Surface Modification of an Ethylene−Acrylic Acid Copolymer Film

Jul 16, 2004 - These reaction steps were then repeated to create the linear and branched architectures on the EAA film surface. The film surface resul...
0 downloads 8 Views 118KB Size
Download PDF
Langmuir 2004, 20, 7151-7158

7151

Surface Modification of an Ethylene-Acrylic Acid Copolymer Film: Grafting Amine-Terminated Linear and Branched Architectures Amol V. Janorkar, Ning Luo, and Douglas E. Hirt* Department of Chemical Engineering and Center for Advanced Engineering Fibers & Films, Clemson University, Clemson, South Carolina 29634-0909 Received February 2, 2004. In Final Form: June 7, 2004 Polymer films can be tailored for a specific application by modifying their surface properties. In this study, linear and branched architectures were grafted to ethylene-acrylic acid (EAA) copolymer films using the so-called grafting from approach. Dicyclohexylcarbodiimide was used to activate the carboxylic acid functionality on the surface of the EAA copolymer film before reacting it with selected di- and triamine compounds. The carboxylic acid functionality was subsequently regenerated by reacting the aminegrafted film with succinic anhydride. These reaction steps were then repeated to create the linear and branched architectures on the EAA film surface. The film surface resulting from each reaction step was analyzed using attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy and contact angle measurements. A systematic analysis of the ATR-FTIR results was performed to estimate the average conversion of the reaction schemes and to explain the observed contact angle results. A significant reduction in water contact angle for the EAA film grafted with a branched architecture was observed. The EAA film grafted with a linear architecture showed a marginal reduction in water contact angle when ethanol was used as a solvent for ethylenediamine. When the solvent for ethylenediamine was changed to water, the contact angle decreased noticeably. However, analysis of control films showed that the reduction in the contact angles was due to the solvent treatment. In the case of branched architectures, such reduction in contact angle due to the solvent treatment was not observed. Several control experiments were performed to ensure that the reduction in the contact angles was in fact due to the grafted species and not due to exposure to various solvents used in the reaction scheme.

Introduction Polymer films possessing acceptable physical properties can be tailored for specific applications by modification of their surfaces without affecting their bulk properties. Several nonselective techniques for polymer film surface modification have been described in the literature. These techniques mainly include plasma treatments,1-4 corona discharge,5-7 and UV irradiation.8-11 Inclusion of migratory additives in the film bulk and covalent attachment of molecules to the polymer film surface present selective ways to tailor film properties by introducing specific functional groups. Use of migratory additives, however, tends to be temporary because these compounds can potentially be removed from the polymer film surface due to friction, heat, or solvent treatment.12,13 Covalent attachment of polymer brushes, macromolecules such as (1) Hasirci, V.; Tezcaner, A.; Hasirci, N.; Suzer, S. J. Appl. Polym. Sci. 2002, 87, 1285-1289. (2) Grace, J. M.; Zhuang, H. K.; Gerenser, L. J.; Freeman, D. R. J. Vac. Sci. Technol., A 2003, 21, 37-46. (3) Riccardi, C.; Barni, R.; Selli, E.; Mazzone, G.; Massafra, M. R.; Marcandalli, B.; Poletti, G. Appl. Surf. Sci. 2003, 211, 386-397. (4) Oiseth, S. K.; Krozer, A.; Kasemo, B.; Lausmaa, J. Appl. Surf. Sci. 2002, 202, 92-103. (5) Foldes, E.; Toth, A.; Kalman, E.; Fekete, E.; Tomasovszky, A. J. Appl. Polym. Sci. 2000, 76, 1529-1541. (6) Suzer, S.; Argun, A.; Vatansever, O.; Aral, O. J. Appl. Polym. Sci. 1999, 74, 1846-1850. (7) Sun, Q. C.; Zhang, D. D.; Wadsworth, L. C. Tappi J. 1998, 81, 177-183. (8) Chen, J. X.; Tracy, D.; Zheng, S.; Xiaolu, L.; Brown, S.; VanDerveer, W.; Entenberg, A.; Vukanovic, V.; Takacs, G. A.; Egitto, F. D.; Matienzo, L. J.; Emmi, F. Polym. Degrad. Stab. 2003, 79, 399-404. (9) Athanassiou, A.; Georgiou, S.; Fotakis, C. Appl. Surf. Sci. 2002, 197-198, 757-763. (10) Kaczmarek, H.; Kowalonek, J.; Szalla, A.; Sionkowska, A. Surf. Sci. 2002, 507-510, 883-888. (11) Sionkowska, A. Polym. Degrad. Stab. 2000, 68, 147-151.

hyperbranched molecules, or smaller-sized molecules to film surfaces provides more permanence for surface modification.14-17 In this study, amine (sNH2) groups were chemically grafted to the surface of an ethylene-acrylic acid (EAA) copolymer film. The reaction scheme involved activation of the carboxylic acid functionality (sCOOH) with dicyclohexylcarbodiimide (DCC) and subsequent coupling with selected di- and tri-amine compounds. The reaction chemistry involving activation of carboxyl groups and subsequent coupling with amine groups has been used extensively for peptide synthesis in bulk solution,18,19 binding biomolecules to gold supports,20 grafting amino acids to poly(ethylene terephthalate) film surfaces,21 and grafting hyaluronic acid to polypropylene, polystyrene, and poly(tetrafluoroethylene) film surfaces.22 The resulting amine groups were then reacted with succinic anhydride (SAN) to regenerate the carboxylic acid function(12) Janorkar, A. V.; Hirt, D. E.; Wooster, J. J. Polym. Eng. Sci. 2004, 44, 34-44. (13) Shuler, C. A.; Janorkar, A. V.; Hirt, D. E. Proceedings of the 61st Annual Technical Conference & Exhibition, Nashville, 2003; Society of Plastics Engineers: Brookfield, CT, 2003; Vol. 49, pp 2724-2728. (14) Luzinov, I.; Julthongpiput, D.; Malz, H.; Pointeck, J.; Tsukruk, V. V. Macromolecules 2000, 33, 1043-1048. (15) Tsukruk, V. V.; Luzinov, I.; Julthongpiput, D.; Malz, H.; Pointeck, J. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 1999, 40, 118119. (16) Cohn, D.; Stern, T. Macromolecules 2000, 33, 137-142. (17) Nakayama, Y.; Matsuda, T. Langmuir 1999, 15, 5560-5566. (18) Anderson, G. W.; Zimmerman, J. E.; Callahan, F. M. J. Am. Chem. Soc. 1964, 86, 1839-1842. (19) Yan, H.; Zhao, Q.; Yuan, J.; Cheng, X.; He, B. Biotechnol. Appl. Biochem. 2000, 31, 15-20. (20) Patolsky, F.; Tao, G.; Katz, E.; Willner, I. J. Electroanal. Chem. 1998, 454, 9-13. (21) Marchand-Brynaert, J.; Deldime, M.; Dupont, I.; Dewez, J.; Schneider, Y. J. Colloid Interface Sci. 1995, 173, 236-244.

10.1021/la049715w CCC: $27.50 © 2004 American Chemical Society Published on Web 07/16/2004

7152

Langmuir, Vol. 20, No. 17, 2004

Janorkar et al.

Figure 1. Reaction scheme for generating linear and branched architectures on the EAA film surface. (i) DCC, dioxane, 3 h; (ii) EDA, ethanol or water, 3 h; (iii) SAN, anhydrous DMF, 4 h; (iv) TREN, ethanol, 3 h. Spectra a-d are shown in Figure 2.

alities. This reaction sequence was then repeated to generate a covalently bonded linear or branched structure on the film surface. The EAA film selected for this study provided a limited number of sCOOH groups on the surface that were ineffective in rendering hydrophilic properties to the film. Bruening and co-workers used an approach to graft branched structures of poly(acrylic acid) to gold surfaces to greatly increase the number of carboxyl groups in the bonded layer to improve wettability as well as other properties.23 Our strategy used smaller molecules to produce branched architectures that would provide multiple sNH2 or sCOOH groups from each sCOOH group on the polymer-film surface. These groups can also then be used as centers for further reactions if desired. As shown in Figure 1, the linear architecture was generated using ethylenediamine (EDA) while tris(2-aminoethyl)amine (TREN) was used to generate the branched architecture. The resulting film surface after each reaction step was analyzed by attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy. Static water contact angle measurements were performed to study the effect of grafting on the hydrophilicity of the film surface. Several control experiments were performed to ensure that the reduction in contact angle was due to the grafted species and not due to the solvent treatment. Experimental Section Materials. EAA copolymer with a composition of 9.7 mol % acrylic acid (PRIMACOR 1410 from The Dow Chemical Co.) was cast-extruded into a film with a nominal thickness of 75 µm at (22) Mason, M.; Vercruysse, K. P.; Kirker, K. P.; Frisch, R.; Marecak, D. M.; Prestwich, G. D.; Pitt, W. G. Biomaterials 2000, 21, 31-36. (23) Bruening, M. L.; Zhou, Y.; Aguilar, G.; Agee, R.; Bergbreiter, D. E.; Crooks, R. M. Langmuir 1997, 13, 770-778.

the Cryovac Division of Sealed Air Corp. (Duncan, SC). DCC, EDA, SAN, TREN, ethanol, dioxane, and anhydrous N,Ndimethylformamide (DMF) were purchased from Aldrich Chemicals and used as received. Experiments Conducted. (i) To synthesize various “generations” of surface-grafted layers, EDA and TREN were used to form linear and branched architectures on the EAA surface, respectively. Table 1 shows the sequence of reaction steps used for the linear architecture (for the reaction sequence to produce a branched architecture, replace EDA with TREN in Table 1). (ii) Several control films were prepared by following the same reaction steps except eliminating the use of reactants in the final step. For example, a control film for step 6 was formed by processing a film through steps 1-5 and then stirring it in anhydrous DMF (solvent for step 6) without adding SAN (reactant for step 6). (iii) To investigate the effect of solvents in the reaction scheme, the linear architecture was also generated by replacing ethanol with water as a solvent for EDA. (iv) A solvent-treated EAA film was obtained by exposing a piece of neat EAA film to all of the solvents in the scheme without using any of the reactants. Reaction Procedures. Activation of Carboxylic Acid Functionality. The film with carboxylic acid functionality on the surface (e.g., EAA film or film resulting from steps 3 and 6) was stirred in a 4% (w/w) solution of DCC in dioxane for 3 h at room temperature. The film was then washed with pure dioxane to remove any excess DCC. The film was allowed to stand at room temperature for another 1 h to ensure evaporation of excess dioxane. Grafting of Amine Compounds to the Activated Film. The DCCactivated film was stirred in an 8% (v/v) solution of a prescribed amine for 3 h at room temperature. Ethanol was used as a solvent for EDA and TREN. Water was also used as a solvent for EDA to investigate the effect of solvents in the reaction scheme. The film was washed with the respective pure solvent to remove any excess amine. The film was then allowed to stand at room temperature for 24 h to ensure evaporation of excess solvent.

Surface Modification of an EAA Copolymer Film

Langmuir, Vol. 20, No. 17, 2004 7153

Table 1. Reaction Sequence for Producing a Third-Generation Linear Architecture on the EAA Film Surfacea step

film before reaction

film after reaction

reactant

solvent

reaction time (h)

1 2 3 4 5 6 7 8

EAA EAA-DCC EAA-EDA EAA-EDA-SA EAA-EDA-SA-DCC EAA-EDA-SA-EDA EAA-EDA-SA-EDA-SA EAA-EDA-SA-EDA-SA-DCC

EAA-DCC EAA-EDA EAA-EDA-SA EAA-EDA-SA-DCC EAA-EDA-SA-EDA EAA-EDA-SA-EDA-SA EAA-EDA-SA-EDA-SA-DCC EAA-EDA-SA-EDA-SA-EDA

DCC EDA SAN DCC EDA SAN DCC EDA

dioxane ethanol anhydrous DMF dioxane ethanol anhydrous DMF dioxane ethanol

3 3 4 3 3 4 3 3

a Key: DCC ) dicyclohexylcarbodiimide; EDA ) ethylenediamine; SAN ) succinic anhydride; SA ) succinic acid. For the reaction sequence to generate a branched architecture on the EAA film surface, replace EDA with TREN.

Figure 2. Representative ATR-FTIR spectra for the (a) neat EAA film, (b) film with activated sCOOH functionality, (c) film grafted with the first-generation branched architecture, and (d) film with regenerated sCOOH functionality. Spectrum a shows the “acid peak” at 1710 cm-1 (b). Spectrum b shows a decrease in the acid peak. The characteristic amide-I peak at 1645 cm-1 (9) appeared in spectrum c. An increase in both the acid and amide-I peaks was observed in spectrum d. Regeneration of Carboxylic Acid Functionality. The aminegrafted film was stirred in a 4% (w/w) solution of SAN in anhydrous DMF for 4 h at room temperature. The film was washed with pure solvent to remove any excess SAN. The film was then allowed to stand at room temperature for 24 h to ensure evaporation of excess solvent. These steps were repeated to achieve the linear and branched architectures. Three or five generations of these architectures were attempted. Analytical Techniques. ATR-FTIR spectroscopy was used to monitor the grafting reactions on a film surface. ATR uses an evanescent field, which decays exponentially into the film, so the absorbance spectrum corresponds to a penetration depth that depends on the wavelength, angle of incidence, and the refractive indices of the sample and crystal.24 The variables used in the experiments gave a penetration depth of approximately 0.4 µm. A Nicolet Avatar 360 with a horizontal, multibounce ATR attachment was used. The peaks at wavenumbers 1645 and 1710 cm-1, corresponding to the sCdO stretches in the amide (amide-I peak) and carboxylic acid (“acid peak”), respectively, were used for characterization. The preferred way of reporting the ATRFTIR results is in terms of peak area (PA) ratio, that is, PA at 1645 cm-1 or at 1710 cm-1 divided by the PA at 1465 cm-1, corresponding to the sCH2sCH2s stretching vibrations in polyethylene and serving as an internal standard. However, the species used in the reaction scheme (EDA, TREN, and SAN) also contain sCH2sCH2s linkages, and so the 1465 cm-1 peak cannot be used as an internal standard. One is then left to use PA for the band of interest with which to quantify the amount of a chemical species present. The limiting factor is the ability to reliably reproduce the same degree of contact between the (24) Urban, M. W. Attenuated Total Reflectance Spectroscopy of Polymers: Theory and Practice; American Chemical Society: Washington, DC, 1996.

specimen and the crystal, which is important if comparisons between specimens will be made. In this research, an aluminum block was placed on top of the film specimen, and a disk-shaped load cell (Entran Devices) was inserted between the block and the ATR clamp to measure the applied force, which reproducibly provided the same degree of contact between the film specimen and the crystal for all measurements. The limits for PA integration were 1594-1683 cm-1 and 1683-1727 cm-1 for the 1645 and 1710 cm-1 peaks, respectively. All of the reported PA values are an average of at least three readings with (95% confidence intervals. Contact angle measurements were performed on a Kruss G10 static contact angle apparatus. The data were analyzed by calculating the water contact angles using the sessile drop method. All of the reported contact angle values are an average of 10 readings with (95% confidence intervals.

Results and Discussion As discussed earlier, EDA was chosen to form a linear architecture on the EAA surface. TREN was chosen to generate branched structures because it had the same chain length (sCH2sCH2s) in each of its arms as that for EDA and had three equivalent terminal amine groups (see Figure 1). Initially, three sequences of reaction steps, or generations, were attempted for both linear and branched architectures. Analysis of ATR-FTIR Results. Figure 2 shows the representative ATR-FTIR spectra for the neat EAA film (a), the film with activated sCOOH functionality (b), the film grafted with the first-generation branched architecture (c), and the film with regenerated sCOOH functionality (d). Spectrum (a) for the neat EAA film showed

7154

Langmuir, Vol. 20, No. 17, 2004

Janorkar et al.

Figure 3. PA of the (a) peak at 1645 cm-1 and (b) peak at 1710 cm-1 obtained for the EAA film grafted with the linear (gray bars) and branched (black bars) architectures. The error bars represent 95% confidence intervals.

the “acid peak” at 1710 cm-1. Spectrum (b) for the film with the activated sCOOH functionality showed a decrease in the acid peak. Additionally, spectrum b shows peaks at 1818 and 1745 cm-1 that may represent the Cd O stretching vibrations of anhydride, which is formed as an intermediate during the activation step, especially in the absence of a nucleophilic species (e.g., an amine).21,25-27 Because spectrum b was collected 24 h after the activation step to ensure evaporation of the excess solvent (dioxane), it is likely that the activated acids were converted to anhydrides. To minimize anhydride formation and briefly allow dioxane evaporation, the subsequent amine-grafting step was performed a relatively short time (1 h) after the activation step. The characteristic amide-I peak at 1645 cm-1 appeared in spectrum c for the film grafted with the first-generation branched architecture. Spectrum c also shows an amide-II peak at 1550 cm-1, which is the combination band of NsH bending and CsN stretching vibrations. The peak at 1550 cm-1 may also contain CdO stretching vibrations from the ionized carboxylic acid groups (COO-) that arise from the interaction of unreacted sCOOH groups with the amine compounds. An increase in both the acid and amide-I peaks was observed in spectrum d for the film with regenerated sCOOH functionality. All of the observations were expected on the basis of the reaction scheme in Figure 1. Figure 3a summarizes the PA results for the peak at 1645 cm-1 for both linear and branched architectures when ethanol was used as a solvent for both EDA and TREN. It shows that the PA at 1645 cm-1 increased for both architectures as more amide linkages were added with each successive coupling step. The increase in PA was higher for the branched structure as expected because it should have a higher number of amide linkages per generation. There was no change in PA for the peak at 1645 cm-1 after the DCC-activation steps 1, 4, and 7 (results not shown in the figure), because it did not result in any amide coupling. The PAs for steps 6 and 8 for the branched architecture did not show any significant change. These results indicated that, in the case of the branched architecture, the conversion of the reaction decreased with each successive generation and a modified method would be required for achieving a higher number of generations. Figure 3b shows the PA for the peak at 1710 cm-1, that is, the “acid peak”. The increase in the PA for the peak at 1710 cm-1 for steps 3 and 6 suggests regeneration of the carboxyl functionality after reacting the amine-grafted EAA film with SAN, as expected. However, undue (25) Lloyd, D. R.; Burns, C. M. J. Polym. Sci., Polym. Chem. Ed. 1979, 17, 3473-3483. (26) Mikolajczyk, M.; Kielbasinski, P. Tetrahedron 1981, 37, 233284. (27) Yan, L.; Marzolin, C.; Terfort, A.; Whitesides, G. M. Langmuir 1997, 13, 6704-6712.

importance should not be given to the changes in the acid peak because it is likely to be influenced by the formation of ionized carboxylic acid groups (COO-). The ATR-FTIR results were studied in more detail to further evaluate the formation of the linear and branched architectures on the surface. We employed a modeling approach where the PAs for the acid and amide peaks were predicted through the stoichiometry of the reaction steps shown in Table 1. The steps involved in the modeling are as follows: 1. The PA for a particular peak produced by a specific functional group (e.g., the amide peak at 1645 cm-1) is dependent on the quantity of the functional group analyzed.

(PA)i ) f(ni)

(1)

where ni is the number of functional groups of species i. A calibration curve is, therefore, necessary to directly convert the observed PA to the number of functional groups. Gaboury and Urban grafted acrylamide and maleic anhydride on a silicone elastomer surface using a microwave-plasma-induced grafting procedure. To quantify the amount of the grafted species, a calibration curve was developed by analyzing several known-concentration solutions of acrylamide and maleic anhydride with transmission-FTIR spectroscopy. The calibration curve generated in this manner was linear, that is, IR absorbance for the sCdO peak was proportional to concentration.28 In our research, a calibration curve for each of the chemical species would be needed if the number of functional groups formed after each reaction step were to be quantified; however, it is not feasible to generate such calibration curves because the separation of various reacting species after each generation is difficult. We, therefore, assumed that a generic linear relationship between the number of functional groups analyzed by ATR-FTIR and the respective PA would be valid in the present study, namely,

(PA)i ≈ kini

(2)

where ki is a proportionality constant. To estimate the values of the proportionality constants for the sCOOH and amide peaks, myristic acid [CH3s(CH2)12sCOOH] and erucamide [cissCH3(CH2)7CHdCH(CH2)11CONH2] were deposited on a polyethylene film surface. Though a primary amide compared to the secondary amides formed during the reaction scheme, erucamide was used for the calibration because it showed a distinct amide-I band at 1645 cm-1, similar to the secondary amides. ATR-FTIR analysis performed on these films showed that a linear relationship existed between the number of functional (28) Gaboury, S. R.; Urban, M. W. Langmuir 1994, 10, 2289-2293.

Surface Modification of an EAA Copolymer Film

Langmuir, Vol. 20, No. 17, 2004 7155

groups deposited and the respective PA, namely,

(PA)COOH ) 1.2 × 10-18‚nCOOH (PA)Amide ) 5.7 × 10

-18

‚nAmide

(3) (4)

2. At this point, we attempt to estimate an average conversion for these reaction schemes. The average conversion (R) was defined as the fraction of the activated sCOOH (for steps 2, 5, and 8) or sNH2 (for steps 3 and 6) groups that underwent the grafting reactions. However, reactions of two sNH2 groups of the same EDA (or TREN) molecule with neighboring activated sCOOH groups or reaction of the SAN molecule with neighboring sNH2 groups could occur. These reactions were referred to as “bridging”. Such bridging reactions can reduce the overall reaction conversion and hinder formation of successive generations. A bridging factor (β) was defined as the fraction of the activated sCOOH (or sNH2) functionality available for the grafting reactions that underwent bridging. 3. The number of sCOOH, sNH2, and amide groups present after an amine grafting step (i.e., steps 2, 5, and 8) could, therefore, be calculated as

(nCOOH)current ) (nCOOH)previous - (nCOOH)reaction (5)

Table 2. Model Calculations for ATR-FTIR PA Results for the EAA Film Grafted with Linear and Branched Architecturesa Linear Architecture average conversion

bridging factor

depth of reaction zone

1.00

0.13

130 nm

experiment step

PA @ 1645

EAA 2 3 5 6 8

0.00 0.69 0.86 0.96 1.14 1.29

(nNH2)bridging (6) (nAmide)current ) (nAmide)previous + (nAmide)reaction (7) where (nCOOH)current, (nNH2)current, and (nAmide)current ) number of sCOOH, sNH2, and amide groups present after an amine grafting step; (nCOOH)previous, (nNH2)previous, and (nAmide)previous ) number of sCOOH, sNH2, and amide groups already present before an amine grafting step; (nCOOH)reaction ) number of activated sCOOH groups reacted with either EDA or TREN; (nNH2)reaction and (nAmide)reaction ) number of sNH2 and amide groups generated due to reaction of the activated sCOOH with either EDA or TREN; and (nNH2)bridging ) number of sNH2 groups that could not be generated due to bridging reactions. From the definitions of R and β, and the reaction stoichiometry, expressions for various terms in equations 5-7 in the case of the linear architecture could be written as

(nCOOH)reaction, (nNH2)reaction, and (nAmide)reaction ) R(nCOOH)previous (8) (nNH2)bridging ) β(nCOOH)previous

(9)

The bridging term appears only while calculating the number of sNH2 groups, that is, the functionality resulting from the amine grafting step, and does not appear while calculating the number of sCOOH groups, that is, the reacting functionality. This follows directly from the definitions of average conversion and bridging factor. 4. Similarly, the number of sCOOH, sNH2, and amide groups present on the surface after a sCOOH regeneration

nAmide

PA @ 1645 cm-1

nNH2

0.00 0.10 0.19 0.26 0.33 0.39

0.00 0.34 0.64 0.90 1.13 1.33

0.00 0.09 0.00 0.07 0.00 0.05

Branched Architecture average conversion

bridging factor

depth of reaction zone

1.00

0.31

130 nm

experiment

model results

step

PA @ 1645 cm-1

nAmide

PA @ 1645 cm-1

nNH2

EAA 2 3 5 6 8

0.00 0.48 0.86 1.26 1.64 1.71

0.00 0.10 0.25 0.36 0.52 0.63

0.00 0.34 0.87 1.24 1.80 2.18

0.00 0.15 0.00 0.16 0.00 0.17

a

(nNH2)current ) (nNH2)previous + (nNH2)reaction -

model results

cm-1

The quantities of nAmide and nNH2 are in units of micromoles.

step (i.e., steps 3 and 6) were calculated by

(nCOOH)current ) (nCOOH)previous + (nCOOH)reaction (nCOOH)bridging (10) (nNH2)current ) (nNH2)previous - (nNH2)reaction (nAmide)current ) (nAmide)previous + (nAmide)reaction

(11) (12)

The expressions for various terms in equations 10-12 in the case of the linear architecture could then be written as

(nCOOH)reaction, (nNH2)reaction, and (nAmide)reaction ) R(nNH2)previous (13) (nCOOH)bridging ) β(nNH2)previous

(14)

Similar equations were derived for the branched architecture after considering the stoichiometry. The number of amide groups calculated in such a manner are shown in column 3 in Table 2. Equation 4 was used to convert the calculated number of amide groups to the corresponding PAs (column 4 in Table 2). 5. The above equations were formulated to provide further insight into the ATR-FTIR data. The analysis may be more complicated if the reactions do not occur solely on the EAA film surface because the reacting molecules, namely, EDA, TREN, and SAN, may penetrate into the near-surface region of the film.29 This leads to three possibilities for the ATR-FTIR analysis: (1) the reaction depth, dR, is less than the depth of penetration for the ATR analysis, dATR () 400 nm); (2) dR ) dATR; and (3) dR > dATR. It can be expected that the bulky DCC molecules (29) Zhang, P.; He, C.; Craven, R. D.; Evans, J. A.; Fawcett, N. C.; Wu, Y.; Timmons, R. B. Macromolecules 1999, 32, 2149-2155.

7156

Langmuir, Vol. 20, No. 17, 2004

Janorkar et al.

Figure 4. Model fit for PA at 1645 cm-1 for the EAA film grafted with the (a) linear and (b) branched architecture. Experimental data are indicated by the symbols, and the model results are shown by the solid line.

would penetrate to a lesser extent into the EAA film compared to the smaller EDA, TREN, and SAN molecules, thereby limiting the reactions to the near-surface region. Therefore, it would be inaccurate to estimate the reaction conversion on the basis of the number of sCOOH groups present, because in case 1 the ATR-FTIR results would be skewed by “seeing” sCOOH groups that are inherent to EAA beyond the reaction zone (dR < d < 400 nm). Additionally, the nonactivated sCOOH groups present beyond the reaction zone could be ionized as a result of interaction with the amine compounds. To develop a consistent method for estimating the average conversion of the reaction scheme, the amide functionality was chosen as a basis for analysis because the amide linkages were formed only as a result of the reaction scheme and were not present in the neat EAA film. 6. For the model calculations, the depth of penetration for the ATR analysis (dATR ) 400 nm) was divided into 10-nm sections and the amide PAs for each reaction step were calculated for each section by varying R and β. For each cumulative reaction depth (dR), the best-fit R and β were then determined by performing a least-squares fit between the model predictions and experimental data. The formation of the sCOO- was not explicitly accounted for in the model as a separate term because the majority of these functionalities were expected to be beyond the reaction zone and the ones present within the reaction zone would be accounted for by the average conversion term. To simplify the model, R and β were assumed to be constant throughout the entire reaction scheme. In reality, the conversion and bridging factor may have different values for each reaction step. However, incorporating too many adjustable parameters would have allowed us to easily obtain a perfect fit to the experimental data. Therefore, we limited the number of adjustable parameters to three (R, β, and dR) and proceeded with the least-squares fit. Figure 4 shows that reasonably good fits to the experimental data were determined for the linear and branched architectures. The model parameters as well as the model results are summarized in Table 2. For both architectures, the best fit was achieved for a reactionzone depth (dR) of 130 nm and R ) 1, while the β (bridging factor) values were 0.13 and 0.31 for the linear and branched architectures, respectively. The dR value suggested that the reactions were confined to a near-surface region of the EAA film, which seems reasonable considering that the bulky DCC molecules may penetrate the film to only a limited extent to activate the sCOOH groups. The β values indicated that the branched architecture showed a greater propensity for bridging. This predicted

behavior is not unexpected because branching with each successive generation would increase the number of reactive groups and bring them in closer proximity, leading to an increased possibility of bridging. A value of R ) 1 means that every available functional group within the reaction zone reacted in some way, that is, sCOOHs were activated by DCC, the activated groups were reacted with EDA or TREN, amine groups were reacted with SAN, or there were bridging reactions. Although it is expected that a large fraction of the functional groups would react, it is unlikely that every available functional group reacted. However, further analysis of the model parameters revealed that the model is not extremely sensitive to high values of R. For example, at given values of dR and β, the results corresponding to R values ranging from 0.85 to 1 are nearly superimposable in Figure 4, whereas low values of R yield significantly different results. The important point is that the predictions matched with experimental data reasonably well, even with a relatively simple model. Analysis of Contact Angle Measurements. Figure 5 shows the contact angle results for the three generations of the linear and branched architectures. The contact angle for the neat EAA film was 97 ( 1°. The black bars represent the grafted films while the gray bars represent the control films, which were formed by following the same reaction steps but eliminating the reactant in the final step. These results show that the EAA film surface-grafted with the amine compounds led to a successive decrease in contact angle with each generation. For the linear architecture, Figure 5a shows that the contact angle of the EAA film decreased up to the second generation but showed no improvement in the third generation. At this point, it should be appreciated that the contact angle of the EAA film grafted with linear or branched architectures is a complex function of the collective number of sNH2 groups, sCOOH groups, and amide linkages formed on the surface for a particular generation. In the case of the linear architecture, the model results predicted an increase in the number of amide linkages but a decrease in the number of sNH2 groups (due to bridging) with each successive generation (compare steps 2, 5, and 8 in columns 3 and 5 in Table 2); consequently, the contact angle of the EAA film grafted with the linear architecture showed an initial small decrease and a plateau. In addition, it can be seen in Figure 5a that the control films have similar contact angle values compared to those obtained by grafting three generations of EDA. This suggested that the solvent (ethanol) treatment might be causing some contact angle changes. Effects of solvent treatment will be discussed later. Overall, the EAA film grafted with the third generation of EDA (contact angle of 86 ( 1°) showed a decrease of about 10° in the

Surface Modification of an EAA Copolymer Film

Langmuir, Vol. 20, No. 17, 2004 7157

Figure 5. Water contact angle measured for the EAA film surface-grafted with the (a) linear and (b) branched architecture (gray bars, control film; black bars, grafted film). The error bars represent 95% confidence intervals. Generations 1-3 correspond to steps 2, 5, and 8 in Figure 3.

Figure 6. PA of the peak at 1645 cm-1 obtained for the EAA film grafted with the linear architectures using ethanol (gray bars) and water (black bars) as the solvent for EDA. The error bars represent 95% confidence intervals.

Figure 7. Water contact angle measured for the EAA film surface-grafted with linear architecture using water as solvent for EDA (gray bars, control film; black bars, grafted film). The error bars represent 95% confidence intervals.

contact angle from the neat EAA film (97 ( 1°). On the other hand, in the case of the branched architecture, Figure 5b shows that there was no plateau in the contact angle values but instead there was a marked decrease in the contact angle of the EAA film with an increase in the number of generations of the branched architecture, while the contact angles of the control films remained constant. This decrease in the contact angles can be attributed to the increased number of amide linkages and sNH2 groups introduced onto the surface by the branched structure (columns 3 and 5 in Table 2). Effect of Solvents on the Generation of Linear Architectures. For each case of EDA, both ethanol and water were used to investigate the effect of solvent on the generation of the linear architecture. The reaction scheme shown in Table 1 was repeated by replacing ethanol with water as a solvent for EDA. The comparison of the ATRFTIR PA for the peak at 1645 cm-1 for the reaction scheme involving ethanol and water is shown in Figure 6. Clearly, on the basis of the PA results, it can be concluded that both systems gave similar conversions for the reactions irrespective of the solvent used. Like Figure 5a, Figure 7 shows that for the linear architecture the control films had similar contact angle values compared to those obtained by grafting three generations of EDA. This suggested that the linear architecture did not show any significant increase in hydrophilicity over the control films. However, it should be noted that the reaction scheme using water as a solvent for EDA reduced the contact angle to a significantly lower value of about 75° compared to 86-89° when using ethanol (Figure 5a).

Effect of Solvents on the Neat EAA Film. Pieces of neat EAA film were soaked in different solvents, namely, dioxane, ethanol, water, and anhydrous DMF, for the same reaction times for the various steps shown in Table 1. The water contact angle of the solvent-treated film was then measured. The results indicated that the contact angle of the EAA film was not affected significantly when soaked in the organic solvents (water contact angle ≈97° in all cases). The contact angle, however, decreased to about 75° when the EAA film was soaked in water for 3 h. This clearly established that the contact angle reduction observed for the linear architecture generated using ethanol as a solvent for EDA (Figure 5a) was primarily due to the sNH2 groups, sCOOH groups, and amide linkages formed on the surface during the grafting process, whereas the contact angle reduction observed for the linear architecture generated using water as a solvent for EDA (Figure 7) was primarily due to the water treatment. Such a reduction in water contact angle of the EAA film surface upon exposure to water may be due to accentuating amorphous acid-rich domains versus semicrystalline polyethylene domains, thereby increasing the number of sCOOH groups on the EAA film surface.30 It should be emphasized that such an effect of the water treatment occurs predominantly after longer exposure times (i.e., 3 h in this case) and not during the water contact angle measurement where the exposure time was only 30 s. To further strengthen the point that the contact angle reduction achieved for the branched architecture was due to the grafted generations of TREN (i.e., the increased number of sNH2 functionalities on the surface), the neat (30) McEvoy, R. L.; Krause, S.; Wu, P. Polymer 1998, 39, 52235239.

7158

Langmuir, Vol. 20, No. 17, 2004

Figure 8. Dependence of water contact angle on the number of generations of the branched architecture grafted to the EAA film surface ([). Films were sonicated in ethanol for 5 min before measuring the water contact angle. Neat EAA film represents zeroth generation. The error bars represent 95% confidence intervals. The solid line represents the linear best fit for the data with the Y intercept at 98.3°, that is, the contact angle of the neat EAA film sonicated in ethanol for 5 min. It should be noted that sonication in ethanol increased the contact angle of the EAA film by approximately 2°.

EAA film was soaked sequentially in all of the solvents for the scheme shown in Table 1 without using any of the reactants. The results showed that the contact angle of the EAA film did not change significantly (water contact angle ≈95° in all cases), establishing that the decrease in hydrophobicity achieved for the branched architecture was due to the sequential grafting of TREN. In addition to the chemical components present on the EAA film surface, the water contact angle also depends on surface roughness, changes in molecular orientation, and swelling of the near-surface region caused by the solvent exposure. To investigate the effect of solvent exposure on surface roughness, atomic force microscopy (AFM) was performed using a Digital Instruments Nanoscope IIIa in the tapping mode. The reaction scheme shown in Table 1 was extended to produce a fifth-generation branched architecture to investigate the surface properties of the EAA film grafted with higher generations. The rootmean-square roughness (RMSR) was evaluated from AFM images of the selected surface-modified and control films for the fifth-generation branched architecture. The plain EAA film had a RMSR of about 1.1 nm (water contact angle ) 97 ( 1°). The RMSR increased to about 6 nm for

Janorkar et al.

the EAA film grafted with the fifth-generation branched architecture (water contact angle ) 72 ( 1°) while the RMSR was about 6.3 nm for the corresponding control film (water contact angle ) 88 ( 3°). Such comparison of the control film prepared by following the same reaction steps but eliminating the reactant in the final step with the corresponding surface-modified EAA films ensured that the water contact angle of the surface-modified EAA film was indeed the result of the grafted species. Before measuring the contact angle of the EAA films grafted with up to five generations of the branched architectures, the films were sonicated in ethanol for 5 min to remove any adsorbed species, if any, and the results are shown in Figure 8. It should also be noted that the contact angle values obtained for the first three generations without sonication (shown in Figure 5b) did not show any significant differences from their sonicated counterparts in Figure 8. The water contact angle of the EAA film grafted with the branched architecture decreased approximately 5° with each successive generation. Summary Surface modification of the EAA film was attempted by generating linear and branched architectures through covalent grafting. The films were characterized by ATRFTIR spectroscopy and contact angle measurements. The ATR-FTIR results indicated formation of the attempted generations on the EAA film surface. However, the conversion decreased with each successive generation. Both linear and branched architectures showed a reduction in water contact angle with each successive generation. However, the contact angle analysis of the control films showed that a similar reduction could be achieved by solvent treatment alone in the case of films grafted with a linear architecture. The films grafted with a linear architecture did not show any significant change in contact angle from their respective control films even when the solvent for EDA was changed from ethanol to water. This indicated that EDA used to generate the linear architecture was not the primary factor in contact angle reduction. The branched structure obtained using TREN showed a marked decrease in contact angle with each successive generation, and the contact angle reduction was not influenced by solvent exposure. LA049715W