Great influences of the lowest unoccupied molecular orbital (LUMO) and highest occupied molecular orbital (HOMO) energy levels of the analyte and their alignments compared to the Fermi level of the substrate on the charge transfer (CT) process, and consequently, on the surfaceenhanced Raman scattering (SERS) phenomenon have been described via theoretical calculations. To provide experimental evidence, in this study, two antibiotics, chloramphenicol (CAP) and amoxicillin (AMX), were investigated as analytes in SERS sensors based on electrochemically synthesized colloidal silver nanoparticles (eAgNPs) as the substrate. Despite the same experimental condition, similarities in analyte structure, and in the ability of absorbing onto eAgNPs, the detection of the two antibiotics showed obvious distinction. While CAP was able to be detected using eAgNPbased SERS sensors at concentrations down to 1.2 × 10−9 M, there were no characteristic peaks observed in the SERS spectra of AMX even at a high concentration of 10−3 M. The LUMO and HOMO energy levels of the two analytes were measured using electrochemical cyclic voltammetry. The obtained results showed that the LUMO levels of both analytes were higher than the Fermi level of Ag, and the LUMO level of AMX was higher than that of CAP. The larger gap between the LUMO level of AMX and the Fermi level of Ag might have prevented the metaltomolecule CT process, which is related to the Raman signal enhancement in both chemical and electromagnetic mechanisms. In contrast, the smaller energy gap in the case of CAP might have allowed the transfer of hot electrons from the Fermi level of the eAgNPs to the LUMO level of the analyte. Therefore, CAP could experience an SERS effect on the eAgNPs under the excitation of a 785 nm laser source, while AMX could not. The hypothesis was then confirmed using three other organic compounds, including furazolidone, 4nitrophenol, and tricyclazole. The results revealed a clear correlation between the LUMO level of the analytes and their SERS signals.
Trang 1Silver Nanoparticles-Based SERS Platform towards Detecting
Chloramphenicol and Amoxicillin: An Experimental Insight into the
Signal and Charge Transfer Process
Quan Doan Mai,# Ha Anh Nguyen, *, # Thi Lan Huong Phung, Ngo Xuan Dinh, Quang Huy Tran, Tri Quang Doan, and Anh-Tuan Le *
Cite This: https://doi.org/10.1021/acs.jpcc.2c01818 Read Online
ACCESS Metrics & More Article Recommendations *s ı Supporting Information
molec-ular orbital (LUMO) and highest occupied molecmolec-ular orbital
(HOMO) energy levels of the analyte and their alignments
compared to the Fermi level of the substrate on the charge transfer
(CT) process, and consequently, on the surface-enhanced Raman
scattering (SERS) phenomenon have been described via theoretical
calculations To provide experimental evidence, in this study, two
antibiotics, chloramphenicol (CAP) and amoxicillin (AMX), were
investigated as analytes in SERS sensors based on electrochemically
synthesized colloidal silver nanoparticles (e-AgNPs) as the
substrate Despite the same experimental condition, similarities in
analyte structure, and in the ability of absorbing onto e-AgNPs, the
detection of the two antibiotics showed obvious distinction While CAP was able to be detected using e-AgNP-based SERS sensors
at concentrations down to 1.2× 10−9M, there were no characteristic peaks observed in the SERS spectra of AMX even at a high concentration of 10−3M The LUMO and HOMO energy levels of the two analytes were measured using electrochemical cyclic voltammetry The obtained results showed that the LUMO levels of both analytes were higher than the Fermi level of Ag, and the LUMO level of AMX was higher than that of CAP The larger gap between the LUMO level of AMX and the Fermi level of Ag might have prevented the metal-to-molecule CT process, which is related to the Raman signal enhancement in both chemical and electromagnetic mechanisms In contrast, the smaller energy gap in the case of CAP might have allowed the transfer of hot electrons from the Fermi level of the e-AgNPs to the LUMO level of the analyte Therefore, CAP could experience an SERS effect on the e-AgNPs under the excitation of a 785 nm laser source, while AMX could not The hypothesis was then confirmed using three other organic compounds, including furazolidone, 4-nitrophenol, and tricyclazole The results revealed a clear correlation between the LUMO level of the analytes and their SERS signals
1 INTRODUCTION
Surface-enhanced Raman spectroscopy (SERS) is a Raman
technique based on plasmonic materials, which has been
developed for ultrasensitive detection of various analytes at low
concentrations, and even at single-molecule level.1−3 Under
the excitation of light, collective oscillations of the conductive
electrons of the plasmonic materials generate a strong
electromagnetic field on the material surface, which couples
with the vibrational modes of the analyte adsorbed on the
surface and leads to enhancement of its characteristic Raman
signal.1,4 A strong electromagnetic field and formation of
chemical complexation of the analyte to the metallic surface are
both essential for the SERS phenomenon They are also the
main principles of the two fundamental mechanisms explaining
this giant enhancement of Raman signal: electromagnetic and
chemical mechanisms (EM and CM), respectively.1,4
There-fore, to optimize the SERS performance, many strategies have been used targeting these two mechanisms, focusing on material fabrication and surface modification For example, on one hand, EM enhancement increases in several specific regions, called hotspots, including the sharp tips of nano-particles (NPs) and the nanogaps (<10 nm) between them.5 Therefore, many studies have focused on designing and developing plasmonic NPs with sharp tips, such as nanostars,6 nanodendrites,7nanotriangles,8etc Moreover, other shapes of
Received: March 16, 2022 Revised: April 16, 2022
Article
pubs.acs.org/JPCC
© XXXX American Chemical Society https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX
Trang 2NPs, such as nanocubes and nanowires, were also promising
SERS substrates thanks to their potential of creating nanogaps
via self-assembly.9 These nanogaps could be created using
polymer-assisted assembly,10DNA origami,11protein
anchor-ing,12 etc On the other hand, researchers have also paid
attention to improving the formation of molecule−metal
complexes by modifying the NPs with capping agents allowing
the binding of the desired analytes via electrostatic
interaction,13 steric effect,14
host−guest interaction,15
bio-logical recognition,16etc
With the aim of improving the efficiency of SERS sensing
systems, recently, other factors concerning SERS signal
enhancement have been investigated For instance,
Álvarez-Puebla reported on the effects of excitation wavelength on the
SERS spectrum, based on both theoretical calculations and
experimental data The study indicated that light excitation has
impacts on not only the electromagnetic field around the
excited silver nanoparticles (AgNPs) but also the charge
transfer process via CM between the analyte and the metallic
surface.2 Besides, another factor, substrate surface oxidation,
was reported to reduce SERS performance in both CM and
EM.17,18 In addition, analyte molecular structure can also
influence the performance of an SERS sensor as a bulky
structure and large steric hindrance might limit chemical and
electromagnetic enhancements.19
The Raman signal of an organic molecule is the result of an
electron relaxation from the excited state down to the ground
state These excited and ground states of an organic molecule
are also known as the lowest unoccupied molecular orbital
(LUMO) and highest occupied molecular orbital (HOMO)
energy levels, respectively Therefore, the SERS signal obtained
from an organic molecule is also directly related to its
HOMO−LUMO energy levels Through theoretical
calcu-lations, many authors have described the importance of
HOMO−LUMO energy levels of the analyte in the
enhance-ment of its Raman signal, in both CM and EM.2,20In a 2017
study, Otero et al reported on the relationship between the
Fermi level of the metal, HOMO−LUMO levels of the organic
molecule, and the charge transfer (CT) process between
them.21 CT is related to the chemical complexation of the
analyte and the metal surface, involving a transition between
the Fermi level of the metal and the LUMO level of the analyte
due to laser excitation.2,20 Those studies have revealed the
undeniable importance of HOMO−LUMO levels of the
analyte and their alignment compared to the Fermi level of
the substrate in the enhancement of Raman signal in both CM
and EM However, to the best of our knowledge, there have
been no experimental model generated to confirm the effects
of HOMO−LUMO energy levels on SERS performance
In this study, we experimentally investigated the detection of
two antibiotics, chloramphenicol (CAP) and amoxicillin
(AMX), using SERS sensors based on electrochemically
synthesized colloidal AgNPs (e-AgNPs) With molecular
weights of 365 and 323 g/mol, respectively, both CAP and
AMX are large molecules Furthermore, their pharmacological
activities require the presence of many functional groups in the
two analytes, which results in their high steric hindrance
However, the primer amine and carboxyl groups in AMX, as
well as electron-rich regions in CAP, allow them to adsorb on
e-AgNPs Their adsorption onto e-AgNPs was confirmed by
ultraviolet−visible (UV−vis) Unexpectedly, despite their
similarities, CAP could be detected by e-AgNP-based SERS
sensors at concentrations down to 1.2 × 10−9 M, while no
characteristic peak was observed in the SERS spectra of AMX, even at a high concentration of 10−3M It is worth stressing that the SERS measurements for CAP and AMX were carried out under the same conditions To analyze whether this significant difference is due to the difference in LUMO and HOMO energy levels of the two analytes, we measured their LUMO and HOMO energy levels using electrochemical cyclic voltammetry (CV) and compared them to the Fermi level of
Ag Our calculations showed that both of the LUMO energy levels were higher than the Fermi level of Ag and the LUMO level of AMX is higher than that of CAP, so the LUMO of CAP is closer to the Fermi level of Ag than that of AMX The results confirmed the CT transition between the Fermi level of the metal nanostructure and the LUMO of the adsorbed molecules The large energy gap between the LUMO level of AMX and the Fermi level of e-AgNPs might have prevented the CT process In contrast, the small gap in the case of CAP might have allowed CT transition, leading to the enhancement
in the SERS spectra Therefore, an analyte can be detected
effectively when its LUMO is close to the Fermi level of the substrate This hypothesis was then confirmed using three other organic compounds, including furazolidone, 4-nitro-phenol, and tricyclazole The HOMO−LUMO levels of these compounds were calculated via CV measurements, based on which we made some predictions about their SERS signals SERS measurements were, subsequently, carried out, and the actual signals were revealed to be in agreement with the predicted ones Hence, the step of measuring the LUMO level
of the analytes is recommended when choosing the appropriate substrates to minimize the energy gap and develop highly
effective SERS sensors in the future
2 METHODS
2.1 Chemicals Sodium citrate (Na3C6H5O7, 99.9%), chloramphenicol (C11H12Cl2N2O5, ≥98.0%), and amoxicillin (C16H19N3O5S, ≥98.0%) were purchased from Shanghai Chemical Reagent and used directly without further purification Two silver plates (purity: 99.99%) were prepared, with dimensions of 100 mm × 5 mm × 0.5 mm Double distilled water was used throughout the experiments
2.2 Electrochemically Synthesized Silver Nanopar-ticles and Their Characterizations The electrochemical syntheses of e-AuNPs were carried out in a beaker containing
200 mL of 0.1% Na3C6H5O7 (acting as both surfactant and electrolyte) in distilled water The electrodes were mechan-ically polished and washed with distilled water to eliminate the oxides of the surfaces before the electrochemical processing Electrolysis was done at room temperature (RT) under uniform magnetic stirring at 200 rpm for 2 h A solution of colloidal e-AgNPs was then obtained with a dark gray color The shapes and sizes of the e-AgNPs were analyzed by a scanning electron microscope (SEM, Hitachi S-4800) operated
at an acceleration voltage of 5 kV, which revealed the spherical shape of the NPs with the average diameter of 24 nm UV−vis absorption spectra were recorded using a JENWAY 6850 spectrophotometer, and 10 mm path length quartz cuvettes were used for the measurements of absorption ranges The as-prepared e-AgNPs exhibited a plasmonic band at ∼404 nm Full information about the NP characterization was reported in our previous study.22
2.3 SERS Substrate Preparation and Measurement Each aluminum (Al) substrate was prepared with dimensions
of 1 cm× 1cm × 0.1 cm, containing a surface-active area with
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX B
Trang 3a diameter of 0.2 cm The substrate was cleaned with ethanol,
then dried naturally at RT Some droplets of the colloidal
e-AgNP solution were placed on the surface-active area, followed
by natural evaporation of water at RT
To evaluate the sensitivity, reproducibility, and repeatability
of the sensors, solutions with various concentrations of CAP
(10−3−10−8M) and AMX (10−3−10−5M) were prepared in
distilled water The SERS sample was prepared as follows: 5μL
of analyte solution was dropped directly onto the e-AgNP/Al
substrate and dried naturally at RT SERS spectra were
monitored by a MacroRamanTM Raman spectrometer
(Horiba) with a 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 to be 45 mW at 45°
contact angle, with a diffraction-limited laser spot diameter of
1.1μm (1.22λ/NA) and focal length of 115 nm The exposure
time for each measurement was 10 s with three accumulations
The final spectrum was obtained after baseline calibration
2.4 Adsorption of Analytes The adsorption ability of
CAP and AMX on the surface of e-AgNPs was monitored
using a UV−vis spectroscopy Five hundred microliters of
CAP/AMX solution with different concentrations (100−500
μM) was added to 1 mL of e-AgNP solution and incubated for
30 min at RT before the absorption spectra of the samples
were recorded
2.5 Electrochemical Measurements Electrochemical
measurements in this work were evaluated on the Palmsens 4
electrochemical workstation under ambient condition The
electrochemical measurements to determine the HOMO−
LUMO energy levels of the several analytes, including CAP,
AMX, furazolidone (FZD), 4-nitrophenol (4-NP), and
tricyclazole (TCZ), were set up based on the established
study with a Pt working electrode and an Ag/AgCl reference
electrode.23 The 0.1 M phosphate buffered solution (PBS) served as the electrolyte All electrochemical potentials were referenced to an Fc/Fc+ internal standard Cyclic voltammo-grams of the analytes were performed at a scan rate of 50 mV
s−1in the potential range from−2 to 2 V
3 RESULTS AND DISCUSSION
3.1 SERS Sensing Performance of e-AgNPs to Detect Chloramphenicol and Amoxicillin To evaluate the sensitivity of e-AgNPs in the detection of CAP, 7 samples of CAP in water were prepared at different concentrations, from
10−3to 10−9M The SERS spectra of those samples on the e-AgNP/Al substrate were measured under the same conditions
as described in ref22 On the Al substrate, in the absence of e-AgNPs, the CAP solution (10−3 M) shows no characteristic band in the Raman spectrum In contrast, in the presence of e-AgNPs, that CAP solution exhibits characteristic peaks at 865,
1108, 135, and 1600 cm−1, which are in agreement with those
in the Raman spectrum of the CAP powder (Figure S1) The band at 865 cm−1is assigned to the stretching mode of C−H
on the phenyl group.24−26The band at 1108 cm−1represents the 14C−18O bending vibration and C−H stretching mode.24−26The band at 1350 cm−1is assigned to the bending vibration of−NO2and the stretching vibration of C−H on the phenyl group.24−26 The band at 1600 cm−1 is assigned to 24C−25O and 24C−22N stretching modes.24 , 25
The molec-ular structure of CAP is presented in Figure S2 Other characteristic bands are listed inTable S1.Figure 1a shows the SERS spectra of CAP at different concentrations (10−3−10−9
M) on e-AgNP/Al substrates Obviously, the intensity of the characteristic peaks of CAP decreases with decreasing concentration However, the intensity of those peaks does not exhibit a high linearity against the CAP concentration It is
Figure 1 (a) SERS spectra of CAP (10−3−10 −9 M) on the e-AgNP/Al substrate (b) Plot of log of SERS intensity vs concentrations at 1350 cm−1 (slope 0.354 ± 0.02, intercept 4.54 ± 0.01) (c) Repeatability and (d) reproducibility of the SERS sensor for CAP detection based on the e-AgNP/
Al substrates.
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX C
Trang 4reflected by the relatively low linear regressions when plotting
the logarithmic SERS intensity at 865, 1350, and 1600 cm−1
against the CAP concentration as R2values were estimated to
be 0.93, 0.80, and 0.86, respectively (Figure S3) This poor
linearity may be due to the adsorption of CAP on the e-AgNP
surface It was reported in our previous study that the
adsorption of CAP on the surface of the electrochemically
synthesized gold nanoparticles (e-AuNPs) obeyed the
Freundlich isotherm with two possible adsorbing sites.19 It
could have also occurred in this case, as the variation in
adsorption sites and orientation of the analyte on the e-AgNPs
can be the reason for the low linearity of SERS intensity against
CAP concentration.26The best linearity was observed at the
band of 1350 cm−1with a linear regression of 0.95 (Figure 1b)
Thus, the limit of detection (LOD) and limit of quantification
(LOQ) of the CAP sensor were calculated based on this plot
to be 1.2× 10−9and 1.93× 10−9M, respectively Besides, the
enhancement factor (EF) of the sensor at the band of 1350
cm−1 was estimated to be 4.6 × 104 and 3.3 × 105 when
detecting CAP at concentrations of 10−7 and 10−8 M,
respectively The calculations for LOD, LOQ, and EF are
fully explained in theSupporting Information
In addition to the sensitivity, the performance of the sensing
system was investigated in terms of repeatability and
reproducibility The repeatability experiment was carried out
by recording the SERS signal of CAP (10−6 M) from 5
randomly chosen spots on the substrate (Figure 1c) The
relative standard deviation (RSD) was calculated based on the
SERS intensity at the band 1350 cm−1to be 13.21% Besides,
the reproducibility of the method was evaluated by preparing
five different e-AgNP/Al substrates and employing them for
detection of CAP (Figure 1d) The sensor showed good
reproducibility with an RSD of 8.66% The calculation for RSD
is as explained inSupporting Information
The e-AgNP/Al substrates were subsequently employed for
SERS detection of the other antibiotic, AMX, whose molecular
weight is close to that of CAP, as shown in Table 1
Surprisingly, the SERS spectra of AMX do not show any
characteristic peaks, even at a high concentration of 10−3M
(Figure 2) The experiments were repeated 3 times, obtaining
similar results Therefore, it can be said that AMX was not
capable of being detected using a SERS sensor based on
e-AgNPs under 785 nm laser irradiation This obvious difference
between the two antibiotics urged us to answer the question
why some analytes could be determined using SERS sensing
systems while many others could not It is worth reminding
that all of the SERS experiments described above were carried
out under the same conditions, using the same substrate
fabrication method Thus, the difference should be due to the
distinction in the intrinsic property of the analytes
In a previous study, we reported that analyte molecular
structure had great effects on its adsorption on the NP surface
and the nanogaps between them, which was known as one of
the decisive factors for SERS detection.19 As a result, analyte molecular structure could influence SERS performance An analyte may be detected effectively using an SERS sensing system when it is a small molecule possessing at least one functional group that can form a specific or strong interaction with the noble metallic surface and has low steric hindrance Regarding the two antibiotics discussed in this study, AMX has
an advantage over CAP because of its functional groups With the presence of the carboxyl (−COOH) and primer amine (−NH2) groups, it is more convenient for AMX to attach on the surface of e-AgNPs However, it exhibits the disadvantages
of molecular weight and steric hindrance AMX is 13% larger than CAP in molecular weight Moreover, because of more functional groups (Figure 3a), AMX also shows larger steric hindrance These structural features could reduce the adsorption ability of AMX onto the e-AgNP surface, in comparison to that of CAP The adsorption ability of organic compounds on the surface of plasmonic nanoparticles could be investigated using the absorption spectra of nanoparticle solutions in the presence of analytes, as described in our previous study It was reported that the level of adsorption of analytes on the nanoparticle surface was responsible for the level of decrease in the absorption intensity of nanoparticle solutions.19 Figure 3b,c shows the absorption spectra of e-AgNPs in the presence of AMX and CAP at varied concentrations It is undeniable that in both cases, the intensity
of the plasmonic band decreases with the increase of antibiotic concentration This may be explained by the aggregation of NPs in the presence of these residues Furthermore, red shifts
of 6 nm and 4 nm can be observed on the addition of CAP and AMX, respectively, indicating the replacement of citrate by the antibiotic residues Therefore, both CAP and AMX absorbed onto the surface of e-AgNPs Definitely, the adsorption abilities
of CAP and AMX were not completely the same, as the addition of CAP led to a more dramatic decrease in the absorption intensity of the e-AgNPs compared to that of AMX Hence, the large steric hindrance of AMX might have lowered its level of adsorption onto the e-AgNP surface However, the
difference in adsorption ability cannot fully explain why AMX could not be detected by an SERS sensing system, especially when it was shown in our previous study that the lower adsorption ability of an analyte only resulted in a poorer SERS performance of the sensing system.19 Therefore, it should be another reason for this phenomenon other than the adsorption
of the analyte
The intensity of SERS is a function of the enormous enhancement of Raman intensity due to the joint contribution
of CM and EM, as described by the formula4
PSERS= PRaman×GSERS=PRaman×GSERSCM×GSERSEM
(1)
where PSERS is the enhanced Raman intensity (i.e., SERS intensity), PRaman is the Raman signal, and GSERS is the total enhancement factor, consisting of GSERSCMand GSERSEMas the enhancements caused by CM and EM, respectively
CM is a short-range effect, which requires a tight contact (or chemical bond) between the metal surface and the analytes It
is associated with the molecular orbital interaction level of the analytes and metal nanomaterials, which has been divided into three contributions: molecular resonance Raman scattering (RRS) mechanism, nonresonant chemical (CHEM) mecha-nism, and CT mechanism (Figure 4).20The RRS mechanism occurs when the incident laser is resonant with an
Table 1 Comparisons of CAP and AMX
CAP (C11H12Cl2N5O)
AMX (C16H19N3O5S)
functional groups to absorb on
adsorption (via UV −vis results) yes yes
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX D
Trang 5intermolecular excitation of analytes The resonance state is
achieved when the energy of the incident laser is equal to the
energy of the HOMO−LUMO transition of the analyte RRS
was reported to be enhanced on a metal surface.20The CHEM
mechanism involves the relaxation of the electronic structure of
the molecule, arising due to a ground-state interaction between
this molecule and a metal surface.20 The CT mechanism is
related to plasmon-induced hot-electron transfer from the
metal to the molecule via surface plasmon resonance
(SPR).20,27 This process has been reported to be dominated
by the energy-level alignment at the metal/molecule
inter-face.21The overall enhancement of CM has been reported in
many studies, and it can be as high as 102−104.20,27
EM, in contrast, is a long-range effect, which is a result of the interaction between the optically excited collective electron oscillation and the analytes This mechanism can be briefly explained by the local electromagneticfield distribution in the proximity of the metallic surface, which leads to increasing Raman cross section, hence improving the Raman output signal.20The EM enhancement can be explained via a model including a three-step processstep 1: local field enhance-ment, step 2: radiation enhanceenhance-ment, and step 3: the resulting mutual excitation induces a direct charge exchange between the metal nanomaterials and the HOMO−LUMO state of the analytes (Figure 4).20,27,28 Step 3 of the EM enhancement process denotes a direct CT between the Fermi level of the
Figure 2 SERS spectra of AMX (10−3−10 −5 M) obtained from three different e-AgNP/Al substrates prepared with the same protocol.
Figure 3 (a) Molecular structures and functional group of CAP&AMX UV−vis spectra of e-AgNPs in the presence of AMX (b) and CAP (c) at varied concentrations.
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX E
Trang 6metal nanostructure and the LUMO of the adsorbed
molecule.20
It is noticeable that all of the three contributions of the CM
and EM concern the HOMO−LUMO levels of the analyte
Therefore, understanding the HOMO−LUMO levels of CAP
and AMX might provide the answer for the question of the
ability to be detected by the SERS sensors of these analytes
3.2 Determination of HOMO−LUMO Levels of CAP &
AMOX Using Cyclic Voltammetry The HOMO and
LUMO energy levels (EHOMO and ELUMO) of CAP and AMX
were calculated based on onset oxidation and reduction
potentials (ϕox and ϕred), respectively, in cyclic
voltammo-grams, using the equations23,29,30
EHOMO= −e(ϕox +4.8 −ϕFc/Fc+) (2)
ELUMO= −e(ϕred+4.8−ϕFc/Fc+) (3)
where ϕFc/Fc+ is the redox potential of the ferrocene/ ferrocenium (Fc/Fc+) couple in the electrochemical measure-ment system, assuming the energy level of Fc/Fc+to be −4.8
eV below the vacuum level In this study, we set up an electrochemical measurement system similar to that described
in the study of Bin et al.23with a Pt working electrode and Ag/ AgCl as the reference electrode Thus,ϕFc/Fc+was assumed to
be 0.44 V versus Ag/AgCl
The CV measurements of CAP and AMX were carried out
in 0.1 M PBS (Figure 5) In the absence of analyte, no redox peak was observed In the presence of CAP, an anodic signal
Figure 4 Mechanisms of chemical enhancement (CM) and electromagnetic enhancement (EM) in SERS Therein, CM enhancement includes three contributions: RRS mechanism, CHEM mechanism, and CT mechanism; EM enhancement is described through a three-step process, where
EFis the Fermi level of the metal, ω is the excitation frequency, and HOMO and LUMO are the highest occupied molecular orbital and lowest unoccupied molecular orbital, respectively.
Figure 5 Cyclic voltammograms of the PBS solution, CAP dissolved in PBS, and AMX dissolved in PBS.
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX F
Trang 7can be detected in an irreversible way at 1.47 V It is the onset
oxidation peak of CAP Besides, two irreversible cathodic peaks
appear at−0.65 and −0.77 V Based on the scanning direction,
the peak at −0.77 V is the onset reduction peak of CAP
Obviously, this cyclic voltammogram also shows a couple of
well-defined reversible peaks, which are marked as A1 and A2
However, measuring the HOMO−LUMO levels of an organic
molecule only concerns onset redox peaks; therefore, these
peaks will not be further discussed in this study Similarly, we
determined the onset oxidation and reduction of AMX to be
1.33 and −1.07 V, respectively Based on these onset redox
peaks, we measured the onset redox potential values of CAP
and AMX as shown in Figure 6a,b, respectively The onset
oxidation potential (ϕox) values for CAP and AMX were
measured to be 1.25 and 1.17 V, respectively Besides, the
onset reduction potential (ϕred) values were estimated to be
−0.52 and −0.99 V, respectively The HOMO energy level
(EHOMO) and LUMO (ELUMO) of CAP were calculated to be
−5.61 and −3.84 eV, respectively, while the EHOMOand ELUMO
of AMX were−5.53 and −3.37 eV respectively The values are
averaged over 10 cycles of CV scans (Figure S4)
3.3 Effects of the HOMO−LUMO Energy Levels of
CAP & AMOX on their SERS Signals Our SERS
experiments were performed under the excitation of a 785
nm laser source; thus, the laser energy (E) was calculated to be
1.58 eV, using the well-known formula
where λ is the wavelength, c is the speed of light in vacuum,
and h is the Planck constant Our calculations from the
electrochemical measurements showed the HOMO−LUMO
gaps of CAP and AMX to be 1.77 and 2.16 eV, respectively Obviously, the incident laser energy was not large enough to achieve the resonance state for the RRS mechanism The CHEM mechanism refers to the relaxation of some electrons from the LUMO level to the HOMO level of the analyte adsorbed on a metal surface, leading to Raman signal emission
It was reported to be the weakest among the mechanisms, which contributes∼10−100 times of enhancement, being only
a small part in the gigantic enhancement of the Raman intensity in the SERS phenomenon.20Thus, this contribution
is not large enough to cause such huge difference in the ability
to be detected of CAP and AMX Instead, involving both CM and EM, the SPR-induced hot-electron transfer process, also known as CT, is the dominant mechanism Figure 7 depicts three possible cases for CT between the metal and the
Figure 6 Cyclic voltammograms of CAP (a), AMX (b), and their magnified images to determine the onset oxidation potential (ϕ ox ) and onset reduction potential (ϕ red ) values.
Figure 7 Schematic illustration of the Fermi energy-level position of the metal (left) relative to the HOMO−LUMO energy level of the analyte (right) in three cases: Φ m > EHOMO(a), EHOMO> Φ m > ELUMO (b), Φ m < E LUMO (c), where Φ m is the work function of the metal surface.
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX G
Trang 8absorbed molecule: (a) the work function of the metal surface
(Φm) is larger than the HOMO energy level of the analyte
(Figure 7a); (b)Φmis less than the HOMO energy level of the
analyte and larger than the LUMO energy level of the analyte
(Figure 7b); and (c)Φmis less than the LUMO energy level of
the analyte (Figure 7c) Otero et al proposed that the electron
transfer from the metal to analyte or the analyte to metal
would occur even without any laser excitation in two cases:Φm
< ELUMOandΦm> EHOMO.21The Fermi energy level of Ag was
reported to be approximately −4.26 eV.31
Thus, the Fermi energy level of e-AgNPs and HOMO−LUMO levels of CAP
and AMX is in the case (b) Consequently, to achieve SERS
effect, e-AgNPs have to be excited by laser irradiation, leading
to a transition between the Fermi level of the metal and the LUMO level of the analyte The difference between the Fermi energy level of e-AgNPs and the LUMO energy level of CAP was calculated to be 0.42 eV, which was obviously lower than that of AMX (i.e., 0,89 eV) This significant distinction may be the explanation for the difference in the SERS results of the two analytes described in Section 3.1 The gap of 0.89 eV might have been so large that it was impossible for a CT transition from the Fermi level of the e-AgNPs to the LUMO level of AMX On the other hand, a smaller gap of 0.42 eV allowed the occurrence of this transition in the case of CAP
We propose hot-electron transfer processes from the e-AgNPs
to CAP and AMX as shown in Figure 8, exhibiting the
Figure 8 Proposed hot-electron transfer process from the e-AgNPs to the analyte in two cases: hot electrons from the surface of the e-AgNPs can transfer to CAP, with a low energy di fference between E F and ELUMO(ELUMO−CAP− E F = 0.42 eV) (panel 1), and hot electrons from the surface of e-AgNPs cannot transfer to the AMOX, with a large energy di fference between E F and ELUMO(ELUMO−AMOX− E F = 0.89 eV) (panel 2).
Figure 9 Determination of the HOMO−LUMO energy levels of FZD, 4-NP, and TCZ analytes via cyclic voltammetry measurements.
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX H
Trang 9importance of the gap between the LUMO level of the analyte
and the Fermi energy level of the substrate to the SERS signal
However, our experimental data did not show any significant
evidence for the effects of HOMO level of the analyte on its
SERS signal
3.4 Confirmation for the Effects of HOMO−LUMO
Energy Levels of the Analyte on the SERS Signal:
Examples of Several Other Analytes The experimental
model of CAP and AMX has shown the importance of the
LUMO energy level of the analyte on SERS signal However,
HOMO level did not show any significant effects To better
confirm this hypothesis, we carried out some “quick tests” on
several organic compounds, in which their HOMO−LUMO
levels were, at first, estimated via CV measurements From
these energy levels, several predictions of their SERS signals
could be made, and comparison between the predicted and actual signals might be a confirmation for the correlation between the HOMO−LUMO energy levels of the analytes and their SERS signals To be convenient for SERS detection, the three selected compounds, including furazolidone (FZD), 4-nitrophenol (4-NP), and tricyclazole (TCZ), were all reported
to be able to adsorb onto the e-AgNP surface.32−34
CV measurements of those organic compounds were carried out in 0.1 M PBS (Figure 9) In the absence of analyte, no redox peak was detected In the presence of FZD, 4-NP, and TCZ, irreversible anodic peaks appear at 1.28, 0.88, and 1.76
V, respectively They are the onset oxidation peaks of FZD,
4-NP, and TCZ, respectively The onset oxidation potential (ϕox) values of FZD, 4-NP, and TCZ were measured to be 1.1, 0.7, and 1.14 V, respectively Usingeq (2), the HOMO levels
Figure 10 SERS spectra of FZD (10−3−10 −5 M) obtained by preparing three di fferent e-AgNP/Al substrates.
Figure 11 (a) SERS spectra of TCZ (10−3−10 −10 M) obtained on e-AgNP/Al substrates (b) Plot of log of SERS intensity −concentrations at 1360
cm−1(slope 0.337 ± 0.02, intercept 4.53 ± 0.01) (c) SERS spectra of TCZ (10 −6 −10 −10 M).
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX I
Trang 10of FZD, 4-NP and TCZ were calculated to be−5.46, −5.06,
and −5.75 eV, respectively On the other hand, the onset
reduction peaks of FZD, 4-NP, and TCZ were determined
thanks to the appearance of irreversible cathodic peaks at
−1.53, −0.94, and −0.65 V, respectively The onset reduction
potential (ϕred) values of FZD, 4-NP, and TCZ were estimated
to be−1.1, −0.85, and −0.4 V, respectively The LUMO levels
of FZD, 4-NP, and TCZ were calculated to be −3.26, −3.55,
and−3.98 eV, respectively The values were averaged over 10
cycles of CV scans (Figure S5)
Obviously, the LUMO level of FZD is even higher than that
of AMX; therefore, it was expected that FZD would not be
detected using a SERS sensor based on e-AgNPs under 785 nm
laser irradiation This prediction was then confirmed by the
SERS measurements of FZD on e-AgNPs (Figure 10) Similar
to the results obtained in the presence of AMX, no
characteristic peak was observed in the SERS spectra of
FZD, even at a concentration as high as 10−3 M The
experiments were repeated 3 times, obtaining similar results It
is interesting that the SERS signal of FZD is similar to that of
AMX, while FZD is 30% smaller than AMX in molecular
weight, and possesses lower steric hindrance and a different
functional group (−NO2) (Figure 13) This result stresses the
effect of the LUMO level of analytes on their SERS signals
In contrast, the LUMO level of TCZ is lower than that of
CAP The smaller gap between the LUMO level of TCZ and
the Fermi level of Ag was expected to be convenient for CT,
and consequently, TCZ was predicted to be detectable using
the e-AgNP-based SERS sensing system The actual result was
in agreement with the prediction.Figure 11a,c shows the SERS
spectra of TCZ at different concentrations (10−10−10−3M),
with characteristic peaks at 430, 566, 592, 890, 1283, and 1360
cm−1 The band at 430 cm−1is assigned to the deformation of
the C−N−C vibration.35 − 37
The band at 566 cm−1represents the deformation of C−C bending.35 − 37
The band at 592 cm−1
is associated with the C−S−C vibration mode.35 − 37
The band
at 890 cm−1represents the symmetric stretching mode of C
C vibration.35−37 The bands at 1283 and 1360 cm−1 are
assigned to the C−N stretching vibration.35 − 37
At the same concentration, the SERS spectra of TCZ exhibit a higher
intensity in comparison to that of CAP.Figure 11b shows the
plot of the logarithmic SERS intensity at 1360 cm−1 against TCZ concentration, based on which LOD was calculated to be 5.86 × 10−10 M, which is significantly lower than that when detecting CAP Therefore, the smaller gap between the LUMO level of the analyte and the Fermi level of the substrate resulted
in a higher intensity of the SERS signal, leading to better sensitivity of the SERS sensor
As a more complicated case, the LUMO level of 4-NP is higher than that of CAP but lower than that of AMX Thus, there could be two possibilities for the SERS signal of 4-NP (1) In case the gap between the LUMO level of the analyte and the Fermi level of Ag was still too large for the CT process
to occur, no SERS signal would be obtained (2) In case the gap was small enough for the CT process, the SERS signal of
4-NP on e-Ag4-NPs could be recorded, but with lower intensity compared to that of CAP The SERS results were in agreement with the second possibility, as at high concentrations of 4-NP (10−3and 10−4M), SERS spectra show characteristic bands at
858, 1270, 1330, 1486, and 1580 cm−1(Figure 12) The band
at 858 cm−1 represents the bending mode of the nitro group.38,39 The band at 1330 cm−1 is assigned to the symmetric stretching of the nitro group, while the band at
1270 cm−1is attributable to a ring deformation mixed with the stretching mode of the nitro group.38,39 The band at 1486
cm−1 represents the bending mode of C−H.39
The band at
1580 cm−1is assigned to the stretching mode of the ring.38,39 However, the SERS intensity is much lower than that of CAP
at the same concentration These peaks disappear at the concentration of 10−5M Similar to the experiments on FZD, the SERS measurements of 4-NP were also repeated 3 times, revealing the same results (Figure 12) Furthermore, it is worth mentioning that CAP is a nitrophenyl-substituted molecule The high intensity of the band at 1350 cm−1 in the SERS spectra of CAP and the bands at 1330 and 1270 cm−1in the SERS spectra of 4-NP suggests that the adsorption of both CAP and 4-NP occurs via their nitro groups.19,39 Moreover, the smaller molecular weight and steric hindrance should have been the advantages of 4-NP to adsorb onto the e-AgNP surface to experience a better SERS effect However, the actual SERS result is the opposite It should be explained by the
difference in LUMO levels of these two compounds
Figure 12 SERS spectra of 4-NP (10−3−10 −5 M) were obtained by preparing three different e-AgNP/Al substrates.
https://doi.org/10.1021/acs.jpcc.2c01818
J Phys Chem C XXXX, XXX, XXX−XXX J