J. MATER. CHEM., 1994, 4( 5), 683-687 Studies of Vapour-phase Chemical Derivatisation for XPSAnalysis using Model Polymers Ian Sutherland, Enshan Sheng, Derek M. Brewis and Richard J. Heath Loughborough University of Technology, Loughborough, Leicesfershire LEI 7 3TU, UK Vapour-phase chemical derivatisation carried out on a vacuum adsorption rig has been investigated. Model polymers were used to evaluate the reactivity and selectivity of trifluoroacetic anhydride (TFAA) and hydrazine. Results have shown that TFAA is a good derivatising reagent for hydroxyl groups, while hydrazine is not suitable for carbonyl groups. Chemical derivatisation with TFAA of flame-treated polypropylene surfaces has shown that under various flame conditions an approximately constant proportion (ca.20%) of the oxygen introduced by flame treatment was present as hydroxyl groups. Selective removal of OH groups by derivatisation has a large effect on the adhesion of a flame-treated polypropylene surface to a reactive polyurethane paint, evidence of chemical reaction at the interface. X-Ray photoelectron spectroscopy (XPS or ESCA) is one of the most widely used surface analysis techniques available for the study of polymers. It has been shown to be a valuable surface analysis tool for elucidating molecular structure from chemical-shift data. However, even for very-high-energy reso- lution XPS, many functional groups cannot always be ident- ified owing to the overlap of chemically shifted peaks or to their low concentrations on the surface.To overcome this problem, chemical derivatisation can be used. It uses a chemi- cal reagent to react with a specific functional group, with the derivative having a unique element which is not previously present on the surface. Reagents containing fluorine are often used because of its large photoelectron cross-section. Various chemical derivatisation reactions have been reviewed by Briggsl and also Batich.2 Chemical derivatisation has been employed by many work- ers to analyse various functional groups after surface modifi- cation of a polymer. These include: trifluoroacetic anhydride (TFAA) for hydroxyl hydra~ine~'~and penta- fluorophenylhydrazine (PFPH)12 for carbonyl groups; sodium hydroxide (NaOH),11,12 trieth~lamine~.~ and trifluoroethanol (in the presence of a coupling agent and a catalyst) for carboxylic acid gro~ps.~~,~~,'~ Many of these reactions are solution-phase derivatisation with which some problems may be associated.Solvents may extract from a polymer lower molecular weight materials, which may be produced during a surface treatment, and at the same time may cause migration of functional groups away from the surface into the bulk of the polymer. The removal of the solvent after the derivatisation could be particularly problematic if ionic derivatives are employed as in the case of carboxylic acid derivatisation with NaOH where the amount of sodium incorporated has been found dependent on the washing conditions." Considering these possible problems associated with solu- tion-phase derivatisation, vapour-phase derivatisation can be used to eliminate these problems.Vapour-phase derivatisation for XPS analysis was first reported by Hammond et a1.: followed by several other worker^.^.^^ Successful chemical derivatisation requires that the derivatising reagent selectively and ideally completely reacts with the functional group. In this work, the reactivity and the selectivity of TFAA towards Table 1 Model polymers used for vapour-phase derivatisation polymer poly(viny1 alcohol) poly(acry1ic acid) poly(viny1 methyl ketone) poly(ethy1ene terephthalate) abbreviation supplier PVA Aldrich PAA Aldrich PVMK Aldrich PET ICI hydroxyl and hydrazine towards carbonyl groups were exam- ined using several model polymers.Effects of hydroxyl groups on the adhesion of flame-treated polypropylene (PP) with a polyurethane (PU) paint was investigated by vapour-phase chemical derivatisation. Experimental Poly(viny1 alcohol), poly(acry1ic acid), poly(viny1 methyl ketone) and poly(ethy1ene terephthalate) (PET) were used as model polymers for hydroxyl, carboxylic acid, carbonyl and ester groups. Details of these polymers are listed in Table 1. PET was supplied as 100pm film. It was cleaned in an ultrasonic bath with trichloroethylene for 30s and used directly as a model polymer for ester groups. Other model polymers were prepared by solution coating onto this cleaned PET film. The coating method involved dipping a clean piece of PET film into the solutions, i.e.10% PVA aqueous solution, 4% PAA aqueous solution or 4% PVMK solution in dimethylformamide (DMF). The film pieces were dried and kept in a desiccator. XPS and attenuated total reflection (ATR) infrared analysis showed that the PET film was com- pletely covered by the model polymers. Vapour-phase deriv- atisation reagents, trifluoroacetic anhydride (TFAA 99 +%) and hydrazine (anhydrous 98 YO)were purchased from Aldrich. Vapour-phase derivatisation was performed on a vacuum adsorption rig (Fig. 1). Vapour-phase derivatisation on a vacuum adsorption rig is considered superior to that under ambient conditions. Water vapour in the air and adsorbed molecules on sample surfaces may affect the derivatisation reaction. Adsorbed vapour on derivatised sample surfaces can be conveniently removed by pumping under the vacuum adsorption rig.Also, since it is an enclosed system it is safer to operate. There are three sub- rigs under the main adsorption rig. Each sub-rig was used exclusively for one derivatisation reaction. Two liquid-nitrogen traps were used to prevent hydrocarbon vapour from reaching the adsorption system. To avoid cross-contamination, different derivatisation reactions were carried out at a time interval of at least 1 day. Air in the flask (250 cm3) containing ca. 10 cm3 derivatising reagent was removed by pumping after the reagent had been frozen with liquid nitrogen. The sample (1.5 cm x 4.0 cm) was evacuated to a base vacuum of ca.10-5Torr and exposed to the derivatising reagent. During the reaction, the flask containing the reagent was kept at 20 "C (room temperature, ca. 25 "C) using a water bath during the reaction to avoid the conden- sation of the reagent on the sample surface. The water-bath temperature was found to have a significant effect on the rate J. MATER. CHEM., 1994, VOL. 4 vacuumTmain vacuum adsorption rig I I sub-rig 1 i P empty flask liquid nitrogen rotary traps gasoutlet Fig. 1 Schematic of vacuum adsorption rig used for vapour-phase derivatisation of the derivatisation rea~ti0n.I~ After the reaction, the flask was sealed and the specimen pumped for ca. 24 h to expel any physically adsorbed vapour.XPS analysis was normally performed on the same day. X-Ray photoelectron (XPS) spectra were recorded on a VG ESCALAB MKI spectrometer using an Al-Ka X-ray source at a pass energy of 85eV and a take-off angle of 90" with respect to the polymer surface. Elemental compositions were calculated by measuring the peak areas following the subtrac- tion of a Shirley-type background. Scofield photoelectron cross-sections15 were used. Correction was made for inelastic mean-free path,16 transmission of the energy ana1yse1-I~ and angular asymmetry in photoemission'* (where appropriate). Radiation damage was not found to be significant under the conditions used (X-ray source power, 200 W; acquisition time, <5 min). Flame treatment of a propylene homopolymer film (100 pm) manufactured by Neste was carried out as described in a previous paper.Ig A two-pack polyurethane paint (M615-122+M210-763, ICI) was air-sprayed onto the PP surface and cured at 90 "C for 30 min.Tensile adhesion was assessed using a composite butt adhesion test.19 In the test, the sample was bonded by an epoxy adhesive (AV100+HV100, Ciba-Geigy) between two cylindrical steel butts with a length of 50 mm and a diameter of 28 mm, and the tensile adhesion testing was carried out after the adhesive was cured. Results and Discussion Surface oxygen concentrations of model polymers, i.e. PVA, PAA, PVMK, and PET, were analysed with XPS. Results are listed together with their stoichiometric values in Table 2.Experimental values agreed broadly with stoichiometric ones. The slight difference (ca. 15% lower in all cases) might be attributed to the orientation of oxygen-containing groups away from the near surface or to the overestimation of the oxygen sensitivity factor. Derivatisationwith TrifluoroaceticAnhydride Trifluoroacetic anhydride (TFAA) was used to derivatise hydroxyl groups. Its reaction with PVA to form an ester can be written as +CH,--FH* OH PVA model polymer ~~~~~ ~ poly(viny1 alcohol) (PVA) poly(acry1ic acid) (PAA) poly(viny1 methyl ketone) (PVMK) poly(ethy1ene terephthalate) (PET) 00 ::II II + CF~-C-O-C-CF~ -+CH,--yHk + CF3-C-OH 0 ITFAA O=C-CF3 Table 2 XPS elemental compositions of model polymers oxygen concentration (atom%) functional group stoichiometric experimental -F-OH 33.3 28.7 0-&-OH 40.0 33.9 0 20.0 17.4-8-P 28.6 23.6-c-0 J.MATER. CHEM., 1994, VOL. 4 401 401 ’”i ./ 0 0.0 0.2 0.4 0.6 0.8 1.o conversion factor x Fig. 2 Theoretical oxygen (0)and fluorine (x) concentrations of PVA derivatised with TFAA according to eqn. (1) and (2) Assuming that all elements are homogeneously distributed within the XPS sampling depth both before and after the derivatisation, the relationship between the conversion factor and the surface atomic concentrations of oxygen and fluorine can be derived as follows. Neglecting hydrogen atoms, which are not detected, there is an original atomic percentage composition of PVA of [Cl0=71.3% and [010=28.7% as detected by XPS.As the sample is derivatised, each hydroxyl group that reacts introduces six extra atoms: 10, 2C and 3F. Taking the conversion factor as x (x =O no reaction, x =1 complete reaction), then the atomic percentage compo-sition of oxygen and fluorine measured by XPS can be written colo+xcolo = [C], +[010+6x [010 3x cola x 100cF1= [C]o+[O]o+6~[O]o [O] and [F] are plotted against conversion factor x in Fig. 2. This shows that the extent of reaction does not vary linearly with [F] detected, as is often mistakenly assumed. Fig. 2 also shows that [F] changes more rapidly than [O] as the reaction proceeds, indicating that [F] should be used to estimate the conversion factor.PVA was allowed to react with TFAA for various durations. Experimental [O] and [F] as detected by XPS are plotted against reaction time in Fig. 3. TFAA was found to react rapidly with PVA. Limiting fluorine incorporation on the surface was obtained at a reaction time of ca. 2 h. No significant increase in [F] was observed even after 16 h reaction. At 16 h reaction, the conversion factor was calculated to be ca. 80% according to eqn. (2) assuming that all the original oxygen existed as hydroxyl groups. This demonstrates that TFAA has a good reactivity towards hydroxyl groups. To evaluate the selectivity of TFAA, other model polymers were reacted with TFAA for 2 h under the same conditions. Results are shown in Table 3. Table 3 shows that only small amounts of fluorine were detected on PAA and PVMK surfaces after 2 h reaction with TFAA, indicating a good selectivity of TFAA towards hydroxyl groups in the presence of carboxylic acid and carbonyl groups.The reaction of TFAA with PAA is thought to be via the substitution of acid H atom and the reaction with PVMK could be through the reaction with enol groups. t O l d 0 4 8 12 16 reaction time/h Fig. 3 Experimental oxygen (0)and fluorine (x) concentr&ons of PVA as a function of reaction time with TFAA Table 3 XPS elemental compositions of model polymers reacted with TFAA for 2 h elemental compositions (atom%) polymer C 0 F PVA 51.3 23.3 25.4 PAA 65.2 31.2 3.6 PVMK 78.9 17.2 3.9 HPP untreated 97.8 1.1 1.1 Derivatisation with Hydrazine A hydrazone is formed as hydrazine reacts with PVMK: SCH2-FHk + NH2NH2 -SCHZ-THk + H20 c=o C=NNHa I I CH3 CH3 PVMK hydrazone The conversion factor (x) can be related in a similar way, as in the case of TFAA derivatisation, to [O] and [N] on derivatised surfaces by the following equations: (3) (4) where [010=17.4 is the original oxygen concentration before derivatisation.A similar plot as in Fig. 2 shows an almost linear relationship between x and [O] or [N]. Fig.4 shows the experimental [O] and [N] against derivatisation duration of PVMK with hydrazine. Again, there exists a limiting [N] value which corresponds to a conversion factor of .Y 0.5. Further prolonging of the reaction was found to have no significant effect on x.While an incomplete derivatisation of PVMK with hydra- zine was observed, the derivatisation with hydrazine towards carbonyl groups was found to be far less selective than that of TFAA towards hydroxyl groups. XPS analysis results of hydrazine-derivatised model polymers are shown in Table 4. Although the reaction of hydrazine with PVA is negligible, substantial reaction with PAA and particularly PET was found. White powder products were observed on the PET 2o T 0 2 4 6 8 10 12 reaction timeh Fig. 4 Experimental oxygen (0)and nitrogen (x) concentrations of PVMK as a function of reaction time with hydrazine Table 4 XPS elemental compositions of model polymers reacted with hydrazine for 2 h elemental composition (atom%) polymer C 0 N PVMK 75.0 8.9 16.1 PVA 70.6 28.5 0.9 PAA 62.7 27.0 10.3 PET 61.9 14.1 24.0 surface after the reaction. This is thought to be due to the chain-scission reaction of PET with hydrazine. Chemical Derivatisation of Flame-treated PP Polypropylene was treated at various air-to-gas (natural gas) ratios, flame intensities and distances from the polymer surface to the flame inner-cone tip.No oxygen was detected on untreated PP surface. After flame treatment, various oxygen concentrations up to 15 atom% were obtained. Except for those with very low oxygen concentrations (< ca. 2 atom%), the tensile adhesion level between flame-treated PP and polyurethane (PU) paint was so high (ca.26 MPa) that the failure of the test joint was cohesive within the PP film. Derivatisation of flame-treated PP surfaces with TFAA has shown that a substantial amount of oxygen was present as hydroxyl groups. Fig. 5 shows the concentration of oxygen present as -OH groups against the total oxygen concen- tration on the surface. It shows that ca. 20% of the total 4r oxygen concentration (atom %) Fig. 5 Correlation between -OH concentration and total oxygen concentration on flame-treated polymer surfaces J. MATER. CHEM., 1994, VOL. 4 oxygen on flame-treated PP surfaces was present as -OH groups. The role of hydroxyl groups on adhesion was investigated using chemical derivatisation with TFAA. Untreated PP did not show any fluorine on the surface upon TFAA derivatis- ation.PP was treated with a mild flame, and hydroxyl groups on the surface were derivatised with TFAA. Water contact angles (advancing and receding) on PP were observed to increase after the derivatisation. Significant decrease in adhesion level (from 26 to 13 MPa) was found with a two- pack polyurethane (PU) paint after the removal of hydroxyl groups on the PP surface by derivatisation with TFAA. This suggests that the existence of hydroxyl groups on PP is crucial in attaining very high adhesion when the PU paint is used. Hydroxyl groups may have reacted with the PU paint since the paint is a highly reactive system containing isocyanate. A fairly good adhesion level (13 MPa) with the PU paint was retained after the removal of the hydroxyl groups on the flame-treated PP surface, indicating that other functional groups may have some contribution to the level of adhesion achieved by flame treatment.Conclusions Vapour-phase chemical derivatisation with trifluoroacetic anhydride (TFAA) and hydrazine using a vacuum adsorption rig were evaluated using model polymers. The following conclusions can be drawn. (1) TFAA was found to derivatise hydroxyl groups to a high degree. The derivatisation was found to be selective in the presence of carboxylic acid and carbonyl groups. (2) Derivatisation with hydrazine for car- bony1 groups has been found problematic. Incomplete reaction and poor selectivity were found. This casts doubt on the validity of the use of hydrazine to derivatise carbonyl groups as previously rep~rted.~.~ Hydrazine reacts significantly with carboxylic acid and aromatic ester groups.It has been found by Chilkoti and Ratner6 that it does not react with aliphatic esters. (3) An approximately constant percentage (ca. 20%) of oxygen on flame-treated PP surfaces has been found to exist as hydroxyl groups. It has been further demonstrated that these hydroxyl groups are important in the adhesion with a two-pack polyurethane (PU) paint. Adhesion level was shar- ply reduced upon removal of these groups by TFAA derivatis- ation. While hydroxyl groups are important in the adhesion with the PU paint, they may not be so in other cases as of the adhesion between PP and an epoxy adhe~ive.'~ References 1 D.Briggs, in Encyclopaedia of Polymer Science and Engineering, ed. J. I. Kroschwitz, Wiley, Chichester 2nd edn., 1989,vol. 16. 2 C. D. Batich, Appl. Surf. Sci., 1988,32, 57. 3 J. S. Hammond, Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., 1980,21(1), 149. 4 R. A. Dickie, J. S. Hammond, J. E. de Vries and J. W. Holubka, Anal. Chem., 1982,542045. 5 E. Sheng, I. Sutherland, D. M. Brewis and R. J. Heath, Surf. Interface Anal., 1992, 19, 151. 6 A. Chilkoti and B. D. Ratner, Surf. Interface Anal., 1991,17, 567. 7 L. J. Gerenser, J. F. Elman, M. G. Mason and J. M. Pochan, Polymer, 1985,26, 1162. 8 J. M. Pochan, L. J. Gerenser and J. F. Elman, Polymer, 1986, 27, 1058. 9 W. R.Gombotz and A. S. Hoffman, J. Appl. Polym. Sci. Appl. Polym. Symp., 1988,42,285. 10 Y. Yakayama, T. Takahagi, F. Soeda, K. Hatada, S. Nagaoka, J. Suzuki and A. Ishitani, J. Polym. Sci. Part A: Polym. Chem., 1988,26, 559. 11 D. S. Everhart and C. N. Reilley, Anal. Chem., 1981,53,665. 12 D. Briggs and C. R. Kendall, Int. J.Adhesion Adhesives, 1982, 13. J. MATER. CHEM., 1994, VOL. 4 687 13 A. Chilkoti, B. D. Ratner and D. Bnggs Chem. Muter., 1991, 3, 18 R. F. Reilman, A. Msezane and S. T. Manson, J. Electron 14 51. E. Sheng, Ph.D. Thesis, Loughborough University of Technology, 1992. 19 Spectrosc. Relat. Phenom., 1976,8, 389. I. Sutherland, D. M. Brewis, R. J. Heath and E. Sheng, Surf. Interface Anal., 1990, 17, 507. 15 16 17 J. H. Scofield, J. Electron Spectrosc. Relat. Phenom., 1976,8, 129. M. P. Seah and W. A. Dench, Surf. Interface Anal., 1979,1,2. M. P. Seah, Surf. Interface Anal., 1980,2,222. Paper 3/07050F; Received 29th Novemher, 1993