Introduction

Decay of surface plasmon to hot carriers has found wide applications in energy conversion, photocatalysis and photodetection1,2. Among them, surface plasmonic resonance (SPR)-catalysed reactions have attracted increasing attention due to their green and convenient process3,4. Moreover, compared with the thermocatalysis, the SPR-catalysis often results in a unique selectivity5,6,7. However, due to the rapid relaxation process of hot carriers (hole and electron) on a short timescale ranging from femtosecond to picosecond, it is still hard to flexibly control the path/degree of SPR-catalysed reactions8,9,10.

Deep and accurate understanding about the mechanism of SPR-catalysed reaction is highly essential for designing an efficient reaction system, which requires a sensitive and in-situ analysis strategy. Thanks for the same origin of SPR-catalysed reaction and surface-enhanced Raman spectroscopy (SERS) with an ultrahigh sensitivity up to single-molecule level, SERS is allowed to in-situ monitor the SPR-catalysed reaction process through providing fingerprint spectra of surface species11,12,13,14,15,16. Some theoretical and indirect experiment proofs from SERS have assumed the oscillated hot electrons and molecular O2 are responsible for the SPR-catalysed redox reactions17,18. However, the active species determining the primary reaction step have still not been determined due to a lack of direct experimental proof.

Here, in-situ SERS technique aided with radical capturers for hot hole, electron and secondary radicals (•OH and •O2−), generated from the plasmon decay of Ag nanoparticles (NPs), has unprecedentedly been used to determine the active species by adopting the oxidation of p-aminothiophenol (PATP) as a model reaction19,20,21,22. Hole is revealed to be directly responsible for the oxidation of PATP to p, p′-dimercaptoazobenzene (4, 4′-DMAB)23,24. Chu prepared Ag/TiO2 system to investigate this reaction. The result shows that hot electrons generating from plasmonic decay would transfer to the conduction band of TiO2, which enables the holes to trigger the immediate conversion25. The role of O2 is determined as an electron capturer to produce isolated holes. PATP can be further oxidized to p-nitrothiophenol (PNTP) within 10 s through a joint utilization of electron capturers of AgNO3 and atmospheric O226,27. To the best of our knowledge, this is the first time to unambiguously reveal the active species for the SPR-catalysed redox reaction through the in-situ SERS technique.

Result and Discussion

Assembled Ag nanoparticle layer spin-coated from 50 nm of Ag NPs on a glass slide is used as plasmon-active substrate both for SERS analysis and catalysis reaction, which shows a wide plasmon resonance absorption band centred at 450 nm (Figures S1–S4). For assembled Ag nanoparticle layer adsorbed with PATP (Ag-PATP), Ag NPs and PATP are premixed before spin-coating. As shown in Fig. 1a, the SERS signal is collected under the irradiation of a 532 nm laser within 10 s. Compared with the Raman signal of PATP on the glass slide, three new peaks (blue line) assigned to the bend vibration of CH (β(CH)) at 1145 cm−1 and the stretching vibration of N = N (ν(N = N)) at 1390 and 1437 cm−1 are observed, implying PATP is oxidized to 4, 4′-DMAB as driven by the SPR effect of Ag nanoparticle layer28,29,30,31,32,33,34. All the three strongly enhanced peaks of DMAB represent symmetric ag vibrational modes, strongly indicating the formation of DMAB from PATP through the N = N bond. These peaks are assigned to the ag12, ag16, and ag17 symmetric vibrational modes, respectively35,36. However, it is found that when ammonium oxalate (AO), a commonly used hole capturer is present, the reaction from PATP to 4, 4′-DMAB is completely quenched since no signal from 4, 4′-DMAB is observed. This result indicates the capturer-assisted strategy actually allows SERS to in-situ explore the reaction process in spite of the ultrafast relaxation process of hot carriers. Since both hole and its secondary radical •OH are strong oxidants, it still cannot be distinguished that PATP is oxidized by hole or •OH. Therefore, t-butanol (TBA) as the capturer for •OH is further adopted to understand its influence on the reaction37. However, strong peaks attributed to 4, 4′-DMAB is still observed with preserved intensity, suggesting the reaction from PATP to 4, 4′-DMAB is not altered by •OH (Fig. 1b). Therefore, it is undoubted that the oxidation of PATP is directly related to the hole decayed from the surface plasmon resonance of Ag nanoparticle layer. To get the original spectrum of PATP, it also can be detected on the Ag substrate by the irradiation of 785 nm on lower laser powers such as 0.50 μW and 0.25 mW (Figure S5).

Figure 1
figure 1

(a) Raman signal of PATP solid (black line), SERS signals of PATP on Ag nanoparticle layer in the absence and presence of AO; (b) SERS signal of PATP on assembly Ag nanoparticle layer in the presence of TBA; (c) Schematic diagram of plasmonic reactions on Ag layer in the presence of AO or TBA. Laser wavelength, 532 nm; Power, 2.5 mW; Integration time, 10 s.

Full size image

The above results seem to be inconsistent with the current reports, where the hot electron together with oxygen is generally considered to be responsible for the plasmon-driven oxidation reaction38,39. To reveal the real role of oxygen and electron during the oxidation of PATP, the reaction was further carried out in N2 atmosphere (Fig. 2a), which is significantly prohibited, suggesting O2 indeed contributes to the oxidation of PATP. It is highly possible that O2 may be reduced by hot electron to •O2− with strong oxidation capacity, which further cause the oxidation of PATP. To check this conjecture, a typical •O2− capturer, p-benzoquinone (BQ), is further applied in the SERS analysis. However, the synthesis conducted in the atmospheric environment in the presence of BQ does not cause any variation of 4, 4′-DMAB signals (Fig. 2b), excluding the possible effect of •O2− on the oxidation of PATP. Intriguingly, when AgNO3 is adopted as an electron capturer in N2 atmosphere, the oxidation of PATP to 4, 4′-DMAB occurs even in the absence of oxygen. A similar experiment was conducted in the water solution40,41,42,43 and the result confirm our suppose (Figure S6). As figure shows, in the water solution, the hole sacrificial agent AO is useful to prohibit the generation of DMAB. And the signal intensity of detected molecule seems weaker for the poor plasmon in the solution.

Figure 2
figure 2

(a) SERS signals of PATP on Ag and Ag-AgNO3 substrates in N2 atmosphere; (b) SERS signals of PATP in the presence of BQ; (c) EPR spectra of DMPO on Ag nanoparticles in the absence and presence of AgNO3; (d) UV-Vis spectra of Ag-AgNO3 substrates before and after the light irradiation. The characteristic peaks of 4, 4′-DMAB labelled by ▼.

Full size image

Electron spin resonance (ESR) is then further adopted to analyse the function of AgNO3 in the oxidation of PATP by using 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) as the indicator of •O2−. It is obvious that the intense signal appears on Ag nanoparticle layer under the laser irradiation, but the presence of AgNO3 causes the decreased peak intensity of DMPO-•O2− (Fig. 2c), which thus proves two facts as follows. First, O2 adsorbed on the surface of Ag nanoparticle layer can actually be reduced by SPR-derived electron. However, the presence of superoxide radical has no effect on the oxidation of PATP; Second, hot electrons can be indeed captured by AgNO3 according to the retarded formation of •O2−. Based on these two facts, the oxidation of PATP in the absence of O2 should be attributed to the improved concentration of holes due to the capture of electron by AgNO3 (eqs 1–3, SI).

Generally, the SPR-derived hole-electron exciton from Ag nanoparticle layer resides in the Fermi energy level, which is harder to be separated than that formed from semiconductor44,45,46,47,48,49. Since the hole has been revealed as the exclusively active species for the oxidation of PATP instead of O-containing oxidants, the retarded oxidation from PATP to 4, 4′-DMAB in the absence of O2 should be attributed to the inefficient separation of hole from electron. As such, molecular O2 should actually function as an electron capturer, which consumes electrons and produce enough holes to initiate the oxidation of PATP (eq 4, SI). Moreover, it is noted the plasmonic absorption of Ag nanoparticle layer in the presence of AgNO3 (Ag-AgNO3) is enhanced under UV-light irradiation (500 W Xe light, Fig. 2d), implying a possible reduction of AgNO3 to Ag during the SERS analysis. To understand the influence of enhanced plasmon resonance intensity on the reaction, the SERS of PATP on the pre-irradiated Ag-AgNO3 assembled nanoparticle layer has been further investigated in N2 atmosphere. The signal intensity attributed to 4, 4′-DMAB seems too low to be detected (Figure S7), thus excluding the contribution from improved plasmon resonance intensity to the conversion efficiency.

Furthermore, it is found that when both of AgNO3 and atmospheric O2 are present, PATP is unprecedentedly transformed into a mixture of PNTP and 4, 4′-DMAB within 10 s as characterized by the appearance of ν (NO2) peak at ca. 1330 cm−1 (Fig. 3a), demonstrating the oxidation efficiency can be enhanced by improving the separation degree of hole-electron exciton. Both of AgNO3 and O2 should be involved in the oxidation of 4, 4′-DMAB to PNTP since no PNTP can be produced when either of them is absent. The influence of AgNO3 density (ρAg) on the conversion efficiency was further investigated. It is found from Fig. 3a that 4, 4′-DMAB is the dominant product at a laser power of 0.5 mW and ρAg of 2.3*10−6g/cm2, as evidenced by the strong peaks at 1440, 1380 and 1140 cm−1 from 4, 4′-DMAB and a weak peak at 1330 cm−1 due to the ν(NO2) of PNTP. The peaks of 4, 4′-DMAB decreases when the laser power and ρAg increase to 2.5 mW and 2.3*10−5g/cm2, along with an increasing intensity of ν (NO2) peak. The intensity ratio between peaks at 1330 and 1390 cm−1 (I1330ν(NO2)/I1390ν(N = N)) was plotted to more clearly demonstrate the co-effect of AgNO3 and laser power (Fig. 3a, red line). The ratio of I1330ν(NO2)/I1140β(CH) was also plotted and used as a reference (Fig. 3b, black line). The I1330ν(NO2)/I1390ν(N = N) and I1330ν(NO2)/I1140β(CH) obtained at 2.5 mW are almost doubled compared with those formed at 0.5 mW when ρAg is 2.3*10−6g/cm2, which are further improved for ca. 30% and 80% when ρAg is increased to 2.3*10−5g/cm2. A higher ρAg leads to the decreased intensity of both 4, 4′-DMAB and PNTP (not shown), implying the shielding of Ag nanoparticle layer by overmuch addition of AgNO3. The actual composition of PNTP should be higher as valued from the peak intensity ratio between the a1 and ag modes since the intensities of the ag modes of 4, 4′-DMAB are significantly stronger than those of the a1 modes of PNTP50, where a small amount of 4, 4′-DMAB may already produce observable signal in SERS spectra. What’s more, to investigate the role of laser power and exposure time51, Ag-AgNO3 substrate was taken to detect PATP under 0.5 and 0.25 mW with different exposure times as Figure S8. The laser is both used for the light source of the SPR reaction and SERS analysis. The decreasing of the laser power decreases the reaction efficiency and so does the SERS sensitivity. We further investigated the reaction under 0.25 mW irradiation, where the Raman intensity is decreased without obvious variation of the reaction efficiency. The variation of the irradiation time to 5 s or 3 min does not obviously cause the change of the reaction process. However, when the exposure time extends to 10 min, the sample seems to be destroyed by the strong laser power, weakening the signal of the PNTP and DMAB on the substrate.

Figure 3
figure 3

(a) SERS signals of PATP on Ag-AgNO3 substrates under the irradiation of 532 nm laser light; (b) Relation between the density of AgNO3 and the composition ratio of PNTP/4, 4′-DMAB.

Full size image

As a further check for the feasibility of radical-capture strategy to the in-situ SERS analysis of other SPR catalytic reactions, the SPR-catalysed reduction of PNTP, another typical reaction model has been further adopted here31. The results shown in Fig. 4a indicate the reduction of PNTP is retarded when AgNO3 is present, accordant with the commonly-accepted understanding about the function of electrons in the SPR-catalysed reduction reaction3,4. On the contrary, the reduction can be promoted by conducting the reaction in N2 atmosphere (Fig. 4b), which should be attributed to the eliminated consumption of electron by molecular O2. Even more, a higher reduction degree of PNTP is achieved when AO is adopted to improve the electron concentration through consuming more holes (Fig. 4b).

Figure 4
figure 4

(a) Raman spectra of PNTP on the glass slide, SERS signals of PNTP on assemble Ag layer in the absence and presence of AgNO3. (b) SERS signals of PNTP on the Ag and Ag-AO substrates in N2 atmosphere.

Full size image

Conclusions

In summary, we have explored the mechanism of SPR-catalysed reaction by a capturer-assisted SERS strategy using the oxidation of PATP on Ag nanoparticle layer as the model reaction. The adoption of AO and AgNO3 as the capturers for hole and electron effectively leads to the separation of SPR-derived hot holes and electrons. The hot hole is directly responsible for the oxidation of PATP to 4, 4′-DMAB. The oxidation of PATP is prohibited in N2 atmosphere but occurs when AgNO3 is further present. O2 plays the role as an electron capturer in promoting the separation of hole-electron. The oxidation of PATP to PNTP has been unprecedentedly achieved in the atmospheric environment when the reaction is assisted by AgNO3. This study provides a novel way to deeply understand the mechanism of plasmon-related photocatalysis and photochemical reactions, which is expected to substantially push the development of SPR-induced green synthesis forward through rational and scientific design.

Method

Experimental Section

Preparation of Ag Nanoparticles with Diameter of ca. 50 nm

The preparation of Ag nanoparticles is adopted from a previously reported method52. In a typical synthesis, 85.0 mg of PVP was dispersed into 20.0 mL of water under magnetic stirring. After the complete dissolution of PVP, 85.0 mg of AgNO3 and 200 μL of 5.0 M NaCl were successively added under rapid stirring. The mixture was kept stirring in the dark for 15 min to form AgCl colloid. The freshly prepared AgCl colloid was then used as the precursor for Ag nanoparticles to achieve a size-controlled synthesis. First, 20.0 mL of 50.0 mM ascorbic acid was added to 2.5 mL of 0.5 M NaOH under magnetic stirring, and 2.5 mL of freshly prepared AgCl colloid is also added. The mixture was then stirred for 2 h in the dark. The products are collected by centrifugation and washed with water and kept in an ethanol solution.

Preparation of assembled Ag nanoparticle layer

Assembled Ag nanoparticle layer was spin-coated from the ethanol solution of Ag nanoparticles (1.0 mL, 0.05 M) and used for SEM, AFM and UV-Vis diffuse reflectance analyses. PNTP or PATP (200 μL, 10−2M) was dispersed on Ag nanoparticle layer by premixing the molecules with the Ag ethanol solution and the mixture was spin-coated using the above procedure. For the study in the presence of capturing agent, 200 μL ethanol solution of AgNO3, AO, CH3OH and BQ (10−3–10−2M) was mixed with 200 μL PNTP or PATP solution (10−2M) before the spin-coating process.

Characterization

Scanning electron microscopy (SEM) analysis was performed using a TESCAN nova III scanning electron microscope. Transmission electron microscopy (TEM) analysis was performed using a JEOL 2100 LaB6 TEM, at a 200 kV accelerating voltage. Raman spectra were recorded on a Renishaw inVia-Reflex Raman microprobe system equipped with Peltier charge-coupled device (CCD) detectors and a Leica microscope. Spectra were collected from the nanoparticle layer with an accumulation time of 10 s. Lasers with wavelength of 532 nm and 785 nm were used as the excitation light source, and a 50× objective with a numerical aperture (NA) of 0.75 was used to get the laser spot diameter of ~1 μm. The electron spin resonance (ESR) technique (with DMPO) was used to detect the radical species over the catalyst on a Bruker EMX-8/2.7 spectrometer. DMPO was added to the suspension system before testing, and then the system was irradiated by visible light using a Xenon lamp. Electron spin resonance (ESR) technique is a very powerful and sensitive method for the characterization of the electronic structures of materials with unpaired electrons. By investigating the resonance line can obtain the information about status of the unpaired electrons in radical and its surrounding environmental, thereby obtaining information about the structure and chemical bonding of the substance, in order to identify the different types of free radicals and their levels.

Additional Information

How to cite this article: Yan, X. et al. Surface-Enhanced Raman Spectroscopy Assisted by Radical Capturer for Tracking of Plasmon-Driven Redox Reaction. Sci. Rep. 6, 30193; doi: 10.1038/srep30193 (2016).