Surface-enhanced Raman scattering of 4-mercaptopyridine on thin films of nanoscale Pd cubes, boxes, and cages

Article abstract of DOI:10.1016/j.cplett.2005.10.028

We have synthesized a variety of Pd nanoparticles of 8-50 nm in size including solid cubes, hollow boxes, and porous cages. Using 4-mercaptopyridine (4MP) as a probe molecule, we have characterized thin films for surface-enhanced Raman scattering (SERS) activity, and have found a significant level of enhancement (with factors ranging from 170 for 8-nm cubes to 1.3 × 10⁴ for boxes). For the cubes and boxes, we observed a trend of stronger enhancement with more red-shifted SPR bands. We evaluated the sensitivity of this approach, and also used SERS to monitor monolayer formation on these particles.

Full text of DOI:10.1016/j.cplett.2005.10.028

Chemical Physics Letters 417 (2006) 230–234
www.elsevier.com/locate/cplett
Surface-enhanced Raman scattering of 4-mercaptopyridine on thin
films of nanoscale Pd cubes, boxes, and cages
*
Joseph M. McLellan, Yujie Xiong, Min Hu, Younan Xia
Department of Chemistry, University of Washington, Seattle, WA 98195-1700, USA
Received 23 September 2005
Available online 27 October 2005
Abstract
We have synthesized a variety of Pd nanoparticles of 8–50 nm in size including solid cubes, hollow boxes, and porous cages. Using 4-
mercaptopyridine (4MP) as a probe molecule, we have characterized thin films for surface-enhanced Raman scattering (SERS) activity,
and have found a significant level of enhancement (with factors ranging from 170 for 8-nm cubes to 1.3 · 104 for boxes). For the cubes
and boxes, we observed a trend of stronger enhancement with more red-shifted SPR bands. We evaluated the sensitivity of this approach,
and also used SERS to monitor monolayer formation on these particles.
ꢀ 2005 Elsevier B.V. All rights reserved.
1. Introduction
ever, the wide application of other noble metals such as
Pd, Pt, and Rh in catalysis, make the extension of SERS
to these metals attractive.
Interest in surface-enhanced Raman scattering (SERS)
has truly blossomed over the last decade and a half. Most
of the efforts have focused on the use of coinage metal sub-
strates (Ag, Au, and Cu) due to their extremely high
enhancement factors, which have even enabled single mol-
ecule detection [1]. It has been widely accepted that there
are two primary enhancement mechanisms in SERS: one
being electromagnetic (EM) enhancement that arises from
the extremely high local fields due to surface plasmon res-
onance (SPR) and the other being chemical enhancement
due to resonance Raman-like interaction between the sub-
strate and the adsorbate [2]. For noble metals, enhance-
ment of SERS signals tends to be primarily caused by the
EM effect. In principle, the position of an SPR band for
a metal nanoparticle can be controlled by varying the com-
position, size, and shape [3]. The SPR bands of the afore-
mentioned coinage metal nanoparticles are typically
located in the visible to near-infrared (NIR) regions, and
thus have been the substrates of choice for conventional
Raman systems, which utilize visible to NIR lasers. How-
To this end, several approaches have been developed.
The first of these can generally be described as a Ôborrow-
ingÕ technique where a less SERS-active material is depos-
ited on a SERS-active substrate and ÔborrowsÕ from the
strong local field generated in the SERS-active material.
This approach has been demonstrated in a variety of ways
ranging from thin metallic films (3–10 atomic layers)
deposited on roughened electrodes [4–6] to core–shell [7–
9] and hybrid [10,11] bimetallic nanoparticles. For all of
these demonstrations, the SERS enhancement rapidly de-
creased with increasing shell or overlayer thickness. Other
approaches have focused on the use of monometallic nano-
particles by synthesizing Pt [9,12,13], Pd [13,14] and Rh [15]
sols, and then using aggregated films of these particles for
SERS. However, in all of these cases the particles were less
than 20 nm and showed no SPR band in the visible or NIR
regions. Since these metals generally have a small imagi-
nary part in their dielectric constants, their SPR bands
are typically located in the UV region. Tian and co-workers
[16] recently demonstrated the application of UV-SERS,
and showed modest enhancement from roughened Rh
and Ru surfaces as compared to Ag particles when
*
Corresponding author. Fax: +1 206 685 8665.
E-mail address: xia@chem.washington.edu (Y. Xia).
0009-2614/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.cplett.2005.10.028

J.M. McLellan et al. / Chemical Physics Letters 417 (2006) 230–234
231
325 nm irradiation was used. Most recently we have
3. Results and discussion
successfully synthesized Pd nanoparticles with edge lengths
of 25 nm and above and further processed them into
hollow structures. All these Pd nanoparticles exhibited
SPR bands in the visible region. Here, we demonstrate
the SERS activities of thin films consisting of these Pd
nanoparticles, and compare their activities as a function
of size, solidity (solid versus hollow), and thus the positions
of SPR bands.
3.1. Comparison of SERS activity
Our group has recently demonstrated shape-controlled
synthesis for a number of noble metals (e.g., Ag, Pt, Pd,
and Rh) with an emphasis on how the positions of SPR
bands are determined by the size, shape, and composition
of the particles [11,17–21]. One particularly interesting
example is Pd, for which we have synthesized solid cubes
[17] with average edge lengths ranging from 8 to 50 nm,
as well as hollow boxes and cages [18] with edge lengths
of ꢀ50 nm. Fig. 1 shows UV–vis absorption spectra and
corresponding TEM images for the Pd nanoparticles that
were used for SERS measurements. Nanocubes with edge
2. Experimental
2.1. Preparation of Pd nanoparticles
The procedures for synthesizing Pd nanocubes, nano-
boxes, and nanocages have already been published
[17,18]. Briefly, these nanoparticles were synthesized using
a polyol method in which Na2PdCl4 was reduced by ethyl-
ene glycol (EG) in the presence of poly(vinyl pyrrolidone)
(PVP). When the same concentration of precursor was
used, the size of the cubes could be controlled by adjusting
the density of seeds formed in the nucleation step. The
synthesis of nanoboxes and nanocages was similar, except
water was added to increase the solubility of oxygen and
thus oxidative corrosion. Cubes were formed initially, but
these could be transformed into boxes and further into
cages with time. Following synthesis, the particles were
washed with acetone and then by ethanol several times to
remove EG and excess PVP, and finally redispersed in
deionized water. Samples were characterized by UV–vis
spectroscopy (HP 8452A diode array spectrometer), and
transmission electron microscopy (TEM). TEM samples
were prepared by drying aqueous suspensions of the
particles on carbon coated TEM grids under ambient
conditions. TEM images were taken on a Phillips 420
transmission electron microscope operated at 120 kV.
2.2. Raman spectroscopic measurement
SERS substrates were prepared by drop-casting 5 lL of
the aqueous sol on a Si wafer and allowed to completely
dry under vacuum. The dry films were immersed in a
4 mM aqueous solution of 4-mercaptopryidine (4MP) for
1 h. The samples were then taken out, rinsed with deionized
water to remove any unadsorbed 4MP molecules, and
dried with a stream of air. For the nanoboxes, two series
of samples were prepared: one with solutions of 4MP that
ranged from 4 mM to 4 pM and another where films were
immersed in a 4 mM solution for different periods of time.
Raman spectra were obtained using a Renishaw inVia
Raman spectrometer (HPNIR785) attached to a Leica
DM IRBE optical microscope. A spot size of ꢀ1.6 lm in
diameter and 785-nm laser excitation from a diode laser
at a power of 3 mW at the sample surface were used to
irradiate the samples. The Raman scattering signals were
collected on a thermoelectrically cooled (ꢁ60 ꢁC) CCD
detector.
Fig. 1. Extinction spectra and corresponding TEM images of Pd
nanoparticles. (a) From top to bottom 50-nm (solid line), 25-nm (dashed),
and-8 nm cubes (doted). (b) Pd nanoboxes (left image; dashed line) and
nanocages (right image; solid line). All scale bars correspond to 50 nm.

232
J.M. McLellan et al. / Chemical Physics Letters 417 (2006) 230–234
lengths of 8, 25, and 50 nm showed SPR bands peaked at
225, 330, and 390 nm, respectively (Fig. 1a). Note that
for the 8-nm cubes, only a ÔtailÕ can be seen extending from
the UV to the visible region. By processing 48-nm cubes
into hollow structures the SPR bands could be further
red-shifted. Fig. 1b shows typical spectra and TEM of such
Pd nanoboxes and nanocages. The positions of all these
SPR bands were consistent with the results of discrete
dipole approximation (DDA) calculations [17,18].
The highest EM field enhancement arises when the irra-
diation wavelength is resonant with the SPR maximum
[22,23]. For SERS measurements we used a 785 nm laser,
so one could expect to see enhanced SERS activity for
those particles with more red-shifted SPR bands. Fig. 2
compares the SERS spectra of 4MP on thin films of 8-,
25-, and 50-nm Pd cubes. We chose 4MP as a probe mole-
cule because: (i) it has been well established in the literature
[24]; (ii) it will form a monolayer on the surface of Pd
nanoparticle making the estimate of surface coverage more
convenient [25]; (iii) it has a large scattering cross-section
[26]. We also investigated the activity of Pd hollow nano-
particles that have even more red-shifted SPR bands.
Fig. 3 shows SERS spectra obtained from films of 50-nm
nanoboxes and nanocages. We found that the nanoboxes
gave the strongest enhancement among all the samples.
The results unambiguously demonstrate the activity of Pd
for SERS. For Pd nanocubes and nanoboxes, the SERS
intensity was greatly enhanced as the SPR band of these
particles was red-shifted. However, the weaker activity of
the nanocages is hard to understand, because the cages
have the most red-shifted SPR band and they are expected
to have the strongest enhancement. A plausible explana-
tion is that a change in the structure of these particles
not only effects the position of the SPR band, but also
the contributions of scattering and absorption to the
Fig. 3. SERS spectra of 4-mercaptopyridine (4MP) adsorbed on thin films
of Pd nanoboxes and nanocages. The inset shows a typical SEM image of
the film containing 50-nm nanoboxes that was used for SERS measure-
ments. The scale bar corresponds to 500 nm.
extinction spectra, as well as overall optical cross-section
[18]. For example, the cages have a thinner wall relative
to the boxes, and this difference in structure could have a
significant effect on the overall EM enhancement. This
observation suggests that while the position of the SPR
band relative to the excitation wavelength is important
for SERS, there may be other significant contributions to
the enhancement mechanism such as aggregation effects.
It also highlights one of the challenges that have plagued
SERS since its infancy, that is, the inconsistent substrate
performance.
To evaluate the limit of detection for this system, we
prepared solutions ranging from 4 mM to 4 pM, immersed
nanobox films in them for 1 h, and then took SERS signals
from the films. The results of this experiment are shown in
Fig. 4, together with a typical SEM image of the films. For
the 4 pM sample no identifiable peaks were observed, how-
ever recognizable to high-quality spectra were obtained for
all other samples. The inset corresponds to a 10· magnifi-
cation of the 1030–1130 cmꢁ1 region for the sample pre-
pared from 1 nM 4MP. Even though the signal-to-noise
ratio of this spectrum was extremely poor, it still seems
to be possible to detect the adsorbed 4MP molecules at this
concentration level, and at higher quality spectra might be
obtainable by optimizing the acquisition time and the num-
ber of accumulations.
The high sensitivity of SERS could also be used to inves-
tigate the formation of 4MP monolayer on the surface of
Pd nanoparticles. Monolayer formation is typically moni-
tored using a combination of techniques such as contact
angle or elipsometry measurements and X-ray photoelec-
tron spectroscopy; results are typically plotted as a func-
tion of immersion time. We demonstrate here that SERS
can also be used in the same way to monitor the formation
kinetics. As shown in Fig. 5, we plotted the integrated
intensity of the 1096 cmꢁ1 band for 4MP adsorbed on Pd
Fig. 2. SERS spectra of 4-mercaptopyridine (4MP) adsorbed on films of
Pd nanocubes with 50, 25, and 8 nm edge lengths. The inset shows a
typical SEM image of the thin film containing 50-nm Pd cubes, where the
scale bar corresponds to 500 nm.

J.M. McLellan et al. / Chemical Physics Letters 417 (2006) 230–234
233
I
surf =Nsurf
G ¼
.
ð1Þ
I
bulk=Nbulk
In Eq. (1), Isurf and Ibulk are the integrated intensities of the
same m1a bands for the adsorbed 4MP and unadsorbed
4MP molecules respectively, and similarly the Nsurf and
Nbulk are the number of adsorbed and unadsorbed 4MP
molecules contributing to the signal. For the bulk (unad-
sorbed) 4MP we used a 0.3 M aqueous solution. In order
to determine the number of molecules that contributed to
the Ibulk signal we had to determine the confocal depth pro-
file of our microscope [26]. For the optical configuration
and microscope employed in this work, the confocal depth
was 13 lm. We used a spot size of 1.6 lm giving a focal vol-
ume of 0.1 pL, and thus Nbulk, for a 0.3 M solution, would
be 2 · 1010 4MP molecules. We estimated the number of
4MP molecules sampled (Nsurf), in the case of the 50-nm
cubes, to be 1.9 · 109. Using the 1096 cmꢁ1 band, Isurf
and Ibulk were 1.7 · 104 and 180, respectively. From these
numbers, we calculated an enhancement factor of
9.3 · 103 for the film of 50-nm cubes, and 2.2 · 103 and
170 for the 25- and 8-nm cubes, respectively. For the nano-
boxes, we calculated an enhancement factor of 1.3 · 104,
while we found an enhancement factor of only 7.5 · 103
for the nanocages. The enhancement factors given here
are for typical samples, and it should be noted that we
did often observe a marked variation in SERS enhance-
ment from sample to sample, even when prepared under
identical conditions. This observation is a good indication
that the way in which the particles aggregate upon drying
may play a critical role in the enhancement mechanism,
as has been seen with Ag and Au nanoparticles
[23,24,27–30].
Fig. 4. SERS spectra from thin films of Pd nanoboxes that had been
functionalized with aqueous 4-mercaptopyridine (4MP) solutions that
ranged from 4 · 10ꢁ3 to 4 · 10ꢁ9 M. A distinct peek at 1096 cmꢁ1 could
still be observed for the 4 · 10ꢁ9 M sample by magnifying the signals in
the 1030–1130 cmꢁ1 region (see the inset). The inset at the top shows a
typical SEM image of the nanobox film that was functionalized with a
4 mM solution. The scale bar corresponds to 200 nm.
30000
25000
20000
15000
10000
5000
4. Summary
0
10
20
30
40
50
60
Immersion Time (min)
A variety of Pd cubic nanoparticles have been synthe-
sized using the polyol method, and characterized by UV–
vis spectroscopy and TEM. Thin films of each of type of
particles were prepared, functionalized with monolayers
of 4MP, and then used for SERS measurements with a con-
focal Raman microscope. We found that (i) these particles
exhibited extraordinary high SERS activity, as compared
to previous demonstrations of SERS on Pd nanoparticles
[14] and (ii) there was a strong correlation between the
SERS intensity and the size or solidity of the particles.
The activity is most likely due to a combination of the
enhancement arising from the significant EM field
enhancement due to SPR and aggregation effects, as indi-
cated by the correlation of SERS activity to SPR position
and variation of the SERS intensity from samples prepared
under identical conditions, respectively. For the nanocubes
and nanoboxes a trend of higher SERS activity with red-
shifted SPR peak was observed. However, in the case of
nanocages, which had the most red-shifted SPR band, we
found a surface enhancement factor that was almost half
that of the nanoboxes (1.3 · 104 versus 7.5 · 103). We are
Fig. 5. Plot of the relative intensity of the 1096 cmꢁ1 band in the SERS
spectra of 4-mercaptopyridine (4MP) adsorbed on Pd nanoboxes as a
function of the immersion time. A 4-mM aqueous solution of 4MP was
used.
nanoboxes as a function of immersion time. The shape of
the curve is consistent with a typical Langmuir adsorption
process. The major advantage of using SERS to monitor
the kinetics is that unlike elipsometry or contact angle mea-
surements, one obtains structural and compositional data
too. In addition, a stronger enhancement might enable
monitoring of molecular adsorption on individual parti-
cles,
techniques.
a task not readily accomplished using other
3.2. Estimate of the surface enhancement factor
To quantify the results, we calculated surface enhance-
ment factors, G, for the samples shown in Figs. 1 and 2
using methods described by Tian and coworkers [26]

234
J.M. McLellan et al. / Chemical Physics Letters 417 (2006) 230–234
[9] Z.Q. Tian, Z.L. Yang, B. Ren, J.F. Li, Y. Zhang, X.F. Lin, J.W. Hu,
D.Y. Wu, Faraday Discuss. 132 (13) (2005), preprint.
[10] B. Pergolese, A. Bigotto, G. Muniz-Miranda, G. Sbrana, Appl.
Spectrosc. 591 (2005) 194.
[11] J. Chen, B. Wiley, J. McLellan, Y. Xiong, Z.Y. Li, Y. Xia, Nano Lett.
5 (2005) 2058.
[12] N.H. Kim, K. Kim, Chem. Phys. Lett. 393 (2004) 478.
[13] R. Go¨mez, J.M. Pe´rez, J. Solla-Gullo´n, V. Montiel, A. Aldaz, J. Phys.
Chem. B 108 (2004) 9943.
currently working to better understand this result as it ap-
pears to be contrary to predictions based on DDA calcula-
tions for Au nanocages [31], and thus could shed light on
the SERS enhancement mechanism for Pd nanoparticles.
We also looked at films of Pd nanoboxes that were func-
tionalized with very dilute 4MP solutions (4 nM), and with-
out optimization, we were able to obtain recognizable
spectra. Finally, we also demonstrate how SERS could be
used to follow the formation kinetics of monolayers on
Pd nanoparticles.
´
ˇ
´
[14] I. Srnova, B. Vlckova, V. Baumruk, J. Mol. Stuct. 410–411 (1997)
201.
[15] W.L. Parker, R.M. Hexter, Chem. Phys. Lett. 107 (1984) 96.
[16] B. Ren, X.F. Lin, Z.L. Yang, G.K. Liu, R.F. Aroca, B.W. Mao, Z.Q.
Tian, J. Am. Chem. Soc. 125 (2003) 9598.
Acknowledgments
[17] Y. Xiong, J. Chen, B. Wiley, Y. Xia, Y. Yin, Z.Y. Li, Nano Lett. 5
(2005) 1237.
[18] Y. Xiong, B. Wiley, J. Chen, Z.Y. Li, Y. Yin, Y. Xia, Angew. Chem.
Int. Ed. (2005) (in press).
[19] N. Zettsu, J.M. McLellan, B. Wiley, Y. Yin, Z.Y. Li, Y. Xia, Angew.
Chem. Int. Ed. (under review).
[20] Y. Sun, Y. Xia, Analyst 128 (2003) 686.
[21] B. Wiley, Y. Sun, J. Chen, H. Cang, Z.Y. Li, X. Li, Y. Xia, MRS
Bulletin 30 (2005) 356.
[22] N. Halas, MRS Bull. 30 (2005) 362.
[23] A.M. Scwartberg, C.D. Grant, A. Wolcott, C.E. Talley, T.R. Huser,
R. Bogomolni, J.Z. Zhang, J. Phys. Chem. B 108 (2004) 19191.
[24] Z. Wang, S. Pan, T.D. Krauss, H. Du, L.J. Rothberg, Proc. Natl.
Acad. Sci. USA 100 (2003) 8638.
This work was supported in part by the MLSC Program
at the UW and a fellowship from the David and Lucile
Packard Foundation. Y.X. is an Alfred P. Sloan Research
Fellow and a Camille Dreyfus Teacher Scholar. M.H. is an
INEST Postdoctoral Fellow supported by Phillip Morris
USA. Instrumentation was provided by the Nanotech User
Facility (NUTF), a member of the National Nanotechnol-
ogy Infrastructure Network (NNIN) supported by NSF.
References
[25] S. Chen, K. Huang, J.A. Stearns, Chem. Mater. 12 (2000) 540.
[26] W.B. Cai, B. Ren, X.Q. Li, C.X. She, F.M. Liu, X.W. Cai, Z.Q. Tian,
Surf. Sci. 406 (1998) 9.
[27] P.M. Tessier, O.D. Velev, A.T. Kalambur, J.F. Rabolt, A.M.
Lenhoff, E.W. Kaler, J. Am. Chem. Soc. 122 (2000) 9554.
[28] D.M. Kuncicky, S.D. Christesen, O.D. Velev, Appl. Spectrosc. 59
(2005) 401.
[29] A.R. Tao, P. Yang, J. Phys. Chem. B 109 (2005) 15687.
[30] A. Tao, F. Kim, C. Hess, J. Goldberger, R. He, Y. Sun, Y. Xia, P.
Yang, Nano. Lett. 3 (2003) 1229.
[1] S.M. Nie, S.R. Emory, Science 275 (1997) 1102.
[2] Z.Q. Tian, B. Ren, D.Y. Wu, J. Phys. Chem. B. 106 (2002) 9463.
[3] A.J. Hayes, C.L. Haynes, A.D. McFarland, G.C. Schatz, R.P. Van
Duyne, S. Zou, MRS Bull. 30 (2005) 368.
[4] M. Fleischmann, Z.Q. Tian, L.J. Li, J. Electroanal. Chem. 217 (1987)
397.
[5] Q. Feng, T.M. Cotton, J. Phys. Chem. 90 (1986) 983.
[6] Y. Zhang, X. Gao, M.J. Weaver, J. Phys. Chem. 97 (1993) 8656.
[7] J.W. Hu, Y. Zhang, J.F. Li, Z. Liu, B. Ren, S.G. Sun, Z.Q. Tian, T.
Lian, Chem. Phys. Lett. 408 (2005) 354.
[31] E. Hao, S. Li, R.C. Bailey, S. Zou, G.C. Schatz, J.T. Hupp, J. Phys.
Chem. B 108 (2004) 1224.
[8] B. Zhang, J.F. Li, Q.L. Zhong, B. Ren, Z.Q. Tian, S.Z. Zou,
Langmuir 21 (2005) 7449.