1. Introduction

Bromate (BrO₃⁻) is formed during the disinfection of water typically at the last stage of the water treatment process. Bromate anions appear in conversion of bromide to bromate in the course of water ozonation, Br⁻ + O₃ → BrO₃⁻ [1]. Under certain conditions, bromate may also be formed in concentrated hypochlorite solutions used to disinfect drinking water [2]. This reaction becomes possible owing to the presence of bromide in the raw materials (chlorine and sodium hydroxide) used in the manufacture of sodium hypochlorite and to the high pH of the concentrated solution [3]. There are also industry-related paths for bromate to appear in food and water.

Bromate is suspected to be a carcinogenic agent to humans. To date, sufficient evidence in animals have been reported though no data for humans available yet [4]. Therefore the bromate content in water should be carefully controlled to minimize suspected cancer risks. The currently recommended upper level of 10 μg/L in drinking water has been set as a reasonable compromise of measurement techniques limitation like ion chromatography (IC), ultraviolet/visible absorbance detection, and inductively coupled plasma–mass spectrometry (ICP-MS) [3]. The health-based value should be reasoned well below 10 μg/L levels. For these health-based limits to be attainable, the concentration of bromate in drinking water should be carefully monitored and controlled. A number of methods can analyze bromate with very low detection limits and without chloride interferences, but involve multistep reactions containing removal of free bromide or suffer from interference in chlorinated waters. This has urged the need for more versatile and cost-effective methods for on-site detection of bromate in water.

Surface enhanced Raman scattering (SERS) is the ultrasensitive detection technique based on enhancement of Raman scattering down to a single molecule detection limit [5, 6]. This huge enhancement caused molecules which occupies certain “hot spots” on a metal singular-topological structure where the most favourable combination of enhancement factors occurs at the wavelength of light used to excite a molecule and the wavelength of scattered light. In typical experiments, ensemble averaged enhancement factors about 10⁶ have been reported by many groups [7]. The phenomenon in general is rather impressive and gives readily a possibility to detect molecules at concentration where Raman signals from a solution are not detectable at all [8, 9].

In spite of the numerous experimental reports on enormous Raman intensity enhancements in metal nanostructures, SERS application in analytical practice makes only the first steps. The main obstacle is quantification of the data in terms of analyte concentration, which is mandatory in most of analytical applications. Another principal problem is non-specificity of enhancement. In many cases, not only Raman signals from target molecules rise up but the Raman signatures from all presenting compounds in a probe experience enhancement as well.

During last years, SERS comes actually closer to routine practical usage. Pesticide detection by SERS at 10 ppm level with silver nanoparticles (AgNPs) has been reported [10]. SERS has been also applied to detect and quantify the concentration of glycated albumin, an important glycemic marker for long-term diabetes monitoring [11]. The ultrasensitive technique for 2,4,6-trinitrotoluene (TNT) detection using SERS has been elaborated [12]. Recently, SERS-based technique of antimony detection has been presented [13]. Similarly, SERS has been suggested for determination of nitrous oxide in air. Trace NO₂^- had strong catalytic effect on the slow redox reaction of bromate with Rhodamine 6G (Rh6G) that caused the SERS quenching at 1507 cm⁻¹ band of R6G adsorbed on the triangular plate-type AgNPs aggregates [14]. Also, some works have been already done using the chromophore fluorescence detection of bromates [2, 4].

In the present article, the catalytic reaction between potassium bromate (KBrO₃) and Rh6G as SERS/fluorescent probe is tested on applicability for ultimate bromate detection in water. We consider it as a very promising approach for the bromate determination of the concentration much lower than the detection limits provided by IC, ICP-MS and colorimetric determination methods. Plasmonic Ag films consisting of the electrostatically immobilized AgNPs were used as SERS substrates. AgNPs are widely used for SERS active plasmonic substrates due to the facile synthesis [15–17]. Plasmonic Ag films are fabricated by means of electrostatic immobilization of AgNPs on glass substrates. Similar films have been previously used in SERS experiments with semiconductor colloidal quantum dots [17] and also for fluorescence enhancement of labeled proteins [18]. To the best of our knowledge, this is the first report on investigating the feasibility of using SERS for bromate detection in water.

2. Experimental

AgNO₃, trisodium citrate, poly(diallyldimethylammonium) chloride (PDADMAC, Mw = 200,000), Rh6G, NaNO₂, H₃PO₄, KBrO₃, KClO₃, KIO₃ and all other reagents were of analytical grade. All the solutions were prepared in the deionized (DI) water.

Ag sol was synthesized by AgNO₃ citrate reduction technique [18]. Briefly, 9 mg of AgNO₃ was dissolved in 50 mL of DI water, and 1 mL of 1% sodium citrate was added dropwise when the mixture began to boil. The reaction vessel was boiled for 1 h, ending with formation of green-brown sols. To make SERS-active substrates, glass substrates of 1 × 2 cm² in size were cleaned by means of ultrasonic treatment in isopropyl alcohol and then kept in a mixture of H₂O-H₂O₂-NH₃ (1:1:1) at 70° C for 15 min. After washing with water, substrates were covered with a polycation layer (PDADMAC, 1 g/L in 0.5M NaCl) during 20 min to develop a positive charge on a glass surface. Since Ag particles were negatively charged they were deposited on PDADMAC-modified substrates by dipping the substrate surface in Ag sol for 24 h.

A series of solutions was prepared by the following procedure. 200 μL of 0.5 × 10⁻⁵ M Rh6G solution, 200 μL of 10⁻⁴-10⁻¹⁰M KBrO₃ solution, 100 μL of 1M H₃PO₄, and 100 μL of NaNO₂ were mixed and diluted to 2 mL. The reference Rh6G solution containing H₃PO₄ and NaNO₂ was prepared without bromate. The mixtures were treated for 15min at 50°C, and then cooled. 5 μL of each mixture was dropped to the plasmonic Ag film, dried and the SERS spectra were recorded. The analytical value of light intensity (ΔI) was calculated as the difference between I₁₅₁₂ Rh6G band of the reference sample without KBrO₃ and I₁₅₁₂ of each sample containing the bromate.

The neodymium-doped lanthanum scandium borate (Nd:LSB) laser (531 nm) was used to excite the Raman spectra. The registration system consisted of a spectrograph with a 1200 lines/mm diffraction grating (Solar TII S3801, Belarus) and a cooled CCD matrix (Princeton instruments, USA). Raman measurements were performed in the backscattering configuration at room temperature with the typical acquisition time of 20 s.

Optical extinction of Ag films, and sol have been characterized by optical density (OD) spectra, D = − log(I/I₀) with I(I₀) being transmitted (incident) light intensity, using a Cary-500 spectrophotometer (Varian, USA). Electron microscopy has been performed using a Zeiss Gemini instrument.

3. Results and discussion

An example of typical electron microscope image of Ag film on glass substrate along with the OD spectra of Ag-sol and Ag-covered glass substrate is given in Fig. 1. SEM image in the inset show that film contains poly-dispersed spherical particles with the average diameter of approximately 45 nm, and a small portion of particles featuring rod-like shape. One can see that the principal optical extinction maximum of nanoparticles in a film at about 400 nm features blue shift with respect to sol. This may result from the effect of dielectric environment on surface Plasmon frequency. Notably, film features higher long wave extinction (600-700 nm) which is attributed to aggregated nanoparticles.

 

Fig. 1 Optical density spectra of Ag sol and Ag film. The electron microscope image of Ag-film is presented in inset.

Download Full Size | PDF

Our first attempts to direct detection of bromates by SERS method using such Ag substrates have not been successful. Also to the best of our knowledge, there are no works concerning direct determination of bromate in water with SERS. So, we attempted to use the probe molecule for this purpose.

Rhodamine 6G (Rh6G) well adsorbs on AgNPs, exhibits pronounced SERS effect and therefore was chosen as probe molecule. It has also strong fluorescence band at 564 nm. The latter was supposed to experience quenching at direct contact with AgNPs [20] and therefore will not affect Raman line detection.

Since bromate has been reported as Rh6G oxidant [14], we expected that the ratio of oxidized/non-oxidized Rh6G forms will depend on bromate concentration and therefore on the Rh6G final SERS intensity. Concentrations of the reaction catalyst NaNO₂ and other presenting substances in solution were kept constant. Concentration of Rh6G and phosphoric acid were adapted as described previously by liu et al. [14] to get optimal conditions for Rh6G reaction with bromate to arrive at maximal change in SERS intensity.

Figure 2 shows results for SERS experiments containing different concentration of bromate in the range of 10⁻⁴−10⁻¹⁰ М along with the reference spectrum without bromate. Each of Stokes-shifted Raman peaks can be assigned to R6G vibrational modes [19–22]. Peaks at 1318, 1512, 1577 and 1652 cm⁻¹ correspond to symmetric modes of aromatic C-C stretching vibrations, the band at 617 cm⁻¹ corresponds to in-plane C-C-C bending mode, the band near 776 cm⁻¹ corresponds to out-of-plane bending mode of C-H bonding, the peak at 1131 cm⁻¹ is associated with C-H in-plane bending, and 1188 cm⁻¹ line is symmetric mode of C-C stretching [19–22].

 

Fig. 2 SERS spectra of Rhodamine 6G (Rh6G) on silver film in presence of 10⁻⁴ - 10⁻¹⁰ M KBrO₃. Rh6G concentration is 0.5 μmol/L. The uppermost spectrum label with (*) corresponds to the reference sample without bromate.

Download Full Size | PDF

The spectra presented in Fig. 2 clearly show that presence of bromate at concentrations of 10⁻⁴ - 10⁻¹⁰ М actually have effect on Rh6G Raman band intensity up to complete quenching of Raman signals at high concentrations, while no shift of the peaks position was observed. Upon increase in bromate concentration, every Raman band of Rh6G becomes less intense and the fluorescence background goes down as well. In [14], Q. Liu et al, bromate concentration was kept constant whereas NaNO₂ concentration was variable and the decrease in peak intensity at 1507 cm⁻¹ was used as indicator of Rh6G oxidation. The oxidation process was also confirmed by IR-spectra of Rhodamine indicating the characteristic absorption peak of Ar-N(H)-R groups at 1312cm⁻¹ were disappeared as well as the characteristic absorption peaks of C-N = O groups at 1287 cm⁻¹ and C-NO₂ groups at 1339 cm were developed [14]. The position of Rh6G peaks in SERS spectra can shifts on a few cm⁻¹ because of difference in surface roughness of the SERS substrates as well as by different spectrometer adjustment [21]. Our peak at 1512 cm⁻¹ in the R6G SERS spectra indicates R6G oxidation [14] and was chosen to draw a calibration curve of bromate concentration.

In our experiments, for bromate concentration range 0-10⁻⁷ М the calibration curve can be plotted since in this range SERS signal changes whereas spectral line position remains constant. This curve is presented in Fig. 3. One can see linear dependence in the above range. For higher bromate concentrations 10⁻⁶ М −10 ⁻⁴ M, SERS bands can hardly be detected due to the high fluorescent background and possibly completely vanish. Therefore presence/absence of Rh6G Raman bands can be used as qualitative validation of water against the allowable bromate level since World Health Organization (WHO) suggests a provisional guideline value of 10 μg/L (0.8*10⁻⁷ М).

 

Fig. 3 Variation of ΔI₁₅₁₂ line intensity as a function of KBrO₃ concentration.

Download Full Size | PDF

The sensitivity of the proposed technique has been compared with the existing ones recommended by the US Environmental Protection Agency (EPA) [3] as well as European Commission [23]. The most sensitive ICP-MS and colorimetric methods provide bromate detection limit of 0.1-0.3 μg/L while the practical quantification level is 1 μg/L. Figure 3 presents the analytical curve and shows that one can detect bromate in water using the SERS method at concentrations low as 10⁻¹⁰ M (nearly 0.01 μg/L).

4. Conclusion

In the present work, SERS detection has been applied to analyze the bromate at low concentration in water. It is based on the catalytic redox reaction between KBrO₃ and Rh6G causing the decrease in Rhodamine SERS as well as Rh6G fluorescent band intensity on the surface of plasmonic silver film. Compared to other works for the trace analysis of bromate anions, the detection sensitivity was enhanced up to 1-2 orders of magnitude. So the method is promising toward the cheap sensors development for the bromate analysis in drinking water. This could lead us to an ultimate goal of adapting and optimizing the emerging Raman nanosensors to develop portable and cost-efficiency test kit for the remote monitoring of bromate in drinking water with the practical quantification level well below 10 μg/L.