Surfaceenhanced Raman scattering (SERS) is a powerful analysis technique that allows both the identification and detection of analytes at trace levels. However, the low rate of charge transfer (CT) between noblemetal nanoparticles and several analytes prevents them from being effectively detected by SERSbased sensors. They are regarded as low Raman crosssection molecules. In this study, we enhanced the performance of the silver nanoparticles (AgNPs)based SERS sensing platform for a low Raman crosssection molecule, urea, focusing on improving the rate of CT. First, a set of Agtitanium dioxide (TiO2) nanocomposites were synthesized. The presence of TiO2 improved the intensity of the SERS signal of urea, in comparison to the use of bare AgNPs. Second, a photoinduced enhanced Raman spectroscopy (PIERS) technique was employed to further elevate the Raman signal of urea. Thanks to the step of preirradiation using UV light at λ = 365 nm, with the use of the substrates containing 25%, 33%, and 50% TiO2 content, enhancements of 1.93, 3.42, and 7.45 times were achieved, respectively, compared to the use of AgTiO2 composites without UV irradiation. Through modification of the substrate, combined with application of the PIERS technique, the SERS system for urea detection using Ag3TiO2 (50% TiO2) achieved a competitive detection limit of 4.6 × 10−6 M. It also allowed the detection of urea in milk at concentrations down to 10−5 M. This substrate modification and PIERS technique are promising for improvement of the sensing performance of other lowcrosssection molecules.
Trang 1Photoinduced Enhanced Raman Spectroscopy for the Ultrasensitive Detection of a Low-Cross-Section Chemical, Urea, Using
Silver−Titanium Dioxide Nanostructures
Quan Doan Mai,† Ha Anh Nguyen, *,† Thi Lan Huong Phung, Ngo Xuan Dinh, Quang Huy Tran,
Tri Quang Doan, Anh Tuan Pham, and Anh-Tuan Le *
Cite This:https://doi.org/10.1021/acsanm.2c03524 Read Online
powerful analysis technique that allows both the identification and
detection of analytes at trace levels However, the low rate of
charge transfer (CT) between noble-metal nanoparticles and
several analytes prevents them from being effectively detected by
SERS-based sensors They are regarded as low Raman
cross-section molecules In this study, we enhanced the performance of
the silver nanoparticles (AgNPs)-based SERS sensing platform for
a low Raman cross-section molecule, urea, focusing on improving
the rate of CT First, a set of Ag/titanium dioxide (TiO2)
nanocomposites were synthesized The presence of TiO2improved
the intensity of the SERS signal of urea, in comparison to the use
of bare AgNPs Second, a photoinduced enhanced Raman spectroscopy (PIERS) technique was employed to further elevate the
Raman signal of urea Thanks to the step of preirradiation using UV light at λ = 365 nm, with the use of the substrates containing
25%, 33%, and 50% TiO2content, enhancements of 1.93, 3.42, and 7.45 times were achieved, respectively, compared to the use of Ag/TiO2 composites without UV irradiation Through modification of the substrate, combined with application of the PIERS technique, the SERS system for urea detection using Ag/3TiO2(50% TiO2) achieved a competitive detection limit of 4.6 × 10−6M
It also allowed the detection of urea in milk at concentrations down to 10−5M This substrate modification and PIERS technique are promising for improvement of the sensing performance of other low-cross-section molecules
1 INTRODUCTION
Surface-enhanced Raman spectroscopy (SERS) has been
recognized as one of the most sensitive analysis techniques
that magnifies Raman signals of analytes, allowing them to be
detected even at the single-molecule level.1,2 Moreover,
originating form inelastic light scattering of analytes, SERS
provides information about their molecular structures,
there-fore, working as a powerful identification tool in analytical
chemistry.1,3 Various SERS-based applications have been
developed for environmental,4,5food,6,7 and health safety8−10
monitoring The adsorption of analytes on the surface of noble
metals such as Au and Ag has been proven to enhance their
Raman signals, thanks to localized surface plasmon resonance
of those metals.11 Under light excitation, the collective
oscillation of conductive electrons of the plasmonic materials
creates a strong magnetic field on the materials surface,
coupling with the vibrational modes of the analytes and,
therefore, enhancing their Raman signals.1,11 It is also the
principle of the most important contribution of Raman
enhancement in the SERS effect, the electromagnetic
mechanism In addition, despite a smaller contribution, the formation of a chemical complex between the analytes and metal surface, also known as the chemical mechanism (CM), cannot be ignored because charge-transfer (CT) transitions between the substrate and the molecule are also essential for the SERS phenomenon to occur.12−14These two mechanisms generate a giant enhancement in the Raman signal, allowing many kinds of analytes, including pesticides, food additives, biomarkers, etc., to be detected at trace levels.15−17
Unfortunately, such a powerful analytical tool is not effective
at detecting every organic analyte In a recent study, we reported on the role of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO)
Received: August 9, 2022
Accepted: September 28, 2022
© XXXX American Chemical Society
A
https://doi.org/10.1021/acsanm.2c03524
ACS Appl Nano Mater XXXX, XXX, XXX−XXX
Trang 2energy levels of the analyte on the SERS signal and CT
process.13 Our study proved that a large gap between the
LUMO energy of the analyte and the Fermi level of the
substrate could prevent the CT process For example,
4-nitrophenol (4-NP), which possesses a nitro group to directly
adsorb onto silver nanoparticles (AgNPs), did not show a large
enhancement in the SERS spectra, while a significant
enhancement was observed in the SERS spectra of another
nitrophenyl-substituted molecule, chloramphenicol (CAP)
The LUMO levels of 4-NP and CAP were calculated to be
−3.55 and −3.84 eV, respectively Therefore, the gap between
the LUMO level of 4-NP and the Fermi level of Ag (−4.26 eV)
was 0.71 eV, which was larger than that of CAP (0.42 eV).13
With a low level of CT, 4-NP can be regarded as a low Raman
cross-section molecule.18,19 Several other molecules such as
cysteine, CO2, adenosine triphosphate, and epidermal growth
factor receptor peptide have been reported to be
low-cross-section, resulting in low SERS signals.14,20,21
However, improving the SERS enhancement of the
molecules with low CT levels is not impossible In the report
mentioned above, we suggested modification of the
nanoma-terial to increase the Fermi level.13In a 2018 study, Tao et al
were successful in improving the rate of CT and enlarging the
cross section of the analyte by modifying the substrate with
WTe2, instead of employing the initial substrate of graphene.22
In addition, in a 2016 study, Parkin’s group reported on a
technique called photoinduced enhanced Raman spectroscopy
(PIERS), in which a step of preirradiation by ultraviolet (UV)
light for a period of time on gold nanoparticles (AuNPs) or
AgNPs deposited on a titanium dioxide (TiO2) rutile surface
was carried out before the SERS experiments.23 This
preirradiation step led to a several times enhancement of the
SERS signal This phenomenon is based on the interaction
between UV light and TiO2, leading to oxygen vacancy states
in the metal oxide semiconductor and an electron donor below
the conduction band edge Under laser excitation, electrons
can be injected into the Fermi level of the metal, increasing the
electron density on the surface of the nanoparticles (NPs) and allowing more electrons to be transferred to the analytes via the CT process.23
Urea, CO(NH2)2, is a redundant product of many organisms; therefore, it is naturally created inside their bodies and then released to the environment However, containing a high content of nitrogen, it has been synthesized and utilized in both agriculture and industry In agriculture, urea has been widely employed as a nitrogen-releasing fertilizer In industry,
it is used to produce cleaning products As a result, a high amount of urea has been released into the aquatic environ-ment, including both surface water and groundwater, every year.24 More dangerously, urea has been added illegally into dairy products to inflate the protein contents.25 Excessive amounts of urea in water and milk may cause kidney diseases
in human.26 In contrast, in the body, urea is released into blood from the liver and transferred to the kidneys, becoming a part of the urine In humans, the normal level of urea in blood
is reported to be 2.5−6.7 mM, while a high level of 30−150
mM represents pathophysiological states.26 Therefore, urea also plays the role of a biomarker for kidney diseases in humans Overall, the detection of urea is important for both environmental and food safety and health monitoring Analytical methods, such as gas chromatography27 and calorimetry,28 can detect urea at low levels Although they can provide sensitive, accurate, and stable results, they are time-consuming and laborious and require expensive instru-ments For more cost-effectiveness, electrochemical sensors for urea detection have been designed based on the use of urease.29 In the presence of this enzyme, the sensors are sensitive and specific; however, strict requirements of the experimental environment such as the temperature and pH have prevented them from being applied for a wide range of samples Nickel-based materials are excellent for nonenzymatic urea sensors; however, they were reported to involve reducing and expanding structure during the oxidation reaction to the target molecules.30Containing two primer amino groups and a Scheme 1 Schematic Illustration of the Synthesis Process of Ag/TiO2Nanocomposites
https://doi.org/10.1021/acsanm.2c03524
B
Trang 3low steric hindrance, urea possesses structural features that
allow them to effectively adsorb on the surface of noble-metal
NPs Therefore, it is expected that it can be detected at trace
levels by SERS sensors Surprisingly, in the literature, there are
a only few urea sensors based on the SERS technique.25,31−34
Moreover, their detection limits were not impressive
There-fore, urea should be one of the low-cross-section molecules,
which can be “unsuitable” for SERS detection
In this study, we aim to improve the performance of the
AgNPs-based SERS sensing platform for urea detection The
LUMO level of urea was calculated using electrochemical
cyclic voltammetry (CV) and compared to the Fermi level of
Ag and the LUMO levels of the published organic molecules to
prove that urea is a low Raman cross-section molecule This
was confirmed by the SERS signal of urea on AgNPs To
improve the enhancement of the SERS signal of urea, we first
modified the substrate by using Ag/TiO2composite
nanoma-terials instead of single AgNPs Four types of Ag/TiO2 with
varied TiO2contents, including 25%, 33%, 50%, and 75%, were
synthesized and employed for SERS measurement of urea The
utilization of composite nanomaterials containing 25%, 33%,
and 50% TiO2resulted in better enhancement compared to the
use of AgNPs These functional nanomaterials were then
employed for the PIERS technique to further improve their
SERS performance The results showed enhancements of 1.93,
3.42, and 7.45 times the SERS intensity when substrates
containing 25%, 33%, and 50% TiO2were preirradiated with
UV light of 365 nm, respectively Therefore, by using a noble
metal−metal oxide nanocomposite, combined with the PIERS
technique, we have improved the performance of SERS sensors
for urea, a low Raman cross-section molecule Urea could be
detected at concentrations as low as 4.6 × 10−6and 10−5M in
distilled water and milk, respectively The results confirmed the
importance of the CT process in the SERS signal and
suggested a direction for evaluating the rate of CT based on
material modification and preirradiation to generate advanced
SERS sensors for low Raman cross-section molecules
2 MATERIALS AND METHODS
2.1 Chemicals Silver nitrate (AgNO3, ≥99.0 wt %), sodium
borohydride (NaBH 4 , 99 wt %), titanium tetrachloride (TiCl 4 , ≥99.8
wt %), ammonium hydroxide (NH 4 OH, 28.0−30.0% NH 3 ), ethanol
(C 2 H 5 OH, 98 vol %), and urea (CH 4 N 2 O, 99 wt %) were purchased
from Shanghai Chemical Reagent and used directly without further
purification Double-distilled water was used throughout the
experi-ments.
2.2 Synthesis of Ag/TiO 2 Nanocomposite Materials and
Their Characterizations Ag/TiO2nanocomposites were prepared
via a facile wet chemistry method, as described in Scheme 1 and in
detail in our previous study.35 TiO 2 NPs were synthesized using a
modified sol−gel method from TiCl 4 precursors.36 An AgNO 3
solution was added to a set of variously prepared mixtures containing
different amounts of crystalline TiO 2 NPs calcined at 400 °C and 50
mL of C 2 H 5 OH Subsequently, NH 4 OH was slowly dropped into the
solution to completely reduce Ag+to Ag0 Finally, a set of Ag/TiO 2
nanocomposites was obtained with various of TiO 2 contents including
25, 33, 50, and 75 wt % and named Ag/1TiO 2 , Ag/2TiO 2 , Ag/3TiO 2 ,
and Ag/4TiO 2 , respectively.
The morphology of Ag/TiO 2 nanocomposites was studied using
scanning electron microscopy (SEM; Hitachi S-4800) operating
under an acceleration voltage of 5 kV The crystal structure of Ag/
TiO 2 was analyzed via X-ray diffraction (XRD; Bruker D5005 X-ray
diffractometer, Cu Kα, λ = 1.5406 Å) under a voltage of 40 kV and a
current of 30 mA The composition and chemical properties of Ag/
TiO 2 were investigated by Raman spectroscopy (Horiba
Macro-RAM) with 785 nm laser excitation The optical properties of Ag/ TiO 2 were also performed using photoluminescence (PL) spectros-copy with a 380 nm excitation wavelength.
2.3 CV Measurement CV measurements were performed using
a Palmsens 4 electrochemical workstation under ambient conditions The experiment was set up based on an established study with a Pt working electrode and an Ag/AgCl reference electrode.37 A 0.1 M phosphate buffer solution (PBS) served as the electrolyte All electrochemical potentials were referenced to an Fc/Fc+ internal standard Cyclic voltammograms of the analytes were performed at a scan rate of 50 mV s−1in the potential range from −2 to +2 V.
2.4 Substrate Preparation and SERS and PIERS Measure-ments Aluminum (Al) substrates were fabricated with dimensions of
1 cm × 1 cm × 0.1 cm with a surface-active area with a diameter of 0.2 cm The substrates were washed with ethanol and then dried naturally at room temperature (RT) A SERS active substrate (AgNPs
or Ag/TiO 2 ) solution was then drop-casted onto the surface-active area and dried at RT.
Solutions with various concentrations (10−6−10 −3 M) of urea were prepared in distilled water For each SERS measurement, the sample
was prepared as follows: 5 μL of an analyte solution was dropped
directly onto the prepared substrate and dried naturally at RT SERS spectra were recorded by a MacroRaman Raman spectrometer (Horiba) with 785 nm laser excitation Raman measurements were acquired by means of a 100× objective with a numerical aperture of 0.90 The laser power was set as 45 mW at a 45° contact angle, with a
diffraction-limited laser spot diameter of 1.1 μm (1.22λ/NA) and a
focal length of 115 nm For each measurement, the exposure time was
10 s with three accumulations The final spectrum was obtained after baseline calibration.
For each PIERS measurement, the prepared substrate was preirradiated using an UV source of 365 or 400 nm in 30 min Subsequently, the analyte solution was dropped directly onto the substrate and dried naturally Measurement was then carried out with the same procedure as that for the SERS measurements.
For real samples, bottled milk was purchased from a local supermarket in Hanoi, Vietnam, and utilized directly without further preparation Urea-spiked milk was obtained by adding an appropriate volume of an urea solution into the milk samples Subsequently, the spiked samples were dropped onto the prepared substrates for PIERS measurements as described above.
3 RESULTS AND DISCUSSION
3.1 Urea, a Low-Cross-Section Molecule The HOMO
and LUMO energy levels of urea can be estimated based on its
onset oxidation and reduction potentials (ϕox and ϕred), respectively, using the equations37−39
in which ϕFc/Fc+ is the redox potential of the ferrocene/ ferrocenium couple (Fc/Fc+) in the electrochemical system, assuming that the energy level of Fc/Fc+is −4.8 eV below the vacuum level Similar to the previous study, we set up an electrochemical system based on the description of Bin et al with a Pt working electrode and a Ag/AgCl reference electrode.37 Therefore, ϕFc/Fc+ was assumed to be 0.44 V versus Ag/AgCl.37 However, as discussed in the previous study, while the distinction in the LUMO levels of the analytes led to a significant difference in their SERS intensities, the effects of the HOMO level were not clear.13Therefore, in this study, we only focus on the reduction potential to calculate the LUMO level of urea
CV measurements of urea were performed in 0.1 M PBS
observed In the presence of urea, an irreversible cathodic peak
https://doi.org/10.1021/acsanm.2c03524
C
Trang 4appears at −1.32 V It is the onset reduction peak of urea The
onset reduction potential (ϕred) of urea was estimated to be
−0.86 V Usingeq 2, we determined the LUMO level of urea
to be −3.50 eV The values are averaged over 5 cycles of CV
scans (Figure S2)
The LUMO level of urea forms a relatively large gap (0.76
eV) with the Fermi level of Ag (−4.26 eV) This gap is even
larger than that of 4-NP (0.71 eV), which was reported to have
a low SERS signal on the Ag substrate.13 This large gap
suggested a low rate of CT between AuNPs and urea in SERS
experiments, which limits the cross section of the molecule In
other words, urea is a low-cross-section molecule Another
example of a low-cross-section molecule, cysteine, was
mentioned in a study of Fu et al., in which the authors
claimed to know about its weak SERS signal.20 Therefore,
SERS measurements were performed to confirm our
hypothesis Figure S1b shows the SERS spectra of urea on
AgNPs As a simple organic molecule containing one carbonyl
group (>C�O) and two amino groups (−NH2), the presence
of urea absorbed on SERS substrates resulted in simple SERS
spectra with only one characteristic band at 1010 cm−1,
representing the C−N stretching mode.40 Although the
presence of −NH2 groups allowed urea to bind directly on
the surface of the Ag substrate, the low intensity of this
dominant peak can be observed, even at a relatively high
concentration of 10−3 M Hence, urea can be regarded as a
low-cross-section molecule
However, the cross section can be enlarged by increasing the
rate of CT.22 In a 2021 study, we designed an experimental
model using methylene blue as the analyte to prove that Ag/
TiO2 contacts in Ag/TiO2 nanocomposites could enhance
electron transfer, leading to improvement of the SERS
intensity.35Therefore, in the effort of improving the
perform-ance of SERS sensors for urea, we utilized a set of Ag/TiO2
nanocomposites with different TiO2contents
3.2 Characterization of Ag/TiO 2 Nanocomposites.
Four types of Ag/TiO2 composite nanomaterials, including
Ag/1TiO2, Ag/2TiO2, Ag/3TiO2, and Ag/4TiO2, containing
25%, 33%, 50%, and 75% TiO2 content, respectively, were
synthesized as described insection 2 The content of TiO2in
the nanocomposites was calculated based on the weight of
TiO2nanostructures and AgNO3precursors added during the
synthesis Their SEM images show the difference in the TiO2/
Ag ratio of distinct nanocomposites (Figure 1a−d) Two types
of materials with different sizes of 43 and 125 nm were detected in the SEM images As revealed by the small size of the AgNPs inFigure S3, the smaller nanoobjects in these SEM images can be AgNPs while the larger ones can be TiO2NPs From Ag/1TiO2 to Ag/4TiO2, the density of the larger nanostructures increases while that of the smaller ones decreases, implying a rise in the TiO2/Ag ratio in those nanocomposites Moreover, their energy-dispersive spectros-copy (EDS) spectra provided qualitative and semiquantitative results of TiO2and Ag in the nanostructures.Figure S4shows the EDS spectra of Ag/1TiO2, Ag/2TiO2, Ag/3TiO2, and Ag/ 4TiO2with the contents of TiO2estimated to be 23%, 34%, 45%, and 78%, respectively, which are in agreement with the initially calculated TiO2contents The XRD results of the as-prepared materials also indicated the presence of Ag and TiO2
in the nanocomposites Figure 2a shows the XRD pattern of Ag/TiO2 nanocomposites in comparison to the reference patterns of TiO2and Ag The main peaks corresponding to Ag (ICDD 01-087-0597) as well as TiO2(ICDD 01-086-1157), both anatase and rutile, appear in the XRD pattern of Ag/ TiO2 This is evidence for the presence of AgNPs in the TiO2 crystal matrix On the basis of their diffraction peaks, the average crystal grain sizes of AgNPs and TiO2were calculated
to be 14 and 19 nm, using the Scherrer formula.41 The presence of AgNPs in the TiO2 crystal matrix was also
confirmed by the PL measurements (λext = 380 nm; Figure
2c,d) Upon a decrease of the TiO2content, from pure TiO2 (100% TiO2) to Ag/1TiO2 (25% TiO2), the Ag content increases, resulting in a decrease of the luminescence intensity
of the composite materials The contact between AgNPs and TiO2NP should have led to luminescence quenching These results also confirmed the difference in the TiO2 (and Ag) content in the composite materials In addition, Figure 2b demonstrates the Raman spectra of Ag/TiO2nanocomposites
in comparison to that of TiO2NPs A red shift of 4−6 nm can
be observed at the dominant peak of TiO2(146 cm−1in the TiO2Raman spectrum) in the presence of AgNPs at different ratios This band represents the Eg optical Raman mode of anatase TiO2.42 Upon an increase of the Ag content in the nanocomposites, this peak is shifted further toward longer wavelengths This indicates the incorporation between AgNPs and the crystal structure of TiO2NPs Information about other characteristic peaks of the nanocomposite materials will be further discussed in the following part
3.3 SERS Sensing Performance of Ag/TiO 2 Nano-composites to Detect Urea. Figure 3a shows the SERS spectra of urea on five different substrates, including AgNPs and Ag/1TiO2, Ag/2TiO2, Ag/3TiO2, and Ag/4TiO2 composites Compared to AgNPs, with all TiO2-containing substrates, the spectra show characteristic bands of TiO2NPs The strong band at 150 cm−1represents the Egoptical Raman mode of anatase TiO2 In addition, the bands at 398, 516, and
638 cm−1are assigned to the B1g, A1g, and Egmodes of anatase TiO2, respectively.42,43 A band at 238 cm−1 appears in all samples It could be associated with the Ag−N stretching mode.44,45 It can be observed that from Ag/1TiO2 to Ag/ 4TiO2, the intensity of the band at 150 cm−1increases with a decrease of that at 238 cm−1 These changes are associated with the ratio of the TiO2 and Ag contents in the nanocomposites because Ag/1TiO2, Ag/2TiO2, Ag/3TiO2, and Ag/4TiO2 contain 25%, 33%, 50%, and 75% of TiO2, respectively Concerning the characteristic band of urea at
1010 cm−1, it is obvious that employing Ag/1TiO2, Ag/2TiO2,
Figure 1 SEM images of (a) Ag/1TiO 2 , (b) Ag/2TiO 2 , (c) Ag/
3TiO 2 , and (d) Ag/4TiO 2 powders.
https://doi.org/10.1021/acsanm.2c03524
D
Trang 5and Ag/3TiO2nanocomposites as SERS substrates resulted in
a higher intensity of the SERS signal of urea, in comparison to
the use of AgNPs The presence of TiO2 is the key for
improvement of the CT rate because employing substrates
containing high percentages of TiO2 (i.e., 33% and 50%)
resulted in a higher SERS intensity than using a substrate
containing a lower percentage of TiO2 (i.e., 25%) However,
on the Ag/4TiO2 substrate, we could not detect the
characteristic peak of urea By increasing the TiO2 content
up to 75%, we decreased the Ag content Because the presence
of AgNPs is essential for urea to experience the SERS effect,
this low Ag content might have limited the SERS performance
of the material The disappearance of the band at 238 cm−1
also represents this low Ag content Because the substrate of
Ag/4TiO2did not improve the SERS performance of the urea
sensors compared to AgNPs, this material was not employed
for the following experiments
Because the difference in the SERS intensity of urea while
using Ag/2TiO2and Ag/3TiO2is not clear, we selected one of
those two, Ag/3TiO2, to evaluate the sensitivity of Ag/TiO2
substrates in the detection of urea Five samples of urea in
water were prepared at different concentrations, from 10−3to
10−5 M Subsequently, their SERS spectra were recorded
cm−1increased with an increase of the urea concentration The
plot of the logarithmic SERS intensity at 1010 cm−1against the
urea concentration within that range is shown in Figure 3c,
representing a good linear relationship with a linear regression
of 0.97 Thanks to the presence of TiO2, the urea sensor based
on the Ag/3TiO2 nanocomposite had a limit of detection (LOD) of 4.2 × 10−5M, which is lower than those of most of the reported noble-metal-based SERS sensors for urea (Table
1) The calculation of the LOD is shown in the Supporting
In addition, the reproducibility of the method was studied as five Ag/3TiO2substrates were prepared independently, using the same protocol, to measure the SERS signal of urea (10−3
calculated to be 8.53%, indicating good reproducibility of the sensor
3.4 PIERS Sensing Performance of AgNPs and Ag/ TiO 2 Nanocomposites to Detect Urea In a continuation of
the effort to further improve the SERS performance of the urea sensors, the PIERS technique was applied The TiO2 -containing substrates, including Ag/1TiO2, Ag/2TiO2, and Ag/3TiO2, experienced UV exposure for 30 min Subse-quently, the SERS signals of urea (10−3M) were collected on those UV-preirradiated substrates To confirm the importance
of TiO2, a similar experiment was carried out on the substrate
of AgNPs
In the absorption spectrum (Figure S5), TiO2NPs exhibit a broad spectrum in the UV region In several previous reports, authors selected UVC light (254 nm) for UV irradiation for TiO2-containing nanocomposites.23,46 In this study, we
selected the UV source with λ = 365 nm, which lies within
the absorption region of TiO2NPs In addition, the success of
Figure 2 (a) XRD patterns of Ag/3TiO 2 (b) Raman spectra of TiO 2 , Ag/1TiO 2 , Ag/2TiO 2 , Ag/3TiO 2 , and Ag/4TiO 2 (c) PL spectra of TiO 2 , Ag/1TiO 2 , Ag/2TiO 2 , Ag/3TiO 2 , and Ag/4TiO 2 (d) Zoomed-in view of the PL spectra of Ag/1TiO 2 and Ag/2TiO 2
https://doi.org/10.1021/acsanm.2c03524
E
Trang 6Ke et al using a xenon ozone-free arc lamp with λ > 400 nm
for irradiation of an Au-TiO2NPs system47triggered us to try
another UV light of longer wavelength Thus, we also selected
UV light with λ = 400 nm, which is the longest wavelength
within the UV range However, the absorption of TiO2at this
wavelength was not as good as that at λ = 365 nm The PL
results for each UV source are presented inFigure S6
Concerning the substrates irradiated by UV light with λ =
365 nm, similar to the SERS signals without UV irradiation
effect in comparison to AgNPs (Figure 4a) However, it is
noticeable that, after UV exposure, the SERS performance of
Ag/3TiO2 significantly surpassed that of Ag/2TiO2 This
difference was not observed without UV exposure Therefore,
it should have been the effect of this UV-irradiation step
To clarify the effect of UV irradiation on each substrate, we compared the SERS signal of urea on each substrate with and without the UV-irradiation step Figure 4b demonstrates that
UV exposure did not have any effect on AgNPs because no significant change was observed in the SERS signal In contrast,
a significant increase in the intensity was observed in the SERS spectra of urea on three TiO2-containing substrates, including Ag/1TiO2, Ag/2TiO2, and Ag/3TiO2, thanks to the UV-irradiation step (Figure 4c−e) Therefore, the PIERS phenomenon has occurred Moreover, this phenomenon is directly related to the presence of TiO2in the nanocomposite materials The intensities of the peak of 1010 cm−1of urea on the Ag/1TiO2, Ag/2TiO2, and Ag/3TiO2substrates after UV irradiation were calculated to be 1.93, 3.42, and 7.45 times higher than those without UV irradiation, respectively A
Figure 3 (a) SERS spectra of urea on five substrates of AgNPs and Ag/1TiO 2 , Ag/2TiO 2 , Ag/3TiO 2 , and Ag/4TiO 2 nanocomposites (b) SERS spectra of urea (10−5−10 −3 M) on Al/3TiO 2 (c) Plot of the logarithmic SERS intensity at 1010 cm−1against the urea concentration (slope 0.71 ± 0.02; intercept 5.58 ± 0.02) (d) Reproducibility of the SERS sensor for urea detection based on Ag/3TiO 2 /Al substrates.
Table 1 Several Reported SERS-Based Urea Sensors
material functionalization LOD (M) linear range (M) laser (nm) ref
Ag dendrite none 3.3 × 10 −3 3.3 × 10 −3 −1.7 × 10 −2 532 31
Au nanostar 11- mercaptoundecanoic acid + urease 3.3 × 10−2 3.3 × 10−2−3.3 × 10 −1 525 32
Ag/3TiO 2(UV irradiation, λ = 365 nm) none 4.6 × 10 −6 10 −6 −10 −3 785 this work Ag/3TiO 2(UV irradiation, λ = 400 nm) none 4.28 × 10−5 10−5−10 −3 785 this work
https://doi.org/10.1021/acsanm.2c03524
F
Trang 7higher content of TiO2resulted in better enhancement in the
SERS signal after UV irradiation With the most impressive
enhancement, Ag/3TiO2 was selected for the following
experiments
Thanks to the PIERS effect, Ag/3TiO2 preirradiated with
UV light (λ = 365 nm) exhibited better sensitivity than that
without UV exposure The SERS spectra of seven samples of
urea in water at different concentrations, from 10−3to 10−6M,
are shown in Figure 5a The plot of the logarithmic PIERS
intensity at 1010 cm−1against the urea concentration within that range is demonstrated inFigure 5b with a linear regression
of 0.96 The equation inFigure 5b was used to calculate the LOD of this sensor, resulting in a LOD of 4.6 × 10−6M, which
is lower than that without the PIERS effect Moreover, this is also a competitive LOD in comparison to other SERS-based sensors for urea detection (Table 1), and the linear range was also enlarged compared to that of the initial SERS sensors
Figure 4 (a) PIERS spectra of urea (10−3M) on five substrates: AgNPs, Ag/1TiO 2 , Ag/2TiO 2 , and Ag/3TiO 2(UV irradiation, λ = 365 nm).
Comparison of the PIERS spectra of urea on AgNPs (b), Ag/1TiO 2 (c), Ag/2TiO 2 (d), and Ag/3TiO 2 (e) with (green) and without (blue) UV irradiation.
https://doi.org/10.1021/acsanm.2c03524
G
Trang 8In addition, the reproducibility of the sensor was also
investigated using five independently prepared Ag/3TiO2
substrates (Figure 5c) The sensor exhibited good
reproduci-bility because the RSD was calculated to be 8.56%
However, with the other UV source (λ = 400 nm), the
PIERS phenomenon did not occur Figure S7a compares the
SERS spectra of urea (10−5−10−3 M) on Ag/3TiO2 without
nm) No significant change was observed In fact, the LOD of
urea detection in this UV-irradiated sensor was calculated to be
4.28 × 10−5M, which was only slightly higher than that of the
non-UV-irradiated sensor Therefore, UV light with λ = 400
nm could not trigger the PIERS effect in our Ag/TiO2-based
sensing system
In addition to the preirradiation source, the effects of the
preirradiation time were also investigated Figure S8
demonstrates the Raman spectra of urea (10−3 M) on Ag/
3TiO2 substrates without and with UV irradiation (λ = 365
nm) for 10, 20, 30, 40, and 60 min It is obvious that
elongating the UV exposure time from 10 to 30 min improved
the performance of Ag/3TiO2substrates because the intensity
of the band at 1010 cm−1increased rapidly However, longer
exposure times, such as 40 and 60 min, did not cause any
significant enhancement in the Raman signal, in comparison to
30 min of preirradiation The experiments were repeated three
times (Figure S9) Hence, elongating the preirradiation time
might enhance the PIERS effects on the Ag/3TiO2 nano-composite; however, it could reach the limit in 30 min The persistence of the PIERS phenomenon on the Ag/ 3TiO2 nanocomposite was studied by collecting a series of PIERS spectra of urea at 5, 15, 30, 45, and 60 min of UV preirradiation It took at least 5 min for the urea solution to completely dry on the substrate; therefore, the first measure-ment was performed 5 min after UV exposure Subsequently, the intensity of the PIERS signal decreased with relaxing time
phenomenon nearly disappeared because the PIERS intensity was then only slightly higher than the SERS intensity (Figure S10b) The experiments were repeated three times (Figure
3.5 Proposed Mechanism of the PIERS Phenomenon
on Ag/TiO 2 Nanocomposites and Effects of the Irradiation Wavelength. Figure S12a demonstrates the structure of TiO2 before it was irradiated with UV light Its conduction band minimum (CBM) and valence band maximum (VBM) energy levels are −4.3 and −7.5 eV, creating a band gap of 3.2 eV.23 It has been reported that
irradiating UV light (λ = 365 nm) could create “line defects”
along the (001) direction of TiO2 because of cooperative oxygen removal,48causing oxygen vacancies on the surface of the TiO2semiconductor (Figure S12b) Subsequently, electron donor states are created at approximately 0.7 eV below its conduction band.23 However, only UV sources with the
Figure 5 (a) PIERS spectra of urea (10−6−10 −3 M) on Ag/3TiO 2after UV irradiation (λ = 365 nm) (b) Plot of the logarithm of PIERS intensity
versus concentration at 1010 cm−1(slope 0.71 ± 0.02; intercept 5.58 ± 0.02) (c) Reproducibility of the PIERS sensor for urea detection based on Ag/3TiO 2 /Al substrates.
https://doi.org/10.1021/acsanm.2c03524
H
Trang 9appropriate wavelength could separate the electron−hole pairs
and send electrons to these states Almohammed et al
proposed that the energy of the incident photons had to be
equal or larger than the band gap of the semiconductor to
achieve an enhancement in the Raman signal after
preirradiation with UV light and then laser excitation in the
SERS experiment.49
The energies of two UV sources with λ = 365 and 400 nm
were calculated to be 3.4 and 3.1 eV, respectively, using the
well-known formula
=
where λ is the wavelength, c is the speed of light in a vacuum,
and h is the Planck constant The first one (λ = 365 nm) has a
photon energy larger than the band gap of TiO2 (3.2 eV),
while the energy of the other (λ = 400 nm) is lower than that
band gap This explains the reason why, despite a small
difference in the wavelength (44 nm), UV light at λ = 365 nm
could trigger the PIERS effect, while the other could not
Preirradiation with UV light at λ = 400 nm can be suitable for
other semiconductors with narrower band gaps such as WO3
(∼2.6 eV),50WS2(1.0−2.1 eV),51etc
After UV irradiation at a suitable wavelength, the substrate
then experienced excitation of a laser source (785 nm) Hence,
the CT process in the PIERS phenomenon followed two steps,
as demonstrated inFigure 6b (1) Thanks to UV irradiation,
the electrons jump from the oxygen vacancy states to the
conduction band of TiO2, and then they are injected into the
Fermi level of Ag.23,52As a result, the density of hot electrons
on the surface of AgNPs increases (2) Thus, more electrons
can be transferred to the analyte on the surface of the
plasmonic NPs Therefore, the obtained Raman signal is
magnified compared to the normal SERS phenomenon (Figure
6a) Because cooperative oxygen removal only occurs on the
surface of the nanocomposites, it gradually increases, leading to
a rise in the PIERS intensity However, it reaches the
saturation state by 30 min of UV irradiation Further
elongating the exposure time leads to a negligible change in
the intensity of the signal In addition, the defects on the
surface of the nanocomposites caused by UV irradiation can be
gradually healed upon exposure of the substrate to air, leading
to a slow decay of the PIERS phenomenon.23 This surface
healing may also have occurred because of the addition of an
analyte solution However, this step was unavoidable because
UV exposure for 30 min might have damaged the organic analyte if the analyte solution was drop-casted onto the substrate prior to UV irradiation Therefore, to achieve the most effective enhancement of the PIERS signal, measurement should be performed as soon as the analyte solution is dried on the preirradiated substrate
Concerning the CT process, the PIERS effect relies on CM enhancement and contributes to part to this enhancement Thus, the PIERS enhancement could not be too high (less than 10 times); however, it obviously improved the sensitivity (i.e., LOD and linear range) of the SERS-based sensing platform for the detection of a low Raman cross-section molecule, urea
3.6 Selectivity and Practicability of a Ag/3TiO 2 -Based PIERS Sensor for Urea Detection To investigate the
selectivity of a Ag/3TiO2-based PIERS sensor for urea detection, we performed PIERS measurements of urea in the presence of interfering compounds that are possibly present in food and clinic samples, including glucose, ascorbic acid, and hydrogen peroxide (H2O2) A solution containing those three interfering compounds was prepared in distilled water with a concentration of 10−4 M for each compound Subsequently, urea was added to obtain concentrations of 10−4and 10−5M Because the nanocomposite was synthesized using a simple procedure without any specific functionalization, the PIERS sensor was not expected to exhibit high specificity and selectivity However, within the PIERS spectrum of the mixture containing 10−4 M urea, the characteristic band of urea (1010 cm−1) was still detectable in the presence of characteristic peaks of the interfering compounds, such as
H2O2(885 cm−1),53ascorbic acid (1127 cm−1),54and glucose (928, 1090, and 1127 cm−1)55 (Figure S13a) The PIERS spectra of the mixtures containing 10−4and 10−5M urea were compared with the PIERS spectra of urea in water at the same concentrations (10−4and 10−5M), as shown in parts b and c
can be detected in the complex spectra of the mixture, representing the presence of urea Although the presence of interfering compounds might have prevented urea from accessing the Ag/3TiO2 surface, leading to a slight decrease
in the intensity of the 1010 cm−1band, the nanocomposite still exhibited the ability to detect the target analyte of urea at Figure 6 SERS (a) and PIERS (b) phenomena with substrates of the Ag/TiO 2 nanocomposite materials.
https://doi.org/10.1021/acsanm.2c03524
I
Trang 10concentrations as low as 10−5M in the presence of interfering
compounds
The practicability of a Ag/3TiO2-based PIERS sensor was
studied by spiking urea into bottled milk samples to obtain
different concentrations (10−6−10−3M) Before spiking with
urea, the Raman spectra of the milk sample were recorded In
the absence of the Ag/3TiO2 substrate, the milk sample
showed no characteristic peak in the Raman spectrum (black;
several vibrational bands can be detected in the spectrum (red), including characteristic bands of TiO2 and a few new ones at 238, 308, and 551 cm−1 They might be due to the presence of protein in milk Upon the addition of urea, the characteristic band at 1010 cm−1 appeared (Figure 7b) This
Figure 7 (a) Raman spectrum of a milk sample (red) on a Ag/3TiO 2 substrate compared to those of TiO 2 (blue) and a milk sample in the absence
of Ag/3TiO 2 (black) (b) PIERS spectra of urea at different concentrations (10−6−10 −3 M) in milk on Ag/3TiO 2 substrates.
Figure 8 (a) SERS spectra of 4-NP (10−5−10 −3 M) on Ag/3TiO 2 substrates (b) Comparison of the PIERS and SERS spectra of 4-NP on Ag/ 3TiO 2 (c) PIERS spectra of 4-NP (10−6−10 −3 M) on Ag/3TiO 2 substrates (d) Plot of the logarithmic PIERS intensity of 4-NP against the 4-NP concentration at 640 cm−1.
https://doi.org/10.1021/acsanm.2c03524
J