Hydrogen sulfide-induced desorption/reorganization of self-assembled monolayers of alkanethiol and its derivatives

Article abstract of DOI:10.1021/ja000278r

Self-assembled monolayers of alkanethiolate on gold were shown to desorb or be replaced from the gold surface by exposure to H₂S vapor. Under the same conditions, the desorption occurred faster with shorter chain length thiols than with longer ones. The thiolate moiety desorbed mainly in the form of a free thiol, which underwent various structural reorganizations depending on the chain length as well as the terminal functional group it carries. For nonpolar, methyl-terminated thiols, the assembly either randomized to give a disordered liquid state (for chain length

Full text of DOI:10.1021/ja000278r

7072
J. Am. Chem. Soc. 2000, 122, 7072-7079
Hydrogen Sulfide-Induced Desorption/Reorganization of
Self-Assembled Monolayers of Alkanethiol and Its Derivatives
Yu-Tai Tao,* Kannaiyan Pandian, and Wen-Chung Lee
Contribution from the Institute of Chemistry, Academia Sinica, Taipei, Taiwan 11529, Republic of China
ReceiVed January 25, 2000. ReVised Manuscript ReceiVed April 28, 2000
Abstract: Self-assembled monolayers of alkanethiolate on gold were shown to desorb or be replaced from the
gold surface by exposure to H2S vapor. Under the same conditions, the desorption occurred faster with shorter
chain length thiols than with longer ones. The thiolate moiety desorbed mainly in the form of a free thiol,
which underwent various structural reorganizations depending on the chain length as well as the terminal
functional group it carries. For nonpolar, methyl-terminated thiols, the assembly either randomized to give a
disordered liquid state (for chain length <C16) or maintained a well-ordered but physisorbed monolayer state
(for chain length gC18). With polar terminal groups such as hydroxyl, ester, acid, amide and, cyano groups,
the desorbed molecules reorganized into discrete clusters of the corresponding thiol compound. The results
demonstrate the important role of the terminal functional group at the film-air interface in determining the
ultimate structure of a molecular assembly.
Introduction
readily perturbed by simple exposure to certain chemical vapor.8
An interesting structural reorganization was manifested by the
n-alkanoate monolayer on silver surface. The process is believed
to originate from the competition for the silver ion by the
chemical agent (for example, HCl or H2S) so that free n-alkanoic
acids were released from the surface binding sites and underwent
further reorganization. Similar reorganization was found for the
thioacid monolayer on gold surface, after ambient hydrolysis
of the gold thiocarboxylate group to a carboxyl group and gold
sulfide.9 In all these cases, the ultimate structure of the assembly
was determined by the balance of various interactions, including
intermolecular interactions, substrate-adsorbate interactions, and
the H-bonding of carboxylic acid molecules.
The alkanethiolate on gold surface is generally considered
stable and robust because of the strong polar covalent Au-S
bond. Nevertheless, it has been noted that a trace amount of
sulfide impurity present in the solution can compete with
4-pyridinethiol effectively during the monolayer adsorption of
the pyridinethiol onto gold surface and even replace the
monolayer completely.10 It was contemplated that an alkanethiol
monolayer may also be vulnerable to a competing hydrogen
sulfide vapor at the gold surface. It is of interest to understand
the molecular process for the thiolate monolayer under H2S
atmosphere. Here we wish to report the observation of displace-
ment of the alkanethiolate moiety from the gold surface by
exposure to H2S vapor and the reorganization of the monolayer
assembly thereafter. It was found that the displacement/
reorganization process depends on the chain length of the
thiolate involved, as well as the terminal functional group it
carries. With polar terminal functional groups such as hydroxyl,
ester, acid, amide, and cyano groups, the thiolate moieties
desorbed and reorganized into discrete clusters of the corre-
The self-assembled monolayers (SAMs) of alkanethiol on
metal surfaces have been developed into a useful system both
for fundamental structural studies of interfacial phenomena as
well as technological applications of miniaturized dimensions.1
An important feature of this system is the robustness of the
adsorbed film due to the polar covalent bonding between the
sulfur atom and the underlying metal (mainly gold, silver, and
copper).2 Nevertheless, the Au-S bond is not as sturdy as it
was generally perceived. For example, it has long been known
that the thiolate adsorbate can be exchanged by a second thiol
molecule present in a contacting solution.3 The thiolate can also
be removed readily by electrochemical oxidation or reduction.4
Upon UV irradiation, a much less adhering sulfonate species
can be generated.5 Oxidation by a trace amount of ozone in air
to give sulfonate has been the subject of many studies recently.6,7
The stability of the monolayer film under various environments
is certainly an important issue that should be faced.
We reported previously that the structure of a monolayer
formed from n-alkanoic acid on metal oxide surfaces could be
* Address correspondence to this author.
(1) Ulman, A. Chem. ReV. 1996, 96, 1533-1554 and references therein.
(2) (a) Li, Y.; Huang, J.; McIver, R. T., Jr.; Hemminger, J. C. J. Am.
Chem. Soc. 1992, 114, 2428-2432. (b) Scott, J. R.; Baker, L. S.; Everett,
W. R.; Wilkins, C. L.; Fritsh, I. Anal. Chem. 1997, 69, 2636-2639.
(3) (a) Bain, C. D.; Evall, J.; Whitesides, G. M. J. Am. Chem. Soc. 1989,
111, 7155-7164. (b) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.;
Mujsce, A. M. J. Am. Chem. Soc. 1990, 112, 4301-4306. (c) Collard, D.
M.; Fox, M. A. Langmuir 1991, 7, 1192-1197.
(4) Widrig, C. A.; Chung, C.; Porter, M. D. J. Electroanal. Chem. 1991,
310, 335-359.
(5) (a) Huang, J. Y.; Dahlgren, D. A.; Hemminger, J. C. Langmuir 1994,
10, 626-628. (b) Tarlov, M. J.; Burgess, D. R. F., Jr.; Gillen, G. J. Am.
Chem. Soc. 1993, 115, 5305-5306.
(6) (a) Li, Y.; Huang, J.; McIver, R. T., Jr.; Hemminger, J. C. J. Am.
Chem. Soc. 1992, 114, 2428-2432. (b) Tarlov, M. J.; Newman, J. G.
Langmuir 1992, 8, 1398-1405. (c) Horn, A. B.; Russell, D. A.; Shorthouse,
L. J.; Simpson, T. R. E. J. Chem. Soc., Faraday Trans. 1996, 92, 4759-
4762. (d) Zhang, Y.; Terrill, R. H.; Bohn, P. W. Chem. Mater. 1999, 11,
2191-2198.
(8) (a) Tao, Y. T.; Hietpas, G. D.; Allara, D. L. J. Am. Chem. Soc. 1996,
118, 6724-6735. (b) Tao, Y. T.; Lin, W. L.; Hietpas, G. D.; Allara, D. L.
J. Phys. Chem. B 1997, 101, 9732-9740.
(9) Tao, Y. T.; Pandiaraju, S.; Lin, W. L.; Chen, L. J. Langmuir 1998,
14, 145-150.
(7) Schoenfisch, M. H.; Pemberton, J. E. J. Am. Chem. Soc. 1998, 120,
4502-4513.
(10) Yoshimoto, S.; Yoshida, M.; Kobayashi, S.; Nozute, S.; Miyawaki,
T.; Hashimoto, Y.; Taniguchi, I. J. Electroanal. Chem. 1999, 473, 85-92.
10.1021/ja000278r CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/07/2000

Desorption/Reorganization of Self-Assembled Monolayers
J. Am. Chem. Soc., Vol. 122, No. 29, 2000 7073
at 20 eV for sulfur and 50 eV for others. A freshly prepared Au substrate
was used as a reference. All peaks were corrected against Au 4f7/2 (BE
) 84.0 eV). Four elements were collected on each sample: Au, C, S,
and O with an acquisition time of 1000, 400, 12000, and 400 s for
each respective element. Curve fitting was done with Omicon-Presents
using Marquardt-Levenburg method.
sponding bulk thiol materials. For nonpolar methyl-terminal
thiols, the assembly either randomized to a disordered liquid
state (for chain length <C16) or maintained a well-ordered
physisorbed monolayer state (for chain length gC18), which is
similar to that of a physisorbed Langmuir-type film transferred
from the air/water interface. The results are rationalized as a
minimization of surface energy at the substrate surface.
Quartz Crystal Microbalance. A SEIKO EG&G quartz crystal
analyzer (Model QCA 917) was used to record the mass changes upon
adsorption/desorption. The 9 MHz AT-cut quartz crystals (supplied by
SEIKO EG&G) with a surface area of 0.3 cm2 Au electrodes were
used for all the studies. The crystals were cleaned for a few seconds
with a 1:1 mixture of hot concentrated H2SO4 and H2O2 (30%) solution,
followed by washing with Milli-Q water, THF, and hexane and finally
drying in a stream of nitrogen. Monolayers were prepared by immersing
the gold-coated quartz crystal in a thiol solution for 1 day and rinsing
with ethanol and drying with nitrogen. The H2S-induced desorption
studies were conducted by the ex situ method. The monolayer-modified
quartz crystals were placed in an H2S chamber for different periods of
time and the frequency change was measured after the sample was
washed with ethanol and THF. The resolution of the frequency change
is within (2 Hz. All experiments were conducted in a closed Faraday
cage system to avoid noise and external disturbance.
Experimental Section
Materials. Gold (>99.99%) and chromium (>99.99%) were ob-
tained from Johnson Matthey Company. One-side polished silicon
wafers were purchased from Semiconductor Processing Company.
Absolute ethanol and GR grade THF were obtained from Merck and
used as received. Saturated alkanethiols were obtained from Aldrich.
Hydrogen sulfide (99.5+% lecture bottle) was obtained from Aldrich.
The thiols 16-mercaptohexadecanoic acid, methyl 16-mercaptohexa-
decanoate, 16-mercapto-1-hexadecanol, and 16-mercaptohexadecana-
mide were prepared according to literature procedures.11,12 17-
Mercaptoheptadecanenitrile was prepared from 16-bromohexadecanol
by reacting with sodium cyanide in DMSO. The resulting 17-
hydroxyheptadecanenitrile was treated with p-toluenesulfonyl chloride
and then sodium thioacetate. Hydrolysis of the thioester gave the desired
product. Fractional recrystallization gave the 17-mercaptoheptade-
canenitrile as a white solid: 1H NMR δ 2.52 (2H, q), 2.34 (2H, t),
1.26 (28H, br s), mp 38-39 °C.
Monolayer Preparation. The gold substrates were prepared by
vacuum deposition of ∼7 nm of chromium and then 200 nm of gold
onto 2-in. silicon wafers under a vacuum of 2 × 10-6 Torr at a rate of
∼0.5 nm/s. An ethanolic solution (∼1 mM) of each thiol was prepared
under ambient condition. A monolayer of respective thiol was prepared
by soaking the gold substrate in the solution for at least 4 h before
being withdrawn from the solution and rinsed three times with pure
ethanol and once with THF then blown dry with a stream of pure
nitrogen. The thiol monolayer was subjected to IR or other characteriza-
tion within 3 h after preparation. No sign of structural change could
be detected by reflection-absorption IR during this period of ambient
storage.
Reflection-Absorption IR Spectroscopy. Reflection-absorption
IR spectra were taken on a Digilab FTS60 Fourier transform IR
spectrometer equipped with a liquid nitrogen-cooled MCT detector.
Single reflection with a grazing angle of 86° of the incident beam was
used. Spectra were taken with p-polarized light. An ozone-cleaned gold
wafer was used as the reference sample.
Results and Discussion
The displacement of alkanethiolate from the gold surface by
hydrogen sulfide was evidenced by a variety of techniques,
which are presented in the following.
Electrochemical Reductive Desorption. It has been estab-
lished that alkanethiolate can be reductively desorbed from the
surface of a gold electrode.5 The potential at which the adsorbate
desorbs depends on the packing as well as the chain length of
the thiolate moiety and the pH of the medium. Figure 1 shows
the typical reductive wave for a hexadecanethiol monolayer on
gold (abbreviated as Au/SC15CH3) at -1.36 V vs Ag/AgCl (3M
NaCl) in 0.5 M KOH in the first scan. The sharp peak will
shift to lower potential and broaden in successive cycles, due
to the partial loss of the thiolate moiety into the electrolyte
solution during the desorption-readsorption cycle.13 In contrast,
the cyclic voltammogram of a gold surface exposed to H2S vapor
gave a very sharp reductive desorption wave at -0.88 V. A
small anodic peak was observed during the reverse scan. It is
noted that a similar reductive peak (-0.9 V) was observed for
the Na2S-treated gold surface, and it was attributed to atomic
and/or oligomeric sulfur adlayer.10 The reduction peak current
increases linearly with increasing scan rate, indicating a surface-
confined species.14 When Au/SC15CH3 was exposed to H2S
for various periods of time and rinsed, the cyclic voltammogram
showed that a peak around -0.9 V was growing with time,
while the peak associated with a closely packed C16 thiolate
was lost within the first 2 h of exposure. The peak at around
-0.9 V gained its full intensity after exposure to H2S for around
6 h, together with some broad peaks present between -1.1 and
-1.2 V. This clearly suggests the replacement of the long-chain
thiol by the small H2S molecules.
Exposure Experiment. The film-covered substrate surface was
placed in a 1-L chamber and purged with a slow stream of N2 for 10
min. The H2S vapor (10 mL) was injected via a syringe to make up a
1% H2S atmosphere in the chamber. After the sample was exposed for
various amounts of time, it was retrieved and characterized directly
(as in the case of IR) or soaked in ethanol and THF consecutively before
characterization for the presence of the monolayer material.
Electrochemical Studies. All electrochemical studies were carried
out using a BAS 100B Electrochemical Analyzer (Bioanalytical System
USA). A three-electrode-cell system with a single compartment was
used. A 2000 Å gold-coated silicon wafer with ∼70 Å chromium as
an adhesion layer was used as a working electrode and a platinum spiral
wire served as a counter electrode. An Ag/AgCl (3M NaCl) electrode
acted as the reference electrode. A 0.5 M KOH solution was used as
the supporting electrolyte. Samples were purged with purified argon
gas for 10 min before starting the experiments. All potentials were
recorded against the Ag/AgCl reference. For electrochemical study, the
freshly prepared Au substrates were annealed at 300 °C for 3 h in a
muffle furnace followed by treatment with UV irradiation for 1 h to
remove organic impurities on the gold surface. After the UV treatment,
substrates were washed with ultrapure water and then rinsed with
ethanol and dried in a stream of nitrogen prior to monolayer formation.
For the H2S-induced desorption studies, the monolayer-modified Au
substrates were exposed to the H2S for various periods of time and
then rinsed with ethanol before the cyclic voltammogram experiments.
X-ray Photoelectron Spectroscopy. The X-ray photoelectron
spectroscopy experiments were performed on an Omicon ESCA system
with an Al KR monochromator X-ray source and EA 125 Hemispherical
analyzer. The X-ray power suppliers were run at 14 kV and 12 mA at
a pressure of 2 × 10-10 mbar. The pass energy of the analyzer was set
(12) Nuzzo, R. G.; Dubois L. H.; Allara, D. L. J. Am. Chem. Soc. 1990,
112, 558-569.
(13) (a) Lamp, B. D.; Hobara, D.; Porter, M. D.; Niki, K.; Cotton, T. M.
Langmuir 1997, 13, 736-741. (b) Wierse, D. G.; Lohrengel, M. M.;
Schultze, J. W. J. Electroanal. Chem. 1978, 92, 121-131.
(14) Murray, R. W. In Electroanalytical Chemistry; Bard, A. J., Ed.;
Marcel Dekker: New York, 1984; Vol. 13.
(11) Bain, C. D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G.
M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-335.

7074 J. Am. Chem. Soc., Vol. 122, No. 29, 2000
Tao et al.
Figure 2. QCM frequency change as a function of the exposure time
for the alkanethiol-cover gold surface.
by the adsorption of H2S. In the earlier stage, the mass loss of
alkanethiol outweighed the mass gain from the adsorption of
the small H2S molecules and thus a net loss of mass was noted.
After much of the long chain has been displaced, net mass
increase due to multilayer adsorption of H2S resumed. From
the slope of the curve, it is suggested that the rate of the
frequency change and thus the displacement rate decreased in
the order of C12 > C16 > C20. The longer chain thiol resisted
the displacement more effectively than the shorter chain thiol.
This chain length effect is similar to the anticorrosion effect of
thiol monolayers against aqueous Br- ion.18
Figure 1. Cyclic voltammogram of (a) H2S-modified gold, (b) Au/
SC15CH3, (c) the monolayer exposed to H2S for 2 h, and (d) the
monolayer exposed to H2S for 6 h.
X-ray Photoelectron Spectroscopy. The displacement pro-
cess for Au/SC15CH3 by H2S was also examined by XPS
spectroscopy. Before being exposed to H2S, the monolayer gave
Quartz Crystal Microbalance (QCM). QCM provides a
convenient way to observe the mass change occurring on the
gold/quartz surface as the change of vibration frequency for the
oscillating quartz crystal is linearly related to the mass change
resulting from the adsorption of other species.15 Use of QCM
to follow the growth kinetics of the monolayer formation of
alkanethiol has been documented.16 In the case of C16 thiol,
the monolayer adsorption process is completed within 20 min,
shown by a stabilized frequency following a quick drop. Control
experiments with bare gold surface exposed to H2S vapor
exhibited a frequency decrease over a much longer period and
leveled off only after several hours. The mass increase corre-
sponds to multilayer adsorption of the hydrogen sulfide species
on the surface. This agrees with the previous observation that
a roughening effect resulted from H2S exposure.17 The displace-
ment of alkanethiolate was monitored for SAMs of three
different chain length by following the frequency change upon
H2S exposure. A fast increase of frequency was noted for all
these SAM-covered surfaces after exposure and rinsing with
organic solvent. In the case of C12 thiol, the frequency started
to decrease again after 2 h. The C16 thiol gave similar behavior
only at a later stage. In the case of C20 thiol, the gradual increase
of frequency continued after 10 h of exposure. In all cases, the
maximum mass loss estimated from the frequency increase is
less than that expected for the loss of a full monolyaer. The
interpretation of these data has to take into account the fact
that the displacement of a long-chain alkanethiol is accompanied
S(2p) peaks at 161.5 eV for S(2p3/2) and 162.8 eV for S(2p1/2
)
respectively in about a 2:1 ratio (Figure 3a), which is in
agreement with the reported binding energies of the surface-
bound sulfur from alkanethiolate.19,20 A strong C(1s) peak
occurred at 284.7 eV for the alkyl chains (Figure 4a). No O(1s)
peak was detected, which indicated the thiolate was the only
binding headgroup species. In contrast, the H2S-exposed bare
gold gave a broad and complicated S(2p) region, with two
apparent peaks at 161.1 and 162.1 eV (Figure 3c). Deconvo-
lution gave two sets of split peaks, each separated by ∼1.2 eV
with a 2:1 ratio, which implies more than one sulfur species
may be present. This may correlate to a mixture of atomic and/
or oligomeric surfur species and sulfhydryl groups, the latter
of which was suggested to exist in the H2S-exposed Au(110)
surface.21 A carbon (1s) peak at 284.0 was found, presumably
due to the residue carbonaceous contaminants on the surface
(Figure 4c). The higher binding energy for C(1s) for ordered
SAM than that of disordered carbon species was also docu-
mented.19 After the SAM surface was exposed to H2S vapor
for 6 h and rinsed with ethanol, the C(1s) shifted to 284.2 eV,
a position similar to the C(1s) peak observed for contaminant
carbonaceous species, yet the intensity has dropped to a similar
level as the H2S-exposed bare gold surface (Figure 4b). The
sulfur region also gave a similar pattern as that of H2S-exposed
(15) Buttry, D. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel
Dekker: New York, 1991; Vol. 17.
(18) Zamborini, F. P.; Crooks, R. M. Langmuir 1998, 14, 3279-3286.
(19) Ishida, T.; Hara, M.; Kojima, I.; Tsuneda, S.; Nishida, N.; Sasabe,
H.; Knoll, W. Langmuir 1998, 14, 2092-2096.
(20) Castner, D. G.; Hinds, K.; Grainger, D. W. Langmuir 1996, 12,
5083-5086.
(21) Fruhberger, B.; Grunze, M.; Dwyer, D. J. J. Phys. Chem. 1994, 98,
609-616.
(16) (a) Karpovich, D. S.; Blanchard, G. J. Langmuir 1994, 10, 3315-
3322. (b) Kim, H. J.; Kwak, S.; Kim, Y. S.; Seo, B. I.; Kim, E. R.; Lee, H.
Thin Solid Film 1998, 327-329, 191-194. (c) Chance, J. J.; Purdy, W. C.
Thin Solid Film 1998, 335, 237-244.
(17) Touzov, I.; Gorman, C. B. Langmuir 1997, 13, 4850-4854.

Desorption/Reorganization of Self-Assembled Monolayers
J. Am. Chem. Soc., Vol. 122, No. 29, 2000 7075
Figure 3. XPS spectra of the C(1s) region of the (a) Au/SC15CH3
monolayer before H2S exposure, (b) after H2S exposure for 6 h, (c)
bare gold exposed to H2S, and (d) the bare gold surface.
Figure 4. XPS spectra of the S(2p) region of (a) the Au/SC15CH3
monolayer before H2S exposure and (b) after H2S exposure for 6 h, (c)
the bare gold exposed to H2S, and (d) the bare gold surface.
bare gold surface (Figure 3b). It is concluded that the monolayer
material was lost after exposure to H2S vapor.
The Reflection Absorption IR Spectroscopy. While the
above characterizations require or result in the removal of
desorbed species from the surface, the reflection-absorption
IR spectroscopy examines the surface as is after H2S exposure.
It provides rich information regarding the detailed structural
change of the assembly upon H2S exposure. Depending on the
group exposing at the surface, two types of behavior were
observed:
(a) Monolayers with a nonpolar terminal group: Figure
5a shows the spectra for the monolayer of Au/SC15CH3 before
and after H2S exposure for various intervals of time. Before
the exposure, the vibrational frequencies for CH2 symmetric
and asymmetric stretches occurred at 2850.4 and 2818.5 cm-1
,
respectively, indicative of a densely packed and crystalline
assembly of the alkyl chains.12 With ambient storage, the
spectrum and thus the structure is stable to organic solvent wash.
However, upon exposure to H2S, the peaks start broadening on
the higher side. After several hours, the peaks shifted to 2855.1
and 2926.5 cm-1, respectively, indicating a disordered, liquidlike
film has resulted. Washing the surface with organic solvent
removed much of the material from the surface, as indicated
by ellipsometry measurement. The IR spectrum remained that
of a disordered film; nevertheless, the intensities did not decrease
by much. It should be noted that in the reflection mode, the
peak intensity is highly orientation-dependent.22 The intensity
change is complicated if both the amount and the orientation
are changing during the washing process.23 For C18 and longer
thiols, the situation is somewhat different. As shown in Figure
5b for C20 thiol, after exposure to H2S the intensities of the
Figure 5. Reflection-absorption IR spectra as a function of H2S-
exposure time for the monolayer of (a) Au/SC15CH3 and (b) Au/
SC19CH3.
two methylene stretches decreased, but the frequencies very
much maintained at the same positions as the starting film. This
implies that the film retained a crystalline packing and ordered
(22) Parikh, A. N.; Allara, D.L. J. Chem. Phys. 1992, 96, 927-945.

7076 J. Am. Chem. Soc., Vol. 122, No. 29, 2000
Tao et al.
Figure 6. Reflection-absorption IR spectra as a function of H2S-
exposure time for the monolayer of Au/SC15COOH.
Figure 7. Reflection-absorption IR spectra as a function of H2S-
exposure time for the monolayer of Au/SC15COOCH3.
structure. The decreasing intensity can be attributed to orien-
tational change, with the alkyl chains being even closer to the
surface normal.22 When the film surface was washed with
organic solvent, a disordered film resulted as suggested by the
frequencies at 2856.5 and 2926.4 cm-1. Material loss is again
indicated by ellipsometry measurement.
spectrum varied. With C12 acid, the spectrum after 1 h of
exposure is similar to that after 24 h, implying the desorption
process may have completed in a rather short period (1-2 h),
while for C20 acid, the spectra kept changing even after 20 h
of exposure.
The dramatic spectral change was reminiscent of the spectral
change when a SAM of n-alkanoic acid on Ag was exposed to
HCl or H2S vapor.8 In those cases, the adsorbed carboxylate
molecules were believed to be protonated and released from
the specific surface sites and reorganized into discrete, hydrogen-
bonded dimer crystallites, all resulting from a competition for
silver ion by the HCl or H2S. The dramatic increase in the
intensities of methylene stretching modes was due to a
preferential orientation of the chain axis of the crystallites to
be parallel to the surface.8 The similar observation seen here
implied that the carboxyl-terminated thiol underwent similar
desorption/aggregation as happened on the acid/silver surface.
Other SAMs with polar terminal functional groups, such as
ester, amide, cyano, and hydroxyl, gave similar spectra changes
upon H2S exposure, i.e., a substantial increase in the C-H
stretch intensities with peak frequencies characteristic of crystal-
line chain packing. Although these polar functional groups do
not form dimeric forms such as that of the carboxyl group,
analysis of the IR spectra also indicates that the polar end group
has experienced a change in the local environment, from that
in an isolated environment to that in a bulk material. For
example, Figure 7 shows the spectral change of methoxy-
carbonyl-terminated monolayer upon exposure to H2S. The
ν(CdO) for the ester group in the monolayer occurs at 1745.7
cm-1, 13 wavenumbers higher than that observed for bulk
crystals of the parent compound. This has been attributed to
the fact that the ester functionality is exposed to ambient in the
oriented monolayer.25 Groups in such an environment are
surrounded by fewer neighboring groups, while after exposure
(b) Monolayer with a polar terminal group: In contrast to
the saturated alkanethiols, monolayers carrying a polar terminal
group gave quite different results. For example, the result of a
monolayer of 16-mercaptohexadecanoic acid (Au/SC15COOH)
is shown in Figure 6. The film before H2S exposure gave the
expected spectrum, with the carbonyl stretch occurring at 1719.1
and 1741.8 cm-1, representing the hydrogen-bonded and isolated
carbonyl group, respectively.12 The C-H stretches occur at
2850.7 and 2918.8 cm-1, respectively, indicative of a well-
ordered packing of the chains. Upon exposure to H2S, the peaks
for the C-H stretches broaden toward the higher side in the
beginning but then narrow down again, with the peak position
maintained around 2919 cm-1 and intensities growing substan-
tially with time. The absorption for the carbonyl group also grew
and shifted to a single peak at around 1697.8 cm-1, which is
assigned as the carbonyl stretch for the H-bonded dimer of
carboxylic acid.24 Also growing in intensities are the deformation
bands for CH2 at 1472.6 and 1465.6 cm-1 and a series of
“progressional bands” between 1300 and 1150 cm-1. These peak
positions virtually match with the peak positions of the KBr
transmission spectrum of the bulk sample of the original 16-
mercaptohexadecanoic acid. The assembly was apparently
physisorbed as the material on the surface can be readily
removed now by organic solvent. Acid-terminated thiols of
different chain length (C12 and C20) gave similar spectral
change, except that the time required to reach the ultimate
(23) In fact, a bare gold surface went through similar H2S exposure and
solvent wash yielded a similar spectrum, presumably due to the tenaciously
bound contaminant. Thus it is not possible to assess quantitatively the loss
of material from the IR spectra.
(24) Bellamy, L. J. The Infrared Spectra of Complex Molecules; Wiley:
New York, 1975.
(25) An extreme limit of such frequency shift due to neighboring
molecules is the generally observed decrease in frequency when an oscillator
is taken from a gas-phase to a condensed-phase environment.

Desorption/Reorganization of Self-Assembled Monolayers
J. Am. Chem. Soc., Vol. 122, No. 29, 2000 7077
Figure 8. Reflection-absorption IR spectra as a function of H2S-
exposure time for the monolayer of Au/SC15CONH2.
Figure 9. Reflection-absorption IR spectra as a function of H2S-
exposure time for the monolayer of Au/SC16OH.
to H2S the ν(CdO) shifted to 1733.0 cm-1, the same as the
carbonyl stretch frequency in bulk sample. This implies the ester
group is in the local environment of a bulk state. In conjunction
with the intensity change and crystalline-like peak positions
(2917.7 and 2848.0 cm-1 for νa(CH2) and νs(CH2), respectively)
of the methylene vibration modes, formation of discrete crys-
tallites of bulk material is suggested. Again, all of these
characteristic peaks were lost after solvent wash, apparently due
to the loss of film materials by washing. Figure 8 shows the
spectral change of amide-terminated alkanethiol monolayer Au/
SC15CONH2. The ν(NH) for the terminal amide group changed
from 3497.2 cm-1 for the free state to 3394.2 and 3195.3 cm-1
for the associated state. Shifts in amide I and II bands in the
1600-1700 cm-1 region (smaller separation of the two bands
in the associated state compare to the wider separation in the
isolated state) lead to the same suggestion that the amide
functional group experienced a local environment change from
that of the interfacial state to that of a bulk state.24 Substantial
increase of the methylene intensities again implied molecular
chains lying closer to the surface. The IR spectra for Au/
SC16OH and Au/SC16CN are shown in Figures 9 and 10.
Substantial growth of methylene intensity was also observed.
On the basis of the above results, we propose that the
alkanethiolate moiety was displaced by H2S from the surface
and then reorganize according to the nature of the end group
and chain length. The observation that a small H2S molecule
can displace long-chain alkanethiolate is interesting. The
concentration effect (large excess of H2S present) should not
be the major reason. It has been known that the monolayer of
a long-chain thiol is more stable than that of the shorter chain
thiol, which is reflected by the reduction desorption potentials.
The potential required to desorb alkanethiolate shifted from
-0.77 V for ethanethiolate to -1.36 V for hexadecanethiolate.
The more negative potential is due to the contribution of the
van der Waals interaction of the alkyl chain. However, it is
noted that the reductive desorption potential for the sulfide
species resulting from H2S is around -0.89 V, more negative
than that of the short ethanethiolate. This would suggest the
Figure 10. Reflection-absorption IR spectra as a function of H2S-
exposure time for the monolayer of Au/SC16CN.
binding of the sulfide species is intrinsically stronger than that
of an alkylthiolate.
In the exchange study of alkanethiolate by another alkanethiol
in solution, it was suggested that the exchange start from defect
sites and there are sites that are not exchangeable and sites that
are exchangeable.3b,c Disulfide was implied, but not rigorously
proven, to be the form desorbed under exchange conditions.26
(26) Schlenoff, J. B.; Li, M.; Ly, H. J. Am. Chem. Soc. 1995, 117,
12528-12536.

7078 J. Am. Chem. Soc., Vol. 122, No. 29, 2000
Tao et al.
Figure 11. Schematic representation of reorganization of the self-assembled monolayer of alkanethiol derivatives as induced by H2S exposure.
In the current case, we found from XPS data that most of the
thiolate can be desorbed by H2S. To delineate the form for which
thiolate is desorbed, we used the closely related system, the
SAM-capped gold nanoparticles,27 which allows the use of NMR
spectroscopy to characterize the product. Close analogy in
structure between the SAM on a gold nanoparticle surface and
on a planar surface was suggested.28 Thus a similar exposure
experiment was carried out on the C16 thiol-covered nanopar-
ticle system. After H2S exposure, the nanoparticles were washed
with CDCl3 and filtered. The organic filtrate collected was
analyzed by the NMR technique. The NMR results29 indicated
that the alkanethiol was the major component released from
the surface, together with disulfide as a minor product. It is
noted that a similar conclusion was made when the place-
exchange reaction was carried out on a nanoparticle system.30
Thus the following reaction is implied
the thiol molecules were pinned on the surface via a strong
covalent Au-S bond, which dictates the spacing of the
headgroups. The chain-chain interaction provides the driving
force for a well-ordered packing of the assembly. The end
groups constitute the third contributing interaction. For nonpolar
end groups, such as methyl, a low interfacial energy is implied
whereas for polar end groups, the interfacial energy will be
higher. When the molecules are replaced by H2S, the strong
pinning interaction no longer exists. The system is open to a
new equilibrium state. Here the end groups apparently become
a determining factor. For all systems carrying a polar functional
group, the assembly undergoes a de-wetting type of reorganiza-
tion to give discrete crystallites, the size of which is determined
by the mobility of the molecule and the intermolecular forces.
In the case of nonpolar end group such as methyl, chain length
and thus the interchain interaction is the prevailing factor. With
C16 and shorter chains, a liquidlike disordered film was
obtained, reflecting the physical state of the parent molecule.
There is no strong enough interaction to sustain a well-ordered
packing. For C18 and longer chains, the oriented monolayer
state is still the most favorable state. Without a strong pinning
to specific binding site, an even more perpendicular alignment
of the chains and thus higher interchain interaction is implied
by the lower intensity yet crystalline-like IR characteristics. Such
is also the most favored orientation by calculation.31 Thus the
chemisorbed monolayer literally changed into a physisorbed,
Langmuir-type monolayer. The tendency for various function-
alized monolayers to reorganize can also be understood in terms
of optimization in interfacial energy. Those monolayers exposing
the polar terminal functional group have a higher surface free
energy. If the molecules are released from the binding sites, a
AuSR(s) + H2S(g) f AuSH(s) + RSH
where s means a surface adsorption site.
More interesting is the fate of the assembly after desorption.
An overall picture is shown in Figure 11. Before desorption,
(27) (a) Brust, M.; Walker, M.; Bethell, D.; Schiffrin, D. J.; Whyman,
R. J. Chem. Soc., Chem. Commun. 1994, 801-802. (b) Hostetler, M. J.;
Wingate, J. E.; Zhong, C. J.; Harris, J. E.; Vachet, R. W.; Clark, M. R.;
Londono, J. D.; Green, S. J.; Stokes, J. J.; Wignall, G. D.; Glish, G. L.;
Porter, M. D.; Evans, N. D.; Murray, R. W. Langmuir 1998, 14, 17-30.
(28) Terrill R. H.; Postlethwaite, T. A.; Chen, C. H.; Poon, C. D.; Terzis,
A.; Chen, A.; Hudchison, J. E.; Clark, M. R.; Wignall, G.; Londono, J. D.;
Superfine, R.; Falvo, M.; Johnson, C. S., Jr.; Samulski, E. T.; Murray, R.
W. J. Am. Chem. Soc. 1995, 117, 12537-12547.
(29) A quartet at δ 2.50 was assigned as the CH2 next to the SH of the
thiol whereas a triplet at δ 2.66 was assigned as the CH2 next to the S-S
of a disulfide. The two sets of signals exist in a ratio of ∼8:2.
(30) Hostetler, M. J.; Templeton, A. C.; Murray, R. W. Langmuir 1999,
15, 3782-3789.
(31) Ulman, A.; Eilers, J. E.; Tillman, N. Langmuir 1989, 5, 1147-
1152.

Desorption/Reorganization of Self-Assembled Monolayers
J. Am. Chem. Soc., Vol. 122, No. 29, 2000 7079
lower free energy state is achieved by having the molecules
reorganizing into bulk material, where the polar end groups are
buried in the bulk, a phenomena suggested to occur even in the
covalently attached state of a self-assembled monolayer upon
ambient storage.32 Attempts to image the reorganization process
by AFM have not been successful, presumably due to the
inherent rough morphology of polycrystalline substrate used in
the study. One may argue the role of the roughened surface on
the structure. It is contended that the roughened surface might
influence the kinetics of a surface reaction.33 Our conclusion
on the structure change was drawn mainly from thermodynamic
considerations and should hold irrespective of how fast it is
attained.
surface by exposure to H2S vapor. The process requires the H2S
to penetrate down to the metal surface and thus exhibit a chain
length effect, where a longer chain provides better protection
against the attack. The desorbed molecules seek a new equi-
librium structure, which very much depends on the terminal
functional group. With polar end groups such as carboxyl, ester,
amide, and hydroxyl groups, a reorganization of the assembly
into discrete but bulk-like clusters is suggested. This can be
seen as a process of minimization of surface energy, such that
the polar group can be removed from the very outer surface.
With a nonpolar end group such as the methyl group, the
oriented assembly is still the favored structure given enough
chain length.
Conclusion
In summary, it is shown that the self-assembled monolayer
of alkanethiol derivatives on gold can be displaced from the
Acknowledgment. We thank the National Science Council
of the Republic of China and Academia Sinica for financial
support. We also thank Dr. Shu-Hua Chien, Ms. Pei-Fen Ho,
and Been-Chih Liou for the XPS measurements and discussion.
(32) Evans, S. D.; Sharma, R.; Ulman, A. Langmuir 1991, 7, 156-161.
(33) For example, the rate of oxidation of alkanethiolate by ozone
depends on the roughness of the gold surface: Lee, M. T.; Hsueh, C. C.;
Freund, M. S.; Ferguson, G. S. Langmuir 1998, 14, 6419-6423. The growth
rate of polymethylene on Au depends on the roughness of gold: Seshadri,
K.; Sundar, V. A.; Tao, Y. T.; Lee, M. T.; Allara, D. L. J. Am. Chem. Soc.
1997, 119, 4698-4711.
JA000278R