Self-assembly of gold nanoparticles utilizing a charge-transfer interaction between carbazolyl and dinitrophenyl units

Article abstract of DOI:10.1246/bcsj.78.501

Colloidal gold nanoparticles self-assembled into macroscopic aggregates by a charge-transfer interaction between a 9-carbazolyl unit as an electron-donor immobilized on the surface of the gold nanoparticles and a bivalent linker containing two dinitrophenyl units as electron-acceptors. Transmission electron microscopy and scanning electron microscopy images showed spherical aggregates consisting of the gold nanoparticles.

Full text of DOI:10.1246/bcsj.78.501

ꢀ 2005 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 78, 501–505 (2005)
501
Self-Assembly of Gold Nanoparticles Utilizing a Charge-Transfer
Interaction between Carbazolyl and Dinitrophenyl Units
ꢀ
ꢀ
Kensuke Naka, Hideaki Itoh, and Yoshiki Chujo
Department of Polymer Chemistry, Graduate School of Engineering, Kyoto University,
Katsura, Nishikyo-ku, Kyoto 615-8510
Received March 17, 2004; E-mail: ken@chujo.synchem.kyoto-u.ac.jp
Colloidal gold nanoparticles self-assembled into macroscopic aggregates by a charge-transfer interaction between a
9-carbazolyl unit as an electron-donor immobilized on the surface of the gold nanoparticles and a bivalent linker con-
taining two dinitrophenyl units as electron-acceptors. Transmission electron microscopy and scanning electron micros-
copy images showed spherical aggregates consisting of the gold nanoparticles.
Metal nanoparticles have been extensively studied for dec-
ades. They have a wide functional diversity very different from
bulk materials, with much of their electronic,1 optical,2 and
r. t.
catalytic3 properties originating from their quantum-scale di-
mensions. Their unique properties are of potential utility in
chemical sensing, linear and nonlinear optics, and in a variety
of nanoscale electronic devices. That is, metal nanoparticles
should become essential building blocks for advanced devices.
In order to progress toward these applications, ways to organ-
ize the nanoparticles into controlled architectures, such as one-,
two-, and three-dimensional arrangements, must be found. One
of the more efficient methodologies for organizing metal nano-
particles involves self-assembly based on selective control
of non-covalent interactions. Several approaches (hydrogen-
bonding interaction,4 ꢀ–ꢀ interaction,5 host–guest interaction,6
van der Waals forces,7 electrostatic forces,8 antibody–antigen
recognition9) have been described concerning the formation
of arrays of metal nanoparticles. These results have provided
a powerful method for employing pre-programmed materials
with the potential for multi-dimensional ordering for the crea-
tion of well-defined structures at the molecular level. The or-
ganization of metal nanoparticles in superstructures of desired
shape and morphology is a challenging research area.
Recently, we have reported on the self-assembly of colloidal
gold nanoparticles into macroscopic aggregates by a charge-
transfer interaction (CT interaction) between a pyrenyl unit
as an electron-donor immobilized on the surface of gold nano-
particles and a bivalent linker containing two dinitrophenyl
units as electron-acceptors.10 This methodology allows the
gold nanoparticles to self-assemble reversibly into aggregates
by controlling the temperature. One of the advantages of this
system is that the binding strength can be easily tailored by
modifying the structure of the electron-donor and electron-
acceptor moieties, since the stability of CT complexes also de-
pend on the structure of the components of the CT complexes.
We expected that the size and shape of a nanoparticle assem-
bly would be controlled through turning the structure of the
components of the CT complexes.
N
SH
:
O2N
O2N
NO2
NO2
:
Fig. 1. Schematic illustration for thermally reversible self-
assembly of metal nanoparticles by charge-transfer inter-
action.
CT complex for a self-assembly process, here, we utilized dif-
ferent combinations of CT complexes between a 9-carbazolyl
unit as an electron-donor and dinitrophenyl units as electron-
acceptors in this system (Fig. 1). The addition of a bivalent
linker to a solution of individual gold nanoparticles resulted
in the formation of macroscopic aggregates comprising nonco-
valently linked gold nanoparticles through the CT interaction.
Results and Discussion
Synthesis of Gold Nanoparticles.
Gold nanoparticles
were prepared by the reduction of HAuCl4 (20.0 mg, 0.048
mmol) with NaBH4 in the presence of 11,110-dithiobis(undeca-
noic acid 2-(9-carbazolyl)ethyl ester) (1) (10 mg, 1:2 ꢁ 10ꢂ2
mmol). The UV–vis absorption spectrum of a red solution after
chemical reduction with NaBH4 indicated the formation of
gold nanoparticles with a surface plasmon absorption band at
500 nm.
The average diameter of the 1-modified gold nanoparticles
was 2:12 ꢃ 1:1 nm, as measured by transmission electron mi-
croscopy (TEM) (Fig. 2).
Figure 3 shows the 1H NMR spectra of 1 and the 1-modified
gold nanoparticles in CDCl3. In the 1H NMR of the 1-modified
In order to obtain detailed knowledge about the effect of a
Published on the web March 11, 2005; DOI 10.1246/bcsj.78.501

502 Bull. Chem. Soc. Jpn., 78, No. 3 (2005)
Self-Assembly of Gold Nanoparticles
NO2
O2N
O2N
CH2
CH2
S
N
OOC
COO CH2
OOC
10
2
10
NO2
2
2
1
O2N
NO2
COO
CH2
OOC
O2N
COO CH2
OOC
NO2
10
10
3
4
Chart 1.
(a)
S
CH2
CH2
CH2 CH2 COO
CH2
N
8
2
(A)
(B)
(C) (D)
CDCl3
H2O
(C) (D)
(A) (B)
(b)
50 nm
Fig. 2. TEM image of the 1-modifed gold nanoparticles.
CDCl3
H2O
gold nanoparticles signals appeared corresponding to those of
the proton signals of the parent 1 (Fig. 3b). The signals of the
methylene protons in the 1-modified gold nanoparticles are
broadened and shifted compared to those of 1. The signals
of methylene protons closest to the thiolate/Au interface can
be hardly recognized, since the motion of the methylene
groups close to the surface of the gold nanoparticles was con-
strained. This type of signal broadening was observed in bipyr-
idyl or alkanethiolate-modified metal nanoparticles.11 These
results indicate clear evidence for the attachment of thiolated
carbazole to the surface of the gold nanoparticles.
ppm
Fig. 3. 1H NMR spectra of (a) 1 and (b) the 1-modified gold
nanoparticles.
Assembly of Nanoparticles in the Presence of Linkers.
A bivalent 3,5-dinitrophenyl linker (2) was added to a benzene
solution of 1-modified gold nanoparticles at room temperature.
The solution color gradually faded with increasing the incuba-
tion time. The reaction of the 1-modified gold nanoparticles
with 2 was followed as a function of time through optical
changes in the surface plasmon band in the UV–vis absorption
spectrum. After the addition of 2, a broadening in the surface
plasmon resonance was observed. The absorption of the sur-
face plasmon band was almost saturated at 24 h after the addi-
tion of 2. Figure 4 shows the UV–vis absorption spectrum re-
corded at 24 h after the addition of 2 with varying concentra-
tions to colloidal solutions of the 1-modified gold nanoparti-
cles. Decreasing in the surface plasmon band became more
significant with increasing the concentrations of 2, indicating
the formation of particle aggregates.6 The UV–vis absorption
(a)
(b)
(c)
(d)
400
500
600
700
Wavelength / nm
Fig. 4. UV–vis absorption spectra by changing the weight
ratio of 2 against the 1-modified gold nanoparticles. 2/1-
modified gold nanoparticles = (a) 0, (b) 1, (c) 5, (d) 10.

K. Naka et al.
Bull. Chem. Soc. Jpn., 78, No. 3 (2005)
503
(a)
(b)
40
30
20
10
0
(1)
(2)
(3)
(4)
(5)
I0/I
0
5
10
15
20
25
330
380
430
480
[ Q ] × 104
Wavelength / nm
Fig. 5. (a) Fluorescence quenching of carbazolyl units on the surface of the gold nanoparticles by 5 in toluene (ꢁex ¼ 330 nm). The
concentration of 5 at decreasing emission intensities are (1) 0.0 M, (2) 1:76 ꢁ 10ꢂ4 M, (3) 7:08 ꢁ 10ꢂ4 M, (4) 1:15 ꢁ 10ꢂ3 M, (5)
1:93 ꢁ 10ꢂ3 M. (b) Stern–Volmer plots of the fluorescence quenching by 5 in toluene in the presence of the 1-modified gold nano-
particles. I0 and I represent the emission intensities in the absence and presence of 5, respectively.
spectrum showed no change for several days. This indicates
that the obtained aggregates maintained their size at room tem-
perature.
nism was substantiated by studying the temperature effect on
the fluorescence quenching efficiency.
TEM Image of the Assembly of Gold Nanoparticles.
TEM was used to characterize the aggregates. One drop of a
solution containing the obtained product was placed on a cop-
per grid and allowed to evaporate the solvent under atmo-
spheric pressure at room temperature. Although the TEM im-
age before the addition of 2 showed well-separated gold nano-
particles with an uniform size distribution (Fig. 2), the TEM
image measured at 24 h after the addition of 2 (5 mg) to a tol-
uene solution containing the 1-modified gold nanoparticles (1
mg) showed the formation of large, spherical aggregates with a
diameter of 1:5 ꢃ 0:7 mm (Fig. 6a). The microscopic aggre-
gates consisted of individual gold nanoparticles. When a
TEM image was measured at 6 h after the addition of 2 (5
mg), a spherical aggregate with an average diameter of 200
nm, in addition to smaller aggregates, was observed. The addi-
tion of 1 mg and 10 mg of 2 resulted in the formation of spher-
ical aggregates with diameters of 34 nm and 3:5 ꢃ 1:2 mm, re-
spectively. Thus, the degree of colloidal association would be
controlled by adjusting the concentration of 2 in the medium.
Equilibrium Constants. The stability of the CT com-
plexes depends on the structure of the dinitrophenyl unit. To
confirm the effect of the acceptor strength, 1,10-decanediyl
bis(3,4-dinitrobenzoate) (4) with o-dinitrophenyl units at each
end was used. In contrast to the results obtained with 2, the ad-
dition of 4 to the solution of the 1-modified gold nanoparticles
showed no change in the UV–vis absorption spectrum, indicat-
ing no aggregate formation of the gold nanoparticles by 4. The
absorption of the CT complexes was neglected compared with
that of the gold nanoparticles in a UV–vis absorption measure-
ment. To determine the equilibrium constants (Kc) of the CT
complex, 5, 6, and 7 were, thus, employed as model com-
pounds for 2, 4, and 1, respectively. The absorption spectra
of mixed solutions with a fixed concentration of 7 and varying
concentrations of 5 and 6 in toluene at room temperature are
shown in Fig. 7. The absorbance of the broad absorption band,
which can be seen up to 500 nm in Fig. 7, increased with an
increase in the concentration of 7 as the donor compound. This
band is ascribed to the CT complex of 5 or 6 with 7. A com-
Two control experiments provided additional insights into
the correlation of the spectral evolution. First, to confirm the
recognition property between the carbazolyl units and the dini-
trophenyl units, we used 1,10-decanediyl dibenzoate (3) with
two benzoate groups, which showed no interaction with the di-
nitrophenyl units. In contrast to the results obtained with 2, the
addition of 3 to a solution of the 1-modified gold nanoparticles
caused no color change and no change in the UV–vis absorp-
tion spectrum. This result indicates that the self-assembly of
the 1-modified gold nanoparticles proceeded through a specific
CT interaction. Second, a colorimetric change was not detect-
able using 1,3-dinitrobenzene instead of 2 for the solution of
the 1-modified gold nanoparticles, which showed a require-
ment of the bivalent linker for aggregation. These results clear-
ly indicate that the 1-modified gold nanoparticles aggregated
through cross-linking by 2, which recognized carbazolyl resi-
dues on the surface of the gold nanoparticles.
Fluorescence Quenching. According to literature,12 the
CT band between carbazole and dinitrobenzene is observed
around 450 nm, which shows a complete overlap with the
spectra of the surface plasmon band of the gold nanoparticles.
Thus, the UV–vis absorption spectrum gave no clear proof of
the presence of the CT complex. To confirm the presence of
the CT interaction in the present system, a fluorescence meas-
urement was carried out. Figure 5a shows the quenching of the
carbazolyl units on the gold nanoparticles by 5 as a model
ꢄ
compound of 2 in toluene at 25 C to eliminate the effect of
the particle aggregates. As the concentration of 5 increased,
the emission intensity decreased without any change in the
spectral pattern. This indicates that the quenching process oc-
curred via charge transfer from the carbazolyl units to the di-
nitrophenyl units. No exciplex emission was detectable in the
spectral region under these experimental conditions. Stern–
Volmer plots of the carbazolyl units of the fluorescence
quenching using 5 are shown in Fig. 5b. The plots are not
linear, indicating a static-type quenching mechanism via a
ground-state complex formation. The static quenching mecha-

504 Bull. Chem. Soc. Jpn., 78, No. 3 (2005)
Self-Assembly of Gold Nanoparticles
(a)
(b)
1 µm
200 nm
Fig. 6. (a) TEM image and (b) magnified TEM image of the spherical aggregates of the 1-modified gold nanoparticles.
(a)
(b)
(4)
(3)
(2)
(1)
(4)
(3)
(2)
(1)
400
450
500
400
450
500
Wavelength / nm
Wavelength / nm
Fig. 7. (a) UV–vis absorption spectra of the CT complexes between 5 and 7. The concentrations of 7 (solid line) are (1) 0 M, (2)
3:42 ꢁ 10ꢂ2 M, (3) 6:92 ꢁ 10ꢂ2 M, (4) 12:6 ꢁ 10ꢂ2 M. The concentration of 5 is 8:85 ꢁ 10ꢂ3 M. (b) UV–vis absorption spectra of
the CT complexes between 6 and 7. The concentrations of 7 (solid line) are: (1) 0 M, (2) 3:51 ꢁ 10ꢂ2 M, (3) 7:12 ꢁ 10ꢂ2 M, (4)
12:4 ꢁ 10ꢂ2 M. The concentration of 6 is 8:85 ꢁ 10ꢂ3 M.
bly process by turning the structure of the acceptor com-
O2N
O2N
O2N
O2N
pounds. Based on the Kc value of the CT complex of 5 and
7, we presume that 4% and 1% of the carbazolyl units on
the gold nanoparticles formed the CT complex with 2 after
the addition of 5 mg and 1 mg of 2, respectively.
N
CH2 CH3
COOCH3
COOCH3
5
6
7
Effect of the CT Complex for Self-Assembly Process.
The present result was compared with a previous result using
the CT complex between the pyrenyl unit and the dinitrophen-
yl unit. The equilibrium constant (Ks) of the CT complex be-
tween the carbazolyl unit and the dinitrophenyl unit is 2.2.
This value is smaller than that of the CT complex between
the pyrenyl unit and the dinitrophenyl unit (Ks ¼ 11:2) in the
previous result.10 Based on the comparison of these equilibri-
um constants, we expect that the aggregation of the 1-modified
gold nanoparticles via the CT complex between the pyrenyl
unit and the dinitrophenyl unit is efficient compared with the
CT complex between the carbazolyl unit and the dinitrophenyl
unit. A TEM investigation using the CT complex between the
pyrenyl unit and the dinitrophenyl unit demonstrated that
Chart 2.
parison of Figs. 7a and 7b revealed that the increase of the ab-
sorption band in Fig. 7a was larger than that in Fig. 7b. These
results suggest that the Kc values of the CT complexes with a
carbazolyl group decreased in the order of 3,5-dinitrophenyl
group > 3,4-dinitrophenyl group. The Rose and Drago equa-
tion13 was used to calculate the Kc value of the CT complex
from the spectrophotometric data. The order of the Kc values
of the CT complexes of different acceptors with 7 appeared
to be 5 (Kc ¼ 2:2) > 6 (Kc ¼ 1:1). This difference in the equi-
librium constants of the CT complex reflects the assembly
process. These results clearly show control of the self-assem-

K. Naka et al.
Bull. Chem. Soc. Jpn., 78, No. 3 (2005)
505
3.38 (2H, t), 4.14 (2H, t), 7.84–8.26(9H, m). Anal. Calcd for
C50H64 N2O4S2 (821.17): C, 73.13; H, 7.86; O, 7.79; S, 7.81;
N, 3.41%. Found: C, 73.08; H, 7.88; O, 7.74; S, 7.88; N, 3.42%.
1,10-Decanediyl Bis(3,5-dinitrobenzoate) (2), 1,10-Decane-
diyl Dibenzoate (3), 1,10-Decanediyl Bis(3,4-dinitrobenzoate)
(4), Methyl 3,5-Dinitrobenzoate (5), and Methyl 3,4-Dinitro-
benzoate (6). These compounds were synthesized according to
a previous paper.10
Colloid Synthesis. Gold nanoparticles were prepared accord-
ing to a procedure described by Brust.14 HAuCl4 (20.0 mg, 0.048
mmol) was dissolved in 30 mL of H2O. Tetraoctylammonium bro-
mide (26.5 mg, 0.048 mmol) was then added as a solution in 30
mL of toluene, and the reaction mixture was stirred until the yel-
low aqueous layer was clear and the organic layer was red. 1 (10
mg, 1:2 ꢁ 10ꢂ2 mmol) was then added, followed by the dropwise
addition of NaBH4 (5.23 mg, 0.14 mmol) as a solution in 5 mL of
H2O. This caused an immediate color change to dark black. The
reaction mixture was stirred for 10 min, and the organic phase
was collected and added to MeOH (100 mL). The precipitates
were isolated by centrifugation. Gold nanoparticles stabilized by
1 were obtained as a black powder.
spherical aggregates with a diameter of 1 ꢃ 0:7 mm were ob-
served at a feed weight ratio of the 2/1-modifed gold nanopar-
ticles = 1. Compared with the previous result, the size of
spherical aggregates (34 nm) at the same feed weight ratio is
small. That is, the stronger CT interaction results in the forma-
tion of larger spherical aggregates. The described experimental
results were in accordance with the expectation based on the
equilibrium constants of the CT complex, demonstrating con-
trol of the aggregate process by turning the CT complex.
Conclusion
In conclusion, we have demonstrated the self-assembly of
nanoparticles into macroscopic aggregates by the CT interac-
tion. The shape of the macroscopic aggregates formed using
this strategy is regular and spherical. The ability to tailor the
binding strength by this methodology would be highly desira-
ble in many applications. The present results suggest that the
self-assembly process of gold nanoparticles is controlled
through turning the structure of the acceptor compounds. We
expect that this concept represents a powerful and general
strategy for the creation of highly structured multifunctional
materials.
References
1
2
R. F. Khairutdinov, Collloid J., 59, 535 (1997).
a) P. Mulvaney, Langmuir, 12, 788 (1996). b) A. P.
Experimental
Materials. All solvents and reagents were obtained from com-
mercial sources and used as supplied except for the following.
Tetrahydrofuran (THF) was distilled under nitrogen.
Alivisatos, J. Phys. Chem., 100, 13226 (1996). c) K. G. Thomas,
B. I. Ipe, and P. K. Sudeep, Pure Appl. Chem., 74, 1731 (2002).
3 A. Roucoux, J. Schulz, and H. Patin, Chem. Rev., 102,
3757 (2002).
Measurement. 1H NMR spectra were obtained with a JOEL
JNM-EX270 spectrometer (270 MHz for 1H NMR) in chloro-
form-d. UV–visible spectra were measured on a Jasco V-530
spectrometer. Transmission electron microscopy was performed
using a JOEL JEM-100SX operated at 100 kV. Scanning electron
microscopy was performed using a JOEL JNM-5310/LV system.
Fluorescence emission spectra were recorded on a Perkin-Elmer
LS 50B luminescence spectrometer.
11,110-Dithiobis(undecanoic Acid). To a solution of 11-mer-
captoundecanoic acid (200 mg, 0.92 mmol) in H2O (100 mL) was
added sodium hydroxide (500 mg, 12.5 mmol). A hydrogen perox-
ide solution (2 mL) was added to a well-stirred resulting mixture.
After being stirred for 30 min, concentrated HCl (2 mL) was add-
ed and the mixture was extracted with ethyl acetate several times.
The organic portion was washed with saturated aqueous NaCl and
dried over MgSO4. After removing the solution, disulfide was
obtained as a colorless solid (110 mg, 55% yield). 1H NMR
(DMSO): ꢂ 1.23–1.35 (12H, m), 1.46 (2H, t), 1.57 (2H, t), 2.26
(2H, t), 2.69 (2H, t).
11,110-Dithiobis(undecanoic Acid 2-(9-Carbazolyl)ethyl Es-
ter) (1). To a solution of 1-pyrenebutanol (1.0 g, 3.64 mmol)
in CHCl3 was added 11,110-dithiobis(undecanoic acid) (0.7 g,
1.61 mmol), dicyclohexylcarbodiimide (1.0 g, 4.85 mmol), and
4-dimethylaminopyridine (0.06 g, 0.49 mmol). The mixture was
stirred at room temperature for 24 h. After removing the precipi-
tate by filtration, the solution was evaporated under reduced pres-
sure. The remaining white solid was subjected to column chroma-
tography. The first fraction containing the product was concentrat-
ed by evaporation and dried under reduced pressure; 1 was ob-
tained (20%). 1H NMR (CDCl3) ꢂ 1.12–1.41 (12H, m), 1.65
(4H, m), 1.80 (2H, t), 1.94 (2H, t), 2.27 (2H, t), 2.64 (2H, t),
4
a) A. K. Boal, F. Llhan, J. E. Derouchey, T. Thurn-
Albrecht, T. P. Russell, and V. M. Rotello, Nature, 404, 746
(2000). b) A. K. Boal and V. M. Rotello, J. Am. Chem. Soc.,
122, 734 (2000). c) J. J. Storhoff, A. A. Lazarides, R. C. Mucic,
C. A. Mirkin, R. L. Letsinger, and G. C. Schatz, J. Am. Chem.
Soc., 122, 4640 (2000).
5
J. Jin. T. Iyoda, C. Cao, Y. Song, L. Jiang, T. J. Li, and
D. B. Zhu, Angew. Chem., Int. Ed., 40, 2135 (2001).
a) J. Liu, S. Mendoza, E. Roman, M. J. Lynn, R. Xu, and
6
A. E. Kaifer, J. Am. Chem. Soc., 121, 4304 (1999). b) S. Y. Lin, S.
W. Liu, C. M. Lin, and C. H. Chen, Anal. Chem., 74, 330 (2002).
7
V. Patil, K. S. Mayya, S. D. Pradhan, and M. Sastry, J. Am.
Chem. Soc., 119, 9281 (1997).
F. Caruso, R. A. Caruso, and H. Mohwald, Science, 282,
1111 (1998).
8
9 W. Shenton, W. A. Davis, and S. Mann, Adv. Mater., 11,
449 (1999).
10 K. Naka, H. Itoh, and Y. Chujo, Langmuir, 19, 5496
(2003).
11 a) M. J. Hostetler, J. E. Wingate, J. E. Harris, R. W.
Vachet, M. R. Clark, J. D. Londono, S. J. Green, J. J. Stokes,
G. D. Wignall, G. L. Glish, M. D. Porter, N. D. Evans, and R.
W. Murray, Langmuir, 14, 17 (1998). b) K. Naka, M. Yaguchi,
and Y. Chujo, Chem. Mater., 11, 849 (1999).
12 Y. Shimazaki, M. Mitsuishi, S. Ito, and M. Yamamoto,
Langmuir, 14, 2768 (1998).
13 N. J. Rose and R. S. Drago, J. Am. Chem. Soc., 81, 6138
(1959).
14 M. Brust, M. Walker, D. Bethel, D. J. Schiffrin, and R.
Whyman, J. Chem. Soc., Chem. Commun., 1994, 801.