an image analyzing software (Scion Images for Windows, Scion, Frederick, MD, USA). 3.7. Measurement of ablation characteristics During excimer laser irradiation on polymer surfaces, ablation phenomenon may be the most unique feature, which is manifested by weight loss and generation of a plume, which is ejected from the surface at high velocity with an explosive sound. 3.7.1. Ablation depth and weight measurements The ablation depth is a key parameter in determining the ablation threshold. By using a Hommel T500 profilometer, ablation depth was determined for different energies and pulse numbers. Laser frequency was set to 1 Hz to prevent heat accumulation between pulses and, as shown in Fig. 4, a screening mask was utilized to obtain uniform laser energy. The mass per unit area is linearly related to the ablation depth, and weight change due to ablation was also measured using a microbalance subsequent to excimer laser irradiation. To obtain more accurate values for small pulse numbers, more than 20 samples were irradiated with the same conditions and weighed together, and averaged. 4. MEASURED CHARACTERISTICS OF POLYMER FILMS
4.1. Orientation measurements Birefringence was measured for all polymer films and WAXS measurements were carried out only for PBT films, since PET films show weak X-ray scattering and PS films are amorphous.
48
J. Kim et al.
4.1.1. Birefringence measurements The main purpose of film stretching process is to induce chain orientation in the machine (stretching) direction. The film’s mechanical properties increase with increasing chain orientation. Birefringence measurements for PET stretched films at different temperatures are shown in Fig. 5. By increasing the stretching ratio, both ∆ n13 and ∆ n23 increased. Both birefringence values for PET films stretched at 85oC showed higher values compared with that of the film stretched at 95oC.
Figure 5. Birefringence (a) and refractive indices (b) of PET films with different stretching ratios (1, machine; 2, transverse; 3, normal directions).
Figure 6. Birefringence (a) and refractive indices (b) of PBT films with different stretching ratios (1, machine; 2, transverse; 3, normal directions).
Effects of excimer laser irradiation on surface morphology
49
When PET films were stretched at 95oC, birefringence values for 100%, 200% and 300% stretched films were not much different. In order to investigate the relation between orientation and ablation characteristics, stretching temperature was set to 85oC for PET, which is 10oC higher than its glass-transition temperature, and this condition was also applied for PBT and PS films. The birefringence and reflective indices for stretched PBT and PS films are shown in Figs 6 and 7, respectively. For the case of PET and PBT films, increasing the stretching ratio caused a rapid increase in nMD and a substantial decrease in nTD and nND, as shown in Figs 5 and 6. 300%-stretched PET film had refractive indices of nMD = 1.63, nTD = 1.57 and nND = 1.55, revealing optical anisotropy. 300%-stretched PBT film showed stronger optical anisotropy than PET film. The refractive index values were measured as nMD = 1.72, nTD = 1.56 and nND = 1.50. The differences in birefringence and refractive indices of polyester and PS films are summarized below: – PET and PBT films showed increases in birefringence, while PS-stretched films showed decreasing negative birefringence with increasing stretching ratio. Both increasing and decreasing negative birefringence values represent an increase in polymer chain orientation. – With increasing stretching ratio, nMD increased rapidly for PET and PBT films, while for PS films the increase was in nND. – With increasing stretching ratio, only nMD increased for PET and PBT films with the other two refractive indices decreasing, but for PS films all three refractive indices increased with increasing stretching ratio.
Figure 7. Birefringence (a) and refractive indices (b) of PS films with different stretching ratios (1, machine; 2, transverse; 3, normal directions).
50
J. Kim et al.
These differences can be explained by considering the molecular structure of PS having bulky phenyl side groups, which will align parallel to the surface. The refractive index of PS along the normal direction will be higher than that along the machine direction resulting in negative birefringences. 4.1.2. WAXS WAXS flat film photographs for stretched PBT films are shown in Fig. 8. Diffraction patterns from (0 1 0) and (1 0 0) planes are observed in the pictures of 100% (Fig. 8a) and 200% stretched films (Fig. 8b). With further increase in stretching ratio to 300%, diffraction patterns from (0 1 1) and (1 1 1) planes are also observed (Fig. 8c). The diffraction arc shapes became narrow with increasing stretching ratio, revealing more uniaxial orientation in polymer chains with stretching.
Figure 8. WAXS flat-film photographs for PBT films stretched at 75oC by (a) 100%, (b) 200% and (c) 300%.
Figure 9. 2θ (degree) scan patterns for stretched PBT films. The upper curves are for 100%stretched PBT films and the lower ones are for 200%-stretched PBT films (left, normal direction; right, transverse direction).
Effects of excimer laser irradiation on surface morphology
51
As mentioned earlier, PBT has a triclinic unit cell structure. The d-spacing for lattice planes can be calculated from equation (7), and can be determined from the WAXS flat film photographs from which diffraction angles for lattice planes can also be determined. Equation Section 4
1 1 = 2 d hkl (1 + 2 cos α cos β cos γ − cos 2 α − cos2 β − cos 2 γ ) sin 2 α k 2 sin 2 β l 2 sin 2 γ 2hk + + + (cos α cos β − cos γ ) a2 b2 c2 ab 2kl 2lh + (cos β cos γ − cos α ) + (cos γ cos α − cos β )) bc ca
(h
2
(7)
a, b and c are the three axis dimensions of crystalline lattice unit cell; α, β and γ are the angles between b and c, a and c, and a and b axes, respectively; h, k and l are the Miller indices representing the crystallographic plane with any positive or negative integers (or zero); and dhkl is the interplanar spacing for (hkl) crystallographic planes. Diffraction angles and d-spacings for (1 0 0) and (0 1 0) planes were measured as shown in Table 4. Diffraction angle values are in good agreement with the values observed from 2θ scanning which is shown in Fig. 9. Bragg spacings calculated from the triclinic unit cell (α-form) of Hall and Pass [92] were 5.14 Å and 3.822 Å, and the diffraction angles were 17.25o and 23.27o for (0 1 0) and (1 0 0) lattice planes, respectively. For the stretched PBT films, WAXS pole figures of (1 0 0) and (0 1 0) planes were constructed using X-ray analysis to understand the state of orientation more clearly (Fig. 10). With increasing stretching ratio, the poles of (1 0 0) and (0 1 0) planes tended to concentrate in film normal and transverse directions, respectively. 100%-stretched film had (1 0 0) planar orientation, and the poles of (1 0 0) plane concentrated in normal direction (ND), while the poles of (0 1 0) plane spread uniformly in the film plane. Significant concentrations of the (1 0 0) and (0 1 0) poles, corresponding to the ND and TD directions, respectively, were observed when the stretching ratio was increased to 200% and 300%. From the results of WAXS pole figures and flat film photographs, we conclude that increasing stretching ratio of PBT films will cause anisotropy in polymer crystallographic orientation. Table 4. Bragg spacing and diffraction angle for PBT lattice planes Lattice plane
d-Spacing (Å)
Diffraction angle (o)
(0 1 0) (1 0 0)
5.09 3.82
17.42 23.28
52
J. Kim et al.
Figure 10. WAXS pole figures for stretched PBT films, showing stereographic projections for the density of crystallographic poles of (010) and (100) planes. A pole point represents the intersection of the normal to a crystal plane with the surface of a sphere, having the crystal at its center. In this case, as the stretching ratio increases, the orientation of (010) crystal planes along the TD and that of (100) crystal planes along the ND increases, revealing that the orientation along the chain direction increases with the increase of stretching ratio.
Effects of excimer laser irradiation on surface morphology
53
Figure 11. White–Spruiell biaxial orientation factors for PBT films. Uniaxial orientation with respect to TD and MD gives the points on the TD and MD axes, respectively: Point (1,0) represents a perfect uniaxial orientation along the TD and point (0,1) represents a perfect uniaxial orientation along the MD. Biaxially equal orientation along the TD and MD gives the points on the line of TD = MD. Point (-1,-1) represents a perfect uniaxial orientation along the ND, and each side of the isosceles triangle represents a planar orientation. As the stretching ratio increases, the chain axis preferentially orients along the MD.
The orientation in the films was quantified using the White–Spruiell biaxial orientation factors [93]. For this calculation, PBT triclinic unit cell was approximated as pseudo-orthorhombic in which the c-axis is along the chain axis, the aaxis is parallel to the phenyl ring normal and the b-axis is taken to be orthogonal to the a–c plane. Using Wilchinsky’s method for pseudo-orthorhombic unit cells [93], we have:
( f12 − e12 )cos2 φi ,010 − ( f 22 − e22 )cos2 φi ,100 + e12 f 22 − e22 f12 cos φi ,c = e12 ( f 22 − e22 ) − e22 ( f12 − e12 ) 2
(8)
cos2 φi ,a =
f 22 cos2 φi ,100 − f12 cos2 φi ,010 , f 22 (e12 − f12 ) − e22 (e22 − f 22 )
where φi,x is the angle between the i laboratory axis (1 = machine direction and 2 = transverse direction) and x-crystallographic axis; φi,hkl is the angle between a diffracting hkl plane and the laboratory axis i; and e and f are geometric constants of unit cell (e1 = cos21.65o, e2 = cos82.68o, f1 = cos68.36o, and f2 = cos7.32o from the geometry of pseudo-orthorhombic unit cell).
J. Kim et al.
54
Using these constants and equation (8), we have
cos 2 φi ,c = 1 − 1.14126 cos 2 φi ,100 − 0.85874 cos 2 φi ,100 cos 2 φi , a = 1.16041 cos 2 φi ,100 − 0.16041 cos 2 φi ,100
(9)
The White–Spruiell biaxial orientation factors were determined from equations (9), which are represented in Fig. 11 for PBT films. Cast films revealed f i ,Bj ≈ 0 , representing no orientation. With increasing stretching ratio the orientation factors of the c and a-axis increased. 4.2. Thermal properties For analyzing ablation of polymer surfaces by excimer laser irradiation, researchers have used photochemical and photothermal mechanisms, which are related to electron excitation and the transformation of energy to the vibrational form, respectively. In order to study the thermal effect of excimer laser ablation, polymer thermal properties were measured using DSC and TGA scannings. DSC scans for PET films are presented in Fig. 12. With increasing stretching ratio, there is increasing glass-transition temperature and almost no change in crystalline melting temperature.
Figure 12. DSC scans for PET films stretched at 85oC.
Effects of excimer laser irradiation on surface morphology
55
As shown in Fig. 13, higher stretching ratio produces higher crystallinity (Xc) in PET films. When the PET film was stretched 300% at 85oC, crystallinity increased from 10% to almost 35% compared to the cast film.
Figure 13. Glass-transition temperature (Tg) (top graphs), cold crystallization temperature (Tc) (middle graphs) and the degree of crystallinity (Xc) (bottom graphs) for PET films as a function of stretching ratio (the stretching ratio numbers correspond to multiples of 100%).
56
J. Kim et al.
While glass-transition temperature and crystallinity increased with increasing stretching ratio, cold crystallization temperature (Tc) decreased, as stretching enhanced crystallization by creating favorable environment for crystallization by orientation and/or compaction of polymer chains. DSC scanning results for stretched PS films are presented in Fig. 14. The glasstransition temperature of PS increased from 98oC for cast film to 106oC at 300% stretching ratio. Highly-stretched films show higher orientation, and this makes polymer chain movement more difficult.
Figure 14. DSC scans for PS films stretched at 110oC.
Figure 15. DSC scans for PBT films stretched at 75oC.
Effects of excimer laser irradiation on surface morphology
Figure 16. TGA scans for cast and stretched (300%) films (from the top: PET, PBT and PS).
57
58
J. Kim et al.
DSC scans were also performed on PBT films, and the results are shown in Fig. 15. The DSC scan patterns show only little difference from those of PET in cold crystalline temperature peaks. As shown in Fig. 15, the peaks for cold crystallization temperature were not as clear due to the fact that PBT crystallizes faster than PET. Glass-transition temperature and crystallinity increased when PBT films were stretched. But crystalline melting temperatures were almost the same for all films. Figure 16 shows TGA scans for PET, PBT and PS cast and stretched films. There were no differences between the cast and stretched films of all materials. Since all films were stretched at about 10oC higher than their glass transition temperatures, there was no thermal degradation of stretched films during the stretching process. Thermal degradation onset points were measured by TGA scans. The values were 421oC, 390oC and 399oC for PET, PBT and PS, respectively. If we assume that excimer laser ablation occurs only by thermal degradation, PBT and PS materials should have higher ablation rates than PET. 5. ABLATION CHARATERISTICS OF POLYMER FILMS
Material ablation during laser irradiation depends on both material and laser parameters. Material parameters include UV absorption coefficient, density, reflectivity and thickness. Laser parameters include wavelength, pulse number, pulse duration, repetition rate and energy fluence. In this section, some experimental results will be presented to help understand the ablation phenomenon.
Figure 17. Typical profilometer scan for ablation depth measurements.
Effects of excimer laser irradiation on surface morphology
59
5.1. Ablation depth By using profilometry, ablation depths for various polymer films were measured. In order to avoid heat accumulation during excimer laser irradiation on polymers, the repetition rate was chosen as 1 Hz for most cases. A typical profilometer scan representing polymer surfaces after excimer laser irradiation is presented in Fig. 17. To measure ablation depth, a baseline was drawn between points a and b (Fig. 17). Ablation depth was measured by averaging the results obtained by subtracting every point between points c and d from the linear equation representing the line ab (Fig. 17). 5.1.1. Ablation depth and excimer laser parameters relationships In order to investigate the relationship between ablation rate and excimer laser parameters, ablation depths were measured with different excimer laser fluences, pulse numbers and repetition rates. 5.1.1.1. Ablation depth and laser fluence relationship. Ablation depths on PBT, PET and PS cast films were measured with excimer laser fluences varying from 100 mJ/cm2 to 1000 mJ/cm2, using 100 pulses at room temperature. For the case of polystyrene, laser fluences were varied from 300 mJ/cm2 to 1000 mJ/cm2, since the shallow ablation depths obtained with 100 mJ/cm2 and 200 mJ/cm2 fluences were not easy to measure. Ablation depths are presented in Fig. 18 with different fluences for various polymer films. These results reveal that ablation depth increased logarithmically with increasing laser fluence for all materials within 1000 mJ/cm2 fluence. Our experimental results confirmed that excimer laser ablation on PBT, PET and PS films followed Beer’s law within our experimental conditions. In other words, the slopes in Fig. 18 are related to UV absorption coefficients of the ablated materials.
Figure 18. Dependence of ablation depth on laser fluence.
60
J. Kim et al.
5.1.1.2. Ablation depth and laser pulse number relationship. The relationship between ablation depth and irradiated laser pulse number is presented in Fig. 19 for PBT, PET and PS cast films. Ablation depths were measured using a profilometer and the repetition rate was set to 1 Hz for all laser fluences. Experimental results
Figure 19. Relationship between ablation depth and number of excimer laser pulses with different laser fluences (from the top: PBT, PET and PS).
Effects of excimer laser irradiation on surface morphology
61
reveal a near-linear relationship between ablation depth and total pulse number for all materials when excimer laser fluence was above the threshold. 5.1.1.3. Ablation depth and laser repetition rate. Increasing laser repetition rate reduces cooling time between laser pulses and causes heat accumulation in surface region. This accumulated heat increases the temperature and complicates the surface temperature profile, thus making ablation depth measurement by profilometer difficult. Therefore, in order to investigate the relationship between ablation depth and laser repetition rate, the ablation weight was measured by a balance accurate to 0.00001 g. In this experiment PET and PS cast films were used. Since repetition rates of 1 Hz and 5 Hz show practically no difference in ablation weight, 5, 10, 25, 50, 80 and 100 Hz repetition rates were selected with different excimer laser conditions. In this experiment, irradiation fluences were selected as 150 and 200 mJ/cm2, and the total numbers of pulses were 200 and 400 for PET and PS, respectively. Experimental results are shown in Fig. 20. Figure 20 reveals that increasing repetition rate causes an increase in ablation weight for PS, but a decrease in ablation weight for PET. Increase in ablation weight for PS with increasing laser repetition rate can be explained by the heat accumulation effect due to the shortening of cooling time between the laser pulses. For PET, the ablation weight decreases abruptly when the repetition rate varies from 5 Hz to 10 Hz, and after that it decreases gradually or stays the same. Srinivasan and Braren [94] explained that such decrease was due to the shielding of the incoming pulse by the ablation products that cause a plume. In our experiments, the ablation weight was lower at 10 Hz repetition rate, which means that the period for each laser pulse was about 0.1 s. Since earlier studies [88] showed that plume could exist for only a few ms, Srivinasan and Braren’s explanation
Figure 20. Relationship between ablation weight and laser repetition rate for PS and PET.
62
J. Kim et al.
does not look plausible in our case. If the incoming excimer laser light can be affected by the existing surface plume or by the products of previous pulse, such plume or products should be active for more than 0.1 s. We could find alternative explanation of this behavior for excimer laser ablation in PET films. For this purpose, the surfaces were observed after excimer laser irradiation with high repetition rate for both PET and PS. While the PS surfaces showed clear or somewhat yellow color, the PET surfaces showed black color, which means that some carbonated materials were present on PET surfaces as created by excimer laser ablation. As shown in Table 1, the UV absorption coefficient for PET is much higher than that for PS. The material with a higher UV absorption coefficient can absorb more energy in the surface region than the material with a lower UV absorption coefficient. Furthermore, with higher excimer laser repetition rate, the heat will be accumulated more, causing much higher temperatures in the surface region. Consequently, some carbonaceous materials could be generated on PET surfaces as a result of thermal degradation or burning. These carbonaceous materials are known to resist excimer laser ablation more than common polymers [77]. Taking this finding into account, it can be concluded that the main reason for decreasing ablation weight with increasing repetition rate for PET is the creation of carbonaceous materials on the surface, which could shift the ablation threshold to higher values. 5.1.2. Ablation depth and material parameter relationship To investigate ablation depth for different material conditions, cast film samples were stretched or annealed to give different orientations and crystallinities before irradiating with an excimer laser.
Figure 21. Ablation depth difference between cast and 300%-stretched PET films.
Effects of excimer laser irradiation on surface morphology
63
5.1.2.1. Ablation depth and stretching ratio relationship. Ablation depths were measured for cast and 300% stretched PET films, and the results are shown in Fig. 21, which indicates that the ablation depth on cast films is higher than that on 300%-stretched PET films. As shown in Section 4, 300%-stretched PET films show higher orientation and crystallinity compared with cast film. With this kind of orientation and crystallinity, a stretched cast film can have much more compact chains and less free volume, which probably make the stretched film more resistant to laser ablation. 5.1.2.2. Ablation depth and crystallinity relationship. In 1988, Novis et al. [35] reported that when PET films were irradiated by ArF excimer laser (193 nm), the etch rate of the amorphous polymer was about 60% larger than that for the semicrystalline PET. They used Mylar films (DuPont) for their measurements. One sample was cast from the melt on quenching rolls, having 1-mm thickness and the other was biaxially stretched, having 12-µm thickness. Estimated crystallinities were 1% and 50% for amorphous and semicrystalline PET films, respectively. In our experiment, in order to exclude any other effect except that of crystallinity on ablation rate, cast films with the same thickness were annealed at different temperatures to induce crystallinity in the films. Annealing conditions were 125oC, 140oC, 160oC and 180oC for 10 min. After annealing the films, DSC scans were carried out to determine crystallinity of each sample. The average crystallinity of the annealed samples was measured as 37±2% for all annealed conditions. Sample films were irradiated at 208 mJ/cm2 of laser fluence using 100 pulses. Ablation depth measurement data are shown in Fig. 22. The annealed samples showed almost the same ablation depth within the error ranges, and non-annealed cast films showed higher ablation rate than annealed films, which confirms that the crystalline material is more resistant to excimer laser ablation.
Figure 22. Ablation depth for annealed PET films.
J. Kim et al.
64
If we consider that ablation depth is the sum of ablation from crystalline and amorphous regions, the following relation can be written: Xtotal = (Xc) xc + (1 - Xc) xa
(10)
where Xc, Xtotal, xc and xa represent % crystallinity, total ablation depth, ablation depth from crystalline region and ablation depth from amorphous region, respectively. From equation (10) and data of Fig. 22, the ablation depth for the amorphous region is calculated to be about 40% higher than that of the crystalline region. The difference in etch rate for crystalline and amorphous regions in PET films can be explained by the difference in their reflectivity and/or absorbance for laser light. Higher energetic stability of the crystalline region also supports this idea. 5.2. Ablation threshold The determinations of ablation threshold fluence, as well as etch rate are of interest mainly for understanding the fundamental ablation behavior, and also are relevant in applications. 5.2.1. Threshold fluence determination using ablation depth measurement Light transmission follows the well-known Beer’ law. If we consider that the ablation depth is directly related to a certain light intensity level, we can calculate ablation threshold fluence from Beer’s Law. For this purpose, equation (3) can be rewritten as:
xth =
1
α
ln( F0 ) −
1
α
ln( Fth )
where Fth is the ablation threshold fluence.
Figure 23. Ablation depth vs. ln(fluence) relations to calculate ablation threshold fluences.
(11)
Effects of excimer laser irradiation on surface morphology
65
By plotting ablation depths for various excimer laser incident fluences (Fig. 23), slope and intercept values can be calculated to determine the ablation threshold fluences. The slope in Fig. 23 represents the reciprocal of UV absorption coefficient or the penetration depth of excimer laser light. So, a steep slope indicates low UV absorption coefficient and high penetration depth. From Fig. 23 one can find the following relations:
α PET > α PBT > α PS χpenetration-PET < χpenetration-PBT < χpenetration-PS
(12)
The ablation threshold values calculated in this fashion are presented in Table 5. The experimental results showed good agreement with data from the literature, except for PS. Table 5. Ablation threshold fluences (in mJ/cm2) for various polymers Method
PBT
PET
PS
From measurement
34 ± 8
31 ± 7
195 ± 15
From literature
38 ± 12 [45]
36 [42]
140 [26]
Figure 24. Method for determination of ablation threshold fluence by measuring ablation weight; the ablation weight–laser fluence line for each film material is extrapolated to the laser fluence axis for this purpose.
66
J. Kim et al.
5.2.2. Threshold fluence determination using ablation weight measurements Ablation threshold fluence can also be determined from the measurements of ablation weight with various incident excimer laser fluences (Fig. 24). Using this method, ablation threshold fluences were determined as 34 ± 10, 42 ± 12 and 208 ± 18 mJ/cm2 for PBT, PET and PS, respectively. These results showed good agreement with the ablation threshold fluences which were determined using ablation depth. 5.3. Surface morphology development Surface microstructure development is one of the most unique results of excimer laser irradiation on polymer surfaces. However, not all polymers produce such
Figure 25. Different microstructures created by excimer laser irradiation. (a) Roll-type, (b) conetype, (c) ripple-type.
Figure 26. Photograph of irradiated PET surface with 20 pulses and different laser fluences (300% uniaxially-stretched PET film at 95oC).
Effects of excimer laser irradiation on surface morphology
67
microstructures on their surfaces. The creation of surface microstructures depends on the material and excimer laser parameters. Furthermore, the microstructure shapes are not identical in every material. They also depend on the processing conditions of the material, as well as laser irradiation conditions. Some different structures produced after excimer laser irradiations are presented in Fig. 25. They are called roll-, cone- and ripple-types in the literature. 5.3.1. Structures on PET and PBT surfaces After irradiating the surfaces of stretched PET films by an excimer laser, photographs were taken (Fig. 26) using a digital camera. In order to see the image more clearly, black background was chosen. Because PET films were transparent initially, black and grey colors in Fig. 26 represent transparent and white (or hazy), regions, respectively. In Fig. 26, the circular shape is the irradiated area. At low fluence, the area looks transparent but as the laser fluence increases, the irradiated area becomes hazy. The hazy color (gray color in Fig. 26) is due to the scattering effect between the visible light and the microstructures developed after laser ablation. When SEM pictures of the hazy circular region in Fig. 26 were taken, microstructures were observed. The appearance of the microstructures depends on polymer film preparation and excimer laser irradiation conditions. While we could find this kind of hazy surface after excimer laser irradiation on stretched films, the unstretched cast PET films had transparent surfaces, indicating the absence of microstructure development. If the films were stretched more, and/or the laser fluences were higher, we could produce hazy surfaces more easily, using less pulse numbers.
Figure 27. The dependence of microstructure type on mode of stretching. When the film was stretched uniaxially (left), we observed roll-type structures aligned perpendicular to the stretching direction, and when the film was stretched in both directions (biaxially), we observe ripple-type structures (right).
68
J. Kim et al.
When the microstructures were investigated by SEM, the peak-to-peak distances along the stretching direction were observed in the 3–5 µm range and the peak-to-valley depth was about 1–4 µm. As mentioned earlier, we could generate surface microstructures only on the surfaces of stretched films. And, it was found that the microstructures showed directionality, with alignment in the direction perpendicular to the stretching direction. If the film was stretched in one direction only, we observed roll-type structures aligned perpendicular to the stretching direction, and if the film was stretched in both directions (biaxially), we observed ripple-type structures (Fig. 27). Using different excimer laser variables, stretched PET and PBT films were irradiated at 248 nm. Laser fluences were set to 100, 150 and 200 mJ/cm2, pulse frequency was 1 Hz and the total numbers of pulses were 5, 10, 15 and 30. The microstructures developed as a result of irradiation were investigated by SEM. Silver coating was deposited on irradiated surfaces before taking SEM pictures. Irradiated PET and PBT surfaces showed similar behaviors in terms of excimer laser and material parameter dependency of microstructure growth. Increases in fluences and total number of pulses resulted in wider peak-to-peak distances (Figs 28 and 29). The SEM pictures of the irradiated surfaces for films with higher stretching ratios also showed wider peak-to-peak distances (Fig. 29).
Figure 28. Surface patterns on 200%-stretched PET film irradiated using different laser conditions (fluences from the top: 100 mJ/cm2, 150 mJ/cm2 and 200 mJ/cm2).
Effects of excimer laser irradiation on surface morphology
69
The peak-to-peak distance increased significantly with increasing film stretching ratio from 100% to 200%, but further increases in film stretching ratio did not affect this distance significantly. By using Scion image analyzing software, peakto-peak distances were measured (Tables 6 and 7). As shown in Table 6, the peak-to-peak distances increased about 250% when stretching ratio increased from 100% to 200%, but the distances were almost the same or slightly decreased when stretching ratio increased from 200% to 300%. This behavior was also observed when stretched films were irradiated with lower fluences, such as 100 and 150 mJ/cm2. Measurement results are presented in Fig. 29.
Figure 29. Surface patterns on stretched PET films after excimer laser irradiation with a fluence of 200 mJ/cm2 (from the top: 100%, 200% and 300% stretched PET films).
Table 6. Peak-to-peak distances (micrometer) for microstructures developed by excimer laser irradiation on PET films with different stretching ratios Stretching ratio (%) 100 200 300 Fluence= 200 mJ/cm2.
No. of pulses 5
10
15
30
– 2.03 ± 0.33 2.08 ± 0.21
– 3.46 ± 0.22 3.07 ± 0.36
1.65 ± 0.12 4.56 ± 0.31 4.25 ± 0.35
2.76 ± 0.24 6.87 ± 0.35 6.64 ± 0.29
J. Kim et al.
70
Table 7. Peak-to-peak distances (µm) for microstructures developed by excimer laser irradiation on PET films stretched 200% using different laser conditions Fluence (mJ/cm2) 100 150 200
No. of pulses 5
10
15
30
1.54 ± 0.12 1.98 ± 0.23 2.03 ± 0.33
2.84 ± 0.09 3.03 ± 0.25 3.46 ± 0.22
3.58 ± 0.33 4.14 ± 0.46 4.56 ± 0.31
5.07 ± 0.48 6.01 ± 0.32 6.87 ± 0.35
Table 8. Peak-to-peak distance (mm) measurements for microstructures developed by excimer laser irradiation on stretched PET films with different laser fluences Stretching ratio (%)
No. of pulses 5
10
15
30
– 1.54 ± 0.12 1.92 ± 0.23
– 2.84 ± 0.09 2.68 ± 0.32
1.33 ± 0.09 3.58 ± 0.33 3.44 ± 0.29
2.69 ± 0.28 5.07 ± 0.48 5.08 ± 0.27
– 1.98 ± 0.23 1.91 ± 0.18
– 3.03 ± 0.25 3.24 ± 0.28
1.55 ± 0.15 4.14 ± 0.46 3.71 ± 0.25
2.77 ± 0.26 6.01 ± 0.32 5.52 ± 0.24
2
Fluence= 100 mJ/cm 100 200 300 Fluence= 150 mJ/cm2 100 200 300
While the peak-to peak distances of microstructures developed by excimer laser irradiation on 200% and 300% stretched PET films showed similar results (Table 8), those on stretched PBT films showed clear tendency of increasing with increasing film stretching ratio (Table 9). As shown in Table 9, an increase of peak-to-peak distance is achieved more effectively by increasing the number of pulses using the same fluence rather than by increasing the excimer laser fluence using the same number of pulses. This behavior is similar to that observed with ablation depth measurement results discussed in Section 5.1. There is a non-linear relationship between the excimer laser fluence and the ablation depth, and a linear relationship between the number of pulses and the ablation depth. Considering the non-linear characteristic of excimer laser light absorption in polymer media, this result looks reasonable. Another important observation from this analysis is the dependence of microstructure dimensions on total number of pulses. As shown in Tables 6 and 9, an approximately linear relationship can be found between microstructure dimensions and number of pulses. In other words, the microstructures from previous excimer laser ablation pulses were not removed by subsequent pulses, but remained during the continuing ablation process, and grew further. More detailed evidence
Effects of excimer laser irradiation on surface morphology
71
Table 9. Peak-to-peak distance (mm) measurements for microstructures developed by excimer laser irradiation on stretched PBT films with different laser fluences Stretching ratio (%)
No. of pulses 5
10
15
30
Fluence= 100 mJ/cm2 100 200 300
– 2.76 ± 0.12 2.81 ± 0.23
2.25 ± 0.09 3.43 ± 0.09 3.99 ± 0.32
2.82 ± 0.09 3.91 ± 0.33 4.99 ± 0.29
3.67 ± 0.28 5.02 ± 0.48 6.42 ± 0.27
Fluence= 150 mJ/cm2 100 200 300
– 2.90 ± 0.12 3.25 ± 0.23
2.71 ± 0.09 3.73 ± 0.09 4.88 ± 0.32
3.42 ± 0.09 4.28 ± 0.33 5.35 ± 0.29
3.89 ± 0.28 5.98 ± 0.48 7.36 ± 0.27
– 3.08 ± 0.12 3.42± 0.23
2.78 ± 0.09 4.00 ± 0.09 5.53 ± 0.32
3.47 ± 0.09 4.16 ± 0.33 6.22 ± 0.29
4.26 ± 0.28 6.67 ± 0.48 7.66 ± 0.27
Fluence= 200 mJ/cm2 100 200 300
Figure 30. Lines mechanically induced on 100%-stretched PET film, which were further enhanced by subsequent excimer laser irradiation.
on this phenomenon is presented in Fig. 30. 100%-stretched PET films were scratched with a piece of paper before irradiating with an excimer laser to make fine lines on their surfaces. By irradiating these PET films with an excimer laser using a fluence of 200 mJ/cm2 and 20 pulses we could generate aligned line structures as indicated with arrows in Fig. 30. Note that these lines were already on the film before laser ablation. Obviously, regardless of whether structure formation is
72
J. Kim et al.
initiated by the excimer laser or artificially induced by mechanical scratching, these structures are enhanced by subsequent excimer laser ablation. Microstructures developed on biaxially stretched (300% ¥ 300%) PET films were also investigated. These structures are different from those on uniaxially stretched films. As shown in Fig. 31, the structures are initially segmented rolltype, but as the pulse number increases to more than 50, structures are transformed to ripple-type. When the film surface was irradiated at high repetition rate with more than 100 pulses, these ripple-type structures turned into cylinder-type structures as shown in Fig. 31. Upon irradiation, a plume originates from the target polymer film surface accompanied by an explosive sound. Due to this plume and explosion, the film surface experiences high pressure. We think the previously mentioned cylinder-type structure is an evidence of surface melting followed by deformation of the formed structures by such pressure generation.
Figure 31. Surface morphology of biaxially stretched PET films due to excimer laser irradiation with different conditions (from the top: 10, 50, 100 and 150 pulses).
Effects of excimer laser irradiation on surface morphology
73
Number densities (1/µm2) for ripple-type and cylinder-type structures were measured for the microstructures shown in Fig. 31 and the results are presented in Table 10. As shown in Table 10, microstructure number density decreased with increasing laser irradiation pulse number, as well as fluence. It can also be observed that these values decreased more dramatically when sample films were irradiated at higher repetition rate. The reason for this decrease in number density of microstructures is thought to be the merger of small structures during excimer laser irradiation. This should be considered a strong evidence for surface melting. Figure 32 shows the microstructure of uniaxially stretched PET film, which was irradiated using 106 mJ/cm2, 1 Hz and 200 pulses. As indicated with arrows, one can Table 10. Microstructure number density (1/µm2) developed by excimer laser irradiation on biaxially stretched PET No. of pulses
Laser conditions 60 mJ/cm2
50 100 150 200 300 500 800
106 mJ/cm2
1 Hz
50 Hz
1 Hz
50 Hz
15.77 7.50 3.75 – 2.24 1.97 1.62
5.65 2.47 1.31 – – – –
4.99 2.78 1.93 1.43 1.12 0.77 –
3.71 1.54 0.43 – – – –
Figure 32. Growth of bigger microstructure by merger of small structures on uniaxially stretched PET film. The arrows indicate such merger locations.
74
J. Kim et al.
observe the merger of 2–5 small structures leading to bigger structures. The mechanism of this structure merger is still not clear, but we think that lateral melt flow due to material ejection and viscous forces in the melt play important roles. In the pictures of microstructures on uniaxially-stretched films, one can see that the microstructures are not generated newly from each laser irradiation pulse, but initially formed structures grow with subsequent pulses. For the biaxially stretched PET film, a similar behavior was also found. A biaxially stretched PET film was irradiated by excimer laser while tilting the film sample at about 45o to the beam direction. 300 mJ/cm2 of fluence and 100 pulses were used to produce the microstructures shown in Fig. 33. These microstructures exhibit directionality commensurate with the 45o tilt angle used. This indicates that the microstructures generated by previous laser pulses influence subsequent structure formation by blocking the incoming excimer laser radiation. During 100-pulse irradiation, explosive sound and plume coming out from the surface with each excimer laser pulse were observed, and the direction of this plume was perpendicular to the film surface and not along the laser beam axis. As mentioned before, microstructures developed on PET and PBT films only when the films were stretched either uniaxially or biaxially. Stretched films possess higher crystalline percentage and chain orientation than unstretched cast film for both PET and PBT films. In 1988, Novis et al. [35] reported that the surface structures created on PET could be interpreted as resulting mainly from the difference in the etch rates between the amorphous and crystalline regions in PET films. Our experimental results also show that the ablation depth in the amorphous region is higher than that in the crystalline region (see Section 5.1.2.2). In order to investigate the possibility of origin of microstructure development due to existing crystalline and amorphous regions in the polymer, PET cast films were
Figure 33. Microstructures developed on biaxially stretched PET film by excimer laser irradiation when the sample film was tilted 45º to the laser beam.
Effects of excimer laser irradiation on surface morphology
75
annealed at 150oC for 20 min before excimer laser irradiation. Annealed PET cast films showed about 37% crystallinity, which is similar to that of 300% stretched PET films. Upon investigating the ablated surface by SEM, no structures were found on the annealed surface. So, the existence of both crystalline and amorphous regions does not appear to be the main reason for microstructure development. 5.3.2. Morphology of PS surface after excimer laser irradiation The surface morphologies of PS films were investigated after excimer laser irradiation. Since the measured ablation threshold for this material was around 110 mJ/cm2, ablation fluences were chosen between 120 and 200 mJ/cm2 with 1 Hz repetition rate. When SEM pictures were taken after excimer laser irradiation of film surfaces, no structures were found on cast or 100%-stretched PS films. But on 200% and 300% stretched and irradiated PS films, roll-type microstructures were observed, which were similar to those on stretched PET and PBT films but different in size. While stretched PET and PBT films can produce single-digit micrometer-size structures, PS produced double-digit micrometer-size microstructures, as shown in Fig. 34.
Figure 34. Surface patterns on 200% stretched PS film irradiated using different laser conditions (fluences from the top: 120, 160 and 200 mJ/cm2).
76
J. Kim et al.
PS surfaces on excimer laser irradiation have a tendency for structure development similar to that on PET and PBT films: – Increasing pulse number or fluence causes wide structures. – Highly stretched films can produce surface microstructures more easily (need less fluence and low pulse number) The material we used in this experiment was atactic and amorphous linear polymer. DSC scan for PS in Section 4.2 did not show any melting peak which means this material was amorphous. So the microstructures developed on these stretched films also indicate that the microstructure development is not exactly related to the existence of both crystalline and amorphous regions in the polymer. Based on these observations one can say that the film-stretching process controls the formation of microstructures. 5.4. Surface microstructures created by Fresnel diffraction Besides the microstructures mentioned in Section 5.3, we generated well-aligned, controllable microstructures on PET films using Fresnel diffraction. The experimental set-up and the theory of Fresnel diffraction are depicted in Fig. 35. In this experiment, the distance of laser source to the straight edge and the distance of the straight edge to sample film were set to 0.9144 m and 6 mm, respectively. KrF excimer laser (248 nm) was used with conditions of 200 mJ/cm2, 1 Hz and 20 pulses. 300%-stretched PET films were irradiated under the abovementioned excimer laser conditions to generate microstructures presented in Fig. 36. Note that the microstructures are aligned perpendicular to the film stretching direction.
Figure 35. The principle of diffraction at a straight edge (Fresnel diffraction).
Effects of excimer laser irradiation on surface morphology
77
In comparison with the microstructures developed on stretched PET films without Fresnel diffraction, Fig. 36 shows very well-defined line structures, which we called channel type. We could decrease the peak-to-peak distance by moving the sample away from the straight edge. While the peak-to-peak distances in the microstructures developed without Fresnel diffraction increased with increasing film stretching ratio, number of total pulses, and laser fluence, those with Fresnel diffraction showed the same distances. However, the number of channels were different with varying the laser irradiation conditions and film-stretching ratio. Increasing laser fluence, total pulse number and film-stretching ratio all caused increases in the number of channels, but not in the peak-to-peak distance. Peak-to-peak distances for this kind of channel-type microstructures were measured using image analyzing software. Observed and calculated distances are presented in Table 11. Calculated values shown in Table 11 were based on minimum excimer laser intensity condition depicted in Fig. 35. Higher laser intensity resulted in deeper ablation, causing the creation of channels (depressions). The observed and calculated results showed good agreement, indicating that the channel-type microstructures in Fig. 36 were due to the etch rate difference caused by intensity variation in Fresnel diffraction.
Figure 36. Microstructures developed by excimer laser irradiation on 300%-stretched PET film using Fresnel diffraction.
J. Kim et al.
78
Table 11. Peak-to-peak distances (mm) in microstructures developed by excimer laser irradiation using Fresnel diffraction (using the minimum intensity equation in Fig. 35) on 300% stretched PET film Measurements
Calculations
21.67 17.11 15.21 12.55 11.41 .
22.78 17.48 14.74 12.99 11.74 .
.
.
.
.
.
.
4.55 4.4
4.49 4.43
Figure 37. Fresnel diffraction pattern on cast PET film.
As explained above, the microstructures developed with Fresnel diffraction are caused by the etch rate difference due to laser intensity variations. This means that the origin of microstructures developed by this method is not from the variation in material properties on film surfaces but from the optical phenomenon of the excimer laser beam. This conclusion is evident from the Fresnel diffraction pattern on cast PET film (Fig. 37). Without Fresnel diffraction, we did not observe any structures on non-stretched films after excimer laser irradiation. Even though we could generate Fresnel patterns on non-stretched PET films using excimer laser we found that the microstructures developed in stretched and non-stretched cases were somewhat different. The microstructures on stretched PET films are better defined and more clear than those on non-stretched films. While non-stretched PET films did not generate microstructures on their surfaces
Effects of excimer laser irradiation on surface morphology
79
Figure 38. Fresnel diffraction patterns on stretched PET films. The stretching direction is 90, 45 and 0o for (a), (b) and (c), respectively.
without Fresnel diffraction, stretched films generated microstructures on film surfaces in a direction perpendicular to the film stretching direction. Considering these facts we could conclude that microstructures created due to film stretching also contributed to the development of Fresnel diffraction patterns on film surfaces. This means that stretched films have the ability to enhance the development of Fresnel diffraction patterns on their surfaces. This idea is more evident from the experiments involving excimer laser irradiation on stretched PET films with different Fresnel diffraction directions. Experiments were preformed on stretched PET films with Fresnel diffraction of excimer laser light at different angles to the stretching direction. The resulting microstructures are shown in Fig. 38. Microstructure patterns in Fig. 38a were produced from the same experimental conditions as those in Fig. 36, and the microstructures in Fig. 38b and 38c were generated by rotating the sample film 45 and 90o, respectively, before laser ablation. With 90o rotation, the film-stretching direction was aligned perpendicular to the propagation direction of Fresnel diffraction. Since 300%-stretched PET film was irradiated using a fluence of 200 mJ/cm2, we could expect the roll-type structure perpendicular to the stretching direction (explained in Section 5.3.1) in the form of channel pattern from Fresnel diffraction. Observed microstructures were a combination of these two structures, which is clearly seen in Fig. 38. When the stretching directions of the samples were 45 and 0o to the Fresnel slit (Fig. 38b and 38c, respectively), two types of microstructures were generated: one due to excimer laser irradiation on stretched film and the other from the Fresnel diffraction. When the stretching direction of the sample was 90o to the Fresnel slit (Fig. 38a); however, only one type of microstructure (straight channels) was formed, generated from Fresnel diffraction, even though this microstructure is a combination of Fresnel diffraction patterns and the microstructures developed by excimer laser irradiation on stretched film.
J. Kim et al.
80
5.5. Surface morphology and polymer orientation relationship As mentioned in Section 5.4, the microstructure due to excimer laser irradiation on polymer films was observed only if the films were stretched. No structures were observed on cast PS, PET and PBT films. Regarding this microstructure development on stretched films, one can say that polymer orientation might be the main reason for microstructure development due to excimer laser irradiation. In order to investigate the dependence of microstructures development and orientation of polymer films, we measured the orientation of PBT films by birefringence and WAXS (Section 4.1) before irradiating the surfaces by excimer laser. After development of the microstructures on the film surface, SEM pictures were taken and analyzed using an image analyzing software. In addition to measuring the mean roll-toroll distance of microstructures (Section 5.3.1), all line segments that made up the microstructures were analyzed by measuring angles and lengths (Fig. 39) in order to investigate the relationship between microstructure pattern (especially line direction resulted from excimer laser irradiation) and polymer chain orientation factor. The orientation factor (the relative orientation with respect to the stretching direction) for the vector normal to surface structure (Fig. 39) was analyzed in a similar way as proposed by Stein [95] for determination of crystallographic orientation of a crystalline polymer:
cos θ i 2
∑ l cos = ∑l
2
i
i
θi (13)
i
i
fi =
3cos 2 θ − 1 2
(14)
If the microstructures are exactly perpendicular to the film stretching direction (machine direction, MD), their relative orientation f i is 1, and for those structures completely parallel to the film stretching direction f i should have a value of -1/2. The values of f i calculated for the microstructures of PET and PBT are listed in Table 12. The microstructures were generated using 200 mJ/cm2 laser fluence and 30 pulses with 1 Hz repetition rate. While the chain orientation factor discussed in Section 4.1.2 increased with increasing film stretching ratio for PBT films, the relative orientation of microstructures to film-stretching direction did not show any trend beyond 100% stretching, which apparently was a sufficient level of stretching to initiate directional micro-structuring. Higher levels of stretching beyond 100% likely help increase pattern definition and clarity, but the directionality remains the same. This conclusion is supported by the fact that the values in Table 12 average around 0.8,
Effects of excimer laser irradiation on surface morphology
81
Figure 39. Measurements of lengths (l) and angles (θ) for line segments that make up microstructures due to excimer laser ablation. ly and lx are the vector components along the machine direction (MD) and transverse direction, respectively.
Table 12. Relative orientation of the microstructures to film stretching direction for PET and PBT developed by excimer laser irradiation (200 mJ/cm2, 30 pulses)
PBT PET
100%
200%
300%
0.87 0.86
0.85 0.80
0.86 0.76
If the microstructures are exactly perpendicular to the film stretching direction, the relative orientation is 1, and for those structures completely parallel to the film stretching direction it should have a value of -1/2.
which means that most of the microstructures are perpendicular to the filmstretching direction. Furthermore, the microstructures’ peak-to-peak distance discussed in Section 5.3.1 (Fig. 29) increased with increasing film-stretching ratio. 5.6. Discussion on microstructures formation The formation of microstructures due to excimer laser irradiation on a polymer is one of the unique results of excimer laser surface treatment. Many researchers have studied the origins of these microstructures formation. Some researchers [9, 35] sought to find this origin in different etching rates between amorphous and crystalline regions in polymer film. This argument was possibly based on the fact that the ablation rate of amorphous region was higher than that of crystalline region. Our experimental results also support this argument by showing 40% higher ablation depth for amorphous region in comparison to the crystalline region (see Section 5.1.2.2). However, the microstructures on the stretched PS films could not be explained by this argument, because the PS could not have any crystalline region in the films.
J. Kim et al.
82
Table 13. The variation of PET film length due to heat shrinkage on exposure to 140oC for 2 h Cast film
100% PET
200% PET
300% PET
5 cm
2.9 cm
4.6 cm
4.95 cm
Other researchers [32, 36, 37] attribute the origin of microstructures to the stress field existing in polymer films, which results from film-processing conditions. They insisted that the stress released from the internal tension due to high temperature gradients created by excimer laser irradiation might lead to a cooperative movement of polymer chains, and this movement could be the origin for microstructure development. There were other explanations for the microstructure formation [73, 96], which were based on the heat shrinkage of the polymer film. In order to investigate the heat shrinkage behavior of PET films, cast and stretched PET films were cut to a rectangular shape of 1 cm by 5 cm and placed in an oven at 140oC for 2 h. After taking these films out from the oven, the dimensions for each sample were measured. Experimental results are shown in Table 13. As shown in Table 13, stretched PET films showed lower heat shrinkage with increasing stretch ratio beyond 100%. Consequently, the highly stretched films were more stable on exposure to heat, which indicated that polymer chain relaxation or molecular movement were more feasible when films were less stretched. So, if the origin of microstructure development is due to the chain movement with existing temperature gradient or due to the heat shrinkage of the polymer films, one should be able to develop a microstructure more easily on 100%-stretched films than on 200%- or 300%-stretched films. However, our earlier experimental results showed that more stretched films generated the microstructures more efficiently, which could not possibly be explained by our current heat shrinkage experiments. As discussed in Section 5.3, roll-type microstructures were readily developed by excimer laser irradiation on stretched PET and PBT films, and the peak-topeak distance increased with increasing stretching ratio for the films. We also observed that the microstructures on 300%-stretched films were better defined and contrasted compared to those on 100%-stretched films. It is unlikely for polymer chains to relax their orientation during excimer laser ablation because of short excimer laser duration (25 ns) followed by rapid cooling. During the film-stretching process, we observed that the degree of crystallinity increased with increasing stretching ratio. Due to this stretching process, one could expect separation and localization of amorphous and crystalline regions for crystalline materials. One could also expect separation and localization of oriented and unoriented regions for PS material. These crystalline and oriented regions resist laser ablation more than amorphous and unoriented regions in the films, and the uneven etching of the surface could result in uneven pressure generation during excimer laser ablation.
Effects of excimer laser irradiation on surface morphology
83
6. SUMMARY AND CONCLUSIONS
Based on the above considerations, we would like to propose a mechanism for microstructure formation, as follows: – At first, a kind of seed microstructure is generated. This is possible by selective ablation in the stretched films, which have two phases: one is easily ablated by UV light, i.e., UV-weak phase (amorphous or relatively unoriented region in the film) and the other one is not easily ablated by UV light, i.e., UV-strong phase (crystalline or well-oriented region). Further stretching of the film will cause further separation of these two phases. Relatively high ablation rates in the UV-weak phase of the polymer film will result in more ablation and will exert higher pressures to the film in lateral and perpendicular directions in comparison to UV-strong phase due to the explosive ablation nature of the excimer laser process. Such inhomogeneous ablation between two phases will cause pressure difference in film surface and this is thought to be a driving force for chain movement in parallel and perpendicular to the surface, because the pressure exerted due to ablation in high ablation rate region should be higher than that in low ablation region. We think this chain movement due to the different ablation pressures (also perhaps combined with the internal stress field or heat shrinkage of polymer chain) is the main origin of microstructure development. – After their development, these small microstructures could merge together and result in larger structures as excimer laser ablation continues. This also could be possible due to the surface melting followed by inhomogeneous material ejections from the surface during excimer laser irradiation. These inhomogeneous material ejections from the surface will cause pressure gradients, which could cause polymer melt movement. This melt movement can cause the structures to combine together. – In addition to this merger of small microstructures, the larger structures could be contrasted and enhanced in definition by repeated excimer laser irradiation pulses. During excimer laser ablation processes, we could observe the plumes with explosive sound coming from the surface, and the direction of the plumes was always normal to the surface. Such dilatational stress effect of the plume may enhance the microstructural patterns.
REFERENCES 1. R. Srinivasan and W.J. Leigh, J. Am. Chem. Soc. 104, 6784 (1982). 2. R. Srinivasan and V. Mayne-Banton, Appl. Phys. Lett. 41, 576 (1982). 3. G.M. Davis, M.C. Gower, C. Fotakis, T. Efthimiopoulos and P. Argyrakis, Appl. Phys. A 36, 27 (1985). 4. R. Srinivasan and B. Braren, Chem. Rev. 89, 1303 (1989). 5. V. Svorcik, V. Rybka and P. Onheiser, Chem. Listy 84, 1067 (1990).
84
J. Kim et al.
6. M.S. Kitai, V.L. Popkov and V.A. Semchishen, Makromol. Chem. Macromol. Symp. 37, 257 (1990). 7. H. Niino, M. Nakano, S. Nagano, H. Nitta, K. Yano and A. Yabe, J. Photopolym. Sci. Technol. 3, 53 (1990). 8. H. Hiraoka and S. Lazare, Appl. Surface Sci. 46, 264 (1990). 9. J.E. Andrew, P.E. Dyer, D. Forster and P.H. Key, Appl. Phys. Lett. 43, 717 (1983). 10. R. Srinivasan, Stud. Phys. Theor. Chem. 35, 595 (1985). 11. P.E. Dyer, S.D. Jenkins and J. Sidhu, Appl. Phys. Lett. 49, 453 (1986). 12. Q. Lou and L. Zhang, Proc. SPIE 4404, 343 (2001). 13. W.C. Lee, V.W. Lindberg, P.H. Wojciechowski and F.J. Duarte, in: Metallized Plastics: Fundamental and Applied Aspects, Vol. 2, K.L. Mittal (Ed.), pp. 449–460. Plenum Press, New York, NY (1991). 14. J.-Y. Zhang, H. Esrom, U. Kogelschatz and G. Emig, J. Adhesion Sci. Technol. 8, 1179 (1994). 15. H.C. Man, M. Li and T.M. Yue, Proceedings of the 10th International Conference on Composite Materials, Whistler, BC, Vol. 3, p. 559 (1995). 16. M. Stehle, J. Non-Cryst. Solids 218, 218 (1997). 17. G. Fortunato, L. Mariucci, R. Carluccio, A. Pecora and V. Foglietti, Appl. Surface Sci. 154, 95 (2000). 18. D. Knittel, A. Eickmeier, T. Bahners and E. Schollmeyer, Textile. Prax. Int. 45, 236 (1990). 19. H. Watanabe and T. Takata, Sen'i Gakkaishi 49, 274 (1993). 20. S. Liang, X. Zhou, Q. Lou, T. Tu and Z. Xing, Proc. SPIE 4915, 272 (2002). 21. S. Liang, X. Zhou, T. Tu, Z. Xing and Q. Lou, Color. Technol. 119, 19 (2003). 22. S. Mihailov and S. Lazare, Appl. Opt. 32, 6211 (1993). 23. F.H.H. Lin, J.H. Huang, E.H.Y. Chou, C.R. Yang, B.C.S. Chou, R.K.S. Luo, W.K. Kuo, J.W. Chang, M.H. Lu, W.H. Huang and C.J. Chen, Proc. SPIE 3511, 67 (1998). 24. W.-S. Chang, K.-K. Yoon, J. Kim, B. Shin and K.-H. Whang, Proc. SPIE 4760, 1014 (2002). 25. K. Naessens, H. Ottevaere, P. Van Daele and R. Baets, Appl. Surface Sci. 208, 159 (2003). 26. H. Frerichs, J. Stricker, D.A. Wesner and E.W. Kreutz, Appl. Surface Sci. 86, 405 (1995). 27. H. Horn, S. Beil, D.A. Wesner, R. Weichenhain and E.W. Kreutz, Nucl. Instrum. Methods Phys. Res., B 151, 279 (1999). 28. S. Petit, P. Laurens, J. Amouroux and F. Arefi-Khonsari, Appl. Surface Sci. 168, 300 (2000). 29. E. Sancaktar, S.V. Babu, E. Zhang, G.C. D'Couto and H. Lipshitz, J. Adhesion 50, 103 (1995). 30. H. Dodiuk, A. Buchman, M. Rotel and J. Zahavi, in: Mittal Festschrift on Adhesion Science and Technology, W.J. van Ooij and H.R. Anderson, Jr. (Eds.), pp. 387–405. VSP, Utrecht (1998). 31. H. Dodiuk, A. Buchman, S. Kenig, M. Rotel, J. Zahavi and T.J. Reinhart, J. Adhesion 41, 93 (1993). 32. T. Bahners and E. Schollmeyer, Proc. SPIE 1132, 146 (1989). 33. A. Yabe and H. Niino, Eur. Mater. Res. Soc. Monogr. 4, 199 (1992). 34. T. Kimura, H. Tokunaga, E. Ito, H. Niino and A. Yabe, Polym. J. 31, 524 (1999). 35. Y. Novis, J.J. Pireaux, A. Brezini, E. Petit, R. Caudano, P. Lutgen, G. Feyder and S. Lazare, J. Appl. Phys. 64, 365 (1988). 36. B. Hopp, M. Csete, K. Revesz, J. Vinko and Z. Bor, Appl. Surface Sci. 96, 611 (1996). 37. D. Knittel, W. Kesting and E. Schollmeyer, Polym. Int. 43, 231 (1997). 38. R. Srinivasan, Science 234, 559 (1986). 39. S. Rice and K. Jain, Appl. Phys. A 33, 195 (1984). 40. B. Wolff-Rottke, H. Schmidt, A. Scholl and J. Ihlemann, Proc. SPIE 1810, 650 (1993). 41. N.P. Furzikov, Appl. Phys. Lett. 56, 1638 (1990). 42. F. Wagner and P. Hoffmann, Appl. Surface Sci. 154, 627 (2000). 43. S. Lazare and V. Granier, J. Appl. Phys. 63, 2110 (1988). 44. E. Arenholz, J. Heitz, M. Wagner, D. Baeuerle, H. Hibst and A. Hagemeyer, Appl. Surface Sci. 69, 16 (1993). 45. D.A. Wesner, H. Horn, R. Weichenhain, W. Pfleging and E.W. Kreutz, J. Adhesion Sci. Technol. 11, 1229 (1997).
Effects of excimer laser irradiation on surface morphology
85
46. R. Srinivasan and B. Braren, J. Polym. Sci. Part A: Polym. Chem. 22, 2601 (1984). 47. P.E. Dyer and J. Sidhu, J. Appl. Phys. 57, 1420 (1985). 48. J.H. Brannon, J.R. Lankard, A.I. Baise, F. Burns and J. Kaufmann, J. Appl. Phys. 58, 2036 (1985). 49. A. Yabe, Springer Ser. Mater. Sci. 35, 171 (1999). 50. M.C. Burrell, Y.S. Liu and H.S. Cole, J. Vac. Sci. Technol. A 4, 2459 (1986). 51. F. Kokai, H. Saito and T. Fujioka, Proc. SPIE 1190, 95 (1990). 52. S. Lazare, P.D. Hoh, J.M. Baker and R. Srinivasan, J. Am. Chem. Soc. 106, 4288 (1984). 53. S. Lazare and R. Srinivasan, J. Phys. Chem. 90, 2124 (1986). 54. F.B. Marcotte, D. Campbell, J.A. Cleaveland and D.T. Turner, J. Polym. Sci., Polym. Chem. Ed. 5, 481 (1967). 55. M. Day and D.M. Wiles, J. Appl. Polym. Sci. 16, 203 (1972). 56. M. Day and D.M. Wiles, J. Appl. Polym. Sci. 16, 175 (1972). 57. H. Watanabe and M. Yamamoto, J. Appl. Polym. Sci. 71, 2027 (1999). 58. R. Srinivasan and S. Lazare, Polymer 26, 1297 (1985). 59. T. Bahners and E. Schollmeyer, Angew. Makromol. Chem. 151, 19 (1987). 60. H. Niino, M. Nakano, S. Nagano, A. Yabe and T. Miki, Appl. Phys. Lett. 55, 510 (1989). 61. A. Yabe and H. Niino, Mol. Cryst. Liq. Cryst. Sci. Technology A 224, 111 (1993). 62. S. Lazare, P. Benet, W. Guan, M. Bolle and S. Mihailov, in: Excimer Lasers (NATO ASI Series E: Vol. 265), L.D. Laude (Ed.), pp. 201–219. Kluwer, Dordrecht (1994). 63. A. Yabe, H. Niino, S. Ono and Y. Sato, Proc. SPIE 2992, 119 (1997). 64. T. Lippert, T. Nakamura, H. Niino and A. Yabe, Appl. Surface Sci. 109, 227 (1997). 65. T. Kunz, J. Stebani, J. Ihlemann and A. Wokaun, Appl. Phys. A 67, 347 (1998). 66. A. Yabe and H. Niino, Reza Kenkyu 26, 155 (1998). 67. K. Naessens, A. Van Hove, T. Coosemans, S. Verstuyft, L. Vanwassenhove, P. Van Daele and R.G. Baets, Proc. SPIE 4088, 327 (2000). 68. B. Keiper, R. Ebert, R. Bohme and H. Exner, Proc. SPIE 5116, 20 (2003). 69. M. Bolle, S. Lazare, M. Le Blanc and A. Wilmes, Appl. Phys. Lett. 60, 674 (1992). 70. M. Bolle and S. Lazare, Appl. Surface Sci. 65, 349 (1993). 71. M. Csete and Z. Bor, Appl. Surface Sci. 133, 5 (1998). 72. M. Csete, R. Eberle, M. Pietralla, O. Marti and Z. Bor, Appl. Surface Sci. 208, 474 (2003). 73. H. Watanabe and M. Yamamoto, J. Appl. Polym. Sci. 64, 1203 (1997). 74. P.E. Dyer, S.D. Jenkins and J. Sidhu, Appl. Phys. Lett. 52, 1880 (1988). 75. B. Hopp, Z. Bor, E. Homolya and E. Mihalik, Proc. SPIE 3423, 389 (1998). 76. F. Raimondi, S. Abolhassani, R. Brutsch, F. Geiger, T. Lippert, J. Wambach, J. Wei and A. Wokaun, J. Appl. Phys. 88, 3659 (2000). 77. D.J. Krajnovich and J.E. Vazquez, J. Appl. Phys. 73, 3001 (1993). 78. J.-F. Silvain, H. Niino, S. Ono, S. Nakaoka and A. Yabe, Appl. Surface Sci. 141, 25 (1999). 79. H. Hiraoka, S. Lazare and A. Cros, J. Photochem. Photobiol. A 65, 293 (1992). 80. L.I. Kalontarov and R. Marupov, Chem. Phys. Lett. 196, 15 (1992). 81. Y. Tsuboi, S.-I. Sakashita, K. Hatanaka, H. Fukumura and H. Masuhara, Laser Chem. 16, 167 (1996). 82. N. Arnold, N. Bityurin and D. Biuerle, Appl. Surface Sci. 138, 212 (1999). 83. N. Arnold and N. Bityurin, Appl. Phys. A 68, 615 (1999). 84. M. Sadoqi, S. Kumar and Y. Yamada, J. Thermophys. Heat Trans. 16, 193 (2002). 85. J.-Y. Zhang, H. Esrom, U. Kogelschatz and G. Emig, J. Adhesion Sci. Technol. 8, 1179 (1995). 86. R. Srinivasan, Springer Ser. Mater. Sci. 28, 107 (1994). 87. B.J. Garrison and R. Srinivasan, Appl. Phys. Lett. 44, 849 (1984). 88. R. Srinivasan, J. Appl. Phys. 73, 2743 (1993). 89. K. Horiuchi, M. Ishiyama, T. Hasebe, K. Yukimura and Y. Imaida, Mater. Chem. Phys. 54, 201 (1998). 90. F.S. Smith and R.D. Steward, Polymer 15, 283 (1974). 91. K.H. Illers, Colloid Polym. Sci. 258, 117 (1980).
86
J. Kim et al.
92. I.H. Hall and M.G. Pass, Polymer 17, 807 (1976). 93. J.L. White and J.E. Spruiell, Polym. Eng. Sci. 21, 859 (1981). 94. R. Srinivasan and B. Braren, Appl. Phys. Lett. 53, 1233 (1988). 95. R.S. Stein, J. Polym. Sci. 31, 327 (1958). 96. K. Mizoguchi, A. Yamamoto, M. Ishikawa, A. Ohuchi and T. Ohie, Proceedings of the 7th Japan International SAMPE Symposium, Tokyo, p. 801 (2001).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 87–111 Ed. K.L. Mittal © VSP 2007
XeCl excimer laser treatment of Vectran® fibers in diethylenetriamine (DETA) environment J. ZENG and A. N. NETRAVALI* Fiber Science Program, Cornell University, Ithaca, NY, 14853-4401, USA
Abstract—Vectran® fibers, made using high-performance liquid crystal polymer (LCP), were treated with XeCl pulsed excimer laser (308 nm) to improve their adhesion to epoxy resin. The laser treatments were carried out in diethylenetriamine (DETA) environment. Varying numbers of laser pulses at different laser fluences were used. The effects of the laser treatment on the fiber surface topography, chemistry and wettability have been investigated. The surface roughness was characterized using scanning electron microscopy (SEM) and atomic force microscopy (AFM). The fiber/epoxy resin interfacial shear strength (IFSS) was characterized using microbead test. The surface roughness of laser-treated fibers increased by up to 2.5-times the value of the control fiber. However, the characteristic wavy structures on the laser-treated fiber surface that were seen when the treatment was carried out in air were not seen after any laser treatment in DETA. After the laser treatment, the dispersion component of the surface energy decreased, while the acid–base component of the surface energy increased significantly. The Vectran® fiber/epoxy resin IFSS increased by up to 50% after the laser treatment. A large part of this improvement could be attributed to higher surface roughness of the laser-treated fibers. Keywords: Vectran® fiber; XeCl excimer laser; adhesion; DETA; fiber/resin interface; wettability; surface energy; surface roughness; interfacial shear strength.
1. INTRODUCTION
Vectran® fiber is one of the latest entries into the high strength fibers and is manufactured by Celanese Advanced Materials (Charlotte, NC, USA). Figure 1 shows the chemical structure of Vectran® fiber [1]. It is the only commercially available melt-spun thermotropic liquid crystalline polymer (LCP) fiber [1]. It has excellent mechanical properties and creep resistance, good chemical resistance and high thermal stability. These excellent properties are a result of the fiber’s high molecular orientation, high crystallinity and fully aromatic nature [1]. Vectran® fibers have been used in many applications, including high-performance ropes and cables, packaging materials and some sports and leisure equipments, such as racquets and tennis strings [1]. The airbags used in the recent Mars exploration were *
To whom correspondence should be addressed. E-mail:
[email protected]
88
J. Zeng and A. N. Netravali
Figure 1. The chemical structure of Vectran® fiber [1].
made of Vectran® fiber as reported by NASA. Though Vectran® fibers have excellent mechanical and chemical properties, their application in advanced composites has been limited because of their low adhesion to the resins. This is due to their relatively low surface energy, non-polar nature and smooth surface [2]. In fiber-reinforced composites, high performance fibers are the main loadbearing component. However, the fiber/resin interface has been shown to play a critical role in determining the mechanical performance of the composites [3–5]. In fact, some of the critical properties such as toughness of the composites and transverse strength in unidirectional composites are directly related to the fiber/resin interfacial characteristics. High interface strength is generally associated with higher strength of composites, whereas lower interface strength provides energy absorbing modes resulting in higher toughness. Three main factors are considered to contribute to the interfacial adhesion in composites; (a) chemical interaction, (b) mechanical interaction (interlocking) and (c) physicochemical interaction [6]. Enhancement in any of these three factors can improve the fiber/resin interface bonding. Up to now, only little has been published in the open literature on improving the Vectran® fiber/resin interface. However, poly(ethylene terephthalate) (PET), which belongs to aromatic polyester family similar to Vectran®, has been extensively studied for various surface modifications by treatments including different plasmas and excimer lasers. These treatments were employed to improve the surface roughness, wettability, dyeability and adhesion of PET to a variety of materials [7–12]. Heitz et al. [7] used UV excimer laser to irradiate the PET foil surface prior to metal deposition. They found that the adhesion of thin metal films on PET foils was significantly improved by the laser treatment. Le et al. [11] employed nitrogen and oxygen plasma treatments to modify PET film surfaces. Their study indicated that the aluminum/PET adhesion improved after nitrogen plasma treatment. Additionally, they found that the oxygen plasma treatment was even more effective in improving the aluminum/PET adhesion than the nitrogen plasma. The XPS results indicated that the additional reactive and polar sites, especially oxygen containing groups, were created on the PET surface by the plasma treatment. These reactive sites resulted in covalent bonds with aluminum (for example, Al– O), increasing the adhesion significantly. Zhao et al. [12] employed (SO2 + O2)
XeCl excimer laser treatment of Vectran fibers in DETA
89
and (N2 + H2 + He) plasmas on PET fibers to improve their dyeability by watersoluble dyes. The plasma treatments introduced polar functional groups on the PET fiber surfaces and increased the acid–base component of the surface energy of the fibers resulting in improved dyeability. UV excimer laser irradiation has been shown to be an effective method to modify polymeric surfaces [13, 14]. The effects of laser radiation on various polymeric materials have been extensively studied by many researchers [14–22]. UV excimer laser radiation, because of its high energy, can cause photodecomposition, photoablation and melting in most polymers depending on their UV absorption coefficients [14–16, 21, 22]. In the case of ultra-high-molecular-weight polyethylene (UHMWPE) fibers which do not absorb significantly in the UV range, longitudinal groove structures have been seen, due to the fibrillar structure of the fibers [21, 22]. For fibers made of highly-UV-absorbing polymers, such as polyesters (e.g., PET) and polyamides (PAs), the characteristic highly regular roll (wavy) structures in the transverse direction have been found after UV laser treatments [19]. As a result of these wavy structures, the surface roughness was shown to increase significantly [23]. Wong et al. [23] found that changes in the surface morphology of PET fibers were closely related to the laser energy applied. The mean roll-to-roll distance increased with increase in the laser energy. The merging of the ripples was observed and was believed to be a major reason for increased roll-to-roll distance. The surface chemistry of the laser-treated polymers had also changed. Watanabe and Takata [24] found that after KrF excimer laser treatment of PET and poly(p-phenylene-3,4'-oxydiphenylene terephthalamide) (PPODPTA) fibers, the adhesion performance was improved, whereas for poly(pphenylene terephthalamide) (PPTA) fibers it was not. In PET fibers, the surface layer was converted into an amorphous state, and, therefore, the affinities for the treating agents were enhanced [24]. Zeng and Netravali [25, 26] employed KrF (248 nm) and XeCl (308 nm) excimer lasers to treat UHMWPE fibers in air as well as in diethylenetriamine (DETA), a liquid amine, at room temperature. After the laser treatment in DETA, some nitrogen containing groups in addition to oxygen containing groups were observed on the surface of UHMWPE fibers. This resulted in higher wettability of fiber surface than that of corresponding fibers laser-treated in air. The fiber/epoxy resin interfacial adhesion was also found to increase by up to 6-fold. Vectran®, which belongs to aromatic polyester family, just as PET, has a high absorption coefficient in the UV region. As a result, UV excimer laser treatment is an effective method to modify the Vectran® fiber surface. The results of Vectran® fibers treated by pulsed XeCl excimer laser (308 nm) in air have been presented elsewhere [27]. After the laser treatment, the fiber surface roughness increased significantly. The extent of roughness increase was a function of both the laser fluence and the number of pulses. At fluences of 24 and 36 mJ/(pulse◊cm2), smooth and nodular surface structures were obtained on the laser-treated fibers, respectively. At the higher fluence of 60 mJ/(pulse◊cm2), well above the threshold
90
J. Zeng and A. N. Netravali
fluence, wavy surface structures developed on the laser-treated fibers. Also at the fluence of 60 mJ/(pulse·cm2), the mean roll-to-roll distance (mean roll wavelength) depended on the number of pulses. The surface also changed from hydrophobic (non-polar) to hydrophilic after the laser treatment. The IFSS increased by up to 75% compared to the IFSS for untreated fibers, and this enhancement has been mainly attributed to the increase in surface roughness. In the present research, pulsed XeCl excimer laser has been used to treat Vectran® fibers, in diethylenetriamine (DETA), to improve their adhesion to diglycidyl ether of bisphenol A (DGEBA) based epoxy resin. The purpose of using DETA environment is to introduce reactive nitrogen containing groups onto the Vectran® fiber surface and, thus, improve its chemical bonding to epoxy resins. The effects of laser fluence, number of pulses and the DETA environment on the excimer laser treatment of the fiber surface have been investigated. 2. EXPERIMENTAL
2.1. Materials The Vectran® fibers used in this study were obtained from Celanese Advanced Materials (Charlotte, NC, USA). The DGEBA-based epoxy resin (DER 331) and tetraethylenepentamine (DEH 26) curing agent were obtained from Dow Chemical (Midland, MI, USA). HPLC-grade water, formamide and methylene iodide, used as probe liquids, as well as the DETA, a liquid amine, were obtained from Aldrich (Milwaukee, WI, USA). Quartz coverslips, 2.54 cm ¥ 2.54 cm, were obtained from Electron Microscopy Sciences (www.emsdiasum.com). 2.2. Specimen preparation and laser treatment Vectran® fiber samples were irradiated while immersed in DETA, using a Lambda Physik LPX 200 excimer laser apparatus. The laser apparatus was operated with XeCl gas which produces a wavelength of 308 nm. A schematic diagram of the experimental setup for laser treatment is shown in Fig. 2 [21]. The fibers were separately placed in a parallel array between two quartz coverslips and the gap between the coverslips was filled with DETA, fully immersing the fibers.
Figure 2. A schematic diagram of the experimental setup for laser treatment [21].
XeCl excimer laser treatment of Vectran fibers in DETA
91
Different fluences of 24, 36, 48 and 60 mJ/(pulse◊cm2) were used in this study. The laser pulse repetition rate was fixed at 1 Hz and each pulse was of 20 ns duration. For every laser treatment, the fiber specimens were first exposed to half the number of pulses from one side and the remaining half from the opposite side. The laser fluence was measured with a photodiode calibrated using Si crystal melt technique [28]. The UV spectrum of DETA, with two quartz coverslips, showed negligible absorption at 308 nm wavelength. After the laser treatment in DETA, the Vectran® fiber specimens were observed again to make sure that the DETA had not evaporated and the fibers were still immersed in DETA. The treated fibers were carefully washed successively in acetone and water to remove any residual DETA before performing various surface analyses. 2.3. Fiber surface characterization The surface topographies of untreated (control) and treated fibers were characterized using scanning electron microscopy (SEM) (Leica, Model 440X) and atomic force microscopy (AFM) (Digital Instruments, Model DimensionTM 3100). For SEM measurements, an acceleration voltage of 5 kV was used. Specimens were sputter coated with Au-Pd prior to SEM examination. For AFM measurements, images were acquired in the tapping mode and at least 7 specimens were characterized for each treatment condition. Due to the small diameter (high curvature) of the fiber, the whole fiber could not be imaged and only a 3 µm ¥ 3 µm of the top surface could be scanned for all specimens. Wettability properties of treated and control fibers were measured using the Wilhelmy technique with a TRI/Princeton wettability apparatus [21, 22, 25, 26, 29, 30]. In order to obtain the acid–base and dispersion components of the fiber surface energy, three different probe liquids: water, formamide, methylene iodide were used. The three-liquid wettability method has been described in detail by Good et al. [29] and Kamath et al. [30]. The elemental compositions of control and laser-treated Vectran® fiber surfaces were evaluated using X-ray photoelectron spectroscopy (XPS) (Surface Science Instruments, Model SSX100), with Al Kα X-rays (1486.6 eV) at 0 to 1000 eV and 55o take-off angle). 2.4. Fiber/resin interface characterization The microbead test was used to estimate the IFSS of Vectran® fiber/epoxy resin [31, 32]. The experimental details including schematic diagram of microbead test can be found elsewhere [31, 32]. Prior to the microbead test, the fiber diameter, d, and the microbead length, l, on the fiber were measured using an optical microscope. Epoxy resin DER 331 and the curing agent DEH 26 were mixed in stoichiometric ratio. A fine and stiff Si3N4 fiber was dipped into the epoxy mixture. Small amount of the epoxy mixture picked up by the tip of the Si3N4 fiber was transferred to the treated part of the Vectran® fiber, to form a round microdroplet (microbead) having diameter between 200 and 300 µm. The epoxy mi-
92
J. Zeng and A. N. Netravali
crobead on the treated fiber was cured at 80oC for 3 h in an air circulating oven. The oven was then turned off and the specimens were allowed to cool down slowly over 12 h, thus avoiding the buildup of internal stresses. The microbead test was carried out on an Instron Tensile Tester (model 1122) at a cross-head speed of 2 mm/min [31, 32]. The average IFSS, τ , was calculated using the following equation:
τ=
F , πdl
(1)
where F is the debonding force of the fiber/epoxy resin interface. It is assumed that τ is uniformly distributed along the entire length of the embedded fiber, i.e., the length of the bead. 3. RESULTS AND DISCUSSION
3.1. Fiber surface topography Figure 3 shows typical SEM micrographs of the control and laser-treated Vectran® fibers in DETA under different treatment conditions. The control fiber shows a smooth surface with only a few very shallow grooves along the fiber axis, as seen in Fig. 3a. Figure 3b–e shows the surface topographies after laser irradiation at fluences of 24, 36, 48 and 60 mJ/(pulse·cm2), respectively, with 75 pulses on each side (total of 150 pulses). Figure 3b shows a slightly rougher fiber surface compared to control fibers, with narrow and shallow grooves along the longitudinal direction. Figure 3c and 3d shows typical fiber surfaces after laser irradiation at fluences of 36 and 48 mJ/(pulse◊cm2), respectively, and show rather smooth surfaces without the fine grooves. This suggests melting at the surface resulting in smoothening. At the highest fluence of 60 mJ/(pulse◊cm2), the Vectran® fiber surface showed irregular nodular structure, as seen in Fig. 3e. All these surface characteristics seen in Fig. 3b–e for laser-treated fibers are significantly different compared to the surfaces of the Vectran® fibers laser-treated in air at corresponding fluences and pulses [27]. The characteristic periodic roll (wavy) structure seen in the case of Vectran® fibers laser-treated in air was not evident in any of the fibers treated in DETA. One reason for this is that DETA, a liquid amine, with a boiling point of 209oC, is a better conductor of heat than air. Although DETA does not absorb much UV radiation, it can conduct heat effectively, causing the surface temperature of Vectran® fiber to be much lower and more uniform than the corresponding surface temperature in the case of fibers laser-treated in air. As a result, the viscosity of the molten surface layer, if at all molten, is higher and it cannot overcome the viscous forces and flows freely to form the periodic wavy structure as in the case of treatment in air.
Figure 3. SEM micrographs of Vectran® fibers: (a) control; (b) laser-treated at a fluence of 24 mJ/(pulse◊cm2); (c) laser-treated at a fluence of 36 mJ/(pulse◊cm2); (d) laser-treated at a fluence of 48 mJ/(pulse◊cm2); (e) laser-treated at a fluence of 60 mJ/(pulse◊cm2). Fibers in (b), (c), (d) and (e) were treated with 75 pulses on each side.
XeCl excimer laser treatment of Vectran fibers in DETA 93
94
J. Zeng and A. N. Netravali
Figure 4. AFM images of surfaces of Vectran® fibers treated in DETA: (a) control, (Ra = 18.0 nm); (b) fluence: 24 mJ/(pulse◊cm2), Ra = 37.2 nm; (c) fluence: 36 mJ/(pulse◊cm2), Ra = 37.5 nm; (d) fluence: 48 mJ/(pulse◊cm2), Ra = 39.7 nm; (e) fluence: 60 mJ/(pulse◊cm2), Ra = 31.2 nm. Fibers in (b), (c), (d) and (e) were treated with 75 pulses on each side.
XeCl excimer laser treatment of Vectran fibers in DETA
95
XeCl laser has been used previously to treat UHMWPE fibers [25, 26]. The XeCllaser-treated UHMWPE fibers in DETA showed different surface topographies as compared to XeCl-laser-treated Vectran® fibers in DETA [26]. In general, with low number of pulses, such as 50 and 100 pulses (on each side) at all the fluences (200, 400 and 600 mJ/(pulse◊cm2)) used, the laser-treated UHMWPE fiber surface showed longitudinal groove structures, similar to what is seen in Fig. 3b. However, with increasing number of pulses, the grooves on the laser-treated UHMWPE fibers became wider and deeper, rather than being smooth as in the case of the laser-treated Vectran® fibers seen in Fig. 3c–e. Only at the highest fluence of 600 mJ/(pulse◊cm2) and with large number of pulses of 300 or 400 on each side of the fiber, the UHMWPE fiber surface melted and smoothed the surface. This is mainly because the energy absorbed by the UHMWPE fibers from the XeCl laser (308 nm) irradiation is much lower than that absorbed by Vectran®, due to its relatively lower absorption coefficient [33, 34]. AFM was used to quantitatively characterize fiber surface roughness at the nanoscale level. Figure 4 shows typical AFM images of the surfaces of control and laser-treated fibers with 75 pulses on each side at different fluences. The control fiber has a smooth surface, as shown in Fig. 4a. After 75 laser pulses, on each side, at the low fluence of 24 mJ/(pulse◊cm2), the fiber surface showed almost no change, as seen in Fig. 4b, compared to that of the control fiber. At 36 mJ/(pulse◊cm2) fluence and with 75 pulses on each side, the surface showed grooves and bumpy nodular structures (Fig. 4c). When the fluence was increased to 48 and 60 mJ/(pulse◊cm2), with 75 pulses on each side, the grooves disappeared. However, the nodular structures became more obvious as can be seen in Fig. 4d and 4e. To quantify the roughness of the laser-treated fiber surfaces, three separate roughness parameters, RMS, Ra and Rmax, were calculated from the AFM measurements [35]. The RMS (root mean square) value measures the standard deviation of the height values within the scanned area and is given by the following equation:
RMS =
∑(z
i
− zav )2 n
,
(2)
where zi is the height value at point i, zav is the average height value in the scanned area, and n is the number of points within the scanned area. Ra measures the arithmetic average of the absolute values of the surface height deviations measured from the mean plane and is given by the following equation:
Ra =
∑| z
i
− zav |
n
(3)
J. Zeng and A. N. Netravali
96
Rmax signifies the difference in heights between the highest (peak) and the lowest (valley) points within the image. Table 1 lists the RMS, Ra and Rmax values for control as well as laser-treated fibers. Figure 5 presents plots showing the relationship between Ra, fluence and number of pulses. From these plots, it can be seen that both fluence and number of pulses are significant factors in determining laser-treated fiber surface roughness. At the fluences of 24, 36 and 48 mJ/(pulse◊cm2), the Ra increased to between 30 nm and 44 nm from 18 nm for control fibers. The Ra values fluctuated around 35 nm with increasing the number of pulses above 30. This is because DETA is a good heat conductor, and at these low fluences, the heat generated by laser irradiation at the fiber surface is more efficiently conducted away by DETA and is dissipated to the outside environment, compared to air. After additional pulses, though the temperature at the fiber surface increases, the thermal conduction and dissipation rate also increases. As a result, the temperature of the fiber surface cannot increase much with increasing the number of pulses, and the Ra value does not increase significantly. At a fluence of 60 mJ/(pulse◊cm2), the Ra reached the maximum value of 45.2 nm at 30 pulses on each side, and then dropped to 35 nm with additional pulses. The reason is that at the fluence of 60 mJ/(pulse◊cm2) the heat generated is large and the temperature of the fiber surface increases with Table 1. Effect of XeCl laser treatments (in DETA) on Vectran® fiber surface roughness Pulse
RMS (nm)
Ra (nm)
Rmax (nm)
Control 24 24 24 24
Fluence
30 45 60 75
22.0 38.8 44.1 38.9 46.9
18.0 30.9 35.1 32.4 37.2
161.2 213.2 260.2 230.1 272.3
36 36 36 36
30 45 60 75
44.3 44.8 44.8 47.5
34.3 37.0 37.7 37.5
285.3 245.8 230.5 256.5
48 48 48 48
30 45 60 75
37.0 39.7 54.2 53.0
30.5 32.8 44.0 39.7
191.5 225.5 350.8 423.0
60 60 60 60
30 45 60 75
52.3 43.5 42.8 38.8
45.2 35.0 35.5 31.2
269.2 256.8 232.0 209.4
Fluence in mJ/(pulse◊cm2); pulse, the number of pulses on each side.
XeCl excimer laser treatment of Vectran fibers in DETA
97
additional pulses, resulting in more surface melting and smoothing. However, the temperature of the fiber surface cannot be as high as that of the laser-treated Vectran® fiber in air, again because of better heat conduction in DETA environment. As a result, the viscosity of the molten surface layer remains high and the characteristic periodic roll structure [23, 36] cannot develop even at the highest fluence used in the current study. In general, these results indicate that the Vectran® fiber surface changed from smooth, prior to XeCl excimer laser treatment, to rougher after the treatment. However, increasing the number of pulses beyond 30 (on each side) did not increase the surface roughness further at any of the fluences used in this study. The enhancement in surface roughness increases the specific surface (bondable) area which could improve the mechanical interlocking with the epoxy resin. Both increased roughness and specific surface area would be expected to improve the fiber/resin IFSS. The fiber diameter measurements showed no significant decrease indicating that the ablation was not severe. The laser pulse duration in this study was only 20 ns. In this short period of time, the surface temperature of treated fiber increases
Figure 5. The relationship of Vectran® fiber surface roughness (Ra) vs. XeCl laser fluence and number of pulses, in DETA.
98
J. Zeng and A. N. Netravali
to a very high level that could result in melting and/or ablation at the surface. However, the time-scale for the molten polymer to flow, in between two pulses, is much shorter, in the ms or µs range. The molten polymer may or may not flow, depending on the viscosity. Therefore, the heat diffusion and temperature decrease right after the laser pulse duration have to be considered. The heat conductivity values of DETA, Vectran® and air, as per CRC Handbook of Chemistry and Physics, are 0.22, 0.18 and 0.0262 W◊m-1◊k-1, respectively. Although DETA conductivity is similar to Vectran®, the liquid DETA can conduct away heat more efficiently. As a result, the surface temperature of the fiber treated in DETA is expected to be much lower than the fiber treated in air. The heat accumulation in these experiments was obvious from the fact that DETA boiled off after a few tens of pulses depending on the fluence. However, for 5 or less pulses, the fiber surface showed no noticeable change in surface topography. 3.2. Surface chemical composition analysis of Vectran® fibers by XPS The surface elemental composition of control and a few laser-treated fibers was investigated using XPS technique. The XPS spectra of control and laser-treated fibers are shown in Fig. 6. The XPS spectrum for the control fiber as seen in Fig. 6a shows two distinct peaks, corresponding to C1s and O1s, respectively. For fibers laser-treated in DETA, a small peak indicating nitrogen can also be found, as shown in Fig. 6b and 6c. Table 2 lists the elemental composition data obtained from control and lasertreated Vectran® fibers. It can be seen from Table 2 that the O/C ratio increased from 0.28 for the control fiber to 0.36 for the laser treatment at a fluence of 60 mJ/(pulse◊cm2) with 30 pulses on each side. The nitrogen percentage increased from zero for the control Vectran® fiber to up to 6% for laser-treated fibers, indicating that nitrogen containing groups were incorporated onto the fiber surface. The laser irradiation of Vectran® fibers induces photothermal degradation generating free radicals. These radicals can combine with the nitrogen in DETA and attach nitrogen-containing groups onto the fiber surface. After removing the fibers from DETA, the free radicals that have not decayed can combine with the oxygen in the air, and increase the O/C ratio on the fiber surface. Similar results showing free radicals combining with atmospheric oxygen have been obtained from plasma treatments of UHMWPE fibers in ammonia environment as well [37]. The binding energies of the deconvoluted peaks for C1s and O1s and the corresponding relative intensities (%) are also listed in Table 2. In general, for C1s, the 285±0.5 eV peak is assigned to C–H and C–C groups [21, 22, 25, 38–41]. The 286±0.5 eV peak is assigned to groups in which carbon forms single bonds to one oxygen or nitrogen, such as C–O and C–N. The 290±0.5 eV peak is assigned to groups in which carbon makes double bond with one oxygen such as C=O. The satellite peaks at binding energies of 295±0.5 eV result commonly from nonmonochromatic Al Kα X-ray source. The O1s peaks were deconvoluted into two
XeCl excimer laser treatment of Vectran fibers in DETA
99
Figure 6. XPS spectra of Vectran® control and laser-treated fibers in DETA. (a) Control; (b) 30 pulses on each side of the fiber at a fluence of 60 mJ/(pulse◊cm2); (c) 75 pulses on each side of the fiber at a fluence of 60 mJ/(pulse◊cm2).
100
J. Zeng and A. N. Netravali
XeCl excimer laser treatment of Vectran fibers in DETA
101
peaks at 531±0.5 eV and 533±0.5 eV [21, 22, 25, 38–41]. The two binding energies correspond to the C=O and C–O bonds, respectively. From Table 2, it can be seen that the amount of C–C/C–H decreased and the amount of C–O/O–C=O increased with more pulses after the laser treatment. The possible reason is that the C–C bonds in the polymer are broken down by the photon and intense heat from the laser energy, generating free radicals. These free radicals can combine with the oxygen and moisture in the air and form C–O/O–C=O bonds. These results are similar to those obtained for the thermal decomposition of PET, which show increase in the carboxylic acid groups after the laser treatment [15, 42]. 3.3. Wettability properties and acid–base interaction measurements Table 3 lists the work of adhesion, contact angle and surface energy and its components for both control and laser-treated fibers (in DETA). The dispersion and acid–base components of the surface energies were estimated by the Wilhelmy technique using three probe liquids [29, 30]. The work of adhesion and the contact angle were obtained in both advancing and receding modes. Hysteresis was calculated by subtracting the receding work of adhesion from the advancing work of adhesion, which signifies the extent of the surface chemical heterogeneity and the surface roughness [43]. Typical plots of the work of adhesion for control and for laser-treated fibers, in water, are presented in Fig. 7. The lower curve in every plot is for the advancing mode and the upper curve is for the receding mode. Comparing the work of adhesion curves for the control and laser-treated fibers and the corresponding data listed in Table 3, it is clear that the laser-treated fibers show slightly higher work of adhesion than the control fibers. This shows that the energy needed to separate the fiber and water increases after the laser treatment. The liquid epoxy resin mixture used in the present study is also polar, similar to water. Therefore, after the laser treatment, the energy needed to separate the fiber and epoxy would increase as well, indicating better adhesion at Vectran® fiber/epoxy resin interface. At the same time, the contact angle decreased with laser treatment, as shown in Table 3. A lower contact angle means a better spreading of the resin on the fiber surface, which indicates a larger specific contact area between the fiber and the resin. This would also help to increase the adhesion at the fiber/resin interface. The absolute value of the hysteresis for the laser-treated fibers, with the maximum value of 32.8 mJ/m2 at a fluence of 24 mJ/(pulse◊cm2) with 75 pulses on each side, in water, is much higher than 12.7 mJ/m2 obtained for control fibers. Control fibers show smoother surface as seen earlier in SEM micrographs (Fig. 3) and AFM images (Fig. 4). Control fibers are also nonpolar. Both these factors contribute to the lower hysteresis value in water. Lasertreated fibers, in general, showed larger hysteresis due to increasingly polar as well as rougher surface. Introduction of the polar groups increases the chemical heterogeneity of the fiber surfaces and enhances the hysteresis. It can be seen from Fig. 7 that there are two types of fluctuations in the work of adhesion plots, especially in the advancing mode. The first type of fluctuation has high amplitude
102
J. Zeng and A. N. Netravali
XeCl excimer laser treatment of Vectran fibers in DETA
103
Figure 7. Typical advancing (bottom) and receding (top) work of adhesion plots for control and laser-treated Vectran® fibers (XeCl laser treatment in DETA) in water: (a) control; (b) laser-treated fiber (fluence: 60 mJ/(pulse◊cm2), pulses: 75 on each side).
but low frequency, which is attributed to the surface roughness [21, 22, 25, 30, 38]. The second type of fluctuation which has low amplitude but high frequency
104
J. Zeng and A. N. Netravali
is due to the surface heterogeneity. Both these types of fluctuations are much larger in the case of treated fiber (Fig. 7b) compared to the control fiber (Fig. 7a), indicating that both fiber surface roughness and surface chemical heterogeneity increase significantly after the laser treatment. From the data presented in Table 3, it can be seen that, in general, the total surface energy for the laser-treated fibers did not change much, as a result of decrease in the dispersion component and almost equivalent increase in the acid– base component of the surface energy. These data differ from the data obtained for corresponding fibers laser-treated in air [27], which showed that the total surface energies of the laser-treated fibers were slightly lower than the total surface energy for the control fiber. After the laser treatments, at all fluences, in DETA, the dispersion component of the surface energy dropped from 41.3 mJ/m2 for the control fiber to around 30 mJ/m2. At the same time the acid–base component, γ AB , increased significantly from zero for the control fiber, indicating its hydrophobic nature, to the highest value of 18.0 mJ/m2 for fibers treated at a fluence of 60 mJ/(pulse◊cm2) with 75 pulses on each side. In general, the acid–base component as well as the hysteresis increased with increase in the number of pulses. The increase in the acid–base component of the surface energy after the laser treatment suggests that the fiber surface became more polar as was expected. The enhancement of fiber surface polarity can be attributed mainly to the incorporation of oxygen and nitrogen containing groups at the surface. From the XPS analysis presented in Section 3.2, the percentage of –O–C=O groups was shown to increase after the laser treatment. A similar phenomenon has been reported for excimer laser treatment of PET films, and a part of these groups has been shown to be –COOH groups [15, 42]. The increase in –COOH, as well as nitrogencontaining groups explains the increase in the acid–base component of the surface energy of laser-treated Vectran® fiber. The wettability results on UHMWPE fibers treated with XeCl laser in DETA have been somewhat similar [26]. After the laser treatment, the acid–base component of the surface energy of UHMWPE fiber increased, while the dispersion component of the surface energy decreased. However, the maximum acid–base component of the surface energy of XeCl laser-treated Vectran® fibers (in DETA) reached 18.0 mJ/m2 at a fluence of 60 mJ/(pulse◊cm2) with 75 pulses on each side. This value is much larger than the maximum acid–base component for XeCl laser-treated UHMWPE fibers of only 7.0 mJ/m2 at a significantly higher fluence of 600 mJ/(pulse◊cm2) with 400 pulses on each side [26]. The higher fluence and higher number of pulses needed for UHMWPE fibers is because of its significantly lower absorption coefficient at 308 nm wavelength. These results indicate that more polar groups are incorporated onto the Vectran® fiber surface compared to the UHMWPE fiber surface after the laser treatment. One of the reasons is that the polymer used for Vectran® fiber has more conjugated structures than polyethylene [1], which can stabilize the free radicals generated during the laser irradiation. These stabilized free radicals have a longer decay period and, thus, have
XeCl excimer laser treatment of Vectran fibers in DETA
105
higher probability of chemically combining with oxygen and/or nitrogen containing species in the environment. Another possible reason is that the absorption coefficient of Vectran®, as stated earlier, is much higher than that of polyethylene [26, 34] and, thus, the surface temperature of Vectran® fiber is much higher than that of polyethylene during the laser treatment. A higher temperature can usually generate more active species and free radicals, which can react with oxygen and nitrogen containing groups in the environment, thus making the fiber surface more polar. In addition, being polyester, Vectran® can combine with moisture present in the environment and hydrolyze to produce the –COOH groups. 3.4. Interfacial shear strength with epoxy resin In Fig. 8, the IFSS values for control and laser-treated fibers with the epoxy resin are plotted as a function of number of pulses, for all the fluences used in this study. The IFSS data were obtained using the microbead test [31, 32]. The average IFSS value of the control fibers with epoxy resin was measured to be 19.5 MPa. After the laser treatment, the IFSS values increased. The extent of improvement in IFSS is seen to depend both on the laser fluence and the number of
Figure 8. The relationship of Vectran® fiber/epoxy resin IFSS and the number of pulses at different fluences in DETA.
106
J. Zeng and A. N. Netravali
pulses. For fluences of 24, 36 and 48 mJ/(pulse◊cm2), the IFSS value increased, in general, with the increase in the number of pulses. A maximum of 28.8 MPa was obtained at the fluence of 48 mJ/(pulse◊cm2) for 75 pulses on each side. However, at the fluence of 60 mJ/(pulse◊cm2), the maximum IFSS value of 27.7 MPa was obtained when the fiber was irradiated with only 30 pulses on each side. Beyond 30 pulses, the IFSS decreased with increasing number of pulses. One of the reasons for the decrease in the IFSS at the high fluence of 60 mJ/(pulse◊cm2) was the melting at the fiber surface. As discussed earlier, the molten material spreads out, smoothing the fiber surface thus reducing the surface roughness. This decreases the mechanical interlocking between the fiber and the resin and reduces the IFSS. Figure 8 shows the dependence of IFSS on the number of pulses at different fluences. It can be seen that there is no clear relationship between IFSS value and fluence. Figure 9 shows the relationship between IFSS and Ra. It is evident from Fig. 9 that all the points fall along a single regression line. These results indicate that the improvement of the surface roughness of the laser-treated fibers is a major reason for IFSS enhancement. This phenomenon is different from the corresponding plots for XeCl laser-treated Vectran® fibers in air [27]. For Vectran® fibers laser-treated in air, the points were seen to follow two separate regression lines, which was attributed to different patterns of surface topographies resulting
Figure 9. The relationship between Vectran® fiber/epoxy resin IFSS and surface roughness (Ra).
XeCl excimer laser treatment of Vectran fibers in DETA
107
at different fluences. IFSS values for fibers treated at the fluences of 24 and 36 mJ/(pulses·cm2), which are lower than the threshold fluence, followed the line with a lower slope. The fiber surfaces were smooth or had some nodular structures. The IFSS values for fibers treated at the fluence of 60 mJ/(pulse·cm2), which is higher than the threshold fluence, followed a line with higher slope. These fiber surfaces showed periodic wavy structures with significantly higher Ra values, in the range of 60 nm compared to 18 nm for control fibers [27]. For Vectran® fibers laser-treated in DETA at all the fluences, none of the surfaces showed wavy structures, and thus all the IFSS values in Fig. 9 fall on a single line. Figure 10 shows the dependence of IFSS value on the acid–base component of surface energy. From the plot, it can be seen that there is no clear relationship between IFSS values and corresponding acid–base component values. However, there is a general increase in IFSS with increase in acid–base component value up to 10 mJ/m2. Beyond this, the IFSS does not increase with increase in the acid– base component value. Figure 11 shows the surface roughness (Ra) of lasertreated fibers versus the acid–base component of the surface energy. It can be seen that there is no clear relationship between the surface roughness (Ra) and the acid–base component of the surface energy. Figures 10 and 11 together with Fig. 9 suggest that the surface roughness of the laser-treated fiber is perhaps a more
Figure 10. The relationship between Vectran® fiber/epoxy resin IFSS and acid–base component,
γ AB , of the fiber surface energy.
108
J. Zeng and A. N. Netravali
important factor in improving the IFSS than the acid–base component of the surface energy under present treatment conditions. Figure 12 shows typical SEM micrographs after the microbead test for the fiber laser-treated at a fluence of 60 mJ/(pulse◊cm2) with 75 pulses on each side. During the microbead test, the fiber was pulled upward relative to the bead. Figure 12a shows the bottom side of the laser-treated fiber surface which remained unaffected after the microbead test. The fiber surface has similar roughness as what was seen in Fig. 3e. Figure 12b shows the microvise (top) side of the fiber surface after microbead test. Peeling of some fibrils by the bead is clearly evident in this case, indicating that the failure occurred within the fiber instead of at the fiber/bead interface. This was found to be true for all experimental conditions involving high fluence and large number of laser pulses. Similar results were also found for the Vectran® fibers laser-treated in air [27]. These results confirm that the fibril-to-fibril bonding of Vectran® fiber is weaker than the fiber/resin bonding after the excimer laser treatment and may be a controlling factor. While the fiber/resin IFSS increased with the laser treatment, the fibril-to-fibril interaction was limited and could not be increased with the laser treatment.
Figure 11. The relationship between the surface roughness (Ra) and acid–base component, the Vectran® fiber surface energy.
γ AB , of
Figure 12. SEM micrographs of the laser-treated Vectran® fiber after the microbead test. (a) The unaffected (bottom) side; (b) the microvise (top) side.
XeCl excimer laser treatment of Vectran fibers in DETA 109
110
J. Zeng and A. N. Netravali
4. CONCLUSIONS
Vectran® fibers were treated using XeCl excimer laser in DETA environment with different fluences from 24 to 60 mJ/(pulse◊cm2) and varying number of pulses from 30 to 75 (on each side). After the laser treatment: – Fiber surface roughness increased. However, increasing the number of pulses beyond 30, on each side, could not increase the surface roughness further, at all fluences used in this study, because of surface melting. – The O/C ratio of the fiber surface increased. In addition, carboxylic acid groups increased significantly after the laser treatment. The nitrogen percentage increased from zero for the control fiber to up to 6%. – The acid–base component of the surface energy increased significantly from zero for the control fiber to a maximum of 18 mJ/m2 at the fluence of 60 mJ/(pulse◊cm2) with 75 pulses on each side. However, the dispersion component of surface energy decreased at the same time. As a result, the total surface energy did not change much after the laser treatment. The hysteresis, however, increased significantly. The acid–base component of the surface energy increased with increase in the number of pulses, at all fluences. – Interfacial shear strength (IFSS) increased by up to 50% after the laser treatment. Most of this enhancement can be attributed to the increase in fiber surface roughness. However, IFSS seems to be limited by the interfibrillar adhesion within the fiber. Acknowledgements The authors would like to thank Professor Michael Thompson of Cornell University for allowing the use of XeCl excimer laser facility in his laboratory; College of Human Ecology, Cornell University for funding this research; and Cornell Center for Materials Research (CCMR) for the use of the facilities. REFERENCES 1. Hoechst Celanese Vectran Technical Brochure (1999). 2. H. C. Linstid, A. Kaslusky, C. E. McChesney and M. Turano, paper presented at The International Plastics Showcase, Chicago, IL (2000). 3. F. L. Matthews and R. D. Rawlings, Composite Materials: Engineering and Science. Woodhead, Cambridge (1999). 4. C. C. Chamis, Composite Mater. 6, 31-77 (1974). 5. A. N. Netravali, R. B. Henstenburg, S. L. Phoenix and P. Schwartz, Polym. Composites 10, 226241 (1989). 6. M. Nardin and I. M. Ward, Mater. Sci. Technol. 3, 814-826 (1987). 7. J. Heitz, E. Arenholz, T. Kefer, D. Bauerle, H. Hibst and A. Hagemeyer, Appl. Phys. A 55, 391392 (1992). 8. A. Hagemeyer, H. Hibst, J. Heitz and D. Bauerle, J. Adhesion Sci. Technol. 8, 29-40 (1994).
XeCl excimer laser treatment of Vectran fibers in DETA
111
9. Y. Pitton, S. D. Hamm, F. R. Lang, H. J. Mathieu, Y. Leterrier and J. A. E. Månson, J. Adhesion Sci. Technol. 10, 1047-1065 (1996). 10. S. Petit, P. Laurens, M. G. Barthes-Labrousse, J. Amouroux and F. Arefi-Khonsari, J. Adhesion Sci. Technol. 17, 353-368 (2003). 11. Q. T. Le, J. J. Pireaux and J. J. Verbist, Surface Interface Anal. 22, 224-229 (1994). 12. R. R. Zhao, L. C. Wadsworth, D. Zhang and C. Q. Sun, AATCC Rev. 3, 21-24 (2003). 13. R. Srinivasan, Science 234, 559-565 (1986). 14. T. Bahners, T. Textor and E. Schollmeyer, in: Polymer Surface Modification: Relevance to Adhesion, Vol. 3, K. L. Mittal (Ed.), pp. 97-123. VSP, Utrecht (2004). 15. S. Lazare and R. Srinivasan, J. Phys. Chem. 90, 2124-2131 (1986). 16. J. Y. Zhang, H. Esrom, U. Kogelschatz and G. Emig, J. Adhesion Sci. Technol. 8, 1179-1210 (1994). 17. J. E. Andrew, P. E. Dyer, D. Forster and P. H. Key, Appl. Phys. Lett. 43, 717-719 (1983). 18. Y. Novis, J. J. Pireaux, A. Brezini, E. Petit, R. Caudano, P. Lutgen, G. Feyder and S. Lazare, J. Appl. Phys. 64, 365-370 (1988). 19. T. Bahners and E. Schollmeyer, J. Appl. Phys. 66, 1884-1886 (1989). 20. T. Bahners, Opt. Quant. Electron 27, 1337-1348 (1995). 21. Q. Song and A. N. Netravali, J. Adhesion Sci. Technol. 12, 957-982 (1998). 22. Q. Song and A. N. Netravali, J. Adhesion Sci. Technol. 12, 983-998 (1998). 23. W. Wong, K. Chan, K. W. Yeung and K. S. Lau, J. Mater. Process Technol. 132, 114-118 (2003). 24. H. Watanabe and T. Takata, Angew. Makromol. Chem. 235, 95-110 (1996). 25. J. Zeng and A. Netravali, in: Polymer Surface Modification: Relevance to Adhesion Vol. 3, K. L. Mittal (Ed.), pp. 159-182. VSP, Utrecht (2004). 26. J. Zeng and A. Netravali, in: Contact Angle, Wettability and Adhesion Vol. 4, K. L. Mittal (Ed.), pp. 407-436. VSP, Leiden (2006). 27. J. Zeng and A. Netravali, J. Adhesion Sci. Technol. 20, 387-409 (2005). 28. K. K. Dezfulian, J. P. Krusius and M. O. Thompson, Appl. Phys. Lett. 81, 2238 (2002). 29. R. J. Good, M. K. Chaudhury and C. J. van Oss, in: Fundamentals of Adhesion, L.-H. Lee (Ed.), p. 454. Plenum Press, New York, NY (1991). 30. Y. K. Kamath, C. J. Dansizer, S. Hornby and H. D. Weigmann, Textile Res. J. 57, 205-213 (1987). 31. S. Luo and A. N. Netravali, J. Mater. Sci. 34, 3709-3719 (1999). 32. U. Gaur and B. Miller, Composites Sci. Technol. 34, 35-51 (1989). 33. J. Ashok, P. L. H. Varaprasad and J. R. Birch, in: Handbook of Optical Constants of Solids II, E. D. Palik (Ed.), Academic Press, Boston, MA (1991). 34. D. Knittel, W. Kesting and E. Schollmeyer, Polym. Int. 43, 231-239 (1997). 35. Nanoscope III, Command Reference Manual for Version 2.51, Digital Instruments (1993). 36. J. S. Rossier, P. Bercier, A. Schwarz, S. Loridant and H. H. Girault, Langmuir 15, 5173-5178 (1999). 37. Z. F. Li and A. N. Netravali, J. Appl. Polym. Sci. 44, 319-331 (1992). 38. A. Netravali, Q. Song, J. M. Caceres, M. O. Thompson and T. J. Renk, in: Polymer Surface Modification: Relevance to Adhesion, Vol. 2, K. L. Mittal (Ed.), p. 355. VSP, Utrecht (2000). 39. F. P. M. Mercx, Polymer 35, 2098-2107 (1994). 40. Z. F. Li, A. N. Netravali and W. Sachse, J. Mater. Sci. 27, 4625-4632 (1992). 41. C. R. Brundle and A. D. Baker, Electron Spectroscopy: Theory, Techniques, and Applications. Academic Press, New York, NY (1977). 42. T. Bahners, D. Knittel, F. Hillenkamp, U. Bahr, C. Benndorf and E. Schollmeyer, J. Appl. Phys. 68, 1854-1858 (1990). 43. R. E. Johnson, Jr. and R. H. Dettre, in: Wettability, J. C. Berg (Ed.), p. 531. Marcel Dekker, New York, NY (1993).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 113–125 Ed. K.L. Mittal © VSP 2007
Surface modification of polymers by ozone. Comparison of polyethylene and polystyrene treated at different temperatures TAKAOMI KOBAYASHI* and HIDETOSHI KUMAGAI Department of Chemistry, Nagaoka University of Technology, 1603-1 Kamitomioka, Nagaoka, Niigata 940-2188, Japan
Abstract—Surface modification of polymers by ozone (O3) was investigated using in situ FT-IR spectroscopy for polyethylene (PE) and polystyrene (PS). Gaseous O3 with 3026 ppm concentration was used to treat the polymers at different temperatures. It was found that the IR band assigned to C=O stretching appeared in the PE-O3 treated at 65°C, but the O3 treatment at 25°C showed a lower intensity of this band. This indicated that the PE segments were finally oxidized to form C=O groups. Mass spectrometry was employed to detect gaseous components of the O3 atmosphere, when PE was present. The presence of OH+ at m/z 17 indicated that dehydrogenation of PE occurred at 65°C. On the other hand, in the PS-O3 system, the results from the in situ FT-IR spectra indicated that the intensity of the C=O band was enhanced in the O3 atmosphere. Oxidation of PS by O3 was observed regardless of the temperature. In the PS-O3 system, the O3 contained the component O2+• at m/z 32 at a much lower signal intensity. These facts evidenced that O3 directly oxidized the PS ring and formed C=O groups via ozonide intermediate. Keywords: Ozone; polymer surface modification; ozonolysis process; polyethylene; polystyrene.
1. INTRODUCTION
Surface modification of polymers is important in a wide variety of industrial technologies [1, 2]. Particularly, active gaseous species have been widely used to modify surfaces of various polymers, since the treatment by such active species can considerably improve the surface properties. Among them, ozone (O3) is known as a convenient species for surface modification of polymers. This is because of the gaseous phase decomposing to safe O2 by thermal reaction: O3 → O2 + O. Hence, O3 is able to generate active oxygen for polymer surface modification. There have been reports on the reaction of O3 with different polymers: polypropylene [3, 4], polysiloxanes [5] and poly(methyl methacrylate) [6]. This is due to the fact that the O3 treatment creates hydrophilic functional groups on a poly*
To whom correspondence should be addressed. Tel.: (81-258) 47-9326; Fax: (81-258) 47-9300; e-mail:
[email protected]
114
T. Kobayashi and H. Kumagai
mer surface. Thus, the O3 treatment of polymers enhances oxidative reaction, which causes the formation of C=O groups. Rabek et al. [3] and Walzak and coworkers [4, 7] studied O3 treatment of polypropylene. Their data indicated that O3 treatment of polymers depended on the chemical structure of the polymer used. In their works, the mechanism of the reaction was based on atomic oxygen produced from O3. Also, they focused mainly on the UV-O3 effect to produce such reactive species from O3, since UV light effectively decomposes O3 to reactive species. It has been known for the alkane–O3 reaction in the gas phase that the O3 treatment of hydrocarbons produces intermediate radicals which decompose into molecular oxygen and hydroxyl radicals [7–9]. A similar reaction process might proceed in organic polymer systems, when the polymer surface is exposed to O3. However, there are less experimental studies on the reaction process of polymers by O3, especially for polyethylene (PE) and polystyrene (PS). Therefore, in the present study, we focused on O3 treatment of PE and PS at different temperatures. The reaction process was investigated by directly recording the FT-IR spectra of these polymers. Furthermore, the reaction process was followed by ultraviolet (UV) spectroscopy and mass spectrometry for the reaction gas of polymer treatment. 2. EXPERIMENTAL
2.1. Materials and sample preparation Before experiments, polyethylene (PE) (Merck, Germany, MW = 66 000) powder was dried for 24 h at 60°C in vacuum. For sample preparation, PE was heated at 150°C on a glass plate and the molten PE, sandwiched between glass plates, was pressed to form a film of about 150-µm thickness. A polystyrene (PS) (Nacalai Tesque, Japan, MW = 115 000) sample was prepared by the wet casting technique as follows. For a 5 wt% concentration, the PS pellets were dissolved in 50 ml tetrahydrofuran (THF) for 24 h at 40°C. Then, the polymer solution was cast on a glass plate (200 ¥ 200 mm2) at 25°C with about 100 µm thickness. Then, the glass plate with the cast solution was immersed in water and a white PS film was formed on the plate. After the solidification process, resultant polymer film was washed well with distilled water to remove THF solvent and dried at 40°C in vacuo for 24 h. 2.2. O3 generation and surface modification of polymers Figure 1a shows a schematic diagram of O3 flow system for polymer treatment. Gaseous O3 was generated by a silent discharge ozonizer (ZC60-MM, Silver Seiko, Japan), when oxygen (O2) flowed inside the ozonizer at 780 Torr (1 Torr = 133 Pa) at 1.0 l/min flow rate. Polymer sample (10 ¥ 10 mm2, 20-µm thick) was set on the reaction part made of Pyrex glass cylinder (inside the square in Fig. 1a) and O3 flowed. Since the glass cylinder had IR windows made of CaF2 (diameter 25 mm, 1 mm thick), the FT-IR spectrum of the film could be recorded in
Surface modification of polymers by ozone
115
Figure 1. Schematic diagram of O3 flow system for polymer treatment. In situ FT-IR spectrometer (a) and UV cell for O3 decay measurement (b) are included. P, pressure gauge; V, valve; F, flow meter; M, Mass detector.
116
T. Kobayashi and H. Kumagai
presence of O3. Here, the cylinder had an outer diameter of 20 mm and was 70 mm in height. The IR cell was loaded into the FT-IR instrument (FTIR-8200, Shimadzu, Japan) as indicated by the dashed rectangle in Fig. 1a. In order to measure the IR spectrum of the polymer film in O3, the IR beam was passed through CaF windows. The generated O3 was introduced into the IR cell with a controlled flow rate of O2 (99.9%, Taiyo Sanso, Japan). The total pressure of the O2 was measured by a pressure gauge (Baratron 122A). For constant 3026±20 ppm (1.35 ¥ 10-4 mol/l) concentration of O3 in the flow, four ozonizer units were connected in series. The O3 treatments were carried out at atmospheric pressure. The O3 concentration was monitored using a UV spectrophotometer (UV-190, Shimadzu). The UV cell (10 ¥ 10 ¥ 40 mm2) (Fig. 1b) was set in the upper stream of the reactor portion. In the absence of the polymer, the O3 Hartley band [10] at 254 nm was used to estimate the O3 concentration by monitoring the absorbance of the O3. It was confirmed by checking the O3 band that the O3 concentration was constant in the range of 3026±20 ppm. Furthermore, when the O3 concentration was measured with time in the absence and presence of polymer, the experimental set-up was used to monitor O3 decay. The UV cell (Fig. 1b) was set in the O3 line and kept inside with a pressure of 10-2 Pa. After introducing the O3 gas, the system was closed both in the absence and presence of polymer (20 ¥ 10 mm2) in the UV cell. Then, the absorbance of the Hartley band at 254 nm was monitored with increasing time at 25, 45 and 65°C. We decomposed the un-reacted O3 downstream at the exhaust portion; the gas flow was heated at 65°C in the presence of a catalyst (250 ml, SEKAD KW, Shinagawa Kasei, Japan). To analyze the O3-polymer reaction, FT-IR spectra were recorded with a resolution of 4 cm-1 in the spectral range from 1000 to 4000 cm-1. For each spectrum, 200 scans were averaged. X-ray photoelectron spectroscopy (XPS; XP-HSIG, Jeol, Japan) was employed for surface analysis of the treated polymers. In the measurements, a MgKα X-ray source (1253.6 eV) was used. For a qualitative analysis, curve fitting of XPS spectra was performed using the Gaussian distribution. 3. RESULTS AND DISCUSSION
In order to analyze the reactions of O3 with polymers, the FT-IR spectra of the polymers were recorded in the transmission mode. As shown in Fig. 1a, the IR radiation was passed directly through the polymer sample in the IR cell, which was filled with 3026 ppm of O3. Therefore, the in situ FT-IR spectrum of the polymer contained both polymer bands and O3 band near 1100 cm-1. Figure 2 shows the FT-IR spectra of PE (Fig. 2a) and PS (Fig. 2b) in the range of 1000–4000 cm-1. We recorded the IR spectra continuously at different exposure times to O3 for each polymer without exposure to air. In Table 1, the assigned IR bands are listed.
Surface modification of polymers by ozone
Figure 2. FT-IR spectra of (a) PE and (b) PS exposed to ozone at 65°C for various times.
Table 1. Assignment of FT-IR bands
PE PS
O3
Wavenumber (cm-1 )
Assignment
1481 1000–1200 1430–1625 1490 1600–2000 3010, 3080 1000
CH2 bending aromatic C–H in-plane ring C–C stretching C–H stretching (aliphatic chain) C–H out-of-plane aromatic C–H stretching
117
118
T. Kobayashi and H. Kumagai
Figure 2a shows the IR spectra of PE measured in the O3 flow at 65°C for 0, 1, 2 and 3 h. The characteristic IR bands at 2927 and 1481 cm-1 were assigned to C–H stretching and CH2 bending, respectively [11]. In the case of Fig. 2b for PS, characteristic IR bands were for aromatic C–H stretching at 3080 and 3010 cm-1, overtone and combination band due to C–H out-of-plane in the region 2000–1600 cm-1, aliphatic C–H stretching at 1490 cm-1, aromatic ring C–C stretching in the region 1625–1430 cm-1 and aromatic C–H in-plane in the region 1200–1000 cm-1. When the PS sample was treated in the O3 atmosphere, immediately the C=O stretching band appeared at 1748 cm-1. These data apparently showed that the O3–polymer reaction was significant in the case of PS.
Figure 3. Time dependence of FT-IR intensities of C=O stretching band for (a) PE and (c) PS, C–H stretching for (b) PE and (d) PS and phenyl C–C stretching for (e) PS.
Surface modification of polymers by ozone
119
In order to clarify the difference between PE and PS, the transmittance for each IR band was plotted with exposure time. Figure 3 presents time change of the IR intensities of C=O and C–H groups at 1721 cm-1 and 1430 cm-1, respectively. In the plot for PE (Fig. 3, curve a), the C=O transmittance showed only little change until the exposure was 0.5 h. Then, the value of the C=O transmittance decreased to 85% within 3 h. In the plot for PS (Fig. 3, curve c), the C=O band intensity at 1748 cm-1 decreased immediately from 90% to about 85%, when PS was exposed to O3 for 1 h. Furthermore, the IR intensity of the CH band for PS gradually increased as shown in Fig. 3, curve d. These results suggested that the formation
Figure 4. FT-IR spectra of (a) PE and (b) PS in the range from 2000 to 1500 cm-1. Exposure to ozone was for 3 h at various temperatures.
120
T. Kobayashi and H. Kumagai
Figure 5. Comparison of (C1s) XPS spectra of PE (a) and (b) and PS (c) and (d). The (b) and (d) spectra were obtained after ozonolysis for 3 h at 65°C.
process of the C=O group on PS was different from that on PE. We also obtained the time dependence of the IR band of the phenyl ring at 1600 cm-1; the band intensity increased by the O3 exposure (Fig. 3, trace e). This indicated decomposition of the phenyl ring of PS in the presence of O3. To investigate the difference between PE and PS, IR spectra were recorded between 25–65°C. Figure 4 shows the FT-IR spectra in region of 1500–2000 cm-1, when the polymers were treated with O3 for 3 h. In the C=O band for PE (Fig. 4, spectrum a), it was confirmed that there was no change in the IR spectrum at 25°C. As shown in Fig. 4, the treatment of PE at 65°C allowed the C=O band to appear on the PE after the polymer was exposed to O3 for 3 h. For PS (Fig. 4, spectrum b), the C=O band at 1721 cm-1 also appeared at 25°C. The C=O band intensity of the treated PS at 45°C was almost the same as that at 65°C. This fact indicated that the C=O formation in PS was independent of the temperature. In addition, the phenyl band in the 1600 cm-1 region decreased even if the treatment was carried out at 25°C. This also supported that the phenyl ring was attacked by O3 and decomposed regardless of temperature. Figure 5 shows C1s XPS spectra for untreated PE and PS (Fig. 5a and 5c) and after exposure (Fig. 5b and 5d) to O3 for 3 h. The C1s peak for PS (Fig. 5d) showed
Surface modification of polymers by ozone
121
Figure 6. Ozone decay (a) without and with (b) PE and (c) PS. Average initial concentration of O3 is 1.35 ¥ 10-4 mol/l (3026±20 ppm).
two components: the aliphatic C–C component at 285.0 eV and the aromatic C–C component at 284.9 eV [12–14]. The appearance of the shoulder band near 289 eV is assigned to C=O group. In PE–O3 for 1 h, the XPS spectrum obtained at 65°C was almost similar to that of untreated PE. When PE was treated for 3 h (Fig. 5b), the C=O band intensity increased. However, the tendency for C=O formation in the PE–O3 system was not as high as that of the PS–O3 system. It is very interesting to explain the temperature effect on O3 concentration in the absence and presence of polymers. Thus, O3 Hartley band at 254 nm was monitored with treatment time using the UV cell (Fig. 1b). The O3 decay in gaseous O3 profiles are shown in Fig. 6 in the absence and presence of polymers at different temperatures. The ozonolysis was carried out using a quartz UV cell
122
T. Kobayashi and H. Kumagai
(10 ¥ 10 ¥ 40 mm3), as shown in Fig. 1b. In the experiment, the UV cell without and with polymer was introduced in the O3 line (Fig. 1). The absorbance of the gaseous O3 was monitored when O3 flow was stopped by closing the cell top valves. As presented in Fig. 6a, O3 concentration reduced with increasing time. The comparison of O3 decay indicated that the reduction of the O3 band increased at higher temperature. This implied that O3 decomposition was enhanced at higher temperature. In Fig. 6b, the O3 decay was accelerated in the presence of PE. There was a similar tendency with change in temperature. That is, the high decay curve was also observed at 65°C. While the half-lives of O3 were 1.9 h at 25°C, 1.6 h at 45°C and 1.2 h at 65°C in the absence of polymer, the PE–O3 system showed half-lives of 1.3 h at 25°C, 0.96 h at 45°C and 0.42 h at 65°C. This means that self-decomposition of O3 increased with increase in temperature. According to the Arrhenius equation, the activation energy (Ea) was estimated from each decay data point recorded at different temperatures. For the thermal ozone decomposition, O3 → O2 + O, the value of Ea was calculated as 76 kJ/mol. The value obtained was close to that previously reported for O3 thermal decomposition, Ea = 83 kJ/mol [15]. On the other hand, the decay profile at 65°C suggested that O3 was consumed in the presence of PE. In the in situ FT-IR spectrum for the PE–O3 system, a strong C=O band appeared at 65°C. This fact confirmed that the PE consumed O3 and then the C=O groups were formed on PE chains. To consider the PE–O3 reaction, it is important to know the reaction process of O3 with a hydrocarbon in the gaseous phase. In O3-ethane, the reaction was proposed to take place via a radical mechanism as follows:
C2 H 6 + O 3 → C2 H 5·+ HOOO·→ C2 H 5·+ O 2 +·OH Here, the HOOO ⋅ radical was formed by hydrogen abstraction from the hydrocarbon (dehydrogenation). Then, the radical finally decomposed into oxygen and hydroxyl radical HO ⋅ [8, 15–17]. Since the O3 decomposition process depends on temperature, it is considered that O3 → O2 + O reaction is enhanced at higher temperature. The O species formed and reacted with the main chain of PE, forming OH species [18, 19]. In contrast to the PE–O3 system, for the PS–O3 system, the decay of O3 was much faster than the decays as shown in Fig. 6a and 6b. The half-lives of O3 for the PS system were 0.21 h at 25°C, 0.15 h at 45°C and 0.12 h at 65°C. Thus, it was reasonable to consider that the phenyl ring of PS reacted directly with O3. This tendency well corresponded to the FT-IR result for the O3–PS system. The values of Ea calculated for the PE–O3 and PS–O3 systems were found to be 30 kJ/mol and 20 kJ/mol, respectively. The comparison between PE and PS suggested that the ozonization route for the PS–O3 system was via a lower activation barrier. For example, it is well known that O3 reacts with unsaturated carbon bonds, such as –C=C– group, through ozonide intermediate [9, 20, 21]. Thus, the reaction
Surface modification of polymers by ozone
123
Figure 7. Mass spectra of O3 flow in the absence and presence of polymers. The data of (a) and (b) were determined at 65°C in the absence of polymers for oxygen (a) and ozone flow (b). Figures (c) and (d) were for untreated and treated PE and (e) and (f) for PS. The treatment time is indicated in each panel.
between O3 and the aromatic phenyl rings of PS might be caused through ozonide formation on phenyl ring. To obtain information on chemical species in the O3 flow, mass spectrometry was employed in the absence and presence of polymers. The gaseous sample was introduced into the vacuum chamber equipped with a mass detector (M, indicated in Fig. 1a) through a needle valve between the detector and O3 line. Figure 7 shows the mass spectra obtained before (Fig. 7a) and after (Fig. 7b) the silent discharge of oxygen. In the absence of polymers, mass peaks of m/z 32 and 16, assigned to O2+• and O+•, respectively, were observed. However, the intensity of O2+• at m/z 32 was significantly decreased relative to that of O+• at m/z 16 in the O3 flow in Fig. 7b. This strongly supported that the O3 flow contained oxygen atoms generated from O3 by self-decomposition: O3 → O + O2. Figure 7d and 7f presents the mass spectra observed in the presence of PE and PS, respectively. On comparing the mass patterns of Fig. 7d and 7f, the intensity of m/z 17 apparently increased in the presence of PE. The difference means that the O species reacted with PE to generate OH species. In addition, the m/z 2 and 18 components were
124
T. Kobayashi and H. Kumagai
monitored for the PE–O3 system, and the results suggested formation of H2+ and H2O+• in the gaseous flow, when PE was present. For the PS–O3 system (Fig. 7f), the mass spectrum of PS treated with O3 had peaks at m/z 2 (H2+), 16 (O+•) and 17 (OH+). It was noted that only a very weak peak corresponding to m/z 32 was observed in the system. The difference between the PS–O3 and the PE–O3 systems suggested that the oxidation of both polymer surfaces took place via different reaction paths. In the PS–O3 system, it was suggested that ozone consumption was very fast relative to that in the PE–O3 system. This indicated that O3 reacted with the phenyl group via a ozonide process [22– 25]. On the other hand, when O3 oxidizes the PE surface, the thermal decomposition of ozone might be the main process; O3 thermally decomposes to O2 and O. Since the oxidation of PE was strongly dependent on temperature, it was concluded that ozone decomposition was the rate-determining process for the oxidation of PE. 4. CONCLUSIONS
The present work focused on polymer-O3 reaction by employing in situ FT-IR technique. The C=O groups formed on both PE and PS were observed in the O3 flow. However, experimental data strongly indicated that different reaction processes occurred in the PS–O3 and the PE–O3 systems. It was found that the C=O band appeared in PS irrespective of temperature. This indicated that the C=O group was formed without the thermal decomposition of O3, which meants that direct reaction of O3 with phenyl ring of PS had occurred. In the PE–O3 system, the reaction between O3 and PE depended on the temperature. The data showed that a mass peak of OH+ was present in the presence of PE, indicating that the oxidation had proceeded via dehydrogeneration. Acknowledgements This work was partially supported by The 21st Century COE Program (Creation of Hybridized Materials with Super Functions and Formation of an International Research and Education Center) of Nagaoka University of Technology. REFERENCES 1. M. J. Wang, Y. I. Chang and F. Poncin-Epaillard, Langmuir 19, 8325 (2003). 2. M. P. Mendoza, M. D. Garcia and F. J. L. Garzon, Carbon 37, 1463 (1999). 3. J. F. Rabek, J. Lucki, B. Ranby, Y. Watanabe and B. J. Qu, in: Chemical Reactions on Polymers, J. L. Benham and J. F. Kinstle (Eds), pp. 187–200, Symposium Series No. 364. American Chemical Society, Washington, DC (1988). 4. M. J. Walzak, S. Flynn, R. Foerch, J. M. Hill, E. Karbashewski, A. Lin and M. Strobel, in: Polymer Surface Modification: Relevance to Adhesion, K. L. Mittal (Ed.), pp. 253–272. VSP, Utrecht (1996).
Surface modification of polymers by ozone
125
5. M. Ouyang, C. Yuan, R. J. Muisencer, A. Boulares and J. T. Koberstein, Chem. Mater. 12, 1591 (2000). 6. A. B. Ponter, W. R. Jones and R. H. Jansen, Polym. Eng. Sci. 34, 1233 (1994). 7. J. M. Hill, E. Karbashewski, A. Lin, M. Strobel and M. J. Walzak, in Polymer Surface Modification: Relevance to Adhesion, K. L. Mittal (Ed.), pp. 273–289. VSP, Utrecht (1996). 8. Q. K. Timerghazin, S. L. Khursan and V. V. Shereshovets, J. Mol. Struct. 489, 87 (1999). 9. D. Heymann, S. M. Bachilo, R. B. Weisman, F. Cataldo, R. H. Fokkens, N. M. M. Nibbering, R. D. Vis and L. P. F. Chibante, J. Am. Chem. Soc. 122, 11473 (2000). 10. R. W. B. Pearse and A. G. Gaydon, in: The Identification of Molecular Spectra, p. 149. Chapman & Hall, London (1950). 11. R. M. Silverstein and F. X. Webster, Spectrometric Identification of Organic Compounds, pp. 102–131. Wiley, New York, NY (1998). 12. D. B. Mawhinney and J. T. Yates, Jr., Carbon 39, 1167 (2001). 13. S. A. Carr and R. B. Baird, Water Res. 34, 4036 (2000). 14. G. Beamson and D. Briggs, High Resolution XPS of Organic Polymers. Wiley, Chichester (1992). 15. C. Park, J. Phys. Chem. 81, 499 (1977). 16. P. S. Nangia and S. W. Benson, J. Am. Chem. Soc. 102, 3105 (1980). 17. L. A. Hull, I. C. Hisatsune and J. Heicklen, Langmuir 14, 5813 (1998). 18. R. d’Agostino (Ed.), in: Plasma Deposition, Treatment and Etching of Polymers, pp. 329–334. Academic Press, Boston, MA (1990). 19. H. Suhr, Plasma Chem. Plasma Proc. 3, 1 (1983). 20. P. Neeb, O. Horie and G. K. Moortgat, Tetrahedron Lett. 37, 9297 (1996). 21. O. V. Manoilova, J. C. Lavalley, N. M. Tsyganenko and A. A. Tsyganenko, Langmuir 14, 5813 (1998). 22. Y. Hon and J. Yan, Tetrahedron 54, 8525 (1998). 23. C. C. Schubert and R. N. Pease, J. Appl. Phys. 78, 2044 (1956). 24. E. J. Feltham, M. J. Almond, G. Marston, V. P. Ly and K. S. Wiltshire, Spectrochim. Acta A 56, 2605 (2000). 25. G. D. Smith, E. Woods, III, C. L. DeForest, T. Baer and R. E. Miller, J. Phys. Chem. A 106, 8085 (2002).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 127–138 Ed. K.L. Mittal © VSP 2007
Atomic force microscopy based studies of photochemically-modified poly(ethylene terephthalate) surfaces THOMAS BAHNERS,1,* KLAUS OPWIS,1 ECKHARD SCHOLLMEYER,1 SHANG-LIN GAO2 and EDITH MÄDER2 1
Deutsches Textilforschungszentrum Nord-West e.V., Adlerstr. 1, D-47798 Krefeld, Germany Leibniz Institut für Polymerforschung, Hohe Str. 6, D-01069 Dresden, Germany
2
Abstract—The irradiation of a polymer by UV light effects photochemical surface modifications, if the photons are sufficiently absorbed by the substrate. A photochemical process can be considered as irradiation of the substrate in some reactive or inert atmosphere. In the presence of a low or nonabsorbing atmosphere and a strongly absorbing substrate, the actual reaction takes place only at the substrate surface, where radical processes are initiated. Four different types of reactions are possible: (I) recombination of radicals, (II) cross-linking of polymer chains, (III) addition of radicals from the reactive atmosphere and (IV) addition of bi-functional molecules resulting in cross-linking. The scope of this work was to study the occurrence of cross-linking of the polymer itself (reaction type II) and deposition of cross-linked thin layers (reaction type IV) following a photochemical surface treatment in the presence of bi-functional molecules. Topography measurements using atomic force microscopy (AFM) show a decrease in the tortuosity of the surface, which is seen as an indication of photo-induced molecular entanglement/cross-linking. Nanoindentation measurements confirmed that surface stiffness within the indentation depth was strongly affected by UV irradiation in the presence of different bi-functional media. Micro-thermomechanical (µTMA) measurements showed that the surfaces of samples irradiated in octadiene and argon had much less thermal expansion and lower softening/melting temperatures than the control sample, which is an indication of decrease in crystallinity because of the occurrence of cross-linking in the near-surface layer. In the case of film treated in the presence of diallylphthalate (DAP), depending on the local structure, either strong decrease of melting temperature or no melting point was found up to 300°C. Keywords: Photochemistry; cross-linking; polymer; AFM; nanoindentation; µTMA.
1. INTRODUCTION
In the field of technical textiles, both highly hydrophobic and oleophobic finishes are increasingly required to control the wetting properties of the textile or to attain barrier functions as well as self-cleaning properties. A number of physical surface *
To whom correspondence should be addressed. Tel.: (49-2151) 843-156; Fax: (49-2151) 843-143; e-mail:
[email protected]
T. Bahners et al.
128
modifications have been described in the literature, examples of which can be found, e.g., in Refs [1, 2]. The irradiation of a fibrous polymer using UV lamps effects photochemical surface modifications, if the photons are sufficiently absorbed. The energy of the photons in the deep UV, i.e., on the order of or below 200 nm, is sufficient to break bonds in the macromolecules. The use of monochromatic UV sources (e.g., excimer lamps), as reported for poly(ethylene terephthalate) (PET) modification [3–5], has the considerable advantage of allowing the choice of a wavelength which is highly absorbed by the polymer, resulting in radical generation with a very high quantum yield. According to the chemical composition of the PET macromolecule, backbone radicals have a very high probability to be generated, but have an extremely low lifetime in highly crystalline polymers, such as PET fibers, whereas side-chain radicals are less probable but long-lived. Effective surface modifications can be expected from treatments in reactive atmospheres. The general condition to achieve such reactions is a marked difference in the absorbance of a low or non-absorbing atmosphere and a strongly absorbing substrate. The actual radical generation and the ensuing reactions take place only at the very surface of the substrate which forms the boundary between the atmosphere and the activated substrate. Basically, four different types of reactions are possible (Table 1): (I) recombination of radicals, (II) cross-linking of polymer chains, (III) addition of radicals from the reactive atmosphere and (IV) addition of bi-functional molecules resulting in cross-linking between the functional groups (cf., Fig. 1). It has been shown that reactions III and IV could result in photo-induced grafting of functional groups and, thus, could be used to increase both wettability and fluid repellence with regard to water or oil [3, 4]. In addition, the modifications are extremely durable because of covalent bonding to a certain degree, as indicated by X-ray photoelectron spectroscopy (XPS) analysis (cf., e.g., Refs [3, 4]). However, only little is known about cross-linking effects and/or thin-layer deposition. As reported elsewhere [3], an indication for the formation of a layer was provided by studies using Fourier-transform infrared (FT-IR) spectroscopy Table 1. Possible reactions following UV lamp irradiation Type
Reaction type
Reaction scheme
Effect
I II
Recombination Reaction with radical(s) of a neighboring chain Reaction with reactive atmosphere Reaction with bi-functional substance
A1* + *A1 --> A1-A1
None
A1* + *A2 --> A1-A2
Cross-linking
A1* + Z --> A1-Z*
Addition
A1* + Z --> A1-Z*, A1-Z* + *A2 --> A1-Z-A2
Cross-linking, thin-layer deposition
III IV
PET molecules are denoted as Ai, and the reactive substance as Z.
AFM-based studies of photochemically-modified PET surfaces
129
and atomic force microscopy (AFM), but no definite evidence was found. A stronger indication was given by a study of the alkaline hydrolysis of PET fibers (textile fabrics) by Opwis et al. [6]. While storage of the as-received PET in caustic soda solution (25%) at 60°C resulted in total destruction of the fibers with an average weight loss of about 86%, a significant increase in the alkali resistance was found after UV irradiation in presence of various reactive media as well as under inert conditions. Using 1,7-octadiene or diallylphthalate (DAP), the weight loss was determined to be only 25 to 28%. Scanning electron microscopy (SEM) micrographs of the original PET fibers are shown in Fig. 2a (as-received) and 2b (after alkaline hydrolysis). A micrograph of a hydrolyzed fiber which had been irradiated in the presence of DAP before is given in Fig. 2c. Opwis et al. [6] also found that the hydrolytic decomposition of PET fibers was strongly reduced after irradiation in the presence of the hydrophilic compound cyclohexane-1,4dimethanoldivinylether (CHMV). This indicates that the increased resistance to hydrolysis cannot be related only to changes of wettability but may also be attributed to cross-linking at the polymer surface. With this background, the scope of this work was to study the occurrence of cross-linking of the polymer itself (reaction type II) and deposition of cross-linked thin layers (reaction type IV) following a photochemical surface treatment using excimer lamps. With regard to reaction type II, irradiations were performed in inert atmospheres, while allyl compounds served as reactive media for the generation
Figure 1. Potential process for a photochemically-induced generation of a thin layer (reaction type IV). The term double addition is used to describe the addition of a bi-functional molecule to two PET molecules which produces the cross-linking.
130
T. Bahners et al.
of reaction type IV. The reactive compounds were chosen based on their chemical, i.e., bi-functional, structure, allowing cross-linking reactions and a low absorption at the wavelength emitted by the UV lamp. In order to obtain analytical evidence, microscopic/nanoscopic analyses were performed. Besides AFM-based nanoindentation as a sensitive tool for stiffness/hardness/modulus determination of thin layer structures on fiber surfaces (cf., Refs [7, 8]), microthermomechanical analysis (µTMA) was used to characterize the thermomechanical properties of differently modified surfaces.
Figure 2. SEM micrographs of (a) as-received PET fibers, (b) fibers after alkaline hydrolysis in 25% NaOH solution (90 min at 60°C) and (c) photo-chemically modified fibers after alkaline hydrolysis (irradiation in DAP).
AFM-based studies of photochemically-modified PET surfaces
131
2. EXPERIMENTAL
Biaxially drawn PET film (thickness 0.1 mm, Goodfellow, Bad Nauheim, Germany) served as the sample material. The samples were extracted (Soxhlet, ethanol/hexane 20:80 (v/v), 4 h) and subsequently irradiated ‘as received’ or in the presence of bi-functional liquids. In the latter case, 1,7-octadiene (>97%, Fluka) and DAP (>98%, Acros Organics) were used as reactive media and applied before irradiation as follows. The samples were covered with a certain amount (1 ml/50 cm2) of the reactive medium. 1,7-octadiene was applied without any solvent, DAP was dissolved in ethanol (5 vol%). The fluid was then left to dry in air. A uniform wetting of the film surface proved to be a problem when using DAP. To overcome this, a second film was carefully laid on top of the sample for a short time in order to spread the reactive liquid. The irradiation of the various samples was done at 222 nm (KrCl lamp) in argon at temperatures below and above the glass-transition temperature of PET. After the treatment the samples were left in the argon atmosphere for about 3 min and then cleaned from residuals by extraction with petroleum benzine with the exception of the samples treated in DAP which were cleaned with ethanol. An AFM (Digital Instruments D3100, USA) was used both as a surface imaging tool and a nanoindentation device. The topography and phase images of the samples were acquired in the tapping mode, and phase shifts, i.e., changes in the phase angle of the oscillating probe relative to the phase angle of the probe cantilever oscillation, recorded simultaneously with height changes, are presented as a phase image. The RMS roughness (Ra) is calculated as the arithmetic average of the absolute values of the surface height deviations; surface area difference ratio (Sdr) expresses the ratio of the effective area of the rough surface to the scan area, i.e., Sdr ≥ 1. For the measurement of surface stiffness and modulus using AFM-based nanoindentation, small indentations in the specimen with the cantilever tip were made, while continuously recording the indentation force, F, and penetration depth, h, during one complete cycle of loading and unloading with a loading rate of 1 Hz. For all samples, several measurements were made at different locations to verify the reproducibility of the observed features. Local, i.e., confined to a certain small spot, micro-thermomechanical analyses of UV-treated films were performed using a Micro-Thermomechanical Analyzer (µTA 2990, TA Instruments, USA) [9]. In the µTMA mode, the probe is placed on the film surface with a certain force, which leads to a deflection of the cantilever. The probe is then heated at a rate of 10°C/s and its deflection recorded along the z-axis, which is perpendicular to the sample surface. When the material undergoes a phase transition, its mechanical properties change (softening), so that the tip will indent the film and the cantilever deflection will change accordingly, permitting detection of phase transitions. Topographic images of the analyzed area were made at a probe temperature of 60°C.
132
T. Bahners et al.
Figure 3. AFM phase images of (a) control PET film and PET films irradiated in the presence of (b) argon, (c) argon/1,7-octadiene and (d) argon/DAP. The UV irradiation was done at 90oC.
3. RESULTS AND DISCUSSION
AFM micrographs and phase images were recorded in order to study the surface topography of PET films (Fig. 3). There are apparent roughness and phase differences between the control and samples after UV irradiation at 90oC. Both roughness (Ra) and the ratio of actual area of the rough surface to scanned area (Sdr), related to the fractal dimension of the surface, decreased after UV treatment. The related decrease in the tortuosity of the surface indicates photo-induced molecular entanglement/cross-linking. When reactive bi-functional media, i.e., octadiene or DAP, were present during irradiation, the treated samples exhibited markedly smooth surfaces. The phase image emphasizes the different surface or underlying bulk features of the samples. In general, this can be in terms of stiffness, the viscosity of a polymer above its glass transition point, or adhesion between layers [10]. As shown in Fig. 3, the modifications resulting from treatments with the help of bi-functional media were observed to occur at PET surfaces on a nanometer scale in either homogeneous form, in the case of DAP, or showing smooth domains with local irregularities, presumably due to local stiffness variations in the near-surface region.
AFM-based studies of photochemically-modified PET surfaces
133
Figure 4. Exemplary indentation force F vs. penetration depth h curves. The value hmax denotes the maximum depth achieved during indentation with the corresponding indentation force shown as Fmax. The surface stiffness k is derived from the slope of the unloading curve.
Nanoindentation methods were employed to study whether the UV irradiation affected the elastic properties of near-surface polymer layer. The indentation force, F, recorded as a function of probe penetration depth can be used to calculate surface stiffness. An exemplary record of the cantilever z-displacement and deflection curve as the tip approaches/retracts a surface is shown in Fig. 4 [11]. The surface stiffness, k, is the slope of the initial elastic unloading curve, dF/dh, which includes contributions from both the specimen and the indenter (cf., Fig. 4). It can be seen from the graphs given in Fig. 5 that k increases with increasing indentation depth. The relationship is not linear because of continuous changes in the contact area for a parabolic shaped tip. It is clearly evident from Fig. 5b that surface stiffness within the indentation depth (<3 nm) was affected by UV irradiation in the presence of bi-functional media. A modified formalism based on the equation of Sneddon [12] allows a quantitative evaluation of the surface stiffness and modulus of samples Es. The relation between surface stiffness and modulus is given as
k = 2 2 Rh ⋅ E r
(1)
1 1 − ν i2 1 − ν s2 = + , Er Ei Es
(2)
with
134
T. Bahners et al.
where R is radius of curvature of the indenter tip and ν is Poisson’s ratio. The subscripts i and s refer to the properties of the indenter material and the specimen, respectively. Er is an effective reduced elastic modulus which includes contributions from both the specimen and the indenter. The properties of the silicon cantilever tip used were as follows: Poisson’s ratio νi = 0.23; Young’s modulus Ei = 162 GPa [13]. The value of the Poisson’s ratio νs for the PET film is 0.44 [14]. The best-fit of equation (1) to the experimental data gave moduli of 8.3, 8.6, 10.9 and 6.6 GPa for the control and after irradiation in argon, octadiene and DAP, respectively. In conjunction with the aforementioned phase imaging observation (Fig. 3), a comparison of these data suggests that the increase of surface modulus in the case of irradiation in octadiene is caused by the photochemically-induced cross-linking. In sharp contrast, the reverse is true, in the case of irradiation in
Figure 5. Variation of surface stiffness, k, as a function of indentation depth, h, for as received PET film (a), and after irradiation in Ar/octadiene (b). Best-fit curves were calculated using equation (1).
AFM-based studies of photochemically-modified PET surfaces
135
Figure 6. Displacement–temperature curves as recorded on untreated PET film (several measurements are shown). Thermal expansion occurs from point a to point b, softening of the film from point b to point c and melting above point c.
Figure 7. Displacement–temperature curves of PET after irradiation in octadiene atmosphere as compared to the control sample (several measurements are shown in both cases). Thermal expansion occurs from point a to point b, softening of the polymer from point b to point c and melting above point c.
136
T. Bahners et al.
DAP, while the modulus of the argon-irradiated sample is not significantly different from the control specimen. Considering the same bulk mechanical properties of all samples in this study, the implication is that in these cases the modified layer, if any, was very thin down to nanometer/atomic scale and the AFM indentation results do not reflect the elastic deformation of the outermost surface alone. In further measurements, µTMA was used to evaluate the thermomechanical properties of the control and UV-treated samples. During the AFM-based measurements, the contacting probe is heated and it deflects according to the thermal effects occurring. The deflection of the probe is recorded by the AFM cantilever displacement in the z-axis. A typical plot of the resulting cantilever (probe) displacement as a function of probe temperature is shown in Fig. 6. As the material under the probe is heated it expands, deflecting the probe upward (Fig. 6, from point a to point b). With further heating, the surface layer of the sample softens at some point leading to plastic deformation under the pressing probe (point b to point c) and finally melts, whereupon the tip of the probe sinks into the polymer (c to d). The analysis of the samples treated in octadiene (Fig. 7), as well as in argon, showed much less thermal expansion and lower softening/melting temperatures, which is an indication of decrease in crystallinity because of cross-linking in the near-surface layer. As can be seen in Figs 6 and 7, the results were highly reproducible in general. In contrast, the analysis of DAP treated film showed a wide variation in recorded curves, if the measurements were taken at different spots over the sample surface (Fig. 8). There was either a strong decrease of melting
Figure 8. Displacement–temperature curves of PET after irradiation in DAP atmosphere measured at five different spots over the sample area (see inset).
AFM-based studies of photochemically-modified PET surfaces
137
temperature or no melting point was found up to 300°C. At the same time, a large variation of the local structure of the sample surface was found (see inset in Fig. 8). This is possibly attributed to non-uniform cross-linking and/or thickness of the modified layer, arising from the initial non-uniform spreading of the reactive medium/solvent on the PET surface. In particular, the data recorded at points 1 and 2 on the surface (cf., Fig. 8) nevertheless indicate a strong crosslinking effect following irradiation in DAP at these places, since no thermal effect is observed within the temperature range of the measurement. As a final remark it should to be noted that an influence of the sample temperature during irradiation was found only in the case of irradiation in pure argon, where a treatment at room temperature did not affect the surface properties of the films. Given this background, all experiments reported here were conducted at 90°C sample temperature. 4. CONCLUSIONS
Cross-linking effects in the outermost surface layer of PET films and fibers following UV irradiation in argon and various bi-functional reactive media were studied using atomic force microscopy (AFM), AFM-based nanoindentation measurements and micro-thermomechanical analysis (µTMA). Topography measurements showed a decrease in the tortuosity of the surface, which is seen as an indication of photo-induced molecular entanglement/cross-linking. Nanoindentation confirmed that surface stiffness within the indentation depth of a few nanometers was affected by UV irradiation with different bi-functional media. The increase in surface modulus in the case of samples treated in octadiene is caused by the photochemically induced cross-linking. The experimental data taken from the samples which were irradiated in argon, as well as diallylphthalate (DAP), indicate that the modified layer, if any, was very thin down to nanometer/atomic scale and the bulk deformation was dominating in the indentation results. µTMA showed that the surfaces of samples irradiated in octadiene and argon had much less thermal expansion and lower softening/melting temperatures, which is an indication of decrease in crystallinity because of the occurrence of cross-linking in the near-surface layer. In the case of DAP-treated film, depending on the local structure, either strong decrease of melting temperature or no melting point was found up to 300°C. This is attributed to variations in cross-linking, as well as in thickness of the modified layer, arising from the initial non-uniform spreading of the reactive media/solvents on the PET surface. REFERENCES 1. K.L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 2. VSP, Utrecht (2000).
138
T. Bahners et al.
2. K.L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 3. VSP, Utrecht (2004). 3. D. Praschak, T. Bahners and E. Schollmeyer, Appl. Phys. A66, 69 (1998). 4. D. Praschak, T. Bahners and E. Schollmeyer, Appl. Phys. A71, 577 (2000). 5. T. Bahners, T. Textor and E. Schollmeyer, in: Polymer Surface Modification: Relevance to Adhesion, Vol. 3. K.L. Mittal (Ed.), pp. 97–124. VSP, Utrecht (2004). 6. K. Opwis, T. Bahners and E. Schollmeyer, Chem. Fiber. Int. 54, 116 (2004). 7. S.L. Gao and E. Maeder, A. Abdkader and P. Offermann: Langmuir 19, 2496 (2003). 8. S.L. Gao and E. Maeder, Composites A 33, 559 (2002). 9. H.M. Pollock and A. Hammiche, J. Phys. D 34, 23 (2001). 10. D. Sarid, T.G. Ruskell, R.K. Workman and D. Chen, J. Vac. Sci. Technol. B 14, 864 (1996). 11. S.L. Gao, E. Maeder and S.F. Zhandarov, Carbon 42, 515 (2004). 12. I.N. Sneddon, Int. J. Eng. Sci. 3, 47 (1965). 13. H. Guckel, Microelectr. Eng. 15, 387 (1991). 14. http://www.goodfellow.com/csp/active/static/E/Polyethylene_terephthalate.HTML
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 139–156 Ed. K.L. Mittal © VSP 2007
Wool surface modification and its influence on related functional properties P. JOVANCIC,1,* R. MOLINA,2 E. BERTRAN,3 D. JOCIC,1 M. R. JULIA2 and P. ERRA2 1
Textile Engineering Department, Faculty of Technology and Metallurgy, University of Belgrade, Karnegijeva 4, 11120 Belgrade, Serbia and Montenegro 2 Department of Surfactant Technology, IIQAB-CSIC, Jordi Girona 18-26, 08034 Barcelona, Spain 3 Applied Physics Department, Faculty of Physics, University of Barcelona, Avenida Diagonal 647, 08034 Barcelona, Spain
Abstract—A particular combined treatment consisting of specific fibre surface tailoring and activation by low-temperature plasma prior to biopolymer or enzyme post-application is described. Owing to selective modification of wool surface, low-temperature plasma treatment leads to the formation of new surface groups and also to the “step-by-step” removal of the F-layer. The low-temperature plasma/chitosan, as well as the low-temperature plasma/enzyme treatment are presented in this paper. These treatments are compared to the modern chemical treatment of wool combining liquid systems (peroxide, peroxide/enzyme) with a liquid biopolymer system (chitosan). All investigated treatments in this study result in an increase in wettability, dimensional stability and polymer adhesion. The dyeing and sorption properties of thus modified wool are also significantly improved. Keywords: Wool; surface modification; low-temperature plasma; peroxide; enzyme; chitosan; wettability; dimensional stability; dyeing behaviour; sorption properties.
1. INTRODUCTION
The surface properties of textiles are important determinants of their usefulness and many conventional chemical treatments are employed to modify these properties. Important properties of textiles, such as wetting, swelling, penetrability, dyeability and polymer adhesion, are strongly influenced by their surfaces. Also it is well known that there is a specific structure–property relationship in wool and other textile fibres. The wool is a protein, keratin, whose main chains are crosslinked with cystine residues, and also contains a variety of side chains, some of them being basic and some acidic depending on the side-chain groups present. The hydrophobic nature of the wool surface is attributed to the presence of 18*
To whom correspondence should be addressed. Tel.: (381-11) 3303-613; Fax: (381-11) 3370-387; e-mail:
[email protected]
140
P. Jovancic et al.
methylicosanoic acid (18-MEA), which is covalently bonded to the fibre as a thioester to cysteine residues of the wool protein. Fatty acid chains of 18-MEA are oriented away from the fibre to produce a “polyethylene-like” layer at the fibre surface, thus making the epicuticle hydrophobic and resistant to attack by different agents [1–3]. Therefore, by removing the covalently-bonded fatty acid monolayer, an increase in hydrophilicity of the fibre surface occurs, which could enhance dye uptake and polymer adhesion. Today, as a consequence of increased environmental awareness, the common “wet” chemical treatments of wool may soon need to be replaced by more favourable physical or chemical means for fibre surface modification. These treatments of textiles aim to obtain the required level of beneficial effect by exclusive modification of fibre surface, thus minimising fibre bulk attack and avoiding the deterioration in fibre quality. Moreover, the modern surface treatments are considered environmentally acceptable and are energy saving compared to conventional finishing methods. The main goal of wool surface modification is to impart shrink resistance at “machine washable” level as an end-user demands, but this effect cannot be exclusively achieved by a low-temperature, i.e., cold plasma (LTP) physical treatment [4–9]. Therefore, to improve the wool shrink resistance, an additional enzymatic treatment [10] or polymer deposition is required [5, 11–13]. The LTP treatment is the most commonly used method for a surface-specific fibre modification, as it affects only the surface, both physically and chemically, without altering the material bulk properties. Specific modification of the surface layer can be achieved by surface bombardment with ions, electrons, free radicals, metastables and UV photons which directly or indirectly participate in plasmachemical reactions. The chemical composition of the wool fibre surface is altered, facilitating the hydrophilic character of the fibre. This could be a consequence of the formation of Bunte salt and cysteic acid residues on the polypeptide chain [14]. Particularly, when oxidizing gases are used, plasma induces cystine oxidation in the A-layer of the exocuticle, converting it to cysteic acid and thus reducing the number of crosslinks in the fibre surface [15]. Although plasma species can penetrate to a depth of 50–100 nm into the fibre, which is deep enough to remove surface lipids, there are also some indications that the internal lipids of the cell membrane complex are also modified to a certain extent [16]. LTP treatments have been shown to be an effective means for modification of wool, cotton and flax surface properties [17–19]. The use of protease enzymes to achieve wool shrink resistance, better whiteness and improved handle is of considerable interest [20–23]. The term “handle” or “hand” has been defined in various ways. For instance, the handle can be defined as the quality of a fabric or yarn assessed by the reaction obtained from the sense of touch. Besides the subjective evaluation of handle, an effective fabric objective measurement system has been developed by Kawabata and co-workers (Kawabata Evaluation System for Fabric – KES-F). Shrink-resistant wool is the major priority, but if enzymes are applied at levels that provide the required
Wool surface modification and related functional properties
141
shrink resistance, wool fibres are often unacceptably damaged [24–26]. As the proteases generally have a large molecule, they preferentially attack the highly swellable cell membrane complex (CMC) by penetrating between cuticular scales causing stripping and weakening of the wool fiber [22]. To successfully confer shrink resistance without weakening of the wool fiber, an enzyme should act on the outer, sulphur-rich cuticular layer preferably after the wool has been modified to enhance this specificity [27]. Disulphide bond splitting owing to an oxidative or sulfite pre-treatment of wool [28–30] makes the wool fiber surface more accessible and, consequently, the enzymatic action on the cuticle is activated. As a consequence of oxidative pre-treatments, the wettability of wool is increased and thus the effectiveness of polymers used for shrink-proofing is improved [2]. It has been confirmed that improved wetting properties play an important role in the shrink resistance [31], especially when wool is subsequently treated with chitosan [32]. Furthermore, enzymes alone or in combination with peroxide are also successfully employed in wool bleaching [33], as an auxiliary agent in wool dyeing [34, 35] and for wool handle enhancement by reducing wool fibre stiffness [36, 37]. Much attention has also been focused on natural polymers as a possible substitute for synthetic polymers. Polysaccharide chitosan (CHT) offers great promise as its positive charge imparts it unique chemical and biological properties and thus makes it attractive for a wide range of textile applications. The solubility of CHT in acidic solutions makes it particularly suitable for industrial purposes. The chemistry of CHT is similar to that of cellulose, but it reflects also the fact that the 2-hydroxy group of the cellulose has been replaced with an acetamide group, which through N-deacetylation becomes a primary aliphatic amino group. In wool finishing, the CHT has been used as a shrink-resist agent [38, 39] as well as an agent for improving the dyeability of wool [40–42]. During the last decade, we have mainly investigated a number of novel treatment techniques with the purpose to obtain wool shrink resistance. The established concept has been to use LTP plasmas (both microwave and radiofrequency, using different plasma gases) and enzymatic treatment combined with peroxide treatment as a pre-treatment step, followed by CHT post-treatment to stabilise and enhance the effects obtained. In this paper, we also present some results of the enzymatic treatment applied after an initial specific wool surface “tailoring” by LTP treatment. The shrink resistance, wettability, swelling and contact angles have been measured [43, 44]. Fourier Transform Infrared (FTIR-ATR) and XPS analyses have been used to provide evidence about the chemical changes on the surface of the fibre [45, 46]. SEM observation and AFM analysis have been used to gain information about fibre topography [47].
P. Jovancic et al.
142
2. PHYSICAL TREATMENT
2.1. Low-temperature plasma treatment Both microwave (2.45 GHz) (MW) and radio-frequency (13.56 MHz) (RF) plasma treatments have been investigated in order to assess the influence of different types of plasma reactors on wool properties. The experience with these two types of plasma reactors confirmed that they both produced effective surface modification, but we were able to control plasma parameters much better when using RF plasma [48, 49]. The results on area shrinkage for RF plasma treated knitted wool fabric (Table 1) showed that very short N2 plasma treatment time (10 s) produce almost no effect on shrink resistance. However, the air plasma treated, as well as the H2Oplasma-treated samples provide similar area shrinkage values regardless of the treatment time. The plasma treatment times over 40 s do not induce further improvement, suggesting that a treatment time of 40 s was sufficient to increase the Table 1. Area shrinkage of LTP (RF)-treated knitted wool fabrics Sample
Area shrinkage (%) after two 5A wash cycles
Untreated N2, 100 W, 0.75 Torr, 10 s N2, 100 W, 0.75 Torr, 40 s N2, 100 W, 0.75 Torr, 120 s Air, 100 W, 0.75 Torr, 10 s Air, 100 W, 0.75 Torr, 40 s Air, 100 W, 0.75 Torr, 120 s H2O, 100 W, 0.75 Torr, 10 s H2O, 100 W, 0.75 Torr, 40 s H2O, 100 W, 0.75 Torr, 120 s
55.1 49.2 12.3 5.8 11.4 11.4 10.7 16.0 12.0 11.6
Table 2. Elemental concentration and atomic ratio for untreated, air, N2 and H2O plasma-treated wool Sample Untreated Air (40 s) N2 (40 s) H2O (40 s)
Elemental concentration (%)
Atomic ratio
C
O
N
S
Si
C/N
O/C
76.7 62.9 64.9 68.0
13.5 27.0 26.1 22.9
6.4 7.7 7.4 7.9
2.7 1.3 1.2 1.2
0.7 1.1 0.4 <0.1
12.0 8.2 8.7 8.6
0.18 0.43 0.40 0.34
The error associated with each measurement is <5% of the reported value.
Wool surface modification and related functional properties
143
wettability, confer shrink resistance and enhance polymer adhesion to wool fibre. These findings have been confirmed by advancing contact-angle measurements using water as the wetting liquid. A sharp decrease in contact-angle at short treatment time (10 s) suggested that the species present in the plasma reacted very fast with the fatty acid chains on the wool fibre surface. The amount of hydrophilic groups formed on the outermost epicuticle layer increases with treatment time up to 40 s [50]. This improvement in wetting properties of plasma treated wool is due to the existence of new hydrophilic groups on the fibre surface and modification, or even partial removal, of the fatty layer, which have been confirmed by XPS analysis [46]. Air, N2 and H2O plasma treatments result in a decrease in relative atomic concentration of C and an increase in relative atomic concentration of O (Table 2), suggesting that oxidation of the fatty layer present on the outermost part of the epicuticle occurred. Furthermore, the C/N atomic ratio decreases from 12.0 for untreated to 8.2, 8.7 and 8.6 for air, N2 and H2O plasma treated wool, respectively. This could be due to the partial removal of the hydrocarbon chains of the fatty layer [51]. The deconvolution of the C1s photoelectron peaks corresponding to air, N2 and H2O plasma-treated wool fabrics (Table 3) reveals a decrease in adventitious (Cadv) and aliphatic carbon (C–C, C–H), and an increase in C–O, C–N, carbonyl (C=O) and carboxylate (COO-) groups compared to untreated wool. The carbonyl group present at the wool surface has been attributed to amide group of the proteins located below the fatty layer. This indicates that the proteins are more detectable by XPS after plasma treatment due to material removal from the fatty layer present on the wool surface [11]. Splitting of some cystine disulphide linkages occurs due to an oxidative effect of air, N2 and H2O plasmas resulting in formation of sulphonic groups that could be easily detected by XPS as S2p peaks [45, 46]. The wettability, shrink resistance and polymer adhesion properties tended to decay with time after RF plasma treatment, mainly for wool treated for short times [52]. Regardless of the plasma gas used, the advancing contact angle Table 3. Relative intensity data from deconvoluted C1s spectra of untreated, air, N2 and H2O plasma-treated wool Sample Untreated Air (40 s) N2 (40 s) H2O (40 s)
Concentration (%) Cadv
C–C, C–H
C–O, C–N
C=O
COO-
15.2 5.7 8.8 2.7
57.4 51.3 47.3 56.3
18.0 21.7 22.7 22.0
8.4 17.0 17.0 14.3
1.0 4.3 4.2 4.7
P. Jovancic et al.
144
increases as a function of time elapsed after the treatment. The wettability decay rate was similar in case of wool treated for 40 s or 120 s, whereas it was faster for the fibres treated for 10 s. Results show that after 30 days of ageing the advancing contact angle achieves a value that is practically independent of the plasma exposure time, suggesting that the groups on the fibre surface have been totally reoriented [45]. 3. COMBINED PHYSICO-CHEMICAL TREATMENTS
3.1. Low-temperature plasma-chitosan treatment The wettability decay was the principal reason that we focussed on immediate after-treatment with chitosan in order to stabilize the effects obtained by the plasma treatment and to attain the maximum efficiency of the treatment [53, 54]. In order to qualitatively evaluate chitosan adsorption on wool, CHT-treated wool was subjected to the staining process [53] that clearly showed considerably higher chitosan adsorption on plasma pretreated wool compared to untreated wool. This has been additionally confirmed with contact angle measurements. The advancing contact angle tends to decrease with LTP treatment and the hydrophobic untreated fibre surface (θADV>90o) becomes hydrophilic (θADV<90o). Table 4 shows [55] that the CHT treatment does not influence the contact angle in comparison with untreated fibre, but when the fibre is LTP+CHT-treated, the contact angle (78o) considerably increases with respect to the value of LTP-only-treated fibre. This confirms the presence of the chitosan on the fibre surface which remains more hydrophilic with respect to the untreated fibre, but more hydrophobic than the surface of the fibre treated with LTP alone. SEM did not show chitosan on the fibre surface even when high chitosan amounts were used (up to 1%), but the evidence for some interfibre chitosan bonding could be seen [53]. However, AFM section analysis suggests (Fig. 1) that the vertical distance, i.e. the scale height, decreases as a result of LTP treatment due to surface material removal or etching effect. The subtle decrease in scale height after CHT application also implies that chitosan forms a very thin film on wool fibres, but it is sufficient to provide additional shrink resistance. The LTP treatment alone or combined with CHT impaired the handle of the modified wool. Therefore, we explored the enzymatic treatment as a possible alternative to CHT treatment in order to improve handle properties of wool. Table 4. Advancing water contact angle (θADV) values for untreated, CHT, LTP and LTP+CHT treated keratin fibres measured by the Wilhelmy balance method
θADV (∞)
Untreated
CHT
LTP
LTP+CHT
102 ± 4
100 ± 2
55 ± 4
78 ± 4
Wool surface modification and related functional properties
145
Figure 1. AFM section analysis results for (a) untreated, (b) LTP and (c) LTP+CHT-treated wool.
3.2. Low-temperature plasma-enzyme treatment As previously explained, the area shrinkage (after two 5A wash cycles) of LTP treated wool noticeably decreases compared to untreated wool (Table 1 and Fig. 2). This effect could be attributed to changes such as roughening of fiber surface, modification or partial removal of highly hydrophobic F-layer present on the wool surface or oxidation of wool surface that increases the content of oxidized sulfur species [6–8, 10, 45]. The application of enzymes after LTP pretreatment easily imparts machine washability to wool regardless of the enzyme used. Low weight loss after LTP treatment confirmed that only surface of the fiber was affected chemically and physically without altering the bulk properties of wool [10]. After subsequent enzymatic treatment the weight loss tends to increase, thus a higher enzyme concentration and prolonged treatment time should be avoided despite the excellent shrink resistance achieved (Fig. 3). The wool damage could be easily controlled by proper choice of enzyme, enzyme dosage and treatment time. The XPS analysis monitored the surface chemical changes after LTP and LTP+enzyme treatments. Plasma treatment results in a decrease in relative atomic
146
P. Jovancic et al.
Figure 2. Area shrinkage after two 5A wash cycles for different treatments. Enzyme A is a biocatalyst based on selected enzymes which act specifically on protein fibres [56], enzyme B is a multipurpose formulation of wool-specific (proteolytic) enzymes [57].
Figure 3. Weight loss after LTP pre-treatment combined with enzyme A and enzyme B. Enzyme A is a biocatalyst based on selected enzymes which act specifically on protein fibres [56], enzyme B is a multi-purpose formulation of wool-specific (proteolytic) enzymes [57].
concentration of C and an increase in relative atomic concentration of O, suggesting oxidation of the fatty layer present on the outermost part of the epicuticle. LTP+enzyme treatment leads to a further decrease in C/N atomic ratio (Table 5). The values of the C/N ratio 4.1 for LTP+enzyme A treatment and 4.3 for plasmaenzyme B treatment, are similar to the reported C/N atomic ratio of 3.4 for the amino-acid analysis of the epicuticle [58] and to the C/N atomic ratio of 5.1 found for wool treated with potassium t-butoxide in t-butanol [59]. This could be attributed to a complete F-layer removal by enzymatic treatment applied after the LTP pretreatment.
Wool surface modification and related functional properties
147
Table 5. Elemental concentration and atomic ratios for untreated, enzyme A and enzyme B, LTP, LTP+enzyme A (1%, 60 min) and LTP+enzyme B (1%, 60 min) treated wool Sample Untreated Enzyme A Enzyme B LTP LTP+enzyme A LTP+enzyme B
Elemental concentration (%)
Atomic ratio
C
O
N
S
Si
C/N
O/C
76.7 75.4 74.9 68.0 58.6 59.4
13.5 14.0 13.3 22.9 23.1 22.6
6.4 7.9 8.6 7.9 14.2 13.8
2.7 2.2 2.4 1.2 2.9 3.2
0.7 1.1 0.6 <0.1 1.2 0.9
12.0 9.5 8.7 8.6 4.1 4.3
0.18 0.18 0.18 0.34 0.39 0.38
The error associated with each measurement is <5% of the reported value.
The deconvolution of the C1s photoelectron peak corresponding to plasmatreated wool reveals a decrease in the adventitious (Cadv) and aliphatic carbon (C– C, C–H), and an increase in C–O, C–N, carbonyl (C=O) and carboxylate (COO-) groups compared to untreated wool (Table 3). With the partial removal of surface lipids, the proteins beneath the lipid layer become more detectable by XPS after plasma treatment. The carbonyl group present on the wool surface has been attributed to amide group of the proteins located below the F-layer [45]. An increase of the carboxylic acid after the LTP treatment is a result of the oxidation of hydrocarbon chains located at the wool surface, which is in accordance with the decrease of the amount of covalently bonded surface lipids present in the wool. LTP+enzyme treatment additionally decreases the relative content of aliphatic carbon (C–C, C–H) and carboxylate (COO-) groups and increases the relative content of the adventitious carbon (Cadv) and an unidentified carbon (Cunknown) peak on the wool surface compared to plasma treated wool alone [10]. The presence of amide peptide bonds is more obvious although some of them could be hydrolyzed as a result of subsequent enzymatic treatment. This implies that significant changes have occured on the wool surface as a consequence of enzyme treatment applied after the plasma pretreatment. The epicuticle or even exocuticle layer could be affected by plasma-enzyme treatment after the F-layer had been completely removed. The SEM images in Fig. 4 also indicated the changes in the appearance of wool fibre surface. From the relative intensity data from deconvoluted XPS S2p spectra we found that LTP-treated wool produced an increase in signal intensity at 168 eV, indicating oxidation of the disulfide bonds to the S6+ form as a result of formation of cysteic acid and probable presence of more than one sulfur oxidation species [6, 10]. After subsequent enzymatic treatment the intensity of the signal corresponding to oxidized sulfur species decreases significantly which could be related to the increase in the advancing contact angle after the LTP+enzyme treatment (Fig. 5).
148
P. Jovancic et al.
Figure 4. SEM micrographs of wool treated with LTP and/or enzymes. Enzyme A is a biocatalyst based on selected enzymes which act specifically on protein fibres [56], enzyme B is a multipurpose formulation of wool-specific (proteolytic) enzymes [57].
Figure 5. S2p XPS spectra of (a) untreated and LTP and (b) LTP+enzyme (1%, 60 min) treated wool. Enzyme A is a biocatalyst based on selected enzymes which act specifically on protein fibres [56], enzyme B is a multi-purpose formulation of wool-specific (proteolytic) enzymes [57].
Wool surface modification and related functional properties
149
Table 6. Advancing water contact angle (θADV) values for untreated, enzyme, LTP and LTP+enzyme-treated keratin fibres measured by the Wilhelmy balance method
θADV (∞)
Untreated
Enzyme A
Enzyme B
LTP
LTP+Enzyme A
LTP+Enzyme B
103 ± 3
92 ± 4
94 ± 2
59 ± 4
61 ± 4
62 ± 5
The results obtained from XPS analysis were confirmed by advancing contactangle determination (Table 6). The wool treated exclusively with enzymes exhibited poor shrink resistance due to lack of oxidation of the fatty layer, despite the surface material removal comparable to LTP treatment (Table 5). Therefore, the surface of enzymatically treated wool remains hydrophobic as the untreated one. Table 6 shows that enzymatic treatment applied after the LTP pretreatment does not influence significantly the contact angle with respect to the LTP-alone-treated fibre. Some increase in the contact angle after enzymatic treatment could be attributed to the elimination of oxidized sulfur species, basically cysteic acid generated as a result of LTP treatment (Fig. 5). 4. COMBINED CHEMICAL TREATMENTS
4.1. Peroxide-chitosan and peroxide-enzyme-chitosan treatments It is known that the H2O2-treated wool has a lower tendency to shrink compared to untreated wool, owing to an increase in the inter-fibre repulsion as a consequence of cysteic acid formation [60]. Therefore, the purpose of the hydrogen peroxide pretreatment was to promote the formation of cysteic acid and, consequently, increase the anionic character of the fibre surface, which, in turn, could enhance chitosan adsorption. Additionally, it is likely that hydrogen peroxide pretreatment promotes a partial removal of the wool surface fatty acid layer [44], which then improves dyeing properties and leads to increased adhesion of polymers to wool. In any case, alkaline hydrogen peroxide pretreatment confers hydrophilicity to wool, while wool pretreated under acidic conditions has a wettability similar to that of untreated wool [32]. Chitosan application on H2O2-pretreated wool significantly improves the shrink resistance. The chitosan concentration in the treatment bath shows a clear influence on wool shrinkage obtained (Table 7). The increasing amount of chitosan on the wool as the concentration in the treatment bath increases has been confirmed also by a staining technique. As a consequence of treatment with H2O2, the wool is subjected to an oxidation reaction that improves whiteness and modifies its chemical and morphological structure [61]. In order to confirm cystine oxidation, the FT-IR–ATR absorbance ratios of cysteic acid versus the amide III of differently treated wool samples are
P. Jovancic et al.
150
Table 7. Area shrinkage and wetting time of untreated and H2O2-pretreated knitted wool fabrics subsequently treated with CHT (H2O2 pretreatment at pH 9.0, 1 h, 70oC) Sample
Area shrinkage (%) after two 5A wash cycles
Wetting time (s)
Untreated 0.25% CHT 0.5% CHT 0.75% CHT 1% CHT H2O2 H2O2 + 0.25% CHT H2O2 + 0.5% CHT H2O2 + 0.75% CHT H2O2 + 1% CHT
49 49 52 32 33 32 25 16 14 6
>2000 287 177 56 13 51 3 2 4 5
Wetting time was determined as the time needed for complete absorption of a drop of distilled water on the fabric sample, taking a mean value of 6 measurements.
Figure 6. IR/ATR 1040/1232 cm-1 absorbance ratio representing cysteic acid after different treatments of wool.
plotted in Fig. 6 [62]. The cysteic acid content remains approximately four times higher than the corresponding cysteic acid content of untreated wool. The incorporation of enzyme into the peroxide treatment bath results in a significant decrease in intensity of the signal corresponding to oxidized sulfur species [63] in a similar way as enzymatic treatment after the LTP pretreatment.
Wool surface modification and related functional properties
151
Figure 7. SEM micrographs of wool treated with (a) H2O2 (1 vol%, 55ºC), (b1–b3) H2O2 (1 vol%, 55ºC) in presence of 2% enzyme A and (c1–c3) 4% enzyme A. Enzyme A is a biocatalyst based on selected enzymes which act specifically on protein fibres [56].
The knowledge of the specific action of enzymes on substrates with a heterogeneous morphological structure and chemical composition, which are characteristics of wool, is still unsatisfactory. It is apparent that the results of enzymatic treatments, especially with proteases, can be unpredictable and may lead to an unacceptable degradation of wool fibre. Consequently, it is essential to restrict the enzymatic action to the wool fibre surface in order to avoid enzyme diffusion into the wool. In other words, the enzymatic action on wool must be carefully controlled [22]. The enzymatic treatments are only partially effective with respect to wool shrinkage reduction [24]. The results of FT-IR–ATR measurements [64] show that the enzymatically treated wool exhibits almost no changes in the redox state of –S–S– cystine bonds. The peak intensities of cystine monoxide and cystine dioxide are only slightly changed after enzyme treatments. Since the untreated wool had a certain amount of cysteic acid probably due to weathering or photo-oxidation [60], this content was slightly reduced on increasing the enzyme concentration. In order to achieve maximum efficiency of the H2O2 treatment with a minimum degradation of wool fibre, we combined peroxide treatment with enzymatic treatment and the presence of an enzyme in the H2O2 treatment was investigated [62]. Although enzyme supplier [56] claimed that the enzyme used in the same bath with H2O2 remained highly active, we noticed that its activity diminished sharply. It has been reported [33] that the enzyme used retains only 5% of its initial activity after
152
P. Jovancic et al.
15 min in the alkaline H2O2 treatment bath. Thus, the influence of an enzyme on whiteness, wettability and area shrinkage occurs during the first few minutes of the H2O2 treatment. This brief enzymatic treatment seems to be sufficient to alter the appearance of the wool fibre surface, or even remove some cuticle cells especially in the presence of 4% enzyme (Fig. 7). The stripping or removal of some of cuticular scales is not observed after the H2O2 treatment even at 70oC. The area shrinkage results (Table 8) clearly demonstrate the positive influence of the presence of an enzyme in H2O2 treatment. This influence is more pronounced after subsequent chitosan application, even at the lowest H2O2 concentration at 55oC. The shrink resistance of H2O2+enzyme-treated wool is particularly improved by the subsequent application of chitosan. The wool treated exclusively with an enzyme exhibited poor shrink resistance and almost no changes in wettability (Table 8). In any case, the level of shrink resistance achieved with alkaline peroxide treatment of wool in the presence of an enzyme can be significantly enhanced further by the application of the biopolymer chitosan. We assume that enzyme diffusion into the wool is minimized as а consequence of better enzyme adsorption on the wool fibre surface owing to the presence of new cysteic acid groups generated by H2O2 alkaline treatment. Therefore, enzyme Table 8. Area shrinkage and wetting time of untreated, enzyme (2%) and H2O2-pretreated (pH 9.0, 1 h, 55oC) knitted wool fabrics subsequently treated with 0.3% CHT Sample
Area shrinkage (%) after two 5A wash cycles
Wetting time (s)
Untreated CHT Enzyme Enzyme + CHT 1% H2O2 2% H2O2 3% H2O2 1% H2O2 + enzyme 2% H2O2 + enzyme 3% H2O2 + enzyme 1% H2O2 + CHT 2% H2O2 + CHT 3% H2O2 + CHT 1% H2O2 + enzyme + CHT 2% H2O2 + enzyme + CHT 3% H2O2 + enzyme + CHT
53 51 42 34 32 25 27 30 17 23 16 12 7.0 5.6 2.5 4.4
>9000 843 >9000 648 >6000 1680 1284 4665 113 46 298 234 167 127 111 97
Wetting time was determined as the time needed for complete absorption of a drop of distilled water on the fabric sample, taking a mean value of 6 measurements.
Wool surface modification and related functional properties
153
concentration at the wool surface could be increased and the enzyme action might also be limited to the wool surface because of its immobilization by ionic interactions between the newly formed sulphonic groups on the wool surface and the basic groups of the enzyme [62]. 5. SORPTION PROPERTIES OF MODIFIED WOOL
Wool dyeability [65–71] and sorption properties [72–74] are also noticeably improved by physical or combined physico-chemical treatments investigated. In order to assess the dyeing behaviour of modified wool, various factors, mainly pH and temperature, have been considered [42, 55, 71]. LTP + chitosan-treated wool showed improved dyeing properties [75]. This effect is due to increased capacity of treated fibre to adsorb dyes as a consequence of surface structural changes (LTP treatment) and/or creation of additional dye adsorption sites (chitosan treatment). The enhanced chitosan adsorption shown on the samples pretreated with LTP always leads to an improvement of dyeing properties. Given the enhanced sorption of acid dyes [74], metal ions [72] and oils [73], we have developed a recycled wool fibre based nonwoven material as a possible sorbent for the removal of certain pollutants from industrial effluents and contaminated water. The combined LTP and chitosan treatment of the recycled-wool-fibre-based nonwoven material increased the metal ions uptake, as shown in Table 9. Table 9. Sorption capacities for some metal ions after 24 h of sorption process at 20oC Metal ion Pb
2+
Cu2+
Zn2+
Treatment
Initial pH
q100
q1000
Untreated CHT LTP LTP+CHT Untreated CHT LTP LTP+CHT Untreated CHT LTP LTP+CHT
4.98 4.98 4.98 4.98 4.72 4.72 4.72 4.72 5.25 5.25 5.25 5.25
4.76 4.95 4.72 5.00 3.07 3.07 2.42 3.36 1.40 1.93 1.64 2.58
13.00 16.65 14.90 15.75 6.00 10.00 5.43 13.65 3.45 4.05 3.32 4.85
q100 = the equilibrium content of metal ions adsorbed onto the recycled-wool-fibre-based nonwoven material (mg/g) at an initial concentration of metal ions in solution of 100 mg/l; q1000 = the equilibrium content of metal ions adsorbed onto the recycled-wool-fibre-based non-woven material (mg/g) at an initial concentration of metal ions in solution of 1000 mg/l.
154
P. Jovancic et al.
The initial concentration, pH, temperature and mechanical agitation all significantly affect the sorption kinetics of metal ions and acid dyes. The loose form of the recycled wool fibre shows significantly higher sorption capacity for oils than the non-woven material. However, the recycled-wool-fibre-based non-woven material could be used as an efficient oil sorbent even after five cycles of sorption [73]. The recycled wool fibre based nonwoven material has the potential to be useful as a low-cost sorbent for treatment of industrial effluents and other water sources but still it has to be examined in detail. 6. CONCLUSIONS
Wool treatments by low-temperature plasma, peroxide, enzymes and chitosan cause significant changes in its surface chemical composition and topography. These changes provide improved functionality and besides the main benefit of obtaining dimensional stability there are other benefits, such as enhanced wettability, improved polymer adhesion and better sorption properties towards metal ions, acid dyes and oils. The sorption properties of recycled-wool-fibre-based nonwoven material treated with low-temperature plasma and/or chitosan imply that it could be commercially employed as an efficient sorbent. Additional research work is in progress. Acknowledgements The authors acknowledge the financial support for this work from EU project (INCO CT 2004-509188, Reduction of environmental risks, posed by emerging contaminants, through advanced treatment of municipal and industrial wastes). This research is also supported, in part, by the Project No. 7017B – Ministry for Science and Technology of the Republic of Serbia. Special thanks are due to the members of our research groups, who were involved in most of the experimental part of this work: Miss M. Radetic, Miss T. Topalovic (University of Belgrade); Mr. G. Viera (University of Barcelona); Miss S. Vilchez (UPC); Mrs. M. Dolcet, Ms. I. Muñoz, Miss E. Pascual (IIQAB-CSIC) and Mr. J.M. Fortuño (ICMCSIC). REFERENCES 1. J. D. Leeder, Wool Sci. Rev. 63, 3 (1986). 2. A. P. Negri, H. J. Cornell and D. E. Rivett, J. Soc. Dyers Colour. 109, 296 (1993). 3. A. P. Negri, H. J. Cornell and D. E. Rivett, Textile Res. J. 63, 109 (1993). 4. H. Höcker, Pure Appl. Chem. 74, 423 (2002). 5. W. Rakowski, in: Proc. 9th Int. Wool Textile Res. Conf., Biella, Vol. IV, pp. 359–368 (1995). 6. A. Hesse, H. Thomas and H. Höcker, Textile Res. J. 65, 355 (1995).
Wool surface modification and related functional properties
155
7. A. Pavlath and K. S. Lee, in: Proc. 5th Int. Wool Textile Res. Conf., Aachen, Vol. III, pp. 263– 274 (1995). 8. K. S. Lee and A. Pavlath, in: Proc. 5th Int. Wool Textile Res. Conf., Aachen, Vol. III, pp. 275– 285 (1995). 9. C. W. Kan, K. Chan and C. W. M. Zuen, Textile Res. J. 69, 407 (1999). 10. P. Jovancic, D. Jocic, R. Molina, M. R. Julia and P. Erra, AATCC Rev. 3, 25 (2003). 11. H. Thomas, Th. Merten, V. Monser and H. Höcker, in: Proc. 10th Int. Wool Textile Res. Conf., Aachen, H. Höcker and B. Küppers (Eds), PL-3, CD-ROM, pp. 1–8 (2000). 12. I. M. Zuchairan, M. T. Pailthorpe and S. K. David, Textile Res. J. 67, 69 (1997). 13. B. Jansen, F. Kümmeler, H. Thomas and H. B. Müller, in: Proc. 10th Int. Wool Textile Res.Conf., Aachen, H. Höcker and B. Küppers (Eds), PL-6, CD-ROM, pp. 1–6 (2000). 14. T. Klausen, H. Thomas and H. Höcker, in: Proc. 9th Int. Wool Textile Res. Conf., Vol. II, Biella, pp. 241–249 (1995). 15. C. W. Kan, K. Chan, C. W. M. Yuen and M. H. Miao, Textile Res. J. 69, 407 (1999). 16. K. S. Lee, Textile Res. J. 46, 779 (1976). 17. K. K. Wong, X. M. Tao, C. W. M. Yuen and K. W. Yeung, J. Soc. Dyers Colour. 116, 298 (2000). 18. T. Wakida, K. Taheda, I. Tanaka and T. Takagishi, Textile Res. J. 59, 49 (1989). 19. G. A. Burne and K. C. Brown, J. Soc. Dyers Colour. 88, 113 (1972). 20. H. R. Haefely, Textilveredlung 24, 271 (1989). 21. S. Fornelli, Melliand Textilber. 74, 120, E33 (1994). 22. E. Heine and H. Höcker, Rev. Prog. Color. 25, 57 (1995). 23. J. Cegarra, J. Soc. Dyers Colour. 112, 326 (1996). 24. A. Riva, J. Cegarra and R. Prieto, J. Soc. Dyers Colour. 109, 210 (1993). 25. P. Jovancic, D. Jocic and J. Dumic, J. Textile Inst., Part 1: Textile Sci. Technol. 89, 390 (1998). 26. P. Jovancic, D. Jocic, P. Erra, R. Molina and M. R. Juliá, Eur. Microsc. Anal. 51, 15 (1998). 27. R. Levene, Y. Cohen and D. Barkai, J. Soc. Dyers Colour. 112, 6 (1996). 28. L. Coderch, A. Pinazo and P. Erra, Textile Res. J. 62, 302 (1992). 29. R. J. Denning, G. N. Freeland, G. B. Guise and A. H. Hudson, Textile Res. J. 64, 413 (1994). 30. F. J. Douthwaite, D. M. Lewis and U. Schumacher-Hamedat, Textile Res. J. 63, 177 (1993). 31. I. M. Zuchairach, M. T. Pailthorpe and S. K. David, Textile Res. J. 67, 69 (1997). 32. M. R. Juliá, M. Cot, P. Erra, D. Jocic and J. M. Canal, Textile Chem. Color. 30, 78 (1998). 33. R. Levene, J. Soc. Dyers Colour. 113, 206 (1998). 34. L. Wiebke, E. Heine and H. Höcker, DWI Rep. 122, 509 (1999). 35. A. Riva, J. M. Alsina and R. Prieto, J. Soc. Dyers Colour. 115, 125 (1999). 36. D. P. Bishop, J. Shen, E. Heine and B. Hollfelder, J. Textile Inst., Part 1: Textile Sci. Technol. 89, 546 (1998). 37. R. Levene and G. Shakkour, J. Soc. Dyers Colour. 111, 352 (1995). 38. M. R. Julià, I. Muñoz, D. Brunsó, M. Cot, D. Jocic and P. Erra, J. Com. Esp. Deterg. 27, 519 (1997). 39. M. S. Masri, V. G. Randal and A. G. Pittman, in: Proc. 1st Int. Conf. on Chitin and Chitosan, Boston, MA, pp. 306–314 (1978). 40. B. Filipowska, A. Walawska and E. Rybicki, L’industrie Textile 1323, 79 (2000). 41. R. S. Davidson and Y. Xue, J. Soc. Dyers Colour. 110, 24 (1994). 42. D. Jocic, M. R. Julià and P. Erra, J. Soc. Dyers Colour. 113, 25 (1997). 43. Woolmark Test Method 31, Washing of Wool Textile Products. Woolmark, Ilkley (2000). 44. R. Molina, F. Comelles, M. R. Julia and P. Erra, J. Colloid Interface Sci. 40, 237 (2001). 45. R. Molina, P. Jovancic, F. Comelles, E. Bertran and P. Erra, J. Adhesion Sci. Technol. 16, 1469 (2002). 46. R. Molina, P. Jovancic, D. Jocic, E. Bertran and P. Erra, Surface Interface Anal. 35, 128 (2003). 47. P. Jovancic, D. Jocic, B. Tomcik, M. Radetic, Z. Lj. Petrovic, M. R. Juliá and P. Erra, Eur. Microsc. Anal. 74, 5 (2001).
156
P. Jovancic et al.
48. P. Erra, D. Jocic, M. R. Julià, A. Cuesta and J. M. D. Tascón, in: Proc. 7th Mediterranean Congress of Chemical Engineering, Barcelona, Book of Abstracts, p. 407 (1996). 49. R. Molina, P. Jovancic, D. Jocic, E. Bertrán, G. Viera, M. R. Julià and P. Erra, in: Proc. 10th Int. Wool Textile Res.Conf., Aachen, H. Höecker and B. Küeppers (Eds), CD-ROM, PL-P1, pp. 1–10 (2000). 50. D. Jocic, P. Jovancic, Z. Lj. Petrovic, E. Bertran, A. Navarro, M. R. Julia and P. Erra, in: Proceedings of the World Textile Conference, 2nd AUTEX Conference, Bruges, p. 299 (2002). 51. P. Jovancic, D. Jocic, M. Radetic, T. Topalovic and Z. Lj. Petrovic, Mater. Sci. Forum 494, 283 (2005). 52. R. Molina, M. R. Julia, A. Navarro and P. Erra, in: Proc. 8th Mediterranean Congress of Chemical Engineering, Barcelona, Book of Abstracts, p. 236 (1999). 53. P. Erra, R. Molina, D. Jocic, M. R. Julia, A. Cuesta and J. M. D. Tascon, Textile Res. J. 69, 811 (1999). 54. P. Erra, R. Molina, A. Cuesta, J. M. D. Tascon and M. R. Julià, in: Advances in Chitin Science, Vol II, A. Domard, G. A. F. Roberts and K. M. Värum (Eds.) (Proceed. 7th International Conference on Chitin Chitosan and Euchis’97) p. 791. (1997). 55. D. Jocic, S. Vilchez, R. Molina, P. Jovancic, A. Navarro, M. R. Juliá and P. Erra, in: Proc. 4th Conf. Eur. Chitin Soc. (Euchis 2001), May 2001, Senigallia (Ancona), R. A. A. Muzzarelli and C. Muzzarelli (Eds.), pp. 317–326 (2002). 56. Bactosol WO Liquid, Technical Information Bulletin, Sandoz, Basle (1993). 57. Biosoft PW, Product Data Sheet, T.S. Chemicals, Haslingden (1998). 58. N. Brack R. N. Lamb, D. Pham, T. Phillips and P. Turner, Surface Interface Anal. 27, 1050 (1999). 59. N. Brack, R. N. Lamb, D. Pham and P. Tarner, Surface Interface Anal. 24, 704 (1996). 60. A. Hesse, H. Thomas and H. Höcker, Textile Res. J. 65, 355 (1995). 61. J. Cegarra and J. Gacen, Wool Sci. Rev. 59, 1 (1983). 62. P. Jovancic, D. Jocic, R. Molina, M. R. Julia and P. Erra, Textile Res. J. 71, 948 (2001). 63. P. Jovancic, D. Jocic, R. Molina, M. R. Juliá and P. Erra, in: Proc. 10th Int. Wool Textile Res. Conf., Aachen, H. Höecker and B. Küeppers (Eds.), CD-ROM, EN-5, pp. 1–11 (2000). 64. P. Jovancic, D. Jocic, R. Molina, A. Manich, R. Sauri, M. R. Juliá and P. Erra, Indian J. Fibre Textile Res. 27, 408 (2002). 65. T. Wakida, S. Tokino and S. Niu, Textile Res. J. 63, 438 (1993). 66. H. Thomas, J. Herrling, W. Rakowski, R. Kaufmann and H. Höcker, DWI Rep. 111, 315 (1992). 67. H. Höcker, H. Thomas, A. Küsters and J. Herrling, Melliand Textilber. 75, 506 (1994). 68. A. Küsters, J. Herrling, H. Thomas and H. Höcker, in: Proc. 9th Int. Wool Textile Res. Conf., Biella, Vol. II, p. 403 (1995). 69. D. Ganssauge and H. Thomas, in: Proc. 10th Int. Wool Textile Res. Conf., Aachen, H. Höecker and B. Küppers (Eds.), CD-ROM, FI-7, pp. 1–11 (2000). 70. D. Jocic, S. Vílchez, T. Topalovic, R. Molina, A. Navarro, P. Jovancic, M. R. Julià and P. Erra, J. Appl. Polym. Sci. 97, 2204 (2005). 71. D. Jocic, S. Vílchez, T. Topalovic, A. Navarro, P. Jovancic, M. R. Julià and P. Erra, Carbohydr. Polym. 60, 51 (2005). 72. M. Radetic, D. Jocic, P. Jovancic, Lj. Rajakovic, H. Thomas and Z. Lj. Petrovic, J. Appl. Polym. Sci. 90, 379 (2003). 73. M. Radetic, D. Jocic, P. Jovancic, Lj. Rajakovic, Z. Lj. Petrovic and H. Thomas, Environ. Sci. Technol. 37, 1008 (2003). 74. M. Radetic, D. Jocic, P. Jovancic, Z. Lj. Petrovic and H. Thomas, Indian J. Fibre Textile Res. 29, 82 (2005). 75. D. Jocic, T. Topalovic, S. Vilchez, P. Jovancic, M. Radetic, Z. Lj. Petrovic, A. Navarro, M. R. Juliá and P. Erra, in: Proc. 10th Int. Wool Textile Res. Conf., Aachen, H. Höcker and B. Küppers (Eds.), CD-ROM, DY-2, pp. 1–11 (2000).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 157–170 Ed. K.L. Mittal © VSP 2007
Surface modification of polyethylene by photosulfonation SUSANNE TEMMEL,1,∗ WOLFGANG KERN2 and THOMAS LUXBACHER3 1
Polymer Competence Center Leoben GmbH, Parkstrasse 11, A-8700 Leoben, Austria Graz University of Technology, Institute for Chemistry and Technology of Organic Materials, A-8010 Graz, Austria 3 Anton Paar GmbH, A-8054 Graz, Austria 2
Abstract—The present work is focused on the surface modification of low-density polyethylene (LDPE) by UV irradiation with a Hg lamp in the presence of sulfur dioxide and air. This process, also called photosulfonation, results in introduction of sulfonic acid groups onto the polymer surface, as well as several micrometers into the polymer bulk. To characterize the modified LDPE surfaces, contact angle as well as zeta potential measurements were utilized. The contact angle θ of water (sessile drop) decreased from θ = 99° to approx. θ = 30°, indicating highly polar surfaces. The zeta potential ζ of the modified LDPE surfaces shifted to less negative values with increasing UV irradiation time. This result is explained by an increase of the hydrophilicity of the LDPE surface. Concomitantly, the isoelectric point shifted to lower pH values, which indicates an increasing amount of –SO3H groups present at the sample surface. To confirm the presence of sulfonic acid groups, FTIR spectroscopy was used to characterize the composition of the polyethylene samples before and after sulfoxidation. Also, LDPE samples were cross-linked by e-beam irradiation and then subjected to the photosulfonation process. Cross-linked LDPE offers a higher degree of modification with –SO3H groups. Moreover, the hydrophilic property of the modified polyethylene was studied as a function of ageing time (in air). We observed that the hydrophilicity of the photosulfonated LDPE surfaces was gradually lost during storage under ambient atmosphere. Within 30 h, the contact angle of water increased from θ = 30° to θ = 65°. This hydrophobic recovery is due to reorientation of polar groups from the surface into the subsurface layer. By storing under a polar atmosphere (48 h in the presence of CH3OH vapour), the polar surface properties of sulfonated LDPE could be recovered. The degree of hydrophobic/hydrophilic recovery was found to depend also on cross-linking before photosulfonation. Keywords: Polyethylene; cross-linked LDPE; surface modification; photosulfonation, zeta potential; contact angle; ageing; hydrophobic recovery.
1. INTRODUCTION
During the past few decades the use of polymers has found a remarkable success in many industrial fields, such as coatings, paints, inks and adhesives [1, 2]. Al∗
To whom correspondence should be addressed. E-mail:
[email protected]
158
S. Temmel et al.
though plastics have excellent physical and chemical bulk properties, very often they do not possess the surface properties required for different applications. Therefore, polymeric materials usually need a pre-treatment to achieve their surface properties for subsequent process steps without altering their bulk properties. For this reason, surface modification of polymers has become an important research area in the polymer industry [3, 4]. Conventional surface modification techniques [3, 4] such as flame treatment [1], corona discharge [1], plasma treatment [1, 5, 6] or irradiation with UV light in the presence of a UV-sensitive gas [7–10] introduce polar groups onto the polymer surface. This provides significant improvement in wettability, paintability, biocompatibility and also in adhesion to other materials. Sulfonation is the introduction of sulfonyloxy groups in organic molecules, either as sulfonic acid (–SO3H) or as sulfonate (–SO3Na). This results in altering the physical and chemical properties of molecules such as solubility in water, detergency, wettability, emulsifying power and foaming power, which are of great interest. Alkanesulfonates can be generated by two photochemical processes [11]: sulfochlorination and sulfoxidation [12–15]. In both methods a gaseous mixture of SO2 and an oxidant is used. In the sulfochlorination chlorine is used as the oxidizing agent, whereas in the sulfoxidation oxygen is the oxidant. The overall sulfoxidation reaction is represented as:
1 hυ → RSO3 H RH + SO2 + O2 2
(1)
The application of photosulfonation reactions to introduce –SO3H groups onto low-density polyethylene (LDPE) surfaces has already been described in the literature [16–19]. Both reaction pathways and optimization of reaction conditions (e.g., SO2/air volume ratio) have been discussed. The aim of the present study was a closer investigation of the properties of LDPE surfaces subjected to photosulfonation processes. In the present investigation such modified LDPE surfaces were characterized by FT-IR spectroscopy, contact angle, as well as zeta potential measurements. To improve the stability of modified LDPE surfaces, polyethylene samples were cross-linked by e-beam irradiation and subsequently photosulfonated. The surface properties of cross-linked LDPE are compared to those obtained with standard LDPE. In earlier studies [16– 19] the stability of photosulfonated polyethylene during storage in the air (as a non-polar medium) was not investigated. The present investigation also addresses the hydrophobic/hydrophilic recovery both for non-cross-linked and cross-linked types of LDPE.
Surface modification of polyethylene by photosulfonation
159
2. EXPERIMENTAL
2.1. Materials Low-density polyethylene (LDPE) foils (thickness 37 µm) were purchased from a commercial supplier. Some LDPE samples were cross-linked using electron beam irradiation (dose: 250 kGy) under inert gas conditions (N2). SO2 of >99.98% purity, as well as compressed air (nitrogen 78 vol%, oxygen 21 vol%, argon 1 vol%, krypton, neon, xenon < 0.1 vol%) were obtained from Linde AG (Graz, Austria). 2.2. FT-IR spectroscopy Infrared spectra were recorded on a Perkin Elmer Spectrum One FTIR spectrometer. Data collection was performed at 4 cm-1 spectral resolution in the region of 4000–450 cm-1 and averaged over 5 scans. All samples were measured in the transmission mode. 2.3. Contact-angle measurements The surface energy γ of the sample surfaces was determined by measuring contact angles θ with a Drop Shape Analysis System DSA100 (Krüss, Hamburg, Germany) using water and diiodomethane as test liquids (drop volume approx. 20 µl). Based on the Owens–Wendt method [20], the surface energy γ, as well as the dispersion and polar components (γD and γP) were evaluated. The literature values for the surface tension components of the test liquids used are given in Table 1 [21]. The contact angles were obtained using the sessile drop method and were measured within two seconds after deposition of the droplet. The reproducibility was within 2°. Table 1. Surface tension γ, dispersion component γD, polar component γP, density and viscosity for water and diiodomethane
Water Diiodomethane
γ (mN/m)
γD (mN/m)
γP (mN/m)
Density (g/cm3)
Viscosity η (mPa·s)
Temperature (°C)
72.8 50.8
21.8 50.8
51 0
0.998 3.325
1.002 2.762
25 20
Data taken from Ström et al. [21].
2.4. Zeta potential measurements The zeta potential ζ of the sample surfaces was determined by the streaming potential method, using a clamping cell connected to an EKA Electrokinetic Analyzer (Anton Paar, Graz, Austria). In this cell the sample is pressed against a poly(methyl methacrylate) (PMMA) spacer with rectangular channels. The meas-
160
S. Temmel et al.
urements were performed with a KCl electrolyte solution (10-3 M, 500 ml). The pH was adjusted to about 10 by adding NaOH (0.1 M, 2.5 ml) and was then decreased stepwise (0.3–0.4 units) by titration with HCl (0.1 M), until pH 3 was reached. Using the clamping cell, a pressure ramp from 0 to 40 kPa was employed to force the electrolyte solution through the cell along the rectangular channels of the poly(methyl methacrylate) spacer. Streaming potentials were converted to zeta potentials using the Helmholtz–Smoluchowski equation [22] and the Fairbrother– Mastin approach [23]. Each value of the zeta potential at a given pH value represents an average value over at least five individual measurements. 2.5. Cross-linking of polyethylene For additional investigations, LDPE foil samples were cross-linked by e-beam radiation with a dose of 250 kGy (in nitrogen). In order to remove the residual soluble components from the cross-linked LDPE samples, a suitable extraction process was developed. After clamping in a frame, the polyethylene foils were exhaustively extracted in m-xylene for 24 h at a temperature of 80°C. The addition of a small amount of BHT (butylated hydroxytoluene: 2,6-di-tert-butyl-4methylphenol), an antioxidant, prevented unwanted oxidation of the LDPE foils. Afterwards the samples were dried at room temperature for 20 h, and then dried in vacuo (2 kPa, 48 h at 40°C) to constant weight. 2.6. Irradiation setup The LDPE samples to be irradiated were placed in a reaction chamber made of stainless steel (type 316 Ti) equipped with a quartz window in the top cover, a gas inlet and a gas outlet. During irradiation experiments, a gas stream composed of sulfur dioxide and compressed air at a volume ratio of 1:1 was passed through the
Figure 1. Emission spectrum of the Hg lamp as recorded with the spectroradiometer. The power density was measured at the sample position.
Surface modification of polyethylene by photosulfonation
161
reaction chamber (open system, atmospheric pressure). The composition of the gas mixture was controlled with two flow meters. The total flow was about 5 l/h in all cases. Irradiation was carried out with a medium pressure mercury lamp (Heraeus, 1300 W). For all experiments, an etched (i.e., corrugated) quartz plate was positioned as a diffuser element between the lamp and the reaction chamber in order to provide a uniform light intensity over the whole sample area. The distance between the lamp and the sample surface was 7.2 cm. The radiation emitted by the medium pressure Hg lamp was determined by a UV spectroradiometer from Solatell (Croydon, UK) by measuring UV light in the range from 230 nm to 465 nm. The spectral distribution of the radiation is presented in Fig. 1. The light intensity integrated over the whole spectral range (230–465 nm) was about 70 mW/cm² (at the sample surface). 2.7. Irradiation procedure LDPE foil samples (cross-linked, as well as non-cross-linked types) were cleaned in an ultrasonic bath (10 min, 30°C) using a mixture of acetone/water and then dried in vacuo (2 kPa, 1 h at 40°C). The samples were positioned in the reaction chamber, which was then purged with a mixture of SO2 and compressed air (1:1, by volume) for about 10 min for conditioning. Subsequently, polyethylene samples were irradiated with the Hg lamp for different periods of time (1–5 min), while maintaining the gas flow (open system, atmospheric pressure). Finally the chamber was purged for an additional 3 min before opening it. After removal from the chamber the modified LDPE samples were extracted in an ultrasonic bath (10 min, 30°C) using a mixture of acetone/water (1:1) to remove soluble reaction products, excess SO2 and H2SO4. After extraction, the samples were dried in vacuo (2 kPa, 1 h at 40°C) and then subjected to characterization. 2.8. Ageing experiments The modified LDPE samples were stored at 20°C in air (with exclusion of light). At intervals, contact angles θ of water were recorded. The maximum storage time was 180 h. After 180 h, the samples were placed in a compartment which contained air saturated with methanol vapour (20°C). After 48 h of this additional treatment, the samples were dried in vacuo (2 kPa, 1 h at 40°C) and again subjected to contact-angle measurements. 3. RESULTS AND DISCUSSION
3.1. FT-IR results Depending on the process conditions, the depth of photomodification can be up to 7 µm [16]. FT-IR measurements were made to check the composition of unmodified and photosulfonated LDPE. The typical IR bands of LDPE (Fig. 2) are
162
S. Temmel et al.
Figure 2. FT-IR spectrum of the unmodified LDPE sample.
Figure 3. FT-IR spectrum of the e-beam cross-linked LDPE sample (prior to photosulfonation).
observed around 2929 and 2848 cm-1 (C–H stretching vibration), 1472 and 1377 cm-1 (C–H deformation vibration), and 730 and 720 cm-1 (–CH2– rocking vibration). The cross-linked LDPE sample was also analyzed by FT-IR spectroscopy. Its spectrum, shown in Fig. 3, is virtually identical to that of uncross-linked LDPE (cf., Fig. 2). The formation of carbonyl groups (approx. 1720 cm-1) due to the cross-linking process was not evidenced in the FT-IR spectrum. After photosulfonation new absorption signals with maxima at 1170 (asymmetric stretching vibration of SO3 units) and 1037 cm-1 (symmetric stretching vibration of SO3 units) evolve (Fig. 4). These signals are attributed to the sulfonic acid groups, which were generated via the UV treatment. The intensities of these two signals are strongly dependent on the irradiation time. They increase with
Surface modification of polyethylene by photosulfonation
163
Figure 4. FT-IR spectrum of an LDPE sample after UV irradiation (3 min) in the presence of SO2 and O2 and subsequent extraction with acetone/water.
increasing irradiation time but remain constant after 4 min of photosulfonation. The signal around 1720 cm-1 can be attributed to aliphatic aldehydes, which seem to be formed as a by-product during the UV treatment. 3.2. Contact-angle measurements The unmodified LDPE and polyethylene treated by UV irradiation for different times were characterized by contact-angle measurements. As shown in Fig. 5, the water contact angle θ on the unmodified non-cross-linked LDPE surface is very high (θ approx. 99°), indicating the hydrophobic character of this material. After
Figure 5. Contact angle θ of water on non-cross-linked (■) and cross-linked LDPE (○) as a function of the UV irradiation time under SO2/air atmosphere.
S. Temmel et al.
164
the UV treatment for four minutes, the contact angle decreased to θ approx. 30°, implying that photosulfonation made the surface more wettable. The contact angle did not decrease further, even when the irradiation time was extended to five minutes. The introduction of sulfonic acid groups is supported by the surface energy data (Fig. 6). Table 2 includes surface energies calculated from the water and diiodomethane contact angles on the polymers. The surface energy (γ) is made up of two components, the dispersion (γD) and the polar (γP). After the treatments, an increase in the polar component is observed, but the dispersion part of the surface energy remains almost constant. This result indicates that the modification results in formation of sulfonic acid groups on the surface, and that the enhanced wettability is caused by the interaction of the polar groups with water. The water contact angle on pristine e-beam-cross-linked LDPE was slightly lower (θ approx. 93°) than that on non-cross-linked polyethylene. This can be explained by the formation of oxidation products during the e-beam cross-linking
Figure 6. Photosulfonation of LDPE. Surface energy γ versus UV irradiation time for non-crosslinked (■) and cross-linked (○) polyethylene.
Table 2. Contact angle and surface energy data for non-cross-linked LDPE subjected to photosulfonation Photosulfonation time (min)
θ (°) for water
θ (°) for diiodomethane
γ (mJ/m²)
γD (mJ/m²)
γP (mJ/m²)
Surface polarity (%)
0 1 2 3 4 5
99.2 78.5 68.6 48.9 31.5 30.3
55.4 52.2 51.0 46.1 45.1 44.9
35.8 38.6 43.7 56.7 66.9 67.5
35.7 33.0 33.7 36.4 37.0 37.1
0.1 5.5 10.0 20.3 29.9 30.4
0.4 14.3 22.8 35.7 44.7 45.1
γ, surface energy; γD, dispersion component; γP, polar component.
Surface modification of polyethylene by photosulfonation
165
Table 3. Contact angle and surface energy data for cross-linked LDPE subjected to photosulfonation Photosulfonation time (min)
θ (°) for water
θ (°) for diiodomethane
γ (mJ/m²)
γD (mJ/m²)
γP (mJ/m²)
Surface polarity (%)
0 0.5 1 1.5 2 3 4 5
94.8 83.4 44.5 32.5 32.6 31.7 29.8 28.3
54.1 47.0 49.6 46.6 47.0 47.7 48.9 48.2
32.9 39.0 58.4 66.0 65.9 66.2 66.9 67.9
31.9 35.9 34.5 36.1 35.9 35.5 34.9 35.3
0.9 3.1 23.9 29.8 29.9 30.6 32.0 32.5
2.9 8.0 41.0 45.2 45.4 46.3 47.9 47.9
γ, surface energy; γD, dispersion component; γP, polar component.
process. Photosulfonation was then carried out in the same manner as with noncross-linked LDPE. Again, the incorporation of sulfonic acid groups is confirmed by the significant decrease in the water contact angle (Fig. 5). Compared to noncross-linked LDPE, the contact angle was found to be θ approx. 28° after a treatment time of only 90 s. This is explained by the stabilization of the surface layers by the cross-linking pre-treatment. Cross-link formation prevents extraction of soluble sulfonated polyethylene chains after UV irradiation during post-exposure extraction process with acetone/H2O. As a result, a higher amount of sulfonic acid groups is maintained at the surface at a given period of photosulfonation. The measured contact angles and the surface energy γ and its components γP and γD of the cross-linked and subsequently photosulfonated samples are listed in Table 3. 3.3. Zeta potential measurements The repeatability of all zeta potential measurements shown in this paper was within 2 mV. Figure 7 displays the zeta potential versus pH curves for unmodified and photosulfonated non-cross-linked LDPE. As for the unmodified polyethylene, the isoelectric point (IEP) around pH 3.9 and the shape of the curve are typical of polymers having no functional groups. Due to the hydrophobic nature of the LDPE, a preferred adsorption of chloride anions is observed which gives a negative zeta potential in the pH range 4–10. Compared to the untreated LDPE, all modified samples showed a marked shift of the isoelectric point towards lower pH and a shift of the zeta potential to less negative values. The shift of the isoelectric point can be explained by the presence of acidic groups (sulfonic acid groups) on the surface. The decrease in the zeta potential in the alkaline range can be explained by an increasing hydrophilicity of the modified LDPE which leads to increasing water adsorption and swelling during contact with the electrolyte solution. Consequently, the adsorption of chloride
166
S. Temmel et al.
anions onto the surface is decreased which results in a shift of the zeta potential to less negative values [24]. As shown in Fig. 7, the changes in the zeta potential of LDPE sulfonated for four and five minutes is quite small, but still noticeable. The zeta potential versus pH curves of unmodified and photosulfonated polyethylene cross-linked by e-beam are illustrated in Fig. 8. Compared to Fig. 7, a higher amount of sulfonic acid groups seems to be present on the cross-linked
Figure 7. Non-cross-linked LDPE: Zeta potential versus pH curves of unmodified LDPE (■) and photosulfonated LDPE (●, 1 min; □, 2 min; ▼, 3 min; ○, 4 min; ▲, 5 min of UV irradiation).
Figure 8. Zeta potential versus pH curves of e-beam cross-linked LDPE prior to (■) and after photosulfonation (●, 1 min; □, 2 min; ▼, 3 min; ○, 4 min; ▲, 5 min of UV irradiation).
Surface modification of polyethylene by photosulfonation
167
LDPE surface after photo-induced surface modification, since the zeta potential shifts to less negative values in the alkaline range. This effect can be explained by stabilisation of the surface during the pre-treatment cross-linking process. 3.4. Ageing of photosulfonated LDPE samples When a polymer is subjected to surface modification, it is important not only to obtain the minimum values of the contact angles, but also to preserve the enhanced wettability obtained (as expressed by the contact angle θ) for quite long time. In order to investigate the long-term stability of photosulfonated LDPE surfaces, the water contact angle θ was monitored as a function of ageing time in ambient air. LDPE samples (both cross-linked and non-cross-linked types) were photosulfonated for 4 min, extracted and dried as described in the Experimental section. The samples were then stored in ambient atmosphere at 20°C, and at intervals the contact angle θ of water was measured. For both types of samples, θ was approx. 30° at the beginning of the investigation (storage time defined as 0 h). We observe that for photosulfonated LDPE the major changes in θ occur within the first few hours of contact with the air (Fig. 9). After a steep exponential increase in θ the process slows down and reaches a plateau value after 70 h. For the non-cross-linked LDPE sample, the θ value did not exceed 70°, even after 180 h ageing in air at 20°C. The metastability of the modified polyethylene properties can be attributed to the following effect: in general, the wettability induced on a polymer surface after modification can decrease depending on the environment in which the treated samples are kept. This decrease in wettability has been attributed to the reorientation of the hydrophilic groups. When a photosulfonated sample is kept in air atmosphere, the hydrophilic groups tend reversibly to migrate
Figure 9. Photosulfonated polyethylene: dependence of water contact angle as a function of storage time in air for non-cross-linked LDPE (■) and cross-linked LDPE (○).
S. Temmel et al.
168
into or become oriented towards the bulk phase, thus altering the hydrophilicity of the surface. Such “hydrophobic recovery” is well known in the literature [25–28]. Photosulfonated cross-linked LDPE showed hydrophobic recovery to a lower extent than non-cross-linked LDPE, see Fig. 9. The contact angle θ increased from 27° up to 60° in 70 h. The lower water contact angle on these samples suggests that more hydrophilic –SO3H groups were still present at the polyethylene surface compared to non-cross-linked LDPE. This result and the lower aging rate can be explained by the limited diffusional mobility of the polyethylene chains in a densely cross-linked polyethylene. We investigated if the effect of “hydrophobic recovery” could be reverted for these samples by storage in a polar environment. Photosulfonated LDPE samples that had been stored in air for 180 h were subsequently placed in methanol atmosphere for additional 48 h. Non-cross-linked photosulfonated LDPE fully recovered the initial contact angle which had been recorded immediately after surface modification (θ approx. 30°). For cross-linked LDPE the contact angle θ changed to 40° after storage under methanol atmosphere (initial contact angle θ=28°). This indicates that less hydrophilic groups were present at the sample surface after storage in a hydrophilic environment. Again, this result can be explained by the restricted movement of polymer segments in crosslinked polyethylene. 4. CONCLUSIONS
–
–
–
Photosulfonation of LDPE by UV irradiation in the presence of SO2 and O2 results in the formation of sulfonic acid groups. The water wettability of the modified polyethylene samples increased with increasing irradiation time, but remained constant after prolonged UV irradiation. This result was evidenced by contact-angle measurements. Electrokinetic investigations on these surfaces showed that the zeta potential of photosulfonated polyethylene shifted to less negative values in the alkaline range with increasing irradiation time, indicating increased hydrophilicity. At the same time the IEP shifted to lower pH values, indicating the presence of sulfonic acid groups at the LDPE surface. Zeta potential measurements were found to be highly sensitive for this type of modified surfaces. The cross-linking of LDPE by e-beam pre-treatment prevents the extraction of soluble sulfonated polyethylene chains after UV irradiation during the post-exposure extraction process with acetone/H2O. This is caused by a radiation-induced network formation which results in higher amounts of sulfonic acid groups at the surface. This was evidenced both by zeta potential and contact angle measurements. Furthermore, it has been demonstrated that the sulfonated surfaces of LDPE samples largely revert to their initial hydrophobic character when exposed to air for extended periods of time. Cross-linking as a pre-treatment process re-
Surface modification of polyethylene by photosulfonation
169
duces the mobility of chain segments and partially stabilizes the surface against hydrophobic recovery. The effect can be reverted by storage in a polar atmosphere (methanol). Also the degree of hydrophilic recovery is dependent on the pre-treatment by cross-linking. Acknowledgements This study was performed at the Polymer Competence Center Leoben GmbH (PCCL, Austria) within the framework of the Kplus-program of the Austrian Ministry of Traffic, Innovation and Technology with contributions of Graz University of Technology (TU Graz, Austria), Anton Paar GmbH (Graz, Austria) and KE KELIT Kunststoffwerk GmbH (Linz, Austria). The PCCL is funded by the Austrian Government and the State Governments of Styria and Upper Austria. Thanks are also extended to ARC Seibersdorf Research GmbH (Dr. J. Wendrinsky) for performing the e-beam irradiation experiments. REFERENCES 1. C.-M. Chan, Polymer Surface Modification and Characterization, Hanser, Munich (1994). 2. C.-M. Chan, T.-M. Ko and H. Hiraoka, Surface Sci. Rep. 24, 1 (1996). 3. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 3. VSP, Utrecht (2004). 4. K. L. Mittal (Ed.), Polymer Surface Modification: Relevance to Adhesion, Vol. 2. VSP, Utrecht (2000). 5. K. Armbruster and M. Osterhold, Kunststoffe 80, 1241 (1990). 6. T. A. Giroux and S. L. Cooper, J. Appl. Polym. Sci. 43, 145 (1991). 7. J. C. Bevington and L. Ratti, Eur. Polym. J. 8, 1105 (1972). 8. H. Niino and A. Yabe, Appl. Phys. Lett. 63, 3527 (1993). 9. U. Meyer, W. Kern, M. F. Ebel and R. Svagera, Macromol. Rapid Commun. 20, 515 (1999). 10. M. Okoshi, M. Murahara and K. Toyoda, J. Mater. Res. 7, 1912 (1992). 11. M. Braun, M. T. Maurette and E. Oliveros, Photochemical Technology. Wiley, Chichester (1991). 12. L. Orthner, Angew. Chem. 62, 302 (1950). 13. R. Graf, Justus Liebig Ann. Chem. 578, 50 (1952). 14. F. Asinger, G. Geisler and H. Eckholdt, Chem. Ber. 89, 1037 (1956). 15. H. Harting, Chem. Ztg. 99, 175 (1975). 16. T. Kavc, W. Kern, M. F. Ebel, R. Svagera and P. Pölt, Chem. Mater. 12, 1053 (2000). 17. W. Kern and T. Kavc, Patent application EP 890227.2. (2001). 18. M. Kaneko and H. Sato, Macromol. Chem. Phys. 205, 173 (2004). 19. S. Temmel, W. Kern and T. Luxbacher, Prog. Colloid Polym. Sci. 132, 54 (2006). 20. D. K. Owens and R. C. Wendt, J. Appl. Polym. Sci. 13, 1741 (1969). 21. G. Ström, M. Frederiksson and P. Stenius, J. Colloid Interface Sci. 119, 352 (1987). 22. M. Smoluchowski, Handbook of Electricity and Magnetism, Vol. 2. Barth, Leipzig (1921). 23. F. Fairbrother and H. Mastin, J. Chem. Soc. 75, 2318 (1924). 24. K. Stakne, M. S. Smole, K. S. Kleinschek, A. Jaroschuk and V. Ribitsch, J. Mater. Sci. 38, 2167 (2003). 25. I. Banik, K. S. Kim, Y. I. Yun, D. H. Kim, C. M. Ryu and C. E. Park, J. Adhesion Sci. Technol. 16, 1155 (2002).
170
S. Temmel et al.
26. S. Bronco, M. Bertoldo, E. Taburoni, C. Cepek and M. Sancrotti, Macromol. Symp. 218, 71 (2004). 27. B. K. Kim, K. S. Kim, K. Cho and C. E. Park, J. Adhesion Sci. Technol. 15, 1805 (2001). 28. F. Truica-Marasescu, S. Guimond, P. Jedrzeowski and M. R. Wertheimer, Nucl. Instrum. Methods Phys. Res. B 236, 117 (2005).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 171–191 Ed. K.L. Mittal © VSP 2007
Covalent coupling of fluorophores to polymer surface-bonded functional groups R. MIX,* K. HOFFMANN, U. RESCH-GENGER, R. DECKER and J. F. FRIEDRICH Federal Institute of Materials Research and Testing, Unter den Eichen 87, D- 12205 Berlin, Germany
Abstract—Chemical reactions at surface functionalities of various surfaces are of ever increasing importance in many different areas, such as microarray technology or design of sensor materials. Prominent examples are the attachment of sensor or biomolecules to polymeric and glass supports. Such developments require techniques for a controlled generation of different types and densities of surface functionalities, testing protocols for covalent attachment of functional molecules to these reactive groups and last, but not least, methods for proper characterization and quantification of these species at various surfaces. Here, results are presented on covalent coupling of fluorescent labels to functional groups on polymer surfaces based on several synthetic concepts employing different types of surface functionalities, spacers, and fluorescent labels such as commercially available dansyl, rhodamine, and fluorescein dyes. Besides the fluorometric characterization of these surfacelinked fluorophores, emission measurements are compared with XPS results, thereby evaluating the potential of fluorometry for the characterization and quantification of surface functionalities. Keywords: Plasma modification; fluorescence; surface functionalization; reactions at polymer surfaces; spectroscopy; quantification.
1. INTRODUCTION
Controlled surface functionalization, i.e., generation of surface functionalities of defined chemical type and concentration and attachment of, e.g., spacers, sensors or biomolecules, is of ever increasing interest in many different areas, such as microarray technology or chemical sensing. Accordingly, not only synthetic concepts for reactions with functional molecules at surfaces are highly desired, but also simple methods for the characterization and quantification of surface functionalities are required. Here, typically used analytical methods include XPS, FTIR, SIMS, etc. [1, 2].
*
To whom correspondence should be addressed. Tel.: (49-30) 8104-4310; Fax: (49-30) 8104-1637; e-mail:
[email protected]
R. Mix et al.
172
Despite the obvious potential of fluorescence techniques, there exist comparatively few papers describing direct coupling of fluorescent labels to surface functionalities for characterization and quantification purposes. Dansyl groups covalently bonded to the surface of oxidatively functionalized polyethylene films have been reported by Holmes-Farley et al. [3]. Ivanov et al. [4] applied the technique of fluorescence labelling for the determination of functional groups on plasmaand chromic acid-modified polymer surfaces. Griesser and Chatelier [5] reported covalent coupling of fluorescein isothiocyanate (FITC) to heptylamine plasma modified polymers. Henneuse-Boxus et al. [6] attached 7-amino-4-(trifluoromethyl)coumarin and 1-aminopyrene to poly(ether ether ketone) (PEEK) films Table 1. Reported densities of surface functionalities Author(s)
Substrate/treatment
Groups per unit area (reported by the authors)
Calcd. conc. (nmol/cm2)
Method
Rasmussen et al. [9]
PE/chromic acid
2¥1015 COOH groups/cm2
3.3 COOH
HolmesFarley et al. [3] Matveev [10]
PE/chromic acid
16¥1014 COOH groups/cm2
2.7 COOH
Quartz (rough)/γ-APT
40–60 nmol NH2/cm2
40–60 NH2
Puleo [11]
Co-Cr-Mo/γ-APT
92 NH2/nm2 15 NH2 (0.082 M γ-APT in water) 1.1 NH2 6.7 NH2/nm2 (1.02 M γ-APT in organic solvent)
Two independent fluorometric techniques Conversion to glycylamides and titration of glycine NH2 labelled with [3H] acetic anhydride Reaction with trinitrobenzene sulfonic acid
Nashat et al. [12]
Microparticles/SiO2, γ-APT
(3–9) ¥10-4 fluorophores/nm2
Griesser and Chatelier [5] Kühn et al. [8] Friedrich et al. [13]
PE, PP/O2 plasma, wet-chemical reduction Polymers/pulsed plasma polymerization/copolymerization of allyl alcohol, allylamine
(0.5–1.5)¥10-4 Confocal fluorescence fluorophores microscopy 0.6 FITC MO calculations
1 molecule FITC occupies 0.2– 0.24 nm2 10–14 OH/100 C
0.4–0.5 OH
1–30 OH/100 C 1–24 NH2/100 C
<0.1–1.2 OH <0.1–0.9 NH2
Derivatization with TFAA (OH) Derivatization with TFAA (OH). PFBA (NH2)
γ-APT, γ-aminopropyltriethoxysilane; FITC, fluorescein isothiocyanate; TFAA, trifluoroacetic anhydride; PFBA, pentafluorobenzaldehyde.
Covalent coupling of fluorophores to polymer surfaces
173
functionalized with hydroxyl groups. Even though a qualitative fluorescence detection was reported, a quantitative analysis of the surface-attached labels and a correlation between XPS and fluorescence measurements failed. Moreover, reproducibility problems have been reported. 1.1. Functional groups at (flat) surfaces For all methods relying on functionalized surfaces, the characterization of the functional groups and their quantification are essential requirements for further reactions with these groups. In Table 1 the densities of surface functionalities reported in the literature are summarized. Further information concerning primary and secondary amine concentrations generated by plasma polymerization of different amines is recently given by Choukourov et al. [7]. By plasma activation of polypropylene followed by wet-chemical reduction, 10–14 OH groups/100 C-atoms can be generated [8]. This corresponds to 2.4–3.4 OH groups per nm2 or 0.3–0.4 nmol OH per cm2, based on MOPAC or MM2 calculations (1 propylene unit occupies about 0.12 nm2 and, accordingly, 8 propylene units occupy approx. 1 nm2), provided that the surface has low roughness. For coating of a polymer surface by plasma polymers the maximum number of OH or NH2 groups of plasma polymerized allylamine or allyl alcohol layers (assuming 100% retention of the functional groups) amounts to 8 functional groups per nm2 or a concentration of 1.3 nmol OH (NH2) per cm2. The detection limit for fluorometric characterization of labelled polymer surfaces was given by Ivanov et al. [4] as approx. 10-3 nmol/cm2 for dansyl groups. 1.2. Tailoring of surfaces with functional groups by plasma chemical processes An elegant approach to introduce a wide variety of functional groups onto chemically inert polymer surfaces is the application of plasma processes. Especially low-temperature glow-discharge plasmas are well suited to plasma-chemically modify the polymer surface, thereby forming reactive sites that can attach atoms or radical fragments from plasma-forming functional groups, which present anchoring sites for chemical grafting of different labels or molecules. First attempts were made by Hollahan et al. who used a NH3 plasma for the introduction of primary amino groups onto polypropylene surfaces [14]. It was shown later that the reaction of an ammonia plasma with polymer surfaces always led to the formation of co-existing N-functionalities of different types. Among all N-functionalities, the concentration of primary amino groups did not exceed 3 NH2 groups/100 C atoms [15]. Attempts to increase the number of primary amino groups using NH3/H2 gas mixtures were unsuccessful [16]. The same situation is manifested using an oxygen plasma. A wide variety of different O-functional groups are generated at polymer surfaces [17], amongst them the concentration of OH groups, typically amounting to 2–4 OH groups/100 C. The formation of hydroxyl groups requires abstraction of H from the polymer chain, formation of OH species in the plasma phase and their attachment onto the
174
R. Mix et al.
backbone. Since the bond energies of C–H and C–C bonds are similar, both, a simultaneous polymer chain scission and H abstraction are plausible. Thus, a polymer rearrangement is often observed. Therefore, a two-step process can be considered that counteracts the unsatisfactory result of the O2 plasma modification by additional use of a wet-chemical reduction process. This was first attempted by Nuzzo and Smolinsky [18] and it was optimized later [8]. In this way a concentration of 10–14 OH groups/100 C was realized with a selectivity of about 60% for all O-functionalities. Another promising process is the plasma-initiated polymerization of monomers carrying functional groups. An especially mild way is pulsed plasma polymerization and the application of low power. Here, the desired monotype functional groups are provided by the monomer, which is polymerized in the (pulsed) plasma. The challenge in producing such monotype functionalized polymer coatings is to carry out the plasma process under conditions as mild as possible to avoid fragmentation of monomers and to preserve the functional groups of the respective monomers [19]. More recently, application of the continuous-wave plasma mode has allowed polymerization of any evaporable organic substance. This strategy offers a completely different polymerization mechanism in plasmas in comparison to classic polymerizations. Accordingly, this type of plasma polymerization is characterized as “fragmentation-polyrecombination” process or as “atomic” polymerization [20, 21]. It is to be expected that using such a continuous-wave plasma polymerization the retention of functional groups is low and the formation of irregular polymer structures is dominant because of complete fragmentation and random recombination processes. This process can be significantly improved by introduction of the pulsed plasma technique using short plasma pulses (0.1–1 ms) for monomer activation, and long plasma-off periods (1–100 ms) needed for a pure chemical chain propagation instead of monomer dissociation and subsequent random recombination of fragments. By minimizing the plasma power input, monomer fragmentation can be suppressed and the number of retained functional groups in the deposited plasma polymer layer should be increased compared to the number of functional groups originally present in the monomer molecules. 60–95% of the functional groups could be retained at the polymer surface depending on the type of monomer used [13]. Generally, monomers suited for radical polymerization in the gas phase without any direct exposure to the plasma during the plasma-off periods are the same as in classic radical polymerization. Primarily vinyl and acrylic monomers should also be able to undergo a radical polymerization under low-pressure conditions. Also, allyl monomers could be efficiently polymerized under such pulsed plasma conditions [22, 23]. Due to the low monomer density in the gas phase under lowpressure conditions, the termination of chain propagation by radical transfer to the monomer is negligible. This side-reaction dominates in radical allyl polymerization in the liquid phase, thus hindering the formation of high-molecular-weight products [24]. Analogously the classic rules of co-polymerization can also be ap-
Covalent coupling of fluorophores to polymer surfaces
175
plied to the co-polymerization reaction in the plasma-off periods of pulsed plasmas. Thus, efficient co-polymerization is possible if the polymerization tendencies of participating comonomers are comparable (so-called co-polymerization parameters). Nevertheless, the fragmentation of monomer molecules and the recombination of fragments and atoms to a randomly structured polymer during the plasma pulses introduces defects into the resulting polymer layer such as trapped radical sites, auto-oxidation products, cross-linking, irradiation defects, etc. [25]. 1.3. Chemical reactions on surface-bonded functional groups In the absence of humidity, isocyanate-containing substances react with OH groups to form urethane bonds. The reaction is slow and, therefore, it is normally catalyzed, for instance, by dibutyltin dilaurate, and needs higher temperatures (60–80°C). Applying this reaction to surface-bonded hydroxyl groups offers an interesting way to transfer (stable) OH groups into spacer-bonded NCO groups [26–30]. These isocyanate groups can easily be converted into primary amino groups by hydrolysis with water. Ghandini et al. [31] and Noiset et al. [32] have reported proven and tested methods to attach diisocyanates to cellulose fibres and flat PEEK film surfaces, respectively, at room temperature. However, the low temperature demands very long reaction times, in worst cases, up to several days. In this paper different strategies for the controlled functionalization of polymer surfaces by diisocyanates at room temperature are presented for the introduction of functional molecules, e.g., fluorescent labels. PP was chosen as the substrate because of its chemical resistance towards solvents during post-plasma wetchemical processing and its suitable optical properties required for fluorescence investigations. 2. EXPERIMENTAL
2.1. Materials PP foils with a thickness of 100 µm (Goodfellow, UK) were ultrasonically cleaned in a diethyl ether bath for 15 min. Allylamine (>99%) was purchased from Merck, Germany and distilled prior to use. All the solvents (tetrahydrofuran (THF), ethanol, methanol, acetonitrile) were dried and distilled before use. THF was stored over molecular sieves for more than 1 week before use. For the wetchemical reduction of the foils, the borane tetrahydrofuran complex (1.0 M solution in THF, Sigma-Aldrich, USA) was used as received. The derivatization reagents trifluoro ethylamine (TFEA, >99%, Sigma-Aldrich, USA), trifluoroacetic anhydride (TFAA, >99%, Merck, Germany), and pentafluoro benzaldehyde (PFBA, >98%, Sigma-Aldrich, USA), as well as toluene 2,4-diisocyanate (TDI, 80/20 isomer mixture, Merck, Germany), methylene-di-p-phenylene isocyanate (MDI, >98%, Sigma-Aldrich, USA), hexamethylene diisocyanate (HDI, >98%, Fluka, Switzerland) and dibutyltin dilaurate (DBTL, >98%, Merck, Germany)
176
R. Mix et al.
were used as received. The fluorescent labels dansyl chloride (DNS-Cl), dansylcadaverine (DNS-Ca), dansylhydrazine (DNS-H), 5(6)-aminofluorescein (AF) and rhodamine 110 (RHO) (purity of all fluorescence labels >95%) were provided by Fluka (Switzerland). They were used without further purification. Glutaraldehyde solution (assay approx. 25% in water) was supplied by Fluka (Switzerland). A 5% solution was used for spacer coupling. 2.2. Generation of polymer surfaces with functional groups O2 plasma treatments were performed in a cylindrical plasma reactor of 50 dm³ volume. The design of the plasma reactor has been described in detail in Ref. [33]. The reactor was equipped with a radio-frequency (rf, 13.56 MHz) generator with an automatic matching unit and an rf bar antenna (length: 35 cm), mass flow controllers for gases and vapours and a heated gas/vapour distribution in the chamber. The substrate foil was mounted on a continuously rotating, grounded steel cylinder (Ø = 10 cm, length = 35 cm, rotating frequency = 0.5–1 s-1) at a distance of 10 cm from the rf-powered electrode. For production of an oil-free high vacuum an adjustable turbomolecular pump was used. The pressure and monomer flow were controlled by varying the speed of the turbomolecular pump. The flow was adjusted by using flow controllers (MKS, Germany). When using aggressive gases or vapours a cryo-trap was interposed between the reactor and the pump. Surface functionalization of PP foils in the O2 plasma was performed in the continuous-wave (cw) mode. The power input was 100 W at a standard pressure of 10 Pa. The deposition of allylamine was carried out in the cw or pulsed mode. The power input was selected between 20 W and 50 W (cw mode) and 200 W to 500 W (pulsed regime at a duty cycle of 0.1 and a pulse frequency of 103 Hz). The deposited layers were about 50 nm thick. The deposition rate was determined with a quartz microbalance (FTM5 Film Thickness Monitor, BOC-Edwards, UK). 2.3. Reduction of O2 plasma treated polypropylene foils The oxygen-plasma-treated foils (5¥5 cm2) were immersed in 12 ml dry THF charged with 3 ml of diborane solution and stirred under N2 at room temperature for 18 h. The foils were removed and washed twice with 10 ml dry THF followed by treatment with 10 ml saturated NaHCO3 in water/methanol (1:1) for 5 min. After washing with water and methanol (both twice for 5 min each) the OHmodified PP foils (PP-OH) were dried and stored in a desiccator. 2.4. Surface analysis and fluorescence measurements The XPS data were acquired with a SAGE 150 Spectrometer (Specs, Berlin, Germany) using a non-monochromatized MgKα or AlKα radiation with 12.5 kV and 250 W settings at a pressure of ≈10-7 Pa in the analysis chamber equipped with channeltron detectors. XPS spectra were acquired in the constant analyzer
Covalent coupling of fluorophores to polymer surfaces
177
energy (CAE) mode at 90° take-off angle. Peak analysis was performed using the peak fit routine from Specs. The FT-IR spectra were recorded with a Nexus instrument (Nicolet, USA) using the ATR (Attenuated Total Reflectance, 45° angle of incidence) technique with a diamond or a Ge cell (Golden Gate, Nicolet, USA). The IR signal originates from a near-surface layer of the polymer film. The information depth depends on the material used as ATR crystal and amounts to about <1.5 µm using germanium and 2.5 µm using diamond. XPS and IR analyses of polymer films have been described in detail elsewhere [34]. 2.4.1. Steady-state fluorescence The fluorescence spectra of the fluorophore-labeled polymer films were measured in a 90°-geometry between two quartz windows with an excitation polarizer set to 0° and an emission polarizer set to 54.7° with a Spectronics Instruments 8100 spectro-fluorometer. Unless stated otherwise, all fluorescence spectra presented were corrected for the wavelength- and polarization-dependent spectral response of the detection system. Prior to each fluorescence measurement, the fluorophorelabeled films were subjected to Soxhlet extraction (2 h) using ethanol as the solvent to remove the adsorbed dye molecules. For each series of measurements, blank or so-called reference samples were prepared by reaction of the nonfunctionalized polymer film with the respective fluorescent label using the same procedure as employed to covalently attach fluorophores to the surfacefunctionalized films. 2.5. Spectroscopic characterization of surface functionalities using XPS 2.5.1. Derivatization of OH groups Samples with OH groups on the surface were exposed to trifluoroacetic anhydride (TFAA) vapours for 10–15 min after evacuation to 103 Pa in order to transform the hydroxyl group into a fluorinated derivative as shown below:
Afterwards these samples were degassed at a pressure of 10-5 Pa to remove the free trifluoroacetic acid [35]. The derivatization of OH groups had proved to be highly selective (ca. >90% [36]). The yield of this reaction is about 90%. 2.5.2. Derivatization of NH2 groups Primary amino groups were reacted with PFBA, and the reaction product was a Schiff´s base:
178
R. Mix et al.
Similar to OH derivatization, the labelling of NH2 groups was performed as a gas-phase reaction. The samples were exposed to PFBA vapours at 45°C for 2 h and subsequently degassed at a pressure of 10-5 Pa for 8 h to remove the nonconsumed PFBA. This labelling reaction reaches yield in the range of 80% [37]. 2.5.3. Derivatization of NCO groups NCO-modified foils (1¥3 cm2) were placed in a flask with 1 ml trifluoroethylamine in 10 ml THF for 24 h at room temperature. After washing with pure THF, the foils were exposed to vacuum (10 Pa) for 2 h and measured by XPS. The N and F elements act as XPS tags:
2.5.4. Derivatization of CHO groups CHO-modified foils (1¥3 cm2) were shaken in a solution of 10 µl 4-fluoromethylaniline in 10 ml THF for 2 h at room temperature. Then, the foils were washed with pure THF (3¥) and dried in vacuum prior to XPS measurements:
2.6. Chemical reactions of isocyanates with OH groups 2.6.1. Urethane formation on the surface OH-modified polypropylene foils (PP-OH) (5¥5 cm2), as well as the whole equipment were carefully dried at 30°C for at least 1 h, before reaction. Thereafter, the foils were dipped in 10 ml dry THF in a covered reaction vessel under N2 flow. The isocyanate (2 ml TDI or 0.5 g MDI in 10 ml dry THF) and the catalyst (20 µl) were added by a syringe to the reaction mixture. The foils were shaken for various times at room temperature. After completion of the reaction, the NCOmodified PP foils (PP-NCO) were washed twice with dry THF.
Covalent coupling of fluorophores to polymer surfaces
179
The NCO-modified PP was transformed into amino functionalized PP (PPNH2) by reaction with water for 2 h. The foils with amino groups were immersed in 10 ml THF and reacted either with 6 ml dansyl chloride solution (2 mg DNSCl/ml acetonitrile) in the presence of 0.5 ml sodium borate buffer (pH 9) for 1 h or with 2 ml FITC solution (0.2 µg FITC/ml ethanol) for 2 h. The reaction of the NCO-functionalized foils with amino groups containing fluorescent labels was carried out in THF or ethanol, which took about 4 h. After completion of the reaction the foils were washed with ethanol, water and then with acetone. The modified foils were stored in the dark prior to fluorescence measurements. 2.6.2. Introduction of the glutaraldehyde spacer The allylamine modified surfaces (5¥5 cm2) were treated with 5% (v/v) aqueous glutaraldehyde solution (10 ml). The reaction was performed at ambient temperature for 2 h. The film was washed with water (3 times) and once with acetone before drying. The further reaction with 6 ml solution of a NH2-containing fluorescent label (0.5 mg fluorescent dye/ml dry THF) was carried out in 20 ml ethanol or THF for 2 h at room temperature. 3. RESULTS
3.1. Monotype functionalization 3.1.1. Reduction of oxygen plasma introduced oxygen-containing groups at PP surfaces to OH groups Using the oxygen rf pulsed-plasma modification, the attachment of oxygen containing functional groups onto the surface of PP can be demonstrated by a series of ATR–FT-IR spectra (see Fig. 1). A strong carbonyl band (≈ 1700 cm–1), alcoholic structures (≈ 1100–1200 cm–1), a broad OH peak and features related to adsorbed water (≈ 3100–3500 cm–1) appear. Then, the plasma parameters were so adjusted that surfaces with an O/C ratio of about 0.2 were produced. These foils were used for further chemical processing. The XPS spectra of PP at different stages of processing are shown in Fig. 2, including PP as received, PP after low-pressure oxygen plasma pretreatment, diborane reduction and derivatization with TFAA. The plasma treatment produces a strong O1s peak. After TFAA derivatization of OH groups at the PP surface only a small F1s peak at a binding energy of 689 eV is detected, indicating a low concentration of OH groups after O2 plasma treatment. This finding is not surprising because the oxygen plasma does not involve hydrogen. Available hydrogen sources are primarily the hydrogen abstracted from the polymer backbone and maybe traces of hydrogen from the dissociation of desorbed water from the walls of the reactor.
180
R. Mix et al.
Figure 1. ATR–FT-IR spectra after different treatment steps of PP foils using a Ge crystal and 45° angle of incidence.
Figure 2. XPS spectra after different treatment steps of PP foils.
Comparing the intensity of the F1s peaks in the survey scans with and without diborane reduction after derivatization of OH groups formed (cf. Fig. 2) it is quite obvious that the diborane reduction strongly increased the yield of OH groups. 3.1.2. Allyamine polymerized layers on PP In Fig. 3a and 3b the C1s peaks of plasma polymerized allylamine and the PFBAderivatized species deposited on PP in the cw mode are shown. The number of
Covalent coupling of fluorophores to polymer surfaces
181
Figure 3. XPS C1s peaks of plasma-deposited allylamine onto PP before (a) and after (b) PFBA derivatization.
available primary amino groups in the deposited allylamine layers was found to be between 12 and 16 NH2/100 C. 3.1.3. Isocyanate grafting followed by coupling of fluorescent labels O2-plasma-treated and diborane reduced PP foils adjusted to 9–12 OH/100 C were reacted with diisocyanates HDI, MDI and TDI. The subsequent addition of water led to amino modified foils as schematically shown in Fig. 4.
Figure 4. Schematics of the reactions of hydroxylated PP surfaces with diisocyanates and water.
Figure 5. Schematics of the attachment of DNS-Cl (a) and FITC (b) to hydroxylated PP surfaces after reaction with diisocyanates and water.
182
R. Mix et al.
The reaction with HDI was too slow; therefore, the reported results are focused on TDI and MDI. Subsequently, fluorescent labels such as DNS-Cl and FITC were grafted according to Fig. 5. In a second way, different amino-functionalized fluorescent labels such as DNS-H, DNS-Ca, AF and RHO were attached to the NCO-modified surface as shown in Fig. 6.
Figure 6. Schematics of the general reactions of NCO-modified surfaces with amino-functionalized fluorescent labels.
Figure 7. N and F concentrations (measured by XPS) for PP-OH reacted first with TDI and then with TFEA (a, b) or with water (c) in dependence on reaction time.
Covalent coupling of fluorophores to polymer surfaces
183
As is known, the reaction of the PP-OH with the isocyanate component is the most time-consuming reaction step in urethane formation. Because of this the reaction was pursued for different periods to estimate the consumption rate of surface-bonded hydroxyl groups. The NCO-modified surfaces were then reacted with water and trifluoroethylamine, as described before. The N and F elemental concentrations estimated by XPS are shown in Figs 7 and 8 for TDI and MDI, respectively. Figures 7 and 8 show that the reaction rates of surface-bonded hydroxyl groups with TDI and MDI are comparable. This means that there is no difference in reaction rates of surface-fixed or ‘mobile’ hydroxyl groups attached to flexible polymer chains as used in polyurethane synthesis. In Fig. 9 the N and S concentrations (measured by XPS) for the fluorescence label DNS-H attached to PP-OH by TDI and MDI are shown. While the S concentration remains constant after 24 h reaction time, the N concentration is growing with increasing reaction time at long reaction periods. The enhanced N concentration is an indication of incomplete exclusion of humidity. Traces of water react with PP-NCO to form amines. The latter further consume isocyanate component forming urea bonds which increase the N concentration at the surface. An example of corresponding fluorescence measurements with this label (DNS-H linked to the PP-OH by TDI) is displayed in Fig. 10. The goal of these fluorescence measurements is to spectroscopically follow the synthesis steps in the covalent attachment of fluorescent labels to different types of surface functionalities with and without interposed spacer. A detailed spectroscopic investigation of surface-bonded fluorophores is presented elsewhere [38].
Figure 8. N and F concentrations (measured by XPS) for PP-OH reacted first with MDI and then with TFEA (a, b) or with water (c) in dependence on reaction time.
184
R. Mix et al.
Figure 9. S and N concentrations (measured by XPS) for DNS-H bonded to hydroxyl-modified PP in dependence on the reaction time of PP-OH with TDI (a) and MDI (b).
Figure 10. Fluorescence excitation and emission spectra (excitation wavelength 350 nm; uncorrected) for DNS-H bonded to PP-OH, treated for various times with TDI.
Covalent coupling of fluorophores to polymer surfaces
185
Figure 11. Relative fluorescence intensity (spectrally corrected) and S content (measured by XPS) of DNS-H bonded to PP-OH via TDI in dependence on the reaction time (PP-OH + TDI).
Figure 12. S and N concentrations (measured by XPS) for DNS-Ca bonded to PP-OH and treated with MDI, as well as relative surface fluorescence intensity in dependence on reaction time (PP-OH + MDI).
186
R. Mix et al.
The fluorescence excitation spectra (normally correlated to the absorption spectra of emitting species [39]) show the expected maxima at about 347 nm and the uncorrected emission spectra of the surface-bonded dansyl label display emission maxima at 473 nm. The fluorescence intensity increases with the reaction time of PP-OH with TDI. From the weak emission of a reference sample obtained from an untreated PP foil immersed in the dye solution under exactly the same conditions as used for the film carrying surface functionalities, it follows that the measured fluorescence originates from covalently attached fluorophores and not from adsorbed dye molecules. Figure 11 displays the concentration of the S hetero atom of the dansyl label as determined by XPS and the fluorescence intensity from the measurements shown in Figs 9 and 10 as a function of the TDI reaction time. It illustrates a similar curve progression obtained by the two analytical methods. Analogous reaction of DNS-Ca with MDI results in similar curves, as shown in Fig. 12. 3.2. Plasma deposition of primary amino groups carrying polyallylamine layers and subsequent introduction of fluorescent labels 3.2.1. Plasma polymerization of allylamine and reaction with FITC Allylamine was polymerized both in continuous wave and pulsed regime, to vary the number of functionalities in the plasma polymer. By performing derivatization of NH2 groups with pentafluorobenzaldehyde, the NH2 concentration was calculated to be about 12–16 NH2/100 C. The variation in the measured concentration of amino groups is due to the different plasma parameters and modes used. To avoid uncontrolled oxygen incorporation into the deposited layers, known as postplasma auto-oxidation, the coated foils were immersed in the FITC solution immediately after finishing the plasma deposition process. The result of this reaction is an attachment of the fluorescent label with formation of a thiourea bond as demonstrated in Fig. 13. Taking the S concentration of the FITC-bonded PP corresponding to Fig. 13 as an indication for the reaction of the fluorescent label to the deposited amine layers, no clear trend with the applied power (pulsed plasma at duty cycle 0.5) was observed, neither for the S concentration (Fig. 14, curves A, B), nor for the fluorescence intensity (Fig. 14, curve C).
Figure 13. Reaction of amino functionalized PP (PP-NH2) with FITC.
Covalent coupling of fluorophores to polymer surfaces
187
Figure 14 also illustrates the effect of the extraction procedure with methanol. The decreasing S concentration after 5 h extraction is obviously due to a removal of short or oligomeric chains incorporated in the layer (Fig. 14, curve A). This result hints to a nearly constant amino group concentration for allylamine deposits at different wattages at a constant duty cycle of 0.5. This finding is in contrast to the results found by Mueller and Oehr [40] for plasma-polymerized allylamine and diaminocyclohexane samples. They found lower concentrations of primary amino groups with increasing wattage at a constant duty cycle. Choukourov et al. [7] compared the concentration of primary and secondary amino groups of allylamine films deposited in cw and pulsed plasma modes at the same (rf) power input. By derivatization with 4-trifluoromethyl benzaldehyde and TFAA they found that the concentration of NH2 groups was the highest for films deposited in the continuous plasma, while secondary amino groups were preferentially generated by the pulsed plasma deposition. The same power input means that equal doses were applied in both pulsed and cw modes, for example, 100 W rf power and a duty cycle of 0.1 corresponds to a power input of 10 W in the cwmode. The same tendency was described by Hamerli et al. [41] for polyallylamine films deposited in presence of argon in the afterglow region of a microwave (MW) plasma. The total amine concentration (primary, secondary and tertiary) obtained by colorimetric staining with Acid Orange II was higher for those films polymerized at higher power-input in the cw-mode as well as in the pulsed mode at constant duty cycle. In a further study Choukourov et al. [42] showed that both the primary and the secondary amine concentrations of pulsed plasma (duty cycle 0.1) polymerized diaminocyclohexane layers followed the same concentration dependence on power, with a minimum at the plasma-on time ton = 0.5 ms. These results indicate
Figure 14. Fluorescence intensity and S concentration for allylamine layers deposited by pulsed plasma on PP, followed by reaction with FITC.
188
R. Mix et al.
that the duty cycle has only a minor influence on the concentration of primary amino groups in the polymerized layer compared to the pulsing frequency (ton and toff) selected under the chosen deposition parameters. The controversial results reported in the literature may be due to the uncertainty as to how deep the derivatization reaction occurs in the bulk and also how complete is this reaction in relation to the information depth of the XPS method. The information depth of the XPS method is 5–7 nm. The derivatization reaction of allyl alcohol deposits with TFAA results in almost theoretical values of F content as reported by Rinsch et al. [43], indicating a derivatization depth of 5–7 nm. At present there is no absolutely reliable information about the maximum depth of derivatization with aromatic aldehydes in the Schiff’s base formation. Additionally, this reaction is related to the incorporation of oxygen into the surface as displayed by Fally et al. [44]. 3.2.2. Anchoring of fluorescent labels to allylamine plasma polymers by inserting a glutaraldehyde spacer A further reaction scheme involved the introduction of a glutaraldehyde spacer to PP-NH2 as shown in Scheme 1.
Scheme 1.
Figure 15. Uncorrected fluorescence spectra (excitation wavelength 490 nm) of rhodamine 110 attached to PP films modified by plasma-polymerized allylamine layers and subsequent reaction with glutaraldehyde.
Covalent coupling of fluorophores to polymer surfaces
189
Following Scheme 1, the attachment of e.g. amino-functionalized fluorescence labels can be achieved. In Fig. 15 preliminary measurements of the fluorescence emission of an RHO labeled polypropylene film are shown. After excitation of the film at 490 nm, the characteristic RHO emission at 528 nm was observed. The non-plasma-modified reference film shows only a weak fluorescence which is assigned to non-conjugated, only physically adsorbed, dye molecules. These results clearly demonstrate a promising alternative synthesis procedure for the attachment of fluorescent labels via spacer molecules to plasma chemically generated surface functionalities. 4. CONCLUSIONS
Functional groups are a prerequisite for grafting of chemical species such as sensor molecules to polymer surfaces. The generation of reactive groups like OH or NH2 on inert polymer surfaces can be achieved either by a combined process consisting of a plasma oxidation followed by a wet-chemical reduction (OH) or by coating with monotype functional groups carring polymer layers, for instance, polyallylamine. The combined process results in a moderate density of OH groups with concentrations of 10–14 OH/100 C. This process shows a selectivity of ca. 60%, i.e. 60% of all plasma-introduced O-functionalities were converted into OH groups. This process, however, allows to covalently attach the OH groups directly to the polymer backbone. The OH group density at the polymer surface can be controlled, to a certain degree, by varying the oxygen plasma treatment time. As described before [21] the OH-functionalization of polymer surfaces was stable, even if the films were stored in air over long periods (>1 year). The use of monomers with functional groups allows to generate a broad variety of functionalized surfaces. In general, the concentration of surface functions can be controlled through common plasma-initiated co-polymerization using “neutral” monomers without functional groups [13]. Here, the generation of amino groups was carried out by plasma polymerization of allylamine. The wet-chemical attachment of molecules to surface-linked functions, however, normally involves exposure of the polymers to a solvent for some hours. Such solvent treatments applied to plasma deposited layers induce a significant drawback: short polymer chains can be extracted by the solvents resulting in a reduced density of reactive groups and, therefore, of grafted substances at the surface. This effect is illustrated in Fig. 14 showing the decrease in S atomic content of FITC-labeled polyallylamine layers after extraction with methanol for 5 h. This decrease is obviously caused by extractable FITC-modified oligomer chains from the plasma polymer layer. Using allylamine for plasma deposition of amino groups, the amount of NH2 groups generated in the layer is only moderate compared to the OH group density in deposited allyl alcohol plasma polymers under comparable plasma conditions. Here, the pronounced tendency for post-plasma oxidation processes further re-
190
R. Mix et al.
duces the amount of reactive NH2 groups resulting in a continuous change in surface composition. New types of N- and O-containing functional groups are formed at the expense of the number of initially deposited NH2 groups and, consequently, the density of grafted, here fluorophore, molecules is reduced. In order to circumvent the problem with respect to the stability of aminomodified polymer surfaces, the covalent attachment of fluorophores was realized by two ways. The first process started from OH-modified PP substrates, which were reacted with diisocyanates and water yielding amino groups. These primary amines represent ideal and more stable anchoring groups for further grafting reactions. The amino functions were then reacted with fluorophores, e.g., DNS-Cl or FITC. Likewise, the reaction of NCO-modified PP foils with aminofunctionalized fluorophores resulted in fluorescent PP surfaces. The fluorophores DNS-Cl and FITC, which contain sulfur heteroatoms, allow to directly compare fluorescence measurements with XPS data. The result of first attempts to correlate these two analytical methods, for example, for dansyl chromophores linked by the spacers MDI and TDI, reveals the potential of fluorescence spectroscopy as a complementary and very sensitive method for the characterization of surface groups. REFERENCES 1. E. E. Johnston and B. D. Ratner, J. Electron. Spectr. Rel. Phenom. 81, 303 (1996). 2. U. Oran, S. Swaraj, J. F. Friedrich and W. E. S. Unger, Surface Coating. Technol. 200, 463 (2005). 3. S. T. Holmes-Farley, R. H. Reamey, T. J. McCarthy, J. Deutch and G. M. Whitesides, Langmuir 2, 266 (1986). 4. V. B. Ivanov, J. Behnisch, A. Holländer, F. Mehdorn and H. Zimmermann, Surface Interface Anal. 24, 257 (1996). 5. H. J. Griesser and R. C. Chatelier, J. Appl. Polym. Sci., Appl. Polym. Symp. 46, 361 (1990). 6. C. Henneuse-Boxus, A. de Ro, P. Bertrand and J. Marchand-Brynaert, Polymer 41, 2339 (2000). 7. A. Choukourov, H. Biederman, D. Slavinska, L. Hanley, A. Grinevich, H. Boldyryeva and A. Mackova, J. Phys. Chem. B 109, 23086 (2005). 8. G. Kühn, S. Weidner, R. Decker, A. Ghode and J. F. Friedrich, Surface Coating. Technol. 116– 119, 796 (1999). 9. J. R. Rasmussen, E. R. Stedronsky and G. M. Whitesides, J. Am. Chem. Soc. 99, 4336 (1977). 10. S. V. Mateev, Biosens. Bioelectr. 9, 333 (1994). 11. A. D. Puleo, J. Biomed. Mater. Res. 29, 951 (1995). 12. A. H. Nashat, M. Moronne and M. Ferrari, Biotechnol. Bioeng. 60, 137 (1998). 13. J. F. Friedrich, G. Kühn, R. Mix and W. E. S. Unger, Polymer Process. Plasmas 1, 28 (2004). 14. J. R. Hollahan, B. B. Stafford, R. D. Falb and S. T. Payne, J. Appl. Polym. Sci. 13, 807 (1969). 15. J. F. Friedrich, J. Gähde, H. Frommelt and H. Wittrich, Faserforsch. Textiltechn. / Z. Polymerenforsch. 27, 604 (1976). 16. A. Meyer-Plath, PhD-thesis. Ernst-Moritz-Arndt University, Greifswald (2003). 17. K. Rossmann, J. Polym. Sci. 19, 141 (1956). 18. R. G. Nuzzo and G. Smolinsky, Macromolecules 17, 1013 (1984). 19. J. F. Friedrich, J. Gähde, H. Frommelt and H. Wittrich, Faserforsch. Textiltechn. / Z. Polymerenforsch. 27, 517 (1976). 20. H. K. Yasuda, ACS Symp. Ser. 108, 37 (1979).
Covalent coupling of fluorophores to polymer surfaces
191
21. G. Kühn, I. Retzko, A. Lippitz, W. E. S. Unger and J. F. Friedrich, Surface Coating. Technol. 142–144, 494 (2001). 22. A. Harsch, J. Calderon, R. B. Timmons and G. W. Gross, J. Neurosc. Methods 98, 135 (2000). 23. H. Schönherr, M. T. van Os, R. Förch, R. B. Timmons, W. Knoll and G. J. Vancso, Chem. Mater. 12, 3689 (2000). 24. H.-G. Elias, in: Makromoleküle, p. 770. Huethig & Wepf, Basel (1975). 25. J. F. Friedrich, G. Kühn, R. Mix, I. Retzko, V. Gerstung, S. Weidner, R.-D. Schulze and W. E. S. Unger, in: Polyimides and Other High Temperature Polymers: Synthesis, Characterization and Applications, Vol. 2, K. L. Mittal (Ed.), pp. 359–388. VSP, Utrecht (2003). 26. Q. Liu, J. R. de Wijn, K. de Groot and C. A. van Blitterswijk, Biomaterials 19, 1067 (1998). 27. Q. Liu, J. R. de Wijn and C. A. van Blitterswijk, J. Biomed. Mater. Res. 40, 358 (1998). 28. J. Yuan, J. Zhang, X. Zang, J. Shen and S. Lin, Colloids Surfaces B 36, 19 (2004). 29. D. Klee, Z. Ademovic, A. Bosserhoff, H. Hoecker, G. Maziolis and H. Erli, Biomaterials 24, 3663 (2003). 30. G. Dong, J. Sun, C. Yao, G. J. Jiang, C. Hung and F. Linc, Biomaterials 22, 3179 (2001). 31. A. Ghandini, B. Vagner, E. Zeno and S. Bach, Polymer Int. 50, 7 (2001). 32. O. Noiset, Y.-J. Schneider, J. Marchand-Brynaert, J. Polym. Sci. A, Polym. Chem. 35, 3779 (1997). 33. J. F. Friedrich, I. Retzko, G. Kühn, W. E. S. Unger and A. Lippitz, in: Metallized Plastics: Fundamental and Applied Aspects, Vol. 7, K. L. Mittal (Ed.), pp. 117–142, VSP, Utrecht (2001). 34. I. Retzko, J. F. Friedrich, A. Lippitz and W. E. S Unger, J. Electron. Spectr. Relat. Phenom. 121, 111 (2001). 35. A. Chilkoti and B. D. Ratner, in: Surface Characterization of Advanced Polymers, L. Sabbattini and P. G. Zambonin (Eds.), pp. 221–259, VCH, Weinheim (1996). 36. S. Geng, PhD thesis. University of Potsdam, Potsdam (1996). 37. D. E. Everhart and C. N. Reilley, Anal. Chem. 53, 665 (1981). 38. K. Hoffmann, U. Resch-Genger, R. Mix and J. F. Friedrich, J. Fluoresc. 16, 441 (2006). 39. J. R. Lakowicz, Principles of Fluorescence Spectroscopy, 2nd edn. Kluwer/Plenum, New York, NY (1999). 40. M. Mueller and C. Oehr, Surface Coating. Technol. 116–119, 802 (1999). 41. P. Hamerli, Th. Weigel, Th. Groth and D. Paul, Biomaterials 24, 3989 (2003). 42. A. Choukourov, H. Biederman, D. Slavinska, M. Trchova and A. Holländer, Surface Coating. Technol. 174–175, 863 (2003). 43. C. L. Rinsch, X. Chen, V. Panchalingam, R. C. Eberhart, J.-H. Wang and R. B. Timmons, Langmuir 12, 2995 (1996). 44. F. Fally, C. Doneux, J. Riga and J. J. Verbist, J. Appl. Polym. Sci. 56, 597 (1995).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 193–207 Ed. K.L. Mittal © VSP 2007
Functionalization of fiber surfaces by thin layers of chitosan and related carbohydrate biopolymers and their antimicrobial activity DIERK KNITTEL∗ and ECKHARD SCHOLLMEYER Deutsches Textilforschungszentrum Nord-West e.V. (DTNW), Adlerstr. 1, D-47798 Krefeld, Germany
Abstract—Surface chemistry and topography determine almost all properties of textiles in use (for example, adhesion, skin contact, handling, etc.), whereas the fiber bulk provides the strength and thermomechanical stability. Thus, new strategies have to be developed for imparting permanent functionality to textile surfaces. This aim can be reached by using, e.g., biocompatible polymers, such as carrageenans, chitosans or their derivatives for permanent finish. Surface modification of cellulosic and synthetic fibers by thin layers of ionic carbohydrates is described. After derivatization with anchoring groups these molecules are chemically bound to the fiber and result in film-like addon (low weight percentage) with high wash fastness. Results on the permanent fixation of chitosan and carrageenan on fibers using different chemical anchor bonds are presented. The antimicrobial activity of these finished textiles has been determined using laser nephelometry and tetrazolium compounds against some selected bacteria and fungi. In cooperation with dermatologists the harmlessness of these biopolymer finishes has been shown. In addition, there are indications of benefits to patients with atopic dermatologic eczemas. The treated textiles can also be used as an ionexchanger for some heavy metals and, as well as protein-binding surfaces. Keywords: Biopolymer; surface modification; functional textile; antimicrobial textile; ion exchange.
1. INTRODUCTION
There is an increasing demand for imparting active agents to textile materials (fabrics and non-wovens) by chemical means in order to create additional properties ('functional textiles') [1]. With synthetic fibers this may create a better hydrophilic behavior (water retention, sweat transport, etc.). On natural fibers this could include anchoring of bacteriostatic or odour binding agents and related functions. An advantageous strategy is to rely on existing fiber types and to modify only the surface, thus retaining the desirable mechanical properties of the bulk fiber. In addition, such a strategy offers high flexibility to the textile finishing industry. ∗
To whom correspondence should be addressed. Tel.: (49-2151) 843-0; Fax: (49-2151) 843-143; e-mail:
[email protected]
194
D. Knittel and E. Schollmeyer
1.1. Ionic carbohydrates for textile finishing Biopolymers or their derivatives as surface modifiers can offer special properties, such as water retention, hydrogel formation or complexing power. Because of their interesting physiological properties along with the possibility to anchor them in textile finishing steps, these should be investigated in detail. The compatibility of pure biopolymers from the carbohydrate family towards the skin and their wound-healing effect are well known. These properties are well acknowledged in the traditional uses of biopolymers in foodstuffs, cosmetics or ointments. Therefore, an important task lies in evaluation of the methods to anchor such biopolymers permanently onto fiber surfaces in a way that the biopolymers will retain their beneficial bulk properties. The physicochemical interactions utilized for fiber surface modification are summarized in Table 1. Some properties of the biopolymers (in bulk) and those which can be achieved by a permanent textile finish are summarized in Table 2 [2–5]. Table 1. Interactions of (derivatized) biopolymers for permanent anchoring on fibers Mode of anchoring
Fiber type CO/CV
WO
PA
PET
PAN
Cross-linker (resin finish)
y
n
n
n
n
Ionic interaction
n
y
y
n
y
Covalent bonding
y
y
y
n
n
van der Waals interaction
n
n
y
y
y
CO, cotton; CV, viscose; WO, wool; PA, polyamide; PET, poly(ethylene terephthalate); PAN, poly(acrylonitrile); y, possible interaction; n, no or weak interaction.
Table 2. Selected properties of some biopolymers Biopolymer (Derivative)
Property (Bulk material)
Application on textile
Dextrins Chitosan
Hydrophilicity Film formation Hydrophilicity
Alginates
Gel, film
Pectin (s) Carrageenan
Gel, film Protein binding
Regulation of micro-climate Antibacterial, antifungal, wound healing Regulation of micro-climate and pH value on skin Water retention Antiallergic
Functionalization of fiber surfaces by carbohydrate biopolymers
195
The chemical structures of two of these biopolymers mainly used in this work are given in Fig. 1.
Figure 1. Structure of chitosan (CTS) (top) and structure of κ-carrageenan (bottom).
Figure 2. Finishing strategy for synthetic fibers by fiber surface modification (a) followed by chemical bonding of biopolymer (b).
Figure 3. Finishing strategy for synthetic fibers using derivatization of the biopolymer (e.g., CTS) followed by fixation according to the disperse dyeing technique.
196
D. Knittel and E. Schollmeyer
1.2. Strategy for permanent fixation of biopolymers on textile surfaces 1.2.1. Preparation and finishing of cellulosic and proteinic substrates with biopolymers (or with derivatives) At present the fixation of biopolymers onto cellulosics is best done by using different bi- or trifunctional anchoring groups applied from a suitable solution of the carbohydrate [6, 7]. Anchoring will occur by statistical reaction of anchor chemical with the fiber and with the biopolymer. The amount of anchor chemical used will have strong influence on handling of the fabric obtained and on the biochemical activity of the treated fabric. Generally, the amounts of anchor chemicals are calculated for the conversion of active groups of the biopolymer in the range of 5–20% of equivalent monomeric units of the carbohydrate. 1.2.2. Preparation and finishing of polyester with chitosan (or its derivatives) For permanent anchoring of CTS or other polymeric carbohydrates on synthetic fibers two strategies can be followed which are schematically given in Figs 2 and 3. Firstly, PET can be surface-modified with, e.g., dodecylamine using the disperse dyeing technique. Thus amino functions are created which in a second step are treated with CTS solutions using anchor chemicals, such as hydroxy- or methoxydichlorotriazine. The second approach is to modify the biopolymer (e.g., CTS) with a certain amount of long alkyl chains and fixing this derivative again similarily to disperse dyeing techniques on PET (Fig. 3) [7]. 2. EXPERIMENTAL
2.1. Materials and methods Chitosan (Heppe 85/S60/A1, Heppe, Queis, Germany), average molecular mass 250 kDa, Chitosan MBP 21 (Marine Bioproducts, Bremerhaven, Germany), degree of deacetylation 85%, mean molecular mass 40 kDa. Other ionic carbohydrates, like κ-carrageenan and alginate, were from Fluka. Anchoring chemicals used were cyanuric chloride (CNC) (Fluka), 2,4-dichloro6-methoxy-1,3,5-triazine (MCNC, synthesized in our laboratory), hydroxydichloro-1,3,5-triazine (sodium salt) (Na-HDCT, Degussa), butane-1,2,3,4tetracarboxylic acid (BTCA) and glycidoxypropyltrimethoxysilane (both from Aldrich). 2,3,5-Triphenyltetrazolium chloride (TTC) for viability tests on microbes was from Merck. Non-ionic detergent Marlipal® O 13/80 (Degussa-Hüls, Germany) and ECE-1 Test Detergent according to ISO 1105-CO6 were used for wetting and washing. Polyelectrolyte standard solutions, e.g., poly(diallyl dimethylammonium chloride) (Poly-DADMAC) and poly(ethylene sulfonate) sodium salt (PES-Na) were from Mütek (Herrsching, Germany). Other chemicals used such as buffers, ionic materials or cultivation media were of highest available grade (Merck).
Functionalization of fiber surfaces by carbohydrate biopolymers
197
2.2. Fabrics Plain weave cotton (CO) fabric (102 g/m2, thickness 0.42 mm, type S/400 (Testex, Bad Münstereifel, Germany), CO (200 g/m2) and polyester fabric (PET) (104 g/m2) were used for textile treatment. Regenerated cellulose fabric (Tencel, 183 g/m2, plain weave) was from Courtaulds. 2.3. Analytical equipment Particle Charge Detector PCD 03 PH (BTG Mütek, Hersching, Germany) with titration system Dosimat 665 (Metrohm), atomic absorption spectrometer SpectrAA-800 (Varian), microbial cell counting system CASY® 1 (Schärfe System, Reutlingen, Germany), laser nephelometer NEPHELOstar Galaxy® (BMG LABTECH, Offenburg, Germany). All other equipment for padding of textile samples and for squeezing, heating and washing used was standard in textile research. Escherichia coli (E. coli, DSM 498), Candida albicans (DSM 11225) and C. krusei (ATCC 6258) were used as microbes. 2.4. Recipes for treating liquors for biopolymer finishing For the use of chitosan a suitable amount was dissolved in slightly molar excess of acetic acid, cooled to 0–5°C and after addition of the cyanuric chloride or dichloromethoxytriazine solution (in dioxane or acetone, about 50–100 ml organic solvent/l aqueous solution) the pH was adjusted to 5.5 by adding NaOH solution. After warming to room temperature, drops, of detergent solution were added and the fabrics padded, squeezed and dried in air. Thermal fixation was done for 5–20 min at 140°C. After fixation, the samples were washed twice at 40°C in a commercial detergent solution using a laboratory washing machine. Using either the hydroxydichlorotriazine anchor (sodium salt) or polycarboxylic acids no organic cosolvent is required. In case of polycarboxylic acids as anchors 0.6 mol sodium acetate trihydrate for 1 mol anchor was used as a catalyst. For pretreatment of PET with dodecylamine, a solution of 5 g/l in isopropanol/water (1:1) was used at 130°C for 20 min. After repeated washings the pretreated samples were processed the same way as the cellulosic materials. 2.5. Measurement of antimicrobial activity of chitosan and of treated textiles Laser nephelometry using the microtiter plate method for monitoring fungal growth was used to determine the antimycotic influence of chitosan during microbe cultivation which is a turbidimetric method yielding numerous data in relatively short time as described in Refs [8, 9]. This method uses the SabouraudGlucose-Bouillon (SGB) as the cultivation medium and 3¥105 cells/ml (cell forming units, cfu) of fungus as inoculum. The principle of this turbidimetric method is given in Fig. 4.
198
D. Knittel and E. Schollmeyer
Figure 4. Principle of nephelometric measurements of fungal growth. (A) Starting fungal colony, (B) turbid suspension of fungi.
Figure 5. Reaction mechanism of the tetrazolium/formazan redox couple for the determination of the viability of microbes (TTC test method).
This actively growing broth culture was adjusted with sterile saline to a final working concentration of 6¥107 colony forming units (cfu) per ml. The cell cultures of Candida species were counted using the cell counter CASY® 1 and adjusted to a final working concentration of 6¥105 cells per ml in SGB buffer. The triphenyltetrazolium (TTC)/triphenylformazan redox couple is well known in biochemistry for determination of the viability of microorganisms and can be also used on surfaces (Fig. 5) [10]. For testing of finished fabrics, circular swatches of test and control textile materials were cut to identical diameter of 3.8±0.1 cm. 6 test and 6 control swatches were sterilized at 110°C. All swatches were stacked in 40 ml nutrient broth medium containing 10 µl of E. coli (108 cfu/ml) as an inoculum, then all flasks were incubated by shaking at 37°C at 200 rpm for 3–4 h. Then 1 ml from each flask containing the test fabrics and the control was added to sterilized test tubes containing 100 µl TTC (0.5%). All tubes were incubated at 37°C for 20 min. The resulting formazan was centrifuged at
Functionalization of fiber surfaces by carbohydrate biopolymers
199
Figure 6. Measuring cell for the particle charge detector (PCD); 1, Teflon vessel; 2, oscillating piston; 3, electrodes.
4000 rpm for 3 min, followed by decantation of the supernatants. The pellets obtained were resuspended in ethanol and centifuged again. The eventual red formazan solution which indicates the activity and viability of the cells was measured photometrically at 480 nm. Antibacterial efficiency was additionally tested in standard flask shake tests followed by colony counting. 2.6. Other analyses on treated fabrics 2.6.1. Characterization of biopolymer finished fabrics regarding accessible surface charges, by polyelectrolyte titration In this work the amount of charged groups of chitosan added onto cotton was measured by titration with a counter-charged polyelectrolyte solution with a known content of charged groups. The titration was monitored by a streaming potential measuring instrument which detects the sign of the charged groups and shows a streaming potential of the colloidal suspension in arbitrary units. The main element of the particle charge detector (PCD) is a cylindrical vessel (1) with an oscillating displacement piston (2) forcing the colloidal solution through the annular space between piston and bore. The movement of the piston creates a streaming potential between the two electrodes (3) above and below the bore as described in Fig. 6. The adsorption of the polyelectrolyte onto the surface of the vessel and piston is effected by van der Waals forces. The flow forced by the oscillating piston causes the diffuse outer part of the electrical double layer to be sheared off. The induced streaming potential data are processed electronically and displayed on a monitor. Polyelectrolyte titrations were performed with poly-DADMAC solutions in case of anionically modified textiles and with PES-Na in case of cationic systems. The concentration used was 0.001 equivalents/l. The fabric was immersed in the
D. Knittel and E. Schollmeyer
200
polyelectrolyte for at least 2 h and the supernatant solution was back-titrated for the remaining polyelectrolyte. 1–2 g fabric was immersed in 50 ml of oppositely charged polyelectrolyte titer (PES-Na or DADMAC, respectively) and stirred for 2 h. After separation of the fabric an aliquot of the remaining solution was back-itrated with the corresponding oppositely charged polyelectrolyte standard. The zero volt point with the particle charge detector marked the end-point of titration. CTS treated fabrics were preconditioned in solutions of pH 4.66 prior to polyelectrolyte measurement [11]. 2.6.2. Ion exchange properties of biopolymer finished cotton fabric Treated fabric (1 g) was immersed in a solution of the corresponding metal ion (2¥10-4 mol/l, acidified with sulfuric acid to pH 3–4 to avoid hydroxide formation) and left for 2 days. Ion concentration of the remaining solution was determined by atomic absorption spectroscopy or standard titration (Ca2+, Zn2+). 2.6.3. Protein binding/adsorption Measurements of albumin adsorption on carrageenan-treated CO were made with a 0.1% protein solution. After immersing in the solution and washing, adsorbed protein was visualized by the ninhydrin reagent. By intense washing, the original carrageenan finish is regenerated. Quantitative assessment of protein binding was done by the Lowry test. 3. RESULTS AND DISCUSSION
In Tables 3 and 4 results are given for thermal fixation on CO of the model compounds chitosan (CTS) and carrageenan from which a practical fixation temperature for systematic investigations of 140oC was determined. The following studies dealt with permanent fixation of chitosan onto cellulosic material using different anchor molecules for chemical bond formation. The amount of an anchor chemical varied from about stoichiometric (calculated on monomer unit of the biopolymer) to only 5%. It was found that only a few anchoring points were necessary in order to have a high activity of the biopolymer (chitosan) chain [11]. But higher amounts give better graft yields. Table 3. Influence of treatment conditions (concentration of biopolymer in liquor, amount of anchor chemical and fixation temperature) for permanent finish of CTS obtained Sample code
CTS (g/l)
Cyanuric chloride (g/l)
Fixation temperature (oC)
Add-on reached after washing (wt%)
CTS 5 (Fluka) CTS 5 (Fluka) CTS 11 (MBP 21) CTS 12 (MBP 21)
4.4 4.4 6.2 12.4
0.47 0.47 2.0 4.0
100 140 140 140
0.5 0.8 3.2 3.7
Functionalization of fiber surfaces by carbohydrate biopolymers
201
The reactive triazines were expected to give high fixation rates. The cyanuric chloride possibly leads to higher cross-linking than the hydroxy- or methoxydichloro compound. Polycarbonic acids bind via intermediate anhydride formation [12]. They are more expensive but were chosen because of their low toxicity. Other anchors like Na-HDCT or chemicals based on bifunctional trialkoxysilanes (sol-gel process) give similar results regarding add-on (data not shown). Prior to examination of treated fabrics it had to be verified that CTS acted distinctly as antimicrobial. Figure 7 shows antimicrobial results for dissolved CTS at pH 6. These experimental parameters were chosen because some work in the literature had discussed the effect of acetic acid which had been used for dissolution of CTS [5]. The inhibitory effect of chitosan on E. coli, thus, is confirmed, even for almost neutral conditions (Fig. 7) when CTS is protonated only to an extent of around 50% of the maximum. Table 4. Results of permanent finish with carrageenan on CO (200 g/m2) (treatment liquor 20 g/l carrageenan, 2.0 g/l cyanuric chloride, pH of padding liquor 9.5; fixation temperature 140oC) Sample code
Biopolymer (g/l)
Add-on after fixation (wt%)
Add-on after washing (wt%)
Car 3 Car 5-2 Car 5-1 Car 6-1
15.0 12.0 12.0 12.0
– 3.0 – –
1.2 1.8 2.6 2.2
Figure 7. Influence of dissolved CTS on the reduction rate of E. coli (pH adjusted to 6; 25°C).
202
D. Knittel and E. Schollmeyer
Figure 8 shows the antimycotic effect of dissolved CTS as determined by laser nephelometry monitoring the fungal growth in the absence of dissolved CTS as well as with various CTS concentrations. Compared to a quick cell growth of the control sample (no CTS) even medium concentrations of CTS produced a distinct inhibition of the fungal cells. In the following, the results on add-on using different anchor chemicals and antimicrobial efficiency on treated cotton are summarized. Figs 9–11 show the percent add-on (wt%) achieved by the treatment conditions used on cotton with different anchor chemicals.
Figure 8. Influence of dissolved CTS at various dilutions on the growth rate of Candida krusei in dependence of incubation time (pH approx. 6) (nephelometric determination).
Figure 9. Graft yield of chitosan on cotton with cyanuric chloride (CNC) anchoring.
Functionalization of fiber surfaces by carbohydrate biopolymers
203
It is evident that a higher amount of an anchor chemical results in a higher weight gain in all cases. The loss of finish during washing is extremely high when using cyanuric chloride under the conditions used but with other anchor chemicals sufficient add-on remains. It should be noted that a finish of >5 wt% (which is obtainable when using high-molecular-weight CTS or high amounts of anchors) the textile becomes a rather stiff material. Figure 12 shows the antibacterial effect of the CTS-treated surface as determined with the TTC test. The lower the formazan absorption measured, the more bacteria have ceased metabolism. Even if there is no sterilizing effect, a distinct bacteriostatic, permanent textile finish is achievable.
Figure 10. Effect of 2,4-dichloro-6-methoxy-1,3,5-triazine (MCNC) concentration as an anchor group on the graft yield of chitosan on cotton.
Figure 11. Effect of butane-1,2,3,4 tetracarboxylic acid (BTCA) concentration as an anchor group on the graft yield of chitosan on cotton.
204
D. Knittel and E. Schollmeyer
Figure 12. Effect of graft yield (wt%) of chitosan on the reduction of E. coli and Micrococcus luteus as measured by the TTC test method (low absorbance indicates less viability of microbes). Anchor group, butane-1,2,3,4 tetracarboxylic acid.
Figure 13. Bacteriostatic activity of PET modified with biopolymer employing different ways of anchoring (low absorbance indicates less viability of microbes), numbers give add-on before and after washing.
The results for biopolymer-treated polyester are shown in Fig. 13. It can be seen that, even if the biopolymer anchoring onto amino-functionalized polyester or treatment by alkylchitosans shows only low gravimetrically detectable add-ons (< approx. 0.3 wt%), efficient bacteriostatic action is still obtainable. Besides antimicrobial activity, other surface properties were determined. Biochemical activity of carrageenan-treated cotton showed a binding capacity for proteins (e.g., albumin) of about 35–40 mg bound protein per g add-on, which might be interesting for binding of proteinic allergens.
Functionalization of fiber surfaces by carbohydrate biopolymers
205
Table 5. Results of polyelectrolyte titrations on biopolymer-treated fabrics: evaluation of accessible charge on the biopolymer surface layer related to amount of add-on Polymer finish code
Add on (wt%)
Fabric treated (mequiv./g)
Biopolymer add-on (mequiv./g)
Alginate 5-2 (CO) Alginate (Tencel®)
1.4 1.45
0.025 0.8¥10
1.7 0.055
Alginate*
1.4
0.8¥10-3
0.07
2.1
-3
0.07
Carrageenan 19
*
1.6¥10
-3
*
Anchoring by oligomeric diisocyanate; for comparison the maximal available ionic charge in the solution of the pure biopolymer used is 4.6¥10-3 equivalents per g (alginate) and 2.4¥10-3 per g (carrageenan).
Further insight into the properties of biopolymer surface networks obtained, concerning the accessibility of the modified textile surfaces to molecules, can be gained by polyelectrolyte titrations and/or determination of ion binding capacities. Polyelectrolyte measurements made on fabrics can be used to assess the availability of ionic charges on the fabric surface which gives information on the (statistical) anchor sites and on the possible chain segment mobility of the biopolymeron-fiber (see Table 5). It is assumed that polymer chain segment mobility within the surface layer obtained determines the functionality envisaged. The results given show that the hitherto use of the anchor chemicals (mainly cyanuric chloride) results in availability of only one-third of large molecules as used in polyelectrolyte titration after biopolymer fixation onto textile (cf., Table 4) compared to bulk biopolymer charge equivalents. It can be shown that by varying the amount of anchor chemical a higher charge accessabiliy per g biopolymer fixed can be reached when there is lower add-on. 3.1. Ion-exchange properties of biopolymer-finished cotton fabric Since the biopolymers used for surface modification bear ionic functionalities (depending on the pH value) they may be used for complexation of heavy metal ions. Some investigations on the exchange capacity of biopolymer finished cotton were carried out and the results are presented in Table 6. The values for binding of Zn2+ (0.3 to 1.6 mmol/g add-on) are similar. For binding of Ca2+, only on pectin treated fabrics some complexation is observed but to a small degree (approx. 0.25 mmol/g add-on). Using the data for binding of small ions like Cu2+ or Zn2+, e.g., for 14 mg addon of alginate/g CO fabric, 4¥10-5 equivalents can be estimated as possible binding sites. The result for ion uptake is about 2.3¥10-5 equivalents. This means a distinctly higher accessability towards binding sites for these small inorganic ions than the access for polyelectrolyte titrants.
D. Knittel and E. Schollmeyer
206
Table 6. Data for binding capacity of biopolymer treated CO fabric for copper ions Fabric treated
Biopolymer (mg/g CO fabric)
Bound Cu2+ (mmol/g fabric)
Bound Cu2+ (mmol/g biopolymer add-on)
Original CO Alginate Carrageenan Chitosan Pectin
– 14 34 47 26
0.0017 0.025 0.015 0.025 0.035
– 1.73 (110) 0.42 (27) 0.54 (34) 1.35 (86)
3.2. Dermatologic evaluation of biopolymer finished cotton Several CO T-shirts were finished (2.3 wt% alginate, 4–6.2 wt% pectin, 3.7 wt% chitosan and 0.6–1.9 wt% carrageenan finishes) and used for dermatologic studies (Finn-chamber test and wearing test). The T-shirts were evaluated by patients with atopic eczema and were found to be very comfortable. None of the carbohydrate-finished fabrics produced skin irritation. 4. CONCLUSIONS
It was shown that chitosan imparted antimicrobial activity to cotton fabrics even when tested under nearly neutral pH conditions. The finish is expected to be durable because of chemical bonding. Analogous to chitosan finishing, other ionic or neutral carbohydrates, like carrageenan or dextrans, can be permanently bound to fiber surfaces. Future research will be aimed at testing the effect against other microbe types and to introduce chitosans of different molecular weights (the lower ones are preferred because of processing demands such as low viscosity of solutions) and at treatment of other fiber types. Also, work will be directed towards enhancement of ion- and protein binding capacities of biopolymer-treated fabrics. Even if there is a distinct difference in graft yields using different anchors, the degree of antimicrobial activity seems to be similar. As far as textile properties, like handle, have to be maintained add-ons of less than 5% are advisable. No irritation to human skin by biopolymer-treated fabrics was evidenced. The following fields for application of biopolymer treated textiles are envisaged: – Bioactive fibers for medicine, – Creation of textiles with improved properties for wound dressing, – Textiles for sensitive skin (allergics, neurodermitics), – Textiles with ion-binding capacities (e.g., for Ag+ ions), – Polymers on textile surfaces with depot function for drugs for slow relase.
Functionalization of fiber surfaces by carbohydrate biopolymers
207
Acknowledgements We would like to thank the Forschungskuratorium Textil e.V. for funding this research project (AiF No. 13519). This project was funded with financial resources of the Bundesministerium für Wirtschaft und Arbeit (BMWA) with a grant from the Arbeitsgemeinschaft industrieller Forschungsvereinigungen "Otto-vonGuericke" e.V., AiF). Also thanks are given to the Federal Ministry for Bildung und Forschung (BMBF) of the FRG for its financial support, grant No. 033459. REFERENCES 1. 2. 3. 4.
D. Knittel and E. Schollmeyer, J. Textile Inst. 91, 151–165 (2000). G.A.F. Roberts, Chitin Chemistry, MacMillan, London (1992). M.F.A. Goosen, Application of Chitin and Chitosan. Technomic, Lancaster, PA (1997). A.A. Muzzarelli and C. Muzarelli (Eds.), Chitosan in Pharmacy and Chemistry. ATEC, Grottamare (2002). 5. S.-H. Lim and S.M. Hudson, J. Macromol. Sci., Part C, Macromol. Rev. C43, 223–269 (2003). 6. R. Knittel, R. Stehr and E. Schollmeyer, German Patent 196 24 170.7 (1996). 7. D. Knittel and E. Schollmeyer, Proc. 3rd Int. Avantex Symp., held in Frankfurt/Main, V8_KNI.PDF (2005). 8. M. Fouda, D. Knittel, C. Hipler, R. Zimehl and E. Schollmeyer, Adv. Chitin Sci. 8, 418–425 (2005). 9. M. Fouda, A. Nickisch-Hartfiel, D. Knittel, R. Zimehl and E. Schollmeyer, Proc. 2nd Int. Conf. Textile Res. Div. National Research Centre, Cairo, pp. 109–115 (2005). 10. E. Seidler, The Tetrazolium-Formazan System: Design and Histochemistry, pp. 1–86. G. Fischer, Stuttgart (1991). 11. D. Knittel and E. Schollmeyer, Melliand Textilber. 83, E15–E16 (2002). 12. C.Q. Yang, Textile. Res. J. 63, 706–711 (1993).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 209–218 Ed. K.L. Mittal © VSP 2007
Dendrons for surface modification of polymeric materials HANS-JÜRGEN BUSCHMANN∗ and ECKHARD SCHOLLMEYER Deutsches Textilforschungszentrum Nord-West e.V., Adlerstrasse 1, D-47798 Krefeld, Germany
Abstract—Dendrons and dendrimers are molecules with a high but defined molecular weight. They are synthesized in several steps and their molecular shape reminds one of a tree. Due to their molecular architecture, dendrons and dendrimers have three-dimensional cavities of different sizes. Dendrons have been permanently attached to cellulosic and poly(ethylene terephthalate) fibers and foils. Due to the presence of the dendrimers, both the physical and chemical properties of the fiber or foil surfaces are modified. Using fluorescence dyes the presence of the dendrons on the fiber surface is shown. Depending on the environment of the dyes the fluorescence signal either increases or is quenched by interactions with the dendrons. Also differental scanning calorimetry is used to demonstrate the presence of the dendrons. Keywords: Dendrons; dendrimers; surface modification; supramolecular chemistry.
1. INTRODUCTION
The progress in polymer chemistry can be separated in different stages of development. After the first synthesis of linear polymers the next step was the synthesis of cross-linked polymers, followed by branched and hyperbranched polymers. Typical for all of these polymers is the distribution of molecular weights depending on the conditions during the polymerization reactions. However, the synthesis of molecules with a defined high molecular weight was first realized during the synthesis of supramolecular ligands able to form complexes with cations, anions or neutral guest molecules. This work was recognized in 1987 with the awarding of Nobel Prize in Chemistry to C. Pedersen, J.-M. Lehn and D. Cram [1]. Using similar synthesis strategies Vögtle had synthesized the first dendritic molecules already in 1978 [2]. Since that time the number of publications has increased continuously. The most important results with respect to the synthesis [3–6] and applications [7–9] of dendrimers and dendrons have been summarized in books and reviews.
∗
To whom correspondence should be addressed. Tel.: (49-2151) 843-210; Fax: (49-2151) 843-143; e-mail:
[email protected]
210
H.-J. Buschmann and E. Schollmeyer
The definition and differentiation of dendrimers and dendrons can be understood from the schemes given in Fig. 1. Dendrimers are spherical molecules and dendrons are part of them, like a tree and its branches. The dendrimers and dendrons have regular and repeated branched structures. They are obtained from a repeated synthesis starting from a single molecule. The first branched molecule obtained from the first synthesis is, therefore, called the first generation. By a subsequent reaction, molecules belonging to the second generation are obtained and so on. The properties of dendrimers have been studied mostly in solution [7–9]. Dendrimers with nonpolar skeletons are able to enclose hydrophobic substances in their cavities. The presence of polar groups in the skeleton, e.g., amino groups, results in the binding of polar molecules.
Figure 1. Schematic representations of a third-generation dendrimer (left) and dendron (right).
Figure 2. Synthesis of dendronized polymers from dendrons of the third generation.
Figure 3. Fixation of third-generation dendrons on a polymer surface.
Dendrons for surface modification of polymeric materials
211
A different approach for the synthesis of polymeric dendrons has been realized by Schlüter and co-workers [10–13]. They used reactive monomers carrying a dendron as a substituent. After polymerization of these monomers they obtained polymers with structurally perfect dendritic side chains, see Fig. 2. The adsorption of different dendrons on cellulose and polyester has been reported [14]. Up to now no attempts have been reported to fix dendrons on polymer surfaces such as fibers and foils. This reaction principle is schematically shown in Fig. 3. The grafting reaction will not result in completely dendronized surfaces. Depending on the size of the dendrons their numbers of moles present on a given surface will vary. However, one can expect that some dendrons will cover the polymer surface almost completely due to their flexible structure. Thus, the surface properties of the polymers will be modified, depending on the nature of the dendron. The attachment of a hydrophobic dendron onto a cellulosic fiber should result in an increased hydrophobicity of the fiber and the attachment of a polar dendron on a nonpolar fiber should enhance the adsorption of water molecules on the surface. Other changes in the surface properties of the modified polymers are possible. The aim of this work was to attain permanent fixation of dendrons on different fibers. For the fixation of dendrons on cellulosic materials a non-polar dendron and for the fixation on poly(ethylene terephthalate) (PET) a polar dendron was selected. 2. EXPERIMENTAL
Two different types of fabrics were used. The cotton fabric (BW-Echtheitsgewebe S/400, testex Prüftextilien, Bad Münstereifel, Germany) was already bleached and desized. The poly(ethylene terephthalate) (PET) fabric was a commercial sample (mass per unit area: 73.2 g/m2). Both fabrics were extracted using chloroform, hexane, acetone, ethyl acetate and a mixture of methanol and water (50/50 vol%). For the experimental studies two different dendrons, one non-polar and one polar, were chosen, see Fig. 4. All chemicals used for the synthesis were commercially available. The synthesis of the non-polar dendrons has been described in detail in the literature [15, 16]. Starting with 3,5-dihydroxybenzyl alcohol the first-generation dendrimer after reaction with benzylbromide was obtained. This reaction sequence was repeated with the first-generation dendrons to obtain the secondgeneration dendrons. For the fixation of these dendrimers onto cellulosic fibers, the hydroxyl group was reacted with cyanuric chloride. The resulting derivative was dissolved in acetone and the pH value adjusted to pH 5 by the addition of dilute aqueous NaOH solution. The cotton fabric was dipped for 5 min into the solution to achieve complete wetting. Afterwards the impregnated fabric was dried to remove the acetone and thereafter heated at 100°C for 10 min. The fabrics were washed several times with acetone and water to remove the nonfixed
212
H.-J. Buschmann and E. Schollmeyer
dendrons. This procedure is nearly identical with that used for the fixation of monochlorotriazinyl-substitued β-cyclodextrin [17, 18]. The synthesis of the amino group containing dendrons strictly followed the procedure described by Vögtle and co-workers [2]. As starting molecule heptylamine was used. After the reaction with acrylonitrile and reduction of the cyano groups the first-generation dendron was obtained. This procedure was repeated to obtain the second- and third-generation dendrons. Due to the presence of the heptyl group in these dendrons they could be fixed easily onto PET surfaces as the heptylgroup diffused into the fibre. These dendrons were dissolved in aqueous solution and the poly(ethylene terephthalate) fabric was immersed and treated for 1 h at 130°C. The fabric was washed several times with distilled water and dried. The textile samples were immersed into the liquids for several minutes to achieve optimal interactions and complete complex formation between dendrons and solvent molecules. Afterwards the samples were dried at room temperature for several hours to remove adsorbed solvent molecules. Differential scanning calorimetry on these samples was performed using a TA Instruments DSC 2010. The scan rate was 20°C/min under a nitrogen atmosphere. The textile samples were immersed into aqueous solutions of the fluorescence dyes 8-anilino-1-naphthalene sulfonic acid (ANS) (λex= 273 nm) (1 g/l) and Rhodamine B (λex= 264 nm) (1 g/l). The samples with the complexed dyes were washed with water several times, until colorless solutions were obtained and then dried at room temperature before recording the spectra. Fluorescence spectra of the solid samples were recorded using a Shimadzu RF-5001PC spectrometer using a solid sample holder. To study the effects of fixing the dendrons on the wetting behavior of fabrics, a drop penetration test was performed. 10 µl water was placed onto the polymer surface from a pipette at constant distance and the time for penetration was
Figure 4. Chemical structures of a third-generation non-polar dendron D1 and a polar dendron D2.
Dendrons for surface modification of polymeric materials
213
determined [17]. In the case of very hydrophobic surfaces no penetration was observable. The nitrogen content of PET fabrics with different generations of dendron D2 was determined from pH-metric titrations. A textile sample was immersed into diluted hydrochloric acid. Afterwards the fabric was washed with distilled water and dried at 120°C. 1 g of the textile sample was immersed into 20 ml of water and this solution was titrated with 0.02 mol/l NaOH. From the titration curves the amount of protonated nitrogen atoms was calculated. 3. RESULTS AND DISCUSSION
The fixation of nonpolar dendrons onto cellulosic materials is schematically shown in Fig. 5. The strategy for the fixation of dendrons is similar to dyeing of cellulosic materials with reactive dyes. The monochlorotriazinyl group of the dendron reacts with one hydroxyl group of the cellulosic material. As a reasult a new chemical bond is formed. Due to the presence of the nonpolar dendrons on the fiber surface the chemical properties of the surface are altered. A very sensitive indication for the presence of hydrophobic structures on the surface can be obtained from the use of solvatochromic fluorescence dyes like ANS. In a hydrophilic environment no fluorescence signal can be detected. However, in nonpolar surroundings the fluorescence signals increase [19]. Even in an aqueous solution this effect is used to ascertain the formation of inclusion complexes of ANS with cyclodextrins and other guest molecules [20, 21]. In Fig. 6 the fluorescence spectra of ANS adsorbed onto cotton fabrics modified with different generations of the nonpolar dendron D1 are shown.
Figure 5. Fixation of dendrons at cellulosic material using dendrons with a monochlorotriazinyl group.
214
H.-J. Buschmann and E. Schollmeyer
Figure 6. Fluorescence spectra of ANS adsorbed onto a cotton fabric modified with different generations of the dendron D1 (―, untreated cotton; – –, cotton treated with cyanuric acid; ·····, 3,5dihydroxybenzyl alcohol; – • – •, generation 1; – •• – ••, generation 2; , generation 3).
Figure 7. Differential thermal analysis results for cotton fabrics modified with different generations of the non-polar dendron D1 (a, reference; b, first generation; c, second generation; d, third generation) with adsorbed water (──) or toluene molecules (– – –).
Dendrons for surface modification of polymeric materials
215
The intensity of the fluorescence signal increases with the generation number of the dendron fixed onto cellulose. Obviously the polarity around the ANS molecules is reduced due to the presence of the dendrons. The shielding of the dye molecule from the polar environment of the cellulosic material becomes more effective with higher generations of the dendron D1. With the fluorescence dye Rhodamine B one finds the opposite behavior. The fluorescence signal has a high intensity even in polar surroundings. The benzene groups of the dendrons D1 enable an energy transfer and as a result the fluorescence signal is quenched. With an increasing number of benzene groups in the proximity of the dye molecule the quenching becomes more efficient. Using differential thermal analysis one obtains information about the melting or evaporation of substances. Cellulosic material easily adsorbs water molecules. Upon heating of cellulosic materials the vaporization of adsorbed water molecules reaches a maximum value at temperatures slightly higher than 100°C due to their interactions with the hydroxyl groups of the cellulose molecules. As shown in Fig. 7, the vaporization of adsorbed water molecules is influenced by the presence of the nonpolar dendron D1. A similar behavior is also found for the vaporization of toluene. The fixation of the polar dendron D2 onto PET fibers is possible due to the presence of the heptyl group at the central nitrogen atom. The alkyl chain diffuses into the polymer matrix if the temperature is raised above the glass transition point of PET. This procedure is comparable to dyeing of PET fabrics using disperse dyes. The fixation of different generations of the polar dendron D2 onto the surface of a PET fiber is schematically shown in Fig. 8.
Figure 8. Fixation of different generations of the polar dendron D2 onto the surface of a PET fiber.
216
H.-J. Buschmann and E. Schollmeyer
Table 1. The amount of nitrogen atoms nN fixed onto PET fibers modified with hexylamine (R-NH2) and different generations of dendron D2 Material
nN (mol/g fabric)
PET (reference) R-NH2
–
Generation 1
(15.6±0.8) ¥ 10-6
Generation 2
(14.1±0.5) ¥ 10-6
Generation 3
(16.8±0.7) ¥ 10-6
(3.5±0.3) ¥ 10-6
Figure 9. Differential thermal analysis results for PET fabrics modified with different generations of the polar dendron D2 and treated with aqueous formic acid (reference (──), generation 1 (─ ─ ─), generation 2 (· · · ·), generation 3 (─ · ─ · ─)).
The amount of the polar dendron D2 fixed onto the surface of the PET fibers can be calculated from pH-metric titrations. The results are given in Table 1. The results from Table 1 give no information about the degree of coverage of the fiber surface. With increasing number of generation, the dendron D2 requires more space on the PET surface for each individual dendron. As a result, the number of amino groups fixed onto the surface is almost constant. Further evidence for the fixation of the dendrons is obtained from the differential thermal analysis of PET fabrics modified with polar dendrons and treated with a protic solvent. This solvent strongly interacts with the nitrogen atoms of the dendron D2. Thus, different generations of this dendron should influence differ-
Dendrons for surface modification of polymeric materials
217
ently the evaporation of this solvent. Treatment of the modified PET fibers with aqueous formic acid gives the results shown in Fig. 9. The signal below 100°C is related to the evaporation of water molecules from the surface of fabrics. This process is influenced by the presence of the dendrons. In the presence of dendrons a signal above 150°C clearly appears; this signal depends on the thermal treatment of the PET fabrics [22]. 4. CONCLUSIONS
All the results presented show that it is possible to fix dendrons onto polymeric materials. Though the layer formed by the dendrons is thin compared to the dimensions of the fibers it is possible to detect the presence of the dendrons using different techniques. This should lead to interesting applications in the field of technical textiles as garments. Acknowledgements We thank the Ministry of Science and Research of Nordrhein-Westfalen for financial support. This support was granted for the project “DTNW/assistance to fund raising of third-party funds”. REFERENCES 1. 2. 3. 4. 5.
M. Dobler, Chimia 41, 416 (1987). E. Buhleier, W. Wehner and F. Vögtle, Synthesis, 155 (1978). H.-B. Mekelburger, W. Jaworek and F. Vögtle, Angew. Chem. Int. Edn. Engl. 31, 1571 (1992). G.R. Newkome, C.N. Moorefield and F. Vögtle, Dendritic Molecules. VCH, Weinheim (1996). J.M.J. Fréchet and D.A.Tomalia (Eds.), Dendrimers and Other Dendritic Polymers. Wiley, New York, NY (2001). 6. D.A. Tomalia, Aldrichim. Acta 37, 39 (2004). 7. D. Astruc and F. Chardac, Chem. Rev. 101, 2991 (2001). 8. M.W.P.L. Baars and E.W. Meijer, Top. Curr. Chem. 210, 131 (2000). 9. W. Krause, N. Hackmann-Schlichter, F.K. Maier and R. Müller, Top. Curr. Chem. 210, 261 (2000). 10. A.D. Schlüter and J.P. Rabe, Angew. Chem. Int. Edn. 39, 864 (2000). 11. L. Shu, A.D. Schlüter, C. Ecker, N. Severin and J.P. Rabe, Angew. Chem. Int. Edn. 40, 4666 (2001). 12. A. Zhang, L. Shu, Z. Bo and A.D. Schlüter, Macromol. Chem. Phys. 204, 328 (2003). 13. A.D. Schlüter and J.P. Rabe, in: Encyclopedia of Polymer Science and Technology Vol. 2, J.J. Kroschwitz (Ed.), pp. 135–171, Wiley, New York, NY (2003). 14. J.M.J. Fréchet, I. Gitsov, T. Monteil, S. Rochat, J.-F. Sassi, C. Vergelati and D. Yu, Chem. Mater. 11, 1267 (1999). 15. C.J. Hawker and J.M.J. Fréchet, J. Am. Chem. Soc. 112, 7638 (1990). 16. K.L. Wooley, C.J. Hawker and J.M.J. Fréchet, J. Am. Chem. Soc. 113, 4252 (1991). 17. U. Denter and E. Schollmeyer, J. Inclusion Phenom. Macrocycl. Chem. 25, 197 (1996). 18. H.-J. Buschmann, D. Knittel and E. Schollmeyer, J. Inclusion Phenom. Macrocycl. Chem. 40, 169 (2001).
218
H.-J. Buschmann and E. Schollmeyer
19. A. Azzi, Q. Rev. Biophys. 8, 237 (1975). 20. B.D. Wagner and P.J. MacDonald, J. Photochem. Photobiol. A 114, 151 (1998). 21. B.D. Wagner, P.J. MacDonald and M. Wagner, J. Chem. Ed. 77, 178 (2000). 22. H.-J. Bernd and A. Bossmann, Polymer 17, 241 (1976).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 219–227 Ed. K.L. Mittal © VSP 2007
Surface modification of textile materials by dip-coating with magnetic nanoparticles J. ZORJANOVIĆ,1,∗ R. ZIMEHL,2 E. SCHOLLMEYER,3 O. PETRACIC,4 W. KLEEMANN,4 D. KNITTEL,3 T. TEXTOR3 and U. SCHLOßER3 1
Süd-Chemie AG, BAA R&D, Steinbockstr. 12, 85368 Moosburg, Germany Merz Dental GmbH, Eetzweg 20, 24321 Lütjenburg, Germany 3 Deutsches Textilforschungszentrum Nord-West e.V., Adlerstr. 1, 47798 Krefeld, Germany 4 Laboratorium für Angewandte Physik, Universität Duisburg-Essen, 47048 Duisburg, Germany 2
Abstract—In this work an efficient way of dip-coating of textile materials with magnetic nanoparticles is presented. Two ferrofluid samples of iron oxide for coating of polyester filaments were used. One sample was AdNano MagSilica (Degussa), consisting of γ-Fe2O3 (maghemite), Fe3O4 (magnetite) and small amounts of α-Fe2O3 (hematite) nanoparticles 5–20 nm in size. The second sample was synthesized by precipitation from solution which yielded particle sizes of 5–12 nm. In order to prevent the formation of aggregates or agglomerates during the preparation, the particles were stabilised by embedding them in a polysiloxane matrix. Finally the coating of textiles with the nanoparticles was carried out by dip-coating. After heating at 130°C for 3 h the textile materials showed superparamagnetic behaviour. Possible sensor and memory applications are briefly discussed. Keywords: Superparamagnetic; textile materials; polysiloxane; iron oxide nanoparticles.
1. INTRODUCTION
Ferromagnetic materials, such as certain iron oxide compounds, become superparamagnetic (SPM) if the size of the magnetic particle falls below a critical size, usually being in the nanometer range [1]. The critical diameter of, e.g., Fe3O4 particles is 25.5 nm [1]. SPM nanoparticles offer a high potential for several applications in different areas such as magnetic data storage, refrigeration, color imaging, controlled transport of anti-cancer drugs, or for separation of biomolecules from solutions [2, 3]. The individual magnetic moments of the atoms in ferromagnetic materials are coupled into domains, so-called Weiss districts. In these domains all magnetic moments are oriented parallel to each other. Their sum forms a mesoscopic magnetic moment. If one can reduce the particle size of a ferromagnetic material until ∗
To whom correspondence should be addressed. Fax: (49-8761) 82-203; e-mail:
[email protected]
220
J. Zorjanović et al.
only a single magnetic domain remains, to each particle one can assign a single large magnetic moment (superspin), which can either rotate or switch between certain magnetic axes. In SPM particles, these reversals can occur spontaneously due to thermal fluctuations. A stable orientation of the magnetic moment thus is not possible. A characteristic temperature of SPM systems is the so-called blocking temperature (TB), below which the particle moments appear frozen on the laboratory time-scale. In the simplest case of a particle with uniaxial magnetic anisotropy involving a potential energy barrier of height ∆E, it is given by inverting the relaxation time τ = τ0 exp(∆E/kBT) for T ≡ TB, where τ0 ≈ 10-10 s is a typical spin-precession time and τ ≡ τmeas equals the characteristic time of the measurement (e.g., τ ≈ 102 s for a SQUID Measurement). Above TB, the moments are unblocked and, hence, are “switchable” by an arbitrarily weak applied field. There are many modifications of iron oxides, but only two modifications are relevant for this study, viz., γ-Fe2O3 (maghemite) and Fe3O4 (magnetite). In the literature many ways for the production and stabilization of maghemite and magnetite are described [4–10]. Both dry and wet chemical methods are suggested [8]. Furthermore, it is necessary to prevent the formation of aggregates or agglomerates of particles; otherwise no individual SPM but collective behaviours are encountered [1, 9]. Consequently, the particles have to be stabilized by colloid chemistry processes [10, 14]. One possibility is the encapsulation of the particles with polymers [4, 10] or metal oxides, e.g., SiO2 [5, 8]. Only a few studies describe the stabilization of iron oxide particles by embedding them in a matrix [11–13]. In this work we present a method for the stabilization of maghemite and magnetite particles in a polysiloxane matrix for coating of synthetic textile materials (e.g., poly(ethylene terephthalate) (PET)). By this one can produce flexible textile materials with “switchable” magnetic properties. 2. EXPERIMENTAL
For the coating of polyester with iron oxide particles two different samples were used. The sample M1, AdNano MagSilica (Degussa) consists of γ-Fe2O3 (maghemite), Fe3O4 (magnetite) and small amounts of α-Fe2O3 (hematite) in a shell of silica forming the non-magnetic matrix [8]. Sample M1 was synthesized in a hydrogen-air flame, whereas sample M2 was synthesized by wet chemical processes and consisted of Fe3O4 only. 2.1. Synthesis of Fe3O4 The synthesis of Fe3O4 was according to Massarat [8]. 20 ml of a 1 mol/l FeCl3·6H2O (Merck) solution in water was mixed with 5 ml of a solution of a 2 mol/l FeCl2·4H2O (Merck) in 2 mol/l HCl (Merck). Then the mixture was added quickly to 250 ml of 0.7 mol/l ammonia solution (Merck) under stirring. After several minutes a black precipitate was formed. After 60 min of stirring, the pre-
Dip-coating of textiles with magnetic nanoparticles
221
cipitate was separated from the solution by a permanent magnet and washed three times with 2-propanol (Roth). 2.2. Stabilization of particles For the stability of a dispersion, both repulsive and attractive forces between the particles are responsible [13]. According to the DLVO theory the repulsive forces result from the electrostatic repulsion between the particles, if they have the same charge. Attractive forces between the particles originate from the weak van der Waals interaction. By steric stabilization one can influence both the van der Waals and the repulsion interactions between the particles. Steric stabilization occurs if macromolecules surround a particle by adsorption or by covalent bond formation so that they form a layer on the surface of the particles. This layer can prevent the particles to form aggregates and the particles are said to be sterically stabilized [13]. In the following the steric stabilization of iron oxides in 2propanol is outlined. The prepared iron oxide particles have OH groups on their surface (since surface atoms of the iron oxide react with water forming hydroxyl groups, Fig. 1). To prevent the formation of agglomerates or aggregates of the dispersed particles in the 2-propanol, the dispersion was treated as follows: a part of the dispersion was treated in an ultrasonic bath for 30 min with 3-glycidoxypropyltrimethoxysilane (GPTMS, Crompton, Silquest A-187® silane). Then 0.1 mol/l HCl (Merck) was added to the dispersion and stirred at room temperature for 20 h. A model for the reaction between the OH groups of metal oxides and GPTMS is presented in Fig. 2. The first step is the reaction between the OH groups on the particle surface and RO groups of the GPTMS (Fig. 2). Around the particle a layer is formed which partly contributes to its stability. In the presence of HCl as a catalyst the condensation reaction continues between the
Figure 1. Model for formation of OH-groups at the metal oxide surface in an aqueous solution [14].
222
J. Zorjanović et al.
Figure 2. Model of the steric stabilization of iron oxide particle in 2-propanol with GPTMS.
GPTMS monomer at the particle surface and further groups outside in solution (step II of the reaction sequence, Fig. 2). The particles thereby attain a layer of polysiloxane and yield a very stable iron oxide dispersion because the particles are sterically stabilized. 2.3. Pretreatment of polyester surfaces Polyester material (Hostaphan, KoSa, Bobingen, Germany) has to be modified with functional groups at the surface in order to link it with the reactive epoxy functional groups of the iron oxide surface. Since in a basic environment (pH approx. 8.5) an amino group can react with the epoxy group of the GPTMS [15], the polyester was first treated with 1-dodecylamine (Merck) in an autoclave at 130°C. By this process the hydrophobic alkyl chain of the 1-dodecyl amine is embedded in the polyester matrix. According to the principle of disperse dyeing the free hydrophilic amino groups remain on the polyester surface [16]. As the amino groups do not bind tightly at the fabric surface but remain free on the fibre surface, the amine groups bound on the polyester surface were quantitatively determined with TNBA (2,4,6-trinitrobenzene-1-sulfonic acid, (Fluka) 1% in N,N-dimethylformamide). TNBA reacts in partly aqueous solution (at pH 8) at room temperature exclusively with primary amino groups [17]. The regeneration of the amino groups was accomplished by treatment with a 1 M sodium hydroxide (Merck) with release of trinitrophenyl derivative (picric acid anion) whose concentration was determined by spectrophotometry (at 353 nm). The maximum coverage of the amino groups on the polyester surfaces was found to be approx. 10 nmol/mm2.
Dip-coating of textiles with magnetic nanoparticles
223
Figure 3. Coating process of functionalized polyester material (B) with stabilized iron oxide particles (A) and attaching the particles via covalent bonds (C).
2.4. Coating of polyester with particle dispersion The iron oxide dispersion was mixed with TMAH (tetramethyl ammonium hydroxide, Merck, 25% in methanol) under stirring until the dispersion reached pH 8. The dispersion was stirred for 2 h at room temperature. Then the polyester surface was coated with this dispersion by dip-coating. Finally the polyester samples were heated for 3 h at 130°C. The coating process is represented schematically in Fig. 3. 2.5. Characterization The particles sizes were determined by transmission electron microscopy (LEO, Oberkochen, Germany). The X-ray diffraction (XRD) patterns were obtained with a Siemens D5000 system (Cu Kα radiation, λ = 1.544 nm). For determination of trinitrophenyl derivative, a Cary 5E UV-Vis-NIR spectrophotometer (λ= 353 nm) was used. The magnetic properties were measured with a superconducting quantum interference device (SQUID) magnetometer (MPMS-5S, Quantum Design, San Diego, CA, USA) in the temperature range between 10 to 300 K. 3. RESULTS AND DISCUSSION
3.1. Transmission electron microscopy TEM micrographs of both samples M1 and M2 are shown in Fig. 4. The iron oxide particles (sample M1) are in the range of 5–20 nm. They are homogeneously distributed in the amorphous siloxane matrix (Fig. 4a). The iron oxide particles of the sample M2 are in the range 5–12 nm. Figure 4b shows the agglomerates of small particles obtained after drying.
224
J. Zorjanović et al.
3.2. X-ray measurements The sample M1 consist of γ-Fe2O3 (maghemite), Fe3O4 (magnetite) and small amounts of α-Fe2O3 (hematite). In contrast, the XRD measurement on sample M2 (Fig. 5) indicates only Fe3O4 (magnetite). The XRD peaks in Fig. 5 agree very well with the theoretical values for Fe3O4 (magnetite). In Fig. 5 only the hkl values of Fe3O4 are shown. There is no indication for the presence of maghemite. The typical peak of maghemite at 2 Q = 15.00 (hkl = 110) is not visible.
Figure 4. TEM micrographs of the samples M1 (a) and M2(b).
Figure 5. X-ray diffraction pattern of the sample M2.
Dip-coating of textiles with magnetic nanoparticles
225
3.3. Magnetization measurements The magnetization of samples M1 and M2 was studied as a function of temperature at a constant magnetic field strength. Both samples were measured in two forms, viz., on polyester and as powder (Figs 6 and 7). In all four cases one finds a splitting of the ZFC (zero field cooling) and FC (field cooling) curves) which is typical of SPM particles. The procedure was first to cool down the sample in a zero field from 300 K to 10 K, then switch on a magnetic field of 100 G and warm the sample up in the field until 300 K (ZFC curve) and then cool it down again in the field to 10 K (FC). For each sample the two forms show virtually no
Figure 6. Magnetization m of sample M1 as a function of temperature at constant field (B = 100 G).
Figure 7. Magnetization m as a function of temperature at constant field strength for the sample M2 (B = 100 G).
226
J. Zorjanović et al.
difference in magnetization. However, the blocking temperatures, TB (i.e., the characteristic temperature where the particle moment becomes "blocked" or "frozen" for T < TB) are different for M1 and M2. The blocking temperature for M2 is TB approx. 100 K (Fig. 7), while for M1 it is obviously higher than the room temperature (Fig. 6). This is obviously due to smaller particles size, on average, in M2 compared to M1. One should note that the M2 particles on polyester (Fig. 7) exhibit a lower blocking temperature, TB ≈ 95 K, than the same sample as powder, TB ≈ 145 K. This is probably due to a wider particle size distribution in the latter case. This is a direct consequence of agglomeration of particles in the dry condition, while on polyester the particles are more uniformely distributed. As a consequence, fabrics coated with magnetic nanoparticles similar to sample M1 will retain their magnetic moment when magnetized at room temperature, while those coated with nanoparticles like M2-type will respond to an external magnetic field solely by enhanced susceptibility, but will show no remancence. Thus, it should be possible to provide filaments, depending on the magnetic particle size, either with local non-volatile memory (sample M1) or with magnetic switchable sensor properties (sample M2). 4. CONCLUSIONS
We presented a simple way to produce polyester materials with superparamagnetic properties. First, the polyester material was functionalised with amino groups by incorporation of dodecylamine in the polyester matrix. Then the iron oxide particles were stabilized in 2-propanol with GPTMS. Since the amino groups react with the epoxy groups of GPTMS, the polyester material was coated by dip-coating. Finally the polyester material was heated for 3 h at 130°C. The modified samples provide new textile functionalities. Acknowledgements We would like to thank Prof. Dr. Mathias Hannig from the University of Saarland for transmission electron microscopy measurements, Dr. Heiko Frahm from the Christian-Albrechts University of Kiel, Institute for Inorganic Chemistry for the X-ray measurements and Degussa AG for the iron oxide sample. Financial support within the scope of the “DTNW/Unterstützung bei der Einwerbung von Drittmitteln” from the Ministerium für Wissenschaft und Forschung of Northrhine-Westphalia is gratefully acknowledged. REFERENCES 1. D.K. Kim, M. Toprak, M. Mikhailova, Y. Zhang, B. Bjelke, J. Kehr and M. Muhammed, Mater. Res. Soc. Symp. Proc. 704, 369–374 (2002).
Dip-coating of textiles with magnetic nanoparticles
227
2. P. Tartaj, M. del Puertos Morales, S. Veintemillas-Verdaguer, T. Gonzalez-Carreno and C.J. Serna, J. Phys. D: Appl. Phys. 36, R182–R197 (2003). 3. A. Jordan, Der Onkologe 7, 1073–1081 (2001). 4. Sh. Gu, T. Shiratori and M. Konno, Colloid Polym. Sci. 281, 1076–1081 (2003). 5. A.P. Philipse, M.P.B. van Bruggen and Ch. Pathmamanoharan, Langmuir 10, 92–99 (1994). 6. T. Hyeon, S.S. Lee, J. Park, Y. Chung and H. Bin Na, J. Am. Chem. Soc. 123, 12798–12801 (2001). 7. D.K. Kim, Y. Zhang, W. Voit, K.V. Rao and M. Muhammed, J. Magn. Magn. Mater. 225, 30– 36 (2001). 8. R. Massarat, IEEE Trans. Magn. MAG-17, 1247–1248 (1981). 9. D. Hoffmann, K. Landfester and M. Antonietti, Magnetohydrodynamics 37, 217–221 (2001). 10. C. Carcia, Y. Zhang, F. DiSalvo, and U. Wiesner, Angew. Chem. Int. Edn. 42, 1526–1530 (2003). 11. A.C. Finnefrock, R. Ulrich, A. Du Chesne, Ch.C. Honeker, K. Schumacher, K.K. Unger, S.M. Gruner and U. Wiesner, Angew. Chem. Int. Edn. 40, 1207–1211 (2001). 12. C.B.W. Carcia, Y. Zhang, S. Mahajan, F. DiSalvo and U. Wiesner, J. Am. Chem. Soc. 125, 13310–13311 (2003). 13. G. Lagaly, O. Schulz and R. Zimehl, Dispersionen und Emulsionen. Steinkopf, Darmstadt (1997). 14. K. Köhler and C.W. Schläpfer, Chem. Uns. Zeit 27, 248–255 (1993). 15. B.U. Kluß and R. Zimehl, Prog. Colloid Polym. Sci. 111, 144 (1998). 16. E. Schollmeyer, J. Zorjanovic, T. Textor, K. Opwis, D. Knittel and T. Bahners, Proc. 5th AUTEX Conf., Portoroz, 27–29.06.2005, pp. 20–28. 17. W.A. Bubnis and C.M. Ofner III, Anal. Biochem. 207, 129–133 (1992).
Part 2 Adhesion Improvement to Polymer Surfaces
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 231–240 Ed. K.L. Mittal © VSP 2007
Adhesion improvement of epoxy resist to a benzocyclobutene layer: Application to nanoimprint lithography BENOÎT VIALLET1,∗ and EMMANUELLE DARAN2 1
LNMO-INSA de Toulouse, Département de Physique, 135 avenue de Rangueil, 31077 Toulouse Cedex 4, France 2 LAAS-CNRS, 7 avenue du colonel Roche, 31077 Toulouse Cedex 4, France
Abstract—Nanoimprint lithography (NIL) is low cost technique used to reproduce nanometer-sized patterns. In this approach, a curable polymer is used and, thus, the adhesion between this polymer and the underlying layer is an important issue. The nanoimprinted polymer used in this study is epoxy-siloxane and the underlying layer is benzocyclobutene (BCB). The poor adhesion between epoxy-siloxane and untreated BCB does not allow NIL with nanometer size patterns. The effects of different surface treatments, UV/ozone and CF4/O2 plasma followed by deposition of SiOx, on the quality of imprinted patterns have been studied. The degradation of the surface layer of BCB by UV/ozone photochemical treatment has been analysed. Surface treatment based on combination of CF4/O2 plasma and SiOx deposition has been found to be efficient in improving the adhesion of epoxy-siloxane to BCB. Keywords: Adhesion; surface energy; nanoimprint; UV/ozone; plasma; polymer.
1. INTRODUCTION
Different materials are used in the fabrication of microsystems. One of the key issues in the success of the fabrication process is the good control of adhesion between the different materials involved. In this study, a treatment to improve adhesion of an epoxy polymer, used as nanoimprint resist, on benzocyclobutene (BCB) is presented. These two polymers are involved in the fabrication of a Bragg grating for integration into a 1.3-µm planar optical amplifier. One of the process steps is nanoimprint lithography. As will be discussed later, this technique needs good control of adhesion. The structure of the Bragg grating integrated in the amplifier is presented in Fig. 1. It is based on a polymer channel waveguide etched to form the grating. ∗
To whom correspondence should be addressed. Tel.: (33-5) 6155-9672; Fax: (33-5) 6155-9697; e-mail:
[email protected]
232
B. Viallet and E. Daran
Figure 1. Structure of optical amplifier based on a 335-nm period Bragg grating.
Figure 2. Cross-section of imprinted resist showing the residual layer left under the master.
Typical dimensions of this grating are a total length of 700 µm, a period of 335 nm and an etch depth of 500 nm. Polymer materials are interesting for this application because they offer the possibility to realize low-loss waveguides and can be processed without deterioration of active material properties. In particular, benzocyclobutene (BCB) has been used to realize a channel on a Nd3+:LaF3 layer [1] with a propagation loss of 1 dB/cm. BCB resist is a polymer with high planarization level, low dielectric constant, low moisture uptake, and high chemical resistance and thermal stability [2]. Although BCB was originally developed as a thin dielectric coating for use in electronic multi-chip modules, its low optical loss (0.8 dB/cm at 1.3 µm), its transparency up to 1.7 µm, and refractive index close to silica (1.5489 at 838 nm) makes it suitable for combining it with other dielectric materials to produce optical components for integrated circuits [3]. BCB cannot be directly nanoimprinted; therefore, a multi-step process is needeed. This process consists of four steps: deposition of BCB layer, nanoimprint lithography, lift-off of etching mask and reactive ion etching (RIE) of BCB. Nanoimprint lithography (NIL) is a low-cost technique introduced by Chou et al. in 1995 [4] to reproduce nanometer-sized patterns. One possible approach in NIL is to imprint a curable polymer. In this approach an epoxy-siloxane resist has been used successfully with a resolution of 60 nm [5].
Adhesion improvement of epoxy resist to a benzocyclobutene layer
233
Figure 3. Bragg grating imprinted with epoxysiloxane resist on a BCB surface without surface treatment. Grating period is 350 nm and aspect ratio is 1.
Adhesion of nanoimprint resist is critical and depends on many parameters: adhesion properties of both the master and the substrate, aspect ratio and geometry of the pattern of the master, verticality of trenches in the master and residual layer thickness (cf., Fig. 2). In this study, the master was coated with octadecytrichlorosilane (OTS) to obtain anti-sticking properties. Surface energy of the OTS-coated master was determined by the three-liquid contact angle method [6] and was found to be less than 25 mJ/m2, which imparts anti-sticking properties to the master. Anisotropic RIE of the master generates vertical trenches. Under these conditions, adhesion is sufficient to nanoimprint microscale patterns but is not sufficient for nanoscale patterns, as can be seen in Fig. 3. Figure 3 shows nanoscale patterns with an aspect ratio of one nanoimprinted on the BCB surface without any treatment. These patterns had a poor adhesion of epoxy-siloxane resist on the BCB surface. Different nanoimprint tests show that the causes for this lack of adhesion are the very low residual layer thickness and the low surface energy of BCB. When nanoimprinting a polymer, the master does not touch the substrate and a residual layer of polymer remains between the master and the substrate (Fig. 2). This residual layer helps to maintain cohesion in nanoimprinted patterns, as well as to improve adhesion of the resist to the substrate. Residual layer thickness depends on the viscosity of the resist. One reason for using epoxy-siloxane for nanoimprint lithography is its low viscosity which enables nanoimprinting at low pressure and with a residual layer thickness less than 10 nm [5]. It is important to assure a low residual layer thickness to transfer easily the patterns to underlying layers. However, this low residual layer thickness causes adhesion problems, as the surface of the resist in contact with the master is more important than the surface in contact with the substrate. To overcome this problem, adhesion to the BCB surface must
234
B. Viallet and E. Daran
be increased. The originality of this paper is in addressing the problem of adhesion during nanoimprint process using a very low viscosity material. Adhesion promoter treatments based on aminosilane self-assembled monolayers are efficient at improving the adhesion of epoxy-based materials [7, 8] to silica. Unfortunately, these treatments are not applicable to polymer surfaces because silane molecules are difficult to graft covalently onto polymers. Therefore, some other strategy must be adopted, e.g., modification of polymerized BCB surface by plasma or ultraviolet (UV)/ozone photochemical process. These processes have already been proposed to improve adhesion of metals, silicon oxide and silicon nitride [9–12] onto the polymerized BCB surface. The purpose of a plasma or UV/ozone process is to oxidize the surface of the polymer with the view to increase its surface energy. These treatments also introduce reactive groups. UV/ozone [13] treatment is also a highly effective method for removing contaminants from surfaces [14] and is, therefore, commonly used for substrate cleaning prior to thin film deposition by molecular beam epitaxy. It creates a clean and controlled (thickness and composition) oxidized surface layer on an inorganic material. However, the effect of UV/ozone is quite different on polymers and depends on the nature and the composition of the polymer. 2. EXPERIMENTAL
2.1. Sample preparation BCB pre-polymer mixture was purchased as CycloteneTM from Dow Chemical and was used diluted at 66% in mesitylene (1,3,5-trimethylbenzene). The solution was spin-coated onto 1-cm2 silicon substrates using the following procedure: substrate cleaning with acetone and water baths, substrate drying at 150°C, application of adhesion promoter (vinyltriacetoxysilane) by spin coating according to the
Figure 4. (a) Chemical structures of polymerized BCB; (b) and (c) monomers for epoxysiloxane resist: (b) 1,3-bis(aminomethyl)cyclohexane (BAC), (c) epoxypropoxypropyl terminated bis(dimethylsiloxane) (DMS-DGE).
Adhesion improvement of epoxy resist to a benzocyclobutene layer
235
Cyclotene procedure [15] and BCB pre-polymer spin coating. A 3500 rpm speed was selected to produce 1.7-µm-thick BCB films. Samples were cured according to Dow Chemical’s “soft cure” process [15] in an oxygen-free furnace. “Soft cure” was chosen instead of “hard cure” because it improves adhesion between polymer layers [15]. The chemical structure of polymerized BCB is shown in Fig. 4. 2.2. Surface treatment Two different surface treatments were tested. The first consisted of exposing the polymerized BCB surface to ultraviolet (UV) radiation and ozone with a UVOCS T0606B cleaning system purchased from UVOCS. The samples were placed 5 mm from the UV lamp generating 185 nm and 254 nm radiations with a power of 10 mW/cm2. The 184 nm wavelength light is absorbed by the atmospheric oxygen and generates ozone. The 254 nm wavelength light is not absorbed by oxygen but decomposes ozone into atomic oxygen [13]. Ozone and atomic oxygen are known as strongly oxidizing agents. These agents, combined with UV radiation, act to modify the surface of BCB. The second treatment consisted of a CF4/O2 plasma followed by deposition of a 3-nm-thick SiOx layer. A 300 barrel plasma chamber from PVA Tepla was used with the following parameters: power 200 W, O2 flow 800 ml/min, CF4 flow 200 ml/min for 30 s exposure. SiOx was deposited by ion beam sputtering with a precision ion etching and Coating System (PECS) from Gatan with an ion energy of 5 keV and SiO2 target. As the ion beam sputtering process can modify the stoichiometry of oxide materials, the exact composition of deposited layer is not known. A layer thickness of 3 nm was used to minimize any effect on the optical waveguide. 2.3. Nanoimprint lithography Epoxy-siloxane polymer was nanoimprinted according to the process detailed in Ref. [5] using 1,3-bis(aminomethyl)cyclohexane (BAC) and epoxypropoxypropyl-terminated bis(dimethylsiloxane) (DMS-DGE, DGE standing for diglycidyl ether) as monomers. BAC was supplied by Sigma-Aldrich and DMS-DGE by ABCR. The structures of BAC and DMS-DGE are shown in Fig. 4. Nanoimprint masters were fabricated by high-resolution electron-beam lithography (EBL) using poly(methyl methacrylate) (PMMA) resist, lift-off and subsequent RIE. Masters were coated with an octadecytrichlorosilane (OTS) [16, 17] molecular layers, ensuring efficient antiadhesion properties of the surface. The patterns etched in the master are gratings with a period of 350 nm, a depth of 150 nm and a length of 700 µm. Monomers were mixed and deposited onto the master. BCB coated substrate and the master were then pressed together for 4 h at 100°C under a pressure of 1.5 MPa to cure the epoxy-siloxane resist and then the master was removed. When removing the master, substrate and master were pulled apart with a force perpendicular to the surface of the substrate. Furthermore, the aim of this study
B. Viallet and E. Daran
236
was to improve the adhesion of epoxy-siloxane resist to BCB layer for NIL. Therefore, NIL was used as a test to check if the adhesion to BCB layer was sufficient for the application or not. 3. RESULTS AND DISCUSSION
3.1. UV/ozone treatment The first technique to improve the adhesion to polymerized BCB was the UV/ozone treatment. The effects of UV/ozone on the BCB layer have been presented previously [18]. The previously published results are briefly discussed
Figure 5. HF and acetone etched layer thicknesses of BCB vs. UV/ozone exposure time. HF etched thickness corresponds to oxidized layer and acetone-etched thickness corresponds to degraded layer.
Table 1. Contact angles (in degrees) of water, diiodomethane and glycerol on polymerized BCB Water Diiodo- Glycerol methane
Base Acid–base Lifshitz– Acid van der parameter parameter component (mJ/m2) (mJ/m2) Waals (mJ/m2) parameter (mJ/m2)
Total surface energy (mJ/m2)
Untreated 81 39 74 40 0.02 5 0.6 41 BCB 10 min 8 44 13 37 2.3 51.4 22 59 treated BCB Also solid surface energy and its components before and after UV/ozone treatment for 10 min are given.
Adhesion improvement of epoxy resist to a benzocyclobutene layer
237
here. It was shown that the effect of UV/ozone treatment on the BCB was to create an oxidized surface, as well as to degrade the polymer surface. The top surface of the oxidized layer was analyzed by X-ray photoelectron spectroscopy (XPS) and was found to have a composition of SiO1.8C0.5Hx. Depth profile was obtained by secondary ion mass spectrometry (SIMS), which showed that carbon concentration increased with depth while oxygen concentration decreased, creating a layered structure. These layers had different chemical properties and their thicknesses were obtained by selectively etching each layer and by measuring the etched layer thickness. Figure 5 presents etched layer thicknesses by hydrofluoric acid (HF) and acetone versus UV/ozone treatment time. The oxidized layer had a composition near SiO2 and could be etched by HF while the polymerized BCB is highly resistant to HF. By measuring the HF etched layer thickness, it is possible to measure the depth of the oxidized layer which was 250 nm for a 90 min UV/ozone treatment. This result is in agreement with SIMS results [18]. The polymerised BCB is a highly cross-linked polymer [15] and is resistant to organic solvents. However, when polymerized BCB was exposed to UV/ozone treatment, its top layer became soluble in acetone. This acetone sensitivity is interpreted as a degradation of the crosslinked polymer structure by UV radiation and ozone. Figure 5 shows that the degraded layer was thicker than the oxidized layer and had a depth of 350 nm for a 90 min UV/ozone treatment. UV/ozone also has an influence on polymerized BCB surface properties. Surface energy was determined by measuring contact angles of three liquids (diiodomethane, glycerol and water) and using the van Oss acid–base theory [6]. The results are presented in Table 1. For more details see Ref. [18]. The Lifshitz–van der Waals surface energy component of BCB slightly decreased with UV/ozone treatment while both acid and base parameters increased. UV/ozone treatment increases the polar component of the BCB surface energy. Epoxy and amine are polar in nature; therefore, the adhesion of epoxy-siloxane to polymerized BCB should be improved by UV/ozone treatment. Nanoimprint tests were performed and the epoxy-siloxane patterns did not adhere to the polymerized BCB surface. The conclusion from these tests was that the UV/ozone treatment was not suitable for promoting adhesion for nanoimprinting. One reason could be that the surface layer of BCB was degraded. 3.2. Plasma treatment Oxygen plasma treatments expose the polymer surface to ion bombardment, electron bombardment, strong oxidizing agents (such as atomic oxygen, oxygen ions, oxygen radicals...) and UV radiation. The effects of electron and ion bombardments on polymerized BCB have been discussed by studying the effects of argon plasma [19]. The argon plasma treatment of the BCB has been reported to create a 10–200-nm-thick degraded layer whereas degradation of the polymer was caused by ion bombardment and by UV radiation. The effects of the oxidizing agents and
238
B. Viallet and E. Daran
UV radiations in oxygen plasma are the same as in UV/ozone; they degrade the surface layer and modify the composition of the surface. After exposing to oxygen plasma, the surface of polymerized BCB is degraded, oxidized and has a composition near SiO2 [20, 21]. With purely organic polymers, the degraded layer is continuously etched during exposure to oxygen plasma. Since the BCB contains silicon, the degraded layer cannot be efficiently etched by oxygen plasma [20, 21]. To etch the silicon part of BCB, a 20% CF4, 80% O2 gas mixture was used. The purpose of CF4 gas was to etch the silicon part of BCB and to eliminate the degraded layer. In order to further reduce the plasma degraded layer thickness, soft plasma conditions were chosen: a power of 200 W, a treatment time of 30 s, and a total gas flow of 1000 ml/min. A Faraday cage was used to minimize ion bombardment. It has been shown [12] by XPS that after the O2/CF4 plasma treatment, the composition of BCB surface was Si 55%, O 28% and C 15%. The plasma treated surface had a significant polar component of surface energy and we measured a contact angle of water of nearly 0° on this surface. However, after this treatment the patterns did not adhere to the surface during nanoimprinting showing results similar to those presented in Fig. 3. The adhesion of epoxy-siloxane was found not to be sufficient for the nanoimprint process. In order to further improve the adhesion, an interface layer was deposited after the O2/CF4 plasma treatment. We had observed good adhesion of the epoxysiloxane resist to oxidized silicon during nanoimprinting process [5]. Therefore, a thin layer of SiOx was tested. The interface layer was a 3-nm-thick SiOx layer deposited by sputtering. Due to the high surface energy of SiOx and also due to its deposition by sputtering, this material possesses good adhesion to the modified BCB surface. Figure 6 presents the results of a nanoimprint experiment on 350-nm
Figure 6. 350-nm period gratings imprinted on polymerized BCB after CF4/O2 plasma treatment and SiOx deposition observed with optical (left) and electron beam (right) microscopies.
Adhesion improvement of epoxy resist to a benzocyclobutene layer
239
period gratings after O2/CF4 plasma treatment and SiOx deposition. All nanoimprint tests realized using O2/CF4 plasma followed by SiOx deposition showed a good adhesion of the patterns. Adhesion properties were found to be sufficient for the nanoimprint process. 4. CONCLUSIONS
Nanoimprinting with epoxy curable polymer on the polymerized benzocyclobutene (BCB) layer suffers from poor adhesion. Different treatments were tested to improve the BCB adhesion. The effect of UV/ozone was studied in detail: the UV/ozone treatment created an oxidized layer, as well as degraded the BCB structure. While the treatment raised the surface energy and improved wetting, surface degradation precluded improvement of adhesion properties. An efficient process, which increased the BCB surface energy, was developed which consisted of a CF4/O2 plasma treatment followed by an SiOx layer deposition. Plasma conditions were optimised to minimize the degradation of BCB. The purpose of using CF4 and O2 gases during the plasma treatment was to etch the degraded layer. REFERENCES 1. T. Bhutta, A. M. Chardon, D. P. Shepherd, E. Daran, C. Serrano and A. Muñoz-Yagüe, IEEE J. Quant. Electron. 37, 1469 (2001). 2. D. Burdeaux, P. Townsend, J. Carr and P. Garrou, J. Electron. Mater. 19, 1357 (1990). 3. C. F. Kane and R. R. Krchnavek, IEEE Photonics Technol. Lett. 7, 535 (1995). 4. S. Y. Chou, P. R. Krauss and P. J. Renstrom, Appl. Phys. Lett. 67, 3114 (1995). 5. B. Viallet, P. Gallo and E. Daran, J. Vac. Sci. Technol. B 23, 72 (2005). 6. C. J. van Oss, Interfacial Forces in Aqueous Media, Marcel Dekker, New York, NY (1994). 7. D. L. Woerdeman, N. Amouroux, V. Ponsinet, G. Jandeau, H. Hervet and L. Leger, Composites Part A 30, 95 (1999). 8. J. G. Iglesias, J. Gonzalez-Benito, A. J. Bravo and J. Baselga, J. Colloid Interface Sci. 250, 251 (2002). 9. S. Luo and C. P. Wong, IEEE Trans. Component Packag. Technol. 24, 43 (2001). 10. R. Taveli, H. Gundlach, Z. Bian, A. Knorr, M. Van Gestel, S. Padiyar, A. E. Kaloyeros and R. E. Geer, J. Vac. Sci. Technol. B 18, 252 (2000). 11. P. Garrou, D. Scheck, J.-H. Im, J. Hetzner, G. Meyers, D. Hawn, J. Wu, M. B. Vincent and C. P. Wong, IEEE Trans. Component Packag. Technol. 23, 568 (2000). 12. K. W. Paik, H. S. Cole, R. J. Saia and J. J. Chera, J. Adhesion Sci. Technol. 7, 403 (1993). 13. J. R. Vig, J. Vac. Sci. Technol. A 3, 1027–1034 (1985). 14. S. R. Kasi and M. Liehr, Appl. Phys. Lett. 57, 2095–2097 (1990). 15. Utilization Procedure of Cyclotene Resist: “CycloteneTM 4000 Series Advanced Electronic Resins (Photo BCB)”, Dow Chemical Technical Resource. Dow, Midland, MI). 16. X. Zhao and R. Kopelman, J. Phys. Chem. 100, 11014 (1996). 17. Y. Liu, K. Wolf and M. Messmer, Langmuir 17, 4329 (2001). 18. B. Viallet, E. Daran and L. Malaquin, J. Vac. Sci. Technol. A 21, 766 (2003).
240
B. Viallet and E. Daran
19. L. Trabzon and O. O. Awadelkarim, Microelectron. Eng. 65, 463 (2003). 20. M. R. Baklanov, S. Vanhaelemeersch, H. Bender and K. Maex, J. Vac. Sci. Technol. B 17, 372 (1999). 21. S. A. Vitale, H. Chae and H. H. Sawin, J. Vac. Sci. Technol. A 18, 2770 (2000).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 241–250 Ed. K.L. Mittal © VSP 2007
Improvement of adhesion between poly(tetrafluoroethylene) and poly(ethylene terephthalate) films GOKNUR BAYRAM,1,* GURALP OZKOC2 and PINAR KURKCU1 1
Department of Chemical Engineering, Middle East Technical University, 06531 Ankara, Turkey Department of Polymer Science and Technology, Middle East Technical University, 06531 Ankara, Turkey
2
Abstract—Multilayer films from poly(tetrafluoroethylene) (PTFE) and poly(ethylene terephthalate) (PET) were produced using an adhesive layer consisting of ethylene-methyl acrylate-glycidyl methacrylate (E-MA-GMA) terpolymer and low density polyethylene (LDPE) blend. The surfaces of PTFE were chemically modified by Na/naphthalene treatment and subsequent acrylic acid grafting. FT-IR and XPS measurements were performed to characterize the modified PTFE surfaces. These analyses showed defluorination and oxidation of the modified PTFE surface, and confirmed the acrylic acid grafting. To measure the adhesion strength between PET and modified PTFE, multilayers were subjected to T-peel tests. Peel strength increased with respect to increasing E-MA-GMA amount in the adhesive layer. Prolonging the time of Na/naphthalene treatment improved the peel strength for multilayers of both acrylic acid grafted and ungrafted PTFE. Scanning electron microscopy (SEM) analysis showed that the texture of the PTFE surface after modifications became rougher when compared to untreated PTFE. Keywords: PTFE; PET; adhesion; surface modification; peel strength.
1. INTRODUCTION
Fluoropolymer films, especially poly(tetrafluoroethylene) (PTFE) as a heat- and flame-resistant material, are important in many applications in industry. They are very resistant to moisture, providing a good moisture barrier. However, their relatively high cost limits their usage. When combined with other film materials as multilayers, they provide a weatherable surface that is non-adhering, chemically inert and strong. Biaxially-oriented poly(ethylene terephthalate) (PET) film is preferred as one of the film layers because it has high tensile properties, excellent dimensional stability, good barrier properties, high usage temperature, excellent optical properties, recyclability and relatively low cost. The current study focuses *
To whom correspondence should be addressed. Tel.: (90-312) 210-2632; Fax: (90-312) 210-2600; e-mail:
[email protected]
242
G. Bayram et al.
on multilayer films produced from PTFE and PET films, which are a good candidate for packaging, insulation and electronic industries. Because of its low surface energy, bonding PTFE to other materials requires special surface modification techniques without altering the bulk properties. Plasma modification [1–4], γ and excimer laser irradiations [5–9], electron irradiation [10] and alkali metal treatment (sometimes called wet-chemical treatment) [11–17] are commonly used. This study aimed especially to construct a model system to improve adhesion between PTFE and PET films using an epoxymodified polyethylene adhesive and its blends with low-density polyethylene (LDPE). The surface of PTFE film was modified using a wet chemical method involving Na/naphthalene treatment (or etching) and acrylic acid (AA) grafting to incorporate reactivity, for combining the unique properties of PTFE and the low cost of a PET film. As a result, properties similar to PTFE film at the given thickness can be achieved in an economical way. Na/naphthalene treatment of PTFE to etch the polymer surface is a commercially successful surface treatment method due to its high rate of reaction and ease of handling. In this method, equimolar metallic sodium (Na) and naphthalene are dissolved in tetrahydrofuran (THF) for wet chemical treatment of PTFE. It is known that some heat is released in the treatment process and this heat may cause THF to decompose, and as a result the Na/naphthalene complex could be destroyed [12]. To avoid this destruction of the complex, cryo-conditions (-40ºC) should be maintained. With this method, it is also possible to achieve defluorination, oxidation, or decrease in water contact angle, which means an increase in surface energy [14, 15]. Moreover, it is possible to graft monomers utilizing these oxidized and defluorinated active centers. In this study, the wet chemical surface treatment of PTFE films was first carried out by using the Na/naphthalene complex in THF at varying treatment times. Then, AA monomer was grafted from solution. FTIR, XPS and tensile tests were performed to characterize the surface structure of treated PTFE. Adhesion between PET/PTFE films was promoted using a blend of ethylene-methyl acrylateglycidyl methacrylate terpolymer and LDPE as an adhesive layer. The adhesion strength and morphology were examined through peel testing and scanning electron microscopy (SEM) analysis, respectively. 2. EXPERIMENTAL
2.1. Materials PTFE films with a thickness of 0.5 mm were purchased from Polikim Sanayi Ticaret (Turkey) and recycled PET (RPET) films were supplied by Larkan Plastik (Turkey). The melting point of the RPET films was determined by DSC as 249°C. The terpolymer of ethylene-methylacrylate-glycidyl methacrylate (E-MA-GMA), Lotader AX 8900 from Arkema Chemicals (France), was used as the adhesive
Improvement of adhesion between PTFE and PET films
243
layer having an epoxy functional group. This material contains 25 wt% methyl acrylate and 8 wt% glycidyl methacrylate. Low-density polyethylene (LDPE), Petilen G03-5, was purchased from Petkim Petrochemical (Turkey). Tetrahydrofuran (THF) was HPLC grade with a purity of 99.8% and supplied by Lab-Scan Analytical Sciences (Ireland). Synthesis grade benzoyl peroxide (BPO), sodium metal rods and synthesis grade naphthalene were obtained from Merck (Germany). Acrylic acid (AA) with a purity of >99.5% was kindly provided by Organic Kimya (Turkey). 2.2. Defluorination of PTFE films by Na/naphthalene treatment in THF One mol of sliced-Na and 1 mol of naphthalene were dissolved in 1 l THF. After stirring for 6–8 h, the solution was cooled to -40°C in a deep-freezer, then films of PTFE (200 mm¥30 mm¥0.5 mm) were immersed in this solution at this temperature and kept for 1, 5 and 10 min. After each period of wet-chemical treatment in Na/naphthalene in THF, the PTFE films were washed with an excess amount of acetone and then with distilled water, and allowed to dry in an oven at 80°C. For FT-IR studies, the same experiments were performed with about 5 g PTFE in a powder form. 2.3. Acrylic acid grafting onto the defluorinated PTFE films The solution polymerization medium for grafting contained 2 g BPO as an initiator and 200 ml acrylic acid in 2000 ml THF. The films were immersed into a twoneck glass reactor. Temperature of the medium was kept between 65–70°C during the reaction time. The films were removed from the reaction vessel at the end of 150 min. They were washed with THF and water, and then allowed to dry in an oven at 80°C. A similar reaction was also carried out for FT-IR studies in a small reaction vessel using PTFE powder. 2.4. Preparation of multilayers In this study, multilayers were produced from defluorinated and/or AA-grafted PTFE films and RPET films using an adhesive layer consisting of varying amounts of E-MA-GMA terpolymer (25, 50, 75 and 100 wt%) and LDPE. The polymer blends for adhesive layers containing 25, 50 and 75 wt% E-MA-GMA were prepared in a co-rotating twin-screw extruder (Thermo Prism TSE 16 TC, L/D:24, UK). The screw speed was 200 rpm and barrel temperatures from feedzone to die were 170–200–200–200–210°C. The molten blend was quenched with water and then chopped into small pellets. Prior to film production from blends, the pellets were dried at 80°C and the pure E-MA-GMA terpolymer was dried at 50°C under vacuum for 4 h. Adhesive layer films were produced using a hot-plate. The process temperature was 230°C, preheating and press times were 1 min each and compression pressure was 15 N/mm2. The adhesive film thickness was controlled using a metal frame to
244
G. Bayram et al.
Figure 1. Schematic representation of multilayer preparation.
0.5 mm. The adhesive films were allowed to cool to room temperature. Then, strips with the dimensions of 200 mm¥30 mm were cut. Multilayers from defluorinated and/or AA-grafted PTFE films, RPET films and adhesive layer from blend of LDPE and E-MA-GMA were prepared on a hotpress at 230°C. The preheating time was 1 min and press time was 3 min. The schematic representation is given in Fig. 1. After hot-pressing, the samples were left to cool to room temperature. For peel tests, the strips were cut to obtain the standard dimensions. 2.5. Characterization Infrared spectra were obtained from mixtures of PTFE powder sample (pure, defluorinated, and AA grafted) and dry KBr using a Nicolet Impact 400 spectrometer (Nicolet Instruments). The chemical compositions of PTFE film surfaces were determined by performing X-ray photoelectron spectroscopy (XPS) (SPECS ESCATM) analysis using a non-monochromatized Al-Kα X-ray source (1486.5 eV), at a pressure lower than 10-5 Pa. The power of the X-ray source was set to 240 W. The binding energies were previously calibrated against a value of the C1s hydrocarbon component centered at 284.6 eV. By measuring the F1s, C1s and O1s XPS spectra, the fluorine to carbon ratio (F/C), as well as oxygen to carbon ratio (O/C) were determined. T-peel tests were performed according to ASTM D 1876 on multilayer film samples (20 cm long and 2 cm wide) using an Instron Mechanical Testing Machine at a crosshead speed of 10 cm/min. Load (N) versus displacement (mm) curves were obtained for all samples. The average force after the first peak load in the force versus displacement curve was divided by the sample width to calculate the peel strength. The results for each sample were given as the average of five measurements. Tensile tests were performed on defluorinated and pure PTFE films according to ASTM D 658 with an Instron Mechanical Testing Machine. Dogbone samples cut from the films using a punch-knife were tested at a crosshead speed of 5 mm/min. The test results were reported as the average of five samples for each experiment.
Improvement of adhesion between PTFE and PET films
245
The morphologies of pure PTFE, defluorinated PTFE, acrylic acid grafted and peeled surfaces of multilayers were investigated using a Jeol JSM-6400 scanning electron microscope. The samples for SEM analysis were precoated with gold prior to imaging to provide conductive surfaces. 3. RESULTS AND DISCUSSION
3.1. FT-IR and XPS FT-IR spectra of defluorinated (i.e., Na/naphthalene treated) and AA-grafted PTFE powder samples are illustrated in Fig. 2. The FT-IR spectra of untreated, defluorinated only at various times and AA-grafted PTFE powders all showed
Figure 2. FT-IR spectra of: (1) PTFE and after (2) 1 min Na/naphthalene treatment, (3) 5 min Na/naphthalene treatment, (4) 10 min Na/naphthalene treatment, (5) 5 min Na/naphthalene treatment + AA grafting.
Table 1. Elemental surface composition of treated and untreated PTFE films Atom ratio
PTFE
5 min Na/naphthalene treated PTFE
5 min Na/naphthalene treated + AA-grafted PTFE
F/C O/C
1.722 0.055
1.053 0.578
0.842 0.789
246
G. Bayram et al.
the respective rocking and wagging absorptions of CF2 at 503 cm-1 and 620– 640 cm-1, strong stretch vibration absorption of CF2 at 1150 cm-1 and strong stretch vibration absorption of CF3 at 1240 cm-1; however, these regions are not shown in Fig. 2 for clarity. Defluorination, together with oxidation, was observed in the B spectra. The absorption band appearing at 1722 cm-1 in spectra 2, 3 and 4 (B) represents the –CF=CF– stretch, which is a potential chemical site for AA grafting. Intensity of this peak increased when the Na/naphthalene treatment time was increased, because the extent of defluorination increased. In spectra 2, 3 and 4 (B), the absorption band appearing at 1883 cm-1 shows the formation of COF bonds as a result of oxidation of the film surface during defluorination. The absorption band at 3350 cm-1 seen in spectra 2, 3, 4 and 5 (C) is due to adsorbed water on the surface by the doped NaF as the side product of defluorination [12, 17]. Despite the existence of NaF salt on the surface, no peak relevant to naphthalene was observed in the spectra. Absorption bands of functional groups of AA due to out-of-plane deformation of COOH, vibration absorption of C=O and –OH stretching are seen at 800, 1760 and 3200 cm-1, respectively. It was observed from the FT-IR spectrum 5 (B) that the absorption band at 1883 cm-1 relevant to COF disappeared when AA grafting was performed. This can be due to two possible reactions. One of them is the hydrolysis reaction of –COF group in the presence of AA and the other is the possible AA grafting occurring via these oxidized sites of the PTFE films. The elemental surface compositions of untreated, Na/naphthalene treated in THF for 5 min and AA-grafted PTFE surfaces were analyzed by XPS analysis. The data for F/C and O/C ratios are summarized in Table 1. The F/C atomic ratio after Na/naphthalene treatment decreased from 1.722 to 1.053, which supported a defluorination of the surface. On the other hand, the O/C ratio increased from 0.055 to 0.578, indicating oxidation during the Na/naphthalene treatment. After acrylic acid grafting, oxygen content of the surface increased (O/C ratio increased from 0.578 to 0.789) indicating the presence of acrylic acid groups on the defluorinated surfaces. 3.2. T-Peel tests Figure 3 represents the variation of peel strength with respect to E-MA-GMA content in the adhesive layer for RPET/adhesive-layer/PTFE multilayer films. The same figure also shows the peel strength values for the PTFE films which were defluorinated at different times (1, 5 and 10 min) with/without AA grafting. It is seen from the graph that as the E-MA-GMA content increased in the adhesive layer, the adhesion strength values increased. The epoxy groups of the adhesive layer have reaction capability with the acid end-groups of RPET [18], as well as with acid groups of AA grafted onto the PTFE surface. Thus, improved peel strength values were obtained as the epoxy content increased. For the grafted-AA
Improvement of adhesion between PTFE and PET films
247
Figure 3. Variation of peel strength with respect to E-MA-GMA content in the adhesive layer and defluorination (Na/naphthalene treatment) time.
on the defluorinated PTFE surface, the improvement in peel strength may be attributed to the increase in polarity of the surface. As defluorination time increased, peel strength values of multilayers containing 50, 75 and 100% E-MA-GMA in the adhesive layer increased. The peel strength of 25% E-MA-GMA samples were not affected by the treatment time. AA grafting especially gave rise to peel strength for 5 and 10 min Na/naphthalene treated films. When all the peel test results were considered, the minimum peel strength was obtained with the sample produced by 1 min Na/naphthalene-treated PTFE and 25% E-MA-GMA containing adhesive layer. The maximum value was obtained when a 10 min Na/naphthalene treatment, AAgrafted PTFE and 100% E-MA-GMA containing adhesive layer were used to produce multilayers. The improvement from minimum to maximum was about one order of magnitude. 3.3. Tensile tests One of the critical points in surface modification is to alter the chemical structure of the surface only in order to obtain the desired property without altering the bulk properties of the material. In the current study, the effect of the wet chemical treatment process on the PTFE film properties was analyzed by tensile tests. Tensile strength and elongation-at-break can be used to evaluate structural changes of the bulk material. Figures 4 and 5 show the tensile strength and elongation-atbreak values of the PTFE films after surface treatments, respectively. Both properties decreased steadily as the treatment time increased. Elongation-at-break values decreased to 68%, and then almost remained constant. As a result, it can be
248
G. Bayram et al.
Figure 4. Tensile strength of PTFE film with respect to surface treatment.
Figure 5. Elongation-at-break of PTFE film with respect to surface treatment.
said that both defluorination and subsequent AA grafting affected the structure of the bulk PTFE. 3.4. Scanning electron microscopy (SEM) Figure 6 shows the morphology of untreated PTFE film, defluorinated PTFE films for 1, 5 and 10 min Na/naphthalene treatment times, and the 5 min Na/naphthalene-treated+AA-grafted PTFE films. It is seen from the first micrograph that the untreated PTFE film had a very smooth surface. After 1 min of Na/naphthalene treatment, some cracks were
Improvement of adhesion between PTFE and PET films
249
Figure 6. SEM micrographs of (1) untreated PTFE, (2) 1 min Na/naphthalene-treated PTFE, (3) 5 min Na/naphthalene-treated PTFE, (4) 10 min Na/naphthalene-treated PTFE and (5) 5 min Na/naphthalene-treated PTFE + AA grafting.
observed on the surface. Lin et al. [17] called a surface displaying such texture as “groove-like surface texture”. However, after 5 and 10 min of Na/naphthalene treatment, the texture turned to a rough, sponge-like morphology. When adhesion and wettability are concerned, it is important to achieve a large contact area between the adhering phases; therefore, such a rough surface may be an advantage from this point of view. On the other hand, such cracks, waves and curls on the surface may decrease the mechanical strength of the material. Hence, it was reasonable to expect a loss of tensile strength and elongation after Na/naphthalene treatment. It can be seen in the fifth micrograph that AA grafting did not appear to cause any morphological changes.
250
G. Bayram et al.
4. CONCLUSIONS
It was observed from FT-IR and XPS studies that Na/naphthalene treatment resulted in defluorination and oxidation of the PTFE surface. AA groups likely bonded to the defluorinated surface through oxidized sites. Tensile tests showed that wet-chemical treatment of PTFE films affected the mechanical performance adversely. Due to the interaction between the epoxy groups of the adhesive layer and the acid groups of both PET and AA-grafted PTFE, the peel strength increased with both increasing E-MA-GMA concentration and defluorination time for the AA grafted samples. REFERENCES 1. C.Y. Huang and C. Tseng, J. Appl. Polym. Sci. 78, 800-807 (2000). 2. A.K.S. Ang, E.T. Kang, K.G. Neoh, B. Liaw and D.J. Liaw, J. Adhesion Sci. Technol. 14, 897914 (2000). 3. V.N. Vasilets, G. Hermel, U. König, C. Werner, M. Müller, F. Simon, K. Grundke, Y. Ikada and H.J. Jacobasch, Biomaterials 18, 1139-1145 (1997). 4. U. König, M. Nitschke, A. Menning, G. Eberth, M. Pilz, C. Arnhold, F. Simon, G. Adam and C. Werner, Colloids Surfaces B: Biointerfaces 24, 63-71 (2002). 5. S.K. Koh, S.C. Park, S.R. Kim, W.K. Choi, H.J. Jung and K.D. Pae, J. Appl. Polym. Sci. 64, 1913-1921 (1997). 6. T. Jun and X. Qunji, J. Appl. Polym. Sci. 69, 435-441 (1998). 7. B. Hopp, N. Kresz, J. Kokavecz, T. Smausz, H. Schieferdecker, A. Döring, O. Marti and Z. Bor, Appl. Surface Sci. 221, 437-443 (2004). 8. A.V. Fedenev, S.B. Alekseev, I.M. Goncharenko, N.N. Koval, E.I. Lipatov, V.M. Orlovskii, M.A. Shulepov and V.F. Tarasenko, Laser Particle Beams 23, 265-272 (2003). 9. E. Adem, M. Avalos-Borja, E. Bucio, G. Burillo, F.F. Castillon and L. Cota, Nucl. Instrum. Methods Phys. Res. B. 234, 471-476 (2005). 10. L. Haussler, G. Pompe, D. Lehmann and U. Lappan, Macromol. Symp. 164, 411-419 (2001). 11. A.A. Benderly, J. Appl. Polym. Sci. 6, 221-225 (1962). 12. C.Y. Huang and W.Y. Chiang, Angew. Makromol. Chem. 209, 9-23 (1993). 13. W.Y. Chiang and C.Y. Huang, J. Appl. Polym. Sci. 47, 577-585 (1993). 14. C. Combellas, S. Richardson, M.E.R. Shanahan and A. Thiebault, Int. J. Adhesion Adhesives 21, 59-64 (2001). 15. I. Noh, K. Chittur, S.L. Goodman and J.A. Hubbell, J. Polym. Sci., Part A: Polym. Chem. 35, 1499-1514 (1997). 16. M. Keusgen, J. Glodek, P. Milka and I. Krest, Biotechnol. Bioeng. 72, 530-540 (2001). 17. C.W. Lin, W.C. Hsu and B.J. Hwang, J. Adhesion Sci. Technol. 14, 1-14 (2000). 18. H. Durgun and G. Bayram, J. Adhesion Sci. Technol. 19, 407-425 (2005).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 251–262 Ed. K.L. Mittal © VSP 2007
Novel approaches to enhance adhesion of cellulose E. DELGADO,1,* J. A. VELÁSQUEZ,2 G. G. ALLAN,3 A. ANDRADE,1 H. CONTRERAS,1 H. REGLA,1 L. R. BRAVO1 and G. TORIZ1 1
University of Guadalajara, Department of Wood, Cellulose and Paper, P.O. Box 52-93, 45020 Zapopan, Jalisco, Mexico 2 Facultad de Ingeniería Química, Universidad Pontificia Bolivariana, Apartado Aéreo 56006 Medellín, Antioquía, Colombia 3 University of Washington, College of Forest Resources, Seattle, WA 98195-2100, USA
Abstract—This paper examines the fixation of functionalities (namely, zwitterions) to cellulose surfaces as a means to enhance adhesion. The approach is discussed in terms of the attachment of ionic moieties to cellulose and its effects on mechanical properties of paper handsheets, and the rupture force of two overlapped cellophane films. It was found that the mechanical properties of the paper were improved, especially the wet tensile strength of paper (about 16% of the dry strength). In zwitterionic cellophane films, the dry rupture force was more than doubled compared to unmodified films. The wet rupture force was dramatically increased from an undetectable value corresponding to an unmodified film to 24–35% of the dry strength in zwitterionic films. A new system of bonding is thus proposed based on zwitterions and its contribution to enhanced adhesion of cellulose surfaces is analyzed. Keywords: Cellulose; zwitterions; amino acids; ionic bonding; cellophane; adhesion; tensile strength.
1. INTRODUCTION
1.1. Cellulose chemical nature Cellulose is a natural linear polymer composed of anhydroglucose units, whose degree of polymerization can reach up to 10 000 units in, for instance, cotton. Owing to the fact that the monomer unit is bonded via glycosidic β-1-4 linkages, cellulose chains develop a crystal structure with hydrogen bonds within the repeat unit, as well as the cellulose chain. Furthermore, cellulose is biomanufactured in such a way that promotes assembling of individual cellulose chains in a hierarchical fashion that leads to the formation of microfibrils that are held together by intramolecular hydrogen bonds. Evidently, the strong adhesion in the cell wall of *
To whom correspondence should be addressed. Tel.: (52-33) 3682-0110; Fax: (52-33) 3682-0643; e-mail:
[email protected]
252
E. Delgado et al.
cellulosic fibers by the ubiquitous hydrogen bonds results in physicochemical properties displayed by the cellulose polymer, such as the high Young’s modulus, high glass transition, the theoretical high melting point (never reached), relatively low chemical reactivity, difficulty in dissolution, etc. [1]. 1.2. Cellulose to cellulose adhesion As in the case for the adhesion of one material to another, the adhesion of cellulose to cellulose must be due ultimately to interactions between the two surfaces at the molecular level. This is because the interacting forces between surfaces decrease very rapidly with distance. To understand adhesion one must study the inter- and intra-molecular interactions directly. This obviously poses considerable theoretical and experimental difficulties. Instead, the study of the mechanical behavior of a macroscopic body itself is used to deduce indirectly what takes place at the site of failure [2]. Thus, the study of parameters considered responsible for causing a coherent body to break, (i.e., the energy or the force) is a common approach to gain general ideas about how the interactions between the constituents of a material contribute to its strength. A great deal of experimental observations about the nature of cellulose fiber to fiber bonding and its role in paper strength have been analyzed and ordered in a very logical and coherent way by Page [3]. Page developed a semi-empirical equation (equation (1)) considering two main factors: the fiber strength and the fiber-fiber bond. If one of these factors is zero, the paper has no strength.
1 9 12Aρ g = + T 8Z bPL(RBA)
(1)
where: T = tensile strength of paper (expressed as breaking length), Z = zero span tensile strength (i.e., the distance between the grips that hold the paper, strip is zero), A = average fiber cross-section, ρ = density of the fibrous material, g = acceleration of gravity, b = bond strength per unit area, P = perimeter of the fiber cross-section, L = average length of fiber and RBA = relative bonded area of the sheet. Equation (1) shows that, in order to enhance the tensile strength of the paper, one has to use longer and stronger fibers, particularly those with thinner cell wall. Also a greater area of contact in-between fibers should enhance tensile strength. Practical ways for increasing the RBA as used in papermaking are mechanical (beating) and chemical hydration (addition of polymer additives). It is well known that hydrogen bonding is the main force behind the strength properties of paper sheets. Hydrogen bonds develop when a fiber suspension is poured on a wire to form paper. Once the water is drained from the suspension, the surface tension compacts the web, brings the fibers together and, upon drying, the available surface hydroxyl groups approach one another and form hydrogen bonds [4]. Hydrogen bonds, however, are not as available at the surface of cellu-
Novel approaches to enhance adhesion of cellulose
253
lose fibers as one might think. Scattering coefficient studies of paper have revealed that only 0.5–2.0% of all hydroxyl groups participate in interfiber bonding. Nonetheless, this small percentage of available hydroxyl groups is sufficient to achieve good paper strength [5]. 1.3. Chemical modification of cellulose Another approach to strengthen paper (for both dry and wet strengths) is to alter the nature of the physicochemical interactions between surfaces of cellulosic fibers by surface chemical modification. The chemical modification can be done to improve adhesion by increasing the enthalpy of interactions between the fibers, for instance, via chemical bonds such as covalent, ionic and long range interactions. Covalent bonds are the strongest in the category of chemical bonds (209–595 kJ/mol), but the formation of covalent bonds in papermaking involves the use of organic solvents, expensive catalysts, long curing times and high temperatures. There is available an enormous body of research on the chemical modification of fibers to improve interfiber bonding. These modifications include introduction of carboxyl, aldehyde, amide, cyanoethyl, hydroxyalkyl, sulfonic acid groups, etc., with the aim to develop fiber surfaces capable of enhanced bonding between them [4, 6–11]. In spite of these efforts, it is not clear if the improved properties of paper (e.g., tensile strength) were obtained by the chemical modification per se or were due to the morphological modification of the cell wall (by the insertion of bulky substituents), which allowed better hydration of the fibers and better contact between them, increasing the hydrogen bonds and hence the mechanical properties. Allan et al. [4] utilized an interesting technique that involved the use of dichloro-s-triazine dyes under alkaline conditions compatible with conventional papermaking processes, to surface modify lignocellulosic pulp fibers for achieving better mechanical properties, mainly dry and wet strengths. Of particular relevance to adhesion between cellulosic fibers is the work of Allan and co-workers on the introduction of ionic functionalities on fibers [12, 13]. 1.4. Ionic bonding Ionic bonds are also as influential as hydrogen bonds in securing fiber–polymer interactions in papermaking. Electrostatic forces are, in fact, already utilized in the manufacture of paper in the areas of sizing, dyeing, and for the retention of resins and fillers. The use of ionic bonding is as efficient as the hydrogen bonds, with the added advantage of 2–5-fold enthalpy increase (ionic bond strengths range from 41.84 to 125.52 kJ/mol). A bonding system based on electrostatic linkages would benefit from their increased ability to occur over greater intermolecular distances than those necessary for the creation of covalent or hydrogen bonds [12] and a higher enthalpy which would result in a stronger paper. Furthermore, their instantaneous formation in aqueous media would eliminate the
254
E. Delgado et al.
need for prolonged curing reactions. In this paper, we focus on a particular case of ionic bonding that is referred to elsewhere as “zwitterionic bonding” [14, 15]. 1.5. The zwitterionic approach A zwitterion is an ion that has both a localized positive and a negative charge within a single molecule. In such a molecule, one ion is the natural counterion of the other. Zwitterions offer the possibility of creating a pair of ionic bonds with an associated energy of approx. 125.52 kJ/mol per reaction site. Zwitterions are well exemplified by nature in amino-acid molecules. The mechanical properties of a paper produced by the zwitterionic fibers should directly reflect an increase or decrease in adhesion, since these properties are affected by the type and quality of the bonds between fibers. Potential changes in either the bond strength or the bonded area of handsheets produced from zwitterionic pulps can be detected by applying direct tests such as the tensile, tear and burst strengths, complemented with a measure of the sheet density. Allan et al. [13] synthesized zwitterionic fibers by reaction of a chloro-striazine complex of the amino acid lysine with lignocellulosic fibers. Handsheets were prepared from the zwitterionic fibers and their mechanical properties were determined focusing essentially on the dry strength of paper webs. The results showed that zwitterionic modified fibers had a negative influence on the tensile strength of dry paper webs. On the other hand, surprisingly, paper made from zwitterionic fibers retained more than 35% of their dry strength after being immersed in water for more than 2 weeks. In the present paper, the bonding of never-dried wet webs made from zwitterionic cellulose fibers has been investigated in an attempt to better understand both the loss of dry strength and the mechanism of the enhanced wet tensile strength 2. EXPERIMENTAL
2.1. Materials A commercial, bleached but unbeaten, Douglas fir (Pseudotsuga menziesii (Mirb.) Franco) pulp was donated by Weyerhaeuser (Federal Way, WA, USA). Regenerated cellulose films were kindly provided by Celanese Mexicana S.A. de C.V. and were purified by repeated extraction with boiling distilled water until free of extractives. 1,3,5-trichloro-s-triazine was repeatedly re-crystallized from carbon tetrachloride to constant melting point, and thereafter stored under nitrogen in a desiccator. All other chemicals were reagent grade from Sigma-Aldrich (St. Louis, MO, USA) or standard laboratory stock and were used without further purification.
Novel approaches to enhance adhesion of cellulose
255
2.2. Methods 2.2.1. Preparation of zwitterionic pulps The fiber reactive amino-acid derivatives, N6-(4,6-dichloro-1,3,5-triazin-2-yl)-Llysine and O-(4,6-dichloro-1,3,5-triazin-2-yl)-L-tyrosine were synthesized and reacted with the pulp as reported by Allan et al. [13]. A control pulp was prepared similarly except that no triazinyl-amino-acid derivative was added. A lysinebased zwitterionic pulp in which the amino-acid fragments were allowed to remain complexed with copper ions was also prepared. Calculations based on the amount of nitrogen analyzed by ESCA indicated that there was approximately 0.6 mmol amino acid per gram fiber. ESCA results reported elsewhere [13] showed that the concentration of amino-acid molecules on the surface of the fibers was three times higher than in the bulk. 2.2.2. Handsheet preparation Handsheets of 64 g·m-2 were prepared in a British sheet mold, and then pressed and dried according to the Canadian standard procedure C.4 [16]. 2.2.3. Hot and pressed handsheets Dried standard handsheets were saturated with distilled water, either by spraying or by immersion for 1 min, and placed between two Tappi standard stainless steel press plates (diameter 15.9 cm). The plates were then pressed at 115°C (plate temperature) and 4.5 kg·cm-2 for 5 min in a 30-ton hydraulic table-top press (model 0-2300, Pasadena Hydraulics, La Brea, CA, USA) and then conditioned under standard conditions (50% RH and 23°C). 2.2.4. Handsheet testing Measurements of basis weight (i.e., the weight in g per m2 paper), thickness, apparent density, tear, burst and dry tensile strengths were performed as described in Canadian standards D.3, D.4, D.9, D.8 and D.6H, respectively [17–21]. The wet tensile strength was determined according to Canadian standard D.10 [22] after immersion in distilled water at room temperature for 1 h. 2.2.5. Wet web strength of zwitterionic handsheets Wet web strips (2.5¥12.5 cm) were cut from standard handsheets (basis weight, 60 g·m2) as described elsewhere [23]. The solids content in the wet handsheets was adjusted by applying pressure (0, 12.5, 25, 37.5 and 50 kg·cm-2) to remove water from the webs and then were stored in a plastic bag until measuring their tensile strength. Specimens were mounted vertically in an Instron universal testing machine between two small pieces of nonwoven Teflon strips at each end, to prevent sticking or failure at the clamp, at an elongation rate of 5 cm·min-1. The moisture content of each strip was determined gravimetrically after oven drying. 2.2.6. Modification of cellophane films with the copper complex of dichloro-s-triazinyl-L-lysine Cellophane film (10 g) was modified with a copper II complex of dichloro-striazinyl-L-lysine solution [11] in distilled water (500 ml) at 20ºC. The mixture
256
E. Delgado et al.
was carefully stirred (10 min) and the ionic strength adjusted with NaCl (15 g) over the next 15 min, then NaOH (6 M) was added until pH 12. The mixture was stirred for a specified time, with the addition of 6 M NaOH as needed to maintain the pH at 12. After the desired time had elapsed, the film was washed extensively in running tap water. The blue-tinted cellophane was immersed into an aqueous solution (400 ml) of Na2EDTA⋅2H2O (2 g, 5.4 mmol), stirred for 10 min (the cellophane turned colorless) and washed extensively with running tap water. The cellophane was then immersed into an aqueous solution (400 ml) of Na2EDTA⋅2H2O (1 g) and Triton X-100 (0.1%) and boiled for 5 min. Then the modified cellophane was again washed and conditioned with different pH buffer solutions for 30 min, washed sequentially in running tap water and distilled water and finally air dried. 2.2.7. Analysis of amino-acid residues in cellophane A modification of the classical ninhydrin-based analysis method was used [24]. For comparison purposes, the total carbon, hydrogen and nitrogen contents were determined in an elemental analyzer (Fisons Instruments, Crawley, UK, EA 1108CHNS-O). 2.2.8. Dry and wet tensile strengths measurements in bonded pairs of cellophane films This procedure was patterned after the measurement of cohesion between two sheets of paper [25]. Thus, two zwitterion-modified cellophane films (5 cm¥10 cm) were sprayed with distilled water until wet, overlapped 2 cm along their long sides, so as to have an overlapped area (2 cm¥5 cm) and pressed at 10.3 kg·cm-2 between two stainless steel plates at 105ºC for 10 min. These dry samples were then cut into three strips (1.5 cm¥6 cm) and the tensile strength of the overlapping area was determined with a pendulum apparatus (Tappi standard test T404 cm-92). The wet strength was measured after first dipping the samples in distilled water for 2 min and removing the excess water by pressing the strips with a small roller between two dry sheets of paper. 3. RESULTS AND DISCUSSION
3.1. Mechanical properties of zwitterionic pulp handsheets Zwitterionic cellulose pulps were synthesized by reaction of never-dried (unbeaten) alpha-cellulose with a derivative reactant prepared from cyanuric chloride and the amino-acid lysine. Previous to this reaction, the amino-acid function was preserved by blocking it with copper, forcing the reaction of lysine with cyanuric chloride to occur through the free amino group (ε-position). Once the attachment of the lysine–cyanuric chloride complex had taken place onto the cellulose, the amino-acid function was freed by removing the copper by washing the modified pulp with EDTA.
Novel approaches to enhance adhesion of cellulose
257
Figure 1. Mechanical properties of handsheets made from zwitterionic pulps. ( ) Apparent den) burst index (kPa·m2·g-1), ( ) dry tensile strength (kN·m-1) and ( ) wet tensity (g·cm-3), ( sile strength (kN·m-1). Apparent density = (mass · m-2 of paper) divided by the thickness (a typical value for bulky paper is 0.5 g · cm-3). Burst index = force to rupture a paper sample held in annular clamps. It is commonly used in the paper industry as a measure of paper tensile strength. Tensile strengths = the force per unit width required to rupture a paper specimen.
Figure 1 shows the mechanical properties of two sets of handsheets prepared with modified cellulose. One set presents the properties of handsheets manufactured in the regular way (control pulp and zwitterionic-lysine-Cu pulp). In the other set of data, handsheets were pressed (under heat) to increase the contact area between the surfaces (control hot pressed and zwitterionic-lysine hot pressed). For handsheets prepared in the regular way, it can be seen that modification of cellulose fibers with zwitterions resulted in slightly increased burst index, although the apparent density was decreased. The tear index (not shown) was also increased. Even though the wet tensile strength was increased which was almost 7% of the dry strength, but this was not sufficient, since for wet tensile strength to be significant, it needs to be higher than 10% of the dry strength. On the other hand, the wet tensile strength of the control pulp was so low that it could not be measured by the instrument. The low increase in the strength properties of the modified pulp could be associated with the lower apparent density shown by this pulp which, in turn, did not
258
E. Delgado et al.
Figure 2. Effect of solids content on the tensile strength of never-dried pulp handsheets made with zwitterionic fibers. ( ) Alpha cellulose, ( ) zwitterionic pulp.
favor a good contact between fibers. In addition, a loss of flexibility of the modified pulp with respect to the untreated pulp, consistent with a stiffer and bulkier pulp [26], was found. From the bonding point of view, the application of pressure on the zwitterionic pulp handsheets should, in principle, increase the contact area between the surfaces, reducing consequently the bulkiness and increasing apparent density. The practical application of this principle in the form of pressing wet zwitterionic handsheets under heat (hot-pressed pulps) yielded, in fact (see Fig. 1) overall better mechanical properties, especially a noticeable high wet tensile strength (about 16% of the dry strength). To obtain information on the possible reasons for the wet strength increase exhibited by zwitterionic pulp handsheets, the tensile strength of never-dried webs at different solids contents was measured (see Fig. 2). It was observed that the tensile strength of zwitterionic pulps was similar to that of the untreated alphacellulose up to 40% solids content. Above this level, zwitterionic handsheets became stronger up to approximately 60% solids content. It has been speculated that beyond 60% solids content hydrogen bonding occurs extensively throughout the paper web [27]. From these results (Figs 1 and 2) it can be said that zwitterions may contribute positively to the tensile strength of never-dried webs at low solids content and that the effect might be obliterated once the hydrogen bonding takes over. In order to further advance the understanding of the possible contribution of zwitterions to the bonding of cellulose surfaces, the changes in fiber morphology
Novel approaches to enhance adhesion of cellulose
259
brought about by the chemical modification have to be avoided. One way to accomplish this is to study the adhesion between zwitterionic films of cellophane, which is regenerated cellulose and can function adequately as a model cellulose surface. 3.2. Bonding of zwitterionic cellophane films Zwitterionic cellophane films were prepared and the rupture force to separate two films with an overlap area of 3 cm2 was investigated both in wet and dry states, by means of an experimental design (central orthogonal composite design 23 with 9 replicates at the center, i.e., 8 runs at each vertex of the cube, two in the center of each axis (6 runs) and 9 replicates at the center of the cube, which amounts to a total of 23 runs) described elsewhere [28]. A central composite experimental design allows estimation of interactions and even the quadratic effects of the factors involved and, therefore, gives an idea of the (local) shape of a response surface. A response surface is a prediction of the values that variables under study can take within the range of parameters explored. The experimental design allowed to study the amount of amino acid fixed onto the film by varying the amount of added dichloro-s-triazinyl-lysine, reaction time, and pH. Table 1 shows relevant results from the experimental design, and it can be seen that the rupture force both in dry and wet states increased with increasing amount of amino acid fixed. The dry rupture force was more than doubled compared to unmodified films. The wet rupture force was dramatically increased from an undetectable value corresponding to an unmodified film to 24–35% of the dry strength in zwitterionic films. Figure 3 shows the response surface for the dry rupture force with a fixed reactant amount of 2.8 mmol/g. It is interesting to note that the dry rupture force was indeed increased under the experimental conditions. This fact is in contrast to the relatively low increase of dry tensile strength in zwitterionic paper webs. As it was previously shown, the modified fibers were less flexible [26] and the paper was bulkier as a result of the modification of the fibers, which might have impaired a good contact between the fibers and thus resulted in poor hydrogen bonding. Thus, the chemical modification changed the natural ability of the fibers to come into come into close contact under normal paper manufacturing conditions, resulting in a stiffer and bulkier paper, whereas this effect might have not occurred in cellophane films. Figure 4 shows the response surface for wet overlapped zwitterionic cellophane films as a function of reactant added and pH (the reaction time appeared not to be significant in this case). The force to rupture showed a dependency on both factors, being maximum at low to neutral pH and at larger amounts of reactant. This behavior should be expected from both the modified fibers and cellophane films by the introduction of zwitterions as well as the dependence of these functional groups on pH.
260
E. Delgado et al.
Table 1. Wet and dry rupture forces of overlapping zwitterionic cellophane films as a function of the amount of amino acid fixed Amino acid fixed (µmol·g-1)
Dry rupture force (N)
Wet rupture force (N)
– 19 44
26.48 29.42 58.55
ND* 7.16 19.61
* ND, not detected.
Figure 3. Response surface for dry rupture force of two overlapping cellophane films as a function of pH and reaction time (reactant load was 2.8 mmol/g on cellophane).
So far, the zwitterionic concept has been built around the hypothesis that once amino-acid molecules are attached to the cellulose surface, providing that the amino-acid function is free, they can behave as zwitterions under proper pH conditions. In the case of a pure lysine solution the isoelectric point is 9.8, which is the average of 9.0 ( pK a 2 of the amino-group at the ε position) and 10.5 ( pK a 3 of the amino-acid function). However, in lysine molecules that have been attached to cellulose surfaces via formation of a complex involving reaction of a coupling agent such as cyanuric chloride (s-triazine molecules) and the ε-amino group of the amino acid, the balance of charges would be achieved around a pH of 5.6, in absence of other major sources of ionic groups. This corresponds to an average of 2.2 ( pK a1 of the carboxyl group) and 9.0 ( pK a 2 of the α-amino group). In this case, the pH at which the maximum wet tensile strength was observed is close to the theoretical isoelectric point of a reacted lysine. This fact suggests indirectly
Novel approaches to enhance adhesion of cellulose
261
Figure 4. Response surface for wet rupture force of two overlapping cellophane films as function of pH and reactant amount (reaction time was 24.5 min).
that zwitterions are active on the surfaces in contact and that they have an effect on the wet rupture force. These overall results demonstrated that cellophane functioned well as a model for studying adhesion between modified cellulose surfaces. It was possible to observe the effect of zwitterions on the adhesion of cellulose surfaces using cellophane films. On the other hand, modified lignocellulosic fibers showed difficulty in their ability to bond one another due to the chemical modification and the effects of zwitterions were not so clear in the dry strength of paper. 4. CONCLUSIONS
It has been shown that the introduction of zwitterions onto and into fibers affects both their bonding ability and their flexibility. The wet tensile strength of paper made from these modified fibers was improved at solids content levels below 60%. The introduction of zwitterions on cellophane surfaces occurred efficiently while in pulp fibers it was more difficult, possibly due to the well-known presence of micropores in the fiber cell wall. Since the pH influenced the wet rupture force of overlapping cellophane surfaces, zwitterions must have been involved in the adhesion. The incorporation of zwitterions on the surface of cellulose fibers and regenerated cellulose films enhanced adhesion, which was reflected in both improved dry and wet strengths.
262
E. Delgado et al.
REFERENCES 1. G. Toriz, R. Arvidsson, M. Westin and P. Gatenholm, J. Appl. Polym. Sci. 88, 337 (2003). 2. A. H. Nissan, in: The Formation and Structure of Paper, F. Bolam (Ed.), p. 119. Technical Section British Paper and Board Makers Association, London (1962). 3. D. H. Page, Tappi 52, 674 (1969). 4. G. G. Allan, P. Mauranen, M. D. Desert and W. M. Reif, Paperi ja Puu 50, 529 (1968). 5. A. H. Nissan, in: Surfaces and Coatings Related to Paper and Wood, R. H. Marchessault and C. Skaar (Eds.), p. 221. Syracuse University Press, Syracuse, NY (1967). 6. G. G. Allan, F. Liu and P. Mauranen, Paperi ja Puu 52, 403 (1970). 7. G. G. Allan and W. M. Reif, Svensk Papperstidning 74, 563 (1971). 8. I. E. Pensa, G. C. Tesoro, R. O. Rau and P. H. Egrie, Textile Res. J. 36, 279 (1966). 9. R. F. Schwenker, Jr. and L. Lifland, Textile Res. J. 33, 107 (1963). 10. D. Rattee, in: The Chemistry of Synthetic Dyes, K. Venkataraman (Ed.), p. 444. Academic Press, New York, NY (1978). 11. K. Ward, Jr., Chemical Modification of Papermaking Fibers. Marcel Dekker, New York, NY (1973). 12. G. G. Allan and W. M. Reif, Trans. Tech. Sect. Can. Pulp Paper Ass. 1, 97 (1975). 13. G. G. Allan, E. Delgado and F. A. Lopez-Dellamary, in: Products of Paper Making, C. F. Baker (Ed.), p. 1101. Pira, Oxford (1993). 14. E. Delgado, The Potential of Zwitterionic Bonding in Paper, Ph.D. Thesis. College of Forest Resources, University of Washington, Seattle, WA (1994). 15. F. A. Lopez-Dellamary, Surface Modification of Pulp Fibers with Amino Acids for Zwitterionic Bonding, Ph. D. Thesis. College of Forest Resources, University of Washington, Seattle, WA (1991). 16. Forming Handsheets for Physical Test of Pulp. PAPTAC Standard Testing Method C.4 (1997). 17. Grammage of Paper and Paper Board. PAPTAC Standard Testing Method D.3 (1997). 18. Thickness and Apparent Density of Paper and Paper Board. PAPTAC Standard Testing Method D.4 (1993). 19. Internal Tearing Resistance of Paper, Paper Board and Pulp Handsheets. PAPTAC Standard Testing Method D.9 (1993). 20. Bursting Strength of Paper. PAPTAC Standard Testing Method D.8 (1993). 21. Tensile Breaking Strength of Paper and Paper Board. PAPTAC Standard Testing Method D. 6H (1984). 22. Wetted Tensile Breaking Strength of Paper and Paperboard. PAPTAC Standard Testing Method D.10 (1992). 23. R. S. Seth, M. C. Barbe, J. C. R. Williams and D. H. Page, Tappi 65, 135 (1982). 24. S. Moore and W. H. Stein, J. Biol. Chem. 211, 907 (1954). 25. P. Krkoska, P. Misovec and A. Blazej, Cellulose Chem. Technol. 18, 507 (1984). 26. E. Delgado, F. A. Lopez-Dellamary, G. G. Allan, A. Andrade, H. Contreras, H. Regla and T. Cresson, J. Pulp Paper Sci. 30, 141 (2004). 27. R. W. Davison, in: Dry Strength Additives, W. F. Reynolds (Ed.), p. 26. Tappi Press, Atlanta, GA (1980). 28. J. A. Velásquez, J. A. Andrade, S. C. Vargas, G. G. Allan, E. Delgado, F. A. López-Dellamary and L. R. Bravo, Cellulose, submmitted (2006).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 263–283 Ed. K.L. Mittal © VSP 2007
An in-mold application of adhesion promoters to polyolefin substrates THOMAS SCHUMAN,* MANINDER SINGH and JAMES STOFFER Department of Chemistry and Materials Research Center, University of Missouri-Rolla, Rolla, MO 65409, USA
Abstract—Several different treatments of polyolefinic plastics are used to improve adhesion to their low energy surfaces. One conventional method is that of tie-coating, e.g., chlorinated polyolefins applied to the polyolefinic parts as a solution or dispersion and dried as a thin, well-adhered film. Another potential process would involve application of similar coatings at the time of part fabrication, e.g., during a molding process, utilizing the process heat instead of solvents to provide a reduced viscosity and adhesion to the polymer surface. Several commercial chlorinated polyolefins (CPOs) were applied as films using the transfer from a dry mold surface or by a conventional (control) wet film application of the adhesion promoter onto the surfaces of polypropylene (PP) and a commercial thermoplastic polyolefin (TPO). Injection molding conditions were optimized for the neat polymers in a standard ASTM dogbone mold profile. Adhesion of a polyurethane acrylic coating to the molded parts was then tested by a cloth peel method. Results showed that in-mold process coatings’ adhesion was significant but less than that obtained for conventionally applied films. The magnitudes of adhesion were dependent on the composition of the CPO used, as well as of the substrate. A single treatment of CPO to the mold surface was shown to transfer coating to the injection molded polymer surface during each of several consecutive injection mold cycles, although film thickness and adhesion to the applied CPO film gradually decreased over the successive injection cycles. A correlation between adhesion values and the amount of chlorine transferred as a CPO coating during the molding process was observed. Keywords: Polyolefins; adhesion promoters; in-mold application.
1. INTRODUCTION
The low surface energies of polyolefin plastics often necessitate a chemical or mechanical pretreatment of the substrate in order to attain adhesion of paint to the substrate [1]. Chlorinated polyolefins (CPOs) can aid the adhesion of paint to plastic when applied via the conventional process [1, 2], i.e., through spray application of a wet coating film followed by drying and heat treatment [3–7].
*
To whom correspondence should be addressed. Tel.: (1-573) 341-6236; Fax: (1-573) 341-6033; e-mail:
[email protected]
264
T. Schuman et al.
Compared to metals, it is difficult to obtain good adhesion to polymers, in particular polyolefins [1–11], due to their inherently low surface energy and hence low wettability. Semi-crystalline polymers are prone to surface crystalline morphologies and poor cohesive strength in near-surface domains that constitute weak boundary layers in the plastic [6, 7]. Compression and injection molded polymers can have stress concentrations in the coating-plastic composite that affect paintability [6, 7, 11], which can be observed through thermal history. As explained by Kinloch [12], adhesion can involve mechanical interlocking, diffusion, or adsorption mechanisms. Mechanical interlocking involves the penetration of adhesion promoter or paint into the macro- or micro-irregularities of the surface. While it is debatable whether mechanical interlocking occurs for polymer surfaces, the polymer surfaces can be roughened through mechanical and chemical process [2, 6, 11] to enhance what might be thought of as a mechanical interlocking process. Geometric patterns are observed to control polymer adhesion strength and release properties [6]. Plasma and chemical processes can also alter the chemical composition of the polymer surface to enhance its adhesion strength [13]. Polymer adhesion may also be due to either polymer chain entanglements or resistance to chain pull-out. Diffusion theory involves the interdiffusion of polymer chains across a polymer/polymer interface. According to Voyutskii [14] the intrinsic adhesion of polymers to themselves and to each other was due to mutual diffusion of polymer molecules across the interface, which requires that the macromolecules and the chain segments of the polymers (adhesive or coating and substrate) possess sufficient mobility and are mutually soluble. The latter requirement can be restated by the condition that macromolecules possess similar values of solubility parameter. The Hansen solubility parameter (δs, of solvent) is defined as:
δs = ((∆H v – RT )/V )1/2 ,
(1)
where ∆Hv is the molar heat of vaporization, R is the gas constant, T is the temperature in Kelvin (K) and V is the molar volume. Solubility parameters for solids have been calculated from interaction strengths of functional groups within the polymer repeat unit per molar volume of the solid. If the solubility parameters of the coating and substrate polymer are not similar then negligible mixing of the polymer phases occurs, even of the chain ends [15]. If one of two contacting polymers is either highly cross-linked, crystalline or below its glass-transition temperature (Tg) prior to contact, then interdiffusion seems to be an unlikely mechanism for adhesion and is observed to negatively affect paintability [6, 7, 11]. Solvent choice of the coating film then becomes a critical factor in achieving adhesion [6, 7, 11]. Adsorption theory proposes that when sufficient intimate molecular contact is available at the interface, interatomic and intermolecular forces will be established between the atoms and molecules in the surfaces of both adhesion promoter or coating and substrate. These forces may be primary or secondary forces of at-
In-mold application of adhesion promoters
265
traction. Chemisorption is ascribed to primary forces, which for polymers can be ionic or covalent in nature. However, the most common intermolecular forces for polymers are van der Waals forces (secondary forces). Other forces in this category are polar bonding and hydrogen bonding, which are polar interactions stronger than van der Waals. The polar forces can be thought of in terms of acidbase interactions that are stronger for acid-base or base-acid and weaker for acidacid or base-base interactions [16, 17]. Thus proper matching of composition may be necessary to achieve maximum adhesion strength. The topographical profile of surfaces has recently been shown to be significant in the magnitude of adhesion to the surface with little regard for the strength of the adhesion forces involved [11, 18, 19]. Sub-surface, topographical profiles of interfacial features have recently been visualized by laser fluorescence confocal microscopy of dye-labeled CPOs [19]. The compositional structures of the conventionally applied, compared to in-mold applied, coating failure surfaces suggest different subsurface interfacial features. These features might be probed using a confocal microscopy technique to determine how the adhesion quality is affected by the subsurface interfacial structures and how the interfacial structures are controlled by the molding conditions. Molding has previously been utilized for applying surface coatings as a twostep process, where normal injection molding process was followed by postinjecting a coating into the substrate mold [20, 21]. In this study, the heat and pressure experienced during an injection molding process were utilized to transfer a CPO film from the mold surface to the molded part. Then, the magnitude of adhesion for the injection-molded parts was compared to a conventionally applied CPO coating, which was considered an experimental control. Two different polyolefin materials were investigated as the injection molded substrates, a polypropylene and a thermoplastic polyolefin. Energy and production time might be saved if an adhesion promoter can be applied during molding of the plastic part instead of as a separate coating process. 2. EXPERIMENTAL
2.1. Materials Several CPOs from Eastman Chemical (153, 343, 515, 310w and 347w) were used as adhesion promoting primer coatings, which were applied by both the traditional (wet film) application methodology [3, 4] as well as by the experimental injection molding process. All CPOs contained 15 to 23 wt% of chlorine with maleic anhydride functionality. The CPOs had glass transition temperatures of about 15°C and softening points between 80 and 100°C. The three solventborne CPOs, designated A, B, and C, were diluted to 5% solutions in toluene to achieve a fast evaporation rate on the mold surface while the waterborne dispersions
T. Schuman et al.
266
Table 1. Molding conditions for injection molded polymer samples Polymer
Nozzle temperature (°C)
Mold temperature (°C)
Nozzle size (cm)
Pressure (MPa)
PP TPO
216 182
27 29
2.2 2.4
7.7 12.9
Table 2. Two-component polyurethane (2K PUR) cross-linked paint formulation Material
wt%
Rohm and Haas acrysol AU 608 (EW 1000) Rhodia tolonate HDB (EW 179) N-butyl acetate Toluene 1-Methoxy-2-propanol acetate Sn(Bu)2(O2C12H23)2 catalyst
61.61 12.15 15.15 3.92 7.15 0.02
(designated by -W) CPO D-W and CPO E-W were used as received. Toluene of ACS reagent grade was obtained from Fisher Chemical. Polypropylene (11S12A, PP) and a flexible thermoplastic polyolefin (WL107, TPO) beads were obtained from Huntsman Polymers. To generate the sample surfaces utilized in the conventional film application study, the polymers were molded into dogbone strips of 1.3¥25.4¥0.3175 cm in dimension. The optimized molding conditions employed are given in Table 1. An unpigmented, two-component acrylic polyurethane (2K PUR) crosslinked paint (Table 2) was used for measuring adhesion to the molded polymer surfaces via the cloth peel test method [3–5]. The paint was formulated at the stoichiometric ratio NCO/OH = 1.1:1. The cloth was a 250-thread count, 120 g/m2 white 80:20 polyester/cotton blend cloth. Dibutyl tin dilaurate catalyst used in the urethane paint was obtained from Acros Scientific. 2.2. Sample preparation 2.2.1. Traditional (wet film) CPO coating application Molded PP and TPO sheets were cut into 1.3¥20 cm rectangular pieces. It was necessary to clean the surface of the sample, as contamination on the surface can result in the formation of a weak boundary layer [5–7, 22]. Polyolefin strips were cleaned with 5% KOH solution using a lint-free tissue and rinsed with deionized water. After cleaning, CPO solution was applied as thin, 12-µm wet films, metered by a doctor blade during traditional wet film application. Samples were then heat treated in an oven at 110°C for 3 h. Previous research by Stoffer and co-
In-mold application of adhesion promoters
267
workers [3–5] had shown that proper heat treatment generally increased the adhesion. The control for these experiments consisted of the urethane acrylic coating applied to the bare molded substrate polymer. 2.2.2. Molded samples Molded samples were made under the same injection-molding parameters as uncoated dogbones (Table 1). The injection molding cycle was ceased briefly while a CPO solution or waterborne dispersion was applied. An excess amount of CPO solution or dispersion was evenly applied to the mold surface using hand spray, the excess briefly allowed to run-off and the film to dry (approx. 30 s), and then injection molding process was continued at a normal pace. The thicknesses of the films on the mold surface were 4 µm for solventborne CPOs and about 3 µm for waterborne CPO films. After a CPO film had been applied to the mold, nine samples were molded in succession and numbered according to their order. For these experiments, results for the first five samples are reported. An experiment (labeled Exp 1, Exp 2, etc.) consisted of a single application of the CPO polymer to the mold surface and 9 subsequent injection molding cycles, which produced a sequential series of 9 molded parts or nine samples per experiment. Three experiments were made for each CPO, which were then labeled by experiment number and injection cycle (sample) number within the experiment, for instance, Exp 1: sample 1, sample 2, …; and Exp 2: sample 1, sample 2, …, etc. 2.2.3. Application of paint Paint was later applied along with a paint-saturated cloth strip [3–5] in preparation for the cloth peel test. Cloth of width similar to the sample width was cut in length double the length of sample. The cloth strip was dipped into the paint, the excess was removed, and then the paint-cloth strip was applied to the sample surface. The extra length was mounted in the tensile testing grip for peel testing. 2.3. Adhesion and surface composition characterization 2.3.1. Peel adhesion The cloth peel adhesion test was then used to measure adhesion strength after a secondary bake of 100°C for 30 min [5]. Samples were mounted in the grips of an Instron tensile testing device. The average load per unit width was reported as the adhesion strength for regions after the load had reached an initial maximum or had appeared to equilibrate [23] (see Fig. 1). Seven traditionally coated samples were tested for each type of CPO, which is reported as the average value and standard deviation. The statistical performance of in-mold coated samples was assessed via the trend of the molded series and across repeated experiments. Control experiments consisted of the urethane acrylic coating applied directly to the bare polymer surface. An Instron model 4204 with a 1 kN load cell was used to measure the peel strength of the urethane acrylic coatings. The substrate was held in
268
T. Schuman et al.
Figure 1. Peel strength as a function of tensile displacement for the 180° peel geometry for CPO B adhesion promoter applied via the mold application. Arrows mark the onset and endset of data averaging upon removal of grip slack to achieve “steady-state” values. Width of specimen is 1.3 cm.
the lower, stationary clamp and cloth was secured in the upper clamp in a 180° peel geometry. The grips were separated at a speed of 10 cm/min. The surfaces created by peel testing were then analyzed for appearance and composition (Fig. 1). 2.3.2. Scanning electron microscopy (SED)/energy dispersive spectroscopy (EDS) A Hitachi S-4700 Field Emission SEM was used to image the surfaces and perform EDS line scanning and compositional phase mapping of samples as well as to capture images using a magnification of 10 000¥. Cloth and substrate surfaces of the failed interface were analyzed and the results reported as the average of 5 measurement points. Samples were sectioned perpendicular to the CPO-substrate interface using a standard flat utility blade and then sputter-coated with Au/Pd (60:40) for about 60 s at 3 mA. EDS was used to analyze the interfacial regions for their elemental composition. 2.3.3. XPS XPS measurements were made using a Kratos Axis-165 XPS system. The data reported are for sputtered surfaces, which were cleaned using argon ions for 60 s. The excitation source was a magnesium Kα filament (1253.6 eV) at 225 W. Electron detection was for a wide aperture so that a broad area of the surface was analyzed at the take-off angle of 45°. Survey scans were made at a pass energy of 80 eV and high-resolution scans were collected at a pass energy of 20 eV. Corrections for charging and baseline corrections were made using X-Kratos software. Shirley baseline corrections were applied. Gaussian peaks were applied for high resolution spectral fitting and peak assignments. Spectra were surface charge corrected according to the internal standard aliphatic carbon peak, assigned as 285.2 eV.
In-mold application of adhesion promoters
269
2.3.4. Wide angle X-ray scattering (WAXS) Grazing incident angle WAXS was used to analyze the crystallinity in the substrates. Experiments were run on a Philips X-Pert diffractometer using a copper X-ray filament source and PW3011/20 proportional detector. Experiments were performed at decreasing grazing angles to analyze crystallinity at different depths of the substrate. 3. RESULTS AND DISCUSSION
3.1. Peel strength characterization Peel adhesion measurements showed a fairly consistent slip-stick peel behavior (Fig. 1). Averaged values of the peel force measured over the sample length were sed to calculate eel adhesion. Figure 2 shows the peel strengths of the conventionally applied CPO films on PP and TPO. CPO B produced the highest
Figure 2. Peel strengths of traditional wet film application of CPO coatings on PP and TPO. White bars indicate standard deviations.
Figure 3. Peel strengths for solvent-borne adhesion promoters applied by the in-mold process to polypropylene (PP) and thermoplastic polyolefin (TPO) substrates. Traditional values are re-plotted from Fig. 2 (= Conventional) for easier comparison to sample data.
270
T. Schuman et al.
Figure 4. Peel strengths for water-borne adhesion promoters applied by the in-mold process to polypropylene (PP) and thermoplastic polyolefin (TPO) substrates. Traditional values are re-plotted from Fig. 2 (= Conventional) for easier comparison to sample data.
adhesion to the polymer surfaces when applied via the conventional wet film method and adhesion was greater to the PP substrate compared to the TPO substrate. Adhesion by the CPO B to TPO was also similar to adhesion strengths observed for CPO films applied by the conventional method to PP. The controls for both polymer substrates displayed zero adhesion, which may be partly explained by a stiff urethane acrylic test coating and the 180° test geometry [23]. Figures 3 and 4 show the adhesion data for the in-mold application of films of solventborne and waterborne CPOs, respectively, to PP and TPO substrates. The last entry for each experimental series shows the traditional wet film coating adhesion, which acts a control for the experiments. Sample 1 comprises the first sample produced after application of the CPO film on the mold surface, sample 2 the second, and so forth across each experiment. 3.1.1. Mold applied solvent-borne CPO films Figure 3 shows peel strengths for different CPOs applied via the in-mold and control methods to molded PP and TPO polymers. All solvent-borne CPOs adhered to the urethane acrylic test coating better than the control. Results showed that solventborne CPO B films generally adhered well to both PP and TPO. CPO B adhered better than other CPOs applied via either the wet film or mold applied methods. Solvent-borne CPO B applied by injection molding outperformed water-borne CPOs. Sometimes higher peel strengths were observed for the second or later samples than the first sample of an experiment. For in-mold applied CPO A coating on PP, later samples had adhesion greater than the initial samples. This behavior could be due either to the CPO film being applied to the mold too thickly with a high initial transfer efficiency to the molten polymer material or to just a higher initial transfer efficiency. In this case, the coating layers on the first sample might
In-mold application of adhesion promoters
271
have been too thick. For instance, CPOs require relatively thin film applications to avoid brittle fracture and poor adhesion [2, 24]. Films on subsequent samples were thinner and performed better in the 180° peel geometry test. 3.1.2. Mold applied water-borne CPO films Water-borne CPOs, which displayed poor adhesion, did not transfer well to the TPO film by means of the in-mold process. On the other hand, CPO D-W performed remarkably well on PP surfaces. CPO D-W performed similarly to the best solvent-borne CPO, CPO B, on PP substrate but showed no adhesion to the TPO material. CPO E-W provided low but consistent adhesion to PP and no appreciable adhesion to the TPO substrate material. While Perdomo [25, 26] has indicated that solvent-borne CPO increases adhesion more than water-borne CPO where solvent mediates the contact between the CPO and the substrate, there should be no influence of solvent in these tests. Use of a solvent-borne CPO product was insufficient to provide adhesion promotion. The term ‘solvent-borne CPO,’ however, has implications with respect to CPO polymer composition and providing compatibility with a polymer type in the absence of solvent. For instance, the results from Fig. 4 show that waterborne CPOs D-W and E-W worked better in the mold transfer process than solventborne CPO C on the PP substrate. Peel adhesion results typically showed that for injection-molded samples the adhesion strength decreased with consecutive samples (Figs 3 and 4). While the first experiment of CPO E-W samples that were mold applied to TPO showed initially large adhesion with consecutive samples displaying less adhesion, a typical pattern for these samples, all subsequent experiments gave no measurable peel adhesion strength. CPO E-W data, therefore, seem to demonstrate an inconsistency with regard to transfer of the CPO from the mold surface onto the TPO substrate. In the case of CPO E-W on PP (Fig. 4), the adhesion strength of the first sample was 205 g/cm and then decreased uniformly with consecutive samples. Figures 3 and 4 also showed that the conventional wet film applications gave consistently higher adhesion than their in-mold applied counterparts. A reduced peel adhesion of the in-mold films compared to the conventional application process appears to be related to film thickness of the CPO layers (see below). In the case of the in-mold applied coating samples, differences could also be ascribed to a poorer CPO-to-substrate polymer compatibility of a dry CPO composition compared to the wet film composition of a traditionally applied coating sample. Another possible reason for performance differences between these samples might be related to heat treatment of the CPO coating layer. Stoffer and co-workers [3–5] found that heat treatment (3 h at 110°C) was required for an effective application of CPOs, followed by a second heating for 30 min at 100°C for application of the urethane acrylic coating film. Recently, a temperature (baking) requirement for optimization of the adhesion was not observed by Ma et al. [19]. Conventional samples in this study were heat treated for 3 h at 110°C. Other issues not addressed in this work were the effects of process heat, pressure, or time to optimize mixing and annealing of the CPO coating to the molded polymer.
272
T. Schuman et al.
While these results confirmed that the mold application method could be used to apply CPO films to the injection molded polymer during the injection molding process and provide enhanced adhesion compared to the bare polymer, the traditional application method remained superior in adhesion compared to the experimental injection-mold application method. Water-borne CPOs, in particular, displayed extremely poor adhesion performance when applied using the injection mold process to the TPO substrate. Poor performance by the water-borne CPO might be due to the alkylphenol–ethoxylate surfactants used for product dispersion stability, which have poor thermal stability or might worsen compatibility with the TPO surface. However, peel strengths were promising for in-mold applications in case of CPO B to both substrates and of CPO D-W to PP. 3.2. SEM/X-ray EDS characterization SEM was utilized to view the interfacial failure regions (Fig. 5) and to determine surface compositions via X-ray EDS (Figs 6–10). In this way, the appearance
Figure 5. (a) SEM image of cloth-side of injection-mold applied CPO (paint/CPO B/TPO system) peeled sample. (b) The corresponding SEM image of the substrate-side for the same injection mold applied paint/CPO B/TPO sample imaged in (a). These images show protrusions from the surface that were commonly observed to have different compositions compared to that of the surface.
In-mold application of adhesion promoters
273
Figure 6. EDS data for percentages of elements on a wet-film-applied CPO B/PP sample. The (¥10) designation indicates that the atomic percentages of chlorine plotted are 10-times larger than the actual values.
Figure 7. EDS data for the percentages of elements on in-mold applied CPO D-W/PP sample. The values of atomic percentages are written for those less than 4%.
Figure 8. EDS data for percentages of elements on in-mold applied CPO E-W/PP sample. The values of atomic percentages are written for those less than 4%.
274
T. Schuman et al.
Figure 9. EDS data for percentages of elements on an in-mold-applied CPO B/TPO sample. The values of atomic percentages are written for those less than 4%.
Figure 10. EDS data for percentages of elements at the surfaces created upon adhesional failure of a conventionally-applied wet CPO D-W film on TPO sample. The (¥10) designation indicates that the atomic percentages of chlorine plotted are 10-times larger than the actual values.
and compositions at the location of fracture might be devised. Figure 5 shows the appearance for typical samples, i.e., those displaying non-zero adhesion strength, of surfaces which resulted after peel adhesion testing. Cloth side and substrate side analyses were made. In particular, the presence of protrusions from the surface was noted. These protrusions typically had composition different than the underlying substrate. For example, the protrusion from the surface might show different percentages of elements compared to the solid surface at the base of the protrusion. Unfortunately, these protrusions, thus, made a strict analysis of a failure surface composition, or a straightforward assignment of location of the failure within the layered composite, difficult. For systems displaying zero adhesion, generally smooth surfaces of differing compositions resulted, corresponding to the urethane acrylic coating and the molded polymer substrate. In many instances, a cross-
In-mold application of adhesion promoters
275
sectioned failure surface would possess backscattered electron image contrast near the surface that, according to EDS line-scans and compositional mapping, corresponded to the presence of chlorine (data not shown). 3.2.1. PP adhesion failure interface compositions EDS data from the peel adhesion surfaces created upon failure of wet film and inmold applications of CPO B coatings to PP (Experiment 1, sample 5, 180° peel adhesion 197 g/cm) are presented in Fig. 6. The actual percent amount chlorine in the figure was magnified by a factor 10 to permit visualization of the low surface concentrations compared to that of the CPO B. For the conventional application, Fig. 6 shows that a relatively large amount of chlorine was present on the clothside but was not found on the substrate-side surface of the failure interface. Nitrogen was mostly found on the cloth-side surface but was also found on protrusions from the substrate (see Fig. 5). Nitrogen is not present in the neat CPO B sample utilized as a control specimen. The pictures of the failure interface and the compositional information strongly suggested that the failure for the conventional sample took place cohesively within the PP phase for the CPO B-PP system. Given that the penetration depth of the EDS experiment is on the order of a micrometer, not detecting the presence of chlorine on the substrate-side surface suggests that failure was cohesive within the substrate layer. Surface protrusions that had compositions different than the average surface complicated these analyses. The conventionally applied CPO B-PP samples showed relatively large differences in compositions of the cloth-side surfaces compared to substrate-side surfaces (Fig. 6). However, the in-mold applied CPO B-PP specimen showed similar cloth and substrate compositions, which were very similar to the cloth-side surface composition of the conventionally applied sample. In addition, the total amount of chlorine at the fracture surface was much less for the in-mold applied sample compared to the amount of chlorine found on the substrate-side surface of the conventional coating process. The quantity of nitrogen on the in-mold applied sample was less than on the conventionally applied sample and was nearly identical on its cloth and substrate surfaces. Similar compositions of the failure surfaces still suggest a cohesive failure, though not strictly in the substrate as proposed for the conventional substrate. A much thinner CPO film for the in-mold applied coating was apparent. For CPO D-W applied via in-mold method to PP (Fig. 7), the compositional results showed no chlorine at the substrate surface, but chlorine was present at the cloth side of the fracture interface. Again, it is suggested here that the mode of failure was cohesive within the substrate layer since the analyzed surfaces present surface compositions similar to those observed for the conventionally applied CPO B coating on PP. As chlorine was present on the cloth and not on the substrate, cohesive failure within the substrate appears likely. Given that the penetration depth of the EDS experiment is on the order of 1 micrometer and given the presence of surface protrusions after adhesional failure, detecting the presence of chlorine on only one side of the failure interface also suggests that failure was co-
276
T. Schuman et al.
hesive, in this case within the substrate layer. The presence of nitrogen on all surfaces appears inconsistent with cohesive failure within the substrate polymer since nitrogen is only associated with the CPO D-W and the urethane polymers. A cohesive failure behavior within the substrate, though, cannot be excluded due to the observed surface protrusions that were relatively high in nitrogen concentration which will influence the average surface nitrogen concentration. Figure 8 shows compositions of the surfaces created upon adhesional failure of a CPO E-W coating in-mold applied to PP substrate. This sample (Experiment 1, sample #1, 180° peel strength = 256 g/cm) showed more nitrogen content at low chlorine content and less nitrogen for higher chlorine content, supporting cohesive failure across the entire paint/CPO/substrate cross section or mostly cohesive failure within the CPO layer. In general, in-mold applied CPO-PP samples displaying adhesion had chlorine present at only one of the failure surfaces. Protrusions remained different in their composition compared to the surface. 3.2.2. TPO adhesion failure interface compositions Figure 5 shows the SEM image (paint/CPO/TPO) of the cloth and substrate surfaces created by adhesional failure of an injection mold-applied CPO B coating on TPO substrate (Experiment 2, sample 2, peel adhesion 376 g/cm). Composition of the surfaces for the in-mold applied CPO B coating on TPO (Fig. 9) was similarly affected by the composition of the protrusions created during the cohesive failure. The in-mold applied CPO coating sample surfaces created upon adhesional failure of non-zero peel adhesion strength samples showed high concentrations of chlorine on the substrate surface. Similar to CPO B-PP samples, protrusions from the cloth surface of the failure interface showed enhanced concentrations of chlorine compared to the cloth surface as well as enhanced oxygen concentration. It was concluded that the mode of failure was mostly cohesive failure within the TPO layer. The compositions of cloth-side and substrate-side of the failure interfaces for both conventional and in-mold applied CPO D-W coatings applied to TPO are shown in Fig. 10. Similar to the CPO B-PP conventional sample, the CPO D-WTPO conventionally applied coating sample showed durable peel adhesion (360 g/cm) that corresponded with a large chlorine concentration at the cloth surface (3.02%) as compared to the substrate surface (0.49%). Since CPO D-W contains nitrogen, the conventionally applied CPO sample failure interface appears to represent a cohesive failure across multiple layers of the paint/CPO/substrate sandwich. Actually, failure was mostly cohesive within the TPO substrate given the presence of surface protrusions having chlorine content. Taking into account the penetration depth of the EDS method thus suggested failure of the substrate, perhaps within the diffuse layer of CPO–TPO [1, 19, 23]. Surface protrusions again influenced the averaged compositions. For the in-mold applied CPO D-W coating on TPO polymer, neither surface of the failure interface showed evidence of chlorine, which suggested poor transfer of the CPO D-W to the molded polymer during the molding process. The 180°
In-mold application of adhesion promoters
277
peel strength for this sample was zero, which is consistent with little or no CPO being present at the interface to promote the adhesion of the paint to the TPO. As explained by Ryntz [2] the thickness of the CPO layer is critical to adhesion: if too thick, cohesive failure will occur within the ‘tie-coat’ but if too thin, adhesion to the substrate will not be attained. Thus, for a thick CPO layer on the substrate the mode of failure would be expected to be cohesive within the CPO layer. However, in general lower amounts of chlorine were found in comparison to the conventionally applied CPO coatings, which does not explain the reduced adhesion strength for the in-mold applied coatings. Based on analysis of the conventional versus in-mold applied coating adhesional failure interfaces, it appears that higher concentrations of CPO were observed for the conventionally applied materials but these produced cohesive failure within the substrate (beneath the CPO layer) of highest adhesion strength. The thinner, in-mold applied CPO coatings produced cohesive failures within the CPO layer for injection molded samples made just after application of CPO to the mold surface. Failure across multiple layers of the paint/CPO/substrate sandwich resulted for successive molded samples after the first molded sample, which showed the magnitude of adhesion dependent on the amount of CPO acquired during molding. 3.2.3. Adhesion as a function of amount of CPO Injection molded samples showed a decrease in peel strength for consecutive samples. It was assumed that the reason for reduced peel strength for successive samples was a progressively thinner CPO layer after each molding cycle. While making injection-molded samples, CPO was sprayed once on both sides of the molds and then injection molding was done to transfer the CPO to the molded polymer surface. Therefore, with successive samples, CPO would be consumed
Figure 11. Trend lines of peel strength compared to the amount of CPO B adhesion promotes present on the surface, represented by percent chlorine content at the TPO surface, are shown for successively molded TPO samples.
278
T. Schuman et al.
with transfer of CPO from the mold surface to the molded polymer surface. A reduced transfer over the course of the experiment would then result in successively decreased adhesion, which was observed in general. EDS and XPS results appear to support the hypothesis above. Figure 11 shows a trend of peel strength versus XPS-measured atomic percentage of chlorine (C–Cl*) present for successively molded CPO B-PP samples. A decrease in the atomic percentage of chlorine (C– Cl*) groups corresponded with a reduction in peel adhesion strength. 3.3. XPS characterization of failure surfaces Although SEM and EDS data were helpful in showing differences between interfacial structures and their compositions, supporting possible mode(s) of failure, the compositional analyses were limited by a lack of surface homogeneity produced by the cohesive failures, e.g., protrusions from the surface, and a lack of chemical information beyond the empirical chemical compositions provided by EDS. Occasionally the EDS scans would show a higher atomic percentage of oxygen and/or nitrogen than those of the control materials. The XPS technique was used to acquire additional chemical structural information in addition to empirical chemical compositional data acquired using EDS. The XPS technique, compared to EDS, similarly identifies the empirical amount of elements present on the surface but in addition provides chemical information regarding the surface chemical groups in which those elements participate. The high-resolution spectra for individual elements are convoluted photoelectron emission peaks, i.e., the additive emissions of different chemical groups. The chemical groups are deconvoluted into specific emission peaks using peak fitting software and knowledge about the probable surface groups present at the material surface. For example, a urethane sample should contain a peak within the C1s emission spectrum at about 287.3 eV that corresponds to C*=ONH, which is spectrally similar to the urethane functional group. Control surfaces of paint, CPO B and PP were analyzed and their information was applied to the analysis of fracture surface samples. Figure 12 shows the XPS high resolution spectra and chemical group peak fitting for the cloth-side and substrate-side surfaces of the CPO B-PP peel adhesion failure interface. Tables 3 and 4 show the chemical group assignments obtained from the XPS spectral analyses of control samples and the in-mold, CPO B-coated PP sample. XPS provides results for the outermost surface composition, which may or may not be representative of the bulk composition, since the photoelectron escape depth is only about 0.5 nm. XPS is, therefore, surface specific and its results can, therefore, be significantly different than the EDS results shown previously due to the difference in analysis depth. On the cloth-side, we observed five peaks in the C1s emission spectrum at 285.2 eV (C*– C), 286.1 eV (C*– N), 286.5 eV (C*– Cl), 287.3 eV (O–(C*=O)– NH) and 289.2 eV ((C*=O)–O–) that were similar to the peaks observed in the C1s emission spectra of control paint, as well as those for control CPO. The
In-mold application of adhesion promoters
279
Figure 12. XPS emission spectra of the surfaces created upon adhesional failure of CPO B applied during molding to PP substrate, cloth (coating) side spectra are on the left and substrate (PP) side spectra are found on the right.
T. Schuman et al.
280
Table 3. Atomic percent concentrations measured by the XPS emission spectra of PP, CPO B, urethane acrylic paint control surfaces, and also of the coating cloth-side and substrate-side surfaces that resulted after peel testing of the paint-CPO-PP substrate sandwich Binding energy (eV)
Chemical assignment
PP
285.2 286.1 286.5 287.3 289.2 528.3 529.8 531.5 532.5 533 397.8 198.1/200.4 200.3 199.6/202.0 102.2 103.1
C*–C C* = C, C*–N C*–Cl; C*–O (C* =O)–OC, NH(C*=O)–O (C*= O)–OH anion (shift) keto, amide (shift) anion, oxide, carbonate keto, (C = O*)–NH, siloxane (C = O*)–OH, Si–O2 urethane, isocyanate C – Cl* 2p 3/2 C–Cl* 2p 1/2; C–Cl*2 2p 3/2 C–Cl* 2p 1/2 Si*O 2p 3/2 Si*O 2p 1/2
90.49 82.20 9.51 – – 9.31 – 1.76 – 0.86 – – – – – – – – – 1.56 – – – – – 2.77 – 1.53 – – – –
62.03 6.18 – 5.11 5.65 5.28 3.07 6.07 – – – – – – 3.09 6.60 0.23 5.19 1.57 3.00 0.004 0.06 0.04 0.14 0.02 0.07 – 0.15 – 0.12
100
100
Percent total composition
CPO B
100
Paint
Cloth
Substrate
66.09 20.23
100
87.26 8.99 0.56 2.28 – – – – 0.63 – – 0.02 0.07 0.02 0.08 0.09 100
Table 4. Element analysis (at%) of PP, CPO B, paint, cloth and substrate
C O N Cl Si Total composition
PP
CPO B
Paint
Cloth
Substrate
100.00 0.00 0.00 0.00 0.00 100.00
94.14 1.56 0.00 4.30 0.00 100.00
95.04 3.32 1.57 0.07 0.00 100.00
84.67 11.79 3.00 0.27 0.27 100.00
99.10 0.63 0.00 0.11 0.16 100.00
composition as shown for the cloth surface corresponds to a mixture of about 64% CPO B and 36% paint. Chlorine was observed on both cloth and substrate surfaces. No nitrogen was observed at the substrate compared to 3 at% at the cloth surface, which suggested that no paint was present at the substrate surface. Despite the lack of nitrogen at the substrate surface, fitting of the XPS C1s spectrum required a small emission peak at 286.1 eV that was assigned to carbon of the
In-mold application of adhesion promoters
281
Figure 13. WAXS patterns of control TPO as a function of incident (grazing) X-ray beam angle (listed as incident grazing angle, in degrees).
urethane group in addition to that at 286.5 eV. The cloth-side surface data, therefore, showed the presence of chemical groups corresponding to both CPO B and paint while the substrate showed groups corresponding to only CPO B and the substrate. These results indicated that the failure was cohesive, located within the CPO layer or across a diffuse CPO B-PP substrate interface. 3.4. Wide angle X-ray scattering (WAXS) of TPO substrate WAXS was performed (Fig. 13) at different incident grazing angles on TPO samples to analyze crystallinity as a function of depth. A smaller incident grazing angle probes less penetration depth into the substrate. With increasing X-ray incident angle, analysis of the surface can be made at a greater depth. Diffraction peaks were observed but did not change significantly as a function of depth for samples with or without a CPO coating layer. Therefore, grazing angle WAXS data did not appear to show changes in crystallinity as a function of sample depth due to the in-mold CPO application method. 4. CONCLUSIONS
The purpose of this study was to examine the effect of chlorinated polyolefin (CPO) coatings applied during injection molding on the adhesion of paint to polypropylene (PP) and thermoplastic polyolefin (TPO) substrates. CPOs were applied by transfer of the CPO from the mold surface to the molded plastic part. Adhesion promoting CPOs were transferred to the injection-molded plastics and resulted in paintable molded parts. Results showed that the in-mold application method to PP or TPO substrates were similar in efficiency depending on the composition of the CPO utilized. The
282
T. Schuman et al.
in-mold applied coatings resulted in lower peel strength compared to traditionally applied solvent-borne or water-borne CPO adhesion promoters of identical composition. Water-borne adhesion promoters were not efficiently applied to the TPO substrate by the in-mold technique but worked well when applied in-mold to polypropylene. In general, solvent-borne CPOs tended to result in greater peel adhesion strengths than water-borne CPOs, which may be due to the alkylphenol ethoxylate surfactants present in the water-borne dispersions. Both adhesional and cohesive failures were observed for injection-molded and traditional wet-filmcoated samples. For solvent-borne CPOs applied by injection molding to PP, measurable peel strengths were achieved through up to the sixth sample after a single application of CPO to the mold surface, although peel strengths diminished slightly for successive samples. On the surfaces of successive samples, e.g., in-mold applied CPO B coating on TPO substrate, EDS and XPS were performed. The EDS and XPS results obtained suggested that the reason for the decrease in peel strength compared to conventionally applied CPO coatings was a decreased concentration of CPO on the surfaces of successive samples. Different morphologies of the CPO-polymer interface appear to result via the conventionally applied CPO coatings compared to the in-mold applied CPO coatings, which was evidenced by changes in the mode of failure. For conventionally applied CPOs, on both PP and TPO, the observed mode of failure was cohesive within the substrate. However, in case of in-mold coating applications the observed modes of failure were either cohesive within the substrate, as well as within the CPO layer, or through interfacial delamination. Interfacial delamination occurred in the case of waterborne CPOs applied by the in-mold technique to TPO due to an apparent failure of the CPO to transfer from the mold to the molten polymer. TPO samples with an in-mold applied CPO B coating, molded directly after the mold surface was treated with CPO B, failed within the CPO layer while successive samples failed cohesively within the substrate. REFERENCES 1. 2. 3. 4. 5.
R. A. Ryntz, JCT Res. 2, 350 (2005). R. A. Ryntz, Prog. Org. Coating 25, 73 (1994). W. L. Dechent and J. O. Stoffer, Polym. Mater. Sci. Eng. 72, 365 (1995). W. L. Dechent and J. O. Stoffer, Polym. Mater. Sci. Eng. 69, 380 (1993). J. M. Land, M. C. Arjona, W. L. Dechent and J. O. Stoffer, Proc. Waterborne, Higher-Solids, and Powder Coatings Symp. 23, pp. 397-407 (1996). 6. R. A. Ryntz, Prog. Org. Coatings 27, 241 (1996). 7. R. A. Ryntz, Q. Xie and A. C. Ramamurthy, J. Coatings Technol. 67, 45 (1995). 8. S. E. Stanley, US Patent 6,293,312 (2001). 9. P. Duvander and I. Postoaca, PCT International Application WO 9962708 (1999). 10. A. Kruse and G. Kruger, J. Adhesion Sci. Technol. 9, 1611 (1995). 11. T. P. Schuman and S. F. Thames, J. Adhesion Sci. Technol. 19, 1207 (2005).
In-mold application of adhesion promoters
283
12. A. J. Kinloch, Adhesion and Adhesives: Science and Technology. Chapman and Hall, New York, NY (1987). 13. F. Garbassi, M. Morra and E. Ochiello, Polymer Surfaces: From Physics to Technology, Chap. 10. Wiley, New York, NY (1994). 14. S. S. Voyutskii, Autoadhesion and Adhesion of High Polymers. Wiley, New York, NY (1963). 15. P. J. Flory, Principles of Polymer Chemistry, Chapter 8. Cornell University Press, Ithaca, NY (1953). 16. F. M. Fowkes, Indus. Eng. Chem. 56, 40 (1964). 17. K. L. Mittal (Ed.), Acid-Base Interactions: Relevance to Adhesion Science and Technology, Vol. 2. VSP, Utrecht (2000). 18. A. J. Crosby, M. Hageman and A. J. Duncan, Polym. Mater. Sci. Eng. 90, 761 (2004). 19. Y. Ma, M. A. Winnik, P. V. Yaneff and R. A. Ryntz, JCT Res. 2, 407 (2005). 20. E. J. Straus and D. S. McBain, US Patent 6,875,389 (2005). 21. T. Fujishiro, T. Izumida, K. Akahori and Y. Yamamoto, US Patent 5,902,534 (1999). 22. W. J. Bailey and E. T. Yates, J. Org. Chem. 25, 1800 (1960). 23. J. E. Lawniczak, K. A. Williams and L. T. Germinario, JCT Res. 2, 399 (2005). 24. R. J. Clemens, G. N. Batts, K. P. Middleton and C. Sass, Prog. Org. Coatings 24, 43 (1994). 25. G. R. Perdomo, Pitture Vernici Eur. 74, 43 (1998). 26. G. R. Perdomo, Pitture Vernici Eur. 74, 47 (1998).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 285–296 Ed. K.L. Mittal © VSP 2007
Design and control of surface properties of UV-curable acrylic systems ROBERTA BONGIOVANNI and ALDO PRIOLA* Department of Materials Science and Chemical Engineering, Politecnico di Torino, c.so Duca degli Abruzzi 24, 10129 Torino, Italy
Abstract—The surface properties of acrylic UV-cured systems can be modified using small amounts of fluorinated co-monomers, whose performance is strictly dependent on their structure and on the fluorine distribution in the molecule. Different classes of monofunctional fluorinated comonomers were co-polymerised with a typical hydrogenated oligomer via a radical UV-curing process. These include: (meth)acrylates of fluorinated alcohols having the following general structure F–(CF2)m–(CH2)nOH, urethane(meth)acrylates of perfluoropolyether of different molecular weights having the formula CF3CF2O(CF2CF2O)m(CF2O)nCF2CH2O– and urethane(meth)acrylates of perfluoropolypropyleneoxide having different molecular weights and different end groups X (X= Cl or H), whose general structure is XC3F6O-[CF2CF(CF3)O]n-1CF2CH2O–. The main factors controlling the surface activity of the fluorinated co-monomers were studied. With any of the comonomers investigated the surface modification obtained was due to their selective migration to the surface before curing. The curing reaction immobilized the surface structure formed, giving rise to stable properties, independent of the environments to which films were exposed. The surface migration was instantaneous for some co-monomers, while others gave the maximum surface modification after some delay. Factors governing the migration extent and the migration rate are discussed. Keywords: UV curing; acrylic oligomers; fluorinated co-monomers; surface modification; wettability.
1. INTRODUCTION
The tailoring of surface properties of polymeric materials is extremely important, because many applications require specific surface properties. One can recall, among others, the wettability towards liquids, the resistance to chemicals, the durability, optical characteristics, abrasion resistance, adhesion and friction [1]. Also, it is required to modify the polymer surface, without tempering with the bulk properties. In this context, UV-curable systems have been shown to be very versatile: significant surface modifications are obtained by using only small amounts of surface modifiers without changing the bulk properties [2]. *
To whom correspondence should be addressed. Tel.: (39-11) 564-4656; Fax: (39-11) 564-4699; e-mail:
[email protected]
286
R. Bongiovanni and A. Priola
The photopolymerisation process (UV curing) is a technique offering unique advantages, such as fast curing rate, use of solvent-free formulations, low energy consumption and use of simple and efficient equipments. For these reasons, it is having a wide success in different fields, such as coatings, inks, adhesives, dental materials, microoptics and microelectronics [3]. The UV curing can be performed either by a radical or a cationic mechanism. In this work, radical systems containing acrylic or methacrylic double bonds were used. As surface modifiers, fluorinated co-monomers are desirable due to their characteristics, connected with the presence of fluorine. Fluorine guarantees unique chemical properties, chemical and thermal stability, weathering resistance, low surface tension, hydrophobicity and oleophobicity, and interesting optical and electrical performances. Many papers have reported on the outstanding properties of fluorinated polymers and their applications, mainly in the fields of coatings, mechanical devices, optical fibers, and electrical and electronic components [4]. The properties of these polymers depend both on the fluorine content and the structure of the polymeric matrix, as well as on the distribution of the fluorine atoms along the polymeric chains: all these factors play important roles in the final performance of the material. We have focussed our attention on the structure of the surface modifiers and have investigated the influence of their fluorine content and of the fluorine distribution on the properties of the cured materials. A wide range of polymer networks were prepared by co-polymerising different types of fluorinated co-monomers under UV irradiation with a typical hydrogenated resin (namely bisphenol-Adiethylether-diacrylate). Among the fluorinated co-monomers, many were synthesised for a special purpose. The aim was to clarify the factors determining the properties of these surface modifiers in order to design molecular architectures and tailor them for specific surface properties. We have considered surface properties such as surface tension, wettability, hydro- and oleo-repellency, and adhesion. In some cases, abrasion and frictional properties were also evaluated. With all the co-monomers investigated the surface modification was achieved through a selective segregation of low surface energy components: this segregation, which is well documented for polymer blends, polymer solutions and copolymers (mainly thermoplastic) [5–10], takes place before curing and is retained by means of the curing reaction in our systems. Factors governing the segregation phenomenon are discussed. 2. EXPERIMENTAL
Different fluorinated co-monomers were employed in this work and are grouped in three classes as follows: Series A, i.e., acrylates of fluorinated alcohols having the following general structure F–(CF2)m–(CH2)nO–CO–CH=CH2, with m = 4, 8 or 10 when n = 2 or n = 2, 3 or 11 when m = 8; n = 2, 3 and 11 corresponds to
Surface properties of UV-curable acrylic systems
287
ethyl (Et), propyl (Pr) or undecyl (Un) chains, respectively. Products with n = 2 were obtained from Daikin (Japan) and used without further purification; the others were synthesised [11, 12]. Series Z, i.e., urethane(meth)acrylates of perfluoropolyether of general formula CF3O(CF2CF2O)m(CF2O)nCF2CH2O–, viz., CF3O-(CF2CF2O)m-(CF2O)n-CF2CH2O-UMA (m ≈ n, MW = 1079; Z-3) and CF3O-(CF2CF2O)m-(CF2O)n-CF2CH2O-UMA (m ≈ n, MW = 1463; Z-4), where UMA stands for –CONHCH2CH2OCOC(CH3)=CH2. Series Y, i.e., urethane(meth)acrylates of perfluoropolypropyleneoxide XC3F6O-[CF2CF(CF3)O]n-1CF2CH2O-UMA, where XC3F6O = XCF2CF(CF3)O– or CF3CFX-CF2O–, X = Cl or H, viz., XC3F6O-[CF2CF(CF3)O]1CF2CH2O-UMA and XC3F6O-[CF2CF(CF3)O]3CF2CH2O-UMA. The macromers containing a chlorine atom (X = Cl) in the end group are labelled as Cl-n, where n is the number of perfluoropropyleneoxide units (Cl-2, n = 2; Cl-4, n = 4). The homologues containing a hydrogen in the end group (X = H) are labelled as H-n (where H-2 means n = 2; H-4, n = 4). This series also includes CF3O-[CF2CF(CF3)O]3CF2CH2O-UMA, which is labelled as F-3. The products of series Z and Y were kindly supplied by Solvay Solexis (Milan, Italy); their synthesis is reported elsewhere [13–16]. The hydrogenated diacrylate, bisphenol-A-diethyletherdiacrylate (BHEDA, trade name Ebecryl 150TM), having the structure
was supplied by UCB (Belgium) and was used as received. The concentration of the fluorinated co-monomers ranged from 0.1 to 1% (w/w), which corresponds to their maximum solubility in the hydrogenated diacrylate. 2-Hydroxy-2-methyl-1-phenylpropanone was added as the photoinitiator (4%, w/w). It was obtained from Ciba Specialty Chemicals, Switzerland (trade name Irgacure 1173). The polymers were obtained by spreading the co-monomer–BHEDA mixtures on a glass substrate with a calibrated wire-wound applicator. No solvent was used. Before curing, the systems were kept under an inert atmosphere for 15 min, unless differently indicated. The mixtures were cured by UV irradiation in an inert atmosphere, according to the procedure reported elsewhere [17]; the experimental conditions assured a complete crosslinking of the polymers. The complete conversion was monitored by FT-IR analysis by checking for the disappearance of the acrylic double bond band at ν=1650 cm-1. The polymers were obtained in the form of films with a thickness of about 100 µm.
288
R. Bongiovanni and A. Priola
Dynamic contact angle measurements were performed on all films, with a Kruss G10 instrument, equipped with a videocamera and an image analyser, at room temperature with the sessile drop technique (advancing rate 0.3 µl/s ). The films were peeled off the glass substrates and analysed on both sides: the side in contact with the air was labelled as the air side and the opposite one as the glass side. On every surface examined at least five contact angles were measured; the differences from the average value were no more than 2° for the advancing angle and 3° for the receding angle. The measuring liquids were doubly distilled water and hexadecane (γ = 72.1 and 28.1 mNm-1, respectively). The polymers were completely insoluble in these liquids. XPS measurements were carried out on films containing fluorinated additives on a VG Instruments spectrometer using a Mg Kα,1,2 X-ray source (1253.6 eV). The X-ray source conditions were 100 W, 10 kV and 10 mA. The base pressure in the spectrometer was 5¥10-10 Torr and the operating pressure was 2¥10-8 Torr. Pass energies of 100 eV and 50 eV were used for wide scans and narrow scans, respectively. Depth profile information was obtained from measurements at takeoff angles of 10° and 45°. According to the equation d=3λsinθ, the sample depth (d) was lower when the take-off angle (toa) was smaller. The value of λ, i.e., the inelastic mean free path length, for carbon is 1.4 nm. All data analyses (linear background subtraction and peak integration) were carried out using VGX900x (version 6) software. Binding energies were referenced to the C–H level at 285.0 eV. 3. RESULTS AND DISCUSSION
All the fluorinated co-monomers employed possess a low solubility in the acrylic resin BHEDA, even when their fluorine content is small; therefore, the highest concentration of the co-monomers used ranged from 0.5 to 1% (w/w). As a consequence, the co-polymers retain the same bulk properties as the pure hydrogenated homo-polymer based on BHEDA, as shown by thermal and mechanical analyses [17]. Data obtained from FT-IR analyses, by monitoring the double bond conversion, have also shown that the introduction of the co-monomers in such a low amount does not affect the kinetics of polymerisation and the final conversion is always quantitative [17]. In contrast, for every system examined the surface properties have been found to be considerably different. 3.1. Surface modification by fluoroalkylacrylate co-monomers (series A) The first class of co-monomers investigated are the monofunctional perfluorinated acrylates, F–(CF2)m–(CH2)nO–CO–CH=CH2 (series A). The water contact-angle data show that these additives modify the air side (i.e., the side in contact with the air) of the films, while the glass side (i.e., the side in contact with the glass
Surface properties of UV-curable acrylic systems
289
Figure 1. Dependence of water contact angle on the concentration and on the fluorinated chain length of the surface modifiers of the series A (air side, unless indicated otherwise).
substrate) possesses the same properties as the pure resin. In Fig. 1 the wettability curve as a function of the F–(CF2)m–(CH2)nO–CO–CH=CH2 concentration (n=2, m=4, 8 or 10) is reported. It is clear that the films become highly hydrophobic only on their air side, while on the glass side the measured contact angle (66°) is independent of the type of formulation. As expected, the longer the CnF2n+1 chain, the higher the hydrophobicity. In the case of the highest-MW co-monomer, the maximum contact angle is typical of a highly fluorinated surface, but is reached at a very low bulk concentration of the fluorinated co-monomer (118° at 0.4%). The fluorine content on the air side of the films, determined through XPS analyses [11], is an order of magnitude higher than the concentration calculated on the basis of the composition, thus suggesting a selective migration of the fluorinated chains to the outermost layers of the films. Therefore, the surface of the co-polymeric network can be considered as mainly composed of the fluorinated co-monomer. Increasing the fluorine content is not the only parameter to control the surface properties. The length of the spacer groups, i.e., the number of methylene units n in the series A co-monomers, is also important. In Fig. 2 the effect of the different spacer chains on the wettability is reported. It shows that by increasing the length of the hydrogenated group for co-monomers having the same C8F17 tail, a clear increase in surface hydrophobicity is obtained. The effect of the spacer group is clearly evident: the amount of the co-monomer required to make the surface hydrophobic is extremely low for the co-monomer with n=11 (undecyl group as a spacer), while the surface activity with n=2 (ethyl group as a spacer) is lower. The additive containing a propyl spacer (n=3) shows an intermediate behaviour.
290
R. Bongiovanni and A. Priola
Figure 2. Dependence of water contact angle on the concentration and on the spacer length of the surface modifiers of the series A having a fluoro-octyl chain (air side, unless indicated otherwise).
These results are in agreement with those of Pospiech and co-workers [18],who showed that semifluorinated alkyl groups linked as side chains to polyester backbone could self-organize at the polymer surface when a critical MW was reached. Similar trends have been found with co-monomers containing complex spacer groups, such as sulphonamidic units or thioether units. For these films the surface segregation has been studied by Rutherford Back Scattering [19]. The surface modification reported above has always been found to be permanent, indicating that an effective co-polymerisation of the fluorinated additives takes place. In fact selective extractions, to which the films were subjected, have shown that the concentration of the free fluorinated co-monomer is negligible. Moreover, the surface properties did not change by exposing the polymers to different environments. 3.2. Surface modification by perfluoropolyether co-monomers (series Z and Y) Examining the additives belonging to the perfluoropolyether-urethane methacrylate group (series Z), a strong surface activity of the perfluoropolyether (PFPE) chain is evidenced. In fact, besides the higher fluorine content, PFPE chains have peculiar properties: they are very flexible (with Tg as low as – 111°C when MW is 2000), highly apolar, hydro- and oleophobic without any crystallization tendency [20]. As an example the wettability of films containing the Z-3 urethanemethacrylate is reported in Fig. 3. The maximum value of the advancing contact angle is around 121° and, more interestingly, it is obtained at a co-monomer concentration lower than 0.3% (w/w). The selectivity of the surface modification towards the air side vis-a-vis to the glass side is evident, as for the series A comonomers.
Surface properties of UV-curable acrylic systems
291
Figure 3. Effect of the Z-3 co-monomer concentration on the water contact angle.
Table 1. Surface composition from XPS analyses (toa=take-off angle) F1s/C1s atomic ratio - air side of Z-3 co-polymer (0.8% (w/w) Z-3) Calculated value 0.012 Experimental value at toa=45° 1.11 Experimental value at toa=10° 1.70 F1s/C1s atomic ratio - air side of Z-3 homopolymer Calculated value 1.43 Experimental value at toa=45° 1.37 Experimental value at toa=10° 1.78
XPS data confirm the surface enrichment (air side) by the fluorinated comonomer. In Table 1 the F/C atomic ratios evidence that at the very surface of the film air side (toa=10°) the fluorine content is two orders of magnitude higher than the calculated value. When the depth examined is greater (toa 45°), the fluorine amount decreases, indicating the existence of a concentration gradient from the outermost surface towards the bulk. It is interesting to compare these data with those obtained by analysing the Z-3 homo-polymer prepared by photopolymerisation: the outermost co-polymeric film has a fluorine content nearly equal to that of the homopolymer. Thus, one can conclude that it is mainly composed of fluorinated co-monomer. For the other products of the series Z the behaviour is similar; the data concerning the maximum wettability and the critical concentration at which it is shown are collected in Table 2. With respect to the series Y, the wettability data on the air side of the films are reported in Table 3 as a function of the concentration.
R. Bongiovanni and A. Priola
292
Table 2. Wettability of copolymeric systems containing series Z co-monomers Co-monomer
Concentration Advancing contact angle, (%, w/w) air side (∞)
Advancing contact angle, glass side (∞)
Z-3 Z-4
0.25 0.10
66 66
121 125
Table 3. Effect of the series Y co-monomer concentration on the water contact angle (air side) Co-monomer None Cl-2 H-2 Cl-4 H-4 F-3
Concentration (%, w/w) 0 0.20 1.0 0.10 1.0 0.20 1.0 0.20 1.0 0.8
Water contact angle (°) 77 87 96 85 96 94 99 96 101 118
The presence of the fluorinated co-monomers, although introduced in low amounts, i.e., no more than 1% (w/w), drastically and irreversibly modifies the property of the air side surface of the UV-cured films. It is evident that in the presence of any of the fluorinated co-monomers the wettability of the films is strongly reduced. First of all, the hydrophobicity depends on the co-monomer concentration, i.e., on the fluorine content, and the contact angles depend also on the length of the perfluoropolyether chain of the co-monomers. Considering products of similar MW, e.g., Cl-4 and H-4, introduced in the acrylic resin at a 1% concentration, it is clear that the end group plays a role in the surface modification; in fact the monofunctional monomer F-3, containing a perfluoromethyl end-group, is more surface active than the other ones. The contact angle of the CF3-terminated co-monomer, used at 0.8% concentration, is 118°, i.e. much higher than the values for the H- and Cl-homologues, even when their MW is higher. The perfluoromethyl end group guarantees a lower wettability than the hydrogen containing end group, which, in turn, is always slightly higher than the chlorine containing group (however, the difference is within experimental error). The contact angle of hexadecane on the different copolymers was also investigated. This parameter, describing the oleophobicity of the material, controls the
Surface properties of UV-curable acrylic systems
293
Figure 4. Hexadecane contact angle on films containing Cl-4 (air side).
Figure 5. Advancing water contact angle on copolymers made from H-2 co-monomer as a function of time between dispensing the mixtures on the substrate and initiating the UV irradiation.
fundamental characteristics of coatings, (e.g., the stain release and the antigraffiti resistance). In our case very high values of contact angle were observed, meaning that the surfaces were oleophobic. The influence of the fluorinated co-monomers on the oleophobicity is shown in Fig. 4 for co-monomer Cl-4. Similar trends were also found for all the other copolymers. As widely discussed [5–10, 21, 22], the mechanism leading to surface modification of polymers in the presence of fluoro-co-monomers involves migration towards the air interface of the fluorinated co-monomers due to their lower surface tension. This diffusion phenomenon leads to surface enrichment by the fluorinated co-monomers. As the networks under characterisation are always fully cured and no change in surface properties was observed with aging of the polymeric films, the migration should happen in the liquid state. After irradiation, the curing by UV light immobilizes the system. Therefore, we thought it was interesting to carry out further investigations on what happens before curing. Different films containing the
R. Bongiovanni and A. Priola
294
Table 4. Time between dispensing the mixture and irradiation for the maximum surface modification (1% (w/w) co-monomer concentration) Co-monomer
Time (min)
Z-3
<1
Z-4
<1
Cl-2
40
Cl-4
<1
H-2
20
H-4
<1
F-3
<1
same concentration of the different fluorinated co-monomers (1%, w/w) were prepared: the time between dispensing the mixtures on the substrate and irradiating them was varied from a few seconds to hours. Contact angle measurements were performed on the air side of the films after complete curing. The results are plotted for the H-2 co-monomer in Fig. 5. It is evident that increasing the time between dispensing the mixture and irradiating, co-monomers can show different performances in the surface modification of the final co-polymers. The results can be rationalised in terms of time necessary to reach the maximum surface modification in the presence of 1% of the comonomer. The data are reported in Table 4 and compared with the F-3 comonomer. A correlation between the molecular weight of the co-monomers of both series Z and Y and the rate of surface modification exists. Surprisingly the higher the molecular weight, the higher the rate of migration. In fact, one would expect that smaller molecules can diffuse through the bulk towards the surface more easily than the larger ones. Therefore, other parameters should also be taken into account. First of all, the migration rate of the co-monomer could be governed by the differences in the solubility parameters between the different co-monomers. This difference increases with the molecular weight, due to the increase of fluorine content. Therefore, the smaller monomers, having higher affinity for the hydrogenated bulk, require more time to migrate to the surface. 4. CONCLUSIONS
The results obtained by investigating the fluorinated co-monomers of series A, Z and Y as additives in UV-cured systems show a high tendency of these products for surface segregation. The factors influencing their surface activity are related to the length of the fluorinated chain and to the length of the spacer group, linking
Surface properties of UV-curable acrylic systems
295
the fluorinated tail and the reactive functionality. The groups present at the end of the fluorinated chains also influence the surface properties of the co-monomers: the perfluoromethyl end group guarantees a lower wettability than the hydrogen containing end group, which, in turn, is always slightly higher than the chlorine containing group. This is in agreement with the findings of Zisman and coworkers in their early work on fluorocarbons [23]. In conclusion, the surface activity of the fluorinated co-monomers can be interpreted as a result of the selective migration of the co-monomers towards the air surface before the systems are cured. The migration was instantaneous in some cases, while in other cases it required some time. It was found that the rate of migration increased by increasing the length of the fluorinated chains; thus, it is inversely related to the co-monomer molecular weight. Acknowledgements Thanks are due to Dr. B. Ameduri (CNRS, Montpellier, France) and Dr. C. Tonelli (Solvay-Solexis, Milano, Italy) for supplying some of the co-monomers investigated. Prof. A. Pollicino (University of Catania, Italy) is acknowledged for the XPS analyses. REFERENCES 1. F. Garbassi, M. Morra and E. Occhiello, Polymer Surfaces: from Physics to Technology, Wiley, New York, NY (1994). 2. R. Bongiovanni, G. Malucelli, M. Sangermano and A. Priola, J. Fluorine Chem. 125, 345 (2004). 3. P. Fouassier and J. F. Rabek, Radiation Curing in Polymer Science and Technology. Vol. I. Elsevier, London (1993). 4. B. Ameduri and B. Boutevin, Well-architectured Fluoropolymers: Synthesis, Properties and Applications. Elsevier, London (2004). 5. D. H. Pan and W. M. Prest, J. Appl. Phys. 58, 2861 (1985). 6. Q. S. Bhatia, D. H. Pan and J. T. Koberstein, Macromolecules 21, 2166 (1988). 7. R. L. Schmitt, J. A. Gardella, J. H. Magill and L. Salvati, Macromolecules 18, 2675 (1985). 8. J. Höpken and M. Möller, Macromolecules 25, 1461 (1992). 9. R. Bongiovanni, G. Malucelli, V. Lombardi, V. Siracusa, C. Tonelli and A. DiMeo, Polymer 42, 2299 (2001). 10. M. Toselli, M. Messori, R. Bongiovanni, G. Malucelli, A. Priola, F. Pilati and C. Tonelli, Polymer 42, 1771 (2001). 11. B. Ameduri, R. Bongiovanni, G. Malucelli, N. Pollicino and A. Priola, J. Polym. Sci. Polym. Chem. 37, 77 (1999). 12. B. Ameduri, R. Bongiovanni, V. Lombardi, N. Pollicino, A. Priola and A. Recca, J. Polym. Sci. Polym. Chem. 39, 4227 (2001). 13. A. Priola, R. Bongiovanni, G. Malucelli, A. Pollicino, C. Tonelli and G. Simeone, Macromol. Chem. Phys. 198, 1893 (1997). 14. C. Tonelli, A. DiMeo, S. Fontana and A. Russo, J. Fluorine Chem. 118, 107 (2002). 15. C. Tonelli, P. Gavezotti and E. Strepparola, J. Fluorine Chem. 95, 51 (1999). 16. M. Malavasi and D. Sianesi, J. Fluorine Chem. 95, 19 (1999).
296
R. Bongiovanni and A. Priola
17. R. Bongiovanni, G. Malucelli, A. Priola, C. Tonelli, G. Simeone and A. Pollicino, Macromol. Chem. Phys. 199, 1099 (1998). 18. D. Pospiech, D. Jehnichen, L. Häußler, D. Voigt, K. Grundke, C. K. Ober, H. Körner, and J. Wang, Polym. Prepr. 39, 882 (1998). 19. M. G. D. Van Der Grinten, A. S. Clough, T. E. Shearmur, R. Bongiovanni and A. Priola, J. Colloid Interface Sci. 182, 512 (1996). 20. D. Sianesi, G. Marchionni and R. J. De Pasquale, in: Organofluorine Chemistry: Principles and Commercial Applications, R. E. Banks, B. E. Smart and J. C. Tatlow (Eds.), p. 431. Plenum Press, New York, NY (1994). 21. S. Wu, Polymer Interface and Adhesion, Marcel Dekker, New York, NY (1982). 22. R. Mason, C. A. Jalbert, M. O'Rourke, P. A. V. Muisener, J. T. Koberstein, J. F. Elman, T. E. Long and B. Z. Gunesin, Adv. Colloid Interface Sci. 94, 1 (2001). 23. F. Hare, E. G. Shafrin and W. A. Zisman J. Phys. Chem. 58, 236 (1954).
Polymer Surface Modification: Relevance to Adhesion, Vol. 4, pp. 297–305 Ed. K.L. Mittal © VSP 2007
Detection of contaminants on polymer surfaces using laser-induced breakdown spectroscopy (LIBS) S. MARKUS,∗ U. MEYER, R. WILKEN, S. DIECKHOFF and O.-D. HENNEMANN Fraunhofer Institute for Manufacturing and Applied Materials Research, Wiener Straße 12, D-28359 Bremen, Germany
Abstract—For quality assurance of adhesive bonds, the use of laser-induced breakdown spectroscopy (LIBS) as a technique for the detection of contaminants was evaluated. Polycarbonate (PC) surfaces were coated with a release agent (silicone oil) forming thin contaminant films. The corresponding film thicknesses in a range of 5–750 nm were analysed by variable angle spectroscopic ellipsometry (VASE). The contaminated surfaces were characterized by LIBS and X-ray photoelectron spectroscopy (XPS). The LIBS signals were correlated with layer thicknesses and were compared with the results obtained from XPS which is an established method for surface characterization. The effect of laser plasma on the surface morphology was studied by profilometry. It was shown that LIBS was applicable for detecting contaminants of silicone based release agents on polymer surfaces that can cause detrimental effects on adhesion properties. Keywords: LIBS; polymer surface; release agent; silicone; adhesion.
1. INTRODUCTION
As regards bonding and coating processes a maximum degree of surface cleanliness is an important requirement for high adhesion strengths. Monolayers or even sub-monolayers of contaminants, like release agents with corresponding film thicknesses of less than 1 nm, can cause adhesion failures. Therefore, the detection of small amounts of surface contaminants is a requirement for quality assurance in bonding technology [1–3]. In particular in cases of similar chemical compositions of substrates and contaminants (e.g., organic contaminants on polymer surfaces) the detection of these species is hampered. In the field of adhesion research surface analysis techniques such as XPS and TOF-SIMS are well established for both qualitative and quantitative analyses of the outermost surfaces of engineering materials [4]. These methods are characterized by high information content and high accuracy of measurements. However, ∗
To whom correspondence should be addressed. Tel.: (49-421) 2246-612; Fax: (49-421) 2246-77612; e-mail:
[email protected]
298
S. Markus et al.
Table 1. Advantages and drawbacks of LIBS as a tool for in-line analysis [11] Advantages
Drawbacks
LIBS can be applied to solids, liquids and gases
Small spot analysis Some destruction of surface layer
No sample preparation Ambient air and atmospheric pressure
Difficulty in obtaining suitable standards (semi-quantitative)
Portable systems
Safety consideration regarding laser radiation
Automated measurements are feasible Real time analysis Remote analysis Sensitivity in ppm range Simultaneous multi-element analysis Small mass required (0.1 µg–0.1 mg [11]) High spatial resolution (1–100 µm [11])
these techniques are only possible with high operating expense (high manpower requirement and long measurement times of several hours). In addition, none of these approaches is non-destructive because only small sized samples can be analysed; also, none of these approaches can be integrated in automated production lines. Laser-induced breakdown spectroscopy (LIBS) is a method that is frequently used to analyse the elemental compositions of solids, liquids, and gases. A high power laser pulse is focussed onto the sample whereby a small amount of material (0.1 µg to 1 mg) is evaporated and forms a plasma above the surface. The optical emissions from this plasma are analysed by a spectrometer. The atomic spectral lines are used to determine the elemental composition as well as the elemental concentration of the sample [5–10]. In comparison to conventional methods of surface analysis, LIBS requires relatively short measurement times (only a few seconds). In addition, LIBS is available in portable setups that allow measurements on large sized samples and it can even be adapted to in-line applications. The advantages and drawbacks of LIBS as a tool for in-line analysis are listed in Table 1 [11]. In this study the suitability of LIBS as a method to detect contaminants with relevance to adhesion was investigated. The sensitivity was studied by carrying out LIBS measurements on thin films of silicone oil on polycarbonate surfaces. The results are compared to XPS analysis.
Detection of contaminants on polymer surfaces using LIBS
299
2. EXPERIMENTAL
2.1. Sample preparation Thin films of silicone oil with various thicknesses in a range from 5 to 750 nm were applied onto polycarbonate foils (100 µm thick) by an automated spreading knife. To obtain thin films of silicone contamination poly(dimethylsiloxane) (PDMS, DMS-T23E, ABCR-Gelest, Karlsruhe, Germany) was diluted in hexamethyldisiloxane (HMDSO). For profilometry measurements a thin gold layer was deposited onto the polycarbonate surface by sputtering to ensure high reflectivity. 2.2. Measurements The film thicknesses were analysed by a variable angle spectroscopic ellipsometer (VASE; J. A. Woollam, Lincoln, NE, USA). The ellipsometric angles Ψ and ∆ were acquired over a spectral range from 250 to 900 nm (in steps of 25 nm) at angles of incidence of 65 and 75°. The n and k values for PC were taken from the database of the ellispometer. The n and k values for silicone were calculated from ellipsometry measurements on well-known substrates. XPS measurements were carried out with a Kratos Ultra System (Kratos Analytical, Manchester, UK) operated at a base pressure of 10-8 Pa and using a monochromatic Al Kα X-ray source. Spectra were acquired in the constant energy analyser mode with a pass energy of 160 eV for survey spectra. The photoemission angle of photoelectrons was 0°. The analysis area had an elliptic shape with axis lengths of 0.3 and 0.7 mm. The samples were neutralized with 2.6 eV electrons. The sensitivity factors for Si2p, C1s and O1s were calculated by comparisons between the known silicone oil composition and the measured composition values. LIPAN 3002 equipment (LLA Instruments, Berlin, Germany) was used for LIBS measurements. In the setup of LIPAN 3002 a Q-switched Nd:YAG-pulsed laser with an unfocussed laser beam diameter of 3 mm, a pulse width of 6 ns and a repetition rate of 20 Hz was used for excitation. The measurements were carried out with pulse energies from 250 to 260 mJ at 1064 nm. An Echelle spectrometer (LLA Instruments) enabled simultaneous detection of wavelengths from 200 to 780 nm with a linear dispersion of pm resolution. The spectrometer was combined with an ICCD camera (1024¥1024 pixels). The image intensifier was gated with a measurement delay of 2 µs. The spectra of plasmas that were induced on surfaces after a single laser shot were analysed by the spectrometer to ensure a maximum degree of surface sensitivity. Thirty measurements were taken to calculate the result for each level of silicone contamination. The physical surface modification was monitored with a UBM Mikrofocus Profilometer (UBM, Sunnyvale, CA, USA). The measurements were performed with a resolution of 500 pixel/mm and a velocity of 0.1 mm/s.
300
S. Markus et al.
3. RESULTS AND DISCUSSION
The contaminated samples were analysed by XPS. The surface concentration in atom percent represents the elemental surface composition with an information depth of approx. 10 nm. The elemental concentrations in PDMS are 50 at% carbon, 25 at% oxygen and 25 at% silicon. A film thickness of more than approx. 10 nm correlates with an elemental concentration of 25 at% measured by XPS (Fig. 1.). The curve shape of the correlation between the film thickness of the
Figure 1. Silicon concentration determined using XPS versus the thickness of silicone oil film measured by ellipsometry.
Figure 2. Surface morphology of PC film after a single laser shot measured by profilometry.
Detection of contaminants on polymer surfaces using LIBS
301
contaminant and the silicon concentration measured by XPS is typical for a method with high surface sensitivity. After a single laser pulse during LIBS analysis a crater with a depth of 50 µm was measured on the PC surface by profilometry. In addition, elevations of the surface were observed at the edges of the crater (Figs 2 and 3). It is assumed that the craters are caused by thermal evaporation of the material initiated by high energy density at the laser spot. The elevations might originate from deposition of evaporated material or thermal expansion of the PC film. LIBS spectra for wavelengths from 200 to 780 nm are shown in Figs 4 and 5. The main differences between spectra of plasma emissions from PC reference and PC film contaminated by silicone oil (80 nm) are found between wavelengths of 200 and 350 nm. In this range of wavelength most of the spectral lines originate from the optical emission
Figure 3. Profile of surface morphology of PC film after a single laser shot measured by profilometry.
Figure 4. LIBS spectrum: intensity of plasma emission for PC reference versus wavelength.
302
S. Markus et al.
Figure 5. LIBS spectrum: intensity of plasma emission for PC film contaminated by silicone oil (80 nm) versus wavelength.
Figure 6. LIBS spectrum: intensity of plasma emission from PC reference versus wavelength.
caused by relaxation processes of atoms and ions. These lines are clearly separated from each other and are useful for qualitative and semi-quantitative evaluations. The range of longer wavelengths, from 350 to 780 nm, contains mainly molecular bands which interfere with each other and are not useful for evaluation of the silicone contaminant. For analysing the influence of an increase in surface contamination, it is useful to compare the ratios of characteristic spectral lines. The PC as a reference is characterized by carbon lines. The strongest carbon line appears at 247.8 nm (Fig. 6). The spectrum of silicone contaminated sample shows additional silicon lines from 250 to 255 nm and at 288.1 nm (Fig. 7). The ratios of intensities of the strongest silicon line (at 288.1 nm) and carbon line
Detection of contaminants on polymer surfaces using LIBS
303
Figure 7. LIBS spectrum: intensity of plasma emission from PC film contaminated by silicone oil (80 nm) versus wavelength.
Figure 8. Ratio of intensity of the strongest silicon line (at 288.1 nm) and carbon line (247.8 nm) versus the thickness of silicone film measured by ellipsometry.
(at 247.8 nm) were reported as the degree of contamination. Figure 8 shows the ratios of intensities of silicon lines (at 288.1 nm) and carbon lines (at 247.8 nm) depending on the thickness of silicone oil film up to 750 nm. LIBS has been established as a method for bulk analysis. Therefore, a linear correlation between the ratios of LIBS intensities and thicknesses of silicone films would have been expected. Up to a film thickness of approx. 100 nm of silicone oil the ratio of intensities of the considered carbon and silicon lines increased significantly. For film thicknesses of more than 100 nm the ratio of intensities was
304
S. Markus et al.
almost constant (Fig. 8). This shows that LIBS, which creates a crater 50 µm deep in polycarbonate (Figs 2 and 3), appears to give information only on the uppermost 100 nm of the sample surface (Fig. 8). The formation of the plasma might be caused by excitation of the first 100 nm of the sample surface only. In this case material from the uppermost 100 nm of the surface is involved in plasma formation by vaporization and ionization processes and, thus, it is this material that causes the spectra of optical emission. At the end of the plasma formation the dense plasma core is expanding. The fast expansion of the plasma leads to a shock wave that is responsible for driving away the residual molten layer of the surface [7]. The effect of ablation of material has to be distinguished from plasma formation. The effect of measurement parameters on the morphological surface modification and the information depth of LIBS will be studied further. Comparing XPS and LIBS measurements, XPS is more surface sensitive. The information depth of XPS measurements for this test series is about 20 nm (Fig. 1) whilst the information depth of LIBS measurements is about 100 nm (Fig. 8). The standard deviation of the measured data is minor for XPS measurements and the technique is more accurate than LIBS. 4. CONCLUSIONS AND OUTLOOK
LIBS turned out to be suitable for detecting contaminants of silicone based release agents on PC surfaces that can affect adhesion properties adversely. Depending on the measurement parameters, LIBS can be used for surface-sensitive measurements. With the contaminant/surface combination of silicone oil and PC the interpretation of data was simple but for the combination of contaminant layers where the contaminant has a similar chemical composition to the substrate the interpretation of data has to be adjusted to account for more subtle differences. Currently, investigations are being carried out to determine the influence of the morphological modification of polymer surfaces by LIBS on the adhesive strengths and the quality of coatings. In case of adhesive bonds it can be assumed that the influence can be neglected. In applications where these polymer surfaces are coated subsequent to LIBS measurements, a negative effect on the optical properties might occur. For quality control applications, surface contaminants can be observed by monitoring ratios of signal intensities. Current investigations are being carried out to adapt LIBS as an in-line control in different fields involving polymer surfaces. REFERENCES 1. A. Heßland and O.-D. Hennemann, Adhäsion Kleben Dichten 38, 10-15 (1994). 2. G. D. Davis, Surface Interface Anal. 20, 368-372 (1993). 3. B. M. Parker and R. M. Waghorne, Surface Interface Anal. 17, 471-476 (1991).
Detection of contaminants on polymer surfaces using LIBS
305
4. W. Possart (Ed.), Adhesion, Current Research and Applications. Wiley-VCH, Weinheim (2005). 5. K. Laqua, in: Chemical Analysis, J. D. Winefordner (Ed.), p. 47. Wiley, New York, NY (1997). 6. G. M. Weyl, in: Laser Induced Plasmas and Applications, L. J. Radziemski and D. A. Cremers (Eds). Marcel Dekker, New York, NY (1989). 7. Y. W. Kim, in: Laser Induced Plasmas and Applications, L. J. Radziemski and D. A. Cremers (Eds). Marcel Dekker, New York, NY (1989). 8. L. J. Radziemski and D. A. Cremers, in: Laser Induced Plasmas and Applications, L. J. Radziemski and D. A. Cremers (Eds). Marcel Dekker, New York, NY (1989). 9. L. Moenke-Blankenburg, in: Chemical Analysis, J. D. Winefordner and I. M. Kolthoff (Eds). Wiley, New York, NY (1989). 10. N. Omenetto, in: Chemical Analysis, P. J. Elving and J. D. Winefordner (Eds). Wiley, New York, NY (1979). 11. Y. Lee and J. Sneddon, ISIJ Int. 42 (Suppl.), S129-S136 (2002).