Thermal/Photo Conversion of Silver 4-Nitrobenzoate to Nitro/Amine- Terminated Silver Nanoparticles

Article abstract of DOI:10.1021/jp0364250

We have investigated the thermal and photochemical characteristics of silver 4-nitrobenzoate (Ag-4NBA) using various analytical tools, such as, X-ray diffraction (XRD), diffuse reflectance infrared Fourier transform spectroscopy, Raman spectroscopy, thermogravimetric analysis, scanning electron microscopy, and transmission electron microscopy. The XRD pattern was composed of a series of peaks that could be indexed to (0k0) reflections of a layered structure. Upon heating the sample, structural changes took place particularly at two specific temperatures. The binding state of the carboxylate group changed from bridging to unidentate at ～380 K. A second dramatic structural change occurred at ～590 K, producing 4NBA-capped Ag nanoparticles. The latter transition temperature for Ag-4NBA is about 80 K higher than that for silver alkanecarboxylates, and such enhanced thermal stability of Ag-4NBA must be associated with the conjugation of the carboxylate group to the aromatic ring. On the other hand, Ag-4NBA was readily subjected to photolysis even by visible radiation. Under irradiation by a He/Ne laser at 632.8 nm, 4-NBA-capped Ag nanoparticles were produced from Ag-4NBA. Upon irradiation by an Ar⁺ laser at 514.5 nm, nitro-to-amine group conversion took place consecutively, finally producing 4-aminobenzoate-capped Ag nanoparticles. Accordingly, the Raman spectrum of Ag-4NBA taken using the 514.5-nm line as the excitation source became identical to the surface-enhanced Raman scattering (SERS) spectrum of 4-NBA adsorbed on Ag.

Full text of DOI:10.1021/jp0364250

4000
J. Phys. Chem. B 2004, 108, 4000-4007
Thermal/Photo Conversion of Silver 4-Nitrobenzoate to Nitro/Amine-Terminated Silver
Nanoparticles
Yeoung Uk Seo, Seung Joon Lee, and Kwan Kim*
Laboratory of Intelligent Interfaces, School of Chemistry and Molecular Engineering,
Seoul National UniVersity, Seoul 151-742, Korea
ReceiVed: August 14, 2003; In Final Form: February 4, 2004
We have investigated the thermal and photochemical characteristics of silver 4-nitrobenzoate (Ag-4NBA)
using various analytical tools, such as, X-ray diffraction (XRD), diffuse reflectance infrared Fourier transform
spectroscopy, Raman spectroscopy, thermogravimetric analysis, scanning electron microscopy, and transmission
electron microscopy. The XRD pattern was composed of a series of peaks that could be indexed to (0k0)
reflections of a layered structure. Upon heating the sample, structural changes took place particularly at two
specific temperatures. The binding state of the carboxylate group changed from bridging to unidentate at
∼380 K. A second dramatic structural change occurred at ∼590 K, producing 4NBA-capped Ag nanoparticles.
The latter transition temperature for Ag-4NBA is about 80 K higher than that for silver alkanecarboxylates,
and such enhanced thermal stability of Ag-4NBA must be associated with the conjugation of the carboxylate
group to the aromatic ring. On the other hand, Ag-4NBA was readily subjected to photolysis even by visible
radiation. Under irradiation by a He/Ne laser at 632.8 nm, 4-NBA-capped Ag nanoparticles were produced
from Ag-4NBA. Upon irradiation by an Ar+ laser at 514.5 nm, nitro-to-amine group conversion took place
consecutively, finally producing 4-aminobenzoate-capped Ag nanoparticles. Accordingly, the Raman spectrum
of Ag-4NBA taken using the 514.5-nm line as the excitation source became identical to the surface-enhanced
Raman scattering (SERS) spectrum of 4-NBA adsorbed on Ag.
1. Introduction
reflections are associated with the three-dimensionally stacked
silver carboxylate layers.4 As a result of the specific arrangement
of Ag atoms in a layer, AgCO2R materials have been proposed
to possess 2-D conductivity in the solid state.5 AgCO2R
materials are also known to decompose (in an atmosphere of
N2) into alkanecarboxylate-derivatized silver nanoparticles.4d,4f
Perfluorocarboxylate-stabilized silver nanoparticles have simi-
larly been prepared via the thermal decomposition of layered
silver perfluorocarboxylates.6 Because the organic moieties
anchored on the Ag nanoparticles can be completely removed
by heating to 750 K or higher, optically tunable, SERS-active
Ag films can be reproducibly fabricated on glass microscope
slides.7
In conjunction with the above implications, we report herein
the structure and thermal behavior as well as the photoresponse
of silver 4-nitrobenzoate (Ag-4NBA). The first concern in this
work is whether a layered structure is also assumed for silver
aromatic carboxylate. The second concern is to determine, if it
consists of a layered structure, how much the thermal charac-
teristics of such a hybrid material are affected by the incorpora-
tion of aromatic rings instead of hydrocarbon chains. If organic
molecule-derivatized Ag nanoparticles are also produced by the
thermal decomposition of Ag-4NBA, the third concern is to
determine whether the organic moiety is 4-nitrobenzoate; the
conversion of silver compounds to silver nanoparticles must
occur via a redox process; it must then be of interest to find
out whether the nitro group is also converted to an amine, along
with the silver species. The fourth concern is whether photo-
reaction occurs even for Ag-4NBA. In fact, we found that Ag-
4NBA consisted of a layered structure, and 4NBA-derivatized
Ag nanoparticles were produced by the thermal treatment of
Ag-4NBA similarly as for silver alkanecarboxylates. Surpris-
Several surface-enhanced Raman scattering (SERS) studies
regarding the physicochemical properties of aromatic nitro
molecules on silver have been reported in the literature.1 Upon
irradiation by an Ar+ laser, the SERS peaks of the original nitro
molecules usually decline in intensity, and a new set of peaks
appears. The new peaks rapidly increase in intensity as a
function of the laser illumination time, suggesting that the nitro
molecules are subjected to photoreaction on the silver surface.
On the basis of these observations, we have recently reexamined
the SERS characteristics of 4-nitrobenzenethiol (4-NBT)2a,2b and
4-nitrobenzoic acid (4-NBA)2c on silver, and their SERS spectral
features were found to be surprisingly coincident with those of
4-aminobenzenethiol (4-ABT) and 4-aminobenzoic acid (4-
ABA) on Ag, respectively. Much the same conclusion could
be derived from X-ray photoelectron spectroscopy measurements
as well as from a coupling reaction forming amide bonds.2a The
surface-induced photoreaction thereby allowed us to prepare
patterned binary monolayers on Ag that would show different
chemical reactivities. Using the binary monolayers as a litho-
graphic template, we could also conduct site-specific chemical
reactions.2a
In the past decade, the number of organic-inorganic hybrid
materials that exhibit alternating 2-D molecular assemblies of
organic and inorganic constituents have expanded considerably,
encompassing a host of technologically relevant materials.3
Silver alkanecarboxylates (AgCO2R) are one class of organic-
inorganic heterostructured materials that show a well-developed
progression of intense X-ray reflections; these intense X-ray
* Author to whom all correspondence should be addressed. Phone: +82-
2-8806651. Fax: +82-2-8743704. E-mail: kwankim@plaza.snu.ac.kr.
10.1021/jp0364250 CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/05/2004

Silver 4-Nitrobenzoate Conversion to Nanoparticles
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4001
ingly, however, 4-aminobenzoate (4ABA)-derivatized Ag nano-
particles were produced upon irradiation by an Ar+ laser (at
514.5 nm) on Ag-4NBA, whereas 4NBA-derivatized Ag nano-
particles were produced by irradiation with a He/Ne laser at
632.8 nm.
2. Experimental Section
2.1. Chemicals and Synthesis. Silver nitrate (99+%),
4-nitrobenzoic acid (4-NBA, 99+%), 4-aminobenzoic acid (4-
ABA, 99%), and triethylamine (99+%) were purchased from
Aldrich and used as received. Unless specified, other chemicals
were reagent grade, and triply distilled water (resistivity greater
than 18 MΩ‚cm) was used throughout. To prepare silver
4-nitrobenzoate (Ag-4NBA), a methanolic solution of AgNO3
(0.1 M, 100 mL) was added dropwise to an equimolar 1:1
mixture of 4-NBA and triethylamine in methanol (100 mL).
After 2 h of vigorous stirring, the resulting white solid was
filtered, washed thoroughly with water and methanol, and finally
dried in a vacuum; hereafter it will be called Ag-4NBA. The
results of the elemental analysis, performed at the Korea Basic
Science Institute using a CE instruments/ThermoQuest Model
Flash EA1112 elemental analyzer, are as follows. Anal. Calcd
for Ag-4NBA (%): C, 30.68; H, 1.47; O, 23.36; N, 5.11.
Found: C, 30.94; H, 1.56; O, 24.05; N, 5.58. In order to help
interpret the IR and Raman spectral peaks of Ag-4NBA, diffuse
reflectance infrared Fourier transform (DRIFT) and SERS
spectra were obtained for 4-NBA and 4-ABA self-assembled
on powdered Ag (Aldrich, 99%, 2-3.5 µm) in 10 mM
methanolic solution for 3 h.
Figure 1. (a) SEM image of silver 4-nitrobenzoate. (b) XRD pattern
of silver 4-nitrobenzoate. The interlayer spacings derived from different
reflections are listed in the inset.
2.2. Characterization. XRD patterns were obtained on a
Rigaku Model D/Max-3C powder diffractometer for a 2θ range
of 5° to 80° at an angular resolution of 0.02° using Cu KR
(1.5419 Å) radiation. The samples were spread on antireflection
glass slides to give uniform films. A scanning electron
microscope (SEM) image of Ag-4NBA was measured using a
JSM 840-A scanning electron microscope at 20 kV. Transmis-
sion electron microscope (TEM) images were acquired for Ag-
4NBA after thermal and radiation treatment on a JEM-2000EXII
transmission electron microscope at 200 kV after placing a drop
of chloroform/methanol (9:1) mixed solution on carbon-coated
copper grids (200 mesh). Thermogravimetric analysis (TGA)
data were obtained with a TA Instrument 2050 thermogravi-
metric analyzer in a nitrogen atmosphere at a heating rate of
10 K/min.
Infrared spectra were measured using a Bruker IFS 113v FT-
IR spectrometer equipped with a globar light source and a liquid
N2-cooled wide-band mercury cadmium telluride detector. To
record the DRIFT spectra, a diffuse reflection attachment
(Harrick Model DRA-2CO) designed to use 6:1, 90° off-axis
ellipsoidal mirrors subtending 20% of the 4π solid angle was
fitted to the sampling compartment of the FT-IR spectrometer.
The pure powdered sample was transferred to a 4-mm diameter
cup without compression and leveled by a gentle tap. A reaction
chamber, made of stainless steel (Harrick Model HVC-DR2)
and loaded with the powdered sample, was located inside the
reflection attachment. CaF2 crystals were used as infrared
transparent windows. The temperature of the sampling cup was
regulated using a homemade temperature controller, and the
chamber was flushed continuously with dry nitrogen (ca. 10
mL/min) during the measurement of DRIFT spectra. A total of
32 scans was measured in the range of 3500-1000 cm-1 at a
resolution of 4 cm-1 using previously scanned pure KBr as the
background. The temperature of the sampling cup was raised
at a rate of 10 K/min and kept for 5 min at each specified
temperature for the acquisition of the DRIFT spectra. The
Happ-Genzel apodization function was used in Fourier trans-
forming all the interferograms. The DRIFT spectra are reported
as -log(R/Ro), where R and Ro are the reflectance of the sample
and of the pure KBr, respectively.
Raman spectra were obtained using a Renishaw Raman
system model 2000 spectrometer equipped with an integral
microscope (Olympus BH2-UMA). The 514.5-nm radiation
from a 17 mW air-cooled Ar+ laser (Spectra Physics model
163-C4210) or the 632.8-nm radiation from a 17 mW air-cooled
He/Ne laser (Spectra Physics Model 127) was used as the
excitation source. Raman scattering was detected with 180°
geometry using a peltier-cooled (-70 °C) CCD camera (400 ×
600 pixels). A glass capillary (KIMAX-51) with an outer
diameter of 1.5-1.8 mm was used as a sampling device, but
when it was needed to reduce photolysis, samples made into
pellets were spun at 3000 rpm during the Raman spectral
measurement. The data acquisition time was usually 90 s. The
holographic grating (1800 grooves/mm) and the slit permitted
a spectral resolution of 1 cm-1. The Raman band of a silicon
wafer at 520 cm-1 was used to calibrate the spectrometer, and
the accuracy of the spectral measurement was estimated to be
better than 1 cm-1. The Raman spectrometer was interfaced to
an IBM-compatible PC, and the spectral data were analyzed
using Renishaw WiRE software, version 1.2, based on the
GRAMS/32C suite program (Galactic Industries).
3. Results and Discussion
3.1. Structure of Silver 4-Nitrobenzoate. The SEM image
in Figure 1a reveals that Ag-4NBA has a fibrous morphology.
These fibrous samples were identified as possessing a layered
structure from the XRD data shown in Figure 1b. The XRD
pattern shows a well-developed progression of intense reflec-
tions, and these intense reflections can be interpreted in terms

4002 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Seo et al.
Figure 3. DRIFT spectra of silver 4-nitrobenzoate in the temperature
range of 293-623 K. The arrows denote the occurrence of distinct
spectral changes.
Figure 2. Schematic depiction of the layered silver 4-nitrobenzoate.
binding state of the carboxylate group to silver in Ag-4NBA is
presumed to be a bridging one.
of three-dimensionally stacked silver carboxylate layers with a
large interlayer lattice dimension.4 This implies that, as in silver
alkanecarboxylates, each layer of Ag-4NBA is separated from
the neighboring layer by twice the length of the nitrobenzene
moiety. In this sense, all intense reflections can be indexed as
(0k0), and the reflections are assigned in Figure 1b. In the inset
of Figure 1b, we list the interlayer spacings derived from
different reflections. The averaged interlayer spacing is 19.41
( 0.61 Å. The layered structure of Ag-4NBA is schematically
drawn in Figure 2.
The comparable XRD patterns of Ag-4NBA and silver
alkanecarboxylate suggest that the arrangement of silver and
carboxylate must be similar to each other in the two compounds.
In this light, it is informative to recall the Ag K-EXAFS study
of Tolochko et al.,4c which showed that the silver atoms in silver
stearate (Ag-STA) are bridged by the carboxylate in the form
of dimers in an eight-membered ring and that the dimers are
further bonded to each other by longer Ag-O bonds forming
four-membered rings; the reported bond lengths are r(Ag-Ag)
) 2.90 Å, r(Ag-O) ) 2.25 Å, r(C-O) ) 1.20 Å, and r(-
OOC-CH2-) ) 1.54 Å. On the other hand, according to the
photographic X-ray study conducted by Blakeslee and Hoard,8
the bond lengths in silver perfluorobutyrate are r(Ag-Ag) )
2.90 Å, r(Ag-O) ) 2.24 Å, r(C-O) ) 1.25 Å, and r(-OOC-
CF2-) ) 1.54 Å. When these values are used, the thickness of
the Ag-COO slab, that is, 2t1 in Figure 2, is estimated to be
4.83 Å. Considering that the averaged interlayer spacing derived
from Figure 1b was 19.41 Å, the thickness of the nitrobenzene
layer, that is, 2t2 in Figure 2, should then be 14.58 Å.
3.2. Thermal Behavior of Silver 4-Nitrobenzoate. One aim
of this work was to determine how much the thermal charac-
teristics of layered AgCO2R would be affected by the incorpora-
tion of an aromatic ring instead of a simple alkane chain. In
this regard, we have conducted a temperature-dependent DRIFT
spectroscopy study along with TGA measurements.
3.2.1. Temperature-Dependent DRIFT Spectral Pattern of Ag-
4NBA. Figure 3 shows a series of DRIFT spectra obtained as a
function of temperature for Ag-4NBA. The peak positions of
the carboxylate stretching bands provide information on the
effect of temperature on the headgroup structure. In this light,
it is informative that the νas(COO-) peak at 1575 cm-1 is rather
invariant up to 383 K and then becomes weak and disappears
above 443 K. Instead, a new peak develops at 1591 cm-1 from
383 K and is sustained up to 593 K. On the other hand, the
νs(COO-) band decreases gradually from room temperature up
to 623 K and the ν(CdO) band develops around 513 K. The
present observation suggests that two distinct structural changes
take place, one at 383 K and the other at 513 K, in connection
with the Ag-O binding state. Assigning the newly developing
peak at 1591 cm-1 at 383 K to νas(COO-), the binding state of
the carboxylate group must change from bridging to unidentate
at 383 K. On the other hand, free acid seems to be produced at
513 K probably from the thermal decomposition of Ag-4NBA;
the first mass loss commences at ∼520 K in the TGA
measurement (vide infra). It is intriguing that the ν(CdO) band
due to free acid is identified at a somewhat lower temperature
for silver alkanecarboxylates, that is, at 500 K for Ag-STA4d
and at 480 K for silver hydroxyhexadecanoate.4f This implies
that the Ag-O bonds in Ag-NBA are stronger than those in
silver alkanecarboxylates, presumably because of the conjuga-
tion of the carboxylate group with the aromatic benzene ring.
On the other hand, considering the melting point of pure 4-NBA
(512-514 K), the observed spectral change at 513 K must also
be associated with the disordering of nitrobenzene moieties.
The NO2 group related peaks also exhibit temperature-
dependent spectral changes. Namely, both the νs(NO2) and νas-
(NO2) bands in Figure 3 gradually lose intensity from room
temperature up to 623 K along with experiencing moderate band
Consulting the crystal structure of 4-NBA9 and the van der
Waals (vdW) radius of the oxygen atom,10 the molecular length
from the carboxylate carbon to the terminal NO2 group should
be 7.73 Å. Doubling this value yields 15.46 Å, which is 0.88 Å
longer than that derived from the measured XRD data, that is,
14.58 Å. The discrepancy of 0.88 Å is presumed to be due to
the tilting of 4-NBA (by ∼19°) with respect to the Ag plane
and/or the interpenetration of NO2 groups between the adjacent
layers in Ag-4NBA. We cannot derive the intralayer structure
because of the insufficient signal-to-noise ratio for the high-
angle XRD peaks, but the overall structure of Ag-4NBA as
drawn in Figure 2 is presumably similar to that of silver
alkanecarboxylate.
Room temperature DRIFT spectral features indicate that the
as-prepared Ag-4NBA is not contaminated with free acid (see
the Supporting Information). In addition, referring to the
empirical relationship between the frequency difference of the
carboxylate stretching bands, that is, ∆ν (∼190 cm-1) )
νas(COO-) - νs(COO-)11,12 and the types of bonding, the
broadening. Concomitantly, the ring vibration at 1615 cm-1
,
that is, 8a, gradually weakens up to 443 K. A new peak,
however, develops at 1630 cm-1 from 363 K and is sustained
up to 543 K. It then shifts to lower frequency above 553 K and
becomes weak until it disappears above 593 K. The origin of
the new peak at 1630 cm-1 is currently unclear but its position
is close to that of the 8a mode; the position of the 8a mode
may be dependent on the type of bonding of the carboxylate

Silver 4-Nitrobenzoate Conversion to Nanoparticles
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4003
Figure 4. (a) TGA and (b) its first derivative traces of silver
Figure 5. (A) TEM image of silver nanoparticles obtained from the
thermal decomposition of silver 4-nitrobenzoate. (B) DRIFT spectrum
of silver nanoparticles obtained from the thermal decomposition of silver
4-nitrobenzoate (a) and that of 4-nitrobenzoic acid self-assembled on
2 µm-sized silver powder (b).
4-nitrobenzoate (full lines); the dotted lines are for silver stearate.4d
group to silver in Ag-4NBA. All these spectral features
nonetheless support the contention that the 4-NBA moiety is
sustained at least up to 593 K without decomposition.
3.2.3. Identity of the Thermal Decomposition Product of Ag-
4NBA. Recently, we have reported that alkanecarboxylate-
derivatized silver nanoparticles with a uniform size distribution
can be obtained via the thermal decomposition of silver
alkanecarboxylates under N2 atmosphere.4d,4f,6 In this light,
consulting the DRIFT spectroscopy and TGA data, we have
analyzed a sample of Ag-4NBA preheated to 600 K for 5 min
in N2 atmosphere to verify that 4NBA-capped silver nanopar-
ticles are indeed produced by the thermal decomposition of Ag-
4NBA. The heated sample was rinsed thoroughly with ethanol
to remove free acids or possible impurities. The remaining solid
was soluble in a mixed solvent of chloroform and methanol.
Figure 5A shows a typical TEM image of the sample. The image
reveals that silver nanoparticles are indeed formed by the thermal
decomposition of Ag-4NBA. The sizes of the nanoparticles are
quite uniform with an average diameter of 6.0 ( 1.4 nm. The
formation of silver nanoparticles can also be confirmed from
the XRD pattern (see the Supporting Information); all the XRD
peaks can be attributed to the reflections of face-centered cubic
Ag. We can confirm further from DRIFT spectroscopy that the
silver nanoparticles are passivated with 4-nitrobenzoate. In the
DRIFT spectrum of the silver nanoparticles shown in Figure
5B(a), peaks due to νas(COO-), νas(NO2), νs(COO-), and νs-
(NO2) are clearly identified at 1583, 1516, 1395, and 1350 cm-1,
respectively. The presence of these bands supports the contention
that the silver nanoparticles are derivatized with 4-nitrobenzoate.
One intriguing aspect is that only the νs(COO-) and νs(NO2)
bands are identified in the DRIFT spectrum of 4-NBA on
powdered silver shown in Figure 5B(b), whereas the νas(COO-)
and νas(NO2) bands are also identified in the DRIFT spectrum
of nanoparticles in Figure 5B(a). It has been well established
that the usual IR surface selection rule is applicable even to the
surface of fine metal particles.2c,11,13 The appearance of the
νas(COO-) and νas(NO2) bands in Figure 5B(a) may be
indicative of the binding of 4-nitrobenzoate to Ag nanoparticles
via only one of the two carboxylate oxygen atoms, whereas in
the self-assembled monolayers (SAMs) of 4-NBA on powdered
silver, the carboxylate group is bound to silver symmetrically
via its two oxygen atoms. In fact, Merklin et al.14 observed a
similar spectral feature to that shown in Figure 5B(a) in their
surface-enhanced IR absorption study of 4-NBA on silver island
film, but the two antisymmetric bands, that is, νas(COO-) and
3.2.2. TGA Measurement. Figures 4a and 4b (see full lines)
show the TGA and its first derivative traces of Ag-4NBA,
respectively. For comparison, similar data for Ag-STA are also
shown in the figures as dotted lines. It is evident that no mass
loss occurs up to ∼520 K. This implies that the transition
occurring at 383 K in Figure 3 is associated entirely with the
conformational change in Ag-4NBA. In the temperature region
of 520-775 K, two distinct mass loss features are observed in
Figure 4. Most of the mass loss occurs around 593 K, and then
a small amount is subsequently lost around 623 K. During the
interval, the actual mass loss amounts to 58.5%. The first mass
loss event is presumably associated with the formation of 4NBA-
capped silver nanoparticles caused by the thermal decomposition
of Ag-4NBA (vide infra), whereas the second loss event is due
to the desorption of 4-NBA from the nanoparticles as well as
the decomposition of residual free acid. On the basis of the
formula weight of Ag-4NBA (AgCO2(C6H4)NO2) and the data
from the elemental analysis, if the organic moiety is completely
lost, the mass loss has to amount to 62% in total. It is remarkable
that this value is exactly the same as that observed by the TGA
measurement. This indicates that free silver remains exclusively
above 623 K.
It is intriguing that the first mass loss event occurs for Ag-
4NBA at about 80 K higher temperature than for Ag-STA,
whereas the second loss event occurs at about 50 K lower
temperature (see the dotted lines in Figure 4; similar dotted data
are also obtained for silver hydroxyhexadecanoate4f). Because
the first mass loss event is associated with the formation of Ag
nanoparticles, the higher temperature must be a result of the
stronger Ag-COO bond in Ag-4NBA than Ag-STA. As
mentioned previously, the stronger Ag-COO bond in Ag-4NBA
has to be attributed to the conjugation effect of the carboxylate
group to the aromatic ring as well as to the inter-ring ππ
interaction. Considering the fact that the second mass loss event
is associated with the complete loss of organic moieties, the
lower temperature in Ag-4NBA can be understood by presuming
that the adsorption strength of 4-NBA on highly curved Ag
nanoparticles is far smaller than that of STA. At higher
temperatures, the inter-ring ππ interaction will not be favorable
for rigid aromatic molecules, whereas the vdW interaction is
possible for flexible alkanoic acids.

4004 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Seo et al.
Figure 6. Raman spectra of (a) pure 4-nitrobenzoic acid and (b) silver
4-nitrobenzoate taken using the 514.5-nm line of an Ar+ laser as the
excitation source under spinning at 3000 rpm; the power at the sampling
position was 17 mW, and the spectral acquisition time was 5 s.
νas(NO2), disappeared after thorough rinsing with methanol. In
our work, however, the νas(COO-) and νas(NO2) bands barely
disappeared in the DRIFT spectrum of the nanoparticles even
after thorough rinsing with methanol and ethanol.
As mentioned in the previous section, free silver remains
exclusively above 623 K (XRD data not shown). When such a
thermal decomposition is conducted after coating Ag-4NBA on
inflammable solid substrates, highly SERS-active, nanostruc-
tured Ag film can be fabricated as demonstrated in our previous
publication.7
Figure 7. (A) Raman spectra of silver 4-nitrobenzoate taken at static
conditions under irradiation by the 514.5-nm line (17 mW at the
sampling position) for (a) 5 s and (b) 10, 20, and 30 min. (c) Raman
spectrum of Ag-4ABA and (d) SERS spectrum of 4ABA adsorbed on
2 µm-sized silver powder; both are obtained using the 514.5-nm line.
The arrows in (b) denote the occurrence of distinct spectral changes.
(B) TEM image of silver nanoparticles obtained from the photolysis
of silver 4-nitrobenzoate.
and 867 cm-1, respectively, in agreement with the IR results
(see the Supporting Information). (The fact that the ν(CdO)
band for the free acid appears at very different positions in the
IR and Raman spectra indicates that 4-NBA exists as a
hydrogen-bonded dimer. This form has a center of symmetry
so that when the two monomer compounds vibrate in phase,
the vibration will be Raman-active only, and when they vibrate
out-of-phase, the vibrations will be IR-active only. In the Raman
spectrum, the symmetrical CdO stretch is thus seen near 1650
cm-1, and in the IR spectrum, the antisymmetrical CdO stretch
is seen near 1700 cm-1.18)
The Raman spectral features of Ag-4NBA taken in static
conditions were significantly different from those obtained in
spinning conditions. In Figure 7A(b) are shown the Raman
spectra of Ag-4NBA taken consecutively in static conditions
under exposure by an Ar+ laser at 514.5 nm for 10, 20, and 30
min; the sampling area was ∼1 µm2, and the power at the
sampling position was 17 mW. For comparison, the Raman
spectrum taken in spinning conditions (the same as that in Figure
6b) is reproduced in Figure 7A(a). Dramatic differences are
identified in the spectral region of 1300-1500 cm-1. As
indicated by the two arrows in Figure 7A(b), a peak near 1455
cm-1 grows and becomes stronger while a peak at 1340 cm-1
progressively weakens as the laser exposure time increases.
Because the 1340 cm-1 peak can be assigned to νs(NO2), its
decrease in intensity indicates that a certain photoreaction occurs
at the nitro group of 4-NBA. The identity of the 1455 cm-1
peak is rather uncertain at the moment. Nonetheless, because
no spectral change occurs at all for the free acid, that is, 4-NBA,
even after prolonged exposure to an Ar+ laser, the spectral
change in Figure 7A(b) must have been associated with the Ag
ions in Ag-4NBA.
3.3. Photochemical Behavior of Silver 4-Nitrobenzoate.
Bokhonov et al.15 reported that long-chain silver carboxylates
were photochemically decomposed into silver nanoparticles
under UV irradiation. Yonezawa et al.16 showed that a silver
film composed of silver clusters could be obtained from the
photolysis of silver alginate under UV irradiation. These reports
suggest that silver carboxylates can be transformed to nanosized
silver particles upon irradiation by UV light. It is indeterminate,
however, whether Ag nanoparticles can also be produced by
the irradiation of visible light onto Ag-4NBA. If Ag nanopar-
ticles are produced by visible-light irradiation, the organic
moiety, that is, 4-NBA, may also be subjected to photochemical
reaction. It can be expected from our recent SERS observations
that the nitro-to-amine group conversion takes place for 4-NBT
on nanostructured silver by irradiation with an Ar+ laser at 514.5
nm.2a,2b The photoconversion efficiency was found to be highly
dependent on the extent of surface roughness, that is, atomic-
and nanoscale roughness, which is intimately associated with
the SERS enhancement sites, that is, the chemical and electro-
magnetic enhancement sites, respectively.2b Separately, we found
that the Ag nanoparticles physically in contact with 4-NBT
SAMs on Au could also induce the photolytic reduction of the
4-NBT moiety simply by irradiation with 514.5-nm laser
irradiation in ambient conditions.17 Based on these observations,
we have examined the visible-light response of Ag-4NBA by
Raman spectroscopy.
Figures 6a and 6b show the ordinary Raman (OR) spectra of
pure 4-NBA and Ag-4NBA, respectively, taken using the 514.5-
nm line of an Ar+ laser as the excitation source. During the
measurement of the latter spectrum, a pelletized Ag-4NBA
sample was spun at 3000 rpm in order to minimize any possible
photoreaction. In the acid spectrum (Figure 6a), the ν(CdO)
and ν(C-OH) bands derived from the H-bonded carboxylic
group are clearly identified at 1636 and 1289 cm-1, respectively.
The corresponding bands are completely absent in the Ag-4NBA
spectrum. Instead, in Figure 6b, the νs(COO-) and δ(COO-)
bands derived from the carboxylate group are observed at 1392
It is surprising that the Raman spectrum of Ag-4ABA
obtained also in static conditions and shown in Figure 7A(c) is
almost identical to that of Ag-4NBA in Figure 7A(b). Moreover,
it is very noteworthy that these Raman spectral features resemble
very closely the SERS spectral features of 4-ABA self-

Silver 4-Nitrobenzoate Conversion to Nanoparticles
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4005
resonance of silver. Regarding the latter matter, it has been
observed by Fedurco et al.19 that the surface roughening of Ag
results in a drastic increase of the photocurrent for the
wavelength close to the surface plasmon frequency. It is well
conceivable that the condition of surface plasmon resonance
can be satisfied with the 514.5-nm excitation by the aggregation
of Ag nanoparticles. In particular, as shown in Figure 7B, when
the photolyzed Ag nanoparticles are appropriately aggregated
to show a large electromagnetic (EM) effect, the photoelectron
ejection efficiency will reach such a level as to induce chemical
reduction of the nearby functional group(s).
We have to mention that the SERS spectrum is not observable
from the 4NBA-capped Ag nanoparticles produced via the
thermal decomposition of Ag-4NBA; this is why only the
DRIFT spectrum is shown in Figure 5B(a). This is in sharp
contrast with the photolyzed Ag nanoparticles. This apparent
difference can be understood by recalling the fact that the SERS
effect is derived mostly from the EM enhancement that is in
turn strongly associated with the sizes and shapes of nanopar-
ticles as well as nanoaggregates.20 As seen in Figure 5A, the
thermally produced Ag nanoparticles are quite monodisperse
whereas the photoproduced Ag nanoparticles in Figure 7B are
highly irregular. It is presumed that at the initial stage of
photolysis, very small Ag nanoparticles that are too small to
exhibit SERS phenomenon are produced, but further exposure
of light must lead to the growth and agglomeration of Ag
nanoparticles sufficient for the occurrence of the SERS phe-
nomenon. In contrast, during the thermal decomposition of Ag-
4NBA, 4NBA-capped Ag nanoparticles are produced only in a
limited temperature region so that the particles must assume
well-separated positions without aggregation; hence, the EM
enhancement is insufficient to produce the SERS phenomenon.
3.4. Plausible Mechanism for Producing Silver Nano-
particles from Silver 4-Nitrobenzoate. There is no generally
accepted mechanism for the thermal as well as the photolytic
decomposition of silver carboxylates. Nonetheless, it has
been proposed in the literature that both processes must proceed
via the production/annihilation of organic radicals as
follows:15,16,21,22
Figure 8. Raman spectra of silver 4-nitrobenzoate taken at static
conditions under irradiation by the 632.8-nm line (17 mW at the
sampling position) for different durations. The inset shows the variation
of the 8a mode intensity of 4-NBA. The dashed line is drawn only as
a guide to the eye.
assembled on the 2-µm-sized silver powders shown in Figure
7A(d). These observations suggest that Ag nanoparticles have
indeed been produced from Ag-4NBA upon irradiation by an
Ar+ laser and that the nitro-to-amine group photoconversion
has subsequently taken place on nanosized Ag particles.
To confirm the production of nanosized silver particles, a
piece of pelletized Ag-4NBA that had been exposed to the
514.5-nm line of a 17 mW Ar+ laser for 10 min was immersed
in a chloroform/methanol mixture, and then the mixture was
shaken gently after which its decanted solution was dropped
on a copper grid for the TEM measurement; the white-colored
Ag-4NBA turns yellowish-brown upon irradiation by an Ar+
laser. In fact, as can be seen in Figure 7B, nanosized silver
particles are clearly observed. Contrary to the nanoparticles
formed from the thermal decomposition of Ag-4NBA, the sizes
are polydisperse and the shapes are quite irregular. Some
particles are also seen to be present in a highly aggregated state.
These data dictate that the spectra in Figure 7A(b) must be SERS
spectra that have been derived from the photolysis of Ag-4NBA
to give aggregated Ag nanoparticles followed by nitro-to-amine
photoconversion on these particles.
To examine the possible wavelength dependence, the Raman
spectrum of Ag-4NBA has also been obtained in static condi-
tions using the 17 mW He/Ne laser at 632.8 nm as the excitation
source. Figure 8 shows the Raman spectra obtained consecu-
tively under exposure to the light for 2, 80, and 120 min.
Surprisingly, the Raman spectral pattern is nearly invariant,
although the absolute peak intensities are variable with respect
to the time of laser exposure. The Raman spectral pattern in
Figure 8 is seen to be comparable to that in Figure 6b. This
implies that the nitro-to-amine group conversion does not take
place for Ag-4NBA upon exposure to the He/Ne laser at 632.8
nm. Nonetheless, the changes in the absolute peak intensities
in Figure 8 are indicative of the occurrence of some sort of
photolysis. As shown in the inset of Figure 8, the peak intensity
of the 8a mode gradually increases up to 80 min and then
abruptly decreases and reaches a plateau level. These spectral
changes are presumed to be associated with the production and
aggregation of Ag nanoparticles by 632.8-nm radiation. We
could indeed identify nanosized particles from the TEM
measurement of the He/Ne laser exposed Ag-4NBA (data not
shown). We separately confirmed that the nitro-to-amine group
conversion hardly occurs for 4-NBA adsorbed on SERS-active
Ag substrates by 632.8-nm radiation, as found in this work.
This may reflect that photoelectrons are more readily ejected
from Ag (at least in the visible region) as the wavelength of
the light source becomes closer to that of the surface plasmon
R-CO2-Ag f Ag0 + R-COO‚ f (Ag0 + R‚ + CO2)
or (Ag0 + R-COOH)
The carboxyl radical (R-COO‚) can either decompose into a
secondary radical (R‚) and carbon dioxide (CO2) or convert to
polymeric species (Rn) and/or carboxylic acid (RCOOH) by
hydrogen abstraction from nearby organic moieties.
In the present work, the thermal decomposition of Ag-4NBA
was observed to occur above 593 K under N2 atmosphere. The
CdO stretching peak was clearly identified at ∼1700 cm-1
above 513 K (see Figure 3), which must have arisen from a
free acid, that is, 4-NBA. (To form a free acid, a hydrogen
source is required. Its source is not yet clear, but we presume
that hydrogen atoms are provided by the dehydrogenation of
phenyl rings.) Metallic silver was also found to form upon
thermal decomposition of Ag-4NBA. However, we could not
find any evidence that would indicate the presence of silver
oxide as a constituent of the thermal decomposition products.
This implies that metallic silver and 4-NBA are produced as
the primary decomposition products but metallic silver is
immediately stabilized as nanoparticles by capping with 4-NBA
molecules, as shown in Figure 5B.
Although the photolysis of Ag-4NBA was conducted in the
ambient condition, its reaction appeared to proceed via a

4006 J. Phys. Chem. B, Vol. 108, No. 13, 2004
Seo et al.
mechanism similar to thermolysis. As mentioned previously,
Yonezawa et al.16 prepared a silver film via the photolysis of
silver alginate on glass under UV irradiation in air; the film
was confirmed to be composed of pure silver by XRD and X-ray
fluorescence (XRF) analyses. The absence of silver oxide may
indicate that water molecules in air do not participate in the
photolysis of silver alkanecarboxylates (AgCO2R). We presume
that the conditions of inertness of water and/or ambient oxygen
are maintained even for the cases of photolysis by visible light.
It is also expected that the irradiation by visible light does not
cause noticeable damage to the organic moiety of AgCO2R. In
this light, it is not unreasonable to observe that 4NBA-capped
Ag nanoparticles are produced by the visible-light-induced
photolysis of Ag-4NBA. Once the 4NBA-capped Ag nanopar-
ticles are formed, they will be immediately subjected to surface-
induced photoreaction to form 4ABA-capped Ag nanoparticles.
Regarding the surface-induced photoreaction of 4-NBT or
4-NBA on Ag, two specific questions have to be answered,
namely, how the reaction proceeds, and what the hydrogen
source is. Although the detailed mechanism is a matter of
conjecture, the reaction is presumed to be associated with the
charge transfer from silver to the adsorbed molecule. If the
energy difference between the Fermi level of the metal and the
low-lying excited state of the charge-transfer complex matches
the energy of the excitation radiation, a resonant charge transfer
from the metal to the excited state of the complex can take
place.1b The electron-transfer step is supposed to be a direct,
optically induced charge transfer from the Fermi level to the
lowest-lying unoccupied molecular orbital of the adsorbate-
silver complex. On the other hand, it is known that when
photoreaction occurs for aromatic nitro molecules in a solution
phase, a chemical species from which hydrogen atom(s) can
be abstracted is needed.23 For a similar photoreaction to occur
in air, water has been claimed to act as a hydrogen source.24
The photoreaction of 4-NBT on Ag occurs several times faster
in water or ethanol medium than in air.2a The photoreaction of
4-NBA occurs much faster when the molecule is coadsorbed
on Ag with hydrophilic group-terminated alkanethiols than with
hydrophobic group-terminated alkanethiols.2c These observations
suggest that the source of hydrogen atoms in the surface-induced
photoreaction of 4-NBT on Ag to 4-ABT must be water or
solvent molecules trapped inside the 4-NBT monolayers rather
than 4-NBT itself.
study of 4-NBA; hence, 4ABA-capped Ag nanoparticles are
produced via the formation of 4NBA-capped Ag nanoparticles
from Ag-4NBA. From the point of view of application, all of
these observations suggest that, if the terminal functionality of
the precursor molecule is judiciously chosen, the silver nano-
particles derivatized with proper end-functional molecules can
be synthesized not only via thermolysis but also by photolysis
of silver carboxylates. For instance, amine group-terminated
nanoparticles can be easily manufactured via such thermolysis
and/or photolysis, and the amine group may further be modified
to incorporate other molecules or biomolecules to function as
delicate indicators or biosensors. In addition, the wavelength-
dependent photoresponse of Ag-4NBA can be applied to the
read-out methodology in memory devices and/or sensor fields.
Acknowledgment. This work was supported in part by the
Ministry of Health & Welfare (Korea Health 21 R&D Project
01-PJ11-PG9-01NT00-0023) and by the Ministry of Science
and Technology (Nano Project, M10213240001-02B1524-
00210) in the Republic of Korea.
Supporting Information Available: Room-temperature
DRIFT spectrum of silver 4-nitrobenzoate; XRD pattern of silver
nanoparticles obtained from the thermal decomposition of silver
4-nitrobenzoate. This material is available free of charge via
the Internet at http://pubs.acs.org.
References and Notes
(1) (a) Roth, P. G.; Venkatachalam, R. S.; Boerio, F. J. J. Chem. Phys.
1986, 85, 1150. (b) Sun, S.; Birke, R. L.; Lombardi, J. R.; Leung, K. P.;
Genack, A. Z. J. Phys. Chem. 1988, 92, 5965. (c) Bercegol, H.; Boerio, F.
J. J. Phys. Chem. 1995, 99, 8763. (d) Yang, X. M.; Yang, X. M.; Tryk, D.
A.; Hashimoto, K.; Fujishima, A. J. Phys. Chem. B 1998, 102, 4933.
(2) (a) Han, S. W.; Lee, I.; Kim, K. Langmuir 2002, 18, 182. (b) Lee,
S. J.; Kim, K. Chem. Phys. Lett. 2003, 378, 122. (c) Han, H. S.; Han, S.
W.; Kim, C. H.; Kim, K. Langmuir 2000, 16, 1149.
(3) (a) Aksay, I. A.; Trau, M.; Manne, S.; Honma, I.; Yao, N.; Zhou,
L.; Fenter, P.; Eisenberger, P. M.; Gruner, S. M. Science 1996, 273, 892.
(b) Lacroix, P. G.; Clement, R.; Nakatani, K.; Zyss, J.; Ledoux, I. Science
1994, 263, 658. (c) Vermeulen, L. A.; Thompson, M. E. Nature 1992, 358,
656. (d) Laget, V.; Hornick, C.; Rabu, P.; Drillon, M.; Turek, P.; Ziessel,
R. N. AdV. Mater. 1998, 10, 1024. (e) Kagan, C. R.; Mitzi, D. B.;
Dimitrakopulos, C. D. Science, 1999, 286, 945. (f) Mitzi, D. B. Chem. Mater.
2001, 13, 3283.
(4) (a) Vand, V.; Aitken, A.; Campbell, R. K. Acta Crystallogr. 1949,
2, 398. (b) Matthews, F. W.; Warren, G. G.; Michell, J. H. Anal. Chem.
1950, 22, 514. (c) Tolochko, B. P.; Chernov, S. V.; Nikitenko, S. G.;
Whitcomb, D. R. Nucl. Instrum. Methods A 1998, 405, 428. (d) Lee, S. J.;
Han, S. W.; Choi, H. J.; Kim, K. J. Phys. Chem. B 2002, 106, 2892. (e)
Lee, S. J.; Han, S. W.; Choi, H. J.; Kim, K. J. Phys. Chem. B 2002, 106,
7439. (f) Choi, H. J.; Han, S. W.; Lee, S. J.; Kim, K. J. Colloid Interface
Sci. 2003, 264, 458.
(5) (a) Uvarov, N. F.; Burleva, L. P.; Mizen, M. B.; Whitcomb, D. R.;
Zou, C. Solid State Ionics 1998, 107, 31. (b) Chadha, M.; Dunnigan, M.
E.; Sahyun, M. R. V.; Ishida, T. J. Appl. Phys. 1998, 84, 887.
(6) Lee, S. J.; Han, S. W.; Kim, K. Chem. Commun. 2002, 442.
(7) Lee, S. J.; Kim, K. Chem. Commun. 2003, 212.
(8) Blakeslee, A. E.; Hoard, J. L. J. Am. Chem. Soc. 1956, 78, 3029.
(9) Groth, P. Acta Chem. Scand. A 1980, 34, 229.
(10) CRC Handbook of Chemistry and Physics, 71st ed.; Lide, D. R.,
Ed.; CRC Press: Boca Raton, FL, 1990.
(11) Lee, S. J.; Han, S. W.; Yoon, M.; Kim, K. Vib. Spectrosc. 2000,
24, 265.
(12) (a) Gericke, A.; Hu¨hnerfuss, H. Thin Solid Films 1994, 245, 74.
(b) Ohe, C.; Ando, H.; Sato, N.; Urai, Y.; Yamamoto, M.; Itoh, K. J. Phys.
Chem. B 1999, 103, 435.
(13) Han, H. S.; Kim, C. H.; Kim, K. Appl. Spectrosc. 1998, 52, 1047;
Lee, S. J.; Kim, K. Vib. Spectrosc. 1998, 18, 187; Han, S. W.; Han, H. S.;
Kim, K. Vib. Spectrosc. 1999, 21, 133.
(14) Merklin, G. T.; Griffith, P. G. J. Phys. Chem. B 1997, 101, 5810.
(15) Bokhonov, B. B.; Lomovsky, O. I.; Andreev, V. M.; Boldyrev, V.
V. J. Solid State Chem. 1985, 58, 170; Bokhonov, B. B.; Burleva, L. P.;
Whitcomb, D. R.; Usanov, Yu. E. J. Imaging Sci. Technol. 2001, 45, 259.
4. Conclusion
We confirmed by XRD analysis that silver 4-nitrobenzoate
(Ag-4NBA) consisted of a layered structure. The temperature-
dependent DRIFT and TGA measurements indicated that Ag-
4NBA was subjected to three distinct transitions. In the first
transition at 383 K, the binding state of carboxylate is converted
from bridging to unidentate. In the second transition at 593 K,
Ag-4NBA is decomposed to produce 4NBA-capped Ag nano-
particles. This second transition temperature for Ag-4NBA is
about 80 K higher than that for silver alkanecarboxylates, and
such enhanced thermal stability of Ag-4NBA must be associated
with the conjugation of the carboxylate group to the aromatic
ring. In the third transition at 623 K, the organic moieties are
totally lost from the Ag nanoparticles. Separately, SERS-active,
nanosized silver particles or aggregates were produced from Ag-
4NBA by visible-laser irradiation at 514.5 and 632.8 nm. The
particles produced under irradiation by the 514.5-nm light were
derivatized with 4-ABA, whereas those under the 632.8-nm light
were capped with 4-NBA. The nitro-to-amine group conversion
observed herein is exactly in conformity with the recent SERS

Silver 4-Nitrobenzoate Conversion to Nanoparticles
J. Phys. Chem. B, Vol. 108, No. 13, 2004 4007
(16) Yonezawa, Y.; Takami, A.; Sato, T.; Yamamoto, K.; Sasanuma,
T.; Ishida, H.; Ishitani, A. J. Appl. Phys. 1990, 68, 1297; Yonezawa, Y.;
Konishi, Y.; Hada, H.; Yamamoto, K.; Ishida, H. Thin Solid Films 1992,
218, 109.
(17) Kim, K.; Lee, I.; Lee, S. J. Chem. Phys. Lett. 2003, 377, 201.
(18) Lin-Vein, D.; Colthup, N. B.; Fateley, W. G.; Grasselli, J. G. The
Handbook of Infrared and Raman Characteristic Frequencies of Organic
Molecules; Academic Press: San Diego, CA, 1991.
(20) (a) Kneipp, K.; Kneipp, H.; Itzkan, I.; Dasari, R. R.; Feld, M. S. J.
Phys.: Condens. Matter 2002, 14, R597 and references therein. (b) Shipway,
A. N.; Katz, E.; Willner, I. ChemPhysChem, 2000, 1, 18.
(21) Field, E. K.; Meyerson, S. J. Org. Chem. 1976, 41, 916.
(22) Usubaliev, B. T.; Nadjafov, H. N.; Musaev, A. A.; Mamedov, K.
S. Thermochim. Acta 1989, 93, 57.
(23) Barltrop, J. A.; Bunce, N. J. J. Chem. Soc. C 1968, 1467.
(19) Fedurco, M.; Shklover, V.; Augustynski, J. J. Phys. Chem. B 1997,
101, 5158.
(24) Tsai, W.-H. Ph.D. Dissertation, University of Cincinnati, Cincinnati,
OH, 1991.