Molecular Depth Profiling of Multilayer Polymer Films Using Time-of

Dec 30, 2004 - Films Using Time-of-Flight Secondary Ion Mass. Spectrometry ... The low penetration depth and high sputter rates obtained using polyato...
0 downloads 0 Views 1MB Size
Download PDF
Anal. Chem. 2005, 77, 911-922

Molecular Depth Profiling of Multilayer Polymer Films Using Time-of-Flight Secondary Ion Mass Spectrometry M. S. Wagner†

Surface and Microanalysis Science Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20889-8371

The low penetration depth and high sputter rates obtained using polyatomic primary ions have facilitated their use for the molecular depth profiling of some spin-cast polymer films by secondary ion mass spectrometry (SIMS). In this study, dual-beam time-of-flight (TOF) SIMS (sputter ion, 5 keV SF5+; analysis ion, 10 keV Ar+) was used to depth profile spin-cast multilayers of poly(methyl methacrylate) (PMMA), poly(2-hydroxyethyl methacrylate) (PHEMA), and trifluoroacetic anhydride-derivatized poly(2-hydroxyethyl methacrylate) (TFAA-PHEMA) on silicon substrates. Characteristic positive and negative secondary ions were monitored as a function of depth using SF5+ primary ion doses necessary to sputter through the polymer layer and uncover the silicon substrate (>5 × 1014 ions/cm2). The sputter rates of the polymers in the multilayers were typically less than for corresponding single-layer films, and the order of the polymers in the multilayer affected the sputter rates of the polymers. Multilayer samples with PHEMA as the outermost layer resulted in lowered sputter rates for the underlying polymer layer due to increased ion-induced damage accumulation rates in PHEMA. Additionally, the presence of a PMMA or PHEMA overlayer significantly decreased the sputter rate of TFAA-PHEMA underlayers due to ioninduced damage accumulation in the overlayer. Typical interface widths between adjacent polymer layers were 10-15 nm for bilayer films and increased with depth to ∼35 nm for the trilayer films. The increase in interface width and observations using optical microscopy showed the formation of sputter-induced surface roughness during the depth profiles of the trilayer polymer films. This study shows that polyatomic primary ions can be used for the molecular depth profiling of some multilayer polymer films and presents new opportunities for the analysis of thin organic films using TOF-SIMS. Measurement of the composition of thin organic films as a function of depth (“depth profiling”) is possible using several different surface analytical techniques. For example, X-ray photoelectron spectroscopy (XPS) can nondestructively depth profile † Present address: Procter & Gamble Co., 11810 E. Miami River Rd., Cincinnati, OH 45252. Phone: (513) 627-0682. Fax: (513) 627-1233. E-mail: [email protected].

10.1021/ac048945c Not subject to U.S. Copyright. Publ. 2005 Am. Chem. Soc.

Published on Web 12/30/2004

over the outermost 5-10 nm of polymer films with subnanometer depth resolution,1 while optical microscopy techniques based on IR and Raman spectroscopies can profile much deeper into samples (many micrometers) at the expense of decreased depth resolution (∼2 µm).2,3 Incorporation of Ar+ sputtering with XPS has also been used to extend the depth range of XPS depth profiling of polymers; however, ion-induced damage quickly reduced the useful molecular information obtainable from the polymer surface.4 Secondary ion mass spectrometry (SIMS) has also been used to depth profile polymer films over the depth range of 10 nm to several micrometers. SIMS depth profiles of polymer films have been obtained using monatomic (e.g., Ar+ and Cs+) and diatomic (e.g., O2+) primary ions for a variety of applications including the surface and interfacial segregation of polymer blends,5-12 the distribution of ions,13,14 metals,15,16 polymer additives,17 and organic dyes18 in polymer surfaces, and the analysis (1) Ratner, B. D.; Castner, D. G. In Surface AnalysissThe Principal Techniques; Vickerman, J. C., Ed.; John Wiley & Sons: Chichester, U.K., 1997; p 43. (2) Fina, L. J. Appl. Spectrosc. Rev. 1994, 29, 309. (3) Sacristan, J.; Reinecke, H.; Mijangos, C.; Spells, S.; Yarwood, J. Macromol. Chem. Phys. 2002, 203, 678. (4) MacKay, S. G.; Bakir, M.; Musselman, I. H.; Meyer, T. J.; Linton, R. W. Anal. Chem. 1991, 63, 60. (5) Chujo, R. Polym. J. 1991, 23, 367. (6) Schwarz, S. A.; Wilkens, B. J.; Pudensi, M. A.; Rafailovich, M. H.; Sokolov, J.; Zhao, X.; Zhao, W.; Zheng, X.; Russell, T. P.; Jones, R. A. L. Mol. Phys. 1992, 76, 937. (7) Strzhernechny, Y. M.; Schwarz, S. A.; Schachter, J.; Rafailovich, M. H.; Sokolov, J. J. Vac. Sci. Technol., A 1997, 15, 894. (8) Rysz, J.; Ermer, H.; Budkowski, A.; Lekka, M.; Bernasik, A.; Wrobel, S.; Brenn, R.; Lekki, J.; Jedlinski, J. Vacuum 1999, 54, 303. (9) Yokoyama, H.; Kramer, E. J.; Hajduk, D. A.; Bates, F. S. Macromolecules 1999, 32, 3352. (10) Duan, Y.; Pearce, E. M.; Kwei, T. K.; Hu, X.; Rafailovich, M.; Sokolov, J.; Zhou, K.; Schwarz, S. Macromolecules 2001, 34, 6761. (11) Huang, W. Y.; Matsuoka, S.; Kwei, T. K.; Okamoto, Y.; Hu, X.; Rafailovich, M. H.; Sokolov, J. Macromolecules 2001, 34, 7809. (12) Lin, H. C.; Tsai, I. F.; Yang, A. C. M.; Hsu, M. S.; Ling, Y. C. Macromolecules 2003, 36, 2464. (13) Gritsch, M.; Hutter, H.; Holzer, L.; Tasch, S. Mikrochim. Acta 2000, 135, 131. (14) Mattsson, J.; Forrest, J. A.; Krozer, A.; Sodervall, U.; Wennerberg, A.; Torell, L. M. Electrochim. Acta 2000, 45, 1453. (15) Gray, K. H.; Gould, S.; Leasure, R. M.; Musselman, I. H.; Lee, J. J.; Meyer, T. J.; Linton, R. W. J. Vac. Sci. Technol., A 1992, 10, 2679. (16) Chun-Guey, W.; Jiunn-Yih, H.; Shui-Sheng, H. J. Mater. Chem. 2001, 11, 2061. (17) Stein, J.; Leonard, T. M.; Smith, G. A. J. Appl. Polym. Sci. 1991, 42, 2355. (18) Pinto, J. R.; Novak, S. W.; Nicholas, M. J. Phys. Chem. B 1999, 103, 8026.

Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 911

of polymer-metal stacks.19,20 While the molecular ions of some organic compounds (e.g., quaternary ammonium salts) show a saturation of ion-induced damage after extended monatomic bombardment at low impact energy and glancing incidence angle,21 depth profiling of polymers using monatomic primary ions has generally been limited to examining monatomic and diatomic negative secondary ions due to ion-induced damage that occurs under monatomic primary ion bombardment.22-27 Furthermore, substantial topography development during sputtering of polymers severely limits the depth resolution obtainable using monatomic primary ions.4,15,17,20 Retention of characteristic molecular information during the depth profiling of organic materials by SIMS requires the removal of the majority of ion-damaged material in the same primary ion impact that created it. Polyatomic primary ion bombardment results in low damage accumulation and sustained molecular secondary ion generation during the depth profiling of some organic small molecules,28,29 small molecules in polymer matrixes,28,30 and polymers.28,30-37 Polyatomic primary ions have also improved depth resolution during the depth profiling of shallow dopant implants in silicon.29,38,39 Unlike monatomic primary ions, polyatomic primary ions disperse their impact energy among their constituent atoms, resulting in lower penetration depths and increased deposition of their impact energy in the near-surface region.28,33,40 These factors result in much higher sputter rates for polymers under polyatomic primary ion bombardment than for monatomic primary ion bombardment,32,34 removing much of the damaged material as it is created. Optimally, sustained secondary ion intensities characteristic of the polymer (typically fragments of the monomer) are observed as a function of depth, creating a “molecular depth profile”. The composition of the polymer plays a critical role in the sputter rate and sustained (19) Sauer, G.; Kilo, M.; Hund, M.; Wokaun, A.; Karg, S.; Meier, M.; Reib, W.; Schwoerer, M.; Suzuki, H.; Simmerer, J.; Meyer, H.; Haarer, D. Fresenius’ J. Anal. Chem. 1995, 67, 642. (20) Bulle-Lieuwma, C. W. T.; van Gennip, W. J. H.; van Duren, J. K. J.; Jonkheijm, P.; Janssen, R. A. J.; Niemantsverdriet, J. W. Appl. Surf. Sci. 2003, 203204, 547. (21) Gillen, G.; Simons, D. S.; Williams, P. Anal. Chem. 1990, 62, 2122. (22) Briggs, D.; Hearn, M. J.; Ratner, B. D. Surf. Interface Anal. 1984, 6, 184. (23) Simko, S. J.; Griffis, D. P.; Murray, R. W.; Linton, R. W. Anal. Chem. 1985, 57, 137. (24) Briggs, D.; Hearn, M. J. Vacuum 1986, 36, 1005. (25) Leggett, G. J.; Vickerman, J. C. Appl. Surf. Sci. 1992, 55, 105. (26) Gilmore, I. S.; Seah, M. P. Surf. Interface Anal. 1996, 24, 746. (27) Briggs, D.; Fletcher, I. W. Surf. Interface Anal. 1997, 25, 167. (28) Gillen, G.; Roberson, S. Rapid Commun. Mass Spectrom. 1998, 12, 1303. (29) Gillen, G.; King, L.; Freibaum, B.; Lareau, R.; Bennett, J.; Chmara, F. J. Vac. Sci. Technol., A 2001, 19, 568. (30) Mahoney, C. M.; Roberson, S. V.; Gillen, G. Anal. Chem., in press. (31) Brox, O.; Hellweg, S.; Benninghoven, A. In Proceedings of the 12th International Conference on SIMS; Benninghoven, A., Bertrand, P., Migeon, H. N., Werner, H. W., Eds.; Elsevier: Brussels, 1999; p 777. (32) Fuoco, E. R.; Gillen, G.; Wijesundara, M. B. J.; Wallace, W. E.; Hanley, L. J. Phys. Chem. B 2001, 105, 3950. (33) Weibel, D. E.; Wong, S.; Lockyer, N. P.; Blenkinsopp, P.; Hill, R.; Vickerman, J. C. Anal. Chem. 2003, 75, 1754. (34) Wagner, M. S. Anal. Chem. 2004, 76, 1264. (35) Wagner, M. S. Surf. Interface Anal., in press. (36) Wagner, M. S. Surf. Interface Anal., in press. (37) Wagner, M. S. Surf. Interface Anal., in press. (38) Iltgen, K.; Bendel, C.; Benninghoven, A.; Niehuis, E. J. Vac. Sci. Technol., A 1997, 15, 460. (39) Gillen, G.; Walker, M.; Thompson, P.; Bennett, J. J. Vac. Sci. Technol., B 2000, 18, 503. (40) Appelhans, A. D.; Delmore, J. E. Anal. Chem. 1989, 61, 1087.

912

Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

characteristic secondary ion generation under extended polyatomic primary ion bombardment,31,35-37,41 with polymers that can degrade by depolymerization performing significantly better than those that are highly susceptible to intra- and intermolecular crosslinking.35-37 The present study complements previous studies on the depth profiling of polymers using polyatomic primary ions by extending the sample complexity to multilayer films. Dual-beam TOF-SIMS depth profiling38 was used to characterize four bilayer and two trilayer polymer films fabricated by subsequent spin-casting of the layers onto silicon substrates. Doses of 5 keV SF5+ in the range necessary to profile through the polymer films and uncover the silicon substrate (typically 5 × 1014 to 1 × 1015 ions/cm2) were used. Positive and negative ion molecular depth profiles were obtained and compared on the basis of the stability of the characteristic secondary ion intensities, the sputter rates of the polymers, and the interface widths between the polymer layers and between the polymers and the silicon substrate. MATERIALS AND METHODS Sample Preparation. Poly(methyl methacrylate) (PMMA; MW 101 000) and poly(2-hydroxyethyl methacrylate) (PHEMA; MW 300 000) were purchased from Scientific Polymer Products, Inc. (Ontario, NY)42 and used without further purification. PMMA and PHEMA were dissolved at 20 mg of polymer/g of solvent in toluene and methanol, respectively. Polymer multilayers were fabricated by sequential spin-casting (2000 rpm for 30 s) of the different polymers onto 1 cm × 1 cm silicon wafer pieces that were previously cleaned by sequential ultrasonication in methylene chloride, acetone, and methanol. Vapor-phase derivatization of the spin-cast PHEMA films with trifluoroacetic anhydride (TFAA) was performed following the method described by Chilkoti et al.43-46 Spin-cast PHEMA films on silicon substrates were placed on a glass Petri dish in a reaction vessel. TFAA (99% purity, Aldrich, Milwaukee, WI) was then injected below the samples, the reaction vessel sealed, and the derivatization reaction allowed to proceed at room temperature for 15 min. The samples were then recovered for the spin-casting of additional polymer layers on top of the TFAA-derivatized PHEMA (TFAA-PHEMA) layer. Thickness measurements of the polymer films were made using a model M-2000 spectroscopic ellipsometer (J. A. Woollam Co., Lincoln, NE). Ellipsometry measurements were made after the deposition of each polymer layer and assumed a homogeneous, smooth film. Time-of-Flight Secondary Ion Mass Spectrometry. Positive and negative ion TOF-SIMS depth profiles were obtained in the dual-beam mode38 using an Ion-Tof IV (Ion-Tof GmbH, Muenster, (41) Norrman, K.; Haugshoj, K. B.; Larsen, N. B. J. Phys. Chem. B 2002, 106, 13114. (42) Certain commercial equipment, instruments, or materials are identified in this paper to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. (43) Chilkoti, A.; Castner, D. G.; Ratner, B. D.; Briggs, D. J. Vac. Sci. Technol., A 1990, 8, 2274. (44) Chilkoti, A.; Ratner, B. D.; Briggs, D. Chem. Mater. 1991, 3, 51. (45) Chilkoti, A.; Ratner, B. D. Surf. Interface Anal. 1991, 17, 567. (46) Chilkoti, A.; Lopez, G. P.; Ratner, B. D.; Hearn, M. J.; Briggs, D. Macromolecules 1993, 26, 4825.

Germany) equipped with a 10 keV electron impact Ar+ ion source for analysis (∼2 pA pulsed ion current, incidence angle of 45°, raster area of 200 µm × 200 µm) and a 5 keV electron impact SF5+ ion source for sputtering (∼2 nA static ion current, incidence angle of 45°, raster area of 1250 µm × 1250 µm). Depth profiles of the polymer multilayers were obtained by alternately collecting static TOF-SIMS spectra using the Ar+ ion beam with sputtering using the SF5+ ion beam. While an optimal depth-profiling experiment would utilize SF5+ as both the sputter and analysis ion sources, the current experimental setup was chosen due to the design of our TOF-SIMS instrument. The ion dose for each TOF-SIMS spectrum collected using the Ar+ ion source was maintained below the static limit of 1012 ions/cm2,22,26,27 though the total Ar+ ion dose during a depth profile was as much as 1014 ions/cm2. Each spectrum in the positive ion depth profiles was mass calibrated using the CH3+ (m/z ) 15), C2H3+ (m/z ) 27), and C3H5+ (m/z ) 41) peaks, while each spectrum in the negative ion depth profiles was mass calibrated using the CH- (m/z ) 13), OH- (m/z ) 17), and C2H- (m/z ) 25) peaks. A pulsed low-energy electron flood gun was used for charge neutralization during the acquisition of the positive ion static TOFSIMS spectra. The electron flood gun current was kept below 50 nA to minimize electron-induced sample damage,47 although such damage was not specifically characterized in this study. For some bilayer and both trilayer samples, the electron flood gun was run for an additional time between the SF5+ sputtering step and the analysis step in the depth profile to facilitate charge neutralization. Omission of this extra electron flood gun step resulted in a rapid decrease of the total secondary ion intensity and the quality (e.g., mass resolution and peak positions) of the mass spectra due to sample charging, precluding successful acquisition of the depth profile. Data Treatment. Positive and negative secondary ions that were characteristic of the polymer pendant group were used to monitor the different polymer layers during the depth profile. In the positive ion mode, C2H5O+ (m/z ) 45), C2H3O2+ (m/z ) 59), and C4H4O2F3+ (m/z ) 141) were used for PHEMA, PMMA, and TFAA-PHEMA, respectively. In the negative ion mode, C2H3O2(m/z ) 59), CH3O- (m/z ) 31), and C2O2F3- (m/z ) 113) were used for PHEMA, PMMA, and TFAA-PHEMA, respectively. The proposed structures for these fragments are shown in Table 1, and the structures for some of these fragments have been proposed previously.43,48 Si+ (m/z ) 28) and SiO2- (m/z ) 60) were used to indicate the silicon substrate in the positive and negative ion modes, respectively. In this study, two types of data normalization, spectrum normalization and ion normalization, were used to clearly delineate the different polymer layers in the depth profiles. These normalization methods were primarily used to facilitate the measurement of sputter rates and interface widths from the depth profiles. Spectrum normalization refers to the normalization of each spectrum in the depth profile to its own total secondary ion intensity (in the m/z ) 0-200 or m/z ) 26-200 range for the positive and negative ion modes, respectively). Spectrum normalization was performed for two reasons: (1) Spectrum normalization reduced the rate of the decrease of the characteristic (47) Gilmore, I. S.; Seah, M. P. Appl. Surf. Sci. 2002, 187, 89. (48) Hearn, M. J.; Briggs, D. Surf. Interface Anal. 1988, 11, 198.

Table 1. Proposed Structures of the Pendant-Group-Related Secondary Ions Used To Monitor the Different Polymer Layers in This Study

secondary ion intensities of polymers under extended SF5+ bombardment.34-37 (2) The characteristic secondary ion for PHEMA (C2H5O+ at m/z ) 45) also appeared in the positive ion TOF-SIMS spectra of PMMA and TFAA-PHEMA; however, the relative intensity of this secondary ion was much higher for PHEMA than for PMMA or TFAA-PHEMA. Spectrum normalization of the depth profiles made this distinction clear for the different polymer layers. Ion normalization refers to the scaling of the intensity of the spectrum-normalized characteristic secondary ion for each polymer between 0 and 1. This normalization was performed to facilitate the assignment of depth scales and the measurement of interface widths for the depth profiles. More details on spectrum and ion normalization are given in the Supporting Information. After spectrum normalization and ion normalization, assignment of a depth scale was performed using a weighted sputter rate for the transition between two adjacent layers as described previously.34 The sputter rates of the polymers were determined by assigning a depth scale to the depth profile and solving iteratively for sputter rates of the different polymer layers that resulted in layer thicknesses that agreed with the ellipsometry measurements. The sputter rate of the silicon substrate under 5 keV SF5+ bombardment was independently measured and found to be ∼0.14 nm3/ion (0.0056 Å/(nA-s) under the conditions used in this study). This sputter rate was consistent with previous measurements of the sputter rate of silicon under SF5+ bombardment.39 Interface widths between layers were measured after assignment of a depth scale to the depth profiles. The interface width was defined as the depth required for the characteristic secondary ion intensity of the underlying layer to rise from 16% to 84% of its maximum relative intensity, following the previously defined standard procedure.49 More details on the assignment of depth scales and the measurement of interface widths are given in the Supporting Information. RESULTS AND DISCUSSION Molecular Depth Profiling of Bilayer and Trilayer Polymer Films. Bilayer films of PMMA and PHEMA on silicon substrates (49) Standard Guide for Measuring Widths of Interfaces in Sputter Depth Profiling using SIMS; ASTM Standard Designation E 1438-91; American Society for Testing of Materials: West Conshohocken, PA, reapproved 2001.

Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

913

Figure 1. Comparison of the depth profiles for (a-c) PHEMA on PMMA and (d-f) PMMA on PHEMA bilayer films on silicon substrates. (a) and (d) show the average (n ) 8) ellipsometric thickness measurements of the individual layers ((1 standard deviation). (b) and (e) show representative positive ion depth profiles, while (c) and (f) show representative negative ion depth profiles. The sputter ion beam was 5 keV SF5+, and analysis ion beam was 10 keV Ar+.

were depth profiled to assess the effect of the layer order on the resulting depth profiles. Figure 1 shows representative positive and negative ion depth profiles of the (b, c) PHEMA on PMMA and (e, f) PMMA on PHEMA bilayer films. For both bilayer films, the characteristic secondary ion intensity of the outermost layer followed the same trends as seen for 5 keV SF5+ bombardment of single-layer PMMA or PHEMA films: a surface transient followed by a relative stabilization of the characteristic secondary ion intensity.37 Once the underlying layer was uncovered, the characteristic secondary ion intensity for the outermost layer decreased while that of the underlying layer increased. However, for both the outermost and underlying layers, a decrease in the characteristic secondary ion intensities with increasing SF5+ ion dose was observed. This decrease was accompanied by the appearance of polycyclic aromatic ions (e.g., C9H7+, C10H8+, C11H9+, and C12H8+)25 and other unsaturated hydrocarbons in the positive ion mode. The formation of unsaturated hydrocarbons during SF5+ bombardment on the surface is discussed in more detail below. Although SF5+-induced damage accumulated in the film during depth profiling, the rate of damage accumulation was low enough 914 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

that characteristic secondary ions for the polymer layers were obtained throughout the depth of the film, enabling the acquisition of the molecular depth profile. Once the polymer layers were sputtered through, the characteristic secondary ion intensity for the silicon substrate (Si+ or SiO2-) increased and the characteristic secondary ion intensities for polymers decreased. The low Si+ intensity at the beginning of the positive ion depth profile was probably due to slight surface contamination, which was removed by SF5+ sputtering. Despite the similarity in overall appearance of the depth profiles of the PHEMA on PMMA and PMMA on PHEMA bilayer films, several differences were also present. The decrease of the characteristic secondary ion intensity of the outermost layer during the surface transient was higher when PHEMA was the outermost layer than when PMMA was the outermost layer. This is consistent with previous studies showing that PMMA accumulated less SF5+-induced damage during SF5+ bombardment than PHEMA.35,37 In the negative ion mode depth profile of the PHEMA on PMMA bilayer film, the CH3O- intensity did not rise near the silicon substrate, as was seen in the negative ion depth

Figure 2. Comparison of the depth profiles for (a-c) PHEMA on TFAA-PHEMA and (d-f) PMMA on TFAA-PHEMA bilayer films on silicon substrates. (a) and (d) show the average (n ) 8) ellipsometric thickness measurements of the individual layers ((1 standard deviation). (b) and (e) show representative positive ion depth profiles, while (c) and (f) show representative negative ion depth profiles. The sputter ion beam was 5 keV SF5+, and the analysis ion beam was 10 keV Ar+.

profiles of spin-cast PMMA films on silicon substrates.35 The intensity increase in the depth profiles of the spin-cast PMMA films has been attributed to a substrate-related matrix effect50 and may have been damped in the depth profile of the PHEMA on PMMA bilayer film by the accumulation of SF5+-induced damage during the depth-profiling measurement. However, the CH3Ointensity increased slightly at the interface between the PMMA and PHEMA layers in the depth profile of the PMMA on PHEMA bilayer film, indicating that this matrix effect could be observed at a polymer-polymer interface as well as a polymer-silicon interface. This matrix effect was not observed for the C2H3O2intensity from PHEMA, consistent with previous studies of spincast PHEMA films.37 The differences between bilayer films with PHEMA or PMMA as the outermost layer were further examined using PHEMA on TFAA-PHEMA and PMMA on TFAA-PHEMA bilayer films. Figure 2 shows representative positive and negative ion depth profiles for (b, c) PHEMA on TFAA-PHEMA and (e, f) PMMA (50) Schnieders, A.; Mollers, R.; Benninghoven, A. Surf. Sci. 2001, 471, 170.

on TFAA-PHEMA bilayer films. As with the PMMA-PHEMA bilayer films, the characteristic secondary ion intensities were relatively stable during SF5+ bombardment, with some degree of SF5+-induced damage accumulation observed by the appearance of polycyclic aromatic ions in the positive ion mode (data not shown). The rate of damage accumulation in the polymer layers was low enough that characteristic secondary ions were observed throughout the depth profiles. The negative ion depth profiles of the PHEMA on TFAA-PHEMA and PMMA on TFAA-PHEMA bilayer films (Figure 2c,f, respectively) showed trends similar to those of the positive ion depth profiles. The characteristic positive and negative pendant group-related secondary ions for TFAAPHEMA showed a significantly less decrease in intensity during continued SF5+ bombardment, in agreement with previous observations that TFAA-PHEMA was more resistant to SF5+induced damage than PMMA or PHEMA.37 The C2O2F3- intensity did not increase near the silicon substrate for either bilayer film, although this matrix effect was observed in the negative ion depth profiles of TFAA-PHEMA films.37 This may have been due to Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

915

Figure 3. Comparison of the depth profiles for (a-c) PHEMA on PMMA on TFAA-PHEMA and (d-f) PMMA on PHEMA on TFAA-PHEMA trilayer films on silicon substrates. (a) and (d) show the average (n ) 8) ellipsometric thickness measurements of the individual layers ((1 standard deviation). (b) and (e) show representative positive ion depth profiles, while (c) and (f) show representative negative ion depth profiles. The sputter ion beam was 5 keV SF5+, and the analysis ion beam was 10 keV Ar+.

the accumulation of damaged material near the silicon substrate, damping this matrix effect. In summary, the order of the polymer layers in the bilayer films did not have a significant qualitative impact on the depth profiles. The trends in the characteristic secondary ion intensities were not significantly affected by layer order, and all films exhibited evidence of some accumulation of SF5+-induced damage. The rate of damage accumulation, however, was low enough for characteristic secondary ions for each of the polymer layers to be observed in the depth profiles. Quantitative differences in sputter rate and interface widths were observed and are discussed in more detail below. Trilayer polymer films of PHEMA, PMMA, and TFAAPHEMA were depth profiled to further assess the effects of layer order and sample complexity on the resulting depth profiles. Figure 3 shows representative positive and negative ion depth profiles for (b, c) PHEMA on PMMA on TFAA-PHEMA and (e, 916 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

f) PMMA on PHEMA on TFAA-PHEMA trilayer films. Unlike the bilayer films, the layer order in the trilayer films affected the appearance of the depth profiles. In the positive ion mode, the C2H5O+ and C2H3O2+ secondary ions were clearly visible in the depth profiles of both trilayer films. However, the C4H4O2F3+ intensity in either trilayer film was not as prominent as seen in the PHEMA or PMMA on TFAA-PHEMA bilayer films. The C4H4O2F3+ intensity was also weaker in the PHEMA on PMMA on TFAA-PHEMA trilayer film (Figure 3b) when compared with the PMMA on PHEMA on TFAA-PHEMA trilayer film (Figure 3e). This showed that the order of the PHEMA and PMMA layers affected the resulting depth profile, with the presence of PMMA at the outermost surface facilitating the acquisition of strong secondary ion intensities for each of the polymer layers. As with the bilayer films, polycyclic aromatic ions were observed during the depth profiling of both trilayer films. In fact, polycyclic aromatic ions were still observed at the end of these depth profiles.

Furthermore, the Si+ secondary ion intensity did not rise to the same level as seen in the bilayer depth profiles once the silicon substrate was uncovered, suggesting that an accumulation of SF5+damaged material partially covered the silicon substrate. The negative ion depth profiles of the PHEMA on PMMA on TFAA-PHEMA (Figure 3c) and PMMA on PHEMA on TFAAPHEMA (Figure 3f) trilayer films were slightly different from the positive ion depth profiles. In the negative ion mode, the characteristic secondary ions for the three polymer layers were clearly observed for both trilayer films. However, as with the positive ion mode, the appearance of the SiO2- signal was quite weak once the silicon substrate was “uncovered”. This suggests that the silicon substrate may not have been fully uncovered during this depth profile due to the accumulation of damaged material and sputter-induced topography formation at the bottom of the sputter crater. For all of the multilayers in this study, primary ion-induced damage accumulation adversely impacted the ability to obtain characteristic secondary ions throughout the depth profile. Enhancing the probability for depolymerization or main chain scission during the SF5+-induced degradation of the polymers could improve the retention of characteristic secondary ions during the depth profile.35-37 Previous studies have suggested that heating the sample to its glass transition temperature can significantly increase the rate of depolymerization during ion bombardment of polymers.51,52 However, this method must be carefully applied to avoid polymer mixing in the multilayers during sample heating. Leggett and Vickerman have also shown that neutral atom bombardment (as opposed to ion bombardment) reduced the accumulation of chemical damage;53 the use of an uncharged polyatomic projectile may further minimize the accumulation of primary particle-induced damage during depth profiling. Formation of Unsaturated Hydrocarbons during SF5+ Bombardment. The chemical changes to the polymer surface during SF5+ bombardment were monitored using spectra obtained during the depth profile. Figure 4 shows several spectra from the positive ion depth profile of the PHEMA on PMMA bilayer film at different points in the depth profile. The initial spectrum in this depth profile (before SF5+ sputtering) is consistent with that of a spin-cast PHEMA film. After a SF5+ dose of 1.8 × 1014 ions/cm2 (Figure 4b), the characteristic secondary ion of PHEMA (C2H5O+) was still visible. Also present are several unsaturated hydrocarbon fragments that are indicative of SF5+-induced damage.37 Furthermore, the total secondary ion intensity decreased significantly from that of the initial spectrum. Once the PMMA layer was uncovered (SF5+ ion dose 2.9 × 1014 ions/cm2, Figure 4c), the characteristic secondary ion for PMMA (C2H3O2-) appeared and the C2H5O+ intensity decreased significantly. The total secondary ion intensity continued to decrease with the additional SF5+ bombardment, and all of the unsaturated hydrocarbon ions were still present in this spectrum. Additional SF5+ bombardment of the PMMA layer (SF5+ ion dose 3.5 × 1014 ions/cm2, Figure 4d) did not significantly change the spectrum except to decrease the total secondary ion intensity. Finally, once the silicon substrate was uncovered (SF5+ (51) Fragala, M. E.; Compagnini, G.; Torrisi, L.; Puglisi, O. Nucl. Instrum. Methods Phys. Res., B 1998, 141, 169. (52) Fragala, M. E.; Compagnini, G.; Puglisi, O. J. Mater. Res. 1999, 14, 228. (53) Leggett, G. J.; Vickerman, J. C. Anal. Chem. 1991, 63, 561.

dose 1.2 × 1015 ions/cm2, Figure 4e), the polymer-related and unsaturated hydrocarbon ions disappeared and the silicon-related ions (e.g., Si+ and SiOH+) dominated the spectrum. Also present was the SiF+ ion that resulted from the implantation of fluorine during SF5+ bombardment. The unsaturated hydrocarbon ions were formed during the SF5+ bombardment of all of the bilayer and trilayer films in this study (data not shown), indicating the accumulation of SF5+induced damage during the depth profile. For the trilayer films, unsaturated hydrocarbons were still observed at the bottom of the sputter crater after the characteristic secondary ions of the polymers were not visible in the spectra. However, characteristic secondary ions were obtained throughout the depth profiles, indicating that undamaged polymer was present at the bottom of the sputter crater during sputtering of the polymer layers. In addition to the appearance of unsaturated hydrocarbons during the depth profile, the total secondary ion intensity decreased with increasing SF5+ dose. The effect of this decrease in total secondary ion intensity was minimized by the use of spectrum and ion normalization of the depth profiles. Effect of Data Normalization on Polymer Depth Profiles. As described above, spectrum normalization and ion normalization of the depth profiles were used before the sputter rates or interface widths were measured. Spectrum normalization is the normalization of each spectrum in the depth profile to its own total secondary ion intensity. Figure 5a shows an example of spectrum normalization for the positive ion PHEMA on PMMA depth profile shown in Figure 1b. It has previously been noted34-37 that spectrum normalization reduces the slope of the steady-state region of the depth profiles of single-layer spin-cast polymer films; this was also observed for the bilayer and trilayer films. Figure 5a shows that, after a surface transient, the spectrum-normalized C2H5O+ intensity from PHEMA was constant during sputtering of the PHEMA layer. This is unlike the absolute C2H5O+ intensity, which decreased rapidly during the surface transient and then continued to decrease at a significantly slower rate with increasing SF5+ dose. The discrepancy between the trend in the absolute and relative secondary ion intensities is not well understood. While the continued decrease in the absolute C2H5O+ intensity and the appearance of unsaturated hydrocarbons indicate continual accumulation of SF5+-induced damage, the constant relative C2H5O+ intensity indicates that a steady state had been reached after the surface transient. The decrease in the absolute C2H5O+ intensity may have been due to factors other than increasing SF5+-induced damage accumulation, such as sample charging or drift in the primary ion current of the analysis ion beam. Quantitative analysis using X-ray photoelectron spectroscopy, for example, is necessary to determine if SF5+-induced damage continued to accumulate or reached a steady state during SF5+ bombardment. Spectrum normalization is, however, a convenient method for minimizing the appearance of SF5+-induced damage in the depth profiles and for more clearly delineating the different layers in the depth profile. Since some characteristic secondary ions (in particular, C2H5O+) can appear in more than one polymer, a more clear delineation between polymer layers is desirable before sputter rates or interface widths are measured. For this purpose, scaling of the spectrum-normalized characteristic secondary ion intensities Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

917

Figure 4. Spectra from the positive ion depth profile of the PHEMA on PMMA bilayer film shown in Figure 1a: (a) before SF5+ exposure and after a SF5+ dose of (b) 1.8 × 1014 ions/cm2, (c) 2.9 × 1014 ions/cm2, (d) 3.5 × 1014 ions/cm2, and (e) 1.2 × 1015 ions/cm2. The inset shows the approximate depth into the sample at the different SF5+ doses.

between 0 and 1 (“ion normalization”) was used. Figure 5b shows the positive ion depth profile of the PHEMA on PMMA bilayer film after spectrum normalization and ion normalization. Note that these data are displayed on a linear instead of logarithmic intensity scale. After spectrum and ion normalization, the C2H5O+ intensity decreased during an initial surface transient and then was relatively constant during SF5+ bombardment of the PHEMA layer. Once the PMMA layer was uncovered, the C2H3O2+ intensity rose to its maximum, decreased slightly during sputtering of the PMMA layer, and was reduced to background levels after the silicon substrate was uncovered. The decrease of the C2H5O+ intensity below zero after sputtering through both polymer layers showed the relative intensity of the C2H5O+ secondary ion in the PMMA layer. The spectrum- and ion-normalized depth profiles were then used to determine the sputter rates of the polymer layers and assign a depth scale to the depth profile (Figure 5c). 918 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

The spectrum- and ion-normalized depth profiles were used in the calculation of sputter rates for the individual polymers and interface widths. The normalized depth profiles are shown in the Supporting Information for this paper, while the sputter rates and interface widths are discussed in more detail below. Effect of Layer Order on Polymer Sputter Rates. The sputter rates for the different bilayer and trilayer films are summarized in Figure 6. The sputter rates of the polymers in the multilayer films were compared with previous measurements of the sputter rates of spin-cast films of these polymers.35,37 The sputter rates have units of cubic nanometers per ion and represent an average amount of material removed per primary ion impact. Previous studies have suggested that degradation via intra- or intermolecular cross-linking lowers the average sputter rate of a polymer under 5 keV SF5+ bombardment.35-37 Degradation by cross-linking is typically accompanied by the loss of oxygen from

Figure 5. Representative positive ion depth profile of a PHEMA on PMMA bilayer film (a) after spectrum normalization, (b) after spectrum and ion normalization, and (c) after assignment of a depth scale.

Figure 6. Sputter rates of the polymers in the bilayer and trilayer films depth profiled in this study. The sputter rates of spin-cast PMMA, PHEMA, and TFAA-PHEMA films are also shown for comparison. Reported values are averages of measurements over four depth profiles ( 1 standard deviation. For reference, a sputter rate of 30 nm3/incident SF5+ is equivalent to a sputter rate of 1.2 Å/(nA s) under the conditions used in this study.

the polymer surface and subsequent observation of unsaturated aromatic hydrocarbons in the positive ion TOF-SIMS spectra of the ion-damaged material. These unsaturated hydrocarbons (e.g., C7H7+, C9H7+, C10H8+, and C13H9+) were also observed in the positive ion depth profiles of the polymer multilayers in this study. With the accumulation of ion-induced damage, the sputter rate can be expected to decrease with increasing SF5+ ion dose as more damaged material accumulates.

The order of the polymer layers affected the sputter rates of the different polymers in the bilayer films. For the PHEMA on PMMA bilayer film, the sputter rate of the PHEMA layer was approximately the same as the sputter rate for spin-cast PHEMA, while the PMMA sputter rate was about half the sputter rate for spin-cast PMMA. For the PMMA on PHEMA bilayer, the sputter rate of the PMMA layer was approximately the same as the sputter rate for spin-cast PMMA. The PMMA overlayer only slightly decreased the sputter rate of the PHEMA layer compared with the sputter rate for spin-cast PHEMA. The decrease of the C2H5O+ and C2H3O2+ relative intensities during the surface transient suggests that PHEMA accumulated more ion-induced damage under SF5+ bombardment than PMMA, with the increased damage accumulation contributing to the differing sputter rates of PHEMA and PMMA in the bilayer films. This also shows that the sputter rate of the polymers can change with increasing SF5+ dose, in contrast with previous studies which have shown that the sputter rate of PMMA under 5.5 keV SF5+ bombardment was relatively constant with increasing SF5+ dose.32 For the PHEMA or PMMA on TFAA-PHEMA bilayers, the sputter rates of PHEMA or PMMA were slightly higher than for spin-cast PHEMA or PMMA. The roughness of the underlying TFAA-PHEMA layer may have resulted in higher sputter rates for the overlying polymer due to the protrusion of some of the TFAA-PHEMA underlayer into the polymer overlayer. Likewise, mixing of the two polymer films could have occurred, though neither interfacial roughness nor mixing was specifically characterized in this study. Overlayers of PMMA or PHEMA on the Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

919

TFAA-PHEMA layer also reduced the sputter rate of the TFAAPHEMA layer compared with TFAA-PHEMA alone. The sputter rate of the TFAA-PHEMA layer was also slightly higher when PMMA was the outermost layer. The accumulation of ion-damaged material probably resulted in the decrease in the TFAA-PHEMA sputter rate, with less damage accumulation occurring when PMMA was the outermost layer. For the trilayer films, similar results were observed. For the PHEMA on PMMA on TFAA-PHEMA trilayer film, the sputter rate of the PHEMA layer was approximately the same as for the PHEMA on PMMA bilayer film. The sputter rate of the PMMA layer was higher than that measured for the PHEMA on PMMA bilayer film, but lower than that measured for the PMMA on TFAA-PHEMA bilayer film. For the PMMA on PHEMA on TFAA-PHEMA trilayer film, the sputter rate of the PMMA layer was the same as that for the PMMA on PHEMA bilayer film. The sputter rate of the PHEMA layer was higher than that measured for the PMMA on PHEMA bilayer film, but less than that measured for the PHEMA on TFAA-PHEMA bilayer film. The sputter rate of the TFAA-PHEMA layer in either layer was significantly less than for a single TFAA-PHEMA layer or either of the bilayer films containing TFAA-PHEMA. These results were consistent with increasing damage accumulation with increasing SF5+ dose. In summary, the decrease in the characteristic secondary ion intensities and the observation of unsaturated hydrocarbons in the positive ion depth profiles noted for the polymers used in this study suggest the accumulation of ion-induced damage during the depth profile. Ion-induced damage accumulation in the outermost layers of polymer multilayers also reduced the sputter rates of underlying layers. These results were consistent with the damage accumulation rates of the individual polymers. Interface Widths and Sputter-Induced Topography. Many factors can contribute to the observed interface width between adjacent polymer layers: (1) Mixing of adjacent polymer layers during fabrication of the polymer multilayers. For this study, this factor can be neglected due to the poor miscibility of the polymers used. (2) Interfacial roughness that occurred via deposition of a polymer overlayer on top of a rough layer. Before sputtering, the PHEMA films spin-cast from methanol and the TFAA-PHEMA films had substantial topography when viewed using optical microscopy, while the PMMA films cast from toluene were substantially smoother. The topography of the PHEMA film was on a lateral length scale of 50-100 µm. While the topography in the vertical dimension was not specifically characterized, the deposition of PMMA onto the PHEMA or TFAA-PHEMA layers may have resulted in interfacial roughness at the buried polymerpolymer interface. (3) Roughness at the outermost surface, which can act as a template and transfer roughness throughout the depth profile. Due to the roughness of the spin-cast PHEMA films, this factor could contribute to the interface widths of multilayer films with PHEMA at the outermost surface. Additionally, PMMA layers deposited onto PHEMA or TFAA-PHEMA layers may exhibit some of the roughness of the underlying layer at the outermost surface. However, sputtering can result in the smoothing of rough polymer surfaces under some conditions.54-56 (4) Ion-induced (54) Pignataro, S.; Fragala, M. E.; Puglisi, O. Nucl. Instrum. Methods Phys. Res., B 1997, 131, 141.

920 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

Figure 7. Interface widths for the (a) bilayer and (b) trilayer films depth profiled in this study. The polymer layers are numbered from the outermost surface. Reported values are averages of measurements over four depth profiles ( 1 standard deviation.

mixing between adjacent layers during depth profiling. (5) Ioninduced topography formation during depth profiling. While some primary ion conditions can result in the smoothing of polymer films, other conditions lead to substantial roughening.56-60 Factors such as the primary ion mass, energy, dose, and current density can affect whether surface smoothing or roughening occurs. If it occurs, ion-induced topography formation would be expected to increase with increasing SF5+ primary ion dose, resulting in increasing interface widths between lower layers. The polymerpolymer and polymer-silicon interface widths were used to assess depth resolution and ion-induced topography formation during the depth profiling of the polymer multilayers. Figure 7a shows the polymer-polymer and polymer-silicon interface widths for the bilayer films. In all cases, the interface widths between adjacent polymer layers were less than the polymer-silicon interface widths, with the interface width between the TFAA-PHEMA layer and the silicon substrate greater than (55) Lee, J. W.; Kim, T. H.; Kim, S. H.; Kim, C. Y.; Yoon, Y. H.; Lee, J. S.; Han, J. G. Nucl. Instrum. Methods Phys. Res., B 1997, 121, 474. (56) Svorcik, V.; Arenholz, E.; Rybka, V.; Hnatowicz, V. Nucl. Instrum. Methods Phys. Res., B 1997, 122, 663. (57) He, D.; Bassim, M. N. J. Mater. Sci. 1998, 33, 3525. (58) Netcheva, S.; Bertrand, P. Nucl. Instrum. Methods Phys. Res., B 1999, 151, 129. (59) Ektessabi, A. M.; Yamaguchi, K. Thin Solid Films 2000, 377-378, 793. (60) Netcheva, S.; Bertrand, P. J. Polym. Sci., B 2001, 39, 314.

the interface widths between PHEMA or PMMA and the silicon substrate. The interface widths between adjacent polymer layers were