3.1. X-ray Diffraction pattern
The crystal plane of CTAB assisted-BiOBr and theCTAB-AgBr-Bi3O4Br/BiOBr were characterised by X-ray diffraction (XRD). Figure1(a) shows the XRD pattern of CTAB assisted-BiOBr, which exhibits the diffraction peaks of BiOBr at the following 2θ values: 21.88, 25.22, 31.76, 32.28, 33.20, 34.06, 39.34, 44.36, 46.28, 50.7, 57.22, 76.86 which correspond to the indices (002), (011), (012), (110), (003), (111), (112), (004), (020), (014), (212) planes of tetragonal BiOBr (JCPDS no. 73-2061). Further, addition of silver source influences the crystal structure of BiOBr. As silver source increases from Ag0.09BOBto Ag0.15BOB, the diffraction peaks of AgBr and BiOBr increases, while the peaks of Bi3O4Br decreases. In situ growth of AgBr, the characteristic diffraction peaks appeared at 26.76, 30.96, 44.36, 55.04, 64.48, 73.26, which assigned to the planes (111), (200), (220), (222), (400), (420) (JCPDS no. 06-0438) and new diffraction peaks appeared at (112), (114) which can be attributed to Bi3O4Br (JCPDS no.84-0793)as shown in figure 1(b-c). The average crystalline size of BOB, Ag0.09BOB andAg0.15BOB were calculated to be 10.1, 40.8 and 30.4 nm using the Scherrer equation,
D = 0.9λ/(ꞵCos(θ)) (1)
Where, D, λ, β and θ are crystallite size, wavelength of X- rays, full width at half maxima (FWHM) and Bragg’s diffraction angle respectively (27).
3.2. UV-visible spectra
Optical property of CTAB-AgBr-Bi3O4Br/BiOBr with different composition of silver as Ag0.09BOB and Ag0.15 BOB were investigated using UV-visible spectroscopy. Figure 1(b) confirm the existence of silver as shown in the UV- visible absorbance spectra of substrate where, two distinct absorption bands were observed at 260 and 355 nm. The absorption band at 260 nm is attributed to BiOBr and the band at 355 nm belongs to the dipole resonance of silver (28).
3.3. SEM
Figure S1 shows the surface morphology of the prepared samples characterized by scanning electron microscope. Figure S1 (a) shows the flake-like structure of CTAB assisted-BiOBr with the average size of 0.6-0.8 µm. Figure S1 (b) shows the flake and rod-like structure of Ag0.09BOB (CTAB-AgBr-Bi3O4Br/BiOBr) microparticles with average length of 0.55-0.65 µm and 3.2-3.4 µm, respectively. Figure S1(c) shows the aggregated flake and rod-like structure of Ag0.15BOB (CTAB-AgBr-Bi3O4Br/BiOBr) microparticles with length of 0.50-0.55 µm and 2-3.3µm.
3.4. XPS
The surface composition and chemical state of Ag0.09BOB composite was analysed by XPS (figure S2). The survey spectrum indicates the presence of Ag, Bi, O and Br in the microparticles (figure S2 (A)). The peak at binding energy of 285.08 eV corresponds to the appearance of adventitious Carbon species (29) (figure S2 (A)). Figure S2 (B) shows the high-resolution spectrum of Bi 4f consist of two peaks at 158.6 and 163.9 eV corresponds to Bi 4f7/2 and Bi 4f5/2 respectively, which confirmed the presence of Bi3+ state in BiOBr (30). Figure S2(C) shows the XPS spectrum of O1s, the dominant peaks at 529.5 eV and 531.7 eV can be attributed to lattice oxygen in Bi-O-Br and surface adsorbed components of hydroxyl group respectively (29). The Br 3d spectrum in figure S2(D) shows peaks at 67.7 and 68.7 eV corresponding to Br 3d5/2 and Br 3d3/2, which showed the presence of Br¯ in AgBr and BiOBr (24). Figure S2(E) shows the XPS spectrum of Ag 3d at binding energies 367.2 and 373.1 eV are assigned to Ag 3d3/2 and Ag 3d5/2confirm the presence of Ag+ ion in Ag0.15BOB composite (25, 31).
3.5. FTIR Analysis
FTIR analysis was carried out to identify the different functional groups present in the prepared composites and all samples show the almost same appearance of the peaks. The presence of CTAB on the surface of BiOBr was confirmed by FTIR analysis as shown in figure 2 (a).
The broad band at 3429 and 1629 cm-1 can be attributed to the stretching and vibration of –OH group and bending H–O–H respectively, the increased broadness in Ag0.09BOB and Ag0.15BOB arises due to adsorbed H2O on the surface of AgBr (32, 33). As shown in figure of Ag0.15 BOB, the intensity of peaks at 2849, 2917 cm-1are weaker than those of CTAB assisted-BiOBr (BOB) are caused by the symmetric and asymmetric stretching vibration mode of methyl and methylene groups of CTAB, respectively (32).The most intense peak at 1384, 1467, and cm-1 arises due toasymmetric and symmetric C-H scissoring vibration of N+-CH3[33-34]. The observed peak at 720 cm-1can be assigned to rocking mode of methylene chain and the peaks at 814, 909, 962, 1034 cm-1 can be attributed to C-N+ stretching bands (33). The peak at 524 cm-1can be assigned to the Bi-O stretching vibration of BiOBr (34). The new peak at 825 cm-1 in Ag0.09BOB and Ag0.15BOB may be assigned due to the C-Br stretching bond (as shown in inset figure 2 (a)) (30), which did not appear in the spectrum of BOB. These results suggest that CTAB exist at the surface of silver containing bismuth composites.
3.3 Raman study of CTAB-AgBr-Bi3O4Br/BiOBr
Figure 2(b) and 2(c) shows the Raman spectra of pure solid CTAB and CTAB adsorbed on bismuth-based composites, respectively. As shown in figure 2(c), CTAB assisted-BiOBr (BOB) composite showed Raman signal at 957, 1042, 1130, 1297, and 1449 cm-1. The shoulder peaks at 957 cm-1 were attributed to CH3 rocking and C-N+ stretch. Blue shift at 1042 by 11 cm-1and 1449 cm-1 by 18 cm-1 in Raman spectrum of CTAB adsorbed on bismuth composites as compared to pure solid Raman spectrum of CTAB, was due to C-C stretching and due to the combination of CH2 scissors, and CH3 deformation, which demonstrates the interaction of CTAB with bismuth composites. After the introduction of AgBr the vibrational signals of CTAB significantly decreased. This result suggests that CTAB dissociated from metal surface in the presence of AgBr and revealed the adsorption of CTAB on the surface of silver containing bismuth composites, which corroborated well with the FTIR analysis. Ag0.09BOB or Ag0.15BOB demonstrated the characteristic CTAB Raman modes at 1042, 1130, 1297, 1449 cm-1. The most prominent Raman band at 1042 cm-1 corresponds to C-C stretch. The Raman peaks at 1130, 1297 and 1449 cm-1 were due to C-C stretching, CH2 twist, and due to the combination of CH2 scissors, and CH3 deformations, respectively (35).
3.4 SERS Measurement
Ethanolic solution of R6G was used as a standard probe molecule to preliminarily investigate the SERS activity of the substrate. The SERS activity was performed by mixing 100 μL of R6G solution with substrate BOB (CTAB assisted- BiOBr), Ag0.09BOB or Ag0.15 BOB (CTAB- AgBr- Bi3O4Br/BiOBr) via the comparison of solid and aqueous Raman spectrum.
The SERS activities of BOB, Ag0.09BOB or Ag0.15 BOB substrates were investigated by using R6G dye with the laser of 785 nm Wavelength. The characteristic Raman peaks of R6G in solid Raman spectra, shown in figure 3(a) were observed at 607, 768, 1117, 1179, 1302, 1360, 1504, 1559, 1593 and 1664 cm−1. The tentative assignment of these vibrational modes of R6G are summarized in Table ST1 (supporting information). The SERS spectrum of substrate with varying concentrations of R6G is shown in figure 3(b). In the SERS spectrum, intense peaks at 1507, and 1364 cm−1 correspond to the C-C stretching mode, while the peak at 1305 cm−1 was assigned to the N-H in-plane bending mode. The peak at 1194 cm−1, the in plane bending mode of C-H was blue shifted by 15 cm−1 (36-37). The detection limit of R6G on bismuth substrate was 10 µM. Similarly, the detection limit was determined as 100 µM for BOB and Ag0.09BOB substrate as shown in figure S3 (supporting information), which indicate the enhanced activity of Ag0.09BOB than that of BOB. However, BOB itself has intense Raman signal (at 1042 cm-1), which can interfere with the analyte's lower concentration (below 100 µM). On the other hand, after the addition of silver, the Raman signal of substrate is suppressed as shown in figure 2(c). No interference is caused by the Ag0.15 BOB substrate to the dye, and there was sufficient SERS signal to observe the dye.
Further, the most prominent peak at 1507 cm–1 of R6G was used in the calibration plot as shown in figure 3 (c). A good linear response was obtained between the SERS intensity and the logarithmic of R6G concentration in the range from 100 µM to 10 µM with the correlation coefficient of R2 = 0.980. The linear equation can expressed as:
Raman Intensity (I SERS) = 2470 Log (R6G) + 13139 (2)
Four different food colours AR, AM, TR, and SY dyes were analysed using BOB, Ag0.09BOB or Ag0.15BOB substrates. Further, normal Raman spectra of aqueous solution of colorants at 1mM concentration were recorded to compare the peak position in the SERS spectrum. Figure (4-5) shows the normal Raman and SERS spectrum of food colours with different concentrations from 1mM to 40 µM on Ag0.15BOB substrate.
AR is a strongly water soluble azo dye used for coloring food products. Figure 4(a) shows the solid Raman spectrum of AR at 751, 974, 1129, 1188, 1223, 1266, 1326, 1383, 1407, 1496, 1578 and 1604 cm-1 and their assignment are listed in Table ST1. Figure 4(b) shows the SERS spectrum of AR dye on Ag0.15BOB substrate, revealing the observed characteristic peaks at 971, 1121, 1190, 1227, 1271, 1316, 1383, 1411, 1491, and1577 cm-1 which is consistent with the solid Raman spectrum. The most intense peak appeared at 1491 cm-1corresponds to C=C stretching in the benzene ringand the sharp peak at 1316 cm-1was red shifted by 10 cm-1with respect to the solid Raman spectradue to the interaction of dye with substrate. The other characteristic peaks at 1190, 1227, 1271 and 1411 cm–1 were attributed to the SO2 asymmetrical stretching vibration, rocking (H-C-H) the C–O plane stretching vibration on the benzene ring, and benzene ring and the rocking vibration of H-C-H respectively (15). The lowest detection limit for AR was 50 μM. The calibration plot of AR shows (figure 4(c)) that as concentration decreases the intensity of Raman peak at 1316 cm–1decreases. The correlation coefficient R2 = 0.983 with the linear equation as follows:
Raman Intensity (I SERS) = 1192 Log (AR) + 5534 (3)
Figure 4(d) shows the Raman spectrum of AM food color. The characteristic peaks of AM appeared at 738, 1052, 1135, 1172, 1231, 1361, 1401, 1434, 1477, 1511 and 1571 cm–1 and their tentative assignments are shown in Table ST1. Figure 4(e) shows the SERS spectrum of AM dye acquired on Ag0.15 substrates, which reveal that the intensity of characteristic Raman peaks decreased with decreasing concentration. The sharp band observed at 1363 cm–1 was due to C-C stretching vibration. Another intense Raman bands at 1231, and 1572 cm–1 correspond to wagging of O-S, and asymmetric stretching of ring, respectively, which are in good agreement with the previous report (38-39). The lowest detection limit of AM was found to be 50 μM. The most intense peak at 1363 cm–1 was used in the calibration plot (figure 4 (f)) in the range of 1mM to 50 μM, which demonstrates the increase in intensity with increased concentration. The calibration equation with correlation coefficient R2 = 0.972 shown below:
Raman Intensity (I SERS) = 5000 Log (AM) + 22482 (4)
Figure 5(a) shows the solid Raman and figure 5(b) depicts the SERS spectrum of TR dye on Ag0.15BOB substrate. The characteristic Raman peaks of TR were observed at 623, 802, 1044, 1132, 1177, 1217, 1361, 1414, 1499, and 1597 cm-1, their assignments are listed in Table ST1 are match well with the previous report (40). The most intense peak located at 1596 cm-1in the SERS spectrum corresponds to C-C stretching and C-H wagging in benzene ring and other sharp peaks at 1129, and 1356 cm-1 are attributed to N=N stretching, C-H wagging and C-C stretching in naphthalene ring (40). The lowest detection limit of TR was found to be 40 µM. The peak at 1596 cm-1 was selected for quantitative analysis of TR. The calibration plot (figure 5 (c)) demonstrates an excellent linear relation in the range of 1mM to 40 μM with the correlation coefficient R2 = 0.989 and the linear equation is shown below:
Raman Intensity (I SERS) = 14515 Log (TR) + 67145 (5)
SY is a synthetic azo dye, which is frequently used in foods due to its low cost was purchased from local market. Figure 5(d) shows the characteristic Raman modes of solid SY dye at 986, 1123, 1178, 1231, 1336, 1384, 1496, 1545 and 1594 cm–1 and their assignments are given in Table ST1. In the SERS spectrum (figure 5(e)) of SY, the peaks at 1160, 1212, and 1302 cm–1 exhibits red shift by 18 cm–1, 19 cm–1and 34 cm–1 as compared to solid state Raman spectrum due to N=N stretching, C-C bending and C-N stretching, respectively, which are in good agreement with the literature (16). These Raman shift demonstrate that Ag0.15BOB substrate highly influenced on the SERS spectrum of SY. The detection limit was determined to be 40 μM. The quantitative analysis of SY was analyzed via the selected most intense peak 1593 cm–1 with an excellent linear response as shown in figure 5(f) in concentration range of 1mM to 40 μM. The linear equation with the correlation coefficient R2 = 0.988 is given below:
Raman Intensity (I SERS) = 2024 Log (SY) + 9221 (6)
The comparison of the detection limits for considered food colorants with the previous report on different SERS substrates are listed in Table ST2 (supporting information).
Moreover, to validate the stability of the synthesized substrate, 10−4 M ethanolic solution of R6G were used as probe molecule, and the SERS spectrum was recorded from eight randomly selected different spots on Ag0.15BOB substrate under identical experimental condition. Figure 6 (a) demonstrates the practically identical SERS spectra with reasonable relative standard deviation (RSD) 9.75% corresponding to the peak at 1507 cm-1 as displayed in Figure 6 (b). The observed result shows the efficient stability of Ag0.15BOB substrate, and makes reliable SERS substrate.
To explain the SERS mechanism, we chosen R6G as a probe molecule on the considered substrate. It is generally expected that there are two mechanisms such as electromagnetic and charge transfer (CT) mechanism. From the UV-visible spectra (figure 1(d)), the observed peaks at 260 and 355 nm indicates the plasmonic resonance in the near to far ultraviolet region which can exhibits substantial electric field enhancement in the UV region (41). Thus in the enhancement mechanism of Ag0.15BOB substrate, the can contribute by the combination of EM effect (electromagnetic effect) of Ag particles and the CT effect of Ag0.15BOB substrate and the probe molecules.
Nonetheless, earlier study indicates that chemical enhancement is the primary means of improving bismuth oxybromide materials (29). As shown in figure 6(c), the charge transfer mechanism between Ag0.15BOB substrate composite and R6G molecule is mainly determined by the photo-induced hot electron transfer and photo-induced charge transfer, which enhanced the polarizability of molecules by causing a greater increase in the Raman scattering cross section when there is more internal charge separation. An excitation source with the wavelength 785 nm can excite the silver particles to generate hot electron transition to the LUMO (lowest unoccupied molecular orbital) of R6G. Besides, electrons from Ag can also be transferred to CB (conduction band) of bismuth oxybromide are driven to the LUMO of R6G by excitation laser, thereby generating the charge transfer efficiency.
In addition, to calculate the enhancement factor of the substrate Ag0.15BOB, the ratio between SERS and normal Raman of R6G, AR, AM, TR, and SY were estimated using following calculation formula:
![](data:image/png;base64,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)
Where, ISERS and IRaman are the peak intensities of the SERS of analyte adsorbed on susbstrate, and normanl Raman signal, respectively. NSERS and NRaman are the number of analyte molecules in the SERS substrate and normal Raman, respectively. In the SERS measurement, 100 μL of analytes solution of 10 μM R6G, 50 μM AR or AM, and 40 μM TR or SY solutions were deposited onto the area of 16 mm2 on the surface of Ag0.15BOB substrate. While, in case of the normal Raman measurement 100 μL of 1mM considered analytes were deposited over the area of 16 mm2 onto the glass slide. Intensity ratio ISERS/ IRaman was calculated corresponding to the peaks such as 1507, 1316, 1363, 1596, and 1593 for R6G, AR, AM, TR, and SY, respectively. The calculated enhancement factor were for R6G, for AR, for AM, for AR, for TR,for SY.
3.5 Analysis of binary mixture of food colors
The SERS analysis of binary mixture with Sunset yellow/ Tartrazine (SY/TR), Sunset yellow SY/AM, Sunset yellow/ Allura red (SY/AR) and Amaranth/Allura red (AM/AR) on Ag0.15 BOB offer an examples for the condition that may occur where two food colorants may be present simultaneously.
Figure 7 (a-d) demonstrates the SERS spectra using Ag0.15 BOB substrate for the detection of binary mixture of food colors, which was prepared by relative ratio of 10-3 M concentration of 1:3, 3:1, 1:1, 1:2, 2:1 ratio. The lowest relative ratio at which at least two distinguishable peaks were observed as highlighted in figure 7 for each colorant. At higher volume ratio with respect to each colors in binary mixture with Ag0.15 BOB substrate, shows the prominent Raman signature of dye, whereas with low relative volume ratio was not detected. Further, we observed that SERS spectra of binary solution of TR with AM and AR were not significantly distinguishable with the used substrate. The SERS spectrum of SY/TR display the characteristic feature of SY at 1302 and 1392 cm–1 and TR at 1358, 1214, can be well differentiate these dyes in binary solution. In addition, some of the spectral bands observed, which are closely related to the Raman bands of dyes SY and TR by slight difference, located in the mixture of both colorants as shown in figure 7(a). Figure 7(b) demonstrates the SERS spectra of binary mixture of SY/AM on Ag0.15 BOB substrate, the main difference is noticed with the typical 1300, 1590 cm–1signal of SY and 1360, 1474, 1572 cm–1of AM, which are evident in dye mixtures. Similarly, in figure 7 (c) SY/AR and in figure 7 (d) AM/AR, the predominant spectral features were observed at 1300, 1456 cm–1of SY and one peak of AR at 1263 cm–1 in SY/AR mixture and the peak at 1363, 1463, 1507 cm–1 of AM and the fingerprints at 1340, 1492 cm–1 of AR in the mixture of AM/AR can distinguishedthe food colors separately in the binary solution. In addition, some of the spectral bands observed, which appear to be a combination of two bands of SY and TR, SY and AM, SY and AR, AM and AR, located in the mixture of binary colorant solutions and their intensity were also increases as compared to the SERS spectra of individual dye on Ag0.15 BOB substrate as shown in figure 7 (a-d). Overall, it was observed that the SERS spectra of dyes show different characteristic bands with different combination of dyes, which may be used as a reference spectral feature for the identification of illegal addition of colours in food stuffs.