Characterization of self-assembled monolayers of fullerene derivatives on gold surfaces: Implications for device evaluations

Article abstract of DOI:10.1021/ja063451d

The widely employed approach to self-assembly of fullerene derivatives on gold can be complicated due to multilayer formations and head-to-tail assemblies resulting from the strong fullerene fullerene and fullerene-gold interactions. These anomalies were

Full text of DOI:10.1021/ja063451d

Published on Web 09/23/2006
Characterization of Self-Assembled Monolayers of Fullerene
Derivatives on Gold Surfaces: Implications for Device
Evaluations
Yasuhiro Shirai, Long Cheng, Bo Chen, and James M. Tour*
Contribution from the Department of Chemistry, Smalley Institute for Nanoscale Science and
Technology, Rice UniVersity, MS-222, 6100 Main Street, Houston, Texas 77005
Received May 17, 2006; E-mail: tour@rice.edu
Abstract: The widely employed approach to self-assembly of fullerene derivatives on gold can be
complicated due to multilayer formations and head-to-tail assemblies resulting from the strong fullerene-
fullerene and fullerene-gold interactions. These anomalies were not examined in detail in previous studies
on fullerene self-assembled monolayers (SAMs) but were clearly detected in the present work using surface
characterization techniques including ellipsometry, cyclic voltammetry (CV), and X-ray photoelectron
spectroscopy (XPS). This is the first time that SAMs prepared from fullerene derivatives of thiols/thiol esters/
disulfides have been analyzed in detail, and the complications due to multilayer formations and head-to-
tail assemblies were revealed. Specifically, we designed and synthesized several fullerene derivatives based
on thiols, thiol acetates, and disulfides to address the characterization requirements, and these are described
and delineated. These studies specifically address the need to properly characterize and control fullerene-
thiol assemblies on gold before evaluating subsequent device performances.
Introduction
due to the well-defined organosulfur adsorption chemistry on
Au surfaces.4 Thiols, thiol esters, or disulfides are known to be
Advances in molecular science continue to push forward the
miniaturization of devices and the innovation of new molecule-
based functional systems.1 Of particular importance for our
group are the fields with potential applications of fullerenes and
fullerene derivatives. Since the discovery of fullerenes,2 due to
their unusual structure and optical and electrical properties, there
has been a tremendous amount of research aimed at developing
new fullerene-based materials with novel useful applications.3
The combination of self-assembled monolayer (SAM) forma-
tion,4 an important aspect in the field of molecular electronics,5
with the advantageous properties of fullerene derivatives could
lead to advances in this field.3,6 In fact, many groups have
worked to construct SAMs of fullerenes or fullerene derivatives
on a variety of substrates including conductive surfaces such
as gold7,8 and other nonconductive, semiconductive, or transpar-
ent conductive surfaces including indium-tin oxide (ITO).9
Among them, studies on gold surfaces are particularly extensive
absorbed onto a gold surface to form a SAM simply by exposing
(7) (a) Caldwell, W. B.; Chen, K.; Mirkin, C. A.; Babinec, S. J. Langmuir
1993, 9, 1945-1947. (b) Arias, F.; God´ınez, L. A.; Wilson, S. R.; Kaifer,
A. E.; Echegoyen, L. J. Am. Chem. Soc. 1996, 118, 6086-6087. (c)
Dom´ınguez, O.; Echegoyen, L.; Cunha, F.; Tao, N. Langmuir 1998, 14,
821-824. (d) Sahoo, R. R.; Patnaik A. J. Colloid Interface Sci. 2003, 268,
43-49. (e) Pan, G.-B.; Liu, J.-M.; Zhang, H.-M.; Wan, L.-J.; Zheng, Q.-
Y.; Bai, C.-L. Angew. Chem., Int. Ed. 2003, 42, 2747-2751. (f) Patnaik,
A.; Setoyama, H.; Ueno, N. J. Chem. Phys. 2004, 120, 6214-6221. (g)
Yoshimoto, S.; Saito, A.; Tsutsumi, E.; D’Souza, F.; Ito, O.; Itaya, K.
Langmuir 2004, 20, 11046-11052. (h) Bonifazi, D.; Spillmann, H.; Kiebele,
A.; de Wild, M.; Seiler, P.; Cheng, F.; Gu¨ntherodt, H.-J.; Jung, T.;
Diederich, F. Angew. Chem., Int. Ed. 2004, 42, 4759-2763. (i) Zhang, S.;
Palkar, A.; Fragoso, A.; Prados, P.; de Mendoza, J.; Echegoyen, L. Chem.
Mater. 2005, 17, 2063-2068. (j) Zhang, S.; Echegoyen, L. J. Org. Chem.
2005, 70, 9874-9881.
(8) (a) Shi, X.; Caldwell, W. B.; Chen, K.; Mirkin, C. A. J. Am. Chem. Soc.
1994, 116, 11598-11599. (b) Kim, S.; Park, S.-K.; Park, C.; Jeon, C., II.
J. Vac. Sci. Technol., B 1996, 14, 1318-1321. (c) Imahori, H.; Ozawa, S.;
Ushida, K.; Takahashi, M.; Azuma, T.; Ajavakom, A.; Akiyama, T.;
Hasegawa, M.; Taniguchi, S.; Okada, T.; Sakata, Y. Bull. Chem. Soc. Jpn.
1999, 72, 485-502. (d) Imahori, H.; Azuma, T.; Ozawa, S.; Yamada, H.;
Ushida, K.; Ajavakom, A.; Norieda, H.; Sakata, Y. Chem. Commun. 1999,
557-558. (e) Shon, Y.-S.; Kelly, K. F.; Halas, N. J.; Lee, T. R. Langmuir
1999, 15, 5329-5332. (f) Kelly, K. F.; Shon, Y.-S.; Lee, T. R.; Halas, N.
J. J. Phys. Chem. B 1999, 103, 8639-8642. (g) Fibbioli, M.; Bandyo-
padhyay, K.; Liu, S.-G.; Echegoyen, L.; Enger, O.; Diederich, F.; Bu¨hlmann,
P.; Pretsch, E. Chem. Commun. 2000, 339-340. (h) Lee, S.; Shon, Y.-S.;
Lee, T. R.; Perry, S. S. Thin Solid Films 2000, 358, 152-158. (i) Liu,
S.-G.; Martineau, C.; Raimundo, J.-M.; Roncali, J.; Echegoyen, L. Chem.
Commun. 2001, 913-914. (j) Zhang, S.; Dong, D.; Gan, L.; Liu, Z.; Huang,
C. New J. Chem. 2001, 25, 606-610. (k) Du, C.; Xu, B.; Li, Y.; Wang,
C.; Wang, S.; Shi, Z.; Fang, H.; Xiao, S.; Zhu, D. New J. Chem. 2001, 25,
1191-1194. (l) Hoang, V. T.; Rogers, L. M.; D’Souza, F. Electrochem.
Commun. 2002, 4, 50-53. (m) Imai, H.; Iwata, N.; Yamamoto, H.
Nanotechnology 2002, 13, 768-770. (n) Yamada, H.; Imahori, H.;
Fukuzumi, S. J. Mater. Chem. 2002, 12, 2034-2040. (o) Kim, K.-S.; Kang,
M.-S.; Ma, H.; Jen, A. K.-Y. Chem. Mater. 2004, 16, 5058-5062. (p) Kang,
S. H.; Ma, H.; Kang, M.-S.; Kim, K.-S.; Jen, A. K.-Y.; Zareie, M. H.;
Sarikaya, M. Angew. Chem., Int. Ed. 2004, 43, 1512-1516. (q) Gu, T.;
Whitesell, J. K.; Fox, M. A. J. Org. Chem. 2004, 69, 4075-4080. (r) Gru¨ter,
L.; Cheng, F.; Heikkila¨, T. T.; Gonza´lez, M. T.; Diederich, F.; Scho¨nen-
berger, C.; Calame, M. Nanotechnoloty 2005, 16, 2143-2248. (s) Zhang,
S.; Echegoyen, L. Tetrahedron 2006, 62, 1947-1954.
(1) (a) Timp, G. L. Nanotechnology; AIP Press/Springer: New York, 1999.
(b) Nanoelectronics and Information Technology: AdVanced Electronic
Materials and NoVel DeVices Information Technology; Waser, R., Ed.;
Wiley-VCH: Weinheim, Germany, 2003. (c) Joachim, C.; Gimzewski, J.
K.; Aviram, A. Nature 2000, 408, 541-548. (d) Forrest, S. R. Nature 2004,
428, 911-918.
(2) Kroto, H. W.; Heath, J. R.; O’Brien, S. C.; Curl, R. F.; Smalley, R. E.
Nature 1985, 318, 162-163.
(3) (a) Prato, M. Top. Curr. Chem. 1999, 199, 173-187. (b) Konishi, T.; Ikeda,
A.; Shinkai, S. Tetrahedron 2005, 61, 4881-4899.
(4) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M.
Chem. ReV. 2005, 105, 1103-1169 and references therein.
(5) Tour, J. M. Molecular Electronics: Commercial Insights, Chemistry,
DeVices, Architecture and Programming; World Scientific: River Edge,
NJ, 2003.
(6) (a) Mirkin, C. A.; Caldwell, W. B. Tetrahedron 1996, 52, 5113-5130. (b)
Zeng, C.; Wang, H.; Wang, B.; Yang, J.; Hou, J. G. Appl. Phys. Lett. 2000,
77, 3595-3597.
9
10.1021/ja063451d CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 13479-13489
13479

A R T I C L E S
Shirai et al.
Figure 1. Fullerene-oligo(phenylene ethynylene) (OPE) hybrids 1 and 2, fullerene tripod 3, and fullerene disulfide 4 were used in this study.
Figure 2. Two approaches to prepare fullerene SAMs, the direct method (top) and the indirect method (bottom), and possible anomalies introduced during
SAM formation.
the gold surface to a solution of the molecule. In most instances,
this simple view of SAM formation on gold has been assumed
to be true for fullerene derivatives as well. However, we have
found that this is not the case for the fullerene derivatives in
our work (Figure 1) and is likely not the case for other fullerene
derivatives that are similar to those shown in Figure 1. We report
here anomalies associated with SAMs of these fullerene
derivatives on gold using surface characterization techniques
including ellipsometry, X-ray photoelectron spectroscopy (XPS),
and cyclic voltammetry (CV). We also show that active control
of the packing density of fullerene SAMs using the rigid
molecular tripod10 3 is a promising (easy and reproducible) way
to achieve highly electrochemically active fullerene SAMs.
There are two approaches to prepare SAMs of fullerene
derivatives on gold (Figure 2). In the first approach, which we
shall call the “direct” method, fullerene derivatives containing
“alligator clips”11 are synthesized so that they can bind to the
(10) For examples of tripod-shaped thiol derivatives, see: (a) Fox, M. A.;
Whitesell, J. K.; McKerrow, A. J. Langmuir 1998, 14, 816-821. (b) Yao,
Y.; Tour, J. M. J. Org. Chem. 1999, 64, 1968-1971. (c) Hu, J.; Mattern,
D. L. J. Org. Chem. 2000, 65, 2277-2281. (d) Zhu, L.; Tang, H.; Harima,
Y.; Yamashita, K.; Hirayama, D.; Aso, Y.; Otsubo, T. Chem. Commun.
2001, 1830-1831. (e) Hirayama, D.; Takimiya, K.; Aso, Y.; Otsubo, T.;
Hasobe, T.; Yamada, H.; Imahori, H.; Fukuzumi, S.; Sakata, Y. J. Am.
Chem. Soc. 2002, 124, 532-533. (f) Zhu, L.; Tang, H.; Harima, Y.;
Yamashita, K.; Aso, Y.; Otsubo, T. J. Mater. Chem. 2002, 12, 2250-
2254. (g) Otsubo, T.; Aso, Y.; Takimiya, K. J. Mater. Chem. 2002, 12,
2565-2575. (h) Li, Q.; Rukavishnikov, A. V.; Petukhov, P. A.; Zaikova,
T. O.; Keana, J. F. W. Org. Lett. 2002, 4, 3631-3634. (i) Jian, H.; Tour,
J. M. J. Org. Chem. 2003, 68, 5091-5103. (j) Wei, L.; Padmaja, K.;
Youngblood, W. J.; Lysenko, A. B.; Lindsey, J. S.; Bocian, D. F. J. Org.
Chem. 2004, 69, 1461-1469. (k) Park, J.-S.; Nguyen Vo, A.; Barriet, D.;
Shon, Y.-S.; Lee. T. R. Langumuir 2005, 21, 2902-2911. (l) Kitagawa,
T.; Idomoto, Y.; Matsubara, H.; Hobara, D.; Kakiuchi, T.; Okazaki, T.;
Komatsu, K. J. Org. Chem. 2006, 71, 1362-1369. (m) Yam, C. M.; Cho,
J.; Cai, C. Langmuir 2003, 19, 6862-6868. (n) Moore, A. M.; Dameron,
A. A.; Mantooth, B. A.; Smith, R. K.; Fuchs, D. J.; Ciszek, J. W.; Maya,
F.; Yao, Y.; Tour, J. M.; Weiss, P. S. J. Am. Chem. Soc. 2006, 18, 1959-
1967.
(9) (a) Chen, K.; Caldwell, W. B.; Mirkin, C. A. J. Am. Chem. Soc. 1993,
115, 1193-1194. (b) Chupa, J. A.; Xu, S.; Fischetti, R. F.; Strongin, R.
M.; McCauley, J. P., Jr.; Smith, A. B., III; Blasie, J. K. J. Am. Chem. Soc.
1993, 115, 4383-4384. (c) Tsukruk, V. V.; Lander, L. M.; Brittain, W. J.
Langmuir 1994, 10, 996-999. (d) Feng, W.; Miller, B. Langmuir 1999,
15, 3152-3156. (e) Kim, K.; Song, H.; Park, J. T.; Kwak, J. Chem. Lett.
2000, 958-959. (f) Yamada, H.; Imahori, H.; Nishimura, Y.; Yamazaki,
I.; Ahn, T. K.; Kim, S. K.; Kim, D.; Fukuzumi, S. J. Am. Chem. Soc. 2003,
125, 9129-9139. (g) Imahori, H.; Kimura, M.; Hosomizu, K.; Sato, T.;
Ahn, T. K.; Kim, S. K.; Kim, D.; Nishimura, Y.; Yamazaki, I.; Araki, Y.;
Ito, O.; Fukuzumi, S. Chem.sEur. J. 2004, 10, 5111-5122. (h) Gulino,
A.; Bazzano, S.; Condorelli, G. G.; Giuffrida, S.; Mineo, P.; Satriano, C.;
Scamporrino, E.; Ventimiglia, G.; Vitalini, D.; Fragala`, I. Chem. Mater.
2005, 17, 1079-1084.
9
13480 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Fullerene Derivative Monolayers on Gold Surfaces
A R T I C L E S
gold surfaces.8 Alternatively, in an “indirect” method, a gold
surface can be modified with an alligator clip-containing
intermediate to which the fullerene can be attached in a second
step.7 Despite extensive literature on the preparation of SAMs
of fullerene derivatives on gold surfaces, little or no attention
has been paid to the strong interactions between fullerenes them-
selves, or between fullerenes and the gold surfaces, during the
SAM formation except for the work by Kelly, et al.8f Because
of the relatively strong adhesion of fullerenes onto gold surfaces
(30-60 kcal/mol, vide infra),12 which are similar to the Au-S
bond strengths, and the tendency of fullerenes to form clusters,13
one could imagine that fullerenes might form anomalous SAMs
such as head-to-tail assemblies and multilayer formations (Figure
2). Furthermore, some thiol groups can be hampered from
binding to the gold surface due to strong lateral fullerene-
fullerene interactions; this is illustrated in the cartoon depicting
several possible modes of multilayer formation via the direct
SAM formation method (Figure 2). However, these aspects were
not well addressed in previous studies on the self-assembly of
fullerene-derivatized free or protected thiols. The indirect
approach in Figure 2 could eliminate the formation of head-
to-tail assemblies. However, multilayer formation can still occur
if inappropriate solvents are chosen in the fullerene assembly.
Moreover, the displacement of thiols by fullerenes must be con-
sidered. In some applications, a fullerene multilayer is actually
desired and intentionally induced.14 Nevertheless, the direct
approach is more popular than the indirect one presumably due
to the difficulties associated with carrying out chemical reactions
on surfaces. If one could assume that the well-known simple
chemistry of thiol-based SAM formation on gold is appropriate
for fullerene derivatives, then the direct approach would be much
simpler in terms of characterization. Using molecules shown
in Figure 1 and surface characterization techniques including
ellipsometry, XPS, and CV on electrode surfaces, here we show
that this simple view is not appropriate for fullerene derivatives
and the effects from the anomalies on the fullerene SAMs such
as head-to-tail assembly and multilayer formation have to be
considered in the design of fullerene-based functional devices.
Scheme 1. Synthesis of the Fullerene-OPE Hybrids 1-SH and
1-SAc
derivatized OPEs with thiol and protected thiol alligator clips
can be easily prepared using our in situ ethynylation method
(Schemes 1-4).17 This class of fullerene-derivatized OPEs is
unique in that one can expect periconjugation18 effects due to
the close proximity between the fullerene and OPE π systems.
After synthesizing the fullerene-OPE molecules, they were
then assembled onto gold surfaces using common assembly
techniques (see the Experimental Section). However, the unusual
results of the growth kinetics (Figure 3) of the fullerene-OPE
hybrid 1 caught our attention.
Under all conditions except mixtures of 1,2-dichlorobenzene
(ODCB), multilayer formation was evident even though the
concentrations were kept to a minimum (less than 30 µM) to
avoid aggregations of the fullerene derivatives during assembly
periods. The ellipsometric thicknesses for 1 assembled from
ODCB solutions were close to those predicted (1.8 nm). It is
likely that 1 formed aggregates during assembly in the other
solvents since the solubility of fullerene or fullerene derivatives
in ODCB is generally quite high compared to that in the organic
solvents toluene, CH2Cl2, and THF.19 Our results demonstrate
the importance of solvent choice for the assembly of fullerene
Results and Discussion
Approaches to molecular electronics using OPE devices15
have gained considerable attention since the successful dem-
onstration of switching effects using this class of compounds.16
To further advance the development of OPE devices for
molecular electronics applications, we investigated fullerene-
OPE hybrid devices such as 1 and 2 (Figure 1). These fullerene-
(11) Tour, J. M.; Jones, L., II; Pearson, D. L.; Lamba, J. S.; Burgin, T.;
Whitesides, G. W.; Allara, D. L.; Parikh, A. N.; Atre, S. J. Am. Chem.
Soc. 1995, 117, 9529-9534.
(12) (a) Chase, S. J.; Bacsa, W. S.; Mitch, M. G.; Pilione, L. J.; Lannin, J. S.
Phys. ReV. B 1992, 46, 7873-7877. (b) Altman, E. I.; Colton, R. J. Surf.
Sci. 1993, 295, 13-33. (c) Tzeng, C.-T.; Lo, W.-S.; Yuh, J.-Y.; Chu, R.-
Y.; Tsuei, K.-D. Phys. ReV. B 2000, 61, 2263-2272. (d) Hunt, M. R. C.;
Modesti, S.; Rudolf, P.; Palmer, R. E. Phys. ReV. B 1995, 51, 10039-
10047.
(13) Patnaik, A.; Okudaira, K. K.; Kera, S.; Setoyama, H.; Mase, K.; Ueno, N.
J. Chem. Phys. 2005, 122, 154703.
(14) Kamat, V. P.; Barazzouk, S.; Thomas, K. G.; Hotchandani, S. J. Phys.
Chem. B 2000, 104, 4014-4017.
(15) (a) Tour, J. M.; Rawlett, A. M.; Kozaki, M.; Yao, Y.; Jagessar, R. C.;
Dirk, S. M.; Price, D. W.; Reed, M. A.; Zhou, C.-W.; Chen, J.; Wang, W.;
Campbell, I. Chem.sEur. J. 2001, 7, 5118-5134. (b) James, D. K.; Tour,
J. M. Aldrichimica Acta 2006, 39, 47-56. (c) James, D. K.; Tour, J. M.
Top. Curr. Chem. 2005, 257, 33-62.
(16) Chen, J.; Reed, M. A.; Rawlett, A. M.; Tour, J. M. Science 1999, 286,
1550-1552.
(17) Shirai, Y.; Zhao, Y.; Cheng, L.; Tour, J. M. Org. Lett. 2004, 6, 2129-
2132.
(18) (a) Zhao, Y.; Shirai, Y.; Slepkov, A. D.; Cheng, L.; Alemany, L. B.; Sasaki,
T.; Hegmann, F. A.; Tour, J. M. Chem.sEur. J. 2005, 11, 3643-3658. (b)
Hamasaki, R.; Ito, M.; Lamrani, M.; Mitsuishi, M.; Miyashita, T.;
Yamamoto, Y. J. Mater. Chem. 2003, 13, 21-26 and references therein.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13481

A R T I C L E S
Shirai et al.
Scheme 2. Synthesis of the Fullerene-OPE Hybrids 2-SH and
techniques such as ellipsometry or quartz crystal microbalance
(QCM) measurements.7a
2-SAca
The film thickness of all compounds determined by ellip-
sometry is summarized in Table 1. All experiments were
performed in the mixture of ODCB/EtOH (4:1); the EtOH was
added to suppress oxidation of sulfur atoms (see the XPS results
for the effect of EtOH in ODCB solution). In most cases,
ellipsometric thicknesses reasonably correspond to the expected
film thicknesses considering that ellipsometry is only accurate
to within (2 Å.21 When free thiol compounds 1-SH or 2-SH
were assembled in air, the resulting films were consistently
thicker than those assembled in a nitrogen atmosphere. It is
likely that the aryl thiols formed disulfide bonds under ambient
conditions. Our previous studies on multifullerene derivatives,
including difullerenes, determined that they are quite insoluble
even in organic solvents such as CS2 and ODCB.17,18a,22 Thus,
the disulfides generated in situ during the assembly will deposit
on the surfaces. Therefore, we do not recommend the use of
unprotected arylthiol derivatives for fullerene derivative SAM
formation because mitigating disulfide formation is difficult.
Figure 4 shows the assembly of fullerene derivatives 1-4 in
ODCB/EtOH (4:1) solvent mixtures. Aryl thioacetates were
deprotected in situ under acidic conditions23 to avoid disulfide
formation. In all cases, excessive multilayer formations are
reasonably suppressed.
CV Characterization of SAMs. Because fullerene is elec-
trochemically active, the resultant fullerene-OPE SAMs can
be further characterized using CV. Figure 5 shows voltammo-
grams from SAMs of 1-SAc (acid deprotected) and 4 on Au
electrode surfaces. The surface electrochemistry of the fullerene-
derivatized Au electrode exhibited broad featureless spectra for
these compounds, regardless of the different deposition time,
concentration, solvent, and deprotection conditions. In the first
few scans for the SAM of 1-SAc, there were sometimes several
features, as shown in Figure 5A, which are likely the redox
signals from fullerenes. However, these features immediately
disappeared in succeeding scans, and eventually the voltam-
mograms became broad and featureless. The initial features
observed upon CV were probably caused by fullerene deriva-
tives loosely associated with the electrode surfaces by means
of weak interactions such as physisorption. In many cases,
shown in Figure 5B as representative voltammograms, we
observed featureless broad voltammograms even in the initial
scans. We observed similar behaviors for all SAMs prepared
from 1, 2, and 4.
a Reagents: Pd/Cu ) PdCl2(PPh3)2, CuI.
derivatives and also the importance of ellipsometry measure-
ments to detect multilayer formation. It is surprising that, to
our knowledge, this is the first example of the use of ellipsom-
etry to address multilayer formation in the assembly of fullerene-
derivatized thiol or thiol esters.20 For compounds with good
solubility in common organic solvents, such as 3 and 4, one
may be able to assume that there is no significant multilayer
formation during assembly. Otherwise, it is necessary to check
the SAM thicknesses or deposited amounts with appropriate
Mirkin et al. and others previously reported the coverage-
dependent electrochemical behavior of fullerene SAMs.6a,7a,8q
It was shown that a tightly packed SAM could inhibit charge-
compensating ion transport within the SAM film, leading to an
electrochemical behavior that is drastically different from the
one for the same redox-active molecule in solution.24 Because
our fullerene-OPEs 1, 2, and 4 can also form dense monolayers
(21) Azzam, R. M. A.; Bashara, N. M. Ellipsometty and Polarized Light; North-
Holland: New York, 1987.
(22) Shirai, Y.; Osgood, A. J.; Zhao, Y.; Yao, Y.; Saudan, L.; Yang, H.; Yu-
Hung, C.; Alemany, L. B.; Sasaki, T.; Morin, J.-F.; Guerrero, J. M.; Kelly,
K. F.; Tour, J. M. J. Am. Chem. Soc. 2006, 128, 4854-4864.
(23) Cai, L.; Yao, Y.; Yang, J.; Price, D. W.; Tour, J. M. Chem. Mater. 2002,
14, 2905-2909.
(24) (a) Rowe, G. K.; Creager, S. E. Langmuir 1991, 7, 2307-2312. (b) Smith,
C. P.; White, H. S. Anal. Chem. 1992, 64, 2398-2405. (c) Caldwell, W.
B.; Campbell, D. J.; Chen, K.; Herr, B. R.; Mirkin, C. A.; Malik, A.; Durbin,
M. K.; Dutta P.; Huang. K. G. J. Am. Chem. Soc. 1995, 117, 6071-6082.
(19) The solubility of fullerene in CH2Cl2, toluene, and ODCB is 0.3, 2.2, and
27 mg/mL, respectively: (a) Taylor, R. The chemistry of Fullerenes; World
Scientific: River Edge, NJ, 1995. (b) Scrivens, W. A.; Tour, J. M. J. Chem.
Soc., Chem. Commun. 1993, 1207-1209.
(20) Previously, SAMs from fullerene-derivatized disulfides were checked using
ellipsometry. However, the examples are similar to 4 in this work,
structurally quite soluble in organic solutions, and unlikely to form
multilayers; see: refs 8e and 8h.
9
13482 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Fullerene Derivative Monolayers on Gold Surfaces
A R T I C L E S
Scheme 3. Synthesis of the Fullerene Tripod 3a
a Reagents: Pd/Cu ) PdCl2(PPh3)2, CuI.
due to interactions between the neighboring fullerenes, an effect
similar to that reported by Mirkin et al. is probably causing the
broad featureless voltammograms. Our attempts to obtain better-
resolved electrochemical signals from less-than-monolayer-
covered self-assemblies failed even though this strategy was
previously reported to permit the ion transport effects.8a With
our molecules, it was difficult to reproduce such partial coverage
in a reliable manner. Controlling the spacing between the
fullerenes by codeposition of inert compounds such as simple
alkanethiols and arylthiols was also unsuccessful. This result
actually parallels previous studies. When adsorbates can strongly
interact with each other, it is difficult to avoid phase segregation
within the mixed SAMs.10m To address this issue with a unique
molecular design, we synthesized the fullerene tripod hybrid 3
that is designed to isolate the fullerene core from neighboring
molecules using bulky tripod legs (Figure 6), allowing easy
diffusion of charge-compensating ions within a SAM film.10
The two well-resolved reversible redox waves, with reductions
at -0.63 and about -0.99 V versus Ag/AgCl in Figure 7,
indicate that this tripod strategy is in fact effective. As with
other surface-confined fullerene species,7-9 it was difficult to
obtain a stable and reversible third redox wave that is observable
in the solution state CV (see the Supporting Information for
the solution state CV of fullerene and related compounds). For
both of the redox signals, peak currents increase linearly with
the scan rate (0.1-1.0 V/s) (Figure 7B), and the peak-to-peak
separations are small (∆E ) 29.5 ( 3.5 mV) and independent
of scan rates, indicating a typical behavior of surface-confined
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13483

A R T I C L E S
Shirai et al.
Scheme 4. Synthesis of the Fullerene Disulfide 4a
Table 1. Assembly of the Fullerene Derivatives in a Solvent
Mixture of ODCBa
thickness (nm)
assembly
conditionb
time
(h)
compd
exptlc
calcdd
1-SH
1-SH
1-SAc
2-SH
2-SH
2-SAc
3
N2
-
acid
N2
-
acid
-
-
24
24
24
24
24
24
24
24
1.7
1.9
1.7
2.4
2.7
2.3
2.3
3.4
1.8
1.8
1.8
2.6
2.6
2.6
2.6
3.4
4
a Solvent mixture of ODCB/EtOH (4:1). b Acid: in ambient conditions
and a catalytic amount of H2SO4 added in a capped container. -: in ambient
conditions without acid in a capped container. N2: assembly done in a
nitrogen glovebox without acid. c The value measured by ellipsometry with
an error of (0.2 nm. d Distance between the gold atom and the top carbon
atom in the fullerene cage. A linear Au-S-C bond angle with a 2.36 Å
Au-S bond was assumed.
Figure 4. Growth kinetics of a series of fullerene derivatives assembled
on Au in ODCB/EtOH (4:1) solvent mixtures. Film thickness was
determined by ellipsometry. Compounds 1-SAc and 2-SAc were deprotected
in situ using the acid deprotection method, while 3 and 4 were assembled
without acid deprotection. All experiments were done under ambient
conditions in a capped vessel, and the results from two separate experiments
are shown for each compound.
a Reagents: Pd/Cu ) PdCl2(PPh3)2, CuI.
molecule) using the Randles-Sevcik equation for a Nernstian
reaction at the adsorbate monolayer (eq 1) or the integration of
the area under the reduction waves,25
ip ) (9.39 × 105)n2VAΓ
(1)
where ip is the peak current in A, n is number of electrons, v
is the scan rate in V/s, A is the electrode surface area in cm2,
and Γ is the coverage in mol/cm2. For 3, where n ) 1, from a
linear relationship between the scan rate V and the peak current
ip and the predetermined electrode surface area A, Γ was
estimated to be 3.8 ( 0.1 × 10-11 (mol/cm2), and therefore the
footprint of 3 is 4.4 nm2/molecule. The integration of the area
under the reduction waves also gave the coverage Γ in the same
range (2.8∼5.0 × 10-11 mol/cm2). Figure 6 schematically
compares this value to the size of molecule on a surface. For
simplicity, in Figure 6 it has been assumed that the molecules
in the SAM bind to the surface using all three of their thiolate
ligands. However, the actual SAM is probably composed of a
Figure 3. Growth kinetics of C60-OPE hybrid 1 assembled on Au under
different solvent conditions. The film thickness was determined by
ellipsometry. Monolayer assembly would be reached at 1.8 nm. Data points
with thicknesses >4.0 nm were not included in this chart.
electroactive species. With these well-behaved electrochemical
responses and assuming all molecules on an Au electrode are
electrochemically active, we estimated the footprint of 3 (nm2/
(25) (a) Finklea, H. O. In Electroanalytical Chemistry; Bard, A. J., Rubinstein,
I., Eds.; Marcel Dekker: New York, 1996; Vol. 19, pp 109-335. (b) Bard,
A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and
Applications; John Wiley & Sons: New York, 2000.
9
13484 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Fullerene Derivative Monolayers on Gold Surfaces
A R T I C L E S
Figure 5. Cyclic voltammograms of SAMs of (A) 1-SAc on a Au disk
working electrode and (B) 4 on a Au bead working electrode in 0.1 M
TBAPF6 in ODCB/MeCN (4:1) at room temperature. 1-SAc and 4 were
assembled in ODCB/EtOH (4:1) with and without acid deprotection,
respectively. The counter electrode was a Pt wire, and the reference electrode
was Ag/AgCl. Scan rates were 100 mV/s.
Figure 7. (A) Cyclic voltammograms of SAM of 3 on a Au bead working
electrode in 0.1 M TBAPF6 in ODCB/MeCN (4:1) at room temperature.
The counter electrode was a Pt wire, and the reference electrode was Ag/
AgCl. The scan rate was raised from 0.1 to 1.0 V/s with 0.1 V/s increments.
(B) Peak currents at the first redox waves of part A as a function of scan
rate.
Figure 6. Schematics for the density of the tripod molecules 3 on the
surface. It is assumed in this picture that all three thiolates bind to the
surface. The fullerene moiety was not shown in the space-filling models
for clarity. The small and large yellow circles are equivalent to an area of
2.4 and 4.4 nm2, respectively, showing the area/molecule at (A) estimated
minimum size, ∼2.4 nm2/molecule, and (B) estimated molecular footprint
from CV results, ∼4.4 nm2/molecule.
that the strong adhesion between fullerenes and gold surfaces
is either a charge-transfer interaction or covalent in nature with
some ionic features. The adhesion energy was estimated to range
from ∼1.3 eV26 to 2.2-2.6 eV27 (from ∼30 to 50-60 kcal/
mol), which is comparable to ∼50 kcal/mol for the Au-S bond.4
In fact, the adhesion of fullerenes to gold surfaces allowed us
to develop surface-rolling molecules, in which adhesion of the
fullerenes to the gold surface was essential for the demonstration
of rolling versus the more common sliding.22,28 Furthermore,
the fullerene/gold adhesion was not limited to the UHV-
deposited fullerenes; studies in the solution phase also reported
the adhesion of fullerenes onto gold surfaces.29 The packing of
fullerenes deposited from solutions corresponds to either a (2x3
× 2x3)R30 arrangement with respect to the Au(111) lattice or
an in-phase overlayer, which are essentially the same results as
obtained on UHV-deposited fullerene overlayers on the same
mixture of several different conformations because 3 can bind
to the surface using one, two, or three of its thiolate ligands;
this, in combination with various conformations arising from
fullerene-Au bonding, would lead to a loosely packed SAM.10k,m,n
In comparison to the simulated crystalline packing of tripod 3
in Figure 6A, (r ≈ 8.7 Å), the isolation (spacing) of fullerenes
in the tripod SAM is increased according to our CV results (r
≈ 12 Å, Figure 6B). The CV result confirms the noncrystalline,
less-defined liquidlike loose packing of the tripod 3. Our strategy
of using a bulky tripod anchor group as a foundation of an
electrochemically active site may be useful in other applica-
tions.10
XPS Analysis of Fullerene SAMs. The SAMs prepared from
the fullerene derivatives of thiols/thiol esters/disulfides were
further characterized using XPS. As mentioned earlier, fullerenes
tend to form strong bonds on gold surfaces. Experimental12 and
theoretical26 work, where fullerenes were deposited on gold
surfaces in ultrahigh vacuum (UHV), supported the conclusion
(26) Wang, L.-L.; Cheng, H.-P. Phys. ReV. B 2004, 69, 165417-12.
(27) Guo, S.; Fogarty, D. P.; Nagel, P. M.; Kandel, S. A. J. Phys. Chem. B
2004, 108, 14074-14081 and references therein.
(28) Shirai, Y.; Osgood, A. J.; Zhao, Y.; Kelly, K. F.; Tour, J. M. Nano Lett.
2005, 5, 2330-2334.
(29) (a) Yoshimoto, S.; Narita, R.; Tsutsumi, E.; Matsumoto, M.; Itaya, K.
Langmuir 2002, 18, 8518-8522. (b) Marchenko, A.; Cousty, J. Surf. Sci.
2002, 513, 233-237. (c) Uemura, S.; Samor´ı, P.; Kunitake, M.; Hirayama,
C.; Rade, J. P. J. Mater. Chem. 2002, 12, 3366-3367.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13485

A R T I C L E S
Shirai et al.
substrates.29b These studies point to the possibility of the head-
to-tail assembly of fullerene-derivatized organosulfurs on gold
surfaces as shown in Figure 2. Surprisingly, however, the possi-
bility of this head-to-tail assembly was not examined in detail
in previous studies on the SAMs of thiols/thiol esters/disulfides
of fullerene derivatives. In addition, to our knowledge, the XPS
analysis on this class of compounds has not been reported.30
We addressed the possibility of the head-to-tail assembly of
the fullerene derivatives along with other noticeable features
using XPS because the difference in the chemical environments
of the sulfur atoms that are bound or unbound to the gold surface
is significant enough to be detected in a reliable manner.31 Figure
8 shows representative detailed sulfur atom XPS spectra of the
SAMs formed from 1-4. There was a notable decrease in the
ratio of unbound-to-bound sulfur atom signals, from 1.1 to 0.3,
by doing the assembly in a nitrogen glovebox (Figure 8 and
Table 2). This difference is likely the consequence of the
decrease in the disulfide formation under nitrogen conditions,
which is consistent with the ellipsometry data under the same
conditions. A doublet centered at ∼169 eV was observed when
the fullerene derivatives of unprotected thiols 1-SH and 2-SH,
or disulfide 4, were assembled using pure ODCB or a mixture
of ODCB and EtOH solutions, whereas no such peak was
detected when the thiol esters 1-SAc, 2-SAc, and 3 were
assembled in the ODCB/EtOH solution. This high-energy sulfur
signal is characteristic of highly oxidized sulfur atoms (+4 state,
sulfonic acid).32 Since our experimental conditions were con-
sistent throughout all measurements, this feature was likely
caused by either a difference in molecular structures or a
difference in solvent conditions. We first examined the solvent
conditions because the use of ODCB in SAM formation is rare.33
A simple nonfullerene model compound, 4-bromobenzenethiol,
was assembled in pure ODCB, EtOH, or mixtures of both,34
and the resulting SAMs were examined using XPS (see the
Supporting Information for the XPS spectra). Significant sulfur
oxidation took place during assembly in pure ODCB solution,
whereas no such signal was detected when pure EtOH was used.
Because the use of ODCB is essential for the prevention of
fullerene aggregation leading to the multilayer formation
according to the ellipsometry analysis, we used a mixture of
ODCB and EtOH, hoping that good solubility in ODCB paired
with the reducing environment in EtOH would be mutually
effective. We found this to be the case. Next, we looked at
structural differences as a cause of sulfur oxidation and noticed
that thiols 1-SH and 2-SH and disulfide 4 were readily oxidized,
whereas sulfides (thiol esters 1-SAc, 2-SAc, and 3) were less
susceptible to the oxidation. This observation is actually parallel
to the general trend in sulfur chemistry, in which thiols or
disulfides are easily oxidized due to the weak S-H or S-S
(30) XPS analysis on fullerene-derivatized methyl sulfides was previously
reported, see ref 8c. Methyl sulfides have a weak interaction with gold
and are less frequently employed as an alligator clip.
(31) (a) Gastner, D. G.; Hinds, K.; Grainger, D. W. Langmuir 1996, 12, 5083-
5086. (b) Yasseri, A. A.; Syomin, D.; Malinovskii, V. L.; Loewe, R. S.;
Lindsey, J. S.; Zaera, F.; Bocian, D. F. J. Am. Chem. Soc. 2004, 126,
11944-11953.
(32) (a) Rieley, H.; Price, N. J.; Smith, T. L.; Yang, S. J. Chem. Soc., Faraday
Trans. 1996, 92, 3629-3634. (b) Rieley, H.; Kendall, G. K.; Zemicael, F.
W.; Smith, T. L.; Yang, S. Langmuir 1998, 14, 5147-5153.
(33) For an example of the use of ODCB as adsorbate solution, see ref 8i.
(34) We decided to use EtOH because it is the most widely used solvent for
SAM formation and it is used to reduce Au surfaces after oxidative cleaning
such as UV/ozone, thus producing a reducing environment. The mechanism
for EtOH suppression of sulfur oxidation during assembly in ODCB is
unknown at this point.
Figure 8. XPS S2p spectra for 1-SH assembled using ODCB in ambient
(capped vessel) or in glovebox, 1-SAc and 2-SAc assembled in ODCB/
EtOH (4:1) mixture using the in situ acid deprotection method in ambient
conditions, and 3 and 4 assembled in ODCB/EtOH (4:1) mixture in ambient
conditions. The broad peaks from 161 to 165 eV were fit using two S2p
doublets with 2:1 area ratios and splittings of 1.2 eV and one S2p doublet
with 2:1 area ratios and splittings of 1.2 eV for the peak at ∼169 eV. Each
doublet is assigned to surface-bound, surface-unbound, and oxidized sulfur
atoms.
9
13486 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Fullerene Derivative Monolayers on Gold Surfaces
A R T I C L E S
Table 2. Sulfur Atom Binding Energies and Atomic Ratios
binding energy (eV)
estimated atomic ratio
bound
S 2p3
unbound
S 2p3
oxidized
S 2p3
unbound/
bound
oxidized/
bound
SAM
/
2
/
2
/2
1-SHa
1-SHb
1-SAcc
2-SAcc
3d
161.9
161.6
161.6
161.6
161.6
161.9
163.5
163.5
163.5
163.4
163.5
163.4
168.6
168.4
1.1
0.3
0.2
0.3
0.4
1.0
1.2
0.8
0
0
0
Figure 9. Schematics of ideal dense normal and head-to-tail SAMs of
compound 4 on gold surfaces.
4d
169.3
0.8
a SAM was prepared in ODCB in ambient (capped vessel) conditions.
b SAM was prepared in ODCB in a glovebox. c In situ acid deprotection in
a mixture of ODCB/EtOH (4:1) in ambient conditions. d Assembled in a
mixture of ODCB/EtOH (4:1) in ambient conditions.
methylenes g10) form densely packed monolayers on gold
surfaces.8q,37 For these reasons, the disulfide 4 was supposed
to form an ideal fullerene SAM on gold surfaces. However, the
S 2p signals from the SAM of disulfide 4 (Figure 8) clearly
indicate an ∼1:1 mixture of (possibly alternating) head-to-tail
assembled molecules. Ironically, even though the disulfide 4
was designed to minimize the unbound species, it gave the
highest ratio in the unbound-to-bound S 2p signal.38 This finding
is important because fullerene derivatives of thiols/thiol esters/
disulfides are believed to form unidirectionally assembled
monolayers on gold surfaces. The fullerene SAMs were thus
thought to be a useful and simple strategy to mimic the core of
natural photosynthesis, in which the unidirectional multistep
electron transfer occurs along well-arranged assemblies, leading
to the generation of a charge-separated state across the
membrane with quantum efficiency close to unity.39 Because
fullerene can be an excellent electron acceptor in a donor-
acceptor electron-transfer system,40 two-dimensional and uni-
directional arrangement of such fullerene systems in the form
of SAM on gold is attractive for the construction of high-
efficiency artificial photosynthesis membranes.41 However, our
results indicate that a method similar to the indirect approach
in Figure 2 may be preferred in order to attain a complete
unidirectional assembly. It is difficult to rationalize why the
disulfides 4 form SAMs with ∼1:1 unbound-to-bound ratio. One
possibility is that the densely packed SAM of 4 is more stable
with the molecules alternatively arranged on the Au surface to
maximize density or molecule-to-molecule interactions (Figure
9). In this case, the molecules, once trapped head-to-tail into a
SAM, may not be able to flip over. The other fullerene
derivatives 1 and 2 are unlikely to form such organized SAMs;
thus, any head-to-tail packed molecules can flip over to
maximize fullerene-fullerene interactions, thereby reducing the
degree of the head-to-tail assembly. In all cases, the cross section
mismatch between the fullerene head and the tether to the
anchoring sulfur is evident (except for tripod 3); this cross
section mismatch could be a cause for the head-to-tail assembly.
bond, respectively, and the oxidation will not stop until the
sulfonic acid (+4 state) is produced.35 A similar phenomenon
was observed for the assembly of the acetyl-protected R,ω-
dithiols,36 in which no oxidation was detected. It is also possible
that fullerene enhances sulfur oxidation because fullerene itself
can be easily reduced (oxidizing surroundings). As a result, the
mixture of ODCB and EtOH (4:1) can suppress the oxidation
for the thiol esters 1-SAc, 2-SAc, and 3. However, it was not
effective enough for the more oxidation-susceptible thiols 1-SH
and 2-SH and disulfide 4.
For all SAMs of the fullerene-derivatized organosulfurs we
tested, careful examination of the broad peaks from 161 to 165
eV suggests at least two doublets of S 2p signals (each with
area ratios of 2:1 and splittings of 1.2 eV). It is likely that these
doublets correspond to surface-bound (∼161.7 eV for S 2p3/2
)
and surface-unbound (∼163.5 eV for S 2p3/2) sulfur atoms.31
The unbound thiol/thiol ester/disulfide S 2p signals may result
from a number of possible species in the SAMs. The fullerene
derivatives that are structurally poor in solubility, such as 1 and
2, could be deposited on top of the fullerene monolayer, leading
to multilayer formation. Alternatively, they could be trapped
within the monolayer film without forming a sulfur-gold bond
due to the strong lateral interactions between fullerenes. Either
of these structures could contribute to the unbound S 2p signals.
Even though the solubility of fullerene derivatives in the ODCB
mixture is good, the possibility of deposition through precipita-
tion cannot be ruled out because homocoupled difullerenes
would be insoluble, leading to the disulfide S 2p signals. Another
possible cause for the unbound S 2p signal is the head-to-tail
assembly (Figure 2), in which the molecules assembled on Au
via fullerene-Au surface adhesion, not the usual thiolate-Au
bonding. This is especially true for highly soluble fullerene
derivatives and thiol esters that will not form homocoupled
difullerene species in situ. The well-solvated species are less
likely to be trapped in the monolayer film or to be deposited
on the surface of the SAM via precipitation; therefore, the
unbound S 2p signals in these cases likely originate from the
head-to-tail assembly. The disulfide 4 was designed to rule out
the possibility of multilayer formation. Because it has a long
alkyl chain, the molecule is quite soluble in the ODCB mixture
and there is no in situ disulfide formation leading to precipitation
and deposition onto surfaces. Furthermore, it is well established
that alkanethiols with a polymethylene chain (number of the
(37) (a) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am.
Chem. Soc. 1987, 109, 3559-3568. (b) Imahori, H.; Azuma, T.; Ajavakom,
A.; Norieda, H.; Yamada, H.; Sakata, Y. J. Phys. Chem. B 1999, 103, 7233-
7237.
(38) The 1-SH assembled under ambient conditions also had a high ratio of
unbound-to-bound S 2p signal, but ellipsometrioc data indicated that
multilayer formation, rather than head-to-tail assembly, was the main cause
for the high ratio.
(39) (a) Ferreira, K. N.; Iverson, T. M.; Maghlaoui, K.; Barber, J.; Iwata, S.
Science 2004, 303, 1831-1838. (b) Alstrum-Acevedo, J. H.; Brennaman,
M. K.; Meyer, T. J. Inorg. Chem. 2005, 44, 6802-6827.
(40) (a) Fullerenes: From Synthesis to Optoelectronic Properties; Guldi, D.
M., Martin, N., Eds.; Kluwer Academic Publishers: Dordrecht, The
Netherlands, 2002. (b) Guldi, D. M.; Prato, M. Acc. Chem. Res. 2000, 33,
695-703.
(35) Cremlyn, R. J. An Introduction to Organosulfur Chemistry; John Wiley &
Sons: New York, 1996.
(36) Lau, K. H. A.; Huang, C.; Yakovlev, N.; Chen, Z. K.; O’Shea, S. J.
Langmuir 2006, 22, 2968-2971.
(41) (a) Imahori, H.; Fukuzumi, S. AdV. Funct. Mater. 2004, 14, 525-536. (b)
Imahori, H.; Yamada, H.; Nishimura, Y.; Yamazaki, I.; Sakata, Y. J. Phys.
Chem. B 2000, 104, 2099-2108. (c) Imahori, H.; Yamada, H.; Ozawa, S.;
Ushida, K.; Sakata, Y. Chem. Commun. 1999, 1165-1166.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13487

A R T I C L E S
Shirai et al.
It is well established that trithiol species,10 such as tripod 3,
often bind to Au surfaces with only one or two of the possible
three S-Au bonds forming.10k,m,n,31b This could explain the
slightly high unbound-to-bound ratio in tripod 3, compared to
the other monothiol derivatives 1 and 2, because not all three
thiols are bound to the Au surface in the SAM formed from 3.
This is especially true for the fullerene tripod 3 because the
strong fullerene-Au interaction will compete with other S-Au
interactions. Interestingly, a tripod molecule that does not
contain fullerene, compound 19 (Scheme 3), also had a relatively
high unbound-to-bound ratio (0.3) by XPS analysis (see the
Supporting Information for the XPS spectrum of 19).
tions.8 However, most of the previous reports assumed that
fullerene derivatives follow the usual chemistry of SAMs on
gold. Here we have shown that this assumption is inappropriate.
The way that fullerenes are integrated into the device will
seriously affect the overall device properties.
Experimental Section
Materials. The synthesis of fullerene derivatives 1-4 is reported
in the Supporting Information. 4-Bromobenzenethiol was obtained from
Lancaster and used as received. Anhydrous 1,2-dichlorobenzene
(ODCB) and anhydrous THF were obtained from Sigma-Aldrich and
degassed by bubbling nitrogen gas through them before use. Absolute
ethanol (EtOH) was obtained from Pharmco and degassed as above.
All other solvents used for self-assembly were freshly distilled over
CaH2. H2SO4 (96%) was obtained from Fisher and used as received.
Si(100) wafers (p-doped, test grade) were obtained from Silicon Quest
International.
Conclusions
In this article we have demonstrated that the seemingly simple
and widely employed approach to self-assembly of fullerene
derivatives on gold surfaces can be complicated due to multi-
layer formation and head-to-tail assemblies (Figure 2). These
fullerene SAM anomalies have not been previously considered
but were clearly detected using surface characterization tech-
niques including ellipsometry, CV, and XPS. Ellipsometry
analysis showed that the use of a combination of ODCB solvent
and protected thiols was necessary in some cases due to the
low solubility of fullerene derivatives and the insolubility of
byproduct generated by in situ disulfide formation. However,
the enhanced sulfur atom oxidation side effect of ODCB was
detected by XPS analysis. The side effect could be suppressed
by mixing EtOH with ODCB without drastically affecting the
solubility of fullerene derivatives in the ODCB mixture. The
XPS analysis of fullerene SAMs also showed a considerable
amount of surface-unbound sulfur atoms along with surface-
bound atoms. From this observation, and based on literature
indicating the existence of strong fullerene-Au surface adhe-
sion, we have proposed the head-to-tail assembly of fullerene
SAMs (Figure 2), in which molecules are assembled on Au via
fullerene-Au bonding, not the usual thiolate-Au bonding. The
other possible cause of the surface-unbound sulfur atoms is the
deposition of low-solubility fullerene derivatives on top of
SAMs and trapping of the fullerene-containing molecules within
the monolayer film, without forming a sulfur-gold bond, due
to strong lateral fullerene-fullerene interactions. However, the
assembly of disulfide 4, in which the possibility of deposition
of fullerene derivatives on top of SAMs is greatly reduced
because of good solubility and an inability to generate insoluble
disulfides in situ, further supports the head-to-tail assembly.
Interestingly, disulfide 4 shows the highest ratio of unbound-
to-bound S atoms even though it was designed to reduce the
ratio. The cross section mismatch between the fullerene head
and the tether to the anchoring sulfur can be a possible cause
of the head-to-tail assembly. However, it would be challenging
to eliminate the cross section mismatch because of the unusually
large size of the fullerene moiety. Electrochemistry on these
fullerene SAMs confirmed the effect of packing density on
electrochemical responses that has been reported previously.
Only the tripod 3, that is designed to isolate the electroactive
moiety in SAMs, showed ideal CV responses.
Preparation of the Au Substrate. The silicon wafer was cleaned
in a hot fresh Piranha (3:1 H2SO4/30% H2O2, v/v) solution for 10 min,
rinsed with flowing distilled water, ethanol, and dried in flowing pure
nitrogen gas. Caution: Piranha reacts Violently with organic com-
pounds, and care should be taken while handling this mixture. The Au
films were deposited by thermal evaporation of 150-200 nm thick Au
onto the silicon wafer with an ∼10 nm Cr adhesion layer at a rate of
1.2-1.3 Å/s at a base pressure of 1 × 10-6 torr. Before use, the Au
substrates were cut into ∼1 cm2 shards and cleaned by a UV/ozone
cleaner (Boekel Industries, Inc., model 135500) for 10 min, followed
by immersion in EtOH for at least 20 min.42
Chemical Assembly. The cleaned Au substrates were immersed into
the adsorbate solutions at room temperature for a specified period. All
the adsorbate solutions were freshly prepared in ambient or nitrogen
atmosphere and kept in the dark during assembly. Typical concentra-
tions were 15∼30 µM for 1 and 2 and 50∼100 µM for 3 and 4.
Precipitates that formed in the adsorbate solution of 1-SH or 2-SH
were filtered though a plug of silica gel under nitrogen atmosphere
before the immersion of the Au substrates. For in situ acid deprotec-
tion,23 ∼50 µL of H2SO4 (96%) was added to about 20 mL of adsorbate
solution, and the solution was incubated in a capped vessel for 30-60
min in order to deprotect the thiol moiety. After the assembly, the
samples were removed from the solutions and quickly immersed in
pure solvent of the same kind as in the adsorbate solution (for the
mixtures of ODCB/EtOH, pure ODCB was used) to avoid precipitation
of the fullerene derivatives, rinsed thoroughly with ODCB and EtOH,
and finally blown dry with nitrogen.
Ellipsometry. Measurements of surface optical constants and
molecular layer thicknesses were taken with a single-wavelength (632.8
nm laser) LSE Stokes ellipsometer (Gaertner Scientific). Measurements
were carried out before and immediately after monolayer adsorption.
The surface thickness was modeled as a single absorbing layer atop an
infinitely thick substrate (fixed ns). The index of refraction was set at
nf ) 1.55, kf ) 0. For growth kinetics measurements, each experiment
consisted of ∼20 Au substrates, and at each specified period, three
substrates were taken out from adsorption solutions and thicknesses
were measured for each substrate and an average value was reported.
X-ray Photoelectron Spectroscopy (XPS) Measurements. A PHI
Quantera SXM XPS/ESCA system at 5 × 10-9 torr was used to take
photoelectron spectra. All samples were transferred to the UHV system
within 20 min after the end of assembly to avoid excessive oxidation
in air. Typical assembly time was 20-24 h. A 114.8 monochromatic
Al X-ray source was applied for all measurements with an analytical
spot size of 0.10 mm × 0.10 mm, a 45° takeoff angle, and a pass energy
of 26.00 eV. High-resolution spectra of the S 2p region used a 45°
takeoff angle and 13.00 eV pass energy. Unless noted, surface samples
The application of fullerenes in a variety of molecule-based
device applications appears promising due to their advantageous
properties, and in fact, many studies on the SAM formation of
fullerene derivatives on gold have been reported as one of the
methods for the integration of fullerenes into device applica-
(42) Ron, H.; Matlis, S.; Rubinstein, I. Langmuir 1998, 14, 1116-1121.
9
13488 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Fullerene Derivative Monolayers on Gold Surfaces
A R T I C L E S
were referenced against the internal Au 4f7/2 line at 84.0 eV. Powder
XPS photoelectron lines were referenced against the C 1s signal at
284.50 eV.
reliably. The geometric areas of the electrodes were obtained from the
slope of a linear plot of the cathodic current versus (scan rate)1/2 for
the reversible reduction of the Fe(CN)63- in 0.1 M KCl, taking 0.76 ×
10-5 cm2/s as the diffusion coefficient and knowing the exact
concentration from the mass (see the Supporting Information for details).
Electrochemical Measurements. CV experiments were performed
on a BAS-CV 50W instrument. An aqueous Ag/AgCl reference
electrode and a platinum wire as the counter electrode were used. All
experiments were conducted under a N2 atmosphere using mixture of
ODCB/MeCN (4:1) containing 0.1 M Bu4NPF6 as the supporting
electrolyte. The gold substrates for routine CV analysis were prepared
by annealing the tip of a gold wire (99.999%, 0.5 mm diameter, Alfa
Aesar) in a gas-oxygen flame. Subsequently, the hot Au bead was
cooled in deionized water. The geometric area of the Au bead electrode
touching electrolyte solution was typically 0.013-0.024 cm2, which
was used to determine the coverage Γ from the integration of reduction
waves. For the determination of the surface coverage Γ using eq 1,
PEEK-encased Au disk electrodes were used because of the well-
defined surface area (typically 5.74-5.85 × 10-3 cm2), even though
reduction waves obtained with this electrode were too small to integrate
Acknowledgment. The Welch Foundation, DARPA/AFOSR,
and the NSF Penn State MRSEC funded this work. We thank
Drs. I. Chester of FAR Research Inc. and R. Awartani of Petra
Research Inc. for providing the TMSA.
Supporting Information Available: Detailed syntheses of the
new compounds, their NMR spectra, solution state CV of
selected compounds, details of the CV calculations, and XPS
spectra of 4-bromobenzenethiol and compound 19. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA063451D
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13489