pH Dependence of Interparticle Coupling for Gold Nanoparticle Assemblies Formation: Electrostatic Attraction and Hydrogen Bonding

Article abstract of DOI:10.1246/bcsj.77.189

With the addition of ω-sulfanylalkanoic or ω -alkyldisulfanylalkanoic acids to sodium-contained aqueous Au colloidal solutions, close-packed self-assembly occurred UV-vis, TEM, solid-state ¹³C NMR, and FT-IR spectra were employed to study the resulting aggregation arrays. The resulting two-dimensional Au assemblies exhibited interparticle coupling via both hydrogen bonding and electrostatic interactions at pH 6. Either acidic (pH 3) or base (pH 11) generated significant three-dimensional close-packed aggregates. It is suggested that electrostatic interactions might lead to the aggregates during the dry-up process in the preparation of TEM samples of the Au nanoparticles capped with COO^- terminals at pH 11. For the acidic solutions (pH 3), coagulation was governed by hydrogen-bonding forces.

Full text of DOI:10.1246/bcsj.77.189

Ó 2004 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 77, 189–193 (2004)
189
pH Dependence of Interparticle Coupling for Gold Nanoparticle
Assemblies Formation: Electrostatic Attraction and Hydrogen Bonding
ꢀ
ꢀ
Chia-Hao Su, Pei-Lin Wu, and Chen-Sheng Yeh
Department of Chemistry, National Cheng Kung University, Tainan, Taiwan 701, R.O.C.
Received April 24, 2003; E-mail: csyeh@mail.ncku.edu.tw
With the addition of !-sulfanylalkanoic or !-alkyldisulfanylalkanoic acids to sodium-contained aqueous Au colloi-
dal solutions, close-packed self-assembly occurred UV–vis, TEM, solid-state 13C NMR, and FT-IR spectra were em-
ployed to study the resulting aggregation arrays. The resulting two-dimensional Au assemblies exhibited interparticle cou-
pling via both hydrogen bonding and electrostatic interactions at pH 6. Either acidic (pH 3) or base (pH 11) generated
significant three-dimensional close-packed aggregates. It is suggested that electrostatic interactions might lead to the ag-
ꢁ
gregates during the dry-up process in the preparation of TEM samples of the Au nanoparticles capped with COO ter-
minals at pH 11. For the acidic solutions (pH 3), coagulation was governed by hydrogen-bonding forces.
Nanostructured materials have given rise to a wide variety of
catalytic, optoelectronic, and biomedical applications.1,2 Using
nanocrystallites as building blocks to form large self-assembled
and self-organized architectures is an important strategy in
nanometer-size device fabrication. Such closely packed assem-
blies provide the opportunity for optimizing electronic and op-
tical performance with respect to function modulation. Hydro-
phobic alkanethiols have often been utilized to induce nanopar-
ticles toward self-organization through van der Waals interac-
tions when constructing nanostructural arrays in solutions.3
Another approach for generating two- (2D) or three-dimension-
blies through electrostatic attraction.6
It is known that either electrostatic interaction or hydrogen-
bonding has been individually employed to construct desired
assemblies. In this study, we demonstrated that !-sulfanylalka-
noic or !-alkyldisulfanylalkanoic acids can be mixed with Au
colloidal solutions to result in functionalized Au colloids exhib-
iting both bonding properties, electrostatic and/or hydrogen-
bonding interactions, to develop self-assembled architectures.
The closely packed 3D arrays were observed in the presence
of either electrostatic attractions (pH = 11) or hydrogen bonds
(pH = 3). The formation of 2D self-assembled structures was
attained in the presence of both electrostatic and hydrogen-
bonded coupling (pH = 6).
4
al (3D) aggregates is to make use of hydrogen bonds. Electro-
static coupling is the other non-covalent force used to form the
linkages between like-charged particles. Compared with both
hydrophobic and hydrogen-bonding interactions, only a hand-
ful of closely-packed structures have demonstrated the exis-
tence of electrostatic attractions among nanoparticles.5
Results and Discussion
The gold nanoparticles prepared according to Natan’s meth-
14
–7
od were 13 ꢂ 1:5 nm in size. When 5 mL of the gold colloidal
A notable feature of metal nanoparticles capped with carbox-
ꢁ
solution was mixed with 1.1 mmol of HS(CH2)nCOOH or
EtSS(CH2)nCOOH (n ¼ 10, 15), a concentration of 7:4 ꢃ
ylic acid- (–COOH) or carboxylate-terminated (–COO ) thio-
1
2
ꢁ1
lates is water-solubility. Nanoparticles fabricated with biocom-
patible properties are an important concept in biotechnology
applications. Carboxylic acids have been successfully function-
alized onto hydrophobic quantum dots to achieve bioconju-
10 particles mL was estimated on the basis of the average
particle diameter and the HAuCl4 concentration.15
When !-sulfanylalkanoic or !-alkyldisulfanylalkanoic
acids were added to the Au colloidal solutions, the suspensions
remained red in color. Figure 1 shows a typical example of the
UV–vis absorption. As can be seen, the distinct characteristic
Au plasmon band at 520 nm of the dispersed Au particles ap-
pears broadened, and increased in intensity after adding
HS(CH2)10COOH. This might indicate the formation of Au ag-
gregates. It should be mentioned that the prepared Au particles
from Natan’s method exhibited a dispersed morphology. TEM
images show the 2D arrays distributed as isolated domains ac-
companied with the dispersed particles on the grids from the re-
sulting solutions, as observed in Fig. 2.
8,9
gates. This methodology has been developed to synthesize
dispersed carboxylic acid- or carboxylate-capped metal nano-
4
a–c,6,7,10–13
particles.
For example, exchange reactions were per-
formed to substitute the !-carboxylic acid alkanethiols for the
hydrophobic alkanethiolates in the formation of these water-
soluble nanoparticles.4 Kimura and co-workers developed a
methodology to prepare water-soluble gold nanoparticles func-
tionalized with carboxylate groups.1 The development of 2D
or 3D arrays utilizing carboxylate- or carboxylic acid-modified
Au colloids was reported, as well. The groups of both Rotello
and Zhong studied the assemblies of the amphiphilic Au nano-
particles showing –COOH functionalities as hydrogen-bonding
a–c
1,12
Figure 3 displays the higher magnification TEM micro-
graphs corresponding to HS(CH2)nCOOH and EtSS(CH2)n-
COOH (n ¼ 10, 15) functionalized Au nanoparticles. The pH
was measured at 6 for those solutions. As seen in the TEM,
4
a,b
sites in building nanocrystal aggregations.
ꢁ
Adachi reported
that the –COO terminals facilitated Au nanospheroid assem-
Published on the web January 13, 2004; DOI 10.1246/bcsj.77.189

1
90 Bull. Chem. Soc. Jpn., 77, No. 1 (2004)
pH Dependence of Au Assemblies Formation
2
1
0
b
a
3
00
400
500
600
700
Wavelength/nm
Fig. 1. UV–vis absorption spectra of (a) Au colloidal solu-
1
00 nm
tion and (b) Au/HS(CH2)10COOH colloids.
Fig. 2. TEM images of Au/HS(CH2)10COOH colloids at pH 6.
Fig. 3. TEM images of (a) Au/HS(CH2)10COOH, (b) Au/EtSS(CH2)10COOH, (c) Au/HS(CH2)15COOH, and (d) Au/EtSS(CH2)15-
COOH colloids. All samples were controlled at pH 6.
the Au nanoparticles organized into regularly ordered 2D
monolayers exhibiting a local hexagonal order. Some arrays re-
vealed a 3D assembly leading to cross-linked networks. Those
colloidal arrays exhibited an interparticle distance dependence
corresponding to the length of the alkyl chain of mercaptocar-
boxylic acid. For example, the distances between the nanopar-
ticles were evaluated to be 2.0 nm for Au/HS(CH2)10COOH
and 2.7 nm for Au/HS(CH2)15COOH assemblies at pH 6. It
is known that chemisorption of alkanethiols on gold surfaces
13C NMR spectroscopy provided the evidence of the thiolate
bonding by investigating Au/HS(CH2)10COOH colloids (pH
6). As seen in Figs. 4a and 4b, no residual trisodium citrate re-
mained in the nanocrystal assemblies based on the disappear-
ance of the carbon signal at 76 ppm. This indicates that the con-
tribution of the trisodium citrate to formation of the organized
arrays is negligible. Furthermore, the feature (C11) at 24 ppm
(Fig. 4c) was not found in Fig. 4a. The excess !-sulfanyl-
alkanoic acids were washed out completely. On the contrary,
the C11 signal was shifted downfield to 52 ppm, the so called
16
leads to the formation of metal–sulfur bonding. Solid-state

C.-H. Su et al.
Bull. Chem. Soc. Jpn., 77, No. 1 (2004)
191
3
3 ppm
C3
C11
Au
COOH (COO-)
a
S
CH2COONa
COONa
CH2COONa
b
HO
C
C
COONa
CH
2
1
HS
1
3
c
COOH
COOH
C3, C11
d
COONa
NaS
COONa
Fig. 5. TEM images of (a) and (b) show the assemblies of
the colloidal Au/HS(CH2)10COOH at pH 11 and 3, respec-
tively.
Fig. 4. Solid-state 13C NMR spectrum of (a) Au/HS-
(
(
CH2)10COOH at pH 6, (b) trisodium citrate, (c) HS-
CH2)10COOH, and (d) NaS(CH2)10COONa.
a
Knight shift,17 suggesting S–H bonded to Au in the form of a
1
7
18
13
thiolate bond. As reported previously, the C chemical
shift at 33 ppm reflects an all-trans conformation of the alkyl
chains in a crystalline environment. On the basis of Figs. 4b–
1
713
1
559
b
c
1
3
ꢁ
4
d, both C carboxyl resonances (–COOH and –COO ) ap-
1
3
pear around 180 ppm. Therefore, the solid-state C NMR
Fig. 4a) did not allow for a conclusion in regards to the func-
tionalities of the modified Au colloids.
(
1
568
Since the Au nanoparticles can be functionalized with car-
boxylic acid/or carboxylate-terminated thiolates, the distinct
colloidal characteristics might be changed by tuning the pH.
The colloidal solutions (pH 6) were modulated using either
NaOH or HCl solutions with a pH of 11 or 3, respectively. In-
terestingly, all solutions using those four carboxylic acids sus-
tained the reddish color at pH 11 and showed no sign of precip-
itation. However, they turned purple and precipitation occurred
at pH 3. Both Au nanoparticles (pH 11 and 3) showed signifi-
cant 3D aggregation, as depicted in the Au/HS(CH2)10COOH
colloids of Fig. 5. Compared with those observed at pH 11,
the aggregates formed were larger and denser under acidic con-
ditions.
1
712
3
000
2000
1000
-1
Wavelength/cm
Fig. 6. FT-IR spectra attained from Au/HS(CH2)10COOH
colloids at pH (a) 6, (b) 11, and (c) 3.
tion, which opposes hydrogen-bonding attraction in the form of
the carboxylic acid groups. However, charged colloid self-as-
sembly could be achieved through electrostatic coupling origi-
nating in the counterions localized between the nanoparticles,
as shown in Scheme 1. Because the excess trisodium citrate
was introduced to synthesize the Au particles in our prepara-
To characterize the properties of the pH-modulated Au as-
semblies, FT-IR spectra were taken for the above three pH sam-
ples. The collected samples were washed with ethyl acetate and
water before measuring the IR spectra. Figure 6a (pH 6) illus-
þ
tion, it is highly possible that the counter ions, i.e. Na , are
ꢁ1
trates two characteristic peaks near 1710 and 1560 cm corre-
sponding to carboxylic acid dimmer and carboxylate, respec-
tively. The carboxylate groups on the particle surfaces could
lead to negatively charged colloids, resulting in repulsive inter-
actions. Such repulsive forces would disfavor particle aggrega-
present among the nanocolloids. It should be mentioned that
the prepared Au colloids without the addition of !-sulfanylal-
kanoic or !-alkyldisulfanylalkanoic acids showed no signals
from carboxylic and carboxylate peaks. With the addition of
NaOH, the pH rose to 11. Figure 6b shows only the carboxylate

1
92 Bull. Chem. Soc. Jpn., 77, No. 1 (2004)
pH Dependence of Au Assemblies Formation
Conclusions
We have demonstrated that Au nanoparticle assemblies re-
sulted from the simple addition of !-sulfanylalkanoic or !-al-
kyldisulfanylalkanoic acids into sodium-contained aqueous Au
colloidal solutions. The resulting 2D arrays revealed the char-
acteristics of the coexistence of both electrostatic and hydro-
gen-bonding attraction at pH 6. Although Au nanoparticles cap-
Au nanoparticle
electrostatic
ꢁ
ped with carboxylate-terminated (–COO ) thiolates could in-
þ
teract with Na to lead to aggregates formation via electrostatic
interactions in alkaline solutions, the observed 3D arrays might
also be due to coagulation during the dry-up process in the
preparation of TEM samples, considering of colloidal proper-
ties at pH 11. Under acidic conditions (pH 3), the colloidal so-
lutions exhibited precipitation and suggested the formation of
hydrogen-bonding interactions leading to 3D close-packed ag-
gregates. Furthermore, solid-state C NMR showed that the al-
kyl chains of the thiolates were present in an all-trans confor-
mation.
Au nanoparticle
hydrogen bonding
Scheme 1.
ꢁ1
ꢁ
13
peak at 1568 cm . Adachi has reported that the –COO termi-
nals facilitated Au nanospheroid assembly in the presence of
6
þ
counterions. The existence of Na ions indeed provided a
chance to form aggregates via electrostatic attraction in this
case. However, it also should be noted that the prepared Au col-
Experimental
ꢁ
loids carried significant COO -terminated thiols, resulting in
Chemicals. The following materials were obtained from Al-
drich: 11-bromoundecanoic acid, 16-bromohexadecanoic acid,
repulsive interactions under alkaline conditions. This results
in the dispersion of the gold colloids.4 As mentioned early,
we have observed that the resulting suspension remained a red-
dish color and exhibited good stability. Different from the ob-
servation at pH 6, TEM images showed significant 3D aggre-
gating morphology. Based on the aforementioned considera-
tion, it is proposed that the coagulation taking place may result
from the dry-up process in the preparation of TEM samples.
b,c
and thiourea. HAuCl4 3H2O (99.99%) was obtained from Alfa
ꢄ
Aesar. Trisodium citrate (99.0%) was obtained from Showa. The
chemicals and solvents were used as received.
Instrumentation. The transmission electron microscopy
TEM) was performed on a JOEL JEM-1200 EX instrument oper-
(
ating at 80 kV accelerating voltage. The UV–vis spectra were ob-
tained using a HP 8453 UV–Visible spectrophotometer. The IR
spectra were obtained using a Nicolet Magana 550 FT-IR spectro-
Herein, electrostatic interactions may play a role in aggregation
during the dry-up process.19 On the other hand, we saw the col-
loidal solutions turn purple and precipitate at pH 3. Figure 6c
provides the evidence of the carboxylic acid standing up at
1712 cm . In acidic solutions, interparticle interactions were
governed by hydrogen bonding facilitated particle aggrega-
The same consequences were observed in the case of
the Au colloids capped with !-alkyldisulfanylalkanoic acids
at pH 3, 6, and 11 (Fig. 7).
1
photometer. The solution H NMR spectra were obtained a Bruker
AVANCE 300 FT-NMR spectrometer; all chemical shifts were re-
ported in ppm from tetramethylsilane as an internal standard. The
1
3
solid-state C magic angle (MAS) NMR experiment was per-
formed on a Bruker AVANCE 400 spectrometer, equipped with
a Bruker double-tuned 7 mm probe with resonance frequencies
ꢁ
1
4
a,b
tion.
13
of 100.6 MHz for C nuclei. The Hartmann–Hahn condition for
1
13
the H ! C cross polarization (CP) experiment was determined
using admantane. 13C CP/MAS NMR spectra were recorded with
a CP contact time of 3 ms, a repetition time of 4 s, and a spinning
1
3
speed in 7 kHz. The C chemical shifts were externally referenced
with tetramethylsilane (TMS).
a
Synthesis of 11-Sulfanylundecanoic Acid (HS(CH2)10-
CO2H). A solution of 11-bromoundecanoic acid (0.5 g, 1.9
mmol) and thiourea (0.22 g, 2.9 mmol) in water (10 mL) was re-
1
559
b
c
1709
ꢁ3
fluxed for 2.5 h. Then, 5 M NaOHaq (1 M = 1 mol dm ) (4
mL) was added slowly with stirring and refluxed for another 2 h.
The aqueous solution was cooled in an ice bath and acidified with
hydrochloric acid to give a precipitate. After filtrating and washing
with water, 11-sulfanylundecanoic acid (0.37 g, 90%) was ob-
1
539
1
tained as a white powder. H NMR (CDCl3) ꢀ 1.27 (12H, m),
1
.32 (1H, t, J ¼ 7:5 Hz, SH), 1.60 (4H, m), 2.34 (2H, t, J ¼ 7:5
Hz), 2.51 (2H, q, J ¼ 7:5 Hz).
1
700
Synthesis of 11-Ethyldisulfanylundecanoic Acid (EtSS-
(
g, 1.7 mmol) and diethyl disulfide (0.42 g, 3.4 mmol) in 5 mL of
triethylamine was stirred at 70 C for 24 h. The solution was con-
centrated under reduced pressure followed by chromatographic pu-
CH2)10CO2H). A mixture of 11-sulfanylundecanoic acid (0.37
3000
2000
1000
ꢅ
-1
Wavelength cm
Fig. 7. FT-IR spectra attained from Au/EtSS(CH2)10COOH
colloids at pH (a) 6, (b) 11, and (c) 3.
rification over silica gel, eluting with CH3OH–CHCl3 (1:19) to
give 11-ethyldisulfanylundecanoic acid (0.42 g, 89%). H NMR
1

C.-H. Su et al.
Bull. Chem. Soc. Jpn., 77, No. 1 (2004)
193
(
2
CDCl3) ꢀ 1.26 (12H, m) 1.29 (3H, t, J ¼ 7:4 Hz), 1.61 (4H, m),
Mater., 10, 13 (1998). d) Z. L. Wang, S. A. Harfenist, R. L.
Whetten, J. Bentley, and N. D. Evans, J. Phys. Chem. B, 102,
3068 (1998). e) L. Motte, F. Billoudet, and M. P. Pileni, J. Phys.
Chem., 99, 16425 (1995).
.31 (2H, t, J ¼ 7:4 Hz), 2.66 (4H, m).
Synthesis of 16-Sulfanylhexadecanoic Acid (HS(CH2)15-
CO2H). The same procedure for the preparation of 11-sulfanyl-
undecanoic acid was used. 16-bromohexadecanoic acid (0.5 g,
1
fanylhexadecanoic acid (0.41 g, 95%). H NMR (CDCl3) ꢀ 1.26
4
See for example: a) Y. Shiraishi, D. Arakawa, and N.
.5 mmol) reacting with thiourea (0.17 g, 2.2 mmol) gave 16-sul-
1
Toshima, Eur. Phys. J. E, 8, 377 (2002). b) J. Simard, C. Briggs,
A. K. Boal, and V. M. Rotello, Chem. Commun., 2000, 1943. c)
W. Zhong, M. M. Maye, F. L. Leibowitz, and C. J. Zhong, Anal.
Chem., 72, 2190 (2000). d) A. C. Templeton, D. E. Cliffel, and
R. W. Murray, J. Am. Chem. Soc., 121, 7081 (1999). e) S. Gomez,
L. Erades, K. Philippot, B. Chaudret, V. Colli e´ re, O. Balmes, and
J.-O. Bovin, Chem. Commun., 2001, 1474. f) A. K. Boal, F. Ilhan,
J. E. DeRouchey, T. Thurm-Albrecht, T. P. Russell, and V. M.
Rotello, Nature, 404, 746 (2000). g) A. K. Boal and V. M. Rotello,
Langmuir, 16, 9527 (2000). h) L. Cusack, R. Rizza, A. Gorelov,
and D. Fitzmaurice, Angew. Chem., Int. Ed. Engl., 36, 848 (1997).
(
7
23H, m), 1.62 (4H, m), 2.35 (2H, t, J ¼ 7:5 Hz), 2.52 (2H, q, J ¼
:5 Hz).
Synthesis of 16-Ethyldisulfanylhexadecanoic Acid (EtSS-
(
CH2)15CO2H). The same procedure for the preparation of 11-
ethyldisulfanylundecanoic acid was used. 16-sulfanylhexadecano-
ic acid (0.41 g, 1.4 mmol) reacting with diethyl disulfide (0.35 g,
2
8
.8 mmol) gave 16-ethyldisulfanylhexadecanoic acid (0.41 g,
1
5%). H NMR (CDCl3) ꢀ 1.25 (22H, m), 1.31 (3H, t, J ¼ 7:4
Hz), 1.64 (4H, m), 2.33 (2H, t, J ¼ 7:4 Hz), 2.68 (4H, m).
Preparation of Functionalized Gold Nanoparticles. The Au
colloids were synthesized following the method of Natan et al. by
5
6
D. G. Grier, Nature, 393, 621 (1998).
E. Adachi, Langmuir, 16, 6460 (2000).
J. Kolny, A. Kornowski, and H. Weller, Nano Lett., 2, 361
ꢁ4
reacting 50 mL of 9:1 ꢃ 10 M HAuCl4 with an excess of triso-
7
(2002).
14
dium citrate in aqueous solution. The 5 mL of gold colloidal
aqueous solution was mixed with an ethanol solution containing
8
9
W. C. W. Chan and S. Nie, Science, 281, 2016 (1998).
G. P. Mitchell, C. A. Mirkin, and R. L. Lestinger, J. Am.
1
1
.1 mmol of HS(CH2)nCOOH or EtSS(CH2)nCOOH (n ¼ 10,
5). The resulting mixtures were used for UV–vis and TEM mea-
Chem. Soc., 121, 8122 (1999).
10 S. R. Johnson, S. D. Evans, and R. Brydson, Langmuir, 14,
6639 (1998).
11 S. Chen and K. Kimura, Langmuir, 15, 1075 (1999).
12 H. Yao, O. Momozawa, T. Hamatani, and K. Kimura,
Chem. Mater., 13, 4692 (2001).
surements. Electron micrographs of the Au colloids were carried
out using a drop of the sample onto a copper mesh coated with
an amorphous carbon film. This mesh was then dried in a vacuum
1
3
desiccator. For solid-state C NMR and FT-IR analysis, the col-
lected colloidal precipitates were washed three times with ethyl
acetate (5 mL) and water (5 mL) to remove the excess !-sulfanyl-
alkanoic or !-alkyldisulfanylalkanoic acids.
13 C. S. Weisbecker, M. V. Merritt, and G. H. Whitesides,
Langmuir, 12, 3763 (1996).
1
Anal. Chem., 67, 735 (1995).
15 R. H. Terrill, T. A. Postlethwaite, C. H. Chen, C. D. Poon,
A. Terzis, A. Chen, J. E. Hutchison, M. R. Clark, G. Wignall, J. D.
Londono, R. Superfine, M. Falvo, C. S. Johnson, E. T. Samulski,
and R. W. Murray, J. Am. Chem. Soc., 117, 12537 (1995).
4
C. Grabar, R. G. Freeman, M. B. Hommer, and M. J. Natan,
We thank the National Science Council of the Republic of
China for its financial support for this work. We also acknowl-
edge Ms. S. Y. Hsu and Mr. S. Y. Yao for the TEM measure-
ments and Ms. J. Z. Wu and Ms. S. Y. Sun for the solid-state
NMR analysis at Tainan Regional Instrument Center, National
Cheng Kung University.
1
1996). b) A. Badia, L. Demers, L. Dickinson, F. G. Morin, R.
6
See for example: a) A. Ulman, Chem. Rev., 96, 1533
(
B. Lennox, and L. Reven, J. Am. Chem. Soc., 119, 11104 (1997).
c) J. Noh, T. Murase, K. Nakajima, H. Lee, and M. Hara, J. Phys.
Chem. B, 104, 7411 (2000).
17 See for example: a) A. Badia, W. Gao, S. Singh, L. Cuccia,
and L. Reven, Langmuir, 12, 1262 (1996). b) B. S. Zelakiewicz, A.
C. de Diol, and Y. Y. Tong, J. Am. Chem. Soc., 125, 18 (2003).
18 H. Schmitt, A. Badia, L. Dickinson, L. Reven, and R. B.
Lennox, Adv. Mater., 10, 475 (1998).
References
1
A. Majetich, J. O. Artman, M. E. McHenry, N. T. Nuhfer,
and S. W. Staley, Phys. Rev. B: Condens. Matter, 48, 16845
(
1993).
2
J. J. Storhoff, R. Elghanian, R. C. Mucic, C. A. Mirkin, and
R. L. Lestinger, J. Am. Chem. Soc., 120, 1959 (1998).
See for example: a) S. L. Tripp and A. Wei, J. Am. Chem.
3
Soc., 123, 7955 (2001). b) M. Gianini, W. R. Caseri, and U. W.
Suter, J. Phys. Chem. B, 105, 7399 (2001). c) Z. L. Wang, Adv.
19 We acknowledge the editor and reviewer for the insightful
comments.