10646
Langmuir 2008, 24, 10646-10653
Poly(ethylene glycol) Monolayer Formation and Stability on Gold and Silicon Nitride Substrates Marta Cerruti,*,†,‡ Stefano Fissolo,§ Carlo Carraro,† Carlo Ricciardi,§ Arun Majumdar,‡ and Roya Maboudian† Departments of Chemical Engineering and Mechanical Engineering, UniVersity of California at Berkeley, Berkeley, California 94720, and Department of Physical EngineeringsLATEMAR Unit, Politecnico of Torino, 10100 Torino, Italy ReceiVed April 30, 2008. ReVised Manuscript ReceiVed July 8, 2008 Poly(ethylene glycol) (PEG) self-assembled monolayers (SAMs) are extensively used to modify substrates to prevent nonspecific protein adsorption and to increase hydrophilicity. X-ray photoelectron spectroscopy analysis, complemented by water contact angle measurements, is employed to investigate the formation and stability upon aging and heating of PEG monolayers formed on gold and silicon nitride substrates. In particular, thiolated PEG monolayers on gold, with and without the addition of an undecylic spacer chain, and PEG monolayers formed with oxysilane precursors on silicon nitride have been probed. It is found that PEG-thiol SAMs are degraded after less than two weeks of exposure to air and when heated at temperatures as low as 120 °C. On the contrary, PEG-silane SAMs are stable for more than two weeks, and fewer molecules are desorbed even after two months of aging, compared to those desorbed in two weeks from the PEG-thiol SAMs. A strongly bound hydration layer is found on PEG-silane SAMs aged for two months. Heating PEG-silane SAMs to temperatures as high as 160 °C improves the quality of the monolayer, desorbing weakly bound contaminants. The differences in stability between PEG-thiol SAMs and PEGsilane SAMs are ascribed to the different types of bonding to the surface and to the fact that the thiol-Au bond can be easily oxidized, thus causing desorption of PEG molecules from the surface.
I. Introduction Self-assembled monolayers (SAMs) formed on various substrates have been extensively employed in the past1,2 for such applications as functionalizing the surfaces with active molecules and protecting them from adsorption of species present in the environment. Two of the most commonly investigated substrates are gold and silicon. In general, SAMs formed on gold are attached using thiol-modified molecules, whereas for Si substrates a variety of strategies are employed.1 A common one involves the condensation of oxysilane end-modified molecules with hydroxyl groups formed upon oxidation of the Si surface. Poly(ethylene glycol) (PEG) SAMs are used to modify both Au and Si substrates to prevent nonspecific adsorption of proteins and cells.3,4 PEG is an amphiphilic polymer, and when hydrated, it is surrounded by a water cage with an organized structure, such that at least 2-3 water molecules are present per ethylene glycol (EG) unit.5 Thus, PEG monolayers provide a convenient route to hydrophilic surface passivation. PEG-based SAMs are formed on Au using PEG molecules that are end-functionalized with thiol groups,4,6-8 whereas several * To whom correspondence should be addressed at the Department of Chemistry, North Carolina State University, Raleigh, NC 27695. † Department of Chemical Engineering, University of California at Berkeley. ‡ Department of Mechanical Engineering, University of California at Berkeley. § Politecnico of Torino. (1) Aswal, D.; Lenfant, S.; Guerin, D.; Yakhmi, J.; Vuillaume, D. Anal. Chim. Acta 2006, 568, 84. (2) Ulman, A. Chem. ReV. 1996, 96, 1533. (3) Harris, J. M. Poly(ethylenglycol) Chemistry. Biotechnical and Biomedical Applications; Plenum Press: New York, 1992. (4) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10714. (5) Heuberger, M.; Drobek, T.; Spencer, N. D. Biophys. J. 2005, 88, 495. (6) Tokumitsu, S.; Liebich, A.; Herrwerth, S.; Eck, W.; Himmelhaus, M.; Grunze, M. Langmuir 2002, 18, 8862. (7) Zhou, C.; Khlestkin, V.; Braeken, D.; Keersmaecker, K. D.; Laureyn, W.; Engelborghs, Y.; Borghs, G. Langmuir 2005, 21, 5988.
different approaches have been reported to form PEG SAMs on Si,9-12 including the end modification of PEG molecules with oxysilane groups.13-15 Recently, similar PEG-silane molecules have been made commercially available, thus greatly simplifying the procedure required to form PEG monolayers on Si.16 PEG SAMs on both Au and Si have been characterized extensively with many different techniques4,7,8,14,16,17 Many cantilever-based sensing devices are made of a silicon nitride (SiN) beam covered on one side with Au. In such cases, PEG SAMs can be used for their surface modification.18,19 However, in contrast to Si and Au, very limited characterization has been performed for PEG-containing SAMs on SiN.17,20 In the present paper, we provide an extensive XPS analysis, complemented by water contact angle measurements, to characterize PEG SAMs formed on Au and SiN and to compare their stability with respect to storage time and elevated temperature. It is important to understand the differences in terms of packing density and stability of the SAMs formed on the two surfaces, (8) Zwahlen, M.; Herrwerth, S.; Eck, W.; Grunze, M.; Hahner, G. Langmuir 2003, 19, 9305. (9) Sofia, S. J.; Premnath, V.; Merrill, E. W. Macromolecules 1998, 31, 5059. (10) Metzger, S. W.; Natesan, M.; Yanavich, C.; Schneider, J.; Lee, G. J. Vac. Sci. Technol., A 1999, 17, 2623. (11) Zhang, F.; Kang, E. T.; Neoh, K. G.; Wang, P.; Tan, K. L. J. Biomed. Mater. Res. 2001, 56, 324. (12) Popat, K. C.; Sharma, S.; Johnson, R.; Desai, T. Surf. Interface Anal. 2003, 35, 205. (13) Jo, S.; Park, K. Biomaterials 2000, 21, 605. (14) Norrman, K.; Papra, A.; Kamounah, F. S.; Gadegaard, N.; Larsen, N. B. J. Mass Spectrom. 2002, 37, 699. (15) Sharma, S.; Johnson, R.; Desai, T. Biosens. Bioelectron. 2004, 20, 227. (16) Papra, A.; Gadegaard, N.; Larsen, N. Langmuir 2001, 17, 1457. (17) Girones, R. L. M.; Bolhuis-Versteeg, L. A. M.; Wessling, M. J. Colloid Interface Sci. 2006, 299, 831. (18) Wu, G.; Ji, H.; Hansen, K.; Thundat, T.; Datar, R.; Cote, R.; Hagan, M. F.; Chakraborty, A.; Majumdar, A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 1560. (19) Shu, W.; Laue, E. D.; Seshia, A. A. Biosens. Bioelectron. 2007, 22, 2003. (20) Kamruzzahan, A.; Ebner, A.; Wildling, L.; Kienberger, F.; Riener, C.; Hahn, C.; Pollheimer, P.; Winklehner, P.; Holzl, M.; Lackner, B.; Scho¨rkl, D. M.; Hinterdorfer, P.; Gruber, H. J. Bioconjugate Chem. 2006, 17, 1473.
10.1021/la801357v CCC: $40.75 2008 American Chemical Society Published on Web 08/27/2008
PEG Monolayer Formation and Stability on Substrates
Figure 1. Molecular structures and abbreviations of PEG compounds used.
so that spurious differences in surface stress or natural frequency due to changes in SAMs are not mistaken with the signals that one is interested in measuring. In particular, while the bond between oxysilane monolayers and oxidized Si (or SiN) surfaces is a covalent bond formed upon condensation of alkoxy groups from the SAM and hydroxyl groups from the surface, the thiolAu bond is not as strong, and it has been shown that oxidation can occur within a few hours of exposure to the ambient environment.21 It has been hypothesized that PEG SAMs22,23 with a more packed and ordered configuration may be more stable and more resistant to protein adsorption than disorganized, sparse SAMs. For this reason, PEG-thiol SAMs have been prepared using an alkylic chain as a spacer between the thiol end group and the EG units. The presence of the alkylic chain should aid the formation of a tightly packed, and thus more stable, monolayer in proximity of the Au surface, on top of which PEG helixes can stand.22 In this paper, we have analyzed this hypothesis by comparing the resistance to shelf-aging and heating of PEG-silane and PEG-thiol monolayers, with and without the presence of an undecylic chain spacer.
II. Materials and Methods A. Preparation of PEG Monolayers on SiN. Shown in Figure 1, PEG molecules modified with methoxysilane end groups (here referred to as “PEG-silanes”) and three different ranges of EG units (values of x) have been investigated, in particular, [(methoxytriethyleneoxy)propyl]trimethoxysilane (95%, Gelest product no. SIM6493.4, MW ) 327.29), [[2-methoxypoly(ethyleneoxy)6-9]propyl]trimethoxysilane (90%, Gelest product no. SIM6492.7, MW ) 460-590), and [[2-methoxypoly(ethyleneoxy)8-12]propyl]trimethoxysilane (90%, Gelest product no. SIM6492.72, MW ) 596-725). The two latter compounds contain mixtures of PEG with 6-9 and 8-12, respectively, EG units. Hereafter, these molecules will be referred to as Si-PEG3, Si-PEG6/9, and Si-PEG8/12, respectively. To prepare 5 mM solutions, 80, 100, and 120 µL of Si-PEG3, Si-PEG6/9, and Si-PEG8/12, respectively, were dissolved in 50 mL of toluene (Sigma), and 20 µL of HCl was added to all the solutions to catalyze the condensation reaction. Fresh solutions were prepared right before the silanization was started. Low-stress silicon nitride films (thickness of 1.5 µ m) were grown on Si substrates by low-pressure chemical vapor deposition. SiN substrate was cut into 5 mm × 5 mm squares, N2 blown, and then ultrasonicated in acetone (Sigma, HPLC grade) for 15 min. After being dried with N2, the samples were subjected to UV-ozonolysis (21) Willey, T.; Vance, A.; van Buuren, T.; Bostedt, C.; Terminello, L.; Fadley, C. Surf. Sci. 2005, 576, 188. (22) Harder, P.; Grunze, M.; Dahint, R.; Whitesides, G. M.; Laibinis, P. E. J. Phys. Chem. B 1998, 102, 426. (23) Li, L.; Chen, S.; Zheng, J.; Ratner, B.; Jiang, S. J. Phys. Chem. B 2005, 109, 2934.
Langmuir, Vol. 24, No. 19, 2008 10647 (UVO-Cleaner, model no. 42, Jelight Co., Inc.) for 10 min. This procedure led to very clean and oxidized surfaces, with an increased number of surface hydroxyl groups compared to that of the starting substrates. Immediately after removal from the UVO-Cleaner, the samples were immersed in the PEG-silane solutions for 1 h. After this time, the samples were sonicated in acetone for 5 min to remove physisorbed molecules, further rinsed with acetone, and N2 dried. Contact angles were measured immediately after sample preparation, and X-ray photoelectron spectroscopy (XPS) analyses were performed within 24 h. B. Preparation of PEG Monolayers on Au. Shown in Figure 1 are PEG molecules end-functionalized with thiol groups (referred to as “PEG-thiols”) and with and without an alkyl spacer between the PEG molecule and the thiol end, namely, R-methoxy-ωmercaptopoly(ethylene glycol)16 (IRIS Biotech, product no. PEG1172, MW ) 750), R-methoxy-ω-mercaptoheptakis(ethylene glycol) (Biovectra, product no. 11156-0695, MW ) 356.6), and (1-mercapto11-undecyl)tris(ethylene glycol) and 1-mercapto-11-undecyl)hexakis(ethylene glycol) (Asemblon, product nos. 171041-011 and 231043011, with MW ) 336.53 and 468.7, respectively). These molecules will be referred to as SH-PEG7, SH-PEG16, SH-uPEG3, and SHuPEG6, respectively. Solutions (5 mM) of SH-PEG7 and SH-PEG16 were prepared by dissolving 8.9 and 18.7 mg of the two molecules, respectively, in 5 mL of ethanol (Sigma). Concentrated solutions of SH-uPEG3 and SH-uPEG6 were prepared by dissolving 25 mg of material in 1 mL of ethanol. This solution was further diluted with ethanol to obtain the desired solvent condition and a concentration of 1 mM. The part of the concentrated solution that was not used immediately for the experiments was stored under N2. Gold substrates with a final thickness of 25 nm were prepared by evaporation on Si wafers, after deposition of a 5 nm Cr adhesion layer. The subrates were cut in 5 mm × 5 mm squares, N2 blown, and then sonicated in acetone for 15 min. After being dried with a stream of N2, the samples were cleaned with UVO for 10 min. Right after this procedure, the samples were immersed in PEG-thiol solutions prepared as described above for 1 h. Then they were sonicated in acetone for 5 min to remove physisorbed molecules, further rinsed with acetone, and N2 dried. Contact angles were measured immediately after sample preparation, and XPS analyses were performed within 24 h. C. Measurement Techniques. Contact angle experiments were performed using a Rame´-Hart contact angle goniometer (model 10000-115) equipped with a CCD camera and analyzed with Rame´Hart software. Static contact angles were measured by the sessile drop method, using deionized (DI) water with drop volumes of 2 µL. The results shown in this paper are the average of at least three measurements performed on different samples. X-ray photoelectron spectroscopy was performed in an ultrahigh-vacuum chamber with a base pressure of 10-9 Torr, equipped with an Omicron EA125 electron energy analyzer and an Omicron DAR400 source of Al KR X-rays at an energy of 1486.6 eV. The detector angle was 0° to the surface normal. Spectral deconvolution was performed after background subtraction with the Shirley method. Energy calibration by the commonly employed method of using the C1s peak of hydrocarbon contamination as a reference (as explained, e.g., by Clark and Thomas) was not possible in all samples, as surface passivation prevented the adsorption of contaminants on some specimens. However, either the etheric or the aliphatic C peaks (or both) can be observed in every sample, and their positions were used as references for the binding energies of all other atomic species.
III. Results and Discussion PEG-silanes and PEG-thiols shown in Figure 1 were deposited on SiN and Au, respectively. The formation and stability of these monolayers were studied with XPS and water contact angle measurements, as described next. A. Monolayer Formation. 1. X-Ray Photoelectron Spectroscopy Results. XPS spectra of PEG-thiol monolayers prepared on gold by deposition for 1 h in pure ethanol are shown in Figure 2. On bare gold, in addition to a Au 4f doublet at 84 and 87.8
10648 Langmuir, Vol. 24, No. 19, 2008
Cerruti et al.
Figure 2. XPS spectra for carbon 1s (top row), oxygen 1s (middle row), and sulfur 2p (bottom row) regions of bare Au (column A), SH-PEG7 (column B), SH-PEG16 (column C), SH-uPEG3 (column D), and SH-uPEG6 (column E). Experimental spectra are plotted in solid lines, fitted spectra in dashed lines, and fits for each peak component in dashed-dotted lines.
Figure 3. XPS spectra for carbon 1s (A) and oxygen 1s (B) regions for bare SiN UV-ozone-treated substrate, Si-PEG3, Si-PEG6/9, and Si-PEG8/12. Experimental spectra are plotted in solid lines, fitted spectra in dashed lines, and fits for each peak component in dashed-dotted lines.
eV, a peak centered at 285 eV is observed, which is attributed to adventitious hydrocarbon contamination. In addition, the oxygen region of the spectrum shows two peaks, one at 531.1 eV, which is attributed to carbon contamination,25 and one at 533.1 eV, which is attributed to physisorbed water.26 The carbon 1s region of the spectra for SH-PEG7 and SH-PEG16 shows only one peak, centered at 286.5 eV and assigned to the etheric C present in the ethylene glycol chain.7,27 The carbon 1s region for SH-uPEG3 and SH-uPEG6 monolayers can be deconvoluted into two peaks, one at the same energy as found for SH-PEG7 and SH-PEG16 (286.5 eV) and the other peak at 1.6 eV lower energy. This latter peak is related to the presence of the alkylic chain in the SH-uPEG6 and SH-uPEG3 samples28 and possibly to hydrocarbon contamination. Hereafter, the peak at 285 eV will be denoted by “C1”, and the second peak at 286.6 eV, indicative of the presence of PEG, will be denoted by “C2”. The O 1s peak, centered at 533 eV, is indicative of the presence of PEG on all samples. In the S 2p region, a main peak centered (24) Reference deleted in proof. (25) Lilley, C. M.; Huang, Q. Appl. Phys. Lett. 2006, 89, 203114. (26) Andersson, L. P. A. N. K.; Nikitin, A.; Ogasawara, H. Phys. ReV. Lett. 2004, 93, 196101. (27) Chandekar, A.; Sengupta, S.; Barry, C.; Mead, J.; Whitten, J. Langmuir 2006, 22, 8071. (28) Ruiz-Taylor, L. A.; Martin, T. L.; Zaugg, F. G.; Witte, K.; Indermuhle, P.; Nock, S.; Wagner, P. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 852.
at 162.7 eV is observed on all PEG-coated samples. A peak at the same position has been observed on bound thiolates and disulfides on gold.29 Usually a peak at even lower energies (162 eV) is observed for low coverages of thiol SAMs on Au, whereas at high coverage, S-S bonds can be formed and shift the sulfur peak position to 163 eV.30 XPS spectra of representative PEG monolayers formed on SiN are shown in Figure 3. In panel A, the high-resolution C 1s regions of spectra for PEG-silanes with different numbers of EG units are compared to the spectrum of bare SiN. The component relative to the etheric C typical of PEG was found at 286.7 ( 0.1 eV (C2), similarly to what was previously observed on Au (see Figure 2). The peak at 285 eV (C1) found on all PEG-silane samples is related to the presence of the propylic chain that separates the PEG oligomers from the silane terminal group. Some carbonaceous contamination on the sample surface also contributes to this peak, and in fact, it is also observed on bare SiN. This explains why the ratio between the C2 and C1 areas does not directly correlate with the number of PEG units. A third peak at lower energy (283 eV) is found on all samples, as well as on bare SiN, and signals the presence of C-Si species, possibly (29) Bearinger, J.; Terrettaz, S.; Michel, R.; Tirelli, N.; Vogel, H.; Textor, M.; Hubbell, J. Nat. Mater. 2003, 2, 259. (30) Rodriguez, J. A.; Dvorak, J.; Jirsak, T.; Liu, G.; Hrbek, J.; Aray, Y.; Gonzalez, C. J. Am. Chem. Soc. 2003, 125, 276.
PEG Monolayer Formation and Stability on Substrates
Langmuir, Vol. 24, No. 19, 2008 10649
Table 1. Relative Grafting Density for SH-PEG and Si-PEG Monolayersa PEG-thiol
C2/Au
std dev
corrected density
relative density
SH-PEG7 SH-PEG16 SH-uPEG3 SH-uPEG6
0.091 0.086 0.056 0.073
0.0063 0.0014 0.0065 0.011
0.013 0.0054 0.019 0.012
0.69 0.28 1.00 0.64
PEG-silane
C2/Si
std dev
corrected density
relative density
Si-PEG3 Si-PEG6/9 Si-PEG8/12
2.77 0.66 0.66
0.64 0.08 0.19
0.92 0.088 0.066
1.00 0.10 0.07
a For each monolayer, the area of the C 1s peak centered at 286.7 eV (C2) is normalized to the area of the peaks relative to the substrate (Au 4f for SH-PEG and Si 2p for Si-PEG) (second column), then divided by the number of C atoms in the PEG chain (fourth column), and then normalized to the monolayer with the highest corrected density (fifth column).
formed during the preparation of the SiN films.31 Figure 3B shows the high-resolution O 1s spectra for the same samples presented in Figure 3A. On bare SiN, the peak at 532.5 is related to oxidized SiOxNy species and to water,32 whereas the peak at 530.0 eV is due to surface contaminants.25 The spectra for all PEG-silane samples have a very intense peak located at 533 eV, which is related to the etheric O from PEG, as observed above for the SH-PEG monolayers. The oxidized SiN substrate may also partially contribute to this peak. A less intense, broad peak centered at 530 eV is also observed and can be related to surface contamination, as observed on the SiN substrate. To establish a quantitative relation between the grafting density and PEG chain length, we report in Table 1 the ratio between the C 1s peak area relative to grafted PEG (C2) and the peak areas representative of the substrate (Au 4f for SH-PEG SAMs and Si 2p for Si-PEG SAMs). On the basis of these data, a “corrected density” (reported in Table 1) has been calculated for each monolayer, by dividing the C2/Au value by the number of PEG units per chain present in each monolayer. “Relative density” results reported in the last column of Table 1 are obtained by normalizing the data on the basis of the monolayer with the highest corrected density. It can be noted that, for both the SHPEG SAMs and the SH-uPEG SAMs, higher densities are obtained with shorter chain monolayers. This is particularly true for SH-PEG16 as compared to SH-PEG7, which indicates that the PEG chain in SH-PEG16 is not straight, but rather folds on itself, thus lowering the effective coverage. Moreover, the longer, disorganized PEG chains may prevent the formation of a very dense monolayer for SH-PEG16. This has been observed for longer PEGs adsorbed on Au nanoparticles33 and on other surfaces.34 The addition of an undecylic chain helps the formation of monolayers with PEG chains standing up better on the Au surface: the highest density among all SAMs analyzed was obtained with SH-uPEG3, and the difference in density between SH-uPEG6 and SH-uPEG3 is lower than that observed when SH-PEG16 and SH-PEG7 are compared. This observation is in agreement with what was previously observed for similarly alkylated SH-PEG monolayers.23 The results obtained for Si-PEG SAMs are also shown in Table 1. Similar to the results obtained on gold, in this case the highest grafting density is obtained with Si-PEG3, i.e., the shortest monolayer. Moreover, the difference in density observed for (31) Deng, Z.-W.; Souda, R. Diamond Relat. Mater. 2002, 11, 1676. (32) Poon, M.; Kok, C.; Wong, H.; Chan, P. Thin Solid Films 2004, 462-463,
42. (33) Liu, Y.; Shipton, M.; Ryan, J.; Kaufman, E.; Franzen, S.; Feldheim, D. Anal. Chem. 2007, 79, 2221. (34) Kingshott, H. J. G. P.; Thissen, H. Biomaterials 2002, 23, 2043.
Table 2. Water Contact Angle Measurements for PEG-Thiol and PEG-Silane Monolayers PEG-thiol
contact angle
std dev
PEG-silane
contact angle
std dev
SH-PEG7 SH-PEG16 SH-uPEG3 SH-uPEG6
63.1 33.3 36.1 36.2
0.59 1.7 0.41 0.79
Si-PEG3 Si-PEG6/9 Si-PEG8/12
41.5 38.2 40.5
3.8 1.1 0.50
Si-PEG3 and the other monolayers is very high. This indicates that also on the SiN surface longer and folded PEG chains prevented the formation of dense surface layers. Evidently, the propylic chain present between the EG units and the oxysilane group is too short to help straighten the PEG chains and form packed monolayers, while the undecylic chain present in SHuPEG SAMs seemed to be quite effective for such a purpose. 2. Contact Angle Results. Water contact angle measurements provide information on the hydrophilicity of the surfaces covered with PEG molecules. Table 2 shows the contact angles that were measured right after the preparation of the monolayers. SH-uPEG3 and SH-uPEG6 show a water contact angle of 36°. Similar values have been reported22 and are indicative of the presence of a hydrophilic surface layer. PEG-thiols without an alkylic chain between the thiol group and the PEG units behave differently. SH-PEG7 exhibits a much higher contact angle value than the other PEG-thiols. This is expected, because SH-PEG7 molecules have a terminal methoxy group, which is more hydrophobic than the terminal hydroxy group present on SHuPEG3 and SH-uPEG6. Even though SH-PEG16 molecules are also terminated with a methoxy group, the observed contact angle for SH-PEG16 is only 33°, more similar to the hydroxy-terminated SAMs. This result confirms the observation inferred from XPS (Table 1), that the longer SH-PEG16 chains are disordered and folded, so that the hydrophilic etheric groups present in the PEG chains are exposed to the water droplet rather than the terminal hydrophobic methoxy groups. Silane SAMs on SiN substrates show water contact angles between 38° and 41°. These results are consistent with the literature for Si-PEG SAMs of the same lengths grafted on silicon16 and are consistent with the values obtained for hydroxyterminated thiol monolayers on gold. The observation of such low values despite the terminal methoxy group for all PEGsilane SAMs is again indicative of some degree of folding of the PEG chains and exposure of hydrophilic internal etheric groups to water interaction. This is consistent with the lower grafting density deduced by XPS measurements for longer PEG-silane SAMs (Table 1). B. Monolayer Stability with Time. In the following section, we examine the stability of PEG-thiol and PEG-silane SAMs as a function of storage time in ambient air. The stability of thiolated SAMs on Au has been studied in the past with respect to photooxidation and oxidation with ozone or simply in air, and it has been shown that the oxidation of the thiolate-Au bond can occur in air without exposure to light in just a few hours, depending on the alkylic chain length21,35 and on the Au substrate quality.36 Only in one case was the issue of thiolated PEG-containing monolayer stability addressed specifically, and it was shown that PEG-containing block copolymers with sulfide or disulfide bonds increase the resistance of such monolayers to oxidation, which otherwise readily occurs by exposure to air and light.29 Here we compare the degradation, over the course of two weeks, of PEGsilane and PEG-thiol SAMs with and without an undecylic chain (35) Schoen_sch, M.; Pemberton, J. J. Am. Chem. Soc. 1998, 120, 4502. (36) Lee, M.-T.; Hsueh, C.-C.; Freund, M.; Ferguson, G. Langmuir 1998, 14, 6419.
10650 Langmuir, Vol. 24, No. 19, 2008
Cerruti et al.
Figure 4. Aging of SH-PEG SAMs in air: XPS spectra of C 1s (top row), O 1s (middle row), and S 2p (bottom row) for SH-PEG16 (A) and SH-uPEG3 (B) samples measured as-prepared (columns labeled “Fresh”) and after 14 days (columns labeled “After 2 weeks”). Experimental spectra are plotted in solid lines, fitted spectra in dashed lines, and fits for each peak component in dashed-dotted lines.
Figure 5. Quantitative data relative to PEG SAM aging: (A) relative amount of C due to PEG (C2 peak) and of S for SH-PEG16 and SH-uPEG3 samples before and after 14 days of aging, (B) relative amount of C due to PEG (C2 peak) and of total C for Si-PEG3 before and after two months of aging and after the following supercritical drying. For each sample, the area underlying the C (and S) peaks was ratioed to the area underlying the Au or Si signals from the substrate.
spacer between the thiol and the EG units, and we also report on the effect of two months of aging for PEG-silane SAMs. 1. XPS Results. To check the stability of SH-PEG SAMs in air, XPS spectra of SAMs as-prepared and after two weeks were measured. Samples were kept in the dark during this period of time. All the monolayers, both with and without an undecylic chain betwen the thiol group and the EG units, showed signs of degradation, namely, oxidation of the thiol-Au bond and lowered density. Figure 4 shows the detailed spectra for SH-PEG16 and SH-uPEG3, chosen as representative samples for PEG-thiol SAMs with and without an alkylic chain spacer. The most evident feature in the spectra is the shift of the S 1s peak from 163 to 168 eV upon aging. This indicates that the sulfur initially bound to Au was oxidized to sulfoxide groups due to contact with air. Sulfur oxidation was accompanied by a broadening of the O 1s peak and the formation of a new species at 531.8 eV, which must also be related to the formation of sulfoxide species.21 Changes were also observed in the C spectra; namely, a new component at 285 eV, relative to surface contamination, was formed in the spectrum of SH-PEG16, and the component relative to PEG (C2) decreased in intensity with respect to the component relative to surface contamination (C1) on sample SH-uPEG3.
A more quantitative analysis of these spectra is shown in Figure 5A, where the areas of the C2 and S peaks were normalized to the area underlying Au signals coming from the substrate for each sample. It is clear that the amounts of both S and C relative to PEG (C2) decreased during sample aging. In particular, the amounts of C relative to PEG and of S left after two weeks for sample SH-PEG16 were 29% and 34% of the starting C and S, respectively, whereas for sample SH-uPEG3 the amounts of C and S left were 41% and 38%, respectively. On both samples, then, the density of the PEG monolayer decreased to more than half the starting density in two weeks of aging in the dark. To analyze the effect of aging on PEG-silane SAMs, XPS spectra were recorded of a sample of Si-PEG3 as-prepared and aged for two months. The most striking effect observed was a global shift of all the peaks to lower energies. Spectra are shown for Si-PEG3 in Figure 6 without energy scale calibration, simply as they were taken from the XPS instrument. The spectra shown in Figure 6A are the same as those in Figure 3 for Si-PEG3, except for the fact that no energy calibration was performed on those shown in Figure 6A. It can be noted that all peak energies in the spectra shown in Figure 6A are very high, which is indicative of a high surface charging, most likely due to the poor conductivity
PEG Monolayer Formation and Stability on Substrates
Langmuir, Vol. 24, No. 19, 2008 10651
Figure 6. Effect of aging and supercritical CO2 drying on Si-PEG3 SAM: XPS spectra of C 1s (top row), O 1s (middle row), and Si 2p (bottom row) for Si-PEG3 as prepared (column A), after two months (column B), and after CPD (column C). Experimental spectra are plotted in solid lines, fitted spectra in dashed lines, and fits for each peak component in dashed-dotted lines.
of the SiN substrate. After two months of aging, all the peaks shifted by approximately -2 eV (Figure 6B). This implies that the sample aged for two months was a better conductor than the as-prepared sample. We hypothesize that this is due to the presence of water adsorbed on the sample, because of the hydrophilic nature of the PEG SAM. To check this hypothesis, we performed supercritical CO2 drying (SCD) on the sample aged for two months to remove adsorbed water and recorded the XPS spectra afterward. The resulting spectra are shown in Figure 6C. All the peaks were again shifted by +2 eV compared to those recorded after two months of aging. This confirms that, after SCD, the water layer adsorbed on the PEG SAM was removed and the SiN surface was charged again during the XPS measurement. However, it should be noted that the water adsorbed on the PEG monolayer was very strongly bound, as it resisted the high-vacuum conditions that were used during the XPS measurement, and was removed only after supercritical drying. We compared the amount of PEG on the samples as-prepared and aged for two months, before and after CPD, by normalizing the area of the C2 peak to the area of the Si 2p peak for each sample. The results are shown in Figure 5B, where the total amount of C versus Si is also reported. Some PEG molecules desorbed during the two months, although less than those desorbed on PEG-thiol SAMs on Au in only two weeks (see Figure 5A). It is possible that some of the PEG was displaced by hydrocarbons adsorbed from the air (the C2/Si ratio decreased while the total C/Si ratio increased in fresh and aged samples, as shown in Figure 5B). The amount of PEG that was left after two months of aging was strongly bound to the surface, as it did not desorb during the SCD process, whereas surface contamination was eliminated by SCD (the C2/Si ratio remained stable, while the total C/Si ratio decreased in the samples aged for two months before and after CPD, as shown in Figure 5B). 2. Contact Angle Results. The variation of water contact angles measured over a period of two weeks for both PEG-thiol and PEG-silane SAMs is shown in Figure 7. Right after the cleaning procedure, which involved UV-ozonolysis, both Au and SiN substrates were extremely hydrophilic and water contact angles
Figure 7. Variation with time of contact angles for PEG-thiol SAMs compared to the gold substrate (A) and PEG-silane SAMs compared to the silicon nitride substrate (B).
were lower than 10°. As a function of storage time in air, both substrates showed an increase in water contact angle due to contamination with the hydrocarbons present in the air (see filled symbols in both sections of Figure 7). In particular, contact angles on Au quickly approached a plateau value of 80°, whereas the increase in water contact angle was slower for SiN, and a plateau value of only 30° was reached. When Au was covered with SH-PEG SAMs, the initially low contact angles increased with time, approaching the value of the bare gold around 80° after a few days of exposition to air (see Figure 7A). On the contrary, contact angles measured on Si-PEG SAMs grafted on SiN showed stable values over two weeks, higher than the plateau value observed for bare SiN (see Figure 7B). These observations confirmed that SH-PEG SAMs degraded after a few days, leaving some Au surface exposed to hydrocarbon contamination from the atmosphere, whereas the coverage obtained for Si-PEG SAMs on SiN was stable over the course of at least two weeks. C. Monolayer Stability at Elevated Temperatures. To investigate the stability of SH-PEG SAMs at elevated temperatures, the SH-PEG16 sample was heated in air at 120 °C and then at 160 °C, for 30 min at each step. XPS spectra recorded at each step are shown in Figure 8A. Some degradation started occurring already at 120 °C: the O and the C peaks broadened,
10652 Langmuir, Vol. 24, No. 19, 2008
Cerruti et al.
Figure 8. Effect of heating the SH-PEG16 monolayer: XPS spectra of C 1s (top row), O 1s (middle row), and S 2p (bottom row) for SH-PEG16 as-prepared (column A) and after being heated at 120 °C (column B) and at 160 °C (column C). Experimental spectra are plotted in solid lines, fitted spectra in dashed lines, and fits for each peak component in dashed-dotted lines.
and a component at 285 eV relative to hydrocarbon impurities started to show, together with a higher energy component which may be related to PEG degradation. The thiolate-Au bond was partially oxidized, as shown by the formation of a peak at 169 eV, indicative of the presence of sulfoxide groups. More hydrocarbon contamination (peak at 285 eV) was found on the sample heated at 160 °C, possibly replacing desorbed SH-PEG molecules. At this temperature, the S oxidation was completed, and no thiolate groups (peak at 163 eV) were observed. A new component centered at 531.5 eV clearly appeared in the O peak, which resembled the previously observed peak on the two week old SAMs (see Figure 4A). The amount of C relative to PEG (C2 peak) decreased when the sample was heated: the C2/Au ratio found on samples heated at 120 and 160 °C was 37% of that observed on the starting sample. This shows that SH-PEG SAMs can be desorbed at similar temperatures found to desorb alkanethiol monolayers.37 It is known that heating alkylsilane molecules can be beneficial for monolayer formation, because it can help the condensation of hydroxyl groups,38 whereas only at temperatures higher than 200 °C do silane-containing SAMs start degrading.39 Thus, experiments were carried out to check whether the quality of a monolayer prepared with an old precursor solution could be improved by heating. Figure 9A shows the C 1s and O 1s spectra of a Si-PEG6/9 monolayer of poor quality. In fact, in the C 1s spectrum, the component at 286.6 eV indicative of the presence of PEG has a low intensity compared to the other C components relative to hydrocarbon contamination (285 eV) and to C-Si species from the SiN substrate (283 eV). In the O 1s spectrum, (37) Shadnam, M. R.; Amirfazli, A. Chem. Commun. 2005, 39, 4869. (38) Jonas, U.; del Campo, A.; Kruger, C.; Glasser, G.; Boos, D. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5034. (39) Zhuang, Y.; Hansen, O.; Knieling, T.; Wang, C.; Rombach, P.; Lang, W.; Benecke, W.; Kehlen-beck, M.; Koblitz, J. J. Micromech. Microeng. 2006, 16, 2259.
Figure 9. Effect of heating a Si-PEG6/9 monolayer of poor quality: XPS spectra of C 1s (top row) and O 1s (bottom row) for Si-PEG6/9 as-prepared (column A) and after being heated at 160 °C (column B). Experimental spectra are plotted in solid lines, fitted spectra in dashed lines, and fits for each peak component in dashed-dotted lines.
a high-intensity peak is present at 530 eV, which was observed on the bare SiN substrate and was not evident on well-formed PEG-silane monolayers (see Figure 3). The spectra recorded after the sample was heated in air at 160 °C are shown in Figure 9B. In spectrum C, the peak at 286.7 eV relative to that for PEG increased in intensity in comparison to the lower energy components. Major changes in the O spectrum were also observed: the peak at 533 eV relative to that for PEG largely increased in intensity compared to the peak at lower energy, resembling the spectra observed on PEG-silane SAMs prepared from fresh
PEG Monolayer Formation and Stability on Substrates
solutions (Figure 3B). Thus, also the O spectra can be used to check the quality of PEG SAMs on SiN, whereas not many references to O spectra can be found in the literature concerning PEG-SH SAMs on Au or PEG-silane SAMs on Si. This experiment shows that heating PEG-silane monolayers is a good way to selectively desorb surface contamination and leave a cleaner PEGylated surface.
IV. Conclusions We have performed extensive XPS analyses complemented by water contact angle measurements to compare PEG monolayers formed on Au and on SiN substrates, using thiol and oxysilane PEG derivatives, respectively. We showed that monolayers prepared with longer chains of both PEG-thiols and PEG-silanes were less dense than those prepared with shorter chains, because of folding of the longer PEG chains. However, PEG-thiol precursors with an undecylic spacer between the terminal thiol groups and the EG units helped the formation of denser monolayers for longer chain PEGs. The undecylic spacer did not improve the stability of PEG-thiol monolayers upon aging. All PEG-thiol SAMs studied degraded in room air in less than two weeks by thiol-Au bond oxidation, and more than half of the PEG molecules were desorbed. Instead, PEG-silane SAMs were stable during the first two weeks. During two months, a strongly
Langmuir, Vol. 24, No. 19, 2008 10653
bound surface hydration layer was formed, and some PEG-silane molecules desorbed, but fewer than those desorbed from PEGthiol SAMs in two weeks. PEG-thiol monolayers were also much less stable upon heating than PEG-silane monolayers. Oxidation of the SH-Au bond was completed when PEG-thiols were heated at 160 °C for 1/2 h. On the contrary, heating improved the quality of PEG-silane monolayers, leading to the desorption of loosely bound contaminants from the SiN surface. These results show that caution must be used when PEG-thiol SAMs are used to functionalize Au surfaces, as they are easily degraded. The effect of such degradation should be taken into consideration on a case-by-case basis, as some applications may be more negatively influenced by it than others. In particular, applications that require formation of PEG SAMs simultaneously on both Au and SiN surfaces, such as in the case of functionalization of Au/SiN cantilevers, could be adversely affected by the different degradation rates of the monolayers on the two surfaces, which could lead to spurious changes in surface stress. Acknowledgment. This work was performed under the auspices of the National Science Foundation by the University of California at Berkeley under Grant No. 0425914 within the Center of Integrated Nanomechanical Systems (COINS). LA801357V
