J O U R N A L O F C H E M I S T R Y Materials Correlation of friction, adhesion, wettability and surface chemistry after argon plasma treatment of poly(ethylene terephthalate) Ben D. Beake,a John S. G. Lingb and Graham J. Leggett*a aDepartment of Chemistry, The University of Manchester Institute of Science and Technology, PO Box 88, Manchester, UK M60 1QD bBritish Steel, Welsh Technology Centre, Port Talbot, West Glamorgan, UK SA13 2NG Received 17th September 1998, Accepted 12th October 1998 The combination of wettability, X-ray photoelectron spectroscopy and scanning force microscopy has been used to analyse the changes to the surface after plasma treatment of poly(ethylene terephthalate) film. Calculations on contact angle data with a combination of polar and non-polar liquids have shown that argon plasma treatment considerably enhances the work of solid–(polar) liquid adhesion and the surface free energy of the films due to the creation of acidic and basic functions on the polymer surface.In contrast, Lifshitz–van derWaals (apolar) interactions decrease slightly as a consequence of plasma-induced chain-scission. We present the first study of a plasma-treated polymer by chemical force microscopy.Plasma-modified surfaces exhibit substantially higher friction than untreated material and are more easily disrupted by the movement of the tip during scanning. Friction is reduced when methyl-functionalised tips are employed. There is a correlation, on plasma treatment, between the rapid increases in surface friction probed by lateral force microscopy and surface free energy probed by wettability and X-ray photoelectron spectroscopy.The modified mechanical properties and polar group incorporation both result from scission of polymer chains and contribute to the lateral force contrast. terephthalate) [PET] and the thermodynamic work of adhesion Introduction between the surface and polar liquids, and have determined The control of the chemical, mechanical and topographical the eVect of diVerent plasma treatment conditions on these properties of surfaces is relevant in numerous applications of quantities.X-Ray photoelectron spectroscopy has been used polymers in the textiles, adhesives, composites and coatings to investigate the compositional changes in the surface/nearindustries. 1–6 Many of these applications require good surface region that are responsible for the improvement in adhesion between the polymer and a surface coating. Plasma wettability after exposure to argon plasma. The level of oxygen treatment is an eVective method for improving the bondability incorporation with increasing time of argon plasma treatment and wettability of polymer surfaces whilst leaving bulk proper- provides a means to monitor the extent of surface functies unaltered.In a plasma the surface is exposed to a broad tionalisation. spectrum of ions, electrons, excited neutrals, radicals, UV and Recently, scanning force microscopy has been used to VUV radiation.2,7 The predominant reactive species in an examine the changes to surface topography which occur on inductively coupled radio frequency argon plasma are thought plasma treatment of polymers such as polytetrafluoroto be argon ions and VUV photons which produce excited ethylene,12 polypropylene,13,14 polymethylmethacrylate15 and states at the surface, the decay of which leads to the formation poly(ethylene terephthalate) (PET).8 In addition to the topoof radicals.2,6,7 graphical information, by operating the scanning force micro- We have previously investigated the relationship between scope in lateral (or frictional ) force mode the eVects of the changing wettability and surface morphology under diVerent plasma modification on the nanoscale surface properties of plasma conditions.8 Completely hydrophilic surfaces were not the polymer can be investigated.In lateral force microscopy obtained even after several hours treatment. Wettability is (LFM), the torsional or twisting motions of the cantilever are thought to be limited by competition between etching of the recorded with high contrast, and are regarded as being indicasurface (via chain scission) and chemical functionalisation tive of frictional interactions between tip and sample.16 The (incorporation of polar groups).8–10 Scanning force diVerence in the lateral force signal between forward and microscopy (SFM), imaging in an intermittent contact mode, reverse scans is proportional to the friction force during where the tip taps the surface, showed that the formation of imaging.16,17 This friction is thought to correlate with adhesion orientated, ridged surface structures occurred over an extended since on the molecular scale both processes involve bond time scale.8 breaking and formation.18 In order to investigate the eVect of In this paper we aim to analyse the changes to surface tip–sample chemical interactions on the friction force, we have chemistry further and investigate their role in surface friction.used chemical force microscopy (CFM), in which the tip Our objective is the development of SFM-based technologies chemistry is controlled by the deposition of a self-assembled for exploring the nanoscale properties of polymer surfaces monolayer (SAM).To our knowledge this is the first study of that have been modified in plasma. We were interested to a plasma treated polymer by CFM. We hypothesised that if know whether there was a correlation between changes in the increase in friction on plasma-treatment is predominantly wettability and the frictional properties of the modified poly- the result of increased acid–base interactions between tip and mer surface.It is known that acid–base (electron donor– sample then the measured friction force should vary with acceptor) chemistry plays an important role in the interfacial tip chemistry in the following order: acid-terminated interactions of polymers, significantly improving their mixing, tip>uncoated silicon nitride tip>methyl-terminated tip and adhesion, adsorption on fillers and fibres and their solubility the magnitude of the increase in friction on plasma treatin organic liquids.11 Specifically, we have shown how it is ment should be significantly lower when the apolar, methyl-functionalised tips are used.possible to calculate the surface free energy of poly(ethylene J. Mater. Chem., 1998, 8, 2845–2854 2845of Cr and 20–25 nm of Au on the front face, and (b) 18–20 nm Experimental of Au on the back-face of the cantilever. The evaporation rate Melinex ‘O’, an additive-free PET with low surface roughness, for the gold was always below 0.03 nm s-1 to ensure that the was obtained from ICI ( Wilton, UK).Mylar D (manufactured cantilevers did not bend during heating.18,22 Once cool, by Du Pont, USA), a PET film treated to incorporate a the cantilevers were immersed in 1 mM solutions of dodecaneparticulate silicate surface additive, was obtained from thiol (from Fluka) or mercaptoundecanoic acid in degassed Goodfellow Advanced Materials (Cambridge, UK).Both ethanol for at least 18 h for the self-assembly process. The materials were biaxially orientated and were used as received. mercaptoundecanoic acid was synthesised according a pro- Plasma treatments were carried out in an inductively-coupled cedure adapted from the literature.23 All glassware was cleaned radio frequency (13.56 MHz) reactor with a base pressure of with ‘Piranha’ solution (357 mixture of 30% hydrogen peroxide 4×10-2 mbar, constructed following a design by Dr R.D. and concentrated sulfuric acid) before use. (Great care should Short of the Department of Engineering Materials at the be exercised in handling Piranha solution; it is an extremely University of SheYeld. Argon (BOC, special gases, UK) was strong oxidising agent and has been known to detonate flowed through the reactor for 15 min before treatment.Plasma spontaneously on contact with organic material.) The functreatment was carried out at 0.1 or 1.0 mbar argon pressure tionalised tips were kept in the alkanethiol solutions until use.and 10W power. After treatment, the reactor was evacuated down to base pressure before exposing the sample to laboratory Results atmosphere. Static advancing contact angles were measured within 30 Contact angle goniometry minutes of plasma treatment on a Rame–Hart model 100-00 The observed contact angles of water, ethylene glycol, formam- goniometer. Water was triply distilled before being passed ide, diiodomethane and glycerol on Melinex ‘O’ treated in 0.1 through a Millipore ‘Milli-Q’ purification system.and 1.0 mbar argon plasmas are shown in Fig. 1 and 2 Diiodomethane (>99%), 1-bromonaphthalene (>97%), forrespectively. Increasing plasma treatment led to increasing mamide (>99.5%), glycerol (>99.5%), and ethylene glycol wettability of the polymer by all the polar liquids.Near- (>99.5%) were all from Merck (Darmstadt, Germany) and limiting values at 0.1 and 1.0 mbar were obtained after 120 used as supplied. Recorded angles are averages of at least six and 300 s respectively. In contrast, the contact angle of the measurements. apolar liquid diiodomethane was found to increase on short X-Ray photoelectron spectra of PET samples treated with time exposure to plasma.Diiodomethane has a slight c+ argon plasma at 0.1 mbar were recorded using a Vacuum component, but as a first approximation this liquid may be Generators ESCALAB instrument with a base pressure of considered apolar.24 We have found the inclusion or exclusion 1×10-8 mbar. The system was equipped with an unmonochroof this small polar term makes no diVerence to the surface mated twin anode X-ray source and 100 mm radius hemispherical electron energy analyser.The sample area analysed by this system was approximately 9 mm in diameter with a take-oV angle of 60°. Al-Ka radiation (1486.6 eV) was used throughout. Survey scans (at 50 eV analyser pass energy) and C 1s and O 1s scans (at 10 eV pass energy) were recorded. The areas under C 1s and O 1s curves were calculated and the O5C ratios were determined using empirically derived sensitivity factors reported by Briggs and Seah.19 Topographic scanning force microscopy (SFM) images were obtained in ambient conditions with a TopoMetrix Explorer scanning probe microscope (TopoMetrix Corp., SaVron Walden, UK).Contact mode imaging was generally performed using silicon nitride cantilevers (nominal force constant 0.064 N m-1) supplied by the microscope manufacturer.The only exception was for the CFM data where Nanoprobes (nominal force constant 0.12 N m-1, from Digital Instruments) were used. The applied load was thought to be Fig. 1 Contact angles of water (&), ethylene glycol ($), formamide (+), diiodomethane (,) and glycerol (2) on PET plasma modified <10 nN for the contact mode imaging in constant force mode.in 0.1 mbar argon. Lateral force imaging was performed simultaneously with the topographical imaging. The eVect of scan velocity on the observed lateral force contrast has been documented.20,21 Since we are only interested in relative changes to the frictional force signal, this complication has been avoided by acquiring all scans in Fig. 7–10 at a constant scan rate (21 mm s-1). To allow meaningful comparison of the frictional responses of diVerent materials, the same tip was used throughout, and the alignment of the laser on the cantilever was not altered for all of the samples tested. The lateral force signals on line profiles from the forward and reverse lateral force images were compared to produce friction loops.A Nanoscope IIIa MultiMode AFM (Digital Instruments, UK) was used for the chemical force microscopy study. The scope mode of the microscope was utilised to provide friction loops. ‘Nanoprobe’ SFM tips were modified18 with alkanethiol SAMs of the same carbon chain length terminated with either hydrophilic (carboxylic acid) groups or hydrophobic (methyl ) Fig. 2 Contact angles of water (&), ethylene glycol ($), formamide groups. A General Engineering bell jar vacuum system was (+), diiodomethane (,) and glycerol (2) on PET plasma modified in 1.0 mbar argon.used to coat the tip-cantilever assemblies, as follows: (a) 2 nm 2846 J. Mater. Chem., 1998, 8, 2845–2854energy calculations. Similar behaviour was also observed using another apolar liquid, 1-bromonaphthalene.There is close similarity between water and glycerol angles. The contact angles of the five liquids shown in Fig. 1 and 2 have been used in the determination of the surface free energy change on argon plasma treatment using computer programs25 following the method of van Oss et al., using their values24,26,29 for the surface tension parameters of the five liquids.Details of the calculation of surface free energy from contact angle data are given in Appendix 1. The change in surface free energy after plasma treatment at 0.1 and 1.0 mbar argon is shown in Fig. 3 and 4. A significant increase in the polar component can be clearly seen, with changes occurring faster at the lower pressure studied.Fig. 3 and 4 also show that at both argon pressures, the apolar component initially decreases on exposure to the plasma before recovering to a value near Fig. 5 Thermodynamic work of polymer–water (&) and polymer– that on untreated Melinex ‘O’. The increase in basic component formamide (+) adhesion after plasma treatment at 0.1 mbar. The acid–base components to the total work of adhesion are also shown; to the surface free energy after plasma treatment is larger at water ($) and formamide (,).either pressure studied than the increase in acidic component. The variation in the thermodynamic work of solid–liquid adhesion with treatment time has also been calculated. Fig. 5 illustrates the changes at 0.1 mbar for water and formamide. X-Ray photoelectron spectroscopy An O5C ratio of 0.39±0.02 was determined for untreated Melinex ‘O’ in good agreement with the theoretical value of 0.40. Fig. 6 shows C 1s spectra (corrected for charging eVects) for untreated and plasma treated Melinex ‘O’. The data show broadening of the line widths of the C 1s peaks. Similar behaviour has been reported in the XP C 1s spectra of several polymers, including PET, after argon plasma treatment, and is regarded as evidence of an increased variety of carbon species on plasma treatment.7 Notably, a new peak has Fig. 3 Surface free energy change on plasma treatment at 0.1 mbar argon; csLW (&), csAB (,), c+ (+) and c- ($). Fig. 6 Fitted C 1s spectra of PET, untreated (a) and plasma-treated Fig. 4 Surface free energy change on plasma treatment at 1.0 mbar argon; csLW (&), csAB (,), c+ (+) and c- ($).for 10 min (b) and 2 h (c). J. Mater. Chem., 1998, 8, 2845–2854 2847appeared in the C 1s spectrum. Argon plasma treatment at either pressure studied increased the O5C ratio by a similar amount. Table 1 shows the O5C ratio as a function of treatment time for 0.1 mbar argon. Near-limiting values were obtained after 1 min treatment, after which only small changes to the lineshapes occurred and it reached a steady state by about 10 min.Survey scans showed that no nitrogen was present after plasma treatment. Lateral force microscopy The plasma-modified Melinex ‘O’ and Mylar D surfaces were more easily disrupted by the motion of the tip during scanning (at the same applied load) than the untreated polymer.Lateral force imaging of untreated Melinex ‘O’ and Mylar D films showed only small frictional contrast on scanning in forward and reverse directions. After treatment with argon plasma there was a significant enhancement in the observed lateral force contrast. Lateral force images revealed greater detail of the structure of the surface additives on the plasma treated Mylar D surface (Fig. 7) although we have observed that the quality of images of the polymeric regions is poor in contact mode. Illustrative topography, lateral force images and friction loops (see below) of Mylar D and Melinex ‘O’ both plasma treated for 2 min are shown in Fig. 8 and 9. Friction loops have been constructed from line profiles and are shown for plasma-modified Melinex ‘O’ and Mylar D in Fig. 8(d) and 9(d) respectively. Fig. 10 shows the dependence of the frictional force on plasma treatment time for Melinex ‘O’ and over the additive features and the polymer surface for Mylar D. The frictional forces are larger over the polymer than the silicate additives in Mylar D. The diVerence in friction is much greater than on the untreated film. The maximum LFM signal is about seven times greater than that on the untreated Melinex ‘O’ and about five times that on the untreated Mylar D.Chemical force microscopy Imaging plasma treated surfaces with AFM tips functionalised with a hydrophobic methyl-terminated alkanethiol monolayer led to some reduction in sample damage, which was more pronounced on surfaces which were treated for longer periods.Friction loops were constructed from images of plasma-treated Mylar D surfaces (or taken directly from the scope mode of Fig. 7 LFM images of Mylar D plasma treated for 10 s at 0.1 mbar. Images are forward (a) and reverse (b) directions. Z-scale ranges: (a) the Nanoscope) with three chemically distinct types of ‘Nano- 1.66 to -2.55 nA; (b) 6.58–3.17 nA. probe’ AFM tips; (i) unmodified silicon nitride, (ii) methylterminated, (iii) carboxylic acid-terminated.It was found that 0.064 N m-1 tips, as expected since the nanoprobes used to the observed lateral forces did depend on the chemistry of the study the eVect of tip chemistry were much stiVer laterally. AFM tip, for both untreated and plasma-treated Mylar D surfaces, with the unmodified and carboxylic acid-modified tips having greater frictional interaction with the sample than Discussion the hydrophobic tips (Fig. 11). Table 2 shows relative frictional The main mechanisms currently thought responsible for the coeYcients calculated from the slopes of friction vs. applied bondability improvement of plasma treated polymer surfaces load plots. The change in lateral force signal on plasma are interfacial diVusion (aided by an increased molecular treatment was much smaller than that obtained with the mobility caused by chain scission) and increased wettability.30 The increased molecular mobility has been inferred from the observation that plasma treated polymers are bondable below Table 1 XP data, O5C ratio as a function of treatment time (0.1 mbar their melting points.30 The incorporation of polar functionaliargon) ties should result in an improvement in the polymer wettability, and the data in Fig. 1 and 2 show much lower contact angles Treatment time/min O5Ca are obtainable for all the polar test liquids used following plasma treatment. 0 0.39 1 0.50 Surface free energy calculations for untreated Melinex ‘O’ 5 0.52 by the method of van Oss and coworkers revealed a similar, 10 0.48 but slightly higher, value (cs#47 mN m-1) than has been 15 0.52 obtained in recent studies (cs#44 mN m-1).29 Although the 30 0.50 PET was obtained directly from sheets, nominally with clean 60 0.50 sides facing inwards, it is possible that it may pick up some 120 0.51 oleophilic impurities on exposure to atmosphere.31 To test for aTypical error in O5C ratio is ±0.03.contamination the polymer was sonicated in diethyl ether for 2848 J. Mater. Chem., 1998, 8, 2845–2854Fig. 8 LFM images of Melinex ‘O’ plasma treated for 2 min at 0.1 mbar. Topographic (a) and lateral force images in forward (b) and reverse (c) directions. Friction loop (d) constructed from lateral force line profiles. Z-scale ranges: (a) 0–2 nm; (b) -1.74 to-9.28 nA; (c) 14.54–8.85 nA. 10 min and dried in high-purity nitrogen immediately before been reported for corona discharge-treated PET by Briggs et al.33 and these probably contribute to the increase in acidic the contact angle measurements. The angles obtained were the same as those without the cleaning procedure; we conclude interactions reported here. The significant increase in basic interactions on the plasma- that the slight diVerences in wettability between our samples and others more likely reflect details of the polymer manufac- treated surface is notable.We suggest that the increase is due predominantly to the formation of carbonyl groups. XP ture (such as surface roughness) rather than the presence of contamination. XPS showed the ratio of oxygen to carbon in spectra (see later) reveal a new peak which is attributed to the creation of carbonyl functions on the plasma-modified sur- the untreated film to be close (0.39) to the expected value (0.40), and a lack of obvious contamination.face.8,34 Carbonyl groups are thought to exhibit basic character32 and CueV et al. have reported XPS data showing that, Calculations of the variation in surface free energy after plasma treatment at 0.1 and 1.0 mbar argon are shown in for argon plasma treatment under their experimental conditions, the only carbon–oxygen functions to increase in Fig. 3 and 4 respectively. The total surface free energy obtained in this study is 64±2 mNm-1, after a period very much intensity were carbonyl groups.34 Nitrogen and oxygen plasma treatments of polypropylene have both been reported to pro- shorter than that required for the formation of the oriented, ridged structures reported in a previous study.8 As expected duce a predominantly basic surface.35,36 Whilst in the former case incorporation of basic N-containing functionalities is from the contact angles, a significant increase in the polar component can be clearly seen.On plasma treatment the expected, in the latter a similar incorporation of carbonyl functions may be occurring. surface has acquired a pronounced acidic and more notably, basic character. Acidic interactions through formation of van Oss, Chaudhury and Good have noted24 that the occurrence of large basic components of the polar free energy hydroxy and carboxyl groups on plasma treatment were expected,7 and an enhancement in the acidic component of (together with a much smaller c+ value) is not uncommon in polymers and natural compounds such as proteins.These are about 2 mN m-1, very similar in magnitude to that calculated in the present study, has been reported in the argon plasma termed24 ‘monopolar surfaces’, and PET is considered27 to be a moderate c- monopole.Strong hydrogen bonding in the treatment of polycarbonate.32 This was explained by the formation of phenolic hydrogen species, formed by photo- depolymerised surface may increase the pKa of the acid groups formed on plasma treatment, leading to a decrease in the acid Fries rearrangements. The formation of phenolic species has J. Mater. Chem., 1998, 8, 2845–2854 2849Fig. 9 LFM images of Myler D plasma treated for 2 min at 0.1 mbar. Topographic (a) and lateral force images in forward (b) and reverse (c) directions. Friction loop (d) constructed from lateral force line profiles. Z-scale ranges: (a) 0–40 nm; (b) -1.33 to -11.11 nA; (c) 9.46–2.44 nA. Fig. 11 Variation of friction force with AFM tip chemistry for Fig. 10 Variation in friction signal with treatment time.Melinex ‘O’ (+), on polymer background of Mylar D (&), over additives on untreated Mylar D and plasma treated for 20 min at 0.1 mbar. Tipsample combinations: unmodified-plasma treated (+), unmodified- Mylar D ($). untreated (6), methyl-plasma treated ($), methyl-untreated (#). component of the polar free energy. It should also be noted that the exact acidic and basic parameters are dependent on set cw+=cw-=25.5 mN m-1).The authors have suggested that the modified values may allow the acidic properties of the values of the test liquid acid and base components chosen. A revised scale of the acid–base parameters of common polymers to be more correctly expressed.37 The similarity of water and glycerol contact angles on solvents with cw+=65 mN m-1 and cw-=10 mN m-1 has recently been proposed (previously van Oss and co-workers plasma-modified surfaces is notable.It has been reported that 2850 J. Mater. Chem., 1998, 8, 2845–2854Table 2 Dependence of the relative friction coeYcient over polymeric regions of Mylar D on the chemistry of the AFM tip Relative friction coeYcient Tip chemistry Untreated polymer After plasma treatmenta Silicon nitride 0.20±0.02 0.36±0.05 Methyl-terminated 0.07±0.01 0.23±0.05 Carboxylic acid-terminated 0.21±0.04 0.32±0.03 aPlasma treatment for 20 min at 0.1 mbar.Errors shown are standard deviations of values from 3–10 separate determinations. the contact angle of glycerol is virtually identical to that of treatment in agreement with the rapid increase in basic interaction revealed by the contact angle data.Small changes in water for a wide range of biological systems.27 DiVerences in the relative acid–base parameters of the two liquids can the C 1s lineshape do occur beyond this point, notably an increase in the carbonyl peak intensity; however the changes account for the unexpected closeness of angles on two liquids whose surface tensions diVer by 12%.27 PET does not become are smaller than those that occur in the first min of treatment and steady state is reached after 10 min.An exact correlation completely wetted by either water or glycerol on argon plasma treatment. This is indicative of a maximum in surface energy. between the time-scales of the variations in contact angle and XP spectra is not expected due to the diVerent sampling depths As suggested previously,8–10 this is probably due to a steadystate being reached between functionalisation (polar group and hence diVering surface selectivities of the two techniques: wettability is sensitive to the outer 0.5–1.0 nm of the surface40 incorporation) and surface etching.Comparison with data for diiodomethane shows that, sur- whilst XPS data contain contributions from a greater depth, ca. 5–10 nm.19 prisingly, the contact angle of the apolar liquid increases on exposure to plasma. We have previously shown8 that plasma Fig. 7 shows lateral force images of the Mylar D surface after only 10 s exposure to plasma. While we have previously treated material is much more susceptible to tip-induced damage during scanning.We suggest that the observed increase reported that plasma treated PET is disrupted during contact mode SFM,8 the additive particles are imaged here with clarity.in the contact angles with apolar liquids (indicative of a decrease in dispersive interactions between surface and test Some of the additive features are in fact aggregates of several smaller particles.After longer exposure this delineation of the liquid) is related to surface disorder caused by plasma-induced chain-scission. The increase in surface mobility (disorder) is additives became less clear, presumably because they were damaged by the plasma. discussed further below in connection with the SFM data. The contact angle measurements of test liquids on a solid The contrast in LFM arises from twisting motions of the cantilever as it transverses the surface.These twisting motions polymer surface have also been used to calculate the thermodynamic work of solid–liquid adhesion, as illustrated in Fig. 5. arise from forces acting parallel to the plane of the sample surface. It is clear that frictional forces contribute to the LFM The acid–base contribution to the work of adhesion increases on plasma treatment, as expected from the increase in surface signal; however, when the local topography of the surface changes then the LFM signal may also contain contributions free energy.38 The increase in the acid–base component of the work of water–PET adhesion is greater than the corresponding from normal ( load) forces.Appendix 2 shows how the eVects of normal forces on the frictional signal can be eliminated by increase in the work of formamide–PET adhesion.Since water is a much stronger acid than formamide the importance of scanning in forward and reverse directions. The images in Fig. 7 and 9 appear to indicate that an inversion of contrast basic interactions on the plasma-modified surface is clear. As shown in Table 1, XPS reveals a substantial increase in occurs over the additives on reversing the scan direction, implying significant frictional interaction.However, examin- the O5C ratio after plasma treatment and subsequent exposure to atmosphere. The O5C ratio is 0.50±0.03 after 1 min ation of the friction loops of Mylar D plasma-treated for 2 min [Fig. 9(d)] reveals that the contrast inversion in the treatment at 0.1 mbar argon and does not rise significantly thereafter.The magnitude of this increase is in agreement with LFM image is illusory. Careful consideration of the line profiles shows that the lateral force over the additives changes a recent determination (O5C#0.51–0.56) by France and Short7 at 10 W and 2.5×10-2 mbar argon using a similar little, while a large change is seen over the polymeric regions.The magnitude of the change is such that the relative contrast reactor configuration. It has been considered7–10 that plasma attack on the ester functionally would be likely to lead to over the silicates changes in the image; however, the largest frictional interaction, according to the analysis of Grafstro�m chain scission, leading to etching and a relatively low saturation level for oxygen incorporation and a large amount of low et al.17 (see Appendix 2) is over the polymer.Friction loops have shown that the lateral force is constant over an image, molecular weight material.7,9 By comparing the etching rates of several organic materials, Prat et al. have concluded that suggesting that although the plasma-modified surface can be worn by the SFM tip during scanning the tip-induced topo- polymers containing functions such as ester groups are more susceptible to degradation since it can more easily occur by graphic changes (which, in principle, could aVect the friction measurement by making the subtraction inexact) are small.In initial chain scission at the functional groups.10 Indeed, there are reports of PET surface modification by argon plasma an earlier paper41 we observed a similar (although smaller) contrast inversion over the additive particles in untreated which show a small decrease in C–O and ester peak intensity. 34,39 It has been suggested that the breaking of ester bonds Mylar D. However a re-examination of the untreated material suggests that the inversion is illusory there too, and that, for can lead to radicals that are resonance-stabilised over those formed in C–C bond breaking.9 Under conditions where the untreated Mylar D, the largest friction interaction is on the polymer surface.Obviously, great care must be taken in the authors reported a loss of ester oxygen, the only peak found to increase in intensity was a new species of ca. 3.0 eV higher interpretation of image contrast in LFM. For Mylar D, the contrast over the polymer background binding energy than hydrocarbon.34 These authors, and others,7 have assigned this to the creation of isolated car- and over the additive surface both show a sharp initial increase with time of plasma treatment as shown in Fig. 10. The bonyl groups. Peak fitting to the C 1s lineshape, after plasma treatment diVerence in friction between the additives and the polymeric background is much larger than on the untreated film.has also revealed a new peak 3 eV from hydrocarbon in the present study (Fig. 6). The peak appears after only 1 min Illustrative topography, lateral force images and friction loops J. Mater. Chem., 1998, 8, 2845–2854 2851of Mylar D and Melinex ‘O’ both plasma treated for 2 min If acid–base interactions were solely responsible for the increase in friction following plasma treatment, the magnitude are shown in Fig. 8 and 9. Comparison of the lateral force line profiles with the topographic image provides evidence that of the change would be significantly smaller when the apolar, methyl-functionalised tips were used.However, the magnitude the lateral forces are aVected by the local sample slope; the friction signal is clearly altered as the tip encounters the surface of the frictional increase on plasma treatment was similar for all the tip chemistries used. Therefore it is clear that while additives. However, the friction over the central gions of the additives and over the polymeric background is invariant chemical interactions contribute to friction following treatment, there is a substantial additional contribution, which we across the image; the method of Grafstro�m et al.17 is applicable for surfaces of this roughness.attribute to an increase in the tip–sample contact area as a consequence of the mechanical softening of the surface. The The data shown in Fig. 10 have all been recorded using the same SFM tip. Thus, although we do not know the exact observed correlation between surface free energy, probed by wettability and XPS, and surface friction, probed by LFM, is lateral forces (as both the lateral spring constants of our tips and the sensitivity of our microscope to lateral displacements a consequence of the fact that chain scission ( leading to mechanical weakness in the surface layer) and polar group are not accurately known), the relative frictional forces are accurately (±10%) determined.After plasma treatment incorporation ( leading to increasing surface energy) occur on a similar time scale, and are complementary aspects of the Melinex ‘O’ surface has a frictional response up to seven times greater than the virgin material.The surface friction measured same physical process. by LFM reaches limiting values on closely similar time-scales to the wettability and XPS data. If acid–base interactions between tip and sample are import- Conclusion ant in determining the frictional interaction then the measured friction force should vary with tip chemistry in the follow- The combination of lateral force microscopy, wettability, and ing order: acid-terminated tipµunmodified silicon nitride X-ray photoelectron spectroscopy has been used to analyse tip>methyl-terminated tip.Although the bulk composition of changes at the film surface after plasma treatment of polythe unmodified tip is silicon nitride, it is thought that oxides (ethylene terephthalate).Calculations on contact angle data and silanols are present at the surface yielding a polar tip.42 with a combination of polar and non-polar liquids have shown Unmodified and COOH-terminated tips are known to show that argon plasma treatment considerably enhances the work similar frictional characteristics when imaging SAMs in air.43 of solid–(polar) liquid adhesion and the surface free energy of Fig. 11 shows the variation in friction force with load for the films.This is shown to be due to the creation of acidic methyl-terminated and unmodified tips before and after plasma and basic functions on the polymer surface. This is confirmed treatment for 20 min. It can be seen that for a given load, the by XP spectra which show an increased oxygen:carbon ratio friction force measured with the methyl terminated tip for the after plasma treatment.In contrast, the Lifshitz–van derWaals treated polymer is significantly higher than that measured with (apolar) interactions decrease as a consequence of plasmathe same tip for the untreated polymer, but similar to the induced chain-scission. Friction force microscopy has shown force measured for the treated polymer with a bare tip.From that plasma-modified surfaces exhibit substantially higher fricthe slopes of such plots it is possible to determine a relative tion than untreated material and are more easily disrupted by coeYcient of friction for a specific tip–sample combination, the movement of the tip during scanning. Typically, modified and these data are plotted in Table 2 along with measurements surfaces show a maximum frictional response which is about for acid-terminated tips. Both polar (unmodified and car- seven times higher than untreated material when imaging with boxylic acid-modified) tips have greater frictional interaction the less stiV cantilevers.Friction forces on plasma-treated and with the polymeric samples than the hydrophobic (apolar) tips unmodified Mylar D depend on the surface chemistry of the do.There was a noticeable improvement in resolution on AFM tip. There is a correlation, on plasma treatment, between imaging the plasma-treated samples with hydrophobic methyl- the rapid increases in surface friction probed by lateral force coated tips. These results indicate that polar interactions microscopy and surface free energy probed by wettability and between tip and sample make a significant contribution to the X-ray photoelectron spectroscopy.Increased surface disorder friction force measured by LFM. and polar group incorporation both result from scission of The frictional force of an adhesive contact is a function of polymer chains and contribute to the increase in friction.the contact load (here kept constant), the area of contact and the surface free energies of the two surfaces. On plasma treated surfaces both an increase in surface free energy (shown by the improved wettability) and an increase in tip–sample contact Acknowledgements area (treated surfaces are more easily disrupted by the motion The authors are grateful to the EPSRC (grant GR/K/88071), of the tip during contact mode scanning) occur and are the Royal Society and the Society of Chemical Industry for mediated via chain scission processes.Studies by Bar et al. on financial support. J.S.G.L. thanks the EPSRC for a research LFM and force modulation microscopy of self-assembled studentship. G.J.L. thanks the NuYeld Foundation for a monolayers have shown the importance of packing density Science Research Fellowship.The authors would like to thank (and hence tip–sample contact area) in determining the conthe Department of Materials Engineering and Materials trast in LFM images on chemically identical regions, with Design, University of Nottingham, where some of this work sharp contrast observed between heptanethiol and octadecanewas carried out, and acknowledge the assistance of J.C. thiol monolayers.44 We have used LFM to probe the surface Bussey (University of Nottingham) in obtaining some of the friction on regions of diVerent hydrophobicity on photopat- XPS spectra. The authors are most grateful to Dr J. H. Clint terned PET films and photopatterned SAMs.45 Imaging under (Surfactant Science Group, School of Chemistry, University similar conditions led to considerably smaller observed LFM of Hull ) for supplying copies of his programs to obtain surface contrast between hydrophilic and hydrophobic regions on free energy data from contact angles and for useful discussions, these surfaces than was observed here between treated and and to Dr R.D. Short (University of SheYeld) for his untreated polymers. This observation, together with measureassistance in designing the plasma reactor.The authors would ments of frictional forces between single-component selfalso like to thank Professor D. Briggs (ICI, Wilton) for assembled monolayers,45 suggests that the greater frictional supplying Melinex ‘O’ with known orientation and for helpful contrast observed after plasma treatment is not entirely due to diVerences in wettability.and stimulating discussion. 2852 J. Mater. Chem., 1998, 8, 2845–2854contribution to the LFM image may be removed by sub- Appendix 1: Theory and calculation of surface free tracting images recorded in the forward and reverse directions. energies from contact angles These comparisons of forward and reverse scans are often called friction loops and should reflect the frictional force Contact angle measurements of liquids on a solid surface can be used to calculate the thermodynamic work of solid–liquid acting between tip and sample.16 adhesion according to the Young–Dupre� equation, [Wsl= cl(1+cosh)].The acid–base contribution to the work of adhesion can be separated30,46 from the Lifshitz–van derWaals References interactions: 1 E.M. Liston, L. Martinu and M. R. Wertheimer, J. Adhes. Sci. Wsltotal=WslLW+WslAB Technol., 1993, 7, 1091. 2 E. M. Liston, in The Interfacial Interactions in Polymer LWrepresents the sum of the three electrodynamic interactions Composites, ed. G. Akovali, Kluwer Academic, London, 1993, which decay with distance at the same rate (primarily the p. 223.don [dispersion] force, with small contributions from the 3 A. M.Mayes and S. K. Kumar, MRS Bull., 1997, 22, 43. 4 F. D. Egitto and L. J. Matienzo, IBM J. Res. Dev., 1994, 38, 423. Keesom dipole–dipole [orientation] and Debye dipole–induced 5 K. Harth and H. Hibst, Surf. Coat. Technol., 1993, 59, 350. dipole [induction] force), collectively designated24,26 as 6 F. Denes, TRIP, 1997, 5, 23.Lifshitz–van der Waals (‘apolar’) interactions. AB represents 7 R. M. France and R. D. Short, J. Chem. Soc., Faraday Trans., the polar or acid–base interactions. Similarly, it has recently 1997, 93, 3173. been established24,26–29 that the surface free energy of the solid 8 B. D. Beake, J. S. G. Ling and G. J. Leggett, J. Mater. Chem., can be separated into two terms: 1998, 8, 1735. 9 F. Clouet and M. K. Shi, J. Appl. Polym. Sci., 1992, 46, 1955. cstotal=csLW+csAB 10 R. Prat, M. K. Shi and F. Clouet, J. Macromol. Sci–Pure Appl. Chem. A, 1997, 34, 471. A geometric mean approach can be used to determine the 11 Acid–Base Interactions, Relevance to Adhesion Science and apolar component of the solid free energy from the contact Technology, ed.K. L. Mittel and H. R. Anderson, Jr, VSP, angle of an apolar test liquid providing that the liquid is of Utrecht, 1991. 12 J. A. McLaughlin, D. Macken, B. J. Meenan, E. T. McAdams and high enough surface tension that it does not spread completely P. D. Maguire, Key Eng. Mater., 1995, 99–100, 331. over the solid surface.24,26–29 However, such an approach is 13 J. F. Friedrich, W. Unger, A.Lippitz, T. Gross, P. Rohrer, flawed when considering the polar component.29,38 Instead, W. Saur, J. Erdmann and H-V. Gorsler, J. Adhes. Sci. Technol., several authors35,36,38 have suggested the approach of van Oss, 1995, 9, 575. Chaudhury and Good24,26–29 is the most useful in the investi- 14 A. Ringenbach, Y. Jugnet and Tran Minh Duc, J. Adhes. Sci. gation of acid–base properties of plasma-modified polymer Technol., 1995, 9, 1209. 15 P. Gro�ning, M. Collaud, G. Dietler and L. Schlapbach, Vide surfaces. Van Oss et al. have introduced the concept of surface Couches Minces, 1994, 272SS, 140. tension parameters, c+ and c-, due to acidic and basic 16 R. Overney and E. Meyer, MRS Bull., 1993, 18, 26. functionalities respectively.24 These comprise (non-additively) 17 S.Grafstro�m, M. Neitzert, T. Hagen, J. Ackermann, the polar surface free energy of a material i, (ciAB), thus: R. Neumann, O. Probst and M. Wo� rtge, Nanotechnology, 1993, 4, 143. ciAB=2(ci+ci-)1/2 18 A. Noy, C. D. Frisbie, L. F. Rozsnyai, M. S. Wrighton and C. M. Lieber, J. Am. Chem. Soc., 1995, 117, 7943. The key to the modification of the Young’s equation (below) 19 Practical Surface Analysis, ed.D. Briggs and M. Seah, 2nd edn. proposed by these authors is the realisation that acidic func- Wiley, 1990, vol. 1, p. 635. tions in the test liquid interact with only basic functionalities 20 G. Haugstad, W. L. Gladfelter, E. B. Weberg, R. T. Weberg and on the solid surface and vice versa. R. R. Jones, Langmuir, 1995, 11, 3473. 21 J. A. Hammershmidt, B. Moasser, W.L. Gladfelter, G. Haugstad (1+cosh)cl=2[(csLWclLW)1/2+(cs+cl-)1/2+(cs-cl+)1/2] and R. R. Jones, Macromolecules, 1996, 29, 8996. 22 E. W. van der Vegte and G. Hadziioannou, Langmuir, 1997, 13, In this Lifshitz–van der Waals/acid–base approach there are 4357. three unknown values for the surface tension components. 23 C. D. Bain, E. B. Troughton, Y.-T. Tao, J. Evall, G.M. Whitesides When contact angle measurements are carried out with three and R. G. Nuzzo, J. Am. Chem. Soc., 1989, 111, 321. or more liquids (at least two of which are polar) with known 24 C. J. van Oss, M. J. Chaudhury and R. J. Good, Adv. Colloid cLW, c+ and c- values, a standard least-squares method can Interface Sci., 1987, 28, 35. 25 The computer programs used to obtain surface free energy data be used to solve for the three unknown parameters.29 from contact angles were from Dr J.H. Clint ( University of Hull ). 26 C. J. van Oss, M. J. Chaudhury and R. J. Good, Chem. Rev., 1988, 88, 927. Appendix 2: Construction of friction loops to 27 C. J. van Oss, M. J. Chaudhury and R. J. Good, Langmuir, 1988, determine surface friction 4, 884. 28 C. J. van Oss, Colloids Surf.A: Physico. Eng. Asp., 1993, 78, 1. A line profile of the forward and reverse scans, a friction loop, 29 W.Wu, R. F. Giese, Jr. and C. J. van Oss, Langmuir, 1995, 11, 379. can be used to determine the frictional interaction between tip 30 Polymer Interface and Adhesion, ed. S. Wu, Marcel Dekker, New and sample according to the treatment of Grafstro�m et al.17 York, 1982, pp. 298–321. These authors have shown that for a force (FN) normal to the 31 Prof. D. Briggs, personal communication. 32 S. Vallon, B. Dre�villon, F. Poncin-Epalliard, J. E. Klemberg- sample due to the application of a load during imaging, the Sapieha and L. Martinu, J. Vac. Sci. Technol. A, 1996, 14, 3194. lateral force (Fy) in a direction y orthogonal to both the 33 D. Briggs, D. G. Rance, C. R. Kendall and A. R. Blythe, Polymer, cantilever and surface normal is given by 1980, 21, 895. 34 R. CueV, G. Baud, J. P. Besse, M. Jacquet and M. Benmalek, Fy=Fty-syFN J. Adhes., 1993, 42, 249. 35 J. Behnisch, A. Holla�nder and H. Zimmermann, Int. J. Polym. where sy is the slope in the y-direction and Fty is the force Mater., 1994, 23, 215. tangential to the surface normal in the y-direction. When there 36 J. Behnisch, A. Holla�nder and H. Zimmermann, J. Appl. Polym. are substantial variations in surface topography the syFN Sci., 1993, 49, 117. component can become significant. Since the lateral force is 37 C. Della Volpe and S. Siboni, J. Colloid Interface Sci., 1997, 195, determined from the diVerence between signals reaching the 121. left and right halves of a four-segment photodetector, in 38 N. Shahidzadeh-Ahmadi, F. Arefi-Khonsari and J. Amouroux, J. Mater. Chem., 1995, 5, 229. principle ignoring piezoelectric drift) the topographic J. Mater. Chem., 1998, 8, 2845–2854 285339 Q. T. Le, J. J. Pireaux and J. J. Verbist, Surf. Interface Anal., 1994, 44 G. Bar, S. Rubin, A. N. Parikh, B. I. Swanson, T. A. Zawodzinski, 22, 224. Jr. and M.-H. Whangbo, Langmuir, 1997, 13, 373. 40 P. E. Laibinis, C. D. Bain, R. G. Nuzzo and G. M. Whitesides, 45 B. D. Beake and G. J. Leggett, unpublished work. J. Phys. Chem., 1995, 99, 7663. 46 X. Qin and W. V. Chang, J. Adhes. Sci. Technol., 1996, 10, 963. 41 J. S. G. Ling and G. J. Leggett, Polymer, 1997, 38, 2617. 42 J. M. Williams, T. Han and T. P. Beebe, Jr., Langmuir, 1996, 12, 1291. 43 J.-B. D. Green, M. T. McDermott, M. D. Porter and L. M. Siperko, J. Phys. Chem., 1995, 99, 10 960. Paper 8/07261B 2854 J. Mater. Chem., 1998, 8, 2