Rapid on-site detection of paraquat in biologic fluids by iodide-facilitated pinhole shell-isolated nanoparticle-enhanced Raman spectroscopy

Abstract

An on-site direct detection method for paraquat in biologic fluids was developed on the basis of iodide-facilitated pinhole shell-isolated nanoparticle-enhanced Raman spectroscopy. A full removal of matrix interference on the nanoparticle surface from the complicated biologic samples, and a selective attraction of paraquat to “hotspot” via a highly electrostatic adsorption on the iodide-modified nanoparticle surface are accomplished in this method. Four distinguished characteristics were provided as (1) high sensitivity with a limit of detection of 1 μg L⁻¹ for paraquat; (2) free of pretreatment with a direct measurement in plasma or urine finished within 1 minute; (3) on-site detection using a portable Raman spectrometer; and (4) high antifouling stability during measurement with the pinhole shell-isolated nanoparticle structure. This method was further showed its great clinical diagnosis applicability on a rapid, accurate detection of the plasma and urine samples from a paraquat poisoned patient.

---

1 Introduction

Paraquat (1,1′-dimethyl-4,4′-bipyridinium dichloride) is a bipyridylium quaternary ammonium compound, and is a typical nonselective herbicide with acute toxicity. People are well aware that paraquat is an effective herbicide with high mortality.¹ Up to now, hundreds of lethal cases have been reported due to the lack of specific antidotes and effective treatments.^2,3 An on-site, rapid and accurate detection technique is highly demanded for the early clinic diagnosis to guide the therapy with proper treatment, and to increase the survival chance of paraquat poisoned patients.

In the past decades, various methods including high performance liquid chromatography (HPLC),^4,5 gas chromatography-mass spectrometry (GC-MS),⁶ and liquid chromatography-tandem mass spectrometry (LC-MS/MS)⁷ have been established for the clinical diagnosis aim. These chromatographic separation-based methods provided high sensitivity and selectivity, with limits of detection (LODs) ranging from 10 to 100 μg L⁻¹ in biologic samples, but so sophisticated and time-consuming, and cannot meet the demand of rapid test. Enzyme linked immunosorbent assay (ELISA) and colorimetric methods have been established for on-site detection of paraquat.^8,9 For example, an ELISA method offered a limit of quantification (LOQ) of 2 μg L⁻¹ for paraquat in urine,⁸ while the pretreatment procedure is tedious, and the cross-reactivity of antibodies towards small organic molecules is a big problem and inclined to false-positive results in practical use. Colorimetric method is fast but with a limited sensitivity and low specificity due to poor discrimination between paraquat and its analogue diquat.⁹

In recent years, surface-enhanced Raman spectroscopy (SERS) is emerged as a powerful on-site technique for its unique characteristics of rapid detection, high sensitivity and fingerprint identification.^10,11 Benefited from the enhanced Raman scattering of analytes through a localized surface plasmonic resonance provided by nanostructural noble metals, the sensitivity of SERS can even reach a single molecule detection level.¹² Several SERS detections on paraquat for pesticide residue determination have been tested. For instance, Gao et al. reported a LOD at nM level (sub μg L⁻¹) of paraquat in aqueous samples with a microfluidic-based SERS method, and proposed an on-site screening potential in the water contamination.¹³ Tang et al. discriminated tricyclazole, paraquat and flusilazole in water at sub mg L⁻¹ level by SERS.¹⁴ However, while SERS is involved in the condition of biologic fluids, severe interferences from biological matrix with great amount of nitrogen- and sulfur-containing compounds overwhelmed the characteristic Raman shifts of paraquat in the same range of 600 to 1700 cm⁻¹.¹⁵ Therefore, no SERS detection on paraquat in biologic fluids has been reported until now.

To address this issue, here we describe a facile SERS based paraquat measurement in biologic fluids by coupling the advantages from pinhole shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS) and iodide addition, and a quantitative determination is achieved. The pinhole structured ultrathin SiO₂ shell (1–2 nm) isolated Au core nanoparticles (pinSHINs) not only left the active adsorption sites on the Au surface for target molecules, but also provided a far more stability than bare gold nanoparticles (bare AuNPs) during measurement. With the introduction of iodide as the cleaning agent and specific linker between pinSHINs and paraquat, the nonspecific adsorption interference from biologic matrices was effectively eliminated, which makes the impossible possible: a direct detection of paraquat in plasma or urine was successfully achieved without tedious pretreatment using a portable Raman spectrometer. The mechanism of the iodide-facilitated effect was discussed accordingly. A LOD of 1 μg L⁻¹ for either plasma or urine samples was obtained, and a good selectivity towards diquat, the analogue of paraquat was achieved. This method was further applied in authentic plasma and urine samples from a paraquat poisoned patient, and results were displayed with good sensitivity, selectivity and reliability.

2 Experimental

2.1 Chemicals and materials

Hydrogen tetrachloroaurate(III) trihydrate (HAuCl₄·3H₂O), sodium citrate and (3-aminopropyl) trimethoxysilane (APTMS) were purchased from Alfa Aesar (MA, USA). Potassium iodide (KI), magnesium iodide (MgI₂) and sodium silicate were obtained from Sigma-Aldrich (MO, USA). Paraquat dichloride tetrahydrate and diquat dibromide monohydrate were purchased from AccuStandard (CT, USA). Sodium cyanide (NaCN), 4,4′-bipyridine, potassium fluoride (KF), sodium bromide (NaBr), sodium iodide (NaI), sodium iodate (NaIO₃), sodium sulfide (Na₂S), sodium carbonate (Na₂CO₃), sodium bicarbonate (NaHCO₃), potassium thiocyanate (KSCN), sodium sulfate (Na₂SO₄), sodium sulfite (Na₂SO₃), sodium nitrate (NaNO₃), sodium nitrite (NaNO₂), disodium hydrogen phosphate (Na₂HPO₄) hydrochloric acid and sodium hydroxide were obtained from Sinopharm Chemical Reagent Co., Ltd. (Beijing, China). All chemicals were of analytical grade or above and used as received. Saline solution was obtained from Shijiazhuang No. 4 Pharmaceutical Co., Ltd. (Shijiazhuang, Hebei Province, China). All solutions were prepared with 18.2 MΩ ultrapure water, which was generated from a Milli-Q A10 system (Millipore, MA, USA). The 96-well plates were purchased from Corning (NY, USA).

Blank human urine was collected from five healthy volunteers, and then mixed and stored at 4 °C. Blank human plasma, paraquat poisoned plasma and urine samples were provided by the 307 hospital of Chinese People's Liberation Army with a permission of the poisoned patient. Blank human plasma was stored at −80 °C. Paraquat poisoned samples were stored at 4 °C. All procedures were approved by the Ethics Committee of the 307 hospital of Chinese People's Liberation Army.

Caution: Paraquat, diquat and sodium cyanide are extremely toxic agents. Handling of these agents should be carried out in a well-ventilated fume hood with high attention. Gloves and stringent protective measures should be adopted.

2.2 Instruments

Transmission electron microscopic (TEM) measurements were performed in a Hitachi H-7650 instrument (Hitachi Technologies Co., Tokyo, Japan). To clearly observe the ultrathin silica shell outside the Au core, a high-resolution TEM (HR-TEM; Tecnai F30, FEI Co., OR, USA) was also employed. UV-vis spectrometric measurements were performed on a Cary 300 double-beam UV-vis spectrometer (Varian Inc., CA, USA). Zeta potential tests were examined on a Malvern nano-zs90 instrument (Malvern, Worcs, UK). All Raman spectra were collected on a portable Raman spectrometer (i-Raman, B&W TEK Inc., DE, USA) equipped with a video microscope system (BAC151A, B&W TEK Inc., DE, USA). The laser excitation wavelength was 785 nm, and the maximum laser power was 300 mW. The Raman spectra were ranged from 175 to 3200 cm⁻¹ with a resolution of 4.5 cm⁻¹.

2.3 Preparation of SERS substrates

2.3.1 Bare AuNPs substrate. The bare AuNPs with an average diameter of 55 nm were synthesized according to a typical citrate-reduced reaction developed by Frens.¹⁶ Briefly, an aliquot of 100 mL HAuCl₄ (0.01%, w/v) in a round-bottom flask was heated until vigorous boiling under reflux with a continuous stirring, and then 0.7 mL of trisodium citrate (1%, w/v) was rapidly added to the boiling solution. The color of solution was turned to dark violet at first, and then quickly changed to purple red. The solution was then under reflux for another 40 min, and cooled to room temperature with a continuous stirring.

2.3.2 PinSHINs substrate. The pinSHINs were prepared by a chemical growth of an ultrathin layer of silica shell (ca. 1 nm) on the surface of bare AuNPs with an average diameter of 55 nm, according to Li et al.¹⁷ Briefly, the solution of bare AuNPs seed was mixed with 1 mM APTMS solution and stirred for 15 min, then 0.54% sodium silicate solution was added and stirred for another 3 min to coat the ultrathin shell of SiO₂ onto the Au seed. The pH of the sodium silicate solution fitted in a narrow range between 10.2 and 11.0.

2.4 Preparation of samples

Stock solutions of paraquat, diquat and 4,4′-bipyridine were prepared at 1000 mg L⁻¹ in ultrapure water. The working standard solution was then prepared from the stock solution by a serial dilution with ultrapure water. All standard solutions were stored at 4 °C.

The spiked biosamples were prepared by the addition of the standard solution into plasma or urine and then serially diluted, and the final concentrations of paraquat in such biologic fluids were ranged from 0.5 to 200 μg L⁻¹. No pretreatment was needed for either spiked samples or poisoned samples before SERS detection. A serial dilution by saline solution was adopted for the real poisoned samples when the SERS intensity was beyond the linear range.

2.5 SHINERS measurement

For each SHINERS measurement, an aliquot of 1 mL pinSHINs was centrifuged at 5000 rpm for 5 min, and the supernatant was removed. Aliquots of 200 μL plasma or urine samples and 10 μL of 2 mol L⁻¹ KI solution were added to the concentrated pinSHINs, then mixed thoroughly and immediately transferred to a 96-well plate for SHINERS measurements. The Raman spectra were collected with an accumulation time of 10 s, and the laser power was set at 60 mW.

3 Results and discussion

3.1 Characterization of the pinSHINs

The morphology of pinSHINs was characterized by a TEM imaging. As shown in Fig. S1,† the pinSHINs are of spherical shape, with an average size of ca. 55 nm in diameter and an ultrathin shell of ca. 1 nm in thickness. The maximum absorption wavelength of pinSHINs is 535 nm, which indicates a size of about 58 nm of the nanoparticles in diameter.¹⁸

3.2 Iodide-facilitated SERS of paraquat in plasma or urine

The SERS collection of paraquat in water is quite convenient, and we can easily measure paraquat as low as 2 μg L⁻¹ in the case of 100 mM NaCl as an aggregation reagent. The Raman shifts of paraquat in SERS showed similar trends with its previously reported Raman spectrum. Peaks of 839 cm⁻¹, 1183 cm⁻¹, 1293 cm⁻¹ and 1643 cm⁻¹ are the characteristic Raman shifts of paraquat in the SERS mode (Fig. 1A). The band around 839 cm⁻¹ is assigned for the C–N stretching mode, and the sharp strong band around 1643 cm⁻¹ is the C=N stretching mode. The band around 1183 cm⁻¹ is the C=C bending vibration mode, and 1293 cm⁻¹ belongs to the C–C structural distortion mode.¹⁹

Fig. 1 (A) SERS spectra of paraquat in water (a), plasma (b), urine (c) (50 μg L⁻¹) with 100 mM of NaCl; (B) influence of different anions in 100 mM on SERS intensity at 1643 cm⁻¹ of paraquat at concentration of 50 μg L⁻¹ in plasma; (C) SERS spectra of paraquat in water (a), plasma (b), urine (c) (50 μg L⁻¹) with 100 mM of KI.

However, when we performed same measurement towards spiked human plasma or urine samples, we cannot detect paraquat even at a 50 μg L⁻¹ level (Fig. 1A). The interference from abundant salts, proteins and other components existed in the human plasma or urine matrix may be the main reason. Considering that SERS response of molecules is highly related to their distances to the metal surface,^20,21 and that biomolecules such as proteins, nitrogen- and sulfur-containing compounds in plasma and urine are inclined to be adsorbed onto the surface of NPs, as shown as the matrix peaks around 620, 730 and 1350 cm⁻¹ (Fig. 1A), we believed that the paraquat molecules were hampered to reach the surface of NPs by a thick layer of other adsorbed molecules so as to hardly show any efficient SERS response. How could we achieve specific measurement of paraquat in complex biologic fluids?

Up to now, no reports have been published about SERS detection of paraquat in biofluids. In a related pesticide detection field, Clauson et al. employed a liquid–liquid extraction to isolate urea and salts, and achieved a SERS detection of acephate, an organophosphate pesticide in urine.²² The direct and selective detection of analytes in complicated matrix without any pretreatment is still a critical challenge in the on-site detection field. Considering that paraquat is a quaternary ammonium which shows as its cation format in a neutral pH but reveals weaker interaction with Au surface than the matrix interferences, we can introduce a kind of anion, which has a moderate interaction with the surface of Au to directly eliminate interferences, and act as a “bridge” between paraquat and the substrate of pinSHINs, so as to attract paraquat to SERS active sites by electrostatic interaction, and achieve a facile, real-time detection. Herein, we examined different common anions, and found that iodide is the most effective kind (Fig. 1B and S2†). The SERS signals of interferences could be effectively suppressed while that of paraquat was greatly enhanced by iodide adsorption (Fig. 1C). This might be due to the excellent advantage of iodide by coupling its constant electrostatic attraction for positively charged N⁺ quaternary ammonium compounds and its strong adsorption onto the surfaces of Au that expels most interfering substances.^23–25 As some of the anions are basic and S²⁻ is the most basic with a pH of 11.7 when added to the sample at the concentration of 100 mM, we tested the influence of pH and found that it didn't have an obvious effect on the SERS signal of paraquat in the range from 6.0 to 12.0 (Fig. S2A†).

When we replaced KI with NaI or MgI_2, we did not found any obvious signal difference, indicating the most contribution was derived from the I⁻ anion. Among all anions we tested (F⁻, Cl⁻, Br⁻, I⁻, S⁻, IO₃⁻, CN⁻, SCN⁻, SO₄²⁻, SO₃²⁻, NO₃⁻, NO₂⁻, HPO₄²⁻, CO₃²⁻, and HCO₃⁻), we emphasized the effect of S²⁻ and CN⁻ as well as I⁻ due to their well-known affinity with Au. As shown in Fig. 1B, regarding the reference band of 1643 cm⁻¹, S²⁻ showed similar, while CN⁻ showed much weaker effect compared with I⁻. The SERS spectra were listed in Fig. S2.† To explain the different phenomenon, we also examined the zeta potentials (Table 1). Towards all three anions, the pinSHINs@I⁻ or S²⁻ or CN⁻ has similar zeta potentials without paraquat. When the surface was further covered by paraquat, the zeta potentials of pinSHINs@I⁻, S²⁻, CN⁻@paraquat increased by 4–8 mV, following an descending order of pinSHINs@I⁻@paraquat, pinSHINs@S²⁻@paraquat, pinSHINs@CN⁻@paraquat. It is indicated that more paraquat cations were tethered to the layer of I⁻, larger SERS response was revealed. We have also observed same trend when we used bare AuNPs instead of pinSHINs.

Table 1 Zeta potentials of different kind of NPs (n = 3)

	Kind of NPs	NPs	NPs@paraquat	NPs@I^−	NPs@I^−@paraquat	NPs@S^2−	NPs@S^2−@paraquat	NPs@CN^−	NPs@CN^−@paraquat	NPs@I^−@diquat	NPs@I@4,4′-bipyridine
Zeta potential (mV)	pinSHINs	−28	−27	−31	−23	−33	−26	−35	−31	−25	−31
	Bare AuNPs	−38	−33	−37	−21	−37	−28	−33	−32	−24	−35

---

We then investigated the effect of KI concentration on the SERS response. Here we used human plasma containing 50 μg L⁻¹ of paraquat as an example of matrix. When the concentration of KI was below 0.1 mM, the SERS signal of paraquat was overall suppressed by the matrix. As the concentration of KI further increased from 0.1 to 100 mM, the signal of paraquat gradually appeared and increased while the SERS signal of blank plasma gradually decreased, indicating that sufficient amount of I⁻ covered the surface of pinSHINs and the matrix around the NPs were efficiently replaced (Fig. S3†). When the concentration of KI was beyond 100 mM, the pinSHINs were excessively aggregated to a status of agglomeration, leading to a decreased SERS intensity of paraquat. We also observed same trends in urine for the SERS signal versus concentration of KI (Fig. S4†). Herein 100 mM of KI was chosen as an optimized concentration in this work.

3.3 The mechanism of iodide-facilitated SHINERS

As shown in Scheme 1, we proposed here a layer-by-layer attraction and matrix repulsion mechanism of the specific enhancement in SERS signal of paraquat. First layer is a layer of I⁻, which replaced other impurities in the matrix by forming a relative strong Au–I bond.²⁶ The second layer is a layer of paraquat cations, which can be tethered to the first layer of I⁻ through electrostatic interaction so as to get closer to the SERS “hot spots”.

Scheme 1 A layer-by-layer attraction mechanism of iodide-facilitated SERS detection of paraquat.

To confirm this assumption, we further tested two analogues of paraquat, i.e., diquat and 4,4′-bipyridine in our work. Diquat is also a quaternary ammonium herbicide, and 4,4′-bipyridine is a tertiary amine intermediate in the production of paraquat, containing a pair of N-donor sites to the Au acceptors at the surface.²⁷ As shown in Fig. 2A, SERS signals of paraquat and diquat (50 μg L⁻¹) in the presence of ten-fold 4,4′-bipyridine were covered by 4,4′-bipyridine (855 cm⁻¹, 1016 cm⁻¹, 1293 cm⁻¹, 1609 cm⁻¹) in water. However, when 100 mM of KI was added into the mixture, the SERS responses of paraquat and diquat (1383 cm⁻¹, 1524 cm⁻¹, 1570 cm⁻¹)²⁸ were greatly enhanced while that of 4,4′-bipyridine almost disappeared, and paraquat and diquat can then be discriminated very well. In Fig. 2B, with the addition of KI, the SERS signals of paraquat or diquat can be greatly enhanced in plasma and urine while that of 4,4′-bipyridine could not, indicating that the electrostatic interaction between I⁻ and positively charged paraquat or diquat is great helpful.

Fig. 2 (A) SERS spectra of paraquat (a), diquat (b), 4,4′-bipyridine (c) and the mixture of all three compounds (d) in water with 100 mM of NaCl; SERS spectra of the mixture of all three compounds with 100 mM of KI (e); (B) SERS spectra of the mixture of paraquat, diquat and 4,4′-bipyridine in plasma and urine with (a and c) and without (b and d) 100 mM of KI. The concentration of paraquat or diquat is 50 μg L⁻¹, and the concentration of 4,4′-bipyridine is 500 μg L⁻¹. (○) represents the characteristic peaks of paraquat, and (△) represents the characteristic peaks of diquat.

To further illustrate the crucial role of such a layer-by-layer attraction, we performed zeta potential tests towards pinSHINs@I⁻@paraquat or the other compound like diquat or 4,4′-bipyridine. As shown in Table 1, compared with 4,4′-bipyridine, only remarkable increases of the original negative zeta potential occurred when quaternary ammonium paraquat and diquat were added to pinSHINs@I⁻ (from −31 mV to −23 mV for paraquat, and to −25 mV for diquat). It indicates that the layer of the cationic paraquat and diquat could be tethered to the layer of I⁻, while the neutral 4,4′-bipyridine could not. Same trends were observed in the case of bare AuNPs at a subtle differed values, indicating that the layer of I⁻ is tightly adsorbed onto the surface of bare AuNPs, while the ultrathin silica shell did not disturb the Au–I interaction but shield some adsorbed amount of I⁻. Furthermore, from the UV-vis spectra of different NPs presented in Fig. 3, we observed a new peak at 762 nm when pinSHINs@I⁻@paraquat formed, indicating that the interaction between I⁻ and paraquat could induce the aggregation of pinSHINs. All the results indicated that electrostatic interaction between I⁻ and paraquat (as well as diquat) played a dominant role in enhancing their SERS responses when we used I⁻ to replace the matrix from Au surface.

Fig. 3 UV-vis spectra of pinSHINs, pinSHINs@diquat, pinSHINs@I⁻, pinSHINs@4,4′-bipyridine, pinSHINs@Cl⁻@diquat, pinSHINs@I⁻@4,4′-bipyridine, pinSHINs@paraquat, pinSHINs@Cl⁻@paraquat, pinSHINs@I⁻@diquat and pinSHINs@I⁻@paraquat (a–j).

3.4 Measurement of paraquat and the stability of the pinSHINs substrate

Our established method is highly sensitive with a LOD as low as 1 μg L⁻¹ in plasma or urine, which is far below the concentration of paraquat in plasma and urine of poisoned patients.^29,30 We plotted the SERS intensity of the band at 1643 cm⁻¹ versus the concentration of paraquat in plasma or urine, and obtained a linear relationship for paraquat from 2 to 40 μg L⁻¹ (R² = 0.989), and 1 to 20 μg L⁻¹ in urine (R² = 0.995; Fig. 4) with the LODs as low as 1 μg L⁻¹. Good recoveries were also found from 93% to 115% at low, middle and high levels (Table S1†). We performed every measurement in only 10 s, and completed the whole procedure within 1 min on a portable Raman spectrometer. These results demonstrate that this method can be readily used in on-site detection of paraquat in poisoned biologic fluids.

Fig. 4 (A) SERS spectra of paraquat in plasma at concentration from 0 to 100 μg L⁻¹ with 100 mM of KI; (B) calibration curve of SERS intensity at 1643 cm⁻¹ versus concentration of paraquat in plasma from 2 to 40 μg L⁻¹; (C) SERS spectra of paraquat in urine at concentration from 0 to 50 μg L⁻¹ with 100 mM of KI; (D) calibration curve of SERS intensity at 1643 cm⁻¹ versus concentration of paraquat in urine from 1 to 20 μg L⁻¹.

For SERS measurement in practical application, the stability of the substrate is an important issue. In a time-dependent experiment, the plasma or urine samples containing 50 μg L⁻¹ of paraquat were added into the pinSHINs, maintained at room temperature for a certain time, and measured by SERS. The pinSHINs substrate has a quite good stability that the SERS signal of paraquat was steadily hold at least 12 hours in plasma and 20 minutes in urine (Fig. S5A and B†). For comparison, we used bare AuNPs as a substrate instead of pinSHINs, although a slightly improved sensitivity at 0.5 μg L⁻¹ was provided, we found a worse stability of bare AuNPs in biologic fluids. The SERS signal of paraquat on bare AuNPs decreased rapidly in less than 5 min in urine, and decreased more than half after 6 h in plasma (Fig. S5C and D†). We contribute the prolonged detection window on pinSHINs to the protection from the ultrathin silica shell around Au core.³¹

3.5 Measurement of paraquat in real samples

We examined the feasibility of this established SHINERS method on plasma and urine samples from one patient who has similar but not so severe paraquat poisoning symptoms, but the patient denied the paraquat oral administration history. The suspected paraquat-containing samples were collected in the 15^th day of physical discomfort. As shown in Fig. 5, a clear evidence of paraquat poisoning was revealed (about 0.5 mg L⁻¹ of paraquat in plasma and 4.2 mg L⁻¹ of paraquat in urine). It is suggested that our method can be readily applied to on-site detection of paraquat poisoning for clinical diagnosis.

Fig. 5 SERS spectra of the urine (a) and plasma (b) of the paraquat poisoned patient and of saline solution (c). Both urine and plasma samples were diluted 100 times by saline solution.

Other various complicated matrices including plasma and bronchoalveolar lavage fluid (BALF) of poisoned rats, spiked milk and orange juice samples were also tested and all good results were obtained (see Fig. S7, S8 and Table S2†), showing its general and wide applicability.

4 Conclusion

We have developed a feasible iodide-facilitated SHINERS technique for the rapid, sensitive and selective detection of paraquat in biologic fluids for the first time with excellent stability of the pinSHINs substrate. Based on the contribution of iodide as an efficient cleaning reagent, and the characteristics of cationic quaternary ammonium paraquat, we proposed and proved a layer-by-layer attraction mechanism. We have readily obtained quantitative results of paraquat in plasma or urine with a LOD as low as 1 μg L⁻¹, and also achieved a simultaneous detection of paraquat and structural-alike diquat. This rapid on-site detection method on paraquat is also applied to the real poisoned samples and shows excellent sensitivity and reliability. We can complete all procedures including sample pretreatment and measurement within 1 min on a portable Raman spectrometer, which reveals a great promise in on-site detection of paraquat as the first response on poisoning events.

Acknowledgements

This research was supported by the National Major Instrument Project of the Ministry of Science and Technology of China (No. 2011YQ03012411) and the National Natural Science Foundation of China (No. 81302473). The real poisoned samples were kindly provided by the 307 hospital of Chinese People's Liberation Army.

Notes and references

1. Y. Yin, X. Guo, S. L. Zhang and C. Y. Sun, Biomed. Environ. Sci., 2013, 26, 509–512.
2. R. J. Dinis-Oliveira, J. A. Duarte, A. Sanchez-Navarro, F. Remiao, M. L. Bastos and F. Carvalho, Crit. Rev. Toxicol., 2008, 38, 13–71.
3. J. L. Lin, D. T. Lin-Tan, K. H. Chen and W. H. Huang, Crit. Care Med., 2006, 34, 368–373.
4. P. Paixão, P. Costa, T. Bugalho, C. Fidalgo and L. M. Pereira, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2002, 775, 109–113.
5. Y. Zou, Y. Shi, Y. Bai, J. Tang, Y. Chen and L. Wang, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2011, 879, 1809–1812.
6. R. M. de Almeida and M. Yonamine, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2007, 853, 260–264.
7. K. Wunnapuk, G. A. Medley, X. Liu, J. E. Grice, S. Jayasinghe, I. Gawarammana, N. A. Buckley and M. S. Roberts, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2011, 879, 3047–3052.
8. M. E. Koivunen, S. J. Gee, E. K. Park, K. Lee, M. B. Schenker and B. D. Hammock, Arch. Environ. Contam. Toxicol., 2005, 48, 184–190.
9. J.-R. Koo, J.-W. Yoon, S.-J. Han, M.-J. Choi, I.-I. Park, Y.-K. Lee, S.-G. Kim, J.-E. Oh, J.-W. Seo, H.-J. Kim and J.-W. Noh, Am. J. Med. Sci., 2009, 338, 373–377.
10. Q. X. Zhu, Y. B. Cao, Y. Y. Cao, Y. F. Chai and F. Lu, Anal. Bioanal. Chem., 2014, 406, 1877–1884.
11. J. Parisi, Q. Dong and Y. Lei, RSC Adv., 2015, 5, 14081–14089.
12. S. M. Nie and S. R. Emery, Science, 1997, 275, 1102–1106.
13. R. Gao, N. Choi, S.-I. Chang, S. H. Kang, J. M. Song, S. I. Cho, D. W. Lim and J. Choo, Anal. Chim. Acta, 2010, 681, 87–91.
14. H. R. Tang, Q. Q. Li, Y. L. Ren, J. P. Geng, P. Cao, T. Sui, X. Wang and Y. P. Du, Chin. Chem. Lett., 2011, 22, 1477–1480.
15. H. J. Butler, S. W. Fogarty, J. G. Kerns, P. L. Martin-Hirsch, N. J. Fullwood and F. L. Martin, Analyst, 2015, 140, 3090–3097.
16. G. Frens, Nature, 1973, 241, 20–22.
17. J. F. Li, X. D. Tian, S. B. Li, J. R. Anema, Z. L. Yang, Y. Ding, Y. F. Wu, Y. M. Zeng, Q. Z. Chen and B. Ren, Nat. Protoc., 2013, 8, 52–65.
18. W. Haiss, N. T. K. Thanh, J. Aveyard and D. G. Fernig, Anal. Chem., 2007, 79, 4215–4221.
19. H. Fang, X. Zhang, S. J. Zhang, L. Liu, Y. M. Zhao and H. J. Xu, Sens. Actuators, B, 2015, 213, 452–456.
20. B. N. J. Persson, K. Zhao and Z. Zhang, Phys. Rev. Lett., 2006, 96, 207401.
21. S. Lal, N. K. Grady, G. P. Goodrich and N. J. Halas, Nano Lett., 2006, 6, 2338–2343.
22. S. L. Clauson, J. M. Sylvia, T. A. Arcury, P. Summers and K. M. Spencer, Appl. Spectrosc., 2015, 69, 785–793.
23. A. Chaumont and G. Wipff, J. Comput. Chem., 2002, 23, 1532–1543.
24. J.-Y. Huang, C. Zong, L.-J. Xu, Y. Cui and B. Ren, Chem. Commun., 2011, 47, 5738–5740.
25. L.-J. Xu, C. Zong, X.-S. Zheng, P. Hu, J.-M. Feng and B. Ren, Anal. Chem., 2014, 86, 2238–2245.
26. P. Gao and M. J. Weaver, J. Phys. Chem., 1986, 90, 4057–4063.
27. T. C. Strekas and P. S. Diamandopoulos, J. Phys. Chem., 1990, 94, 1986–1991.
28. M. R. Lopez-Ramirez, L. Guerrini, J. V. Garcia-Ramos and S. Sanchez-Cortes, Vib. Spectrosc., 2008, 48, 58–64.
29. H.-W. Gil, M.-S. Kang, J.-O. Yang, E.-Y. Lee and S.-Y. Hong, Clin. Toxicol., 2008, 46, 515–518.
30. J. M. Scherrmann, P. Houze, C. Bismuth and R. Bourdon, Hum. Exp. Toxicol., 1987, 6, 91–93.
31. J. F. Li, Y. F. Huang, Y. Ding, Z. L. Yang, S. B. Li, X. S. Zhou, F. R. Fan, W. Zhang, Z. Y. Zhou and B. Ren, Nature, 2010, 464, 392–395.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra06954a

---