Rapid and sensitive detection of uranyl ion with citrate-stabilized silver nanoparticles by the surface-enhanced Raman scattering technique

Abstract

Uranium contamination poses a huge threat to human health due to its widespread use in the nuclear industry and weapons. We proposed a simple and convenient wet-state SERS method for uranyl detection based on the citrate-stabilized silver nanoparticles. The effect of citrate on the detection performance was also discussed. By using the citrate as an internal reference to normalize the peak of uranyl, a quantitative analysis was achieved and a good linear relationship of uranyl concentration from 0.2 to 5 µM with the limit of detection of 60 nM was obtained. With its simplicity, convenience and cost-effectiveness, this method has great potential for the detection of other molecules also.

1. Introduction

Uranium (with a high density of 19.2 g cm⁻³), as one of the most important nuclear sources, has shown great application in the nuclear industry and weapons [1,2]. The widespread use of uranium produces a large amount of nuclear wastes which may migrate to the groundwater, bringing long-term health damage to human beings due to its radioactive properties and chemical toxicity [3,4]. Trace amounts of uranium have been found in plants, animals and even in human urine. Hence, real-time monitoring of uranium concentration in ecosystems, environmental waters, and even in the human body is of significance—particularly for nuclear workers.

Uranium exists in the environment usually with two redox states: U(VI) and U(IV). U(IV) is relatively insoluble and can be easily precipitated as uranium dioxide (UO₂) or can be oxidized to U(VI) [5]. The most common and thermodynamically stable form of U(VI) is uranyl ion (UO₂²⁺). It is soluble and mobile, which may harm body organs easily and attract more attention by analytical chemists. Traditional instrument-based methods, including mass spectrometry (MS), high-performance liquid chromatography (HPLC), inductively coupled plasma atomic emission spectroscopy (ICP-AES) and X-ray fluorescence spectroscopy [6–8], cannot detect nuclear materials easily. There is a dire need for the development of a simple and portable method for onsite and real-time monitoring of uranium.

Surface-enhanced Raman scattering (SERS) technique [9–13] has the advantages of simplicity, convenience, high sensitivity, non-requirement of pretreatment of analysed sample and has great potential in the rapid and sensitive detection of inorganic ions, organic molecules, even biomacromolecules [14–17]. Many efforts have been made for SERS detection of uranyl ion by fabricating different types of SERS substrates, such as silver-doped sol–gel [18], gold nanostars [19], Al₂O₃-coated silver nanorod [20], silver nanoparticles conjugated reduced graphene oxide nanosheets [21] and so on. SERS measurements of those based on ordered nanostructure substrates (dry-state substrate) fabricated by self-assembly or vapour deposition methods usually need drying the sample onto the substrate, which is difficult to obtain quantitative detection due to the coffee-ring effect and substrate-self defect. What is more, the fabrication process of the dry-state substrate is complex and expensive, which shows no commercial competitive advantages.

In this study, we reported a very simple and convenient wet-state method to detect uranyl ion (UO₂²⁺) by silver colloid by adjusting the amount of citrate to get good sensitivity. Citrate (Cit) has three functions: (1) as the stabilizer to stabilize silver nanoparticles; (2) as the complexing agent to capture uranyl ion through chelation between COO⁻ and UO₂²⁺ (scheme 1) [22] and (3) as the internal molecule to normalize the Raman signal of uranyl ion to eliminate the adverse effects of the matrix and other external factors. This method is user-friendly and inexpensive and can be achieved in any normal laboratories by any manipulators. What is more, the entire sample fabrication and detection process only needs several seconds if silver colloid had been synthesized previously, which makes it possible for onsite and real-time UO₂²⁺ monitoring in the environment.

Scheme 1.

Schematic of the citrate-stabilized AgNPs for SERS detection of UO₂²⁺.

2. Material and methods

2.1. Chemicals and materials

Uranyl nitrate hexahydrate was obtained from the China National Nuclear Corporation (Lanzhou, China). It was dissolved in ultrapure water to make a 10⁻² M stock solution, and then diluted to the final concentration before use. Trisodium citrate dihydrate (99%) and silver nitrate (AgNO₃, ≥99.8%) were purchased from Aladdin. All reagents were used without further purification. Ultrapure water with a resistivity of approximately 18.0 MΩ cm was used throughout the experiments.

2.2. Synthesis of silver colloid

The silver colloid was synthesized by the citrate-reduced method according to Lee & Meisel report [23]. In 100 ml of H₂O, 18 mg of AgNO₃ was dissolved and then heated to boiling. Two millilitres of 38.8 mM sodium citrate was added with fast stirring. The mixture was kept boiling for about 30 min and was allowed to cool at room temperature under stirring.

2.3. SERS sample fabrication

In a typical fabrication process, 8 ml of freshly prepared silver colloid was added into the tube and centrifuged at a speed of 6000 r.p.m. Then 2 ml of water was added to disperse silver nanoparticles (AgNPs) by ultrasonication. The fourfold concentrated AgNP colloid was obtained and used for SERS detection. Citrate cannot be removed completely by this method due to its strong absorption ability on the silver surface. After centrifugation is done, the retained citrate can be absorbed on the surface of AgNPs to protect them from aggregation. For a SERS sample, 10 µl of standard uranyl concentration aqueous solution was added into 90 µl of condensed silver colloid, and the value of Raman spectrum was collected immediately. To characterize the effect of citrate on the intensity of uranyl peak, four silver colloid samples, containing different amounts of citrate, were fabricated and were denoted as A, B, C and D. Eight millilitres of silver colloid was centrifuged and condensed four times by adding: (A) 2 ml of 0.02 wt% trisodium citrate aqueous solution, (B) 2 ml of 0.01 wt% trisodium citrate, (C) 2 ml of water and (D) the colloid was centrifuged and washed with water twice and condensed by adding 2 ml of water. Here, the amount of citrate in sample D with twice centrifugation is lower than that of sample C.

2.4. Measurements

The morphology of silver nanoparticles was characterized by a scanning electron microscope (SEM, JEOL-JMS-7001F), the absorption spectra were obtained by a UV–vis spectrometer (SHIMADZU, UV-1800). SERS measurements were conducted on a LabRam Xplora confocal Raman spectrometer (Horiba Jobin Yvon). The laser with the excitation wavelength of 532 nm was used in the experiment, and the laser power is approximately 2.5 mW with the collection time of 10 s. The Raman scattering signal was collected with a numerical aperture (NA) microscopic objective from Olympus (50×, NA = 0.5). For quantitative detection of uranyl ion, five different positions were randomly selected and the SERS intensity was obtained by averaging the relative Raman signals.

3. Results

Compared with gold, AgNPs are superior in terms of SERS performance and cost-effectiveness [24]. With the fast development of nanotechnique, although the shape of the nanostructure is abundant (nanoplates, nanorods, nanoflowers, nanosatellites, etc.) and can be precisely controlled [25–28], nanosphere is the most easily synthesized and also the most stabilized structure. Seen from the point of ease and convenience, silver nanosphere structure (figure 1a) synthesized by citrate-reduced method was chosen as the SERS substrate here [23,29]. The low-magnification SEM image (see electronic supplementary material, figure S1) shows that the average size of AgNPs with rather narrow size distribution is approximately 60 nm, which guarantees high sensitivity, because larger-sized nanoparticles can produce a more sensitive signal. The aggregates may be produced from the nanoparticle aggregation when the colloid was drying. Excess citrate can be adsorbed on the surface of silver nanoparticle to form an electric double layer to protect silver from aggregation. However, one important point that many researchers may neglect in their SERS studies is the Raman signal interference of citrate. Figure 1b shows the Raman spectrum of solid trisodium citrate; there are five peaks with strong scattering intensities at 848, 957, 1448, 2928, 2969 cm⁻¹. Seen from the SERS of silver colloid (figure 1c), the three C-COO stretching modes (μ(C-COO)) of Cit exhibit strong Raman signal at 930 cm⁻¹ [30], which moves towards lower wavenumbers compared with that of solid citrate at 957 cm⁻¹ (figure 1b) due to the charge transfer between Cit and Ag. The high background signal will cause an adverse effect on molecule detection, especially during trace analysis. However, after centrifugation, the background signal of Cit decreased markedly (figure 1c), thus the amount of citrate in silver colloid for SERS analysis is of significance.

Figure 1.

(a) SEM image of AgNPs; (b) Raman spectrum of sodium citrate crystal; (c) Raman spectra of Ag colloid before and after centrifugation; (d) SERS of 10 µM UO₂²⁺ under the condition of A–D. 8 ml of silver colloid was centrifuged and condensed four times by adding: (A) 2 ml of 0.02 wt% trisodium citrate aqueous solution, (B) 2 ml of 0.01 wt% trisodium citrate, (C) 2 ml of water and (D) the colloid was centrifuged and washed with water twice and condensed by adding 2 ml of water.

To understand the effect of Cit on the SERS of UO₂²⁺, a set of control experiments were performed by adjusting the amount of Cit in silver colloid. Four samples were designed: 8 ml of silver colloid was centrifuged and condensed four times by adding: (A) 2 ml of 0.02 wt% trisodium citrate aqueous solution, (B) 2 ml of 0.01 wt% trisodium citrate, (C) 2 ml of water and (D) the colloid was centrifuged and washed with water for two times and condensed by adding 2 ml of water. Centrifugation wash of silver nanoparticles cannot remove citrate fully, thus the amount of citrate retained in colloid in C is larger than that in D. There is a small change in the λ_SPR of AgNPs from samples A to C (see electronic supplementary material, figure S2) after adding 10 µM of UO₂²⁺, showing little aggregation of AgNPs. However, the intensity of λ_SPR in D (see electronic supplementary material, figure S1) decreased remarkably and a new peak at 700 nm appeared, and the silver colloid changed from grey green to dark simultaneously, indicating the severe aggregation of AgNPs. The aggregation of AgNPs brings difficulty for quantitative analysis of uranyl ion. Figure 1d shows the SERS spectra of 10 µM of UO₂²⁺. The relatively broad and asymmetric peak near 750 cm⁻¹ is attributed to the υ₁ symmetric stretch of the uranyl ion of uranyl complexes. The ligands have a great effect on the Raman shift of O=U=O [31,32]. Uranyl citrate forms trimeric species with the 3 : 3 and 3 : 2 U : Cit complexes at the near-neutral pH region, exhibiting a complicated complex interaction [22]. The SERS band of uranyl is lower (750 cm⁻¹) than uranyl citrate Raman band (832 cm⁻¹) (see electronic supplementary material, figure S3), indicating strong coordination between uranyl and silver along with Cit. The strong chemical interactions between uranyl ion and ligands weaken the axial U=O band intensity due to the extensive electron density transfer from citrate and silver surface to the equatorial plane of absorbed uranyl ions, causing the uranyl band to shift from 870 cm⁻¹ to a lower wavenumber, which has been reported in the previous literature [33–35]. Sample C obtains a good SERS signal of uranyl (figure 1d). The intensity of uranyl band decreases with the increase of citrate (figure 1d, A to C). The increase of citrate will protect AgNPs from aggregation and widen the distance between adjacent AgNPs, and thus decrease the Raman intensity. In addition, the excess citrate free in solution will coordinate with the UO₂²⁺, preventing UO₂²⁺ from close to AgNPs, which has no contribution to the intensity of the uranyl band (Samples A and B). Proper aggregation can minimize the distance between adjacent AgNPs to create a ‘hot spot’ (sample C), but overage aggregation causes the sedimentation of silver, which weakens the signal of uranyl and brings high background signal, for instance, in the circumstance of sample D (figure 1d, C cf. D). So sample C is an appropriate choice for uranyl detection.

To verify the analytical performance of the proposed wet-state method for uranyl ion, different uranyl ion concentration was determined by adding standard UO₂²⁺ in sample C. The SERS signal of uranyl species tended to be stable through initial fluctuation, as shown in figure 2, making quantitative detection possible. The environmental factors (temperature, humidity, etc.), the instrumental conditions (laser intensity, focal distance, etc.), sample distinction (colloid concentration, sample cup size, etc.) and human factor will affect the measurement result, which make quantitative detection difficult (see electronic supplementary material, figure S4) [36]. To eliminate the effects of external factors, here we established an internal reference with the peak of the citrate centred at 930 cm⁻¹ to normalize the peak of uranyl centred at 750 cm⁻¹. Figure 3a shows the SERS spectra of uranyl ion with the concentration from 0.1 to 5 µM. After calibration and normalization by citrate, a good linear relationship (R² = 0.996) can be obtained between the relative Raman intensity (I_uranyl/I_citrate) and the uranyl concentration from 0.2 to 5 µM (figure 3b). The limit of detection (LOD) (60 nM) was calculated by three standard deviations of the blank measurements, showing a good sensitivity. Table 1 lists the comprehensive comparison results of the proposed wet-state method with others reported in the literature [20,34,37–39]. The LOD of our proposed SERS method is higher than those obtained based on DNAzyme or complex nanostructures, while the advantages such as simplicity, rapidness and convenience make it a forceful competitor in real-time environmental monitoring. What is more, the LOD of the proposed strategy is still better than those of photometry and SERS based on modified gold nanoparticles. Although it is difficult to detect the 60 nM of UO₂²⁺ in the experiment due to the intrinsic spectral background of the blank silver colloid, 200 nM was easily detected in this study.

Figure 2.

SERS of 5 µM uranyl with five times of continuous measurement under the condition of sample C.

Figure 3.

(a) SERS of uranyl ion with a concentration from 0.1 to 5 µM. (b) The calibration curve of uranyl ion concentration with its relative SERS intensity (I_uranyl/I_citrate) by citrate as the internal reference.

Table 1.

An overview for uranyl ion detection.

materials used	method applied	LOD	refs
DNAzyme, AuNPs	colorimetry	50 nM	[37]
gold nanoparticles	photometry	500 nM	[38]
gold nanoparticles	SERS	800 nM	[34]
silver nanorod	SERS	1 nM	[20]
DNAzyme, hairpins	electrochemistry	2 pM	[39]
condensed AgNPs	SERS	60 nM	this paper

To further evaluate the selectivity of the proposed wet-state SERS strategy for uranyl ion detection, a series of contrast experiments were conducted, including Na⁺, Zn²⁺, Ca²⁺, Cu²⁺, Fe²⁺, Ni²⁺ and Cd²⁺, as shown in figure 4. These impure metal ions did not lead to the signal enhancement. At the same time, the presence of other competitive metal ions had no effect on the relative Raman intensity of uranyl ion (see electronic supplementary material, figure S5), indicating a good sensitivity. The reproducibility of the proposed strategy was evaluated by six repetitive measurements for 1 µM UO₂²⁺, and the intra-assay relative standard deviation (RSD) was 10% estimated. In addition, the RSD of the inter-assay for UO₂²⁺ was determined (8.7%) by using five batches of different Ag nanoparticle colloids. These results showed a good selectivity and reproducibility of the proposed method for UO₂²⁺ detection.

Figure 4.

The selectivity of the proposed strategy for UO₂²⁺: 5 µM of UO₂²⁺; 500 µM of Na⁺, Zn²⁺, Ca²⁺, Cu²⁺, Fe²⁺, Ni²⁺ and Cd²⁺.

To investigate the validity of the proposed method to environmental sample, a known amount of uranyl ion was added into the tap water sample. These three spiked samples with the uranyl ion concentration of 5, 2 and 1 µM were measured, as shown in table 2. The recoveries and RSDs of three spiked uranyl ion samples in real tap water are from 86% to 105% and 7% to 10%, respectively. These results indicate that our proposed method has great potential for rapid detection of uranyl ion in environmental samples.

Table 2.

Determination of UO₂²⁺ in tap water with citrate-stabilized silver nanoparticles.

sample	add (μM)	found (μM)	recovery (%)	RSD (%)
1	5	5.25	105	10.1
2	2	1.85	92.5	8.22
3	1	0.86	86.0	7.23

4. Conclusion

We reported a very simple SERS method for rapid detection of uranyl ion based on the citrate-stabilized silver colloid. Owing to the strong interaction between uranyl and citrate as well as silver, uranyl can be captured into the ‘hot spot’ zone, and uranyl ion as low as 200 nM was easily detected by the proposed method, showing an acceptable sensitivity. A good linear plot between the relative intensity and the uranyl concentration was attained by using the residual citrate as the internal reference to normalize the peak of uranyl. Our proposed strategy shows great potential for uranyl real-time and fast analysis in the environment. Furthermore, the internal reference strategy adopted in this paper will push SERS quantitative detection to be more veracious and reliable.

Data accessibility

The data supporting this article have been uploaded as part of the electronic supplementary material.

Authors' contributions

J.J. carried out the SERS measurement and data analysis, and drafted the manuscript. S.W. and H.D. carried out the silver nanoparticle synthesis and characterization. H.W. helped to achieve partial SERS measurements, J.C. and J.L. proposed the above idea and helped draft the manuscript. All the authors gave their final approval for publication.

Competing interests

We declare we have no competing interests.

Funding

This work was financially supported by China Academy of Engineering Physics (TCSQ2016203, no. XK909-2), and the Natural Science Foundation of China (no. 21501157).