Ultrastable substrates for surface-enhanced Raman spectroscopy: Al 2O3 overlayers fabricated by atomic layer deposition yield improved anthrax biomarker detection

Article abstract of DOI:10.1021/ja0638760

A new method to stabilize and functionalize surfaces for surface-enhanced Raman spectroscopy (SERS) is demonstrated. Atomic layer deposition (ALD) is used to deposit a sub-1-nm alumina layer on silver film-over-nanosphere (AgFON) substrates. The resulting overlayer maintains and stabilizes the SERS activity of the underlying silver while presenting the surface chemistry of the alumina overlayer, a commonly used polar adsorbent in chromatographic separations. The relative affinity of analytes for alumina-modified AgFON substrates can be determined by their polarity. On the basis of SERS measurements, dipicolinic acid displays the strongest binding to the ALD alumina-modified AgFON among a set of pyridine derivatives with varying polarity. This strong affinity for carboxylate groups makes the SERS substrate an ideal candidate for bacillus spores detection using the dipicolinate biomarker. The SERS signal from extracted dipicolinate was measured over the spore concentration range 10 ^-14-10^-12 M to determine the saturation binding capacity of the alumina-modified AgFON surface. The adsorption constant was determined to be K_spore = 9.0 × 10¹³ M^-1. A 10-s data collection time is capable of achieving a limit of detection of ～1.4 × 10³ spores. The shelf life of prefabricated substrates is at least 9 months prior to use. In comparison to the bare AgFON substrates, the ALD-modified AgFON substrates demonstrate twice the sensitivity with 6 times shorter data acquisition time and 7 times longer temporal stability. ALD expands the palette of available chemical methods to functionalize SERS substrates, which will enable improved and diverse chemical control over the nature of analyte-surface binding for biomedical, homeland security, and environmental applications.

Full text of DOI:10.1021/ja0638760

Published on Web 07/19/2006
Ultrastable Substrates for Surface-Enhanced Raman
Spectroscopy: Al2O3 Overlayers Fabricated by Atomic Layer
Deposition Yield Improved Anthrax Biomarker Detection
Xiaoyu Zhang,† Jing Zhao,† Alyson V. Whitney,† Jeffrey W. Elam,‡ and
Richard P. Van Duyne*,†
Contribution from the Department of Chemistry, Northwestern UniVersity, 2145 Sheridan Road,
EVanston, Illinois 60208, and Energy Systems and Materials Science DiVisions,
Argonne National Laboratory, 9700 South Cass AVenue, Argonne, Illinois 60439
Received June 2, 2006; E-mail: vanduyne@chem.northwestern.edu
Abstract: A new method to stabilize and functionalize surfaces for surface-enhanced Raman spectroscopy
(SERS) is demonstrated. Atomic layer deposition (ALD) is used to deposit a sub-1-nm alumina layer on
silver film-over-nanosphere (AgFON) substrates. The resulting overlayer maintains and stabilizes the SERS
activity of the underlying silver while presenting the surface chemistry of the alumina overlayer, a commonly
used polar adsorbent in chromatographic separations. The relative affinity of analytes for alumina-modified
AgFON substrates can be determined by their polarity. On the basis of SERS measurements, dipicolinic
acid displays the strongest binding to the ALD alumina-modified AgFON among a set of pyridine derivatives
with varying polarity. This strong affinity for carboxylate groups makes the SERS substrate an ideal candidate
for bacillus spores detection using the dipicolinate biomarker. The SERS signal from extracted dipicolinate
was measured over the spore concentration range 10-14-10-12 M to determine the saturation binding
capacity of the alumina-modified AgFON surface. The adsorption constant was determined to be Kspore
)
9.0 × 1013 M-1. A 10-s data collection time is capable of achieving a limit of detection of ∼1.4 × 103
spores. The shelf life of prefabricated substrates is at least 9 months prior to use. In comparison to the
bare AgFON substrates, the ALD-modified AgFON substrates demonstrate twice the sensitivity with 6 times
shorter data acquisition time and 7 times longer temporal stability. ALD expands the palette of available
chemical methods to functionalize SERS substrates, which will enable improved and diverse chemical
control over the nature of analyte-surface binding for biomedical, homeland security, and environmental
applications.
Introduction
analytical tool has been hindered by the lack of SERS substrate
stability. Generally, SERS activity is affected by the oxidation
Surface-enhanced Raman scattering occurs when a molecule
is spatially confined within the zone of enhanced electromag-
netic fields generated by excitation of a localized surface
plasmon resonance (LSPR).1 Consequently, the magnitude of
the induced dipole moment of the molecule increases, and
accordingly, the intensity of the inelastic scattering increases.
Surface-enhanced Raman spectroscopy (SERS) has many
characteristics that can be exploited in biosensing applications,2-6
such as sensitivity, selectivity, and no interference from water
molecules. However, the acceptance of SERS as a general
of silver or the aggregation of noble metal colloidal nanostruc-
tures.7-9 In this work, we demonstrate a simple strategy for
dramatically improving the stability of traditional SERS sub-
strates. An ultrathin alumina layer was coated onto silver film-
over-nanosphere (AgFON) substrate using atomic layer depos-
tion (ALD). ALD utilizes self-limiting surface reactions to
control interfacial thickness and composition with molecular
precision.10 Previous quartz crystal microbalance measurements
have demonstrated highly uniform layer-by-layer growth of the
ALD alumina on Ag nanoparticles with a growth rate of ∼1 Å
per deposition.10 The sub-1-nm thickness is extremely advanta-
geous in preserving sensitivity, because the SERS intensity
decays by approximately 1 order of magnitude for each 2.8 nm
separation between the surface and the scatterer.11 The details
† Northwestern University.
‡ Argonne National Laboratory.
(1) Schatz, G. C.; Van Duyne, R. P. Electromagnetic Mechanism of Surface-
Enhanced Spectroscopy. In Handbook of Vibrational Spectroscopy; Chalm-
ers, J. M., Griffiths, P. R., Eds.; Wiley: New York, 2002; Vol. 1, p 759.
(2) Zhang, X.; Young, M. A.; Lyandres, O.; Van Duyne, R. P. J. Am. Chem.
Soc. 2005, 127, 4484.
(3) Yonzon, C. R.; Haynes, C. L.; Zhang, X. Y.; Walsh, J. T.; Van Duyne, R.
P. Anal. Chem. 2004, 76, 78.
(7) Von Raben, K. U.; Chang, R. K.; Laube, B. L.; Barber, P. W. J. Phys.
Chem. 1984, 88, 5290.
(8) Fang, J. H.; Huang, Y. X.; Li, X.; Dou, X. M. J. Raman Spectrosc. 2004,
35, 914.
(9) Fornasiero, D.; Grieser, F. J. Chem. Phys. 1987, 87, 3213.
(10) Whitney, A. V.; Elam, J. W.; Zou, S. L.; Zinovev, A. V.; Stair, P. C.;
Schatz, G. C.; Van Duyne, R. P. J. Phys. Chem. B 2005, 109, 20522.
(4) Daniels, J. K.; Caldwell, T. P.; Christensen, K. A.; Chumanov, G. Anal.
Chem. 2006, 78, 1724.
(5) Kneipp, J.; Kneipp, H.; Rice, W. L.; Kneipp, K. Anal. Chem. 2005, 77,
2381.
(6) Driskell, J. D.; Kwarta, K. M.; Lipert, R. J.; Porter, M. D.; Neill, J. D.;
Ridpath, J. F. Anal. Chem. 2005, 77, 6147.
9
10304
J. AM. CHEM. SOC. 2006, 128, 10304-10309
10.1021/ja0638760 CCC: $33.50 © 2006 American Chemical Society

Ultrastable Substrates for SERS
A R T I C L E S
of the intensity decay function are determined by the nano-
structure of the underlying silver surface.
chemistry behavior of the ALD alumina-modified AgFON
surfaces based on SEM, LSPR, and SERS results, and (2) the
applicability and efficiency of these substrates as a bacillus spore
sensing platform.
The use of ALD alumina presents several advantages. First,
compared to conventional overlayer materials, the ultrathin
alumina layer is extremely stable to oxidation and high
temperature.12 This helps to maintain the high stability of SERS
activity with minimal decrease in signal. Second, alumina is
commonly used as a polar adsorbent in chromatographic
separations. The relative affinity between Raman scatterers and
alumina-modified AgFON substrates can be predicted on the
basis of their polar interaction, which has been well established
in the chromatography literature. Generally, molecules with
strong polarity, such as carboxylic acids, have high affinity to
alumina.13,14 Therefore, this novel SERS substrate is an ideal
candidate for the detection of carboxylic acids due to the strong
polar interaction. Third, the scope of analytical applications of
SERS has been broadened by modifying noble metal surfaces
with an analyte-specific affinity coating.3,6,15 The coatings used
range from simple alkanes15 to complex macrocycles with the
common theme of containing a thiol group to anchor the coating
to a noble metal substrate. Large partition coefficients on the
coating allow analytes to partition closer to the surface.15
However, the high coverage of the thiolate self-assembled
monolayers (SAMs) is thermodynamically unstable.16 Thermal
desorption17,18 and photooxidation19-21 of the thiolate molecules
result in defects in the coating. In comparison to the previously
used thiolate SAMs, ALD alumina enjoys greater molecular
thickness control, greater physical and chemical stability, more
complete surface coverage, less signal attenuation due to
distance effects, and predictable affinity.
In a previous study, the optically optimized AgFON substrate
had been applied to quantitatively detect a biomarker for anthrax,
calcium dipicolinate (CaDPA), from bacillus spores.2 A limit
of detection (LOD) of ∼2550 anthrax spores was achieved on
the AgFON sensor with a data acquisition period of 1 min and
a laser power of 50 mW. In this study, we have increased the
affinity between CaDPA and the sensor surface using ALD
alumina, thus increasing the SERS signal intensities. The
alumina-modified sensor shows 2-fold improvement in LOD
using 6 times shorter data acquisition time. Lifetime testing
measurements indicate that the SERS intensity is stable on
alumina-coated AgFONs for at least 9 months. These results
indicate that this new SERS sensor has the potential to be
extremely useful for biomedical, homeland security, and envi-
ronmental measurements.
Experimental Section
Materials. Ag (99.99%) was purchased from D. F. Goldsmith
(Evanston, IL). Glass substrates were 18 mm diameter, No. 2 cover
slips from Fisher Scientific (Pittsburgh, PA). Surfactant-free white
carboxyl-functionalized polystyrene latex nanospheres with diameters
of 590 nm were obtained from Interfacial Dynamics Corp. (Portland,
OR). Tungsten vapor deposition boats were purchased from R. D.
Mathis (Long Beach, CA). Water (18.2 MΩ‚cm) was obtained from
an ultrafilter system (Milli-Q, Millipore, Marlborough, MA). All the
other chemicals, reagents, and solvents were purchased from Aldrich
Chemical (Milwaukee, WI) or Fisher Scientic (Fairlawn, NJ) and used
without further purification.
Thin-layer chromatographic (TLC) analyses were performed on
aluminum-backed aluminum oxide 60 F-254 neutral with a 0.2 mm
layer thickness (type E, E. Merck, Darmstadt, Germany) in 10:1 v/v
hexanes:ethyl acetate.
Calcium dipicolinate (CaDPA) was prepared from DPA and calcium
hydroxide according to the method of Bailey and co-workers.22
Bacillus subtilis spore samples were prepared according to the
previously published method.2 Approximately 1 g of sample was
determined to contain 5.6 × 1010 spores by optical microscope
measurements (data not shown). The spore suspension was made by
dissolving spores in 0.02 M HNO3 solution and sonicating for 10 min,
which effectively extracts CaDPA from spores. This concentration of
the HNO3 solution was selected because of its capability for CaDPA
extraction and benign effect on the AgFON SERS activity. The
sonication procedure was performed because no SERS signal of CaDPA
was observed from the spore solution prior to sonication (data not
shown).
AgFON Substrate Fabrication. Glass substrates were pretreated
in two steps: (1) piranha etch (CAUTION: piranha solution should
be handled with great care), in which 3:1 H2SO4:30% H2O2 at 80 °C
for 1 h was used to clean the substrate, and (2) base treatment, in which
5:1:1 H2O:NH4OH:30% H2O2 with sonication for 1 h was used to render
the surface hydrophilic. Approximately 2 µL of the nanosphere
suspension (4% solids) was drop-coated onto each substrate and allowed
to dry in ambient conditions. The metal films were deposited in a
modified Consolidated Vacuum Corp. vapor deposition system with a
base pressure of 10-6 Torr. The deposition rates for each film (10 Å/s)
were measured using a Leybold Inficon XTM/2 quartz crystal mi-
crobalance (QCM) (East Syracuse, NY). AgFON substrates were stored
in the dark at room temperature prior to use.
Atomic Layer Deposition (ALD). Alumina films were fabricated
on the AgFON substrates by ALD. The reactor utilized in these
experiments is similar to that used in previous publications.10 Tri-
methylaluminum (TMA) and deionized H2O vapors were alternately
pulsed through the reaction chamber, utilizing N2 as the carrier gas, at
a mass flow rate of 360 sccm, a pressure of 1 Torr, and a growth
temperature of 50 °C. One complete ALD cycle takes 42 s and includes
four steps: (1) TMA reactant exposure time, 1 s; (2) N2 purge following
TMA exposure time, 10 s; (3) H2O reactant exposure time, 1 s; and
(4) N2 purge following H2O exposure time, 30 s. Long purge times are
necessary at low temperatures to prevent chemical vapor deposition of
alumina.23,24 A previous study indicated nearly ideal layer-by-layer
growth of the ALD alumina on Ag surfaces with an average rate ∼1
Å/cycle.10 This result greatly simplifies the interpretation of the
To frame the achievements reported herein, this paper
addresses (1) the physical characteristics and the surface
(11) Dieringer, J. A.; McFarland, A. D.; Shah, N. C.; Stuart, D. A.; Whitney,
A. V.; Yonzon, C. R.; Young, M. A.; Zhang, X. Y.; Van Duyne, R. P.
Faraday Discuss. 2006, 132, 9.
(12) King, F. In Aluminum and its alloys; West, E. G., Ed.; Ellis Horwood:
New York, 1987; p 313.
(13) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 45.
(14) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 52.
(15) Carron, K.; Peitersen, L.; Lewis, M. EnViron. Sci. Technol. 1992, 26, 1950.
(16) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M.
Chem. ReV. 2005, 105, 1103.
(17) Ishida, T.; Hara, M.; Kojima, I.; Tsuneda, S.; Nishida, N.; Sasabe, H.; Knoll,
W. Langmuir 1998, 14, 2092.
(18) Zhang, Z. S.; Wilson, O. M.; Efremov, M. Y.; Olson, E. A.; Braun, P. V.;
Senaratne, W.; Ober, C. K.; Zhang, M.; Allen, L. H. Appl. Phys. Lett. 2004,
84, 5198.
(19) Lewis, M.; Tarlov, M.; Carron, K. J. Am. Chem. Soc. 1995, 117, 9574.
(20) Schoenfisch, M. H.; Pemberton, J. E. J. Am. Chem. Soc. 1998, 120, 4502.
(21) Zhang, Y. M.; Terrill, R. H.; Tanzer, T. A.; Bohn, P. W. J. Am. Chem.
Soc. 1998, 120, 2654.
(22) Bailey, G. F.; Karp, S.; Sacks, L. E. J. Bacteriol. 1965, 89, 984.
(23) Elam, J. W.; Groner, M. D.; George, S. M. ReV. Sci. Instrum. 2002, 73,
2981.
(24) Groner, M. D.; Fabreguette, F. H.; Elam, J. W.; George, S. M. Chem. Mater.
2004, 16, 639.
9
J. AM. CHEM. SOC. VOL. 128, NO. 31, 2006 10305

A R T I C L E S
Zhang et al.
Figure 2. LSPR reflectance spectra of bare AgFON and alumina-modified
AgFON substrates with different alumina thickness. The ALD cycles of
alumina varied from 1 to 3; D ) 600 nm, dm ) 200 nm. The vertical dotted
line denotes the laser excitation used in the SERS measurements. All
reflectance spectra were collected against a mirror-like Ag film over glass
surface as a reference in air.
(λex) slightly to the blue of the LSPR maximum wavelength
for adsorbate-covered nanoparticle arrays.25 Additionaly, similar
observations have been reported based on plasmon-sampled
surface-enhanced Raman excitation spectroscopy.2,26 This find-
ing yields a general rule for maximizing SERS signals for
chemical and biological detection.
Figure 1. Scanning electron microscope images of alumina-modified
AgFON substrates (D ) 600 nm, dm ) 200 nm, and two ALD cycles of
alumina).
The plasmon position of FONs can be tuned by controlling
the size of the spheres used in the underlying nanosphere mask.
Generally, an increase in polystyrene sphere diameters results
in a red shift in LSPR.2 Here, polystyrene spheres with a
diameter of 590 nm were selected to fabricate SERS substrates
with the LSPR in the near-infrared (NIR) range.2 The resulting
substrates are optimized for the SERS measurements using NIR
laser excitation. NIR laser excitation reduces not only the
interference from biological fluorescent background but also
the potential for tissue damage due to the low water and tissue
absorption of NIR light.
Since AgFONs are not optically transparent, the reflectivity
minimum (λmin) was used to locate the LSPR positions. Figure
2 shows the LSPR reflectance spectra of AgFON coated with
0-3 ALD cycles. For reference, the vertical dotted line in Figure
2 denotes the 785 nm laser excitation used in the following
SERS measurements. The LSPR λmin positions of AgFON
coated with 0-3 ALD cycles remain close to the SERS
excitation wavelength, ranging between 774 and 783 nm, and
are not greatly affected by coating alumina. This result agrees
with a more detailed and rigorous study on NSL-produced Ag
nanoparticles with alumina layers.10
thickness of the alumina overlayers, which can be deduced easily from
the number of ALD cycles.
Scanning Electron Microscopy (SEM). SEM images of alumina-
coated AgFON were observed with a Hitatchi S-4700-II SEM.
LSPR Reflectance Spectroscopy. Reflectance measurements were
carried out using a SD2000 spectrometer coupled to a reflection probe
(Ocean Optics, Dunedin, FL) and a halogen lamp (F-O-Lite H, World
Precision Instruments, Sarasota, FL). The reflection probe consists of
a tight bundle of 13 optical fibers (12 illumination fibers around 1
collection fiber), with a usable wavelength range of 400-900 nm. All
reflectance spectra were collected against a mirror-like Ag film over a
glass surface as a reference in air.
SERS Apparatus. The macro-Raman system consists of an interfer-
ence band-pass filter (Coherent, Santa Clara, CA), a 1-in. long-pass
dielectric edge filter (Sermrock, Rochester, NY), a single-grating
monochromator with the entrance slit set at 100 µm (model VM-505,
Acton Research Corp., Acton, MA), a liquid-N2-cooled CCD detector
(model Spec-10:400B, Roper Scientific, Trenton, NJ), and a data
acquisition system (Photometrics, Tucson, AZ). A compact diode laser
(model RL785, Renishaw plc, Gloucestershire, UK) was used to
generate 785 nm. All the measurements were performed in ambient
conditions.
3. Competitive Adsorption on Alumina-Modified AgFON.
In the analysis of biologically, industrially, and environmentally
relevant samples, multiple analytes with similar molecular
structures are often in competition for the SERS surface. In that
situation, the detection selectivity is improved when the affinity
between the target analyte and the sensing platform is optimized.
Central to the work reported here is our ability to predict the
relative affinity of different molecules on an alumina-modified
AgFON substrate. As a proof-of-concept experiment, we
examine the competitive adsorption of dipicolinic acid (a),
diacetylpyridine (b), and dimethoxypyridine (c) onto an alumina-
modified AgFON surface at room temperature (Figure 3A).
Result and Discussion
1. Structure Characterizations of Alumina-Modified Ag-
FON. Figure 1 shows a SEM image of an alumina-modified
AgFON substrate with a nanosphere diameter (D) of 590 nm,
Ag mass thickness (dm) of 200 nm, and two ALD cycles of
alumina. The low-magnification SEM image (Figure 1) displays
a region of hexagonally packed polystyrene spheres. The inset
in Figure 1, a high-magnification image, shows that the top of
each nanosphere is not smooth but exhibits substructure
roughness features 30-50 nm in size.
2. Alumina Thickness Effect on LSPR. Previously, a
detailed wavelength-scanned SERS study of benzenethiol ad-
sorbed on Ag nanoparticle arrays revealed that the maximum
SERS enhancement factor occurs for excitation wavelengths
(25) McFarland, A. D.; Young, M. A.; Dieringer, J. A.; Van Duyne, R. P. J.
Phys. Chem. B 2005, 109, 11279.
(26) Haynes, C. L.; Van Duyne, R. P. J. Phys. Chem. B 2003, 107, 7426.
9
10306 J. AM. CHEM. SOC. VOL. 128, NO. 31, 2006

Ultrastable Substrates for SERS
A R T I C L E S
Figure 4. (A) SERS spectrum of 2 × 10-5 M CaDPA in 0.2 µL of 0.02
M HNO3 on alumina-modified AgFON substrate (two ALD cycles of
alumina). (B) SERS spectrum of CaDPA on bare AgFON substrate. (C)
The alumina thickness effect on the SERS intensity. The normalized spectral
intensities at 1020 cm-1, I1020, were plotted versus ALD cycles. I1020 was
taken from SERS spectra that correspond to varying alumina thickness on
the AgFON substrates. Each data point represents the average value from
three SERS samples. Error bars show the standard deviations. Laser
excitation, 785 nm; laser power, 50 mW; acquisition time, 10 s; D ) 600
nm; and dm ) 200 nm. * denotes adu/(s‚mW).
4. Bacillus Spore Detection Using Alumina-Modified
AgFON. In our previous study, CaDPA was efficiently extracted
from spores by sonication in 0.02 M HNO3, deposited onto
AgFON substrates, and then detected by SERS.2 In this work,
we follow the same extraction protocol but use ALD alumina-
modified AgFON sensors. The results presented below detail a
substantial advance toward the development of a practical SERS-
based anthrax detection device. The performance of the SERS
sensor in terms of sensitivity and temporal stability will be
discussed.
Figure 3. (A) Notations of dipicolinic acid (a), diacetylpyridine (b), and
dimethoxypyridine (c). (B) TLC chromatogram showing the separation of
a, b, and c. Eluent, 10:1 hexanes:ethyl acetate. Spots were detected by
placing the TLC plate in iodine vapor. (C) SERS spectra of 2 mM a, b,
and c on alumina-modified AgFON substrates. Reference spectra of
individual pyridine derivatives are shown along with the spectra of the
equimolar mixtures. Laser excitation, 785 nm; laser power, 50 mW;
acquisition time, 30 s; D ) 590 nm; dm ) 200 nm; and two ALD cycles of
alumina. The gap in spectral coverage results from changing the grating in
the spectrometer.
4.1. Effect of Alumina Thickness on SERS Detection for
CaDPA. CaDPA was selected as the probe molecule to test
the effect of alumina thickness on SERS detection for bacillus
spores. Figure 4A demonstrates a SERS spectrum of 2 × 10-5
M CaDPA on a AgFON substrate (D ) 590 nm, dm ) 200
nm). The SERS bands at 1020 and 812 cm-1 are associated
with CaDPA, in agreement with the previous Raman studies.
The peak at 1050 cm-1 is from the symmetrical stretching
vibration of NO3- from nitric acid, which is used as an internal
standard to reduce the sample-to-sample deviations. For com-
parison, a parallel SERS experiment was conducted on an
alumina-modified AgFON substrate (D ) 590 nm, dm ) 200
nm, two ALD cycles). The SERS spectrum obtained on the
modified substrate has the same spectral patterns but higher
intensities. A plot of the normalized SERS intensity at 1020
cm-1 as a function of ALD cycles is shown in Figure 4C. Each
data point represents the average intensity at 1020 cm-1 from
three samples, with the standard deviation shown by the error
bars. Interestingly, the SERS intensity of CaDPA achieves a
maximum when the AgFON substrates are modified with two
ALD cycles of alumina. The intensity then decreases with further
addition of alumina.
These pyridine derivative analytes serve as a model system
because they provide the unambiguous SERS spectral signatures
important to differentiate between the relative adsorption
affinities observed during competitive adsorption experiments.
On the basis of a thin-layer chromatography (TLC) experiment
(Figure 3B), the relative affinity between each analyte and
alumina is determined to be a > b> c. Therefore, we expect
that the SERS spectra of mixed analytes on the alumina-
modified AgFON substrates would be dominated by the
strongest adsorbate a. To verify this hypothesis, a series of SERS
measurement were made. We produced the SERS samples by
incubating the alumina-modified AgFON substrates in 2 mM
analyte in ethanol solutions for 5-6 h. Adsorption of each
individual analyte produces a specific SERS signature (Figure
3C). It is worth noting that high signal-to-noise SERS spectra
were obtained on alumina-modified AgFON substrates, even
though the adsorbates were not in direct contact with the silver
surface. The spectra can be compared with those from equimolar
mixtures of pairs of a+b, a+c and b+c. For mixtures that
contain a, the spectra are dominated by peaks characteristic of
a at 1458, 1390, 1042, and 1018 cm-1. Similarly, the spectrum
for b+c combinations is dominated by signatures of b. From
these results, it is clear that a has the highest adsorption affinity
on alumina-modified AgFON surfaces and c has the lowest,
which is consistent with anticipation in light of the TLC results.
The high affinity of dipicolinic acid has numerous practical
implications. As an example, alumina-modified AgFON surfaces
are ideal sensing platforms for calcium dipicolinate, a well-
known biomarker for bacillus spores,22 because the increase in
affinity improves the ultimate LOD.
This result requires special consideration, given that the
surface-enhanced Raman spectral intensities decrease with the
addition of spacers (e.g., a polymer film, cold-condensed
molecular layer, or thiolate SAM) between the Raman scatterer
and the metal surface.27-29 The SERS intensity is essentially
dependent on both the electromagnetic field and the number of
(27) Otto, A.; Mrozek, I.; Grabhorn, H.; Akemann, W. J. Phys.: Condens. Matter
1992, 4, 1143.
(28) Kovacs, G. J.; Loutfy, R. O.; Vincett, P. S.; Jennings, C.; Aroca, R.
Langmuir 1986, 2, 689.
(29) Ye, Q.; Fang, J. X.; Sun, L. J. Phys. Chem. B 1997, 101, 8221.
9
J. AM. CHEM. SOC. VOL. 128, NO. 31, 2006 10307

A R T I C L E S
Zhang et al.
Figure 5. (A) Adsorption isotherm for CaDPA onto alumina-modified
AgFON substrates. I1020 was taken from SERS spectra that correspond to
varying CaDPA concentrations in 0.2 µL of 0.02 M HNO3 on alumina-
modified AgFON substrates. Laser excitation, 785 nm; laser power, 50 mW;
acquisition time, 10 s; D ) 600 nm; dm ) 200 nm; and two ALD cycles of
alumina. The inset shows the linear range that is used to determine the
LOD. Each data point represents the average value from three SERS spectra.
Error bars show the standard deviations. (B) Adsorption data of CaDPA fit
with the linear form of the Langmuir model, eq 2. The slope and intercept
values are used to calculate the adsorption constant.
Figure 6. Adsorption isotherm for B. subtilis spore suspension on alumina-
modified AgFON substrates. I1020 was taken from SERS spectra that
correspond to varying spore concentrations in 0.2 µL of 0.02 M HNO3 on
the substrates. The inset shows the linear range that is used to determine
the LOD. Each data point represents the average value from three SERS
samples. Error bars show the standard deviations. Laser excitation, 785 nm;
laser power, 50 mW; acquisition time, 10 s; D ) 600 nm; dm ) 200 nm;
and two ALD cycles of alumina.
adsorbed molecules.1 According to previous SERS distance
dependence studies, the electromagnetic field decays rapidly
when the distance between the scatterers and the silver nano-
structures increases. The origin of the increase in overall SERS
intensity (Figure 4C) likely comes from an increase in the
number of scattering molecules. In other words, the adsorption
affinity of CaDPA to alumina is anticipated to be greater than
that of CaDPA to silver. To have a clearer picture of this affinity
difference, the adsorption constants must be measured.
4.2. Adsorption Isotherm and LOD for CaDPA on
Alumina-Modified AgFON Substrates. The quantitative re-
lationship between SERS signal intensity and CaDPA concen-
tration is demonstrated in Figure 5. At low concentrations, the
peak intensity increases linearly with concentration. In this study,
the LOD is defined as the concentration of CaDPA for which
the strongest SERS signal of CaDPA at 1020 cm-1 is equal to
3 times the background SERS signal for a 10-s acquisition
period and 50 mW laser power. The background signal refers
to the SERS intensity from a sample with a CaDPA concentra-
tion equal to zero, which is calculated to be the intercept of the
low-concentration end of the adsorption isotherm (Figure 5A).
The LOD for CaDPA, evaluated by extrapolation of the linear
concentration range of the adsorption isotherms (Figure 5A
inset), is found to be 1.9 × 10-6 M (in 0.2 µL of 0.02 M HNO3).
At higher CaDPA concentrations, the response saturates as
the adsorption sites on the alumina-modified AgFON substrate
become fully occupied. Saturation occurs when CaDPA con-
centrations exceed ∼1.5 × 10-4 M (in 0.2 µL of 0.02 M HNO3).
To determine the adsorption capacity of extracted CaDPA on
an alumina-modified AgFON, the Langmuir adsorption isotherm
was used to fit the data:
of CaDPA (M), and KCaDPA is the adsorption constant of CaDPA
to alumina-modified AgFON (M-1). From eq 2, KCaDPA is
calculated from the ratio between the intercept and the slope.
Slope and intercept analyses of the linear fit (Figure 5B) lead
to the value of the adsorption constant KCaDPA ) 4.9 × 104
M-1
.
Previous SERS studies on bare AgFON substrates indicated
that the adsorption constant for CaDPA was 9.0 × 103 M-1
and the LOD was 3.1 × 10-6 M (laser excitation, 750 nm; laser
power, 50 mW; acquisition time, 60 s).2 The affinity between
CaDPA and alumina-modified AgFON is ∼5 times stronger than
that of bare AgFON. As a result of the change in surface
chemistry, the addition of alumina improves the LOD of
CaDPA.
4.3. Adsorption Isotherm and LOD for Bacillus Spores
on Alumina-Modified AgFON Substrates. The SERS spectra
of a bacillus spore suspension are dominated by CaDPA on
alumina-modified AgFON. Illustrated in Figure 6, the parallel
studies of SERS intensities at 1020 cm-1 versus spore concen-
trations indicate that the LOD is 1.4 × 103 spores in 0.2 µL of
0.02 M HNO3 and the adsorption constant for CaDPA extracted
from spores, Kspore, is 9.0 × 1013 M-1. In contrast, the adsorption
constant was 1.7 × 1013 M-1 for extracted CaDPA on bare
AgFON surface and the LOD was 2.6 × 103 spores (laser
excitation, 750 nm; laser power, 50 mW; acquisition time, 60
s).2
4.4. Temporal Stability of Alumina-Modified AgFON. An
ideal sensor should be stable for long periods of time and require
infrequent maintenance. In this work, the temporal stability of
alumina-modified AgFON substrates (D ) 590 nm, dm ) 200
nm, two ALD cycles) was studied over a period of 9 months.
The alumina-modified AgFON substrates were stored in Petri
I1020
K
CaDPA[CaDPA]
θ )
)
(1)
(2)
I1020,max
1 + KCaDPA[CaDPA]
dishes in the dark prior to use. SERS spectra of 3.0 × 10-14
M
1
1
I
1
1
)
+
spores (3.6 × 103 spores in 0.2 µL of 0.02 M HNO3) were
captured on alumina-modified AgFON substrates of different
ages. Figure 7 shows a representative SERS spectrum on a
9-month-old alumina-modified AgFON substrate prior to use.
The intensity ratios between the strongest CaDPA peak at 1020
I1020
K
I1020,max
CaDPA 1020,max [CaDPA]
where θ is the coverage of CaDPA on the alumina-modified
AgFON, I1020,max is the maximum SERS signal intensity at 1020
cm-1 when all the SERS active sites on alumina-modified
AgFON are occupied by CaDPA, [CaDPA] is the concentration
cm-1 and the NO3 peak at 1050 cm-1 (I1020/I1050) were
-
9
10308 J. AM. CHEM. SOC. VOL. 128, NO. 31, 2006

Ultrastable Substrates for SERS
A R T I C L E S
increase due to the greater binding affinity of CaDPA toward
alumina compared to silver and the intensity decrease due to
the decay of the electromagnetic field around silver nanostruc-
tures is demonstrated.
This work also further details our improvements in detecting
bacillus spores using SERS. Most notably, the sensor system is
now more sensitive and maintains its performance over a long
period of 9 months. The SERS signal of extracted CaDPA was
measured over the concentration range 10-14-10-12 M to
determine the saturation binding capacity of the alumina-
modified AgFON surface and calculate the adsorption constant
(KCaDPA ) 4.9 × 104 M-1). At present, a 10-s data collection
time is capable of achieving a LOD of ∼1.9 × 10-6 M using
785 nm laser excitation. For bacillus spores on alumina-modified
AgFON, the adsorption constant of extracted CaDPA was
determined to be 9.0 × 1013 M-1 and the LOD 1.4 × 103 spores.
This experiment demonstrates two very important future
trends in SERS. First, the addition of the alumina layer imparts
a new chemical functionality to the SERS active surface. We
will expand our palette of available ALD layer materials, which
will enable greater chemical control over the nature of the
selectively adsorbed analytes. Second, the novel SERS substrates
reported herein hold enormous promise for combining both TLC
and SERS without extra deposition steps of silver colloids or
positively charged polymer.30,31 This combination, with the
development of miniaturized Raman spectrometers and nano-
fabrication methods, will allow for effective separation and
sensitive identification of components in a mixture analyte,
which is an important task for the development of chip-based
chemical and biological detection systems.
Figure 7. SERS spectrum of 3.0 × 10-14 M spores (3.6 × 103 spores in
0.2 µL of 0.02 M HNO3) on a 9-month-old alumina-modified AgFON
substrate. The inset shows the intensity ratio (I1020/I1050) variation with time.
Laser excitation, 785 nm; laser power, 50 mW; acquisition time, 10 s; D )
590 nm; dm ) 200 nm; and two ALD cycles of alumina.
measured to quantitatively compare the substrates of different
ages (shown in Figure 7 inset). The temporal stability of
alumina-modified AgFON substrates is evident in the fact that
the SERS intensities of extracted CaDPA remained constant over
the course of 9 months. The excellent long-term stability,
coupled with precision and low cost, makes alumina-modified
AgFON substrates ideally suited for potential field sensing
applications.
Conclusions
Acknowledgment. We are grateful to Dr. Jeffrey N. Anker,
Dr. Douglas A. Stuart, and Dr. Kallie Willets for the helpful
comments. This work was supported by the Chemical Sciences,
Geosciences and Biosciences Division, Office of Basic Energy
Sciences, Office of Science, U.S. Department of Energy (DE-
FG02-03ER15457), the National Science Foundation (DMR-
0076097, DMR-0520513, and CHE-0414554), and the Air Force
Office of Scientific Research MURI program (F49620-02-1-
0381). ALD was conducted at Argonne National Laboratory
and was supported by the U.S. Department of Energy, BES-
Materials Sciences (W-31-109-Eng-38). We are also grateful
to the Electron Microscopy Center at Argonne National Labora-
tory, which is supported by the DOE Office of Science
(W-31-109-Eng-38).
Alumina-modified AgFON substrates were fabricated using
atomic layer deposition. The surface chemistry of this new type
of SERS substrates is predictable on the basis of the polar
interaction between the analyte molecules and the stationary
phase, alumina. This hypothesis was verified by both the TLC
and SERS analyses of the competitive adsorptions between
dipicolinic acid, diacetylpyridine, and dimethoxypyridine. As
expected, dipicolinic acid has the highest binding affinity to
alumina-modified AgFON substrates. The high adsorption
affinity of carboxylic acid to alumina-modified AgFON surfaces
was exploited to improve the detection of CaDPA, a bacillus
spore biomarker. The SERS intensity of CaDPA was measured
as a function of alumina overlayer thickness. The AgFON
substrates modified with two ALD cycles of alumina optimize
for the detection of CaDPA. This case demonstrates the
importance of understanding the SERS signal transition mech-
anism. Specifically, the tradeoff between SERS intensity
JA0638760
(30) Sequaris, J. M. L.; Koglin, E. Anal. Chem. 1987, 59, 525.
(31) Seifar, R. M.; Altelaar, M. A. F.; Dijkstra, R. J.; Ariese, F.; Brinkman, U.
A. T.; Gooijer, C. Anal. Chem. 2000, 72, 5718.
9
J. AM. CHEM. SOC. VOL. 128, NO. 31, 2006 10309