Noncovalent self-assembly of silver nanocrystal aggregates in solution

Article abstract of DOI:10.1021/jp0006173

Size monodisperse silver nanocrystals have been stabilized by chemisorption of long-chain alkane thiols incorporating a receptor site. When dispersed in a suitable solvent these nanocrystals recognize and selectively bind a long-chain alkane incorporating two complementary substrate sites and are noncovalently linked. The nanocrystal aggregates formed as a result have been studied by NMR, FT-IR, and dynamic light scattering. As a consequence, it has been possible to characterize the interactions between the receptor and substrate sites that lead to nanocrystal aggregation. It has also been possible to gain insights into the factors that affect the size, shape, and internal structure of the nanocrystal aggregates formed. An important insight is that the kinetics of nanocrystal aggregation, and as a consequence the structure of the nanocrystal aggregates formed, depends on the number of receptor sites at the surface of a nanocrystal.

Full text of DOI:10.1021/jp0006173

6164
J. Phys. Chem. B 2000, 104, 6164-6173
Noncovalent Self-Assembly of Silver Nanocrystal Aggregates in Solution
Stephen Fullam, S. Nagaraja Rao, and Donald Fitzmaurice*
Department of Chemistry, UniVersity College Dublin, Belfield, Dublin 4, Ireland
ReceiVed: February 17, 2000
Size monodisperse silver nanocrystals have been stabilized by chemisorption of long-chain alkane thiols
incorporating a receptor site. When dispersed in a suitable solvent these nanocrystals recognize and selectively
bind a long-chain alkane incorporating two complementary substrate sites and are noncovalently linked. The
nanocrystal aggregates formed as a result have been studied by NMR, FT-IR, and dynamic light scattering.
As a consequence, it has been possible to characterize the interactions between the receptor and substrate
sites that lead to nanocrystal aggregation. It has also been possible to gain insights into the factors that affect
the size, shape, and internal structure of the nanocrystal aggregates formed. An important insight is that the
kinetics of nanocrystal aggregation, and as a consequence the structure of the nanocrystal aggregates formed,
depends on the number of receptor sites at the surface of a nanocrystal.
Introduction
SCHEME 1
The size-dependent properties of nanocrystals have been
studied in detail.1-9 Increasingly, however, it is the collective
properties of organized assemblies of nanocrystals that are being
studied.3,10,11 These studies are being facilitated in large part
by the preparation of superlattices of size-monodisperse
nanocrystals.1,5,10-14 Of particular interest is how the size-
dependent electronic and magnetic properties of the constituent
nanocrystals can be exploited to tune their collective properties.
Generally, nanocrystal superlattices are prepared at an air-
water interface using Langmuir-Blodgett techniques (two-
dimensional)12 or on a suitable substrate by controlled solvent
evaporation (three-dimensional).10 Both approaches, however,
are limited by the fact that only relatively simple nanocrystal
architectures can be realized. For this reason, approaches that
permit the assembly of complex nanocrystal architectures are
of particular interest.
One possible approach is to adsorb stabilizer molecules
incorporating a receptor site at the surface of a nanocrystal. It
is expected that this nanocrystal will recognize and selectively
bind another at which is adsorbed stabilizer molecules incor-
porating a complementary substrate site. By this means, the
assembly of complex nanocrystal architectures in solution can
be programmed.15
To date, TiO2 nanocrystals stabilized by physisorbed long-
chain alkanes incorporating a receptor site (uracil moiety) have
been programmed to recognize and selectively bind in solution
TiO2 nanocrystals stabilized by physisorbed long-chain alkanes
incorporating a substrate site (diaminopyridine moiety).16
Limited ordering of the constituent nanocrystals was apparent
in the mesoaggregates formed as a result.
While encouraging, the specific approach outlined above has
a number of limitations. Among these are the following. First,
the stabilizer molecules, incorporating either a receptor or
substrate site, were not chemisorbed at the nanocrystal surface.
Second, the modified nanocrystals were not size-monodisperse.
To address the first of these limitations, a methodology was
developed that permits the controlled chemisorption of a
stabilizer molecule, incorporating the desired receptor or
substrate site, at the surface of the nanocrystals of a stable
dispersion.17 This methodology was used to prepare size-
polydisperse gold nanocrystals stabilized by a chemisorbed
mixture of a long-chain alkane thiol, dodecane thiol (95%), and
a long-chain dodecane thiol incorporating a receptor site, 12-
mercapto-dodecyl-1-uracil (5%). It was established that these
nanocrystals recognize and selectively bind in solution a long-
chain alkane incorporating a complementary substrate site, N,N-
2,6-pyridinediylbis[undecamide] (Scheme 1).
To address the second of these limitations, size-monodisperse
silver nanocrystals stabilized by a chemisorbed long-chain
alkane thiol, dodecane thiol I (82%), and a chemisorbed long-
chain alkane thiol incorporating a receptor site, N,N-2,6-
pyridinediylbis[undecamide]-4-oxy-[12-mercapto dodecanyl] II
(8%), have been prepared (Scheme 2). These nanocrystals and
their dispersions are denoted Ag-(I+II). Size-monodisperse
silver nanocrystals, stabilized by a chemisorbed long-chain
alkane thiol incorporating a receptor site, II (100%), have also
been prepared. These nanocrystals and their dispersions are
denoted Ag-II.
When dispersed in chloroform, it was expected that Ag-(I+II)
and Ag-II would recognize and selectively bind long-chain
alkanes incorporating two complementary substrate sites, (1,12-
* Corresponding author.
10.1021/jp0006173 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/14/2000

Silver Nanocrystal Aggregates in Solution
J. Phys. Chem. B, Vol. 104, No. 26, 2000 6165
SCHEME 2
Anal. Calcd. for I (C12H26S): C, 71.21; H, 12.95; S, 15.89.
Found: C, 71.19; H, 13.12; S, 15.55. 1H NMR (chloroform-d):
δ 0.88 (t, 3H, J ) 7.0 Hz); δ 1.26 (s, 18H); δ 1.61 (q, 2H, J )
7.3 Hz); δ 2.52 (q, 2H, J ) 7.0 Hz).
Anal. Calcd. for II (C41H75N3O3S): C, 71.48; H, 10.74; N,
6.10; S, 4.65. Found: C, 70.75; H, 10.63; N, 5.59; S, 4.35. 1H
NMR (chloroform-d): δ 0.89 (t, 6H, J ) 6.8 Hz); δ 1.28-
1.44 (m, 48H); δ 1.68-1.79 (m, 8H); δ 2.36-2.39 (t, 4H, J )
7.3 Hz); δ 2.69-2.72 (t, 2H, J ) 7.3 Hz); δ 4.03-4.06 (t, 2H,
J ) 6.3 Hz); δ 7.56 (s, 2H); δ 7.64 (s, 2H).
Anal. Calcd. for III (C26H40N6O6): C, 58.66; H, 7.51; N,
15.78. Found: C, 53.92; H, 6.90; N, 13.87. 1H NMR (chloroform-
d): δ 1.26-1.64 (m, 20H); δ 2.22 (s, 6H); δ 3.35-3.41 (t, 4H,
J ) 7.31 Hz); δ 4.26 (s, 4H); δ 5.60 (s, 2H); δ 9.50 (s, 2H); δ
9.70 (s, 2H).
Anal. Calcd. for IV (C28H51N3O3): C, 70.44; H, 10.69; N,
8.80. Found: C, 71.25; H, 10.95; N, 8.95. 1H NMR (chloroform-
d): δ 0.88 (t, 6H, J ) 7.0 Hz); δ 1.26-1.58 (m, 40H); δ 2.36
(t, 2H, J ) 7.8 Hz); δ 4.16 (s, 2H); δ 5.52 (s, 1H); δ 8.14 (s,
1H, -NH amidic); δ 9.51 (s, 1H, -NH imidic).
acetoamide dodecane) bis-6-methyluracil III (Scheme 3). It was
also expected that the ability of III to be bound by two receptor
sites on different nanocrystals would promote aggregation of
Ag-(I+II) and Ag-II. NMR, FT-IR, and dynamic light scattering
(DLS) have all been used to study the kinetics of aggregation
of these nanocrystals and the structure of the aggregates formed.
The principal objectives of this study were the following.
First, characterize the interactions at the molecular level between
the receptor modified nanocrystals and the bifunctional sub-
strates in solution that lead to nanocrystal aggregation. Second,
to gain insights into the factors that control the size, shape and
structure of the nanocrystal aggregates formed as a result.
Preparation of Receptor-Modified Nanocrystals. Ag-I and
Ag-(I+II) were prepared following the method of Brust et al.18
Briefly, AgNO3 (0.154 g) was dissolved in H2O (30 mL) and
the phase transfer catalyst (C8H17)4NBr (2.228 g), dissolved in
chloroform (30 mL) and added with stirring for 1 h. A solution
of I (0.825 g) or a mixed solution (80:20 mole ratio) of I (0.066
g) and II (0.057 g) in chloroform (8 mL) was then added (Ag/S
mole ratio of approximately 1:1). After stirring for 2 to 3 min,
the reducing agent NaBH4 (0.394 g), dissolved in H2O (24 mL),
was added and stirring continued overnight. The nanocrystals
formed were precipitated, filtered, and washed (using ethanol)
to obtain a dry gray powder redispersible in chloroform.
Ag-II was prepared using a related approach. Briefly, AgNO3
(0.154 g) was dissolved in H2O (30 mL) and the phase transfer
catalyst, (C8H1704)NBr (2.228 g) dissolved in chloroform (30
mL), was added with stirring for 1 h. At this point the silver
salt was reduced by addition of NaBH4 (0.394 g) in H2O (24
Experimental Section
Synthesis of Stabilizer-Receptor and Substrate Molecules.
Compound I was used as supplied by Aldrich. Compound II
was prepared as shown (Scheme 4). Compound III was prepared
as shown (Scheme 5). Compound IV was prepared as described
in detail elsewhere.15 All molecules were characterized by
1
elemental analysis and H NMR.
SCHEME 3

6166 J. Phys. Chem. B, Vol. 104, No. 26, 2000
Fullam et al.
SCHEME 4: Reagents and Conditions: (i) EtOH, 2% H2SO4, Reflux, (12 h); (ii) K2CO3, Br(CH2)12SH, Acetone (dry),
Reflux (72 h); (iii) EtOH, N2H4, Reflux (3 h); (iv) 2M HCl, NaNO2, H2O; (v) EtOH, Reflux (24 h); (vi) Ethanolic KOH,
Reflux (4 h); (vii) Pyridine (dry), Lauryl Chloride, CHCl3 (dry)
SCHEME 5: Reagents and Conditions: (i) H2N(CH2)12NH2, CH3CN, Reflux, (12 h); (ii) Acetyl Chloride, CHCl3 (dry),
Pyridine (dry)
III
mL). The chloroform phase was concentrated prior to precipita-
tion of the silver nanocrystals by addition of a large volume of
ethanol and their recovery by centrifugation. The phase transfer
catalyst stabilized silver nanocrystals (20 mg) were redispersed
in acetone (4 mL) to which II (40 mg) in chloroform (4 mL)
was added. After stirring for 6 days, these nanocrystals were
precipitated by addition of a large volume of ethanol. The
precipitated nanocrystals were filtered and washed with ethanol
to obtain a dry gray powder redispersible in chloroform.
laser was used as the light source (100 mW). The temperature
was maintained at 25 °C ( 0.02 °C throughout by an external
circulator. All samples were filtered through 0.02 µm inorganic
Anotop filters. Having been established that effects due to
absorption at 488.0 nm by the silver nanocrystals are negligible,
a silver nanocrystal concentration of 9.0 × 10-10 mol dm-3 was
used. Additions to nanocrystal dispersions were made using a
calibrated microsyringe and solutions previously filtered through
0.02 µm inorganic Anotop filters.
1
Characterization Techniques. H NMR spectra were ob-
Transmission electron micrographs (TEMs) were obtained
using a JEOL TEMSCAN 2000 EX and graphite coated copper
grids.
tained using either a JEOL JNM-GX270 FT or Varian 500 FT
spectrometer. FT-IR spectra were obtained using a Mattson
Infinity FT spectrometer and a CaF2 cell (0.20 mm path length).
Dynamic light scattering (DLS) studies were performed using
a Malvern PCS-4700 instrument equipped with a 256 channel
correlator. The 488.0 nm line of a Coherent Innova-70 Ar ion
Results and Discussion
Characterization of Receptor-Modified Nanocrystals. Sil-
ver nanocrystals stabilized by the alkane thiol I and the modified

Silver Nanocrystal Aggregates in Solution
J. Phys. Chem. B, Vol. 104, No. 26, 2000 6167
Figure 1. (a) TEM of Ag-(I+II) and histogram nanocrystal diameters determined from image. (b) As in (a) for Ag-II.
alkane thiol incorporating a receptor site II, Ag-(I+II), have
been characterized by TEM and elemental analysis as described
in detail elsewhere.17-19 It has been established by TEM that
these nanocrystals have a spherical silver core that is 71 Å in
diameter, a size polydispersity of less than 10% and, as can be
seen from Figure 1, form a hexagonally close-packed array upon
solvent evaporation.20 It has been further established by
elemental analysis that there are on average 957 molecules
adsorbed at the surface of each nanocrystal and that the ratio
of I:II is 12:1.21 On this basis, and assuming that the average
area occupied by a molecule of II adsorbed at the surface of a
silver nanocrystals is 23 Å2 (see below), it is calculated that
the average area occupied by a molecule of I adsorbed at the
surface of a silver nanocrystal is 16 Å2. This value agrees well
those previously reported for alkane thiols adsorbed at the
surface of a silver nanocrystal.17,19
been established by TEM that these nanocrystals have a
spherical silver core that is 88 Å in diameter, a size polydis-
persity of less than 5% and, as can be seen from Figure 1, form
a hexagonally close-packed array upon solvent evaporation.22
It has been established by elemental analysis that there are on
average 1088 molecules of II adsorbed at the surface of each
nanocrystal.23 On this basis it is calculated that the average area
occupied by the stabilizer molecules adsorbed at the surface of
a silver nanocrystal is 23 Å2, the value assumed above. This
value, as might have been expected, is somewhat larger than
values previously reported for unmodified alkane thiols adsorbed
at the surface of a silver nanocrystal.17,19
Receptor-Substrate Binding in Solution. To establish that
the diaminopyridine receptor site incorporated in II recognizes
and selectively binds the complementary uracil substrate sites
1
incorporated in III to form a 2:1 complex, the H NMR and
Silver nanocrystals stabilized by adsorbed alkane thiol
incorporating a diaminopyridine receptor site II, Ag-II, have
also been characterized by TEM and elemental analysis. It has
FT-IR spectra of II, III, and different mixtures of II and III
were measured in chloroform-d at 25 °C.
Shown in Figure 2 are the H NMR spectra of II (1.45 ×
1

6168 J. Phys. Chem. B, Vol. 104, No. 26, 2000
Fullam et al.
of III are added the imidic resonance of III is shifted upfield
to 10.19. The resonance assigned amidic proton of III is
observed at δ 9.56 throughout.
Also shown in Figure 2 are the 1H NMR spectra of III (1.88
× 10-3 mol dm-3) prior to and following addition of the
indicated number of molar equivalents of II (relative to III) in
0.25 molar equivalent amounts. Also shown is the spectrum of
II (0.47 × 10-3 mol dm-3, 0.25 molar equivalent).
Following each addition of 0.25 of an equivalent of II to III
there is a downfield shift in the imidic proton resonance of III
by about δ 0.25. After 1.25 equivalents of II are added, the
amidic proton resonance of IV is shifted downfield to δ 10.64.
Following addition of 0.25 of an equivalent of II to III there is
a downfield shift in the amidic proton resonance of II to δ 8.88.
Each subsequent addition of 0.25 of an equivalent of II is
accompanied by a small upfield shift. After 1.25 equivalents of
II are added, the amidic resonance of II is shifted upfield to
8.53. The resonance assigned amidic proton of III is observed
at δ 9.56 throughout.
These findings are understood as follows. Upon addition of
0.25 of an equivalent of III to II, a large fraction both substrate
sites of III is complexed to II, while only a small fraction of II
is complexed to III. As a consequence, the resonance assigned
to the imidic proton of III is shifted downfield to a value
characteristic of III in the 2:1 complex shown in Scheme 6
and associated by two triple arrays of complementary hydrogen
bonds.24 Under the same conditions the resonance assigned to
the amidic proton of II is shifted downfield, but only by a small
amount, to a value that remains characteristic of II in solution.24
In the course of four subsequent additions of 0.25 equivalents
of III, the resonance assigned to the imidic proton of III is
observed to shift back upfield as the fraction of uncomplexed
III present in solution increases. Under the same conditions,
the resonance assigned to the amidic proton in II is shifted
downfield to a value characteristic of the complex shown in
Scheme 6. Analogous observations made upon addition of II
to III are consistent with formation the 2:1 complex shown in
Scheme 6 and are associated by two triple arrays of comple-
mentary hydrogen bonds.
1
Figure 2. (a) H NMR spectra of (i) II (1.45 × 10-3 mol dm-3) and
It should be noted that because the dynamics of complex
formation are fast on the NMR time scale, the chemical shifts
observed for the amidic and imidic peak positions are the
population-weighted average of chemical shifts of II and III in
the complexed and uncomplexed states. For this reason the end
point of a titration is not well defined. It should also be noted
that the chemical shifts of the amidic and imidic resonances in
II and III are concentration dependent.
(ii) III (0.36 × 10-3 mol dm-3) in chloroform-d at 25 °C. Also shown
are the H NMR spectra of II (1.45 × 10-3 mol dm-3) in chloroform
1
at 25 °C to which the following molar equivalents of III (0.36 × 10-3
mol dm-3) have been added: (iii) 0.25; (iv) 0.50; (v) 0.75; (vi) 1.00;
and (vii) 1.25. (b) 1H NMR spectra of (i) III (1.88 × 10-3 mol dm-3
)
and (ii) II (0.47 × 10-3 mol dm-3) in chloroform-d at 25 °C. Also
shown are the 1H NMR spectra of III (1.88 × 10-3 mol dm-3) in
chloroform at 25 °C to which the following molar equivalents of II
(0.47 × 10-3 mol dm-3) have been added: (iii) 0.25; (iv) 0.50; (v)
0.75; (vi) 1.00; and (vii) 1.25.
Shown in Figure 3 are the FT-IR spectra of II (2.90 × 10-3
mol dm-3), III (1.45 × 10-3 mol dm-3), and a mixture of II
(2.90 × 10-3 mol dm-3) and III (1.45 × 10-3 mol dm-3) in
chloroform-d at 25 °C. The bands at 3420 and 3392 cm-1 are
assigned to the amidic N-H stretches of II and III, respec-
tively.25 On preparing the above mixture of II and III, the
intensity of the bands assigned to the free amidic N-H stretches
decreases.26 Also on preparing the above mixture of II and III,
a series of bands between 3200 and 3300 cm-1, assigned to the
hydrogen-bonded amidic protons are newly observed.27 These
findings are consistent with formation of a 2:1 complex between
II and III associated by formation of two triple arrays of
complementary hydrogen bonds,28 and therefore are also
10-3 mol dm-3) prior to and following addition of the indicated
number of molar equivalents of III (relative to II) in 0.25 molar
equivalent amounts. Also shown is the spectrum of III (0.36 ×
10-3 mol dm-3, 0.25 molar equivalent).
The resonance assigned to the amidic protons of II is observed
at δ 7.96, while the resonances assigned to the imidic and amidic
protons of III are observed at δ 9.16 and δ 9.56, respectively.
Following each addition of 0.25 of an equivalent of III to II
there is a downfield shift in the amidic proton resonance of II
by about δ 0.1. After 1.25 equivalents of III are added, the
amidic proton resonance of II is shifted downfield to δ 8.61.
Following addition of 0.25 of an equivalent of III to II there is
a downfield shift in the imidic proton resonance of III to δ
10.50. Each subsequent addition of 0.25 of an equivalent of III
is accompanied by a small upfield shift. After 1.25 equivalents
1
consistent with the findings of the H NMR studies reported
above.
It is clear, based on the findings presented above, that a
modified alkane thiol incorporating a diaminopyridine receptor

Silver Nanocrystal Aggregates in Solution
J. Phys. Chem. B, Vol. 104, No. 26, 2000 6169
SCHEME 6
SCHEME 7
site recognizes and selectively binds a modified alkane thiol
incorporating two uracil substrate sites. The question which
arises next is whether silver nanocrystals stabilized by an alkane
thiol incorporating a diaminopyridine receptor site will also
recognize and selectively bind a modified alkane thiol incor-
porating two uracil substrate sites.
Receptor-Modified Nanocrystal-Substrate Binding in
1
Solution. Presented are the findings of H NMR and FTIR
studies which establish that silver nanocrystals stabilized by II,
specifically Ag-(I+II), recognize and selectively bind IV in
solution. That is, findings which establish that the diaminopy-
ridine receptor site incorporated in II continues to recognize
and selectively bind the complementary uracil substrate site
incorporated in IV, even when chemisorbed at the surface of a
silver nanocrystal (Scheme 7).
1
Shown in Figure 4 are the H NMR spectra of Ag-(I+II)
(1.88 × 10-5 mol dm-3 silver nanocrystals, 1.40 × 10-3 mol
dm-3 II) in chloroform-d at 25 °C prior to and following
successive additions of 0.15 molar equivalents of IV (with
respect to II). Also shown is the spectrum of IV (0.21 × 10-3
mol dm-3, 0.15 molar equivalents) in chloroform-d at 25 °C.
Initially, the resonances assigned to the amidic protons of II
in Ag-(I+II) are observed at δ 7.55, while the resonances
assigned to the imidic and amidic protons of IV are observed
at δ 7.65 and δ 9.35, respectively. Following addition of 0.15
molar equivalents of IV to the Ag-(I+II) dispersion, the amidic
Figure 3. FTIR spectra of (i) II (2.90 × 10-3 mol dm-3) and (ii) III
(1.45 × 10-3 mol dm-3) in chloroform-d at 25 °C. Also shown is the
FTIR spectrum (iii) of a 2:1 molar ratio mixture of II (2.90 × 10-3
mol dm-3) and III (1.45 × 10-3 mol dm-3) in chloroform at 25 °C.

6170 J. Phys. Chem. B, Vol. 104, No. 26, 2000
Fullam et al.
The band at 3420 cm-1 is assigned to the amidic stretches of
II. The band at 3392 cm-1 is assigned to the imidic N-H stretch
of IV. It is observed that the intensity of the band assigned to
the amidic N-H stretch of II decreases as the concentration of
added IV increases. At the same time, the increase in intensity
of the band assigned to the imidic N-H stretch of IV is less
that that observed for the same concentration of IV in solution.
The newly observed bands between 3300 cm-1 and 3200 cm-1
are assigned to formation of the hydrogen bonds between the
receptor and substrate.
It is clear, based on the 1H NMR and FT-IR findings
presented above, that a silver nanocrystal stabilized by an alkane
thiol and an alkane thiol incorporating a receptor site, namely
Ag-(I+II), recognizes and selectively binds in solution a
molecule incorporating a complementary substrate site, namely
IV. Furthermore, it is also clear that the resulting 1:1 complex
is noncovalently associated by a triple array of hydrogen bonds
as shown in Scheme 7.
It should be noted that these detailed studies were not
undertaken using the modified alkane thiol incorporating two
uracil substrate states, namely III. The reason for this, as
discussed in detail below, is that the addition of III to Ag-(I+II)
promotes nanocrystal aggregation, thereby precluding detailed
characterization of the molecular interactions leading to ag-
gregation by NMR and to a lesser extent by FT-IR.
Aggregation of Receptor-Modified Nanocrystals in Solu-
tion. A question which arises is whether the ability to program
a size-monodisperse silver nanocrystal to recognize and selec-
tively bind a molecular substrate in solution can be exploited
to assemble complex and technologically relevant nanocrystal
assemblies, also in solution.
With respect to this question, we note that a large number of
studies have been directed toward understanding the mechanisms
of colloid aggregation and their relationship to the structure of
the aggregates formed.29 Generally, two limiting cases have been
identified.30,31 The first, termed slow or reaction-limited ag-
gregation, refers to the case where only a small fraction of the
collisions result in the two colloidal particles involved adhering
to each other.30 It has been found that the growth kinetics of
reaction-limited aggregation are described by eq 1
1
Figure 4. (a) H NMR spectra of (i) Ag-(I+II) (1.88 × 10-5 mol
Rh(t) ) Rh(t0) exp(ct)
(1)
dm-3 nanocrystals, 1.40 × 10-3 mol dm-3 II) and (ii) IV (0.21 × 10-3
mol dm-3) in chloroform-d at 25 °C. Also shown are the H NMR
1
where Rh(t0) is the initial hydrodynamic radius of the aggregate,
typically determined by dynamic light scattering, and c is a
parameter characteristic of the experimental conditions. Related
static light scattering experiments have established that the
fractal dimension of such aggregates is 2.05.
spectra of Ag-II (1.88 × 10-5 mol dm-3 nanocrystals, 1.40 × 10-3
mol dm-3 II) in chloroform at 25 °C to which the following molar
equivalents of IV (0.21 × 10-3 mol dm-3) have been added: (iii) 0.15;
(iv) 0.30; (v) 0.45; (vi) 0.60; and (vii) 0.75. (b) FTIR spectra measured
under same conditions as in (a).
The second case, termed fast or diffusion-limited aggregation,
refers to the situation where each collision results in the two
colloidal particles involved adhering to each other.31 It has been
found that the growth kinetics of diffusion-limited aggregation
are described by eq 2
proton resonance of II is shifted downfield to δ 7.65. After
addition of a total of 0.75 molar equivalents of IV, this resonance
is shifted downfield to δ 8.07. Following addition of 0.15 of
an equivalent of IV to Ag-(I+II), the resonance assigned to
the imidic proton resonance of IV is shifted downfield to δ
8.69. Each subsequent addition of 0.15 of an equivalent of IV
is accompanied by a small upfield shift. After 0.75 equivalents
of IV are added this resonance is shifted upfield to δ 8.64. The
resonance assigned amidic proton of IV is observed at δ 9.35
throughout, implying that it is not involved in complexation.
Also shown in Figure 4 are the FTIR spectra of Ag-(I+II)
(1.88 × 10-5 mol dm-3 silver nanocrystals, 1.40 × 10-3 mol
dm-3 II) in chloroform-d at 25 °C prior to and following
successive additions of 0.15 molar equivalents of IV (with
respect to II). Also shown is the spectrum of IV (0.21 × 10-3
mol dm-3, 0.15 molar equivalent) in chloroform-d at 25 °C.
Rh(t) t1/d
(2)
f
where Rh(t) is the hydrodynamic radius of the aggregate at some
time t, determined by dynamic light scattering, and df is the
fractal dimension which has a value of 1.75.
The associated physical picture can be summarized as follows.
In the reaction-limited regime, the particles have a low prob-
ability of adhering upon contact, sample many possible con-
figurations, and form aggregates which are relatively dense,
possibly locally ordered, and have a fractal dimension of 2.05.
In the diffusion-limited regime, the particles have a high

Silver Nanocrystal Aggregates in Solution
J. Phys. Chem. B, Vol. 104, No. 26, 2000 6171
Figure 5. (a) Hydrodynamic radius of Ag-(I+II) (3.29 × 10-8 mol dm-3 nanocrystals, 2.43 × 10-6 mol dm-3 II) plotted against elapsed time in
chloroform at 25 °C to which has been added 0.5 (1.21 × 10-6 mol dm-3) and 1.0 (2.43 × 10-6 mol dm-3) molar equivalents of III. (b) Semilog
plot of data in (a). (c) Hydrodynamic radius of Ag-(I+II) (3.29 × 10-8 mol dm-3 nanocrystals, 3.58 × 10-5 mol dm-3 II) plotted against elapsed
time in chloroform at 25 °C to which has been added 0.5 (1.79 × 10-5 mol dm-3) molar equivalents of III. (d) Log-log plot of data in (c).
probability of adhering upon contact, sample few possible
configurations, and form relatively diffuse aggregates, probably
locally branch-like, and which have fractal dimension of 1.75.
The above summary description of particle aggregation
assumes that nanocrystal motion is diffusive and that particle-
particle aggregation is homogeneous. Furthermore, it assumes
that particle aggregation is irreversible with little subsequent
reorganization of the aggregate structure. In practice, there are
cases in which two particles have a high probability of adhering
upon contact and in which the kinetics of aggregate growth are
diffusion limited, but in which the strength of the interaction
between aggregated particles is sufficiently small that significant
restructuring leads to a larger fractal dimension.32
The previous sections of this paper have described the
preparation and characterization of size-monodisperse nano-
crystals stabilized by chemisorption of alkane thiols and alkane
thiols incorporating a receptor site (diaminopyridine), namely
Ag-(I+II) or Ag-II. The present paper has also established that
these receptor modified nanocrystals recognize and selectively
bind molecules incorporating a substrate site (uracil), namely
IV. As a consequence, it should prove possible to induce
aggregation of either Ag-(I+II) or Ag-II by addition of 0.5
molar equivalents (with respect to II) of III, corresponding to
1.0 molar equivalent of substrate sites, to these dispersions. It
was expected that, in the case of Ag-(I+II), with an 8%
coverage of receptor sites, aggregation would be reaction-
limited, but that in the case of Ag-II, with a 100% coverage of
receptor sites, that aggregation would be diffusion-limited. To
test these expectations the aggregation kinetics of Ag-(I+II)
and Ag-II dispersions containing added III have been studied
by DLS.
Shown in Figure 5 are the hydrodynamic radii measured by
DLS in chloroform-d at 25 °C for two Ag-(I+II) dispersions
(3.29 × 10-8 mol dm-3 concentration of silver nanocrystals,
2.43 × 10-6 mol dm-3 concentration of adsorbed receptor sites
in II) to which have been added 0.50 (1.21 × 10-6 mol dm-3
)
or 1.00 (2.42 × 10-6 mol dm-3) molar equivalents of III with
respect to II. As there are two substrate sites incorporated in
III, this corresponds to 1.00 and 2.00 substrate site equivalents,
respectively.
Having added 0.50 of a molar equivalent of III with respect
to II, or 1.00 substrate equivalents with respect to the receptor,
to a dispersion of Ag-(I+II) it was expected that a fraction of
III, containing two substrate sites, would bind receptor sites
on different Ag-(I+II) nanocrystals and that this would lead to
nanocrystal aggregation. This expectation is seen to be well
founded as the average hydrodynamic radius increased by more
than 60 nm during 2 h.
Having added 1.00 molar equivalent of III with respect to
II, or 2.00 substrate equivalents with respect to the receptor, to

6172 J. Phys. Chem. B, Vol. 104, No. 26, 2000
Fullam et al.
a dispersion of Ag-(I+II), it was expected that each molecule
of III, containing two substrate sites, would in the majority of
cases bind only a single receptor site on a Ag-(I+II) nanocrystal
and that this would inhibit nanocrystal aggregation. This
expectation is also seen to be well founded as the average
hydrodynamic radius increased by less than 10 nm during 2 h.
It is clear that the kinetics of aggregation of Ag-(I+II)
induced by addition of 0.5 molar equivalents of III are slow
and described by an expression of the general formula given in
eq 1. Specifically, a semilog plot of the data in Figure 5a yields
a straight line, see Figure 5b. On this basis it is concluded that
aggregation is reaction-limited.29,30 This finding is consistent
with a low probability of two receptor-modified silver nano-
crystals aggregating upon collision and suggests that the nano-
crystal aggregate formed most likely has a compact and
relatively ordered structure. It is equally clear that the kinetics
of aggregation of Ag-(I+II) induced by addition of 1.0 molar
equivalents of III are still slower. On this basis it is concluded
that aggregation is largely inhibited. This finding is consistent
with all of the receptor sites being occupied by a molecule of
III, as a consequence of which and as expected, aggregation is
inhibited.
long-chain alkane thiol incorporating a diaminopyridine receptor
site, Ag-II, have also been prepared. When dispersed in
chloroform these “programmed” nanocrystals recognize and
selectively bind a long-chain alkane incorporating two comple-
mentary substrate sites, III, and are noncovalently linked. The
nanocrystal aggregates formed as a result have been character-
ized by NMR, FT-IR, and DLS.
The key findings are that aggregation of Ag-(I+II) in the
presence of added III is reaction limited, while aggregation of
Ag-II in the presence of added III exhibits diffusion-limited
kinetics. These findings demonstrate, for the first time, that the
number of receptor sites present on the surface of a dispersion
of nanocrystal can be used to control their aggregation kinetics.
Angle-dependent static light scattering studies are in progress
to better understand the relationship aggregation kinetics and
the structures of the nanocrystal aggregates formed.
An implication of these findings is that to achieve the desired
level of control over the architecture of structures assembled
from nanocrystals in solution it will be necessary to be able do
the following: first, determine which nanocrystals recognize
and selectively bind to each other and, second, determine the
strength of the forces acting between nanocrystals. These forces
should be such that reaction limited aggregation is followed by
restructuring to yield a nanocrystal aggregate with the desired
architectural properties.
A general insight based on the findings reported here is that
the surface modification can be used to determine the aggrega-
tion kinetics and, as a consequence, the structure of aggregates
formed by colloidal particles. It is expected that this insight
will inform the work of scientists and technologists in diverse
fields, ranging from medical diagnostics to electronics.
Also shown in Figure 5 are the hydrodynamic radii measured
by DLS in chloroform-d at 25 °C for an Ag-II dispersion (3.29
× 10-8 mol dm-3 concentration of silver nanocrystals, 3.58 ×
10-5 mol dm-3 concentration of adsorbed receptor sites in II)
to which have been added 0.50 (1.79 × 10-5 mol dm-3) molar
equivalents of III with respect to II. As there are two substrate
sites incorporated in III, this corresponds to 1.00 of an
equivalent of receptor sites.
Having added 0.50 molar equivalents of III with respect to
II, or 1.00 substrate equivalents with respect to the receptor, to
a dispersion of Ag-II, it was expected that the each molecule
of III, containing two substrate sites, would bind receptor sites
on different Ag-II nanocrystals and that this would lead to
nanocrystal aggregation. This expectation is seen to be well
founded as the average hydrodynamic radius, measured by
dynamic light scattering, increased to 100 nm during 2 h.
It is clear that the kinetics of aggregation of Ag-II induced
by addition of 0.5 molar equivalents of III are fast and described
by an expression of the general formula given in eq 1.
Specifically, a log-log plot of the data in Figure 5c yields a
straight line of slope 0.68, see Figure 5d. On this basis it is
concluded that aggregation is diffusion limited.29,31 This finding
is consistent with a high probability of two receptor-modified
silver nanocrystals aggregating upon collision and suggests that
the nanocrystal aggregate formed most likely has an open and
disordered structure.
It is noted that the reciprocal of the slope of the log-log
plot referred to above yields a value for the fractal dimension
of the aggregate formed.31 The value obtained (1.45) is
significantly less than that expected (1.75). This finding is
consistent with an earlier study of the salt-induced aggregation
of polystyrene particles.32 This study reported that accurate
values for the fractal dimension of the aggregate were obtained
only from variable-angle static light scattering studies. The
possibility that the initially formed aggregates were restructuring
was considered.
Acknowledgment. The authors thank the Petroleum Re-
search Fund of the American Chemical Society for funding the
reported study (ACS-PRF# 32897-ACS).
References and Notes
(1) Alivisatos, A. P. Science 1996, 271, 933 and references therein.
(2) Murray, C. B.; Norris, D. J.; Bawendi, M. G. J. Am. Chem. Soc.
1993, 115, 8706.
(3) Andres, R. P.; Bielefeld, J. D.; Henderson, J. I.; Janes, D. B.;
Kolagunta, V. R.; Kubiak, C. P.; Mahoney, W. J.; Osifchin, R. G. Science
1996, 273, 1690.
(4) Vossmeyer, T.; Katsikas, L.; Giersig, M.; Popovic, I. G.; Diesner,
K.; Chemseddine, A.; Eychmueller, A.; Weller, H. J. Phys. Chem. 1994,
98, 7665.
(5) Harfenist, S. A.; Wang, Z. L.; Alvarez, M. A.; Vezmar, I.; Whetten,
R. L. J. Phys. Chem. 1996, 100, 13904.
(6) Herron, N.; Calabrese, J. C.; Farneth, W. E.; Wang, Y. Science
1993, 259, 1426.
(7) Schmidt, G. Chem. ReV. 1992, 92, 1709.
(8) Moritz, T.; Reiss, K.; Diesner, K.; Su, D.; Chemeseddine J. Phys.
Chem. B 1997, 101, 8052.
(9) Brus, L. J. Phys. Chem. 1986, 90, 2555.
(10) Murray, C. B.; Kagan, C. R.; Bawandi, M. G. Science 1995, 270,
1335.
(11) Kagan, C. R.; Murray, C. B.; Bawendi, M. G. Phys. ReV. B 1996,
54, 8633.
(12) Heath, J. R.; Knobler, C. M.; Leff, D. V. J. Phys. Chem. B 1997,
101, 189.
(13) Motte, L.; Billoudet, F.; Lacaze, E.; Douin, J.; Pileni, M. P. J. Phys.
Chem. B. 1997, 101, 138.
(14) Elghanain, R.; Storhoff, J.; Micic, R.; Letsinger, R.; Mirkin, C.
Science 1997, 277, 1078.
(15) (a) Cusack, L.; Rao, S. N.; Wenger, J.; Fitzmaurice, D. Chem.
Mater. 1997, 9, 624. (b) Cusack, L.; Rao, S. N.; Fitzmaurice, D. Chem.
Eur. J. 1997, 3, 202. (c) Cusack, L.; Marguerettaz, X.; Rao, S. N.; Wenger,
J.; Fitzmaurice, D. Chem. Mater. 1997, 9, 1765.
(16) Cusack, L.; Rizza, R.; Gorelov, A.; Fitzmaurice, D. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 848.
Conclusions
Silver nanocrystals stabilized by a chemisorbed long-chain
alkane thiol and a chemisorbed long-chain alkane thiol incor-
porating a diaminopyridine receptor site, Ag-(I+II), have been
prepared. Silver nanocrystals stabilized only by a chemisorbed
(17) Ahern, D.; Rao, S. N.; Fitzmaurice, D. J. Phys. Chem. B 1999,
103, 1821.

Silver Nanocrystal Aggregates in Solution
J. Phys. Chem. B, Vol. 104, No. 26, 2000 6173
(18) Brust, M.; Walker, M.; Bethell, D.; Schriffrin, D.; Whyman, R. J.
Chem. Soc., Chem. Commun. 1994, 801.
(19) Korgel, B.; Fullam, S.; Connolly, S.; Fitzmaurice, D. J. Phys. Chem.
1998, 102, 8379.
on the percentage composition by weight of N, the number of molecules
of II per nanocrystal of silver was calculated and found to be 1088.
(24) (a) Brienne, M.-J.; Gabard, J.; Lehn, J.-M. J. Chem. Soc., Chem.
Commun. 1989, 1868. (b) Cusack, L.; Rao, S. N.; Wenger, J.; Fitzmaurice,
D. Chem. Mater. 1997, 9, 624. (c) Cusack, L.; Rao, S. N.; Fitzmaurice, D.
Chem. Eur. J. 1997, 3, 202. (d) Cusack, L.; Marguerettaz, X.; Rao, S. N.;
Wenger, J.; Fitzmaurice, D. Chem. Mater. 1997, 9, 1765.
(25) In this context we note that the chemical shifts observed initially
might be due to changes in the acidity of the deuterated solvents used.
(26) Bellamy, L. J. The Infrared Spectra of Complex Molecules;
Matheun: London, 1954.
(27) (a) Feibush, B.; Fiueroa, A.; Charles, R.; Onan, K.; Feibush, P.;
Karger, B. J. Am. Chem. Soc. 1986, 108, 3310. (b) B. Feibush, M. Saha,
Onan, K.; Karger, B.; Giese, R. J. Am. Chem. Soc. 1987, 109, 7531. (c)
Hamilton, A.; Van Engen, D. J. Am. Chem. Soc. 1987, 109, 5035. (d) Bisson,
A.; Carver, F.; Hunter, C.; Waltho, J. J. Am. Chem. Soc. 1994, 116, 10292.
(28) Pimentel, G.; McClellan, A. The Hydrogen Bond; Freeman: San
Francisco, 1956.
(29) Family, F.; Landau, D., Eds. Kinetics of Aggregation and Gelation;
North-Holland: Amsterdam, 1984.
(30) Weitz, D.; Huang, J.; Lin, M. Sung, J. Phys. ReV. Lett. 1985, 54,
1416.
(31) (a) Weitz, D.; Olivera, M. Phys. ReV. Lett. 1984, 52, 1433. (b)
Weitz, D.; Huang, J.; Lin, M.; Sung, J. Phys. ReV. Lett. 1984, 53, 1657.
(32) Zhou, Z.; Chu, B. J. Colloid Sci. 1991, 143, 356. (b) Zhou, Z.;
Wu. P.; Chu, B. J. Colloid Sci. 1991, 146, 541.
(20) To obtain these values, the diameter of 300 silver nanocrystals
modified by chemisorbed I and II were determined from a TEM image
prepared by evaporation of a chloroformic dispersion of these nanocrystals
on a carbon-coated copper grid. It is noted that these nanocrystals show
long range ordering, an observation that strongly supports the assertion that
these nanocrystals are relatively monodisperse.
(21) To obtain these values, a known weight of a sample silver
nanocrystals modified by chemisorbed I and II was analyzed for percentage
composition by weight of the following elements: C, H, N, and S. Based
on the percentage composition by weight of N and S, the numbers of
molecules of I and II per nanocrystal of silver were calculated to be 883
and 74, respectively.
(22) To obtain these values, the diameter of 170 silver nanocrystals
modified by chemisorbed II were determined from a TEM image prepared
by evaporation of a chloroformic dispersion of these nanocrystals on a
carbon coated copper grid. It is noted that these nanocrystals clearly show
long range ordering, an observation that strongly supports the assertion that
these nanocrystals are relatively monodisperse.
(23) To obtain these values, a known weight of a sample silver
nanocrystals modified by chemisorbed II was analyzed for percentage
composition by weight of the following elements: C, H, N, and S. Based