pubs.acs.

org/Langmuir
© 2010 American Chemical Society

Characterization of Alkanethiol Self-Assembled Monolayers on Gold by

Thermal Desorption Spectroscopy
Johanna Stettner* and Adolf Winkler
Institute of Solid State Physics, Graz University of Technology, Petersgasse 16, A-8010 Graz, Austria

Received January 18, 2010. Revised Manuscript Received March 5, 2010

SAM formation of undecanethiol (UDT) and mercaptoundecanoic acid (11-MUA) on Au(111) and on gold foils,
using wet chemical preparation methods as well as physical vapor deposition (PVD) in UHV, has been studied by means
of thermal desorption spectroscopy (TDS), low energy electron diffraction (LEED), and Auger electron spectroscopy
(AES). The main aim of this study was to explore the possible application of TDS to characterize the quality of a SAM
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

and to determine its thermal stability. The influence of various parameters, like substrate pretreatment, film formation
method, and type of the functional end group, has been studied in detail. Three different temperature regimes can be
identified in TDS, which yields specific information about the organic layer: Desorption of disulfides around 400 K can
Downloaded via UNIV GRONINGEN on November 13, 2019 at 14:41:04 (UTC).

be shown to result from standing molecules in a well-defined SAM. Desorption of intact molecules and of molecules with
split-off sulfur is observed around 500 K, resulting from lying molecules. Finally, desorption of an appreciable amount
of gold-containing molecules is observed around 700 K. This is more pronounced for 11-MUA than for UDT, and in
addition more pronounced for solution-based SAMs than for PVD prepared SAMs. These results emphasize the
important role of gold adatoms in SAM formation, as recently discussed in the literature.

Introduction monolayers of the same quality as for methyl-terminated al-

During the past decades, the formation of self-assembled
monolayers (SAMs) on metal surfaces has attracted enormous
interest.1-6 They provide a wide range of applications, among
others in the fabrication of nanoscale materials, as biosensors, or
as highly ordered model systems.7-11 For such applications, a
detailed knowledge of the alignment of the molecules on the
surface and of their thermal stability is crucial.
The most frequently studied SAM system is that of alkanethiols
on gold. There is wide agreement that methyl-terminated, long-
chain alkanethiols form well-ordered monolayers on gold.12,13
However, in order to offer the opportunity for further chemical
modifications, alkanethiols with acid end groups are more
desirable.14-16 Even though extensive investigations on SAM
formation of acid-terminated alkanethiols have been performed
in the past years, it is not yet clear if one can obtain well-ordered
kanethiols.17-20
The common way to produce SAMs is to put the cleaned
substrate in ethanolic solution of the corresponding thiols for
approximately 24 h.2,3,21 Besides the preparation in solution, gas-
phase deposition of alkanethiols can also be performed.22-24
Even though it causes much more experimental effort, physical
vapor deposition (PVD) allows not only the investigation of the
early state of SAM formation but also characterization of the
SAM with surface analytical tools.
In order to investigate SAMs properly, appropriate experimen-
tal methods are required that not only show the existence of small
ordered domains (as provided by STM25,26), but give information
of the SAM arrangement over the whole sample surface, e.g.,
by GIXRD.27,28 A rather infrequently used method in this con-
text is thermal desorption spectroscopy (TDS),21,22,29-32 which,

*Corresponding author. Tel: þ43 316 8967, Fax: þ43 316 873 8466, e-mail:
hanni@sbox.tugraz.at.
(1) Li, X. M.; Huskens, J.; Reinhoudt, D. N. J. Mater. Chem. 2004, 14, 2954.
(2) Love, J. C.; Estroff., L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M.
Chem. Rev. 2005, 105, 1103.
(3) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151.
(4) Schreiber, F. J. Phys. (Paris) 2004, 16, 881.
(5) Evans, S. D.; Williams, L. M. Funct. Org. Polym. Mater. 2000, 149.
(6) Ulman, A. Chem. Rev. 1996, 96, 1533.
(7) Ahn, H.; Cuong, P. D.; Park, S.; Kim, Y.; Lim, J. Sci. Direct 2003, 255, 819.
(8) Cavalleri, O.; Natale, C.; Stroppolo, M. E.; Relini, A.; Cosulich, E.; Thea, S.;
Novi, M.; Gliozzi, A. Phys. Chem. Chem. Phys. 2000, 2, 4630.
(9) Cechet, F.; Marcaccio, M.; Margotti, M.; Paolucci, F.; Rapino, S.; Rudolf,
P. J. Phys. Chem. B 2006, 110, 2241.
(10) Leopold, M. C.; Bowden, E. F. Langmuir 2002, 18, 2239.
(11) Nakano, K.; Yoshitake, T.; Yamashita, Y.; Bowden, E. F. Langmuir. 2007,
23, 6270.
(12) Nuzzo, R. G.; Dubois, L. H.; Allara, D. L. J. Am. Chem. Soc. 1990, 112,
558.
(13) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. J. Am. Chem. Soc.
1987, 109, 3559.
(14) Czanderna, A. W.; King, D.; Spaulding, D. J. Vac. Sci. Technol., A 1991, 5,
2607.
(15) Duevel, R. V.; Corn, R. M. Anal. Chem. 1992, 64, 337.
(16) Frey, B. L.; Corn, R. M. Anal. Chem. 1996, 68, 3187.
(17) Arnold, R.; Azzam, W.; Terford, A.; W€oll, Ch. Langmuir 2002, 18, 3980.
(18) Mendoza, S. M.; Arfaoui, I.; Zanarini, S.; Paolucci, F.; Rudolf, P.
Langmuir 2007, 23, 582.
(19) Wang, H.; Chen, S.; Li, L.; Jiang, S. Langmuir 2004, 7, 2633.
(20) Creager, S. E.; Steiger, C. M. Langmuir 1995, 11, 1853.
(21) K€afer, D.; Witte, G.; Cyganik, P.; Terfort, A.; W€oll, Ch. J. Am. Chem. Soc.
2006, 128, 1723.
(22) Kondoh, H.; Kodama, C.; Sumida, H.; Nozoye, H. J. Chem. Phys. 1999, 3,
1175.
(23) Schreiber, F.; Eberhardt, A.; Leung, T. Y.; Schwartz, P.; Wetterer, S. M.;
Lavrich, D. J.; Berman, L.; Fenter, P.; Eisenberger, P.; Scoles, G. Phys. Rev. B
1998, 19, 12476.
(24) Balzer, F.; Gerlach, R.; Polanski, G.; Rubahn, H.-G. Chem. Phys. Lett.
1997, 274, 145.
(25) Gorman, C. B.; He, Y.; Carroll, R. L. Langmuir 2001, 17, 5324.
(26) Ito, E.; Konno, K.; Noh, J.; Kanai, K.; Ouchi, Y.; Seki, K.; Hara, M. Appl.
Surf. Sci. 2005, 244, 584.
(27) Cossaro, A.; Mazzarello, R.; Rousseau, R.; Casalis, l; Verdini, A.; Kohlmeyer,
A.; Floreano, L.; Scandolo, S.; Morgante, A.; Klein, M. L.; Scoles, G. Science 2008,
321(5891), 943.
(28) Fenter, P.; Eisenberger, P. Phys. Rev. Lett. 1993, 70(16), 2447.
(29) Hayashi, T.; Wakamatsu, K.; Ito, E.; Hara, M. J. Phys. Chem. 2009, 113,
18795.
(30) Lee, J. G.; Lee, J.; Yates, J. T. J. Phys. Chem. B 2004, 108, 1686.
(31) Lee, J. G.; Lee, J.; Yates, J. T. J. Am. Chem. Soc. 2004, 126, 440.
(32) Rzeznicka, I. I.; Lee, J.; Maksymovych, P.; Yates, J. T. J. Phys. Chem. B
2005, 109, 15992.

Langmuir 2010, 26(12), 9659–9665 Published on Web 03/26/2010 DOI: 10.1021/la100245a 9659
Article
however, can be carried out quite easily in a UHV chamber. The
crucial point is the proper analysis of the TD spectra, which in the
case of desorption of large organic molecules can be quite
complicated.
In this work, we focus on the alkanethiols undecanethiol
(UDT) and 11-mercaptoundecanoic acid (11-MUA) on gold.
These molecules differ only in their functional end group (-CH3
and -COOH) and thus provide a suitable pair in order to
investigate its influence on the SAM formation. Furthermore,
we compare the ex situ (from solution) with the in situ (by PVD)
preparation with respect to the SAM formation. Finally, we
report on the importance of the substrate preparation (Piranha
solution, argon ion sputtering) in this context.
The main analytical technique used in this work is TDS. To
facilitate the interpretation of the TD spectra, low energy electron
diffraction (LEED) and Auger electron spectroscopy (AES)
measurements were additionally performed. It will be shown that
TDS can provide a powerful tool to characterize the SAM. This
method is particularly suitable to shed light onto the specific role
of Au-adatoms in the SAM formation.27,33-36
Experimental Section
All experiments were conducted in a UHV chamber with a base
pressure of 1  10-10 mbar, but most of the experiments were
performed without baking the system, leading to a background
pressure in the low 10-8 mbar range. Two different substrates
were used: a Au(111) single crystal (2 mm thick, Ø 10 mm) and
gold foils (0.1 mm thick, 10 mm  10 mm). The single crystal
sample was suspended on thin, 0.2-mm-diameter Ta wires which
were inserted in a thin groove at the rim of the sample and
clamped to the sample holder. The gold foils were fixed to a
stainless steel plate via thin Ta wires, which was also connected to
the sample holder via tantalum wires.37 Both samples could be
heated by resistive heating via the Ta wires up to 1000 K and
cooled by liquid nitrogen down to 100 K. The temperature was
measured by a Ni/NiCr thermocouple attached to the backside of
the samples. Typically, a heating rate of 1 K/s was used. In order
to clean the substrates, Arþ-sputtering with subsequent annealing
was applied. The cleanliness of the sample surfaces was checked
by an Auger electron spectrometer and the surface structure of the
single Au crystal was confirmed by LEED.
In order to perform deposition of UDT on the sample in UHV,
the liquid material was stored in a glass container and the vapor,
according to the vapor pressure at room temperature, was
supplied via a leak valve and a 2 mm ID tube directly onto the
sample (tube exit-sample distance: 20 mm). The actual exposure
is a combination of the direct impingement of the molecules and
the isotropic impingement due to the partial pressure increase of
UDT in the vacuum chamber. Strictly speaking, the description of
the exposure in terms of Langmuir (1 L = 10-6 Torr s = 0.75 
10-6 mbar s) is therefore not appropriate. However, in order to
obtain a good measure of the actual exposure, the following
procedure was applied: In a first experiment, UDT was fed
directly onto the cooled sample (inline dosing) for a certain time,
followed by thermal desorption of the layer. In a second experi-
ment, UDT was fed into the chamber for an equally long time,
while the sample was turned away from the inlet tube (isotropic
dosing) and again the desorption peak was detected. The ratio
between the amounts of desorbing particles yields the so-called
enhancement factor for direct dosing, which turned out to be ∼40
(33) Woodruff, D. P. Phys. Chem. Chem. Phys. 2008, 10, 7211.
(34) Maksymovych, P.; Sorescu, D. C.; Yates, J. T. Phys. Rev. Lett. 2006, 97,
1461031.
(35) Voznyy, O.; Dubowski, J. J.; Yates, J. T.; Maksymovych, P. J. Am. Chem.
Soc. 2009, 131, 12989.
(36) Kautz, N. A.; Kandel, S. A. J. Phys. Chem. C 2009, 113, 19286.
(37) Frank, P.; Hlawacek, G.; Lengyel, O.; Satka, A.; Teichert, C.; Resel, R.;
Winkler, A. Surf. Sci. 2007, 601, 2152.
9660 DOI: 10.1021/la100245a
Stettner and Winkler
in our case. As the isotropic partial pressure increase was typically
in the range of 2  10-8 mbar, one can assume the actual pressure
at the sample surface in front of the inlet tube to be 40 times
higher, leading to a pressure of 8  10-7 mbar. The values of
exposure given in this survey are thus calculated by the pressure
increase, corrected with the enhancement factor, and multiplied
with the time of evaporation.
The 11-MUA is a solid at room temperature and therefore had
to be heated up to about 50
C to get a liquid with an appropriate
vapor pressure. Again, the vapor was fed onto the sample via a
leak valve and a 2 mm ID stainless steel tube. The whole gas inlet
system was permanently held at 150
C in order to avoid
recondensation on the inner wall of the gas inlet tube (tube
exit-sample distance: 20 mm). In this case, no significant iso-
tropic pressure increase in the chamber could be observed, due to
the much smaller vapor pressure. Thus, the actual exposure is just
given by the exposure time and the vapor pressure according to
the temperature of the substance in the container. The vapor
phase was purified by several freeze/thaw cycles and differential
pumping of the gas line. The cleanliness of the vapor phase was
checked by feeding the gas directly into a mass spectrometer.
The ex situ preparation of the SAMs on gold foils was
performed on differently prepared samples: (a) cleaned in Piranha
solution38 and (b) Arþ-sputtered and annealed in UHV. The in-
solution SAM preparation on Au(111) was always performed on
sputter-cleaned samples. AES revealed that some carbon impu-
rities were still present after cleaning the substrates in Piranha
solution, while Arþ-sputtering yields an uncontaminated surface,
which remains clean even after short exposure to air. Subse-
quently, the substrates were put in 1 mM ethanolic solution of
the corresponding molecule for 24 h. Thereafter, the samples were
dried with CO2 spray and installed immediately into the UHV
chamber for further characterization.
Different surface analytical techniques were available: AES
provided the opportunity to determine the chemical composition
of the SAM, though one should keep in mind that impinging
electrons can destroy the organic layer. In order to avoid sig-
nificant damage, the electron emission currents were kept below
0.4 μA. Furthermore, LEED optics with a multichannel plate was
used, which can be operated at very low emission currents to
reduce damaging of the SAM. A quadrupole mass spectrometer
(QMS) with a mass range from 1 to 500 amu was used to control
the residual gas phase as well as to perform TDS. For this
purpose, the desorption flux was monitored by the multiplexed
mass spectrometer in in-line position.
Results and Discussion
Vapor Deposition of UDT and 11-MUA on Gold. SAM
Characterization by LEED. Even though TDS was the method
the focus was put on, the LEED measurements on a Au(111)
single crystal will be described first, as their results are essential to
fully understand the desorption spectra. An important issue to
obtain good LEED patterns is a careful substrate preparation.
Therefore, Arþ-sputtering of the Au(111) crystal and contem-
poraneous annealing up to 900 K was performed for several
hours. Figure 1a shows the LEED pattern which was observed
immediately after an exposure of 50 L of UDT at room tempera-
ture. The LEED observation was performed at liquid nitrogen
temperature, in order to increase the signal/noise ratio. A similar
pattern was observed√for 11-MUA. These patterns can be best
described by a (12  3) rectangular structure (see Figure 1b), as
obtained by an appropriate simulation. The simulation was
performed using LEEDSIM;39 the input parameters were the
Au-Au interatomic distance a = 0.288 nm and the length of the
(38) Caution: Piranha solution (70% H2SO4, 30% H2O2) reacts violently with
most organic materials and must be handled with extreme care.
(39) http://www.sim4tec.com/?Products:LEEDsim_-_LEED_simulation_software.
Langmuir 2010, 26(12), 9659–9665
Stettner and Winkler
Figure 1. Measured LEED pattern immediately after 1 min eva-
poration of UDT on Au(111); beam energy: √ U = 41 V (a). In
addition, the calculated pattern of a (12  3) rectangular struc-
ture is shown, the simulation was performed with LEEDSIM (b).
This pattern can be attributed to molecules lying on the surface, as
shown in the scheme (c).
40
UDT molecule, which is l = 1.7 nm. According to the litera-
ture,33 the unit cell is assumed to contain two lying molecules,
oriented with their carbon chains parallel to the surface along the
Æ1-10æ direction and with their sulfur groups at the opposite ends.
The corresponding coverage is then 0.09 ML with respect to the
Au(111) surface. Further deposition of UDT leads to a weakening
of the LEED pattern, and finally, it vanishes after an exposure of
about 500 L.
However, waiting overnight and repeated evaporation resulted
in a new diffraction pattern in the case of UDT, as √depicted
in Figure 2a. This pattern can be attributed to a (2 3  3)
rectangular structure, which is well-known to be formed by
alkanethiols on Au(111).33,34 The simulated pattern is shown in
Figure 2b. This structure is assumed to stem from 0.33 ML
(40) Hobara, D.; Sasaki, T.; Imabayashi, S.; Kakiuchi, T. Langmuir 1999, 15,
5073.
Langmuir 2010, 26(12), 9659–9665
Article
Figure 2. Measured LEED pattern of UDT on Au(111), two days
after the evaporation of the same amount as in Figure 1, beam
energy U = √28 V (a). The calculated pattern is shown in (b),
revealing a (2 3  3) rectangular structure. The encircled spots are
not visible in the measured LEED pattern due to the existence of
glide planes, as described in the literature.24 This pattern can be
attributed to a SAM of standing molecules, as shown in the scheme,
where the arrangement of the head groups are shown (c).
standing molecules with the carbon chain tilted ∼30
from the
33
surface normal, which implies 4 molecules per unit √ cell.√ The
Wood description of this structure would be a c(4 3  2 3)30

structure, with 8 molecules in the unit cell. Heating such a SAM
up to 450 K resulted again in the pattern of the lying molecules.
Basically, these LEED patterns are in good agreement with the
results obtained by Gerlach et al.41 for decanethiol on Au(111). As
to the 11-MUA molecule, such a pattern could not be observed
under equivalent preparation conditions.
SAM Characterization by TDS. Multilayer Desorption.
The main endeavor of this work was to obtain a deeper under-
standing of the SAM formation using TDS. Most of the TD
experiments were performed on gold foils, but TDS on the
Au(111) single crystal yields almost the same results. In order to
determine the cracking pattern of UDT and 11-MUA in the QMS
(41) Gerlach, R.; Polanski, G.; Rubahn, H.-G. Appl. Phys. A: Mater. Sci.
Process. 1997, 65, 375.
DOI: 10.1021/la100245a 9661
Article
Figure 3. Some relevant masses of the cracking pattern, character-
izing the desorption of a multilayer of UDT (a) and of 11-MUA
(b) from a gold surface, after physical vapor deposition in UHV
at 200 K.
upon ionization, first the desorption of a multilayer was investi-
gated, because it can be assumed that in this case only intact
molecules desorb. For this purpose, the sample was cooled to
200 K and a sufficient amount of molecules was deposited. The
multilayer desorption spectra of UDT and 11-MUA from a gold
foil are shown in Figure 3a,b, respectively. For UDT, besides the
mass of the intact molecule (188 amu), two prominent cracking
products are shown: 55 amu, and 43 amu, produced in the QMS.
These masses can be attributed to CxHy fragments of the alkyl
chain. They are thus representative of the intactly desorbing UDT
molecules in the TD spectra. A detailed description of the
cracking pattern of 11-MUA can be found in a previous work;42
some important cracking products are 41 amu, 55 amu, and
200 amu. The intensity of the intact 11-MUA molecule (218 amu)
in the mass spectrum is very weak and is therefore not used. The
highest mass which is detectable with a sufficient sensitivity is
mass 200. This mass stems most likely from the separation of
the -OH group, in addition to the S-H bond splitting during
ionization.
For the quantitative evaluation of the multilayer TD spectra,
the QMS signal (m = 43 amu (UDT), m = 41 amu (11-MUA))
was calibrated, assuming that the desorption peak area of the
monolayer of lying molecules (β-peak, vide infra) contains 1.3 
1014 molecules/cm2, as deduced from LEED. This corresponds to
a coverage of 9% of the Au(111) surface, which provides 1.4 
1015 surface atoms.
Thereby, the multilayers of UDT and 11-MUA, as shown in
Figure 3a,b, can be calculated by comparing the areas of the
multilayer and monolayer TDS to contain 1  1016 and 5  1016
molecules, respectively. Taking into account the density of UDT
(42) Frank, P.; Nussbacher, F.; Stettner, J.; Winkler, A. In Interface controlled
organic thin films, Rubahn, H. G., Sitter, H., Horowitz, G., Al-Shamery, K., Eds.;
Plenum Press: Berlin, 2009; Vol. 129, p 107.
9662 DOI: 10.1021/la100245a
Stettner and Winkler
Figure 4. Evolution of the desorption spectrum of UDT on Au-
(111) as a function of the waiting time ((a) mass 43, (b) mass 188,
(c) mass 374) in the monolayer regime (adsorption temperature,
300 K; heating rate, 1 K/s). The β-peak can be attributed to lying
molecules, the R-peak stems from standing molecules. The latter
are shown to preferably desorb as disulfides.
and 11-MUA (0.8 g/cm3 and 1 g/cm3, respectively), one can
determine the mean thickness d of the investigated multilayers.
For the films presented in Figure 3a,b, the calculation yields
d = 60 nm in the case of UDT and d = 200 nm in the case of
11-MUA.
From the determination of the adsorbed amount, and hence
the absolute desorption rate, the desorption energy Edes and the
frequency factor ν can be obtained from the multilayer peak,
using the Polanyi-Wigner equation for zero-order desorption.43
This equation connects these quantities with the desorption rate
Rdes, the (constant) number of molecules per square unit Θ, and
the temperature T.
Edes
Rdes ¼ Θ 3 ν 3 eð - kT Þ
An Arrhenius type plot of ln(Rdes/Θ) vs 1/(kT) yields the
desorption energy Edes from the slope of the straight line and
ν from the y-axis intercept. Θ is the coverage of the actual outer-
most layer, which is assumed to contain 3  1014 molecules (con-
sisting of a mixture of lying and standing molecules). This results
in Edes = (0.8 ( 0.1) eV and ν = 6  10(16(1) Hz in the case of
UDT and Edes = (1.0 ( 0.1) eV and ν = 6  10(17(1) Hz in the
case of 11-MUA. The values for ν are significantly higher than the
usually used values of ν = 1  1013 Hz. However, it has already
been shown that this value only holds for atoms and small
molecules, and that for larger organic molecules, sometimes very
high frequency factors are observed.44-47
Monolayer Desorption. In order to grow a monolayer of
UDT and 11-MUA, the sample was held at room temperature,
thus avoiding multilayer formation, but it turned out that again
(as already verified by LEED) the waiting time after the evapora-
tion strongly influences the desorption behavior. In Figure 4,
the desorption spectra of UDT on the Au(111) single crystal are
shown for different preparation procedures. In addition to the
mass of the intact molecule (mass 188, Figure 4a), the cracking
(43) Redhead, P. A. Vacuum 1962, 12, 203.
(44) Winkler, A. In Interface controlled organic thin films; Rubahn, H. G., Sitter,
H., Horowitz, G., Al-Shamery, K., Eds.; Plenum Press: Berlin, 2009; Vol. 129, p 29.
(45) Paserba, K. R.; Gellman, A. J. Phys. Rev. Lett. 2001, 86, 4338.
(46) M€ullegger, S.; Winkler, A. Surf. Sci. 2006, 600, 1290.
(47) Tait, S. L.; Dohnalek, Z.; Campbell, C. T.; Kay, B. D. J. Chem. Phys. 2005,
122, 164708.
Langmuir 2010, 26(12), 9659–9665
Stettner and Winkler Article

mass 43 is shown (Figure 4b). The spectra were obtained after an

exposure of 500 L of UDT at room temperature and different
waiting times: (a) few seconds, (b) one day, and (c) three days after
the deposition, respectively. While the first one is characterized by
a single desorption peak at around 500 K (β-peak), the evolution
of a second peak at around 400 K (R-peak) can be seen in the latter
two spectra. With the results of LEED, these two peaks can be
attributed to standing (R-peak) and lying (β-peak) molecules,
respectively. The area under the desorption curve can be corre-
lated with the amount of desorbing particles. Remarkably, this
area is significantly increased after one day waiting time. This is
explained by postadsorption of UDT due to the fact that UDT is
quite difficult to pump, leading to a background pressure of UDT
over a period of several hours after the deposition. While the Figure 5. Comparison of the mass 188 (intact UDT) and mass 187
SAM (of standing molecules) is formed, bare substrate areas are (corresponding thiolate) in the desorption spectrum of UDT from
generated on which the UDT molecules from the residual gas can Au(111). A heating rate of 1 K/s was applied. In the inset, the
further adsorb. Thus, while the lying molecules of the saturated multilayer desorption of the corresponding masses is shown, which
monolayer slowly form patches of well-aligned standing mole- can be assumed to stem from intact molecules. This demonstrates
cules (SAM) via a thermally activated process, bare substrate that monolayer desorption of UDT mainly proceeds in the form of
thiolates. For more details, see text.
areas are generated on which subsequently molecules from the
still relatively high partial pressure adsorb and are incorporated in
the SAM.
Along with the R-peak, a desorption peak of mass 374 emerges
with increasing waiting time, as shown in Figure 4c. This gives
clear evidence that the standing molecules desorb as disulfides. In
fact, the cracking pattern of the R-peak differs slightly from that of
the β-peak, indicating that not the same molecular species desorb
from the surface in both cases. This implies that mass 43 is the
cracking of a single UDT molecule in the case of the β-peak but the
cracking of a disulfide in the case of the R-peak. Apparently, for a
densely packed monolayer of standing molecules a rather strong
sulfur-sulfur interaction exists, leading to the recombinative
desorption of disulfide molecules. As a consequence, this leads
to a weakening of the S-Au bond and therefore to a lower
desorption temperature of the disulfide than for the single mole-
cules. The tendency of alkanethiols to form disulfides has already
been reported previously in the literature.22,29,48,49
Further detail on these peaks is shown in Figure 5, where the Figure 6. Desorption spectrum of UDT on a gold foil (heating
mass ratio 187/188 is considered in the monolayer and the rate: 1 K/s). The R-peak is again attributed to standing molecules,
desorbing as disulfides, the β-peak stems from the remaining lying
multilayer regime of UDT on a Au(111) surface. The monolayer
molecules. In comparison to the UDT SAM on the Au(111) single
was prepared by an exposure of 500 L of UDT at room crystal, the R-peak is less pronounced.
temperature and a waiting time of 3 days, the multilayer was
prepared by applying an exposure of 1000 L of UDT at a sample after three days waiting. This is explained by the polycrystallinity
temperature of 200 K. The mass 187 corresponds to the thiolate. of the gold foil, which exposes surfaces with different packing
There is a significantly different 187/188 mass ratio in the multi- densities.
layer peak as compared to the R- and β-peaks. As we know that Comparing these results of UDT to the desorption behavior of
the R-peak stems from desorption of disulfides, this is not 11-MUA (Figure 7a,b), one can find similarities as well as
surprising. We would rather assume that in this case hardly any differences. While the desorption of lying molecules again takes
cracking mass 188 should be observed. The fact that mass 188 is place at around 500 K, even after 14 days only a little SAM
detected, nevertheless, is attributed to some cross-talking of the formation (R-peak) could be observed, in agreement with the
QMS. Interestingly, the 188/187 ratio is similar for the β-peak, absence of the corresponding LEED pattern. One interesting
which indicates that almost all lying molecules desorb as thiolates. feature is the desorption of mass 34 (H2S) which takes place at a
This has already been argued for the similar system octanethiol on somewhat higher temperature than the β-peak (designed β0 -peak).
Au(111).29 The concomitant desorption of mass 185 (the sulfur-free 11-
In order to investigate the influence of the substrate on the MUA molecule) indicates the breaking of the sulfur-carbon
SAM formation, TD spectra of UDT on a gold foil are shown in bond, as already described previously.48
Figure 6a,b,c. Again, the desorption spectra after different wait- SAM Characterization by AES. For further characterization
ing times are compared. While the spectra immediately and 1 day of the SAM formation, AES has been performed. AES on
after the deposition barely differ from the desorption of the alkanethiols has the advantage of high sensitivity for the sulfur
Au(111), one can see that the R-peak is much less pronounced (in comparison to XPS). Although one has to keep in mind a
possible influence of the impinging electrons on the SAM ar-
(48) Stettner, J.; Frank, P.; Griesser, T.; Trimmel, G.; Schennach, R.; Gilli, E.; rangement, as already mentioned in the Experimental Section,
Winkler, A. Langmuir 2009, 25, 1427.
(49) Voets, J.; Gerritsen, J. W.; Grimbergen, R. F.; van Kempen, H. Surf. Sci. useful fingerprints can be expected. To check the reliability of the
1998, 399, 316. spectra, we performed repeated measurements on the same

Langmuir 2010, 26(12), 9659–9665 DOI: 10.1021/la100245a 9663

Article
Figure 7. Desorption spectra (heating rate: 1 K/s) of 11-MUA on
gold immediately after deposition at 200 K (a) and after a waiting
time of 14 days (b). Only little SAM formation takes place within
time (R-peak). Desorption of mass 34 (H2S) at a slightly higher
temperature than the β-peak (designated β0 -peak) indicates the
cracking of the sulfur-carbon bond.
Figure 8. AES sulfur signal of three differently prepared UDT
layers by PVD. A clear shift of the Auger signal toward higher
energies, starting from the multilayer (intact molecules), via the
monolayer of lying molecules to the monolayer of standing mole-
cules (SAM) can be observed. This can be used as a fingerprint for
SAM formation.
sample position up to ten times. Thereby we could not find any
change of the peak positions or heights. The interesting finding is
that we could observe a clear shift of the sulfur peak dependent on
the layer preparation: In Figure 8, the Auger spectra in the sulfur
region for a monolayer immediately after the UDT deposition at
room temperature and three days after the deposition are shown.
For comparison, the sulfur signal of a multilayer of UDT is
added. There is an energy shift of 1.1 eV between the multilayer
and a monolayer of lying molecules and a further shift of 1.8 eV
9664 DOI: 10.1021/la100245a
Stettner and Winkler
Figure 9. Desorption spectrum of an UDT SAM prepared ex situ
in ethanolic solution. A heating rate of 1 K/s was used. The
substrate was previously cleaned by Arþ-sputtering and annealed
at 900 K. The spectrum is similar to that of a SAM prepared by
PVD, showing the R-peak of the standing molecules and the β-peak
of the remaining lying molecules. The inset shows the mass of
374 amu, indicating the desorption of disulfides around 400 K
(R-peak). A small H2S peak can also be observed (β0 -peak).
Figure 10. Multiplexed desorption spectrum (heating rate: 1 K/s)
of an in-solution prepared 11-MUA SAM on a sputtered and
annealed gold substrate. The R-peak of the standing molecules and
the β0 -peak of the lying molecules are followed by a third desorp-
tion feature (γ-peak). This feature can be attributed to desorption
of gold containing molecules (mAu = 197 amu).
between a lying monolayer and a standing monolayer. Thus, the
AES S-signal can be used as a fingerprint to characterize the
quality of a SAM.
Solution Preparation of UDT and 11-MUA on Gold.
SAM Characterization by TDS. In Figure 9, the TD spectrum
of the ex situ prepared UDT SAM on a sputtered and annealed
gold foil is shown. The spectrum basically corresponds to that of
an evaporated SAM after long waiting time, characterized by
desorption of standing molecules at 400 K (R-peak), followed by
the β-peak of the remaining lying molecules. The standing
molecules again desorb as disulfides, indicated by desorption of
mass 374 (not shown).
The desorption spectrum of 11-MUA following the same
preparation procedure is depicted in Figure 10. Besides the peaks
of the standing (R-peak) and the lying (β-peak) molecules, a third
desorption peak at 700 K develops, designated as γ-peak. This
peak is characterized by desorption of a significant amount of
mass 197, in addition to other cracking products of 11-MUA.
Since m = 197 is the mass of gold, this is a strong indication that
desorption of gold-containing molecules takes place in this
temperature range. This gives evidence that the acid end group
strongly influences the SAM formation, showing a complex
Langmuir 2010, 26(12), 9659–9665
Stettner and Winkler
Figure 11. Multiplexed desorption spectrum (heating rate: 1 K/s)
of UDT on gold, prepared ex situ in ethanolic solution. The
substrate was cleaned in Piranha solution. The small R-peak
indicates only a small amount of standing molecules; the high
β-peak reveals a much higher amount of molecules lying on the
surface. The γ-peak characterized by mass 197 again denotes
desorption of gold-containing molecules.
binding mechanism leading to gold etching at elevated tempera-
ture. In a previous work,48 we attributed this desorption feature to
a cracking of the 11-MUA. However, since a similar peak also
shows up during desorption of UDT from Au(111) (see below),
we are quite sure that m = 197 can be attributed to Au atoms re-
sulting from the cracking of gold containing organic molecules in
the QMS. In fact, the existence of methylthiolate-Au-adatom
complexes in the context of methylthiolate self-assembling on
Au(111) has already been suggested in the group of Yates34,35
according to their high-resolution STM experiments. Also, resi-
duals of Au-atoms in the ethanolic solution of alkanethiols after
SAM formation were found by Edinger et al.,50 indicating the
existence of such Au-organic molecule complexes. However, the
desorption of such thiol-gold compound molecules at elevated
temperatures has not yet been described in the literature to the
best of our knowledge.
A similar desorption feature was also found for UDT after
cleaning the gold foil in Piranha solution, as shown in Figure 11.
Again, a desorption peak of mass 197 at high temperature
appears. Apparently, during the removal of carbon impurities
(50) Edinger, K.; G€olzh€auser, A.; Demota, K.; W€oll, Ch.; Grunze, M. Langmuir
1993, 9, 4.
Langmuir 2010, 26(12), 9659–9665
Article
in Piranha solution a rough gold surface is generated, which leads
to a strong gold adatom-thiol binding complex. As a result of this
ill-defined surface, also, the SAM formation is less pronounced,
as indicated by the small R-peak. The important message of these
investigations is that a strong correlation exists between the
desorption of gold-containing molecules and both the functional
end group and the surface quality.
Summary and Conclusions
Undecanethiol (UDT) and 11-mercaptoundecanoic acid
(MUA) layers were prepared on gold foils and on a Au(111) single
crystal surface, using wet chemical preparation methods, as well
as physical vapor deposition in UHV. In case of UDT, the
formation of a well-ordered SAM on a properly Arþ-sputtered
and annealed Au(111) substrate could be obtained both by in situ
and ex situ layer preparation, as verified by LEED. The crucial
step is a good surface preparation (cleaning and annealing) and
a long waiting time (1-2 days) after PVD as well as in the case of
ex situ preparation. For √
the as-such prepared layers on Au(111),
the observation of a (2 3  3)rectangular LEED structure is
indicative of a good SAM of standing molecules. Some of
the standing molecules desorb already at 400 K in the form
of disulfides (R-peak). The remaining molecules lay down
and form√ a well-ordered layer of lying molecules, exhibiting a
(12  3)rectangular structure. Most of these molecules then
desorb at around 500 K in the form of thiolates (β-peak), but in
some cases, the C-S bond breaks, leading to subsequent desorp-
tion of the fragment and H2S (β0 -peak).
In the case of 11-MUA on gold, comparable preparation
procedures (ex situ and in situ) yield only little SAM formation.
Following PVD, even 14 days after deposition still mainly lying
molecules are observed. An interesting feature, which is most
pronounced for ex situ prepared 11-MUA layers (but also for ex
situ prepared UDT after Piranha solution cleaning), is the
appearance of a desorption peak at 700 K, which can be
attributed to the desorption of gold-containing thiol molecules.
Thus, thermal desorption spectroscopy is a perfect tool to
characterize the SAM.
Acknowledgment. We are grateful to Gregor Trimmel, Thomas
Griesser and Gregor Hlawacek for helpful discussions. We thank
the Austrian Science Fund (FWF), P19197 for the financial
support.
DOI: 10.1021/la100245a 9665