Self-Assembled Monolayers of Aromatic Thiol and ... - ACS Publications

The Journal of Physical Chemistry C 2017 121 (1), 459-470. Abstract | Full .... ACS Nano 0 (proofing),. Abstract | Full .... Effects of Metal Coating ...
0 downloads 0 Views 92KB Size
Langmuir 2001, 17, 6981-6987

6981

Self-Assembled Monolayers of Aromatic Thiol and Selenol on Silver: Comparative Study of Adsorptivity and Stability Sang Woo Han, Seung Joon Lee, and Kwan Kim* Laboratory of Intelligent Interface, School of Chemistry and Molecular Engineering and Center for Molecular Catalysis, Seoul National University, Seoul 151-742, Korea Received March 26, 2001. In Final Form: June 16, 2001 We have investigated comparative adsorptivities and relative stabilities of self-assembled monolayers of thiol (benzenethiol, BT) vs selenol (benzeneselenol, BSe) on a silver surface by diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy and surface-enhanced Raman spectroscopy (SERS). BT and BSe are chemisorbed on silver as benzenethiolate and benzeneselenolate, respectively, after deprotonation with a tilted orientation with respect to the substrate surface; the benzene rings of BT and BSe are tilted by 25° and 37°, respectively, from the substrate normal. Competitive adsorption experiments show that adsorption of BT is more favorable by 0.3 kcal/mol. Temperature-dependent DRIFT spectra indicated that the monolayer of benzenethiolate on silver is thermally more stable than that of benzeneselenolate. This could be evidenced from the fact that the vibrational peaks of the benzenethiolate species were observed up to 458 K, while those of benzeneselenolate became substantially weakened around 418 K. The more negative desorption potential of the BT monolayer clearly indicates that the benzenethiolate binds more strongly to the silver surface than the benzeneselenolate does.

Introduction Self-assembled monolayers (SAMs) have attracted a great deal of interest in recent years due to their various applications such as in biomimetic films, chemical sensors, nonlinear optical materials, high-density memory devices, protective coatings, lubricants, and photopatterning substrates.1 Although a wide variety of substrates and functional groups are known to form SAMs, the thiol/ disulfide monolayer on Au has received considerable attention due to its simplicity and ease of preparation.2 Most of the thiol/disulfide compounds investigated so far contain a long hydrophobic tail, which enables these compounds to form a compact monolayer on Au surfaces. The bonding pattern in both thiols and disulfides is the same except for an oxidative dissociation of the S-S bond for disulfides.3 In contrast, selenol/diselenide monolayers have not received enough attention despite their promising utility for a variety of applications such as in photoresists, photocatalysts, preparation of semiconductor quantum dots, photoinduced electron-transfer systems, and so on.4 There have been several reports on organoselenium monolayers. However, adsorbate structure, adsorption strength, and the stability of SAMs from organoselenium are somewhat controversial issues. Samant et al.5 were the first to characterize a selenolate monolayer. Using * To whom correspondence should be addressed. Phone: +822-8806651. Fax: +82-2-8743704. E-mail: [email protected]. (1) (a) Hickman, J. J.; Ofer, D.; Laibinis, P. E.; Whitesides, G. M. Science 1991, 252, 688. (b) Mirkin, C. A.; Ratner, M. A. Annu. Rev. Phys. Chem. 1992, 43, 719. (c) Li, D.; Ratner, M. A.; Marks, T. J.; Znang, C. H.; Yang, J.; Wong. G. K. J. Am. Chem. Soc. 1990, 112, 7389. (d) Wollman, E. W.; Kang, D.; Frisbie. C. D.; Larcovic, T. M.; Wrighton, M. S. J. Am. Chem. Soc. 1994, 116, 4395. (e) Tariov, M. J.; Burgess, D. R. F., Jr.; Gillen, G. J. Am. Chem. Soc. 1993, 115, 5305. (f) Kawanishi, Y.; Tamaki, T.; Sakuragi, M.; Seki, T.; Swuzki, Y.; Ichimura, K. Langmuir 1992, 8, 2601. (2) Ulman, A. Chem. Rev. 1996, 96, 1533. (3) Ulman, A. An Introduction to Ultrathin Organic Films from Langmuir-Blodgett to Self-Assembly; Academic Press: San Diego, CA, 1991. (4) Pandey, G.; Poleshwar Rao, K. S. S. Angew. Chem., Int. Ed. Engl. 1995, 34, 2669. (5) Samant, M. G.; Brown, C. A.; Gordon, J. G., II. Langmuir 1992, 8, 1615.

X-ray diffraction, they determined the two-dimensional structure of the dodecaneselenolate SAM on gold. They found that it was distorted ∼3% from the typical (x3×x3)R30° structure of a thiolate SAM, and determined that the molecular tilt angle was 15 ( 1° off-perpendicular. On the basis of published values for Au-Se and Au-S bond strengths, Samant et al. proposed that the thiolate is chemisorbed more strongly than the selenolate. Dishner et al.6 revealed that the two-dimensional structure of benzeneselenolate monolayers on gold follows a (x3×x3)R30° unit cell, using scanning tunneling microscopy (STM). The fact that STM images of well-ordered benzeneselenolate monolayers could be obtained, while STM images of benzenethiolate monolayers cannot, suggests that the Au-Se interaction is stronger than the Au-S interaction. Huang et al.7 determined the compositions of diphenyl disulfide (DPDS) and diphenyl diselenide (DPDSe) mixed monolayers formed by displacement and competitive adsorption by surface-enhanced Raman spectroscopy (SERS). They found that the Se-Se bond is cleaved to form benzeneselenolate upon adsorption, analogous to formation of benzenethiolate monolayers from DPDS. DPDSe displaced benzenthiolate from gold, but DPDS did not displace benzeneselenolate. Competitive adsorption experiments showed that adsorption of DPDSe was more favorable by ∼0.7 kcal/mol. On the other hand, in the study of the adsorption properties of DPDS, DPDSe, and naphthalene disulfide on gold films, Bandyopadhyay et al.8 reported that the Se-Se bond of DPDSe is preserved upon adsorption. In addition, their temperature-dependent SERS studies revealed that the DPDS monolayer is by far the most stable one, being stable up to a temperature of 423 K. Herein we aim to report on the comparative capacity for SAM formation of thiols vs selenols on a silver surface and the relative stabilities of SAMs to thermal and (6) Dishner, M. H.; Hemminger, J. G.; Feher, F. J. Langmuir 1997, 13, 4788. (7) Huang, F. K.; Horton, R. C., Jr.; Myles, D. C.; Garrell, R. L. Langmuir 1998, 14, 4802. (8) Bandyopadhyay, K.; Vijayamohanan, K.; Venkataramanan, M.; Pradeep, T. Langmuir 1999, 15, 5314.

10.1021/la010464q CCC: $20.00 © 2001 American Chemical Society Published on Web 09/29/2001

6982

Langmuir, Vol. 17, No. 22, 2001

electrochemical treatment. Benzenethiol (BT) and benzeneselenol (BSe) were selected as model compounds. Adsorption characteristics were investigated by virtue of diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy and SERS. The composition of BT and BSe mixed monolayers formed by competitive adsorption was determined using SERS. The compositions of the monolayers allowed us to determine the relative adsorptivities of BT and BSe. The relative thermal and electrochemical stabilities of the BT and BSe monolayers were characterized by temperature-dependent DRIFT spectroscopy and potential-dependent SERS, respectively. To the best of our knowledge, this is the first thorough investigation on the adsorptivity and the stability of an organoselenium compound adsorbed on a silver surface. Experimental Section Chemicals. Silver powders with a nominal particle size of 2-3.5 µm (>99.9%) and silver foil (0.05 mm thick, 99.9%) were purchased from Aldrich. BT (TCI, >98%) and BSe (Aldrich, 99%) were used as received. The chemicals otherwise specified were reagent grade, and triply distilled water, of resistivity greater than 18.0 MΩ·cm, was used in making the aqueous solutions. Monolayer Preparation. For adsorption of BT and/or BSe on powdered silver, 0.05 g of silver powder was placed in a clean small vial into which 0.5 mL of the stock solution (1 mM BT or BSe in ethanol) was added. After 3 h, the liquid phase was decanted. The remaining solid particles were washed with excess ethanol and dried in a N2 atmosphere for 2 h. The powdered sample was then transferred to either a Harrick microsampling cup for DRIFT spectral analyses or a thin glass capillary for SER spectral analyses. Substrates for potential-dependent SERS measurements were prepared by immersing silver foils, etched previously by dipping in 30% HNO3 for 2-3 s (this process makes the silver foil SERS active.9,10) and then in 1 mM BT or BSe in ethanol solution for at least 1 h. After the substrates were removed, they were rinsed thoroughly with ethanol and then dried with a high-purity N2 gas stream. Infrared Spectral Measurements. Infrared spectra were obtained using a Bruker IFS 113v Fourier transform IR spectrometer equipped with a globar light source and a liquid N2-cooled wide-band mercury cadmium telluride detector. The method for obtaining the room temperature and temperaturedependent DRIFT spectra has been described previously.11-16 A Harrick model DRA-2CO diffuse reflection attachment was used, and a total of 32 scans were measured at a resolution of 4 cm-1 with the use of previously scanned pure Ag powders as the background. To record the temperature-dependent DRIFT spectra, a reaction chamber, made of stainless steel (Harrick model HVC-DR2), was located inside the reflection attachment. The DRIFT spectra are reported as -log(R/Ro), where R and Ro are the reflectance of the sample and that of the pure Ag powder, respectively. Raman Spectral Measurements. Raman spectra were obtained using a Renishaw Raman system model 2000 spectrometer equipped with an integral microscope (Olympus BH2UMA). The 514.5 nm radiation from a 20 mW air-cooled argon ion laser (Spectra Physics model 163-C4210) was used as an excitation source. Raman scattering was detected with 180° geometry using a Peltier cooled CCD detector. The Raman band (9) (a) Xue, G.; Dong, J. Anal. Chem. 1991, 63, 2393. (b) Xue, G.; Zhang, J. Macromolecules 1991, 24, 4195. (c) Wu, Y.; Zhao, B.; Xu, W.; Li, B.; Ozaki, Y. Langmuir 1999, 15, 1247. (d) Wu, Y.; Zhao, B.; Xu, W.; Li, B.; Jung, Y. M.; Ozaki, Y. Langmuir 1999, 15, 4625. (10) Lee, I.; Han, S. W.; Kim, C. H.; Kim, T. G.; Joo, S. W.; Jang, D.-J.; Kim, K. Langmuir 2000, 16, 9963. (11) Han, H. S.; Kim, C. H.; Kim, K. Appl. Spectrosc. 1998, 52, 1047. (12) Lee, S. J.; Kim, K. Vib. Spectrosc. 1998, 18, 187. (13) Han, S. W.; Han, H. S.; Kim, K. Vib. Spectrosc. 1999, 21, 133. (14) Han, H. S.; Han, S. W.; Joo, S. W.; Kim, K. Langmuir 1999, 15, 6868. (15) Han, H. S.; Han, S. W.; Kim, C. H.; Kim, K. Langmuir 2000, 16, 1149. (16) Lee, S. J.; Han, S. W.; Yoon, M.; Kim, K. Vib. Spectrosc. 2000, 24, 265.

Han et al. of a silicon wafer at 520 cm-1 was used to calibrate the spectrometer, and the accuracy of the spectral measurement was estimated to be around 1 cm-1. The Raman spectrometer was interfaced with an IBM-compatible PC, and the spectral data were analyzed using Renishaw WiRE v. 1.2 software based on the GRAMS/32C suite program (Galactic). A glass capillary (KIMAX-51) with an outer diameter of 1.5-1.8 mm was used as a sampling device for the powdered sample. The HNO3-etched Ag foil clamped to a spectroelectrochemical cell was used for the potential-dependent SERS experiment. Platinum wire was used as the counter electrode. A silver wire was put into the cell as a quasi-reference electrode (AgQRE),17 and its potential was calibrated against the saturated calomel electrode (SCE); all potentials quoted in this work were thus relative to that of the SCE. The electrode potential was controlled by a Bioanalytical Systems CV-27 potentiostat. The electrolyte was NaClO4, and the electrolyte solution was deaerated thoroughly with highpurity N2 gas before spectral measurements.

Results and Discussion We have reported previously that DRIFT and SER spectra can be obtained with a very high signal-to-noise ratio for adsorbates on 2-µm-sized silver particles.11-16 The DRIFT spectral pattern was little different from the external reflection infrared spectral pattern taken for the same molecules on vacuum-evaporated thick silver films.11-13 The usual surface selection rule that only the vibrational modes whose dipole moment derivatives have components normal to the metal surface are exclusively IR active18 thus seemed applicable even to the surface of fine metal particles. (Greenler et al.19 performed classical and quantum mechanical calculations showing that the surface IR selection rule should apply for particles larger than 1.5 nm; in these calculations, the particles were assumed to be smooth and regular. In our previous STM study,11 the 2-µm-sized silver particles were in fact smooth and flat on the several tens of nanometers scale.) The SER spectral pattern on the powdered silver was also found to show little difference from that on vacuumevaporated thin, rough silver films.13 The combination of infrared and Raman spectroscopies should be very useful in investigating molecular adsorption on metal surfaces, especially on SERS-active noble metals such as Ag, Au, and Cu. In this work, we thus investigated the adsorption characteristics, including the comparative ability of SAM formation and the relative thermal stability of BT and BSe, using silver powder as a substrate. Homogeneous Monolayers of Benzenethiolate and Benzeneselenolate. The infrared spectra of the monolayers formed on silver are compared in Figure 1; parts a and b show the DRIFT spectra of BT and BSe, respectively, self-assembled on silver powder from a 1 mM ethanol solution. The major peaks in Figure 1 are collectively summarized in Table 1, along with appropriate assignments made on the basis of published data.20-23 In the infrared spectra of neat BT and BSe (not shown), the S-H and Se-H stretching peaks (ν(SH) and ν(SeH)) appeared at 2566 and 2301 cm-1, respectively, but their counterparts were completely absent in the DRIFT spectra shown in Figure 1. The CSH and CSeH bending bands (δ(CSH) and δ(CSeH)) identified at 913 and 795 cm-1 in (17) Han, S. W.; Seo, H.; Chung, Y. K.; Kim, K. Langmuir 2000, 16, 9493. (18) Greenler, R. G. J. Chem. Phys. 1966, 44, 310. (19) Greenler, R. G.; Snider, D. R.; Witt, D.; Sorbello, R. S. Surf. Sci. 1982, 118, 415. (20) Varsanyi, G. Vibrational Spectra of Benzene Derivatives; Academic Press: New York, 1969. (21) Joo, T. H.; Kim, M. S.; Kim, K. J. Raman Spectrosc. 1987, 18, 57. (22) Whiffen, D. H. J. Chem. Soc. 1956, 1350. (23) Rasheed, F. S.; Kimmel, H. S. Spectrosc. Lett. 1977, 10, 791.

SAMs of Aromatic Thiol and Selenol on Silver

Langmuir, Vol. 17, No. 22, 2001 6983

Figure 1. DRIFT spectra of (a) BT and (b) BSe adsorbed on Ag powder. Table 1. DRIFT and SER Spectral Data (cm-1) and Vibrational Assignments of BT and BSe Adsorbed on Silver Powder BT

BSe

DRIFT

SERS

DRIFT

SERS

assignmenta

3052 1575 1473 1181 1073

3060 1573 1473 1181 1072

3053 1574 1472 1178

3050 1571 1470 1176

1022 1000 733 691

1022 999

1068 1018 998 726 687

1061 1019 997

2 (a1, ν(CH)) 8a (a1, ν(CC)) 19a (a1, ν(CC)) 9a (a1, β(CH)) 1 (a1, β(CCC)) + ν(CS) 1 (a1, β(CCC)) + ν(CSe) 18a (a1, β(CH)) 12 (a1, β(CCC)) 11 (b1, γ (CH)) 4 (b1, γ (CCC)) 6a (a1, β(CCC)) + ν(CS) 6a (a1, β(CCC)) + ν(CSe) 7a (a1, β(CCC)) + ν(CS) 7a (a1, β(CCC)) + ν(CSe)

689 660 417 298 a

Assigned on the basis of refs 20-23.

the neat infrared spectra of BT and BSe, respectively, were also completely missing upon adsorption. This observation indicates, as expected, that BT and BSe chemisorb on silver as benzenethiolate and benzeneselenolate, respectively, after deprotonation. In Figure 1a, the 1073 cm-1 band has contributions from the ring mode 1 and the ν(CS) vibration, and the 3052, 1575, 1473, 1181, 1022, and 1000 cm-1 bands are assigned to in-plane phenyl ring modes. All these bands have dipole moments parallel to the molecular axis. The bands appearing at 733 and 691 cm-1 are ascribed to the out-of-plane C-H and ring deformation modes, respectively, and have dipole moments perpendicular to the molecular plane. Recalling the IR surface-selection rule,18 the presence of both inplane and out-of-plane vibration modes indicates that the benzenethiolate has neither a perpendicular nor a parallel orientation with respect to the silver surface. Since the DRIFT spectral features of BSe on Ag are similar to those of BT (see Figure 1b and Table 1), benzeneselenolate will also have a tilted orientation with respect to the substrate surface. More quantitatively, the tilt angle (θ) between the benzene ring and the surface normal may be estimated using the relationship24

tan2 θ ) [A(a1)/A(b1)]T[A(b1)/A(a1)]DR

(1)

in which A(a1) and A(b1) correspond, respectively, to the infrared absorbances of the a1 (in-plane) and b1 (out-ofplane) modes in the transmission IR and DRIFT spectra.

Figure 2. SER spectra of (a) BT and (b) BSe adsorbed on Ag powder.

Using the intensity ratios of the ring 18a (a1) and 11 (b1) modes, the benzene rings of BT and BSe on Ag are then calculated to be tilted away from the surface normal by 25° and 37°, respectively; in these calculations, the transmission IR spectral data were from either silver benzenethiolate or silver benzeneselenolate (data not shown). These suggest that BSe should assume a more tilted orientation on Ag than BT. Parts a and b of Figure 2 show the SER spectra of BT and BSe, respectively, self-assembled on silver powder. These spectra were obtained from the same samples as were used in obtaining the DRIFT spectra shown in Figure 1. The observed peaks in Figure 2 are collectively summarized in Table 1. As in the DRIFT spectra, the ν(SH) and ν(SeH) bands were completely absent in the SER spectra, and the δ(CSH) and δ(CSeH) bands were also completely missing. The ring 1 and 6a modes, which have contributions from the ν(CS) or the ν(CSe) vibration, are red-shifted by ∼20 and ∼10 cm-1, respectively, upon adsorption. Together these changes indicate that BT and BSe adsorb dissociatively on silver as benzenethiolate and benzeneselenolate, respectively, forming a S-Ag (or SeAg) bond as revealed by DRIFT spectroscopy. The adsorption mechanism of an adsorbate can be deduced from its SER spectrum through a detailed analysis of the peak shift and band broadening caused by the surface adsorption.25 According to the SERS investigation of benzene derivatives by Gao and Weaver,26 the ring modes have to red-shift by around 10 cm-1 and have an increase in their widths insofar as the surface-ring π orbital interaction is the driving force for the surface adsorption; the red shift can be attributed to the bond weakening caused by the electron back-donation from the metal to the antibonding π* orbital of the benzene moiety.26,27 It is worthwhile on this basis to note that the bandwidths of the benzene ring modes are hardly different between the ordinary Raman (OR, not shown) and SER spectra. In (24) Han, S. W.; Ha, T. H.; Kim, C. H.; Kim, K. Langmuir 1998, 14, 6113. (25) (a) Joo, T. H.; Kim, K.; Kim, M. S. J. Phys. Chem. 1986, 90, 5816. (b) Yim, Y. H.; Kim, K.; Kim, M. S. J. Phys. Chem. 1990, 94, 2552. (c) Kwon, Y. J.; Son, D. H.; Ahn, S. J.; Kim, M. S.; Kim, K. J. Phys. Chem. 1994, 98, 8481. (d) Cho, S. H.; Han, H. S.; Jang, D.-J.; Kim, K.; Kim, M. S. J. Phys. Chem. 1995, 99, 10594. (e) Kim, S. H.; Ahn, S. J.; Kim, K. J. Phys. Chem. 1996, 100, 7174. (26) Gao, P.; Weaver, M. J. J. Phys. Chem. 1985, 89, 5040. (27) (a) Moskovits, M.; Dillela, D. P. J. Chem. Phys. 1980, 73, 6068. (b) Avouris, P.; Demuth, J. E. J. Chem. Phys. 1981, 75, 4783.

6984

Langmuir, Vol. 17, No. 22, 2001

addition, the peak positions of the ring modes, except the ones having a contribution from the C-S or C-Se vibration, are also similar to those in the OR spectra. (In fact, the 8a modes of BT and BSe are red-shifted by ∼10 cm-1. Using the downshifts alone as evidence that the ring plane lies flat on the metal surface, however, neglects the effect that bonding to the metal through the headgroup can have on the ring mode vibrational frequencies. Complexation of BT with Ag(I)21 leads to a downshift comparable to that observed when BT adsorbs on Au or Ag, for example.) These observations indicate that surfacephenyl ring interaction may not be a dominant factor in the surface adsorption. According to the electromagnetic surface selection rule, vibrational modes whose polarizability tensor elements are perpendicular to a metal surface should be enhanced much more in an SER spectrum than the parallel ones.28 In this light, the complete absence of out-of-plane modes in the SER spectra presented in Figure 2 may reflect the fact that benzenethiolate and benzeneselenolate should take on a tilted orientation on the silver surface. The noticeable appearance of the C-H stretching bands at 3060 and 3050 cm-1 in parts a and b, respectively, of Figure 2 is also indicative of such an orientation on silver.29 The SER spectral features of BT and BSe on a Ag surface seem thus to be consistent with the DRIFT ones. The estimated molecular orientation in this work is in fact consistent with those of previous works.30,31 Competitive Adsorption Experiments. We performed competitive adsorption experiments to determine the relative adsorption equilibrium constants and the relative free energies of adsorption of benzenethiolate and benzeneselenolate. The experiments were conducted by immersing the silver powder for 3 h in an ethanolic solution containing a mixture of BT and BSe. The mole ratios ranged from 0.10:0.90 to 0.90:0.10 (BT/BSe), with a total adsorbate concentration of 1 mM. After the Ag powder was rinsed with ethanol and then dried, we obtained SER spectra for the mixed monolayers. The spectra were analyzed to obtain the compositions of the mixed monolayers. Figure 3 shows a plot of the mole fraction of benzeneselenolate in the monolayer as a function of the mole fraction of BSe in the immersion solution. The inset shows the SER spectra of the mixed monolayers in the range between 900 and 150 cm-1, where the 6a and 7a modes are observed. The 298 cm-1 band in the SER spectra of benzeneselenolate was used in estimating the monolayer composition of benzeneselenolate. To compare the data obtained from different surfaces, this band was normalized against a band near 1000 cm-1 (ring mode 12) in the SER spectra. In a 50:50 BSe/BT solution mole ratio, the monolayer mole ratio reaches ∼30:70. For a 90:10 solution ratio, the monolayer ratio is about 60:40. The plot of monolayer vs solution composition (Figure 3) shows that benzenethiolate adsorbs more readily than benzeneselenolate, but not overwhelmingly. (In fact, the DRIFT spectra of mixed monolayers did not vary significantly with the mole ratio. This is not unreasonable when the fact that the overall DRIFT spectral features of BT and BSe are nearly the same (see Figure 1) is considered. In the SER spectra of mixed monolayers, however, the 6a and 7a bands of BT and BSe are clearly separated (see the inset of Figure 3). Therefore, we can estimate the (28) Moskovits, M. J. Chem. Phys. 1982, 77, 4408. (29) (a) Creighton, J. A. Surf. Sci. 1983, 124, 209. (b) Moskovits, M.; Suh, J. S. J. Am. Chem. Soc. 1986, 108, 4711. (30) Takahashi, M.; Fujita, M.; Ito, M. Surf. Sci. 1985, 158, 307. (31) Carron, K. T.; Hurley, L. G. J. Phys. Chem. 1991, 95, 9979.

Han et al.

Figure 3. Monolayer mole fraction (θ) of benzeneselenolate as a function of the solution mole fraction (χ) of BSe. The solid curve represents a Langmuir isotherm fit. The inset shows the SER spectra of the mixed monolayers in the range between 900 and 150 cm-1.

compositions of the mixed monolayers from the analysis of the SER spectral features.) The difference in the free energies of adsorption (∆Gads) of benzenethiolate and benzeneselenolate was determined from the adsorption Keq values of each component, obtained by fitting the data in Figure 3 to a competitive adsorption isotherm model. We used a variation of the Langmuir analytical isotherm for a two-component adsorption process:7,32

θA )

θB )

(KAχA)R 1 + (KAχA)R + (KBχB)β (KBχB)β 1 + (KBχB)β + (KAχA)R

(2)

(3)

In this model, θA and χA are the mole fractions of component A in the monolayer and solution phase, respectively. R and β are constants representing the component-specific nonidealities, which can range from 0 to 1.32 A value of 1 follows the ideal Langmuir isotherm, in which all active sites have equivalent adsorption energies.33 A value less than 1 shows deviation from the Langmuir isotherm, with 0 indicating the largest distribution of active-site adsorption energies.33 KA is the equilibrium constant, defined by

KA ) [Aads]/χA[S]

(4)

[Aads] is the surface concentration of one component, and [S] is the concentration of bare sites.34 For our purposes, we substituted χB with 1 - χA and fitted eq 2 to the data in Figure 3, with θA designated as the mole fraction in the monolayer. The solid curve in Figure 3 represents the result of this Langmuir isotherm fit. The fitting involved calculating Keq, R, and β to obtain the best fit. We calculated Keq values of (6.3 ( 0.4) × 104 and (3.9 ( 0.4) × 104 for benzenethiolate and benzeneselenolate, respectively. For the nonideality terms (R and β), we obtained values of 0.7 (32) Koopal, L. K.; Riemsdijk, W. H. v.; Wit, J. C. M. D.; Benedetti, M. F. J. Colloid Interface Sci. 1994, 166, 51. (33) (a) Sips, R. J. Chem. Phys. 1948, 16, 490. (b) Sips, R. J. Chem. Phys. 1950, 18, 1024. (34) Masel, R. I. Principles of Adsorption and Reaction on Solid Surfaces; John Wiley & Sons: New York, 1996.

SAMs of Aromatic Thiol and Selenol on Silver

Langmuir, Vol. 17, No. 22, 2001 6985

Figure 5. Temperature dependence of the intensity of the 18a mode of BT (filled circles) and BSe (open circles) adsorbed on Ag. Figure 4. DRIFT spectra of (a) BT and (b) BSe adsorbed on Ag powder taken as a function of temperature.

( 0.02 (benzenethiolate) and 0.6 ( 0.02 (benzeneselenolate), indicating the deviation from the pure Langmuir isotherm for competitive adsorption. The greater Keq value for benzenethiolate is consistent with our experimental observations that benzenethiolate adsorbs more readily than benzeneselenolate. We use the relationship between Keq and ∆Gads, shown in eq 5, to calculate the ∆Gads for each monolayer component.

∆Gads ) -RT ln K

(5)

The difference in the free energies of adsorption for BT and BSe, ∆∆Gads, was calculated to be 0.3 kcal/mol, favoring BT adsorption. This small difference is sufficient to favor benzenethiolate adsorption over benzeneselenolate adsorption and is consistent with the similar chemical structures of the two compounds. Thermal Stability of BT and BSe on Ag. Parts a and b of Figure 4 show a series of DRIFT spectra obtained as a function of temperature for BT and BSe molecules, respectively, self-assembled on Ag powder. All spectra were measured at the temperature indicated, with the temperature held constant to (1 K for 5 min while the spectra were being recorded. The temperature was raised from 298 to 488 or 448 K in 10 K steps. (In fact, the selected SER spectra are shown in Figure 4 for convenience.) The sample chamber was flushed continuously with dry nitrogen (ca. 100 mL/min) during the measurement of the DRIFT spectra. The temperature dependence of the band intensities of the 18a modes of BT and BSe are plotted in Figure 5. The temperature-dependent DRIFT spectra of the monolayers reveal their comparative thermal stability. As shown in Figures 4 and 5, the major bands are seen up to 458 K for BT, while for BSe the bands are identifiable only up to 418 K. This result unambiguously suggests that the monolayer of benzenethiolate on silver is thermally more stable than that of benzeneselenolate. Detailed information on the structural properties can be obtained from a careful analysis of the DRIFT spectral features. In the previous temperature-dependent DRIFT spectroscopic studies of the SAMs of stearic acid12 and 4-(dimethylamino)benzoic acid,16 obvious structural changes of the adsorbates were observed at certain temperatures. These can be estimated from the variations of the peak positions and certain decreases of the band

intensities of the specific vibrational modes. Contrary to the results of the previous work, the peak positions and the bandwidths of the major bands shown in Figure 4 are rather independent of the thermal treatment. This indicates that the benzenethiolate and the benzeneselenolate do not undergo structural changes until they are desorbed from the silver surfaces. Electrochemical Stability of BT and BSe on Ag. We investigated the stability of monolayers of BT and BSe on silver in electrochemical environments which may be encountered in many applications such as chemical sensing and in the study of fundamental electron-transfer phenomena.35 SERS was selected for evaluating the effects of applied potential on the structural stability of the monolayers for the following reasons. First, water is a weak Raman scatterer, allowing access to all spectral regions in the Raman spectrum containing information about the SAM film structure and integrity. In addition, SERS provides abundant spectral information about the structure of thiol or selenol SAMs, including information in low-frequency regions that is not easily accessible with FTIR.36 Parts a and b of Figure 6 show SER spectra of BT and BSe molecules self-assembled on Ag foil, respectively, obtained in aqueous 0.2 M NaClO4 as a function of applied potential. All spectra were measured at the potential indicated, with the potential held constant for 5 min while the spectra were being recorded. Overall, the SER spectral features shown in Figure 6 are similar to those obtained in air. A previous SER spectroscopic study confirmed that water has little effect on the monolayer structure, even after 1 h of exposure.36 Earlier FTIR results of Porter37 and Anderson38 also indicated that water has no influence on the structure of SAMs. As shown in Figure 6a, most of the ring modes of BT are substantially broadened and their peak positions are red-shifted as the electrode potential goes in the negative direction. The intensity of the C-H stretching band (ν(CH)) appearing at 3060 cm-1 (35) (a) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559. (b) Whitesides, G. M.; Laibinis, P. E. Langmuir 1990, 6, 87 and references therein. (c) Murray, R. W. Molecular Design of Electrode Surface; Techniques of Chemistry Series; John Wiley & Sons: New York, 1992; Vol. XXII. (d) Zhong, C.-J.; Porter, M. D. Anal. Chem. 1995, 67, 709A. (e) Pierrat, O.; Lechat, N.; Bourdillon, C.; Laval, J.-M. Langmuir 1997, 13, 4112. (36) Schoenfisch, M. H.; Pemberton, J. E. Langmuir 1999, 15, 509. (37) Stole, S. M.; Porter, M. D. Langmuir 1990, 6, 1199. (38) Anderson, M. R.; Gatin, M. Langmuir 1994, 10, 1638.

6986

Langmuir, Vol. 17, No. 22, 2001

Figure 6. SER spectra of (a) BT and (b) BSe adsorbed on Ag taken as a function of applied potential.

decreases also at a more negative potential. Such variations of the spectral features mean that the phenyl ring of benzenethiolate is oriented more parallel to the silver surface at a more negative potential. In the SERS study of BT on a Au electrode, Szafranski et al.39 also observed a similar change in orientation along the applied potential. In contrast with benzenethiolate, benzeneselenolate does not show the structural change upon variation of the electrode potential. The peak positions and bandwidths are almost invariant up to -1.3 V (see Figure 6b). The overall band intensity decreases substantially around -1.4 V, however. We interpret this change as resulting from electrochemical reductive desorption of the adsorbate, in agreement with the work of previous investigators.36,39,40 The BSe peaks are not recovered when the potential is stepped back to -0.1 V, indicating that desorption of BSe is irreversible over the time scale of this work. (The resulting solution that contains the desorbed BSe is so dilute that readsorption is not detectable on the laboratory time scale.) Porter and co-workers40 were the first to characterize the chemistry of the bound thiol headgroup and report the reductive desorption of 1-alkanethiol SAMs on evaporated polycrystalline Ag and Au surfaces in 0.5 M KOH, LiOH, and NaOH. The products of this process are proposed to be Ag0 or Au0 and thiolate species, RS-. Reductive desorption of a propanethiol film occurs at ca. -1.0 V. For longer chain alkanethiols, such as dodecanethiol and octadecanethiol, no reductive desorption peak is observed at polycrystalline Ag, regardless of the electrolyte. In these systems, reductive desorption of the film and solvent reduction occur simultaneously at a negative potential of ca. -1.4 V. Consistent with these previous works on longer 1-alkanethiols,40 desorption of benzeneselenolate may be facilitated by electrolysis of water at the interface. The potential at which water reduction begins at a BT- or BSe-modified Ag electrode is ca. -1.4 V in our experimental condition. The residual SER peaks of BSe observed at -1.4 V (see Figure 6b) indicate that the reductive desorption does not result in complete removal of the monolayer from the interface in this experiment, however. Contrary to the case of a BSe monolayer, a substantial decrease of the band intensity (39) Szafranski, C. A.; Tanner, W.; Laibinis, P. E.; Garrell, R. L. Langmuir 1998, 14, 3570. (40) (a) Widrig, C. A.; Chung, C.; Porter, M. D. J. Electroanal. Chem. 1991, 310, 335. (b) Weisshaar, D. E.; Lamp, B. D.; Porter, M. D. J. Am. Chem. Soc. 1992, 114, 5860.

Han et al.

of BT on a silver electrode does not occur around -1.4 V. An irreversible change occurs around -1.8 V. A more negative desorption potential of the BT monolayer clearly indicates that the benzenethiolate binds more strongly to the silver surface than the benzeneselenolate does, in agreement with the competitive adsorption experiment and the temperature-dependent DRIFT spectroscopic study. Until the reductive desorption begins, the structural stability of benzeneselenolate upon variation of the electrode potential is higher than that of benzenethiolate, however (vide supra). Theoretical Consideration of the Relative Adsorptivities of BT and BSe. In principle, a different adsorptivity originates from different adsorption free energies. The adsorption free energy is determined by several factors including the solvation energies of the adsorbate and the surface as well as the adsorbate-surface and the adsorbate-adsorbate interaction energies. However, in this work, we may need to consider only the adsorbate-surface interaction since the geometry of BT on Ag is quite comparable to that of BSe. With a tilted orientation, the direct surface-ring π orbital interaction will be unfavorable for both adsorbates. The higher adsorptivity of BT than BSe may then be attributed mostly to the fact that the bond dissociation energy of the Ag-S bond is greater than that of the Ag-Se bond; in fact, the former is known to be 51.9 kcal/mol while the latter is 48.4 kcal/mol.41 To obtain further insight into the adsorbate-surface interaction, we attempted to evaluate the adsorbatesurface Coulombic and charge-transfer interactions. To this end, ab initio quantum mechanical calculations have been performed for benzenethiolate and benzeneselenolate at the MP2/6-31G level.42 According to our natural population analysis,43 the net charges of the sulfur and selenium atoms in C6H5S- and C6H5Se- were -0.64 and -0.58 electron charge, respectively. Since the potential of zero charge of the silver electrode and of silver powders is the negative of the open circuit potential in air, ethanol, or water,44 the effective charges of the silver surfaces used in this work can all be presumed to be positive. Although this may imply that BT must adsorb on silver more strongly than BSe, the Coulombic interaction does not seem to be a crucial factor responsible for the higher adsorptivity of BT on Ag since the net charge difference between C6H5S- and C6H5Se- is rather insignificant. It has been well-documented that a metal can accept electrons from an electron donor into its lowest unoccupied or partially occupied molecular orbital (LUMO or conduction band). The closer the adsorbate HOMO (highest occupied molecular orbital) and the substrate LUMO are in energy, the larger the magnitude of the charge-transfer interaction.45 In our ab initio calculation, however, the HOMO energies of C6H5S- and C6H5Se- were computed to be -2.18 and -1.86 eV, respectively. The higher HOMO energy of benzeneselonolate than benzenethiolate implies that BSe must adsorb more strongly on Ag than BT in contrast with the actual experimental observation. Hence, the relative HOMO energies seem not to be a factor discriminating the relative adsorptivities of BT and BSe (41) Smoes, S.; Mandy, F.; Vander Auwera-Mahieu, A.; Drowart, J. Bull. Soc. Chim. Belg. 1972, 81, 45. (42) The ab initio quantum mechanical calculation was carried out using the Gaussian 98 program. (43) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F. NBO Version 3.1. (44) Bockris, J. O., Conway, B. E., Eds. Comprehensive Treatise of Electrochemistry; Plenum Press: New York, 1969; Vol. 1. (45) Garrell, R. L.; Chadwick, J. E.; Severance, D. L.; McDonald, N. A.; Myles, D. C. J. Am. Chem. Soc. 1995, 117, 11563.

SAMs of Aromatic Thiol and Selenol on Silver

Langmuir, Vol. 17, No. 22, 2001 6987

on the silver surface. It is unfortunate that the present frontier orbital approach does not clarify the origin of the more favorable adsorption of BT than BSe on Ag. More extensive theoretical work seems necessary, taking into account all the orbital interactions between the donor (adsorbate) and acceptor (surface). Examination of the second-highest occupied molecular orbital as well as the lower-lying orbitals is certainly needed because they may contribute significantly to the charge-transfer energetics.46

that benzenethiolate should possess greater adsorptivity than benzeneselenolate. This implies that selenolates cannot be used as alternatives to thiolates when stronger adsorbate-surface interactions are desired. Using perturbation theory, we attempted to explain the more favorable adsorption characteristics of BT than BSe on silver, but further extensive calculations seemed necessary, taking into account all the orbital interactions between the donor (adsorbate) and acceptor (surface).

Summary and Conclusions We have examined the comparative adsorptivities and the relative stabilities of SAMs from thiol vs selenol on a silver surface through competitive adsorption experiments and temperature- and potential-dependent vibrational spectral measurements. Our results demonstrate

Acknowledgment. This work was supported in part by the Korea Research Foundation (KRF; Grant 042D00073) and the Korea Science and Engineering Foundation (KOSEF; Grant 1999-2-121-001-5). K.K. also acknowledges KOSEF for providing a leading-scientist grant. S.W.H. was supported by KOSEF through the Center for Molecular Catalysis at Seoul National University. S.J.L. acknowledges the KRF for providing the BK21 fellowship.

(46) Han, S. W.; Joo, S. W.; Ha, T. H.; Kim, Y.; Kim, K. J. Phys. Chem. B 2000, 104, 11987.

LA010464Q