Introduction

Surface plasmon resonance (SPR) is a collective oscillation of surface conduction electrons that is excited by incident light at specific photon energies. This phenomenon has been exploited in plasmon-driven chemical reactions1,2,3,4,5,6,7,8,9,10, which have attracted significant attention because of their high throughput and low energy requirements. Hot electrons with high kinetic energy have been generated from plasmon decay11,12. These hot electrons play a key role in plasmon-induced chemical reactions by providing the required electrons and energy for the reactions to proceed.

Recently, extensive studies have demonstrated that p-aminothiophenol (PATP) absorbed on metal nanostructures can be selectively transformed into a new azo species, p,p’-dimercaptoazobenzene (DMAB), by a plasmon-driven catalytic reaction13,14,15,16,17. In this oxidation reaction, two 4-amine (-NH2) groups of PATP are selectively coupled to a –N=N– chemical bond to form DMAB. A similar aromatic compound, 4-nitrothiophenol (4-NBT), can also be catalyzed into DMAB with the assistance of plasmons18,19,20,21. In this reduction reaction, two 4-nitro groups of 4-NBT are selectively reduced to an azo bond. The two aforementioned reactions are strongly affected by the nature of the substrate materials22, the irradiation wavelength and power23,24, the solution pH25 and the ambient atmosphere17 or aqueous enviroment26. However, the plasmon-enhanced catalysed reactions on metallic catalysts can be monitored and controlled by surface-enhanced Raman scattering (SERS)14,21 and tip-enhanced Raman spectroscopy (TERS)27.

2-amino-5-nitrobenzenethiol (2A-5-NBT, see Fig. 1a) is an interesting molecule that contains benzyl, amine, nitro and -SH groups. This molecule can be used to study the competition between plasmon-driven oxidation and reduction in different environments, e.g., aqueous or ambient atmospheric environments. These three different molecules, 2,2’-dimercapto-4,4’-dintroazobenzene, 3,3’-dimercapto-4,4’-diaminoazobenzene and 2,3’-dimercapto-4-nitro-4-amino-azobenzene (see Fig. 1b–d) could be oxidized or reduced by SPR. This molecule can also be used to determine the most stable environment for plasmon-driven chemical reactions.

Figure 1
figure 1

Molecular structures.

(a) 2-amino-5-nitrobenzenethiol, (b) 2,2’-dimercapto- 4,4’-dintroazobenzene, (c) 3,3’-dimercapto-4,4’-diaminoazobenzene and (d) 2,3’-dimercapto-4-nitro-4-amino-azobenzene.

In this study, we used SERS spectroscopy in conjunction with theoretical calculations to investigate the plasmon-driven selective reduction reaction of 2A-5-NBT, which was chemically absorbed on a roughened Ag substrate and was dimerized to a new azobenzene derivative compound, 3,3’-dimercapto-4,4’-diaminoazobenzene. This selective reduction reaction proceeded via the nitro group of 2A-5-NBT and thus, oxidation via the amine group of 2A-5-NBT did not occur. The product was very stable in aqueous environments. However, in ambient atmospheric environments, the product was not stable and reverted to 2A-5-NBT, where the resulting –N=N– chemical bond could be further dissociated by plasmon scissors.

Our findings provide a novel environmentally friendly synthetic method for forming azobenzene derivative compounds, which are important and valuable industrial dyes28 and have potential applications as molecular wires and switches in molecular electronics29,30,31.

Results

Characterisation of SERS-active substrates

The substrate for the SERS measurements was prepared using a previously reported method26. The SEM image (Fig. 2) of the roughened Ag electrode shows three-dimensional (3D) “hot spots” or 3D nanogaps26. The SEM image shows that the substrate was roughened along and perpendicular to the surface. This structure increased the number of “hot spots” and the intensity of the SPR for the plasmon-driven chemical reactions32.

Figure 2
figure 2

SEM image of Ag substrate.

Plasmon-driven chemical reactions in an aqueous environment

Figure 3(a) and (b) shows the measured Raman spectrum of the 2A-5-NBT powder and the corresponding simulated Raman spectrum, respectively, which are in good agreement. The theoretical calculations showed that the strongest Raman peak at 1302 cm−1 in Fig. 3(a) could be attributed to the -NO2 stretching mode of 2A-5-NBT.

Figure 3
figure 3

Molecular Raman spectra.

(a) Raman spectrum of 2-amino-5-nitrobenzenethiol powder, (b) simulated Raman spectrum of 2-amino-5-nitrobenzenethiol, (c) SERS of 2-amino-5-nitrobenzenethiol in an aqueous environment and (d) simulated Raman spectrum of 3,3’-dimercapto-4,4’-diaminoazobenzene.

Figure 3(c) shows the SERS spectrum of 2A-5-NBT that was measured in an aqueous environment. The profiles in Fig. 3(a) and (c) are significantly different from each other. SPR-assisted chemical reactions appear to have occurred for 2A-5-NBT in the aqueous environment. As discussed in the Introduction section, three types of molecules can be synthesised by different reduction or oxidation reactions. The simulated Raman spectrum for 3,3’-dimercapto-4,4’-diaminoazobenzene is shown in Fig. 3(d). When compared, Fig. 3(c) and (d) clearly shows that 2A-5-NBT was selectively reduced to 3,3’-dimercapto-4,4’-diaminoazobenzene via two nitro groups from 2A-5-NBT molecules, whereas the 2-amine group of 2A-5-NBT was unchanged. The vibrational modes of the strong Raman peaks of 3,3’-dimercapto-4,4’-diaminoazobenzene are shown in Fig. 4, where the peak at 1436 cm−1 was attributed to the –N=N– stretching mode.

Figure 4
figure 4

Molecular vibrational modes of 3,3’-dimercapto-4,4’-diaminoazobenzene.

We studied the stability of plasmon-driven selective reductive reactions by measuring the time-dependent sequential SERS spectra in an aqueous environment (see Fig. 5). These plasmon-driven selective reduction catalysis reactions were found to be very efficient and the resulting product was very stable.

Figure 5
figure 5

Time-dependent sequential SERS spectra in an aqueous environment.

The acquisition time of each spectrum is 90 seconds and the time interval is 3 minutes.

Plasmon-driven chemical reactions in an ambient atmosphere environment

We studied the effect of different environments on plasmon-driven selectively catalysed reactions by measuring the time-dependent sequential SERS spectra (see Fig. 6). At the initial stages of the chemical reaction, the reactant 2A-5-NBT was clearly observed (see the upper spectrum in Fig. 6). With increasing time, 2A-5-NBT was selectively catalysed to 3,3’-dimercapto-4,4’-diaminoazobenzene (see the lower spectrum in Fig. 6). Subsequent SERS measurements showed that the product reverted to the reactant via the dissociation of the –N=N– chemical bond by plasmon scissors. There have been several reports on the dissociation of the –N=N– chemical bond by plasmon scissors using SERS33 and TERS34.

Figure 6
figure 6

Time-dependent sequential SERS spectra in an ambient atmospheric environment.

The acquisition time of each spectrum is 90 seconds and the time interval is 3 minutes.

Discussion

2A-5-NBT contains an amine group (–NH2) and a nitro group (–NO2); hence, it is difficult to ascertain which nitrogenous group participates in the selective formation of the azo bond. We considered three potential coupling modes in simulating the Raman spectra of the corresponding products 2,2’-dimercapto- 4,4’-dintroazobenzene, 3,3’-dimercapto-4,4’-diaminoazobenzene and 2,3’-dimercapto-4-nitro-4-amino-azobenzene (see Fig. 7(a)–(c)), which are denoted as molecules A, B and C, respectively. The molecular structure graph of molecule A in Fig. 7(a) shows that a –N=N– chemical bond could form at the 2’2 coupling via two 2-amine groups from 2A-5-NBT molecules. Similarly, a second possible molecule B (see Fig. 7(b)) could be selectively produced through the coupling of two 5-nitro groups of 2A-5-NBT molecules. The plasmon could also convert the 2A-5-NBT to a third possible product, C, (see Fig. 7(c)) by the selective coupling of a 2-amine group and a 5-nitro group of 2A-5-NBT.

Figure 7
figure 7

Simulated Raman spectra.

(a) 2,2’-dimercapto-4,4’-dintroazobenzene, (b) 3,3’-dimercapto-4,4’-diaminoazobenzene and (c) 2,3’-dimercapto-4-nitro-4-amino-azobenzene.

We determined the selected product from the plasmon-driven catalytic coupling reaction by comparing the SERS spectrum of 2A-5-NBT on the roughened Ag substrate with the calculated Raman spectra for the three products. For molecules A and C, the strong Raman peaks at approximately 1330 cm−1 were attributed to the NO2 stretching modes (see Fig. 8) that were not observed in the experimental results in Figs 3(c) and 5, which is a distinct and unique experimental signature. The simulated Raman spectrum in Fig. 7b was consistent with the experimental SERS spectrum in Fig. 3(c). Thus, we concluded that 2A-5-NBT was selectively reduced to 3,3’-dimercapto-4,4’-diaminoazobenzene. The reduction reaction occurred before the oxidation of the molecules with amine and nitro groups. Different environments were found to strongly influence the stability, probability and efficiency of selective catalysis reactions by plasmons.

Figure 8
figure 8

Molecular vibrational modes.

(a) -NO2 vibrational modes of 2,2’-dimercapto- 4,4’-dintroazobenzene and (b) -NO2 vibrational modes of 2,3’-dimercapto-4-nitro-4-amino-azobenzene.

Methods

2A-5-NBT was purchased from Beijing Kaida Co., according to the customer’s requirements. The NMR spectrum of 2A-5-NBT is presented in the supporting information.

The Ag substrate (a single-crystal silver rod of 99.99% purity) was polished with emery paper and cleaned with Milli-Q water in an ultrasonic bath. The Ag substrate was then placed in a typical electrochemical cell containing a 0.1 M Na2SO4 solution for roughening. A double potential step was used to roughen the surface by applying a voltage of +0.25 V for 8 s and then applying a voltage of −0.35 V.

The Raman spectrum of the 2A-5-NBT powder and the SERS spectrum of 2A-5-NBT (1 10−6 M) were measured using Raman spectroscopy (Renishaw inVia system) with a laser incident wavelength of 632.8 nm.

All of the theoretical calculations were performed with the Gaussian 09 suite using density functional theory35 and the PW91PW91 functional36 with b3lyp/6-311++g(d,p) for the reactant and the 6-31G(d) basis set for the three products.

Additional Information

How to cite this article: Ding, Q. et al. Effect of aqueous and ambient atmospheric environments on plasmon-driven selective reduction reactions. Sci. Rep. 5, 10269; doi: 10.1038/srep10269 (2015).