PAPER
www.rsc.org/loc | Lab on a Chip
Additional amplifications of SERS via an optofluidic CD-based platform
Dukhyun Choi,†a Taewook Kang,†b Hansang Cho,a Yeonho Choia and Luke P. Lee*a
Received 14th July 2008, Accepted 1st September 2008
First published as an Advance Article on the web 23rd October 2008
DOI: 10.1039/b812067f
In this paper, signal amplifications of surface-enhanced Raman scattering (SERS) are realized by an
optofluidic compact disc (CD)-based preconcentration method for effective label-free environmental
and biomolecular detections. The preconcentration of target molecules is accomplished through the
accumulation of adsorbed molecules on SERS-active sites by repeating a ‘filling–drying’ cycle of the
assay solution in the optofluidic CD platform. After 30 cycles, the clear and high SERS signal of
rhodamine 6G of 1 nM is readily detected. In addition to the preconcentration-based signal
amplification by the optofluidic SERS system on the CD platform, we introduce a controlled
precipitation of gold nanoparticles by CuSO4 for SERS substrates. This method provides highthroughput, high-sensitive and large-area uniform SERS substrates on the optofluidic CD platform.
The uniform SERS signals from different positions in spots of 3 mm2 on the different CDs gives
us confidence in the reliability and stability of our SERS substrates.
1. Introduction
Surface-enhanced Raman scattering (SERS) has provided the
opportunities to design ultrasensitive biosensors for effective
label-free biomolecular assays. Furthermore, owing to the
versatility of Raman spectroscopy such as a nondestructive
approach to chemical analysis and the fingerprint characteristics
of biomolecules, a wide variety of applications of SERS have
been demonstrated in medicine, biochemistry, electrochemistry,
and ultrahigh vacuum (UHV) surface science.1–7 However, the
development of robust methods for integrating SERS with reliable and reproducible on-chip biosensing systems being capable
of easy sample preparation and dynamic analysis of target
molecules is still challenging.8–13 Here we report an optofluidic
SERS on a compact disk (CD) platform (this platform is referred
to as an optofluidic SERS-CD platform in this paper), which is
designed to preconcentrate the molecule of interest via simply
repeating ‘filling–drying’ cycles of molecular solutions for SERS
signal amplification. We demonstrate the promising capability of
the preconcentration using this SERS-CD platform through
SERS intensity of rhodamine 6G (R6G) with concentrations
ranging from 1 mM to 1 nM.
Assembled metal nanoparticles have so far been successfully
applied to a SERS substrate.14–16 For on-chip SERS-active
sites, accumulated gold nanoparticles in nanochannel10,17 and
Ag/PDMS nanowell structures12 were also used. In this study, we
introduce a controlled precipitation of gold nanoparticles
(GNPs) by CuSO4 as an excellent and suitable SERS substrate
for microfluidic CD platform. It was found that our SERS
substrate has both high sensitivity (down to 1 nM detection of
a
Biomolecular Nanotechnology Center, Berkeley Sensor & Actuator
Center, Department of Bioengineering, University of California at
Berkeley, Berkeley, California, 94720, USA. E-mail: [email protected];
Fax: +1-510-642-5835; Tel: +1-510-642-5855
b
Department of Chemical and Biomolecular Engineering, Sogang
University, Seoul, 121-742, Korea
† These authors contributed equally.
This journal is ª The Royal Society of Chemistry 2009
R6G) and uniformity (only 25% variation in SERS signal at 4
inch wafer scale CD platform level). In order to verify the reliability and stability of our GNP aggregates-based SERS
substrates, we investigated SERS spectra from different positions
within a spot (position-to-position), different spots on the same
CD (spot-to-spot) as well as on different CDs (batch-to-batch).
2. Conceptual overview
The basic concept of our optofluidic SERS-CD platform
including the preconcentration mechanism for the SERS spectra
amplification is illustrated in Fig. 1. The SERS-CD platform
consists of hierarchically-aligned microfluidic channels with
radial direction (Fig. 1a). Sample loaded into either one main
intake (the center of the SERS-CD platform) or twelve secondary
inlets (for various molecular solutions or concentrations) can be
uniformly dispensed to SERS-active sites at the end of each
channel and analyzed by SERS. High-throughput homogenous
sample preparation coupled with identical SERS detection sites
enables us to analyze the sample reliably. Moreover, the injection
of the sample solution and drying the solution onto the SERSactive sites from the main intake can be easily repeated to
preconcentrate the target molecule for SERS enhancement
(Fig. 1b-II, III). In general, the intensity of surface-enhanced
Stokes Raman scattering, ISERS(ns), can be estimated by,
ISERS(ns) f NM$|A(nL)|2$|A(ns)|2$sRads
(1)
where NM is the number of molecules involved in the SERS
process, sadsR is the Raman cross section of the adsorbed
molecule, and A(nL) and A(ns) are the field enhancement factors
at the laser and Stokes frequency, respectively.6 Among these
variables, NM is the only flexible and unlimited control parameter for the signal enhancement because the variables other
than NM are intrinsic factors which are almost fixed for a given
SERS-active substrate and a target molecule. Therefore, as the
proposed ‘filling–drying’ cycles progress, we can accomplish the
Lab Chip, 2009, 9, 239–243 | 239
Fig. 1 Optofluidic SERS-CD platform and preconcentration mechanism. (a) The schematic illustration of the optofluidic SERS-CD platform with
high-throughput sample preparation. (b) The mechanism of preconcentration consisting of 4 steps. I: Fabrication of SERS-active sites, II: Injection of
a sample solution, III: Adsorption of target molecules on the SERS-active site during drying process prior to SERS measurement. IV: Accumulation
of the adsorbed molecules by the repetition of the steps II and III. (c) As the ‘filling–drying’ cycle is progressed, the SERS intensity can be enhanced due
to the preconcentration of the target molecule.
facile preconcentration of target molecule (Fig. 1b-IV) in order
to significantly enhance the SERS signal (Fig. 1c).
3. Materials and methods
treated with oxygen plasma for 20 s, at the pressure of the oxygen
plasma of 450 mTorr, in a vacuum of about 1 Torr, and with
a power of 40 W for bonding the PDMS and the glass wafer and
for treating to be hydrophilic surfaces (Fig. 2b-III).
3.1 Optofluidic SERS-CD platform
We fabricated an optofluidic SERS-CD platform consisting of
a single main intake, 12 secondary inlets, and 84 SERS-active
detection spots (Fig. 2a). Venting outlets were fabricated above
the detection spots in order to reduce the pressure drop during
preconcentration. The SERS-CD platform was constructed
using standard soft lithography techniques (Fig. 2b). The mold
for microfluidic channels with a thickness of 50 mm was made of
a negative photoresist (SU8–2035, MicroChem, MA) (Fig. 2b-I).
Polydimethylsiloxane (PDMS) solution, a mixture of PDMS
prepolymer and a curing agent (SYLGARD 184 A and B, Dow
Corning, MI) in a ratio of 10:1 (v/v) was cast on the mold,
degassed in a vacuum dessicator with a mechanical pump to
remove any air bubbles and then cured for 4 h in an oven at 60 C
(Fig. 2b-II). The cured PDMS replica was peeled from the mold,
and intakes and outlet holes were punched into the cured replica.
Both the punched PDMS replica and a 4 inch glass wafer were
3.2 Precipitation of GNPs for SERS substrate
The salt-induced aggregated nanoparticles such as gold and
silver are well suited to act as SERS-active sites in the optofluidic
SERS-CD platform with respect to several reasons: (i) SERSactive sites can be simultaneously generated at the end of every
channel by injecting the colloidal solution into the main intake
which branches into 84 separate detection spots, (ii) no lithographic techniques are required. The other methods used to
create SERS substrates are unfavorable for integrating with
microfluidic CD platform in terms of the difficulties in bonding
between microchannels of the PDMS and SERS substrates,
and alignment between the SERS-active sites and detection
spots.18–23 Initially, CuSO4 was chosen as an aggregation agent,
since the higher charge on Cu2+ induces stronger aggregation
than monovalent cations, and halide anions strongly bind to
the surface of noble nanoparticles, which repels analytes.24 In
Fig. 2 Design and fabrication of the optofluidic SERS-CD platform. (a) Photograph of a SERS-CD platform which is organized by a single main
intake located in the center of CD, 12 secondary inlets with 7 microfluidic channels respectively for high-throughput assays and 84 SERS-active sites/
venting outlets. (b) I: Patterning of a mold for microfluidic channels and chambers by photolithography, II: Mold transfer with PDMS, III: Punching of
intakes and outlet holes and bonding of the PDMS replica to the glass wafer (4 inches).
240 | Lab Chip, 2009, 9, 239–243
This journal is ª The Royal Society of Chemistry 2009
a typical experiment, the SERS substrates were formed at the end
of every 84 microfluidic channels by loading the mixed solution
(v/v 50:1) of 80 nm diameter GNPs and 10 mM CuSO4 in the
main intake (note, during this process, all secondary intakes were
covered with glass slides to prevent the loading solution from
flowing out). Pressure applied to the main intake forced the
solution to form droplets (4 mL) at the end of each channel
(Fig. 1a).
We used commercially available Raman Systems R-2001
(Ocean Optics Inc.) which is a fully integrated Raman analyzer
for real-time qualitative and quantitative spectral analysis with
a 785 nm, 90 mW diode laser. Because of the large beam size of
the laser (about 200 mm in diameter), samples and SERS
substrates were not burned under 300 W/cm2 laser power. The
acquisition time was 10 sec. In order to confirm the reliablity and
the reproducibility of our SERS-CD platform, we measured
SERS signals from at least 5 different positions in a spot,
3 different spots in one CD and two different CD platforms
(totally, over 30 tests) for each investigation.
We explored SERS intensities by varying the concentrations of
GNPs at constant CuSO4 concentration. At lower concentrations, the CuSO4 aggregated GNPs is 1-dimensional, branchedstructure which is similar to fractal structure formed by kinetic
aggregation of GNPs.25 As the concentrations increases, GNPs
form multilayered aggregates (Fig. 3a). Since this morphological
transition of the aggregates causes a change in the plasmon
resonance of the aggregates, the SERS enhancement is expected
to be related with the extent of aggregation. SERS signal (1509
cm 1) of 500 nM R6G varies with the concentrations of GNPs
and is maximized at a surface coverage of about 40% of GNPs
(Fig. 3b). In order to address both the reliability and reproducibility of our SERS-CD platform, SERS signals were collected
from different positions, spots, and CDs, respectively. As shown
in Fig. 3b (error bars), it can be seen that the variation of SERS
signals in even different SERS-CD platforms was as small as
25%. Of course, the variation of SERS signals at different positions in the same spot (not different spots and CDs) was only
within 10% (data not shown). We also checked the stability of
our aggregated GNPs as a SERS-active site. After the aggregates
were fabricated at the end of each microchannel even though
water and R6G solutions were added in several times through the
secondary inlets, any significant changes of the morphology were
not found (as evident from SEM images).
We further evaluated SERS enhancement of the CuSO4induced aggregates by comparing the aggregates formed by
MgCl2. Since SERS spectra in this study were recorded from
‘‘dried’’ colloids/analyte spots, the effect of anions on SERS
signal can be excluded. The reason we compared the aggregation
properties of CuSO4 with MgCl2 is that our observation of
enhanced SERS intensity from CuSO4 may be attributed to
structural differences between them (not the effect of anions) and
a cation may play an important role in the process of aggregation. Fig. 4a shows the SEM images of the aggregates induced by
either CuSO4 or MgCl2 at the same ionic strength. Interestingly,
we found that CuSO4 generated the aggregates with higher
degree of branching and higher 2-dimensional surface coverage
than MgCl2. SERS spectra of R6G showed that the SERS
intensity at 1509 cm 1 was 5 to 6 times higher in the case of
CuSO4 than MgCl2, which might reflect the morphological
This journal is ª The Royal Society of Chemistry 2009
Fig. 3 Characterization of SERS substrates for the optofluidic SERSCD platform. (a) Representative SEM images of aggregate GNPs
induced by CuSO4 depending on GNPs concentrations. The coverage
values were calculated from the assumption that all GNPs in solution
cover the confined area with monolayer thickness. The inset shows the
corresponding morphology of GNPs aggregates. The scale bars in the
inset indicate 500 nm. (b) Relative SERS intensity of 500 nM R6G as
a function of the volume of GNPs. Changes in SERS signal mirror the
morphological transition of the GNPs aggregates. (*spot: 3.14 mm2
SERS-active site).
difference between them. The details on why Cu2+ is better than
other divalent ions to produce enhanced SERS intensity and
why morphologies of aggregates are quite different depending on
a cation at the same ionic strength are of great interest to us and
require a significant number of experiments; however, is outside
the scope of the current paper. We are currently investigating the
mechanism of aggregation by CuSO4 and its relation with
enhanced SERS intensity in great details.
4. Results and discussion
In order to demonstrate the feasibility of using this platform to
perform effective and high-throughput assays with many
different sample concentrations, the signal reproducibility and
SERS enhancement of the integrated platform were investigated.
The SERS spectra of R6G concentrations in the range from 1
nM to 1 mM were measured. Even though we applied R6G as
a representative, the other biomolecules and chemcials can be
used to our SERS-CD platform. For the other target samples,
drying temperature should be carefully controlled in order not to
lose their characteristics. Based on equation (1), we expected the
Lab Chip, 2009, 9, 239–243 | 241
Fig. 4 Effect of aggregating agents to SERS spectra. (a) SEM images of
the aggregate GNPs based on aggregation agents, CuSO4 and MgCl2.
Despite the same volume of GNPs is used, the coverage of MgCl2aggregated GNPs is less than half of that of CuSO4-aggregated GNPs.
The insets show that MgCl2 induces three dimensional structures of
GNPs rather than two dimensional shapes (The scale bars indicate 500
nm). (b) CuSO4-aggregated GNPs exhibit at least 5 larger SERS intensity
than that of MgCl2-aggregated GNPs. Inset shows the relative intensity
magnitude of the SERS signals at 1509 cm 1 peak.
SERS intensity would be enhanced by increasing the number of
R6G molecule. Therefore, for SERS enhancement, R6G was
preconcentrated via simply repeating ‘filling–drying’ cycles. We
injected R6G solution of 14 mL into each secondary inlet (thus
2 mL at each SERS active site). In order to uniformly distribute
the R6G molecules in a SERS-active site and minimize the
coffee-stain effect on the SERS sites, the platform was heated to
40 50 C by using a hot plate (Thermo Fisher Scientific Inc.).
Fig. 5 shows the SERS spectra of 500 nM R6G as a function of
‘filling–drying’ cycle. Minor background signals from the glass
substrate have been subtracted. As shown in Fig. 5a, the
observed Raman bands were in agreement with the bending and
stretching modes of R6G.26 As the cycle increases, the SERS
signals of R6G were significantly amplified. After 9 cycles, SERS
signals of 500 nM R6G were enhanced by 10-fold than that of
1 cycle. The SERS intensities at 1509 cm 1 for order-of-magnitude changes of R6G concentrations linearly augmented with the
number of ‘filling–drying’ cycle, which verifies high reliability of
our SERS-CD platform as well as the stability of the SERSactive sites (i.e., the aggregated GNPs). It is noteworthy that even
1 nM of R6G could be detected after 30 cycles. We can expect
that our SERS-CD platform is able to apply to much lower
sample concentaration. Considering that the required volume of
the sample per one cycle and one SERS-site is as low as 2 mL
(only 60 mL for even 30 cycles), our SERS-CD platform with
‘filling–drying’ preconcentration can be superior or comparable
to other SERS-based biomolecule detection systems because the
amount of a biological sample (both concentration and volume)
242 | Lab Chip, 2009, 9, 239–243
Fig. 5 SERS intensities with ‘filling–drying’ cycles. (a) SERS spectra of
500 nM R6G molecules as a function of the ‘filling–drying’ cycle by our
optofluidic SERS-CD platform. As the cycle is progressed, the SERS
signal enhancement was observed. (b) SERS signal at 1509 cm 1 of R6G
molecules with concentrations from 1 mM to 1 nM. Once the SERS signal
was observed, the signals increased linearly with the number of the cycle.
After 30 cycles, the clear and high SERS signal of R6G molecules with
1 nM was measured.
required for analysis is very critical in molecular biology and
medicine.3–5,8,9
Conclusions
In conclusion, we have fabricated an optofluidic SERS-CD
platform which is capable of high-throughput biological analysis
and additional SERS signal amplification via optofluidic CDbased preconcentration. By increasing the number of R6G
molecules via simple ‘filling–drying’ cycle, the SERS signal of
R6G linearly augmented, which suggests the high fidelity of
our SERS-CD platform. The experimental results ensure the
systematic enhancement characteristic of the SERS-CD platform
and validate the dependence of the SERS signal on the number
of molecule. In addition, our SERS-CD platform generated
84 identical SERS-active sites at one time via CuSO4-induced
GNPs aggregation. The GNP aggregates formed by CuSO4
showed little variations in SERS intensity from position-to-position, spot-to-spot, and batch-to-batch which verifies the reliability
and stability of our SERS sites. The SERS-CD platform has the
advantages of microfluidic devices such as high-throughput,
This journal is ª The Royal Society of Chemistry 2009
automation and low sample consumption. Furthermore, the
SERS-CD platform can be served as an ultrasensitive label-free
biomolecular detection and sensing devices. Hence, this study
shows the first demonstration of the SERS-CD platform for the
applications into various optofluidic multifunctional on-chip
biodevices.
Acknowledgements
This work was supported by a grant (Code No. HR0011-06-10050) from Micro/Nano Fluidics Fundamentals Focus (MF3)
center under DARPA S&T Fundamentals. The authors (D. C.
and T. K.) also appreciate the support by the Korea Research
Foundation Grant funded by the Korean Government
(MOEHRD) (KRF-2006-D00017 and KRF-2006-D00041).
Moreover, this work was carried out under the support by Intel
Corp. (Integrated Confocal Nanofluidic SERS for Proteomics)
and by a grant (Code No. 08K1501-02300) from Center for
Nanostructured Materials Technology under 21st Century
Frontier R&D Programs of the Ministry of Science and Technology, Korea.
References
1
2
3
4
5
6
C. V. Raman and K. S. Krishnan, Nature, 1928, 121, 501.
Y. Q. Chen, J. Kraut and R. Callender, Biophys. J., 1997, 72, 936.
I. Nabiev, I. Chourpa and M. Manfait, J. Raman Spectrosc., 1994, 25, 13.
H. Fabian and P. Anzenbacher, Vib. Spectrosc., 1993, 4, 125.
D. L. Jeanmaire and R. P. Van Duyne, J. Electroanal. Chem., 1977, 84, 1.
K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari and M. S. Feld, Chem.
Rev., 1999, 99, 2957.
This journal is ª The Royal Society of Chemistry 2009
7 S. Nie and S. R. Emory, Science, 1997, 275, 1102.
8 P. B. Monaghan, K. M. McCarney, A. Ricketts, R. E. Littleford,
F. Docherty, W. E. Smith, D. Graham and J. M. Cooper, Anal.
Chem., 2007, 79, 2844.
9 K. R. Strehle, D. Cialla, P. Rösch, T. Henkel, M. Köhler and J. Popp,
Anal. Chem., 2007, 79, 1542.
10 M. Wang, N. Jing, I-H. Chou, G. L. Cote and J. Kameoka, Lab Chip,
2007, 7, 630.
11 S. Chan, S. Kwon, T. W. Koo, L. P. Lee and A. A. Berlin, Adv.
Mater., 2003, 15, 1595.
12 G. L. Liu and L. P. Lee, Appl. Phys. Lett., 2005, 87, 074101.
13 Y. Lu, G. L. Liu and L. P. Lee, Nano Lett., 2005, 5, 5.
14 R. G. Freeman, K. C. Grabar, K. J. Allison, R. M. Bright,
J. A. Davis, A. P. Guthrie, M. B. Hommer, M. A. Jackson,
P. C. Smith, D. G. Walter and M. J. Natan, Science, 1995, 267, 1629.
15 R. G. Freeman, M. B. Hommer, K. C. Grabar, M. A. Jackson and
M. J. Natan, J. Phys. Chem., 1996, 100, 718.
16 A. Wei, B. Kim, B. Sadtler and S. L. Tripp, Chem. Phys. Chem., 2001,
12, 743.
17 I-H. Chou, M. Benford, H. T. Beier and G. L. Coté, Nano Lett., 2008,
8, 1729.
18 K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan,
R. R. Dasari and M. S. Feld, Phys. Rev. Lett., 1997, 78, 1667.
19 H. Xu, E. J. Bjerneld, M. Kall and L. Borjesson, Phys. Rev. Lett.,
1999, 83, 4357.
20 J. B. Jackson and N. J. Halas, PNAS, 2004, 101, 17930.
21 Y. Lu, G. L. Liu, J. Kim, Y. X. Mejia and L. P. Lee, Nano Lett., 2005,
5, 119.
22 G. L. Liu, Y. Lu, J. Kim, J. C. Doll and L. P. Lee, Adv. Mater., 2005,
17, 2683.
23 H-H. Wang, C-Y. Liu, S-B. Wu, N-W. Liu, C-Y. Peng, T-H. Chan, CF. Hsu, J-K. Wang and Y-L. Wang, Adv. Mater., 2006, 18, 491.
24 S. E. J. Bell and N. M. S. Sirimuthu, J. Am. Chem. Soc., 2006, 128,
15580.
25 D. A. Weitz and M. Oliveria, Phys. Rev. Lett., 1984, 52, 1433.
26 P. Hildebrandt and M. Stockburger, J. Phys. Chem., 1984, 88,
5935.
Lab Chip, 2009, 9, 239–243 | 243