Introduction

Chemiluminescence (CL), a light emission induced by chemical reaction, has been intensively investigated for many years [1]. CL method has many advantages, such as simple equipment, high sensitivity and low background signals. The CL study has been extended to nanoparticles systems from traditional molecular systems among nanoparticles as a catalyst, reductant, luminophor, and energy acceptor [2, 3]. In particular, silver nanoparticles (AgNPs) have a redox potential lower than that of gold, platinum and so on. AgNPs have been used in the luminol [4, 5], bis (2,4,6-tricholorophenyl) oxalate [6, 7], Ce(IV) [8] and KMnO4 [9] system.

Catechol is a natural polyphenolic compound which is ubiquitous in nature since it commonly generates in factory processes and widely exists in higher plants such as teas, fruits, tobaccos and some traditional Chinese medicines. It is widely used in the area of cosmetics, tanning, pesticides, medicines and photography chemicals [10]. Catechol is easily absorbed from the gastrointestinal tract, causing vasoconstriction, renal tube degeneration, liver function decrease, cancers and neurodegenerative diseases, besides accumulating in bone marrow [11]. So, catechol is extremely harmful to animals, plants and aquatic environments even at very low concentration, and is considered one of the environmental pollutants by the US Environmental Protection Agency [12]. Consequently, sensitive and rapid simultaneous determination of catechol is required. Several analytical methods have been used for the determination of catechol, such as spectrophotometry [13, 14], electrochemistry [15–20], electrogenerated CL [21, 22], gas chromatography [23], high-performance liquid chromatography [24] and capillary electrophoresis [25–27].

The study showed that AgNPs induces the CL of the K3Fe(CN)6-fluorescein system but that the CL intensity of K3Fe(CN)6-fluorescein-AgNPs is reduced in the present of catechol and the decrease CL intensity is proportional to the catechol concentration in a certain range. Hence, a CL method has been designed for the analysis of catechol in environment waters. The reaction mechanism is also discussed.

Experimental

Apparatus

CL measurements were carried out on an IFIS-C intelligent flow injection sampler (Xi’an Remax Analytical Instrument Co. Ltd., Shannxi, China, http://www.chinaremex.com/) and BPCL-1-TIC ultra weak CL analyzer (Beijing Jianxin Lituo Company, Beijing, China, http://www.bpcl.com.cn/). CL spectra were measured with Cary Eclipse spectrofluorimeter (Varian, USA, http://varian.lookchem.com/). Absorption spectra were taken on a Cary5000 spectrophotometer (Varian, USA, http://varian.lookchem.com/). The centrifugal separation was carried out on the TGL-16G centrifuge (Shanghai Anting Scientific Instrument Factory, Shanghai, China, http://www.centrifuge.com.cn/).

Reagents

Catechol and silver nitrate were purchased from Tianjin Chemical Reagent Company (Tianjin, China, http://sjyc.company.lookchem.cn/). Poly(vinyl pyrrolidone) (the mean molecular weight is 10,000 g·moL−1), fluorescein, sodium citrate, sodium borohydride and K3Fe(CN)6 were purchased from Sinopharm Group Chemical Reagent Co. Ltd. (Shanghai, China, http://www.reagent360.cnal.com/).

The 0.01 M stock solution of catechol and K3Fe(CN)6 were prepared in water, respectively. The 0.1 mM working solution of K3Fe(CN)6 was prepared by diluting the stock solution in 0.1 M NaOH. The 0.01 M stock solution of fluorescein was prepared by dissolving fluorescein in 0.1 M NaOH and diluting to 100 mL with water. The working solution of fluorescein was prepared by diluting the stock solution in water. All the reagents used were of analytical reagent grades and doubly distilled water was used throughout the whole experiment.

Preparation of AgNPs

AgNPs was synthesized by chemical reduction method according to the literature [28] with some modification. In brief, 0.30 g poly(vinyl pyrrolidone) was dissolved in 10.0 mL water and mixed with the 10.0 mL 0.02 M AgNO3 solution. Then, 2.0 mL 0.1 M NaBH4 was added dropwise to the mixture under vigorous stirring. The mixture color changed to yellow gradually. The reaction was conducted at room temperature for 30 min and aged for 2 days at 4 °C. The raw colloid was centrifuged at 10,376 rcf for 30 min. AgNPs were characterized by UV-vis spectrophotometer and the surface plasmon resonance band of AgNPs appears at 430 nm. Statistical analysis of the TEM image revealed that the average diameter of the AgNPs was about 50 nm ± 5 nm, which was calculated by 50 particles. (Fig. S1, Electronic Supplementary Material, ESM). The concentration of AgNPs was calculated by the number of silver atoms contained in the sample.

Procedure

The flow injection CL detection system employed in this work was described in Fig. 1. The carrier water, K3Fe(CN)6 solution, the mixed solution of fluorescein and AgNPs were continuously pumped at a flow rate of 3.6 mL·min−1 into the flow cell. The catechol standard or sample solution was injected by an eight-way injection valve with a sample loop of 100 μL. The CL signal produced in the flow CL cell was monitored by a photomultiplier tube (PMT, operated at −700 V) and was recorded by a computer equipped with a data acquisition interface. Data acquisition and treatment were performed with BPCL software running under Windows XP. The concentration of catechol was quantified by measuring the decrease CL intensity, ΔI = I 0 – I s, where I 0 and I s were CL intensity in the absence and presence of catechol.

Fig. 1
figure 1

Schematic diagram of the flow injection chemiluminescence system

Full size image

Results and discussion

Kinetic characteristic of CL reaction

The CL intensity-time curve is shown in Fig. S2 (ESM). No detectable CL emission can be recorded for K3Fe(CN)6-fluorescein system in the absence of AgNPs. When K3Fe(CN)6 solution was injected into the mixture solution of fluorescein and AgNPs, a strong CL reaction was initiated immediately and the maximum CL intensity was obtained in 0.5 s (peak a). Therefore, the CL emission was attributed to the catalytic effect of the AgNPs on the K3Fe(CN)6-fluorescein system. Further, a inhibitory CL signal occurred after the catechol was added to the K3Fe(CN)6-fluorescein-AgNPs system (peak b).

Optimization of the reaction conditions

A series of experiments were conducted to select the optimum analytical conditions using a 1.0 μM catechol solution. The optimized parameters included the concentration of K3Fe(CN)6, fluorescein, NaOH and AgNPs, the flow rate of reagents.

The effect of K3Fe(CN)6 concentration on the CL reaction was examined in the range of 0.1–0.5 mM. The experiments showed the maximal CL signal is obtained when the concentration of K3Fe(CN)6 was 0.3 mM (Fig. S3A, ESM). The fluorescein concentration was studied in the range of 0.1–0.5 mM. As shown in Fig. S3B, ESM, the CL intensity increased with increasing the concentration of fluorescein. When the fluorescein concentration was in the range of 0.25–0.35 mM, the CL signal reached a maximum value, and then decreased above 0.35 mM. Therefore, the fluorescein concentration of 0.3 mM was selected for further studies. The effect of NaOH concentration from 0.06 to 0.14 M added in K3Fe(CN)6 solution was investigated (Fig. S3C, ESM). The CL intensity increased with the concentration of NaOH and reached the maximum at 0.1 M NaOH, then decreased. That might be due to the fact that the catalytic ability of AgNPs declined in the high alkaline media. Thus, 0.1 M NaOH was selected as the optimal concentration for further experiments. The AgNPs concentration was varied from 0.3 to 0.8 μM. The CL intensity increased steadily with AgNPs concentration and then decreased probably due to the strong interaction among the particles (Fig. S3D, ESM). It was found that when the concentration of AgNPs solution was 0.55 μM, CL signal reached the highest. Thus 0.55 μM was selected as optimum concentration.

The flow rate of reagents played an important role in CL measurements, as the time consumed to transfer the excited product into the flow cell is critical parameter for achieving maximum amount of the emitted light. Therefore, the flow rate was studied over the range 2.0–4.5 mL·min−1 with equal flows in each channel, while keeping all other conditions as constant. The results showed that the CL signal increased with the increasing of flow rate, then decreased. So, the flow rate of 3.6 mL·min−1 was selected under the present conditions.

Analytical performance

Under the optimum conditions, the relative CL intensity was linear to the concentration of catechol over the range 0.1 to 10.0 μM. The regression equation was ΔI = 261.8c + 69.24 (where c was the concentration of catechol in 1.0 μM) with a correlation coefficient of 0.9994 (Fig. 2). The determination limit was 5.0 nM catechol (3σ). The relative standard deviation was 1.5 % by 11 replicate determinations of 1.0 μM catechol. These indicated that the CL method has good linearity, relatively high sensitivity and precision and acceptable reproducibility. Compared with most of the nanomaterial-based methods [15–19, 21], this method had lower detection limit (Table 1).

Fig. 2
figure 2

Calibration plot for the determination of catechol

Full size image
Table 1 Figures of merit of recently reported nanomaterial-based methods for determination of catechol
Full size table

Interference

The influence of potential interferences was investigated by analyzing a standard solution of 1.0 μM catechol with varying amounts of interfering species. A substance was considered no interference if the variation of the CL intensity was within ±5 %. The tolerable concentration ratios of potentially interfering compounds to 1.0 μM catechol were over 3000-fold for NH4 +, Zn2+, Pb2+, Al3+, CO3 2− and HCO3 −; 1000-fold for K+, NO3 −, Na+, SO4 2−, Ca2+, Mg2+ and Cl−; 300-fold for Cu2+, Mn2+, ether, formic acid and alcohol; 100-fold for Fe3+, Co2+ and benzoic acid; 50-fold for Hg2+, Cd2+ and Fe2+; 10-fold for ortho-hydroxybenzoic acid and phenol; 5-fold for resorcinol and hydroquinone.

Analysis of tap water and lake water

The CL method has been applied to the determination of catechol in tap water and lake water. The results were given in Table 2. In order to evaluate the validity of the analysis method, these samples were prepared by adding the standard solution of catechol and were tested. The recovery experiments were carried out by adding a standard catechol solution to tap water and lake water. The recovery was range from 99 % to 105 %. The results confirmed the accuracy of the method.

Table 2 The determination of catechol in tap and lake water (n = 11)
Full size table

Chemiluminescence mechanism

The CL spectra of K3Fe(CN)6 -fluorescein system in the absence and presence of AgNPs was recorded and was shown in Fig. 3. Almost no CL signal was detected for the reaction between K3Fe(CN)6 and fluorescein. After adding AgNPs into the system a strong CL signal occurred (curve b). The maximum wavelength is 520 nm, which was in agreement with the fluorescent spectrum of fluorescein. This indicated that the CL emission was from the excited state of fluorescein, and AgNPs played an enhanced role in light emission. When catechol was added to the system, the CL intensity was markedly inhibited and the maximum wavelength was still 520 nm (curve a).

Fig. 3
figure 3

Chemiluminescence spectra

Full size image

The UV-vis spectra were investigated. Fig. S4A showed that K3Fe(CN)6 showed two absorption peaks at about 300 nm and 420 nm (curve a), fluorescein showed a absorption peak at about 492 nm (curve b) and AgNPs showed a absorption peak at about 430 nm (curve d). The absorption value of the K3Fe(CN)6 and fluorescein mixed system almost equaled to the sum of two individual systems (curve c), which meant that no obvious reaction occurred. When AgNPs was added to the system, the characteristic absorption peak of hexacyanoferrate and fluorescein disappeared. The absorption peak of the K3Fe(CN)6-fluorescein-AgNPs system (curve e) occurred at 443 nm and the absorption intensity greatly enhanced. Therefore, it is presumed that the reaction take place between K3Fe(CN)6 and fluorescein in the presence of AgNPs. The characteristic absorption peak of catechol displayed at 276 nm (Fig. S4B). When K3Fe(CN)6 was mixed with catechol, the characteristic absorption peak of catechol disappeared.

The solutions of K3Fe(CN)6, fluorescein and AgNPs were saturated with O2 and N2 by bubbling and then the system was examined under saturated conditions. The CL intensity for saturated and normal conditions decreased in the order O2 saturated > normal > N2 saturated, suggesting the participation of O2 in the course of the CL reaction.

Based on the above experiments and discussion, a possible CL reaction mechanism was as follows: under the catalytic action of AgNPs, a portion of fluorescein was oxidized by K3Fe(CN)6 accompanied by a release of energy and other fluorescein absorbed that energy and produced excited state fluorescein. The fluorescein of the excited state returned to the ground state, accompanied by CL. On the other hand, under the catalytic action of AgNPs, dissolved oxygen adsorbed on the surface of AgNPs formed superoxides. Superoxides were reacted with water to generate hydrogen peroxide [29]. The produced H2O2 was mixed with K3Fe(CN)6 in an alkaline medium to produce oxygen [30]. The produced superoxide radical reacted with the fluorescein to yield excited state fluorescein and the CL emission was produced when excited state fluorescein returned to the ground state. Catechol as a reducing agent was able to react with K3Fe(CN)6 and reduced the formation of excited state fluorescein, and then the CL emission was inhibited.

Conclusion

AgNPs are found to enhance the weak CL of the K3Fe(CN)6-fluorescein system. Based on the inhibitory effect of catechol on the CL reaction in alkaline solution, a sensitive CL method was developed for the determination of catechol in tap and lake water samples with satisfactory results. The method is rapid and simple, in addition to high sensitivity, low detection limit and less interference. This work was important for the investigation of efficient catalysts for weak CL system and the comprehension of CL mechanism correspondingly. Further investigation is necessary to decide whether the CL system can be suitable for other phenolic compounds and some drugs which contain phenolic hydroxy groups.