of 8
Bioinspired Disordered Flexible Metasurfaces for Human Tear
Analysis Using Broadband Surface-Enhanced Raman Scattering
Vinayak Narasimhan, Radwanul Hasan Siddique, Haeri Park, and Hyuck Choo
*
Cite This:
ACSOmega
2020, 5, 12915
12922
Read Online
ACCESS
Metrics & More
Article Recommendations
*
s
ı
Supporting Information
ABSTRACT:
Flexible surface-enhanced Raman scattering (SERS) has
received attention as a means to move SERS-based broadband biosensing
from bench to bedside. However, traditional
fl
exible periodic nano-
arrangements with sharp plasmonic resonances or their random counter-
parts with spatially varying uncontrollable enhancements are not reliable
for practical broadband biosensing. Here, we report bioinspired quasi-
(dis)ordered nanostructures pre
senting a broadband yet tunable
application-speci
fi
c SERS enhancement pro
fi
le. Using simple, scalable
biomimetic fabrication, we create a
fl
exible metasurface (
fl
ex-MS) of quasi-
(dis)ordered metal
insulator
metal (MIM) nanostructures with spec-
trally variable, yet spatially controlled electromagnetic hotspots. The MIM
is designed to simultaneously localize the electromagnetic signal and block
background Raman signals from the underlying polymeric substrate
an
inherent problem of
fl
exible SERS. We elucidate the e
ff
ect of quasi-
(dis)ordering on broadband tunable SERS enhancement and employ the
fl
ex-MS in a practical broadband SERS demonstration to
detect human tear uric acid within its physiological concentration range (25
150
μ
M). The performance of the
fl
ex-MS toward
noninvasively detecting whole human tear uric acid levels
ex vivo
is in good agreement with a commercial enzyme-based assay.
1. INTRODUCTION
Surface-enhanced Raman scattering (SERS) has shown great
promise as a technique for molecular
fi
ngerprinting because of
its high sensitivity and selectivity along with its inherent
simplicity.
1
3
In particular, SERS from
fl
exible substrates has
received great attention recently owing to its advantages over
rigid substrates.
4
6
For instance,
fl
exible SERS can be used on
irregular surfaces for
in situ
biosensing thereby overcoming
complex analyte extraction strategies and other sample
preparation steps required while using conventional rigid
substrates. However, the adoption of
fl
exible SERS as a point-
of-care diagnostic tool has been limited by a number of
factors.
6
First, most structures used for this application are
designed either with periodic arrangements because of their
highly predictable plasmonic resonances
7
,
8
or with random
arrangements because of their ease of fabrication.
9
,
10
However,
periodic structures with narrowband resonance pro
fi
les are not
tunable for multiplexing.
6
8
Multiplexed
in situ
SERS for
instance would require a broadband and tunable plasmonic
resonance pro
fi
le to provide uniform enhancement of various
Raman modes occurring at greatly di
ff
ering vibrational energy
states.
11
13
In the same vein, random structures do not
guarantee repeatable SERS performance for a given bandwidth
because of spatially varying enhancements that are not
tunable.
14
Second,
fl
exible SERS platforms are usually made
of polymers that generate a considerable Raman background
signal.
6
Finally, challenges in obtaining reproducible signal can
also be attributed to the di
ffi
culty in homogeneously, scalable
and cost e
ff
ective manufacturing SERS active sites or
hotspots.
15
These issues necessitate the requirement for
fl
exible SERS approaches with spatially uniform broadband,
yet tunable plasmonic resonances that are scalable and reliably
manufacturable.
Inspiration can be sought from nature which boasts a
plethora of biophotonic nanostructures possessing quasi-
(dis)order or controlled diso
rder where both structural
dimensions and periodicity follow unique distributions.
16
,
17
Such an amalgamation of short-range order (
i.e
., periodicity)
with long-range disorder (
i.e
., randomization) along with
variations in the structure size leads to a host of useful
omnidirectional broadband, y
et tunable optical proper-
ties.
18
20
In this work, using a simple biomimetic fabrication
process, we realize a highly scalable and
fl
exible plasmonic
metasurface-based (
fl
ex-MS) SERS platform. The
fl
ex-MS
consists of a dense quasi-(dis)ordered ensemble of gold (Au)
nanodisks on nanoholes separated by a sub-10 nm silicon
Received:
February 14, 2020
Accepted:
May 6, 2020
Published:
May 18, 2020
Article
http://pubs.acs.org/journal/acsodf
© 2020 American Chemical Society
12915
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
This is an open access article published under an ACS AuthorChoice License, which permits
copying and redistribution of the article or any adaptations for non-commercial purposes.
Downloaded via CALIFORNIA INST OF TECHNOLOGY on June 11, 2020 at 20:37:45 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
dioxide (SiO
2
) nanogap in a metal
insulator
metal (MIM)
con
fi
guration. The entire con
fi
guration is created at a wafer-
scale over a 400
μ
m-thick silicone elastomer (PDMS) layer.
The MIM is designed not only for light con
fi
nement and
fi
eld
enhancement, but is also responsible for blocking the Raman
background signal from the underlying PDMS. The fabrication
approach which is spin-coating based allows for a short-range-
ordered Gaussian distribution of MIM nanostructure diameters
resulting in a broadband, yet tunable plasmonic resonance and
corresponding
|
E
/
E
o
|
2
enhancement factor (EF). Through
rigorous numerical simulations and comparative experimenta-
tion, we isolate the contributions of the bioinspired short-range
periodicity (
i.e
., positional disorder) and the diameter
distribution (
i.e
., size disorder) toward SERS EF. As a practical
broadband SERS demonstration of the
fl
ex-MS, we perform
the label-free detection of human tear uric acid (UA) toward
the diagnosis various chronic pathologies. Our platform
e
ff
ectively tracks prominent peaks of UA at large Raman shifts
within its physiological concentration range (25
150
μ
M) in
human tears. Finally, we show that the performance of the
fl
ex-
MS in detecting UA levels in whole human tear samples from
di
ff
erent subjects is in good agreement with a commercial
enzyme-based assay.
2. RESULTS AND DISCUSSION
2.1. Flex-MS Fabrication and Characterization.
The
fl
ex-MS was fabricated using a simple process consisting of
three steps, as shown in
Figure 1
a. On spin-coated PDMS thin
fi
lms, the
fi
rst step involves a biomimetic technique that relies
on nanostructuring through the lateral phase-separation of two
synthetic polymers
polystyrene (PS) and polymethyl meth-
acrylate (PMMA) co-dissolved in methyl ethyl ketone
(MEK).
20
,
21
This approach is analogous to the formation of
quasi-(dis)ordered biophotonic nanostructures on the wings
and scales of birds and insects.
22
The lateral phase separation
of PS and PMMA occurs under spin-coating which results in
densely packed quasi-(dis)ordered distribution of circular
hydrophobic PS islands in a matrix of hydrophilic PMMA
(
Figure S1
). Furthermore, through the control of various
parameters such as the spin-speeds, relative humidities,
polymer weight ratios, and molecular weights, the average
diameter and short-range periodicity of the PS islands can be
e
ff
ectively tuned.
23
,
24
Next, through selective dissolution of
PMMA in acetic acid which leaves behind a nanopillar mask of
PS, SiO
2
is directionally deposited through
E
-beam evapo-
ration. Finally, Au of an appropriate thickness is directionally
evaporated over the SiO
2
nanopillars to create a scalable
fl
ex-
MS with uniform sub-10 nm thick MIM nanogaps. A
photograph and SEM image of the
fl
ex-MS are shown in
Figure 1
b,c. A 2D fast Fourier transform (FFT) of the top view
SEM image as shown in the inset of
Figure 1
c reveals a ring-
shape distribution in the spatial domain that is characteristic of
quasi-(dis)order in nature.
16
,
17
,
20
The average short-range
isotropic periodicity in this case is 318
±
45 nm. Moreover, the
second inset in
Figure 1
c shows a high-magni
fi
cation SEM
image of a single MIM structure con
fi
rming the sub-10 nm
hotspot. Additionally, we have veri
fi
ed the existence of the sub-
10 nm gap in our prior work.
24
Figure 1
d shows a distribution
of MIM structure diameters which can be estimated by a
Gaussian pro
fi
le
fi
t with a mean and standard deviation (SD)
of 101
±
49 nm.
2.2. Single MIM Structure Simulations.
The aforemen-
tioned dimensions were chosen through rigorous
fi
nite-
di
ff
erence time-domain (FDTD) simulations. First, as shown
in
Figure 2
a, a single MIM structure with a 5 nm insulator gap
was simulated to obtain the extinction pro
fi
le and electric-
fi
eld
distribution. The plasmonic behaviors of the two metal layers
(Au nanohole and Au nanodisk) couple with each other
leading to strong electric-
fi
eld con
fi
nement at the MIM
junction.
24
This e
ff
ect has been demonstrated for di
ff
erent
MIM combinations toward broadband SERS applications.
25
,
26
Additionally, we have previously demonstrated the use of
similar MIM structures with higher-order plasmonic gap
modes toward plasmon-enhanced
fl
uorescent detection of
nucleic acids and suppression of
fl
uorescence quenching.
24
A
modal analysis reveals that in the vis
NIR regime, this
coupling results in the formation of a dipolar mode polarized
along the MIM junction (
Figure 2
b). This, in turn, produces a
very tunable and large extinction cross section and a
corresponding localized
|
E
|
2
enhancement.
Figure 2
c presents
the maximum normalized
|
E
/
E
o
|
2
enhancement numerically
obtained for single MIM structures ranging from 60 to 110 nm
in diameter with a 5 nm gap. However, like most ordered
plasmonic structures used for SERS, this mode is fairly
narrowband.
2.3. Flex-MS Ensemble Simulations.
To quantify the
impact of quasi-(dis)ordering on the
|
E
/
E
o
|
2
enhancement, we
generated an ensemble of quasi-(dis)ordered MIM structures
to simulate their optical properties. Two aspects were studied:
(1) short-range periodicity (
i.e
., positional disorder) and (2)
diameter distribution (
i.e
., size disorder). Here, the diameters
of the structures were made to obey a Gaussian distribution as
with the
fl
ex-MS (diameter: 98
±
30 nm) with short-range
periodicity (319
±
37 nm) (
Figure S2
). To isolate the
individual contributions of (1) and (2), periodic structures
(diameter: 100 nm, periodicity: 320 nm) and short-range-
ordered structures of the same diameter (diameter: 100 nm)
Figure 1.
(a) Simple and scalable three-step fabrication process of the
fl
ex-MS. (b) Fabricated
fl
ex-MS sample with (c) quasi-(dis)ordered
MIM nanostructures. Insets correspond to the 2D FFT taken to
determine a short-range periodicity of 318
±
45 nm (left) and a single
MIM nanostructure of diameter 100 nm with the sub-10 nm insulator
nanogap indicated by the white arrow (right). Scale bar: 2
μ
m. (d)
MIM structure diameter distribution with a Gaussian mean and SD of
101
±
49 nm.
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12916
were considered. For the three categories of structures, an
insulator gap of 9 nm was chosen to tune the resonance close
to the laser wavelength (
λ
L
= 785 nm) (
Figure S3
). The
e
ff
ective normalized
|
E
/
E
o
|
2
enhancement in each case
(averaged over 4000 individual hotspots) was computed not
only at the laser wavelength (
λ
L
= 785 nm), but also the
Raman-shifted wavelengths of relevance (
λ
1
,
λ
2
, and
λ
3
= 827,
861, and 902 nm) for this work. For periodic structures, the
resonance pro
fi
le is predictably governed by the collective gap-
plasmon resonance of individual MIM structures that
dominate any weaker in-plane lattice e
ff
ects (
Figure 3
a).
27
This is evidenced by the fact that the periodic array resonance
position and mode shape is identical to that of a single MIM
structure (diameter: 100 nm, gap: 9 nm) (
Figure S4
). Next,
quasi-(dis)ordered structures with the same diameter (
i.e
.,
positional disorder) were studied. The e
ff
ect of the pure
positional disorder retains the collective gap-plasmon reso-
nance of individual MIM structures marked by the same
resonance position and intensity as that of the periodic array
given the same number of averaged hotspots (
Figure 3
b). In
other words, for a periodicity that is large enough to ensure
either weak or no coupling between adjacent nanostructures,
the e
ff
ect of the short-range order (
i.e
., pure positional
disorder) is negligible.
28
,
29
Finally, the introduction of a
Gaussian distribution of diameters along with the short-range
order (
i.e
., size and positional disorder) produces a
considerably more broadband
|
E
/
E
o
|
2
enhancement compared
to the periodic structures as shown in
Figure 3
c because of the
resonance of subsets of MIM structures of a given diameter
(
Figures 3
dand
S5
).
28
The bandwidth of the
|
E
/
E
o
|
2
enhancement of size-disordered structures (
219 nm) was
3.4 times larger than that of the periodic structures (
65 nm).
The fractal, yet controllable nature of the MIM ensemble with
a size and positional disorder produces a broadband
|
E
/
E
o
|
2
enhancement that encompasses
λ
L
and
λ
1
3
. While the periodic
structures and those with a pure positional disorder provide a
high e
ff
ective enhancement at
λ
L
(
i.e
.,
|
E
/
E
o
|
λ
L
2
) of 3358 and
3430, respectively, their enhancements at
λ
1
3
is considerably
lower (periodic:
|
E
/
E
o
|
λ
1
2
1210,
|
E
/
E
o
|
λ
2
2
563, and
|
E
/
E
o
|
λ
3
2
285; pure positional disorder:
|
E
/
E
o
|
λ
1
2
1252,
|
E
/
E
o
|
λ
2
2
585, and
|
E
/
E
o
|
λ
3
2
292). In comparison, the introduction of
the size disorder provides a more uniform enhancement with
|
E
/
E
o
|
λ
L
2
1992,
|
E
/
E
o
|
λ
1
2
1748,
|
E
/
E
o
|
λ
2
2
1305, and
|
E
/
E
o
|
λ
3
2
921.
Figure 2.
(a) Schematic of a single MIM nanostructure on the
fl
ex-MS platform (b) Coupling of the Au nanodisk and Au nanohole across the SiO
2
nanogap results in greatly enhanced electric-
fi
eld environments. (c) Field pro
fi
le of the nanogap shows the presence of a vertically polarized dipolar
mode. (d) Normalized maximum theoretical
|
E
/
E
o
|
2
enhancement (solid curves) and extinction cross section (dotted curves) of a single MIM
nanostructure of varying diameter between 60 and 110 nm with a
fi
xed gap size of 5 nm.
Figure 3.
Ensemble of (a) periodic MIM nanostructures and (b) those with pure positional disorder demonstrating an identical narrowband
e
ff
ective normalized
|
E
/
E
o
|
2
enhancement pro
fi
le. (c) E
ff
ective normalized
|
E
/
E
o
|
2
enhancement pro
fi
le of a quasi-(dis)ordered MIM ensemble with
both positional and size disorder is broadband compared to (a,b). The
|
E
/
E
o
|
2
enhancement pro
fi
le encompasses the excitation and the Raman-
shifted wavelengths (
λ
L
,
λ
1
,
λ
2
, and
λ
3
) of relevance in this work. (d) Field-map spanning the 3.5
×
3.5
μ
m array at
λ
L
,
λ
1
,
λ
2
, and
λ
3
shows the
progressive excitation of
fi
rst small and then larger MIM nanostructures with increasing wavelength. (e) E
ff
ective
|
E
/
E
o
|
4
enhancement from
periodic structures, structures with a positional disorder only and those with both positional and size disorder. This e
ff
ective enhancement was
numerically computed as EF(
ω
L
,
ω
R
).
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12917
2.4. Broadband SERS Enhancement Factor Compu-
tation.
The origin of this broadband e
ff
ect can be of great
bene
fi
t to SERS. As per the frequently used
|
E
|
4
-approximation,
SERS enhancement factor (EF) is typically expressed as
ω
ω
ω
=
r
r
r
E
E
E
F( , )
(,)
(,)
Lm
loc L m
inc L m
4
where
E
loc
(
ω
L
,
r
m
) and
E
inc
(
ω
L
,
r
m
) are the localized and
incident electric
fi
elds at a laser excitation frequency of
ω
L
for a
Raman dipole at position
r
m
.
30
Here, a signi
fi
cant contribution
from the radiative enhancement
M
rad
is overlooked in favor of
electric-
fi
eld enhancement
M
loc
.
31
33
M
rad
measures the
enhancement of the power radiated by a dipole in the
presence of a plasmonic nanostructure. Factoring in this
contribution, the EF can be more precisely expressed as
30
ωωωω
ω
ω
ω
ω
=
=
rrr
r
r
r
r
MM
E
E
E
E
EF(, ,) (,) (,)
(,)
(,)
(,)
(,)
LRm locLmradRm
loc L m
inc L m
2
loc R m
inc R m
2
where
ω
R
is the Stokes-shifted Raman scattering frequency.
The
|
E
|
4
-approximation assumes that
ω
R
ω
L
and as a result,
M
loc
M
rad
. While this approximation is accurate for
ω
L
in the
blue and green regime, it is inaccurate in the red and NIR
regime particularly within the
fi
ngerprint region (500
1500
cm
1
).
32
,
34
This is because for a given vibrational mode, the
di
ff
erence between the vibrational and excitation energies
becomes more signi
fi
cant for lower energy excitations. In
principle, for NIR-based SERS, broadband EFs are greatly
desirable as they better account for
M
rad
.
32
Based on our
ensemble simulations, the e
ff
ective EF(
ω
L
,
ω
1
3
) for the
periodic structures (EF(
ω
L
,
ω
1
)
4.06
×
10
6
, EF(
ω
L
,
ω
2
)
1.89
×
10
6
, and EF(
ω
L
,
ω
3
)
0.96
×
10
6
) and those with a
purely positional disorder (EF(
ω
L
,
ω
1
)
4.29
×
10
6
, EF(
ω
L
,
ω
2
)
2.01
×
10
6
, and EF(
ω
L
,
ω
3
)
1.00
×
10
6
) are evidently
lower for
ω
2
and
ω
3
than structures with a size and positional
disorder (EF(
ω
L
,
ω
1
)
3.48
×
10
6
, EF(
ω
L
,
ω
2
)
2.60
×
10
6
,
and EF(
ω
L
,
ω
3
)
1.84
×
10
6
)(
Figure 3
e).
The e
ff
ect of the bioinspired quasi-(dis)order was then
experimentally veri
fi
ed by comparing the optical properties of
the
fl
ex-MS with periodic MIM structures (diameter: 100 nm,
periodicity: 320 nm, gap: 9 nm) fabricated
via E
-beam
lithography. Using a microspectroscopic setup in dark-
fi
eld
(DF) mode, the scattering cross section of the two
metasurfaces were measured in the NIR regime. This revealed
the considerably more broadband scattering pro
fi
le of the
fl
ex-
MS compared to the periodic array (
Figure 4
a). Furthermore,
the existence of the 9 nm gap was veri
fi
ed through simulation
(
Figure S6
). The SERS performance of the two sets of
structures was compared by detecting UA
a SERS-active
molecule with prominent peaks at 640, 1134, and 1645 cm
1
which originate from the skeletal ring and C
Nbond
deformations.
35
When excited with a 785 nm laser, these
peaks appear at 827, 861, and 902 nm, respectively (
i.e
.,
ω
L
ω
R
). For the same UA concentration in DI water (150
μ
M)
and three di
ff
erent laser powers (0.32, 0.62, and 1.12 mW), the
three peaks under consideration were uniformly enhanced by
the
fl
ex-MS compared to the periodic MIM structures (
Figure
4
b,c). The 640 cm
1
peak was enhanced more signi
fi
cantly by
the periodic structures compared to the
fl
ex-MS as this peak
lies closest to
ω
L
, where the sharp plasmonic resonance of the
former is tuned. However, 1134 and 1645 cm
1
peaks that
were located further away from
ω
L
were enhanced more
signi
fi
cantly by the
fl
ex-MS because of its broadband
enhancement pro
fi
le. This property is particularly useful for
sensing with low power. In our case for instance, when using a
power of 0.32 mW, the 1134 and 1645 cm
1
peaks were
enhanced 3.1 and 5.7 fold by the
fl
ex-MS compared to the
periodic structures. Finally, the background suppression
property of the
fl
ex-MS was also veri
fi
ed through comparative
experimentation (
Figure S7
).
2.5. Broadband SERS Biosensing of Tear UA.
The
broadband enhancement of the
fl
ex-MS platform was used to
detect various concentrations of UA. Hyperuricemia (
i.e
.,
elevated levels of blood UA) has been identi
fi
ed as a biomarker
of various diseases such as gout or gouty arthritis,
36
diabetes,
37
Parkinson
s disease,
38
renal disease,
39
and cardiovascular
disease,
40
to name a few. While continual monitoring of UA
in blood is hindered by the invasive nature of blood collection
and sampling, human tears which are considerably less invasive
to assay thereby proving to be an interesting alternative.
41
Tears are also far less complex in constituents compared to
blood and have large average concentrations of UA (68
±
46
Figure 4.
(a) Scattering intensity of the fabricated periodic and quasi-(dis)ordered array shown as solid lines measured using a microspectroscopic
setup operating in DF mode. The simulated scattering pro
fi
le of a single MIM structure (gap: 9 nm) which most closely matches the experimental
result is also shown as a dotted line. The SEM images of the corresponding structures are shown in the inset. (b) SERS spectra of UA measured at
0.62 mW laser power. While the periodic MIM array enhances the 640 cm
1
peak considerably, the
fl
ex-MS o
ff
ers a more broadband enhancement
of all three peaks. (c) Analysis performed for various laser powers.
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12918
μ
M).
10
Furthermore, a clear correlation between blood and
tear UA levels has been established.
10
As a result, assaying tear
UA levels noninvasively and in a label-free fashion
via
fl
exible
SERS holds great promise. Using the
fl
ex-MS, we
fi
rst tracked
the 640, 1134, and 1645 cm
1
peaks for UA concentrations
ranging from 25 to 150
μ
M in phosphate-bu
ff
ered saline (PBS)
(
Figure 5
a
c). Next, to better simulate the SERS pro
fi
le of UA
in whole tears, we prepared various concentrations of UA
between 25 and 150
μ
M in an arti
fi
cial tear bu
ff
er which
consisted of various prominent tear proteins such as lysozyme,
lactoferrin, albumin, and immunoglobulins, as well as electro-
lytes that maintain tear osmolarity such as Na
+
,K
+
,Cl
, and
HCO
3
(see the Methods Section in
Supporting Information
for additional details). As a practical demonstration of
broadband SERS enhancement, the 640, 1134, and 1645
cm
1
peaks were tracked with excellent linearity being
observed as shown in
Figure 5
d. These measurements from
an arti
fi
cial tear bu
ff
er were used as a characteristic curve to
map measurements taken for whole tears. Finally, we analyzed
the performance of the
fl
ex-MS toward the detection of UA
levels in whole tears. Here, we tested pooled tears from eight
di
ff
erent samples obtained from healthy subjects. As a
comparison, the same measurement was made using a
commercial colorimetric enzyme-based assay. The average
concentration from the
fl
ex-MS for the 640, 1134, and 1645
cm
1
peaks (66, 67, and 67
μ
M, respectively) was in good
agreement with that from the enzyme-based assay (76
μ
M)
which demonstrates the potential of the
fl
ex-MS as an e
ff
ective
label-free SERS diagnostic platform (
Figure 5
e).
3. CONCLUSIONS
In summary, using a simple biomimetic fabrication process, we
have developed a cost-e
ff
ective, scalable, and
fl
exible plasmonic
metasurface-based platform for label-free SERS. The
fl
ex-MS
consists of a dense ensemble of closely coupled Au nanodisks
and Au nanoholes separated by sub-10 nm SiO
2
nanogaps in
an MIM arrangement all on
fl
exible PDMS thin
fi
lms. The
MIM provides a very tunable and large extinction cross section
and a corresponding localized
|
E
|
2
enhancement. When
considering an ensemble of such MIM structures exhibiting
controlled disorder, the resonance pro
fi
le of the entire system
becomes broadband thereby enabling the uniform enhance-
ment of not just the excitation wavelength (
λ
L
), but also large
Stokes-shifted Raman scattering wavelengths (
λ
R
λ
L
).
Finally, using
fl
ex-MS as a broadband diagnostic platform, we
demonstrate the label-free detection of UA in both arti
fi
cial
tear bu
ff
er as well as whole human tear samples. Additionally,
we compare the performance of the
fl
ex-MS with a commercial
UA measurement assay and show that they are in good
agreement. As a result, we envisage that through broadband
SERS enhancement, the
fl
ex-MS can prove to be a reliable and
scalable label-free diagnostic platform for a variety of SERS-
active molecules.
4. EXPERIMENTAL SECTION
4.1. Biomimetic Flex-MS Fabrication.
First, a 400
μ
m
thick PDMS layer (Sylgard 184 elastomer base mixed with
curing agent in a ratio of 10:1, Dow Chemical Co., USA) was
spin-coated on a 4 in. Si wafer and cured at 65
°
C for 12 h.
Figure 5.
(a) Peaks of UA at 640, 1134, and 1645 cm
1
tracked between 25 and 150
μ
M in PBS. (b) SERS surface mapping (UA concentration:
150
μ
M) over a 150
×
150
μ
m area at 640, 1134, and 1645 cm
1
showing spatial uniformity. Scale bars: 20
μ
m. (c) Normalized intensity of each
peak shows excellent linearity. (d) 640, 1134, and 1645 cm
1
peaks tracked between 25 and 150
μ
M in arti
fi
cial tear bu
ff
er. (e) SERS performance
of the
fl
ex-MS using all three peaks is compared for pooled whole tears consisting of 8 individual tear samples with a commercial enzyme-based
assay. The two results are in good agreement.
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12919
Next, poly(methyl methacrylate) (PMMA,
M
w
= 5090,
Polymer Standards Service GmbH, Germany) and polystyrene
(PS,
M
w
= 3250, Polymer Standards Service GmbH, Germany)
were co-dissolved in MEK (Sigma-Aldrich, USA) with mass
ratios of 70%:30%. The solution concentrations were 20 mg
mL
1
. After exposing the PDMS substrate to O
2
plasma for 5
min, the polymer blend solutions were spin-coated at a spin-
speed of 3500 rpm and acceleration of 2000 rpm s
1
for 30 s.
Relative humidity was maintained between 40 and 50% during
the spin-coating. As shown in
Figure S1
, the de-mixing of the
blend components occurs during spin-coating because of the
di
ff
erence in relative solubility of PS and PMMA in MEK.
First, water condensation begins at humidity levels above 35%
forming a water-rich layer at the air/solution interface because
of the di
ff
erence in the evaporation rate between water and
MEK. Water starts to condense from the air into the solution
because of the evaporation of MEK decreasing the temperature
on the top, below the dew point. Because of the high water
concentration, a 3D phase separation occurs between PS/MEK
and PMMA/MEK/water. Upon drying, a purely lateral
morphology was formed with ellipsoidal PS islands in a
PMMA matrix. The samples were then rinsed in acetic acid for
60 s and dried in a stream of N
2
to remove the PMMA matrix
leaving behind the PS islands. The resulting PS nanopillar
mask then served as a template for the
E
-beam evaporation of a
100
±
0.5 nm-thick SiO
2
layer (CHA MK40 E-Beam
Evaporation, CHA Industries, USA). Next, a 91
±
1.1 nm-
thick Au layer was deposited (CHA MK40 E-Beam
Evaporation, CHA Industries, USA) over SiO
2
to generate
the MIM layer with
9
±
1.6 nm gap. Finally, the PDMS
fi
lm
was peeled o
ff
the Si wafer and served as a SERS substrate for
biosensing experiments.
4.2. Periodic MIM Array Fabrication.
PMMA950 A4
(MicroChem, USA)
E
-beam resist was spin-coated at 1500
rpm for 60 s onto an Si substrate cleaned with acetone and
isopropyl alcohol prior to spin-coating. The sample was then
exposed at a dosage of 800
μ
Ccm
2
(Raith EBPG 5200, Raith
Nanofabrication, Germany). Following exposure, development
was carried out using a 1:1 ratio of IPA/MIBK for 1 min. Next,
SiO
2
was directionally deposited
via E
-beam evaporation
(CHA MK40 E-Beam Evaporation, CHA Industries, USA)
after which lift-o
ff
was performed in Remover PG leaving
behind SiO
2
nanopillars. Finally, Au of an appropriate
thickness was deposited (CHA MK40 E-Beam Evaporation,
CHA Industries, USA) over SiO
2
to generate the MIM layer.
4.3. High-Resolution Imaging and Statistical Anal-
yses of the Flex-MS.
SEM imaging was performed on the
fl
ex-MS using a Nova 200 Novalab Dualbeam microscope
(FEI, USA) at 10 kV. ImageJ (National Institutes of Health,
USA), a Java-based public-domain image processing tool was
used to obtain the diameter distribution of the nanostructures.
Every pixel in the acquired SEM images were converted to
black or white based on a thresholding condition that was
obtained by calculating the mean intensity value of all pixels of
the image. Following this, the diameter distribution of the
MIM scatterers was determined. Finally, the FFT analysis to
determine short-range periodicity was performed using
MATLAB (MathWorks, USA).
4.4. Optical Simulations of the Flex-MS.
Optical
simulations of a single MIM nanostructure as well as the
quasi-ordered/periodic ensemble as a whole were performed
using 3D
fi
nite-di
ff
erence time-domain software (Lumerical
Solutions, Canada). A combination of periodic and perfect
matching layer boundary conditions along with a plane wave
source was used. The absorption cross section was obtained in
the total
fi
eld region inside the source while the scattering
cross section was obtained in the scattered
fi
eld region outside
the source. The in
fl
uence of the nanogap was studied as shown
in
Figure S6
. The
|
E
/
E
o
|
2
enhancement was obtained by
placing a frequency-domain
fi
eld monitor at the MIM junction
spanning the entire MIM nanostructure or the ensemble,
respectively. The
fi
eld maps as well as the e
ff
ective
|
E
/
E
o
|
2
enhancement (averaged over 4000 individual hotspots or mesh
cells) in the case of the ensemble were calculated using
MATLAB (MathWorks, USA).
4.5. Spectroscopic Analyses.
An optical microscope
operating in dark-
fi
eld (DF) mode was used for the
microspectroscopic investigation of the fabricated
fl
ex-MS
and periodic MIM array samples. A halogen lamp was used as a
light source using a 50
×
objective. The scattered light was
collected in a confocal con
fi
guration and analyzed using a
spectrometer (AvaSpec-ULS2048x64-USB2). A 400
μ
m core
optical
fi
ber was used to obtain a spatial resolution of 20
μ
mto
characterize the scattering of the samples.
4.6. SERS Measurements.
All SERS measurements were
performed by incubating a 30
μ
L droplet of the reagent on the
fl
ex-MS for 30 min. Following this, the drop was dislodged
from the sample surface using a stream of N
2
. The sample was
then measured dry using a Raman microscope (inVia,
Renishaw, United Kingdom) with a 50
×
objective lens. A
785 nm laser operated at 0.32, 0.62, or 1.12 mW for a duration
of 60 s was used to take the measurements.
4.7. Arti
fi
cial Tear Bu
ff
er Preparation.
Arti
fi
cial tear
bu
ff
er was prepared using previously reported methods.
10
,
42
Brie
fl
y, electrolytes K
+
, HCO
3
,Na
+
, and Cl
were added to
DI water at concentrations of 24, 24, 130, and 130 mM,
respectively. Following this, representative tear proteins such as
lysozyme, lactoferrin, albumin, and IgG were added at
concentrations of 2 mg mL
1
, 2 mg mL
1
, 0.02 mg mL
1
,
and 3
μ
gmL
1
respectively.
4.8. Whole Tear Assay.
Eight individual tear samples were
collected from healthy subjects and pooled for the experiment.
For a comparative study, a commercial enzyme-based
colorimetric UA detection kit (Sigma-Aldrich, USA) was
used. The assay was performed following the manufacturer
s
protocol.
ASSOCIATED CONTENT
*
s
ı
Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acsomega.0c00677
.
Schematic of the di
ff
erent stages of the 3D phase
separation process, characterization of the
fl
ex-MS
FDTD model, extinction cross section as a function of
the nanogap size, comparison of
|
E
/
E
o
|
2
enhancement
from a single MIM structure and a periodic array,
simulated
fl
ex-MS FDTD model, comparison of an MIM
structure with an open and covered nanogap, and
suppression of background signal by the
fl
ex-MS (
PDF
)
AUTHOR INFORMATION
Corresponding Author
Hyuck Choo
Department of Electrical Engineering, California
Institute of Technology, Pasadena, California 91125, United
States; Samsung Advanced Institute of Technology, Samsung
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12920
Electronics, Suwon, Gyeonggi-do 16678, South Korea
;
Email:
hyuck.choo@samsung.com
,
hchoo@caltech.edu
Authors
Vinayak Narasimhan
Department of Medical Engineering,
California Institute of Technology, Pasadena, California 91125,
United States;
orcid.org/0000-0003-4165-402X
Radwanul Hasan Siddique
Department of Medical
Engineering, California Institute of Technology, Pasadena,
California 91125, United States; Image Sensor Lab, Samsung
Semiconductor, Inc., Pasadena, California 91101, United
States;
orcid.org/0000-0001-7494-5857
Haeri Park
Department of Medical Engineering, California
Institute of Technology, Pasadena, California 91125, United
States
Complete contact information is available at:
https://pubs.acs.org/10.1021/acsomega.0c00677
Author Contributions
V.N., R.H.S., and H.C. conceived the study. V.N., R.H.S., and
H.P. performed the necessary simulations and experiments
under the supervision of H.C. The manuscript was written
through contributions of all the authors.
Funding
SAMSUNG Global Research Outreach (GRO) program.
Notes
The authors declare no competing
fi
nancial interest.
ACKNOWLEDGMENTS
The authors acknowledge the
fi
nancial support provided by the
SAMSUNG Global Research Outreach (GRO) program. The
authors are also thankful for the support and resources
provided by the Kavli Nanoscience Institute at Caltech.
REFERENCES
(1) Nie, S.; Emory, S. R. Probing Single Molecules and Single
Nanoparticles by Surface-Enhanced Raman Scattering.
Science
1997
,
275
, 1102
1106.
(2) Sharma, B.; Frontiera, R. R.; Henry, A.-I.; Ringe, E.; Van Duyne,
R. P. SERS: Materials, Applications, and the Future.
Mater. Today
2012
,
15
,16
25.
(3) Langer, J.; Jimenez de Aberasturi, D.; Aizpurua, J.; Alvarez-
Puebla, R. A.; Auguie
́
, B.; Baumberg, J. J.; Bazan, G. C.; Bell, S. E. J.;
Boisen, A.; Brolo, A. G.; Choo, J.; Cialla-May, D.; Deckert, V.; Fabris,
L.; Faulds, K.; García de Abajo, F. J.; Goodacre, R.; Graham, D.; Haes,
A. J.; Haynes, C. L.; Huck, C.; Itoh, T.; Ka
̈
ll, M.; Kneipp, J.; Kotov, N.
A.; Kuang, H.; Le Ru, E. C.; Lee, H. K.; Li, J.-F.; Ling, X. Y.; Maier, S.
A.; Mayerho
̈
fer, T.; Moskovits, M.; Murakoshi, K.; Nam, J.-M.; Nie,
S.; Ozaki, Y.; Pastoriza-Santos, I.; Perez-Juste, J.; Popp, J.; Pucci, A.;
Reich, S.; Ren, B.; Schatz, G. C.; Shegai, T.; Schlu
̈
cker, S.; Tay, L.-L.;
Thomas, K. G.; Tian, Z.-Q.; Van Duyne, R. P.; Vo-Dinh, T.; Wang,
Y.; Willets, K. A.; Xu, C.; Xu, H.; Xu, Y.; Yamamoto, Y. S.; Zhao, B.;
Liz-Marza
́
n, L. M. Present and Future of Surface-Enhanced Raman
Scattering.
ACS Nano
2020
,
14
, 28.
(4) Xu, K.; Wang, Z.; Tan, C. F.; Kang, N.; Chen, L.; Ren, L.; Thian,
E. S.; Ho, G. W.; Ji, R.; Hong, M. Uniaxially Stretched Flexible
Surface Plasmon Resonance Film for Versatile Surface Enhanced
Raman Scattering Diagnostics.
ACS Appl. Mater. Interfaces
2017
,
9
,
26341
26349.
(5) Cui, H.; Li, S.; Deng, S.; Chen, H.; Wang, C. Flexible,
Transparent, and Free-Standing Silicon Nanowire SERS Platform for
in Situ Food Inspection.
ACS Sens.
2017
,
2
, 386
393.
(6) Xu, K.; Zhou, R.; Takei, K.; Hong, M. Toward Flexible Surface-
Enhanced Raman Scattering (SERS) Sensors for Point-of-Care
Diagnostics.
Adv. Sci.
2019
,
6
, 1900925.
(7) Kang, H.; Heo, C.-J.; Jeon, H. C.; Lee, S. Y.; Yang, S.-M. Durable
Plasmonic Cap Arrays on Flexible Substrate with Real-Time Optical
Tunability for High-Fidelity SERS Devices.
ACS Appl. Mater.
Interfaces
2013
,
5
, 4569
4574.
(8) Jeong, J. W.; Yang, S. R.; Hur, Y. H.; Kim, S. W.; Baek, K. M.;
Yim, S.; Jang, H.-I.; Park, J. H.; Lee, S. Y.; Park, C.-O.; Jung, Y. S.
High-Resolution Nanotransfer Printing Applicable to Diverse Surfaces
via Interface-Targeted Adhesion Switching.
Nat. Commun.
2014
,
5
,
5387.
(9) Tian, L.; Jiang, Q.; Liu, K.-K.; Luan, J.; Naik, R. R.; Singamaneni,
S. Bacterial Nanocellulose-Based Flexible Surface Enhanced Raman
Scattering Substrate.
Adv. Mater. Interfaces
2016
,
3
, 1600214.
(10) Park, M.; Jung, H.; Jeong, Y.; Jeong, K.-H. Plasmonic Schirmer
Strip for Human Tear-Based Gouty Arthritis Diagnosis Using Surface-
Enhanced Raman Scattering.
ACS Nano
2017
,
11
, 438
443.
(11) Messina, G. C.; Malerba, M.; Zilio, P.; Miele, E.; Dipalo, M.;
Ferrara, L.; De Angelis, F. Hollow Plasmonic Antennas for Broadband
SERS Spectroscopy.
Beilstein J. Nanotechnol.
2015
,
6
, 492
498.
(12) Li, M.; Xu, J.; Romero-Gonzalez, M.; Banwart, S. A.; Huang, W.
E. Single Cell Raman Spectroscopy for Cell Sorting and Imaging.
Curr. Opin. Biotechnol.
2012
,
23
,56
63.
(13) De Angelis, F.; Malerba, M.; Patrini, M.; Miele, E.; Das, G.;
Toma, A.; Zaccaria, R. P.; Di Fabrizio, E. 3D Hollow Nanostructures
as Building Blocks for Multifunctional Plasmonics.
Nano Lett.
2013
,
13
, 3553
3558.
(14) Lin, X.-M.; Cui, Y.; Xu, Y.-H.; Ren, B.; Tian, Z.-Q. Surface-
Enhanced Raman Spectroscopy: Substrate-Related Issues.
Anal.
Bioanal. Chem.
2009
,
394
, 1729
1745.
(15) Wu, L.; Wang, W.; Zhang, W.; Su, H.; Liu, Q.; Gu, J.; Deng, T.;
Zhang, D. Highly Sensitive, Reproducible and Uniform SERS
Substrates with a High Density of Three-Dimensionally Distributed
Hotspots: Gyroid-Structured Au Periodic Metallic Materials.
NPG
Asia Mater.
2018
,
10
, No. e462.
(16) Johansen, V. E.; Onelli, O. D.; Steiner, L. M.; Vignolini, S.
Photonics in Nature: From Order to Disorder. In
Functional Surfaces
in Biology III: Diversity of the Physical Phenomena
; Gorb, S. N., Gorb,
E. V., Eds.; Biologically-Inspired Systems; Springer International
Publishing: Cham, 2017; pp 53
89.
(17) Kolle, M.; Lee, S. Progress and Opportunities in Soft Photonics
and Biologically Inspired Optics.
Adv. Mater.
2018
,
30
, 1702669.
(18) Wiersma, D. S. Disordered Photonics.
Nat. Photonics
2013
,
7
,
188
196.
(19)Florescu,M.;Torquato,S.;Steinhardt,P.J.Designer
Disordered Materials with Large, Complete Photonic Band Gaps.
Proc. Natl. Acad. Sci. U.S.A.
2009
,
106
, 20658
20663.
(20) Narasimhan, V.; Siddique, R. H.; Lee, J. O.; Kumar, S.;
Ndjamen, B.; Du, J.; Hong, N.; Sretavan, D.; Choo, H. Multifunc-
tional Biophotonic Nanostructures Inspired by the Longtail Glasswing
Butterfly for Medical Devices.
Nat. Nanotechnol.
2018
,
13
, 512
519.
(21) Siddique, R. H.; Donie, Y. J.; Gomard, G.; Yalamanchili, S.;
Merdzhanova, T.; Lemmer, U.; Ho
̈
lscher, H. Bioinspired Phase-
Separated Disordered Nanostructures for Thin Photovoltaic Absorb-
ers.
Sci. Adv.
2017
,
3
, No. e1700232.
(22) Saranathan, V.; Seago, A. E.; Sandy, A.; Narayanan, S.; Mochrie,
S. G. J.; Dufresne, E. R.; Cao, H.; Osuji, C. O.; Prum, R. O. Structural
Diversity of Arthropod Biophotonic Nanostructures Spans Amphi-
philic Phase-Space.
Nano Lett.
2015
,
15
, 3735
3742.
(23) Siddique, R. H.; Kumar, S.; Narasimhan, V.; Kwon, H.; Choo,
H. Aluminum Metasurface with Hybrid Multipolar Plasmons for
1000-Fold Broadband Visible F
luorescence Enhancement and
Multiplexed Biosensing.
ACS Nano
2019
,
13
, 13775.
(24) Narasimhan, V.; Siddique, R. H.; Hoffmann, M.; Kumar, S.;
Choo, H. Enhanced Broadband Fluorescence Detection of Nucleic
Acids Using Multipolar Gap-Plasmons on Biomimetic Au Meta-
surfaces.
Nanoscale
2019
,
11
, 13750
13757.
(25) Yang, Y.; Hu, Z.; Wang, Y.; Wang, B.; Zhan, Q.; Zhang, Y.; Ao,
X. Broadband SERS Substrates by Oblique Angle Deposition Method.
Opt. Mater. Express
2016
,
6
, 2644
2654.
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12921
(26) Wang, D.; Zhu, W.; Best, M. D.; Camden, J. P.; Crozier, K. B.
Wafer-Scale Metasurface for Total Power Absorption, Local Field
Enhancement and Single Molecule Raman Spectroscopy.
Sci. Rep.
2013
,
3
,1
6.
(27) Kravets, V. G.; Kabashin, A. V.; Barnes, W. L.; Grigorenko, A.
N. Plasmonic Surface Lattice Resonances: A Review of Properties and
Applications.
Chem. Rev.
2018
,
118
, 5912
5951.
(28) Auguie
́
, B.; Barnes, W. L. Diffractive Coupling in Gold
Nanoparticle Arrays and the Effect of Disorder.
Opt. Lett.
2009
,
34
,
401
403.
(29) Bertin, H.; Bru
̂
le
́
, Y.; Magno, G.; Lopez, T.; Gogol, P.; Pradere,
L.; Gralak, B.; Barat, D.; Deme
́
sy, G.; Dagens, B. Correlated
Disordered Plasmonic Nanostructures Arrays for Augmented Reality.
ACS Photonics
2018
,
5
, 2661
2668.
(30) Ding, S.-Y.; You, E.-M.; Tian, Z.-Q.; Moskovits, M. Electro-
magnetic Theories of Surface-Enhanced Raman Spectroscopy.
Chem.
Soc. Rev.
2017
,
46
, 4042
4076.
(31) Le Ru, E. C.; Etchegoin, P. G.
SERS Enhancement Factors and
Related Topics
; Le Ru, E. C., Etchegoin, P. G., Eds.; Elsevier:
Amsterdam, 2009.
(32) Stiles, P. L.; Dieringer, J. A.; Shah, N. C.; Van Duyne, R. P.
Surface-Enhanced Raman Spectroscopy.
Annu. Rev. Anal. Chem.
2008
,
1
, 601
626.
(33) McFarland, A. D.; Young, M. A.; Dieringer, J. A.; Van Duyne,
R. P. Wavelength-Scanned Surface-Enhanced Raman Excitation
Spectroscopy.
J. Phys. Chem. B
2005
,
109
, 11279
11285.
(34) Schlu
̈
cker, S. Surface-Enhanced Raman Spectroscopy: Con-
cepts and Chemical Applications.
Angew. Chem., Int. Ed.
2014
,
53
,
4756
4795.
(35) Westley, C.; Xu, Y.; Thilaganathan, B.; Carnell, A. J.; Turner, N.
J.; Goodacre, R. Absolute Quantification of Uric Acid in Human
Urine Using Surface Enhanced Raman Scattering with the Standard
Addition Method.
Anal. Chem.
2017
,
89
, 2472
2477.
(36) Reginato, A. M.; Mount, D. B.; Yang, I.; Choi, H. K. The
Genetics of Hyperuricaemia and Gout.
Nat. Rev. Rheumatol.
2012
,
8
,
610
621.
(37) Bhole, V.; Choi, J. W. J.; Woo Kim, S.; de Vera, M.; Choi, H.
Serum Uric Acid Levels and the Risk of Type 2 Diabetes: A
Prospective Study.
Am. J. Med.
2010
,
123
, 957
961.
(38) Wen, M.; Zhou, B.; Chen, Y.-H.; Ma, Z.-L.; Gou, Y.; Zhang, C.-
L.; Yu, W.-F.; Jiao, L. Serum Uric Acid Levels in Patients with
Parkinson
s Disease: A Meta-Analysis.
PLoS One
2017
,
12
,
No. e0173731.
(39) Kohagura, K.; Kochi, M.; Miyagi, T.; Kinjyo, T.; Maehara, Y.;
Nagahama, K.; Sakima, A.; Iseki, K.; Ohya, Y. An Association between
Uric Acid Levels and Renal Arteriolopathy in Chronic Kidney
Disease: A Biopsy-Based Study.
Hypertens. Res.
2013
,
36
,43
49.
(40) Feig, D. I.; Kang, D.-H.; Johnson, R. J. Uric Acid and
Cardiovascular Risk.
N. Engl. J. Med.
2008
,
359
, 1811
1821.
(41) Hagan, S.; Martin, E.; Enríquez-de-Salamanca, A. Tear Fluid
Biomarkers in Ocular and Systemic Disease: Potential Use for
Predictive, Preventive and Personalised Medicine.
EPMA J.
2016
,
7
,
15.
(42) Culver, H. R.; Wechsler, M. E.; Peppas, N. A. Label-Free
Detection of Tear Biomarkers Using Hydrogel-Coated Gold Nano-
shells in a Localized Surface Plasmon Resonance-Based Biosensor.
ACS Nano
2018
,
12
, 9342
9354.
ACS Omega
http://pubs.acs.org/journal/acsodf
Article
https://dx.doi.org/10.1021/acsomega.0c00677
ACSOmega
2020, 5, 12915
12922
12922