Sodium saccharin in soft drinks analysis
Sodium saccharin is a generally used synthetic sweetener in soft drinks, beverages and so on. Long-term consumption of sodium saccharin will result in all kinds of health problems. A sensitive surface-enhanced Raman spectroscopy (SERS) method with silver nanorod (AgNR) array substrates was explored for quantifica-tion of sodium saccharin in soft drinks. The AgNR array substrates fabricated by an oblique angle deposition technique exhibited an super SERS activity with an enhancement factor of 108. A good linear relationship was demonstrated between SERS peak intensity and Sodium saccharin concentration in the concentration range of 0.5–100 mg/l, and a limit of detection was determined to be 0.3 mg/l. The detection of sodium saccharin in different soft drinks was further explored by combining SERS spectra with the partial least squares discriminant analysis. This method enables the quick and highly sensitive detection of SS with minimal sample pretreatment, and holds great promise in food safety application.
Introduction of sodium saccharin
Sodium saccharin has a long history of use in food processing, feed production, and the com-modity industry, and has been approved in more than 90 countries. Specifically, it is often added into sauces, soft drinks, candy, and other sweet foods to improve the flavor and reduce calories. The acceptable daily intake (ADI) values of sodium saccharin determined by the Joint FAO/WHO Expert Committee on Food Additives are 0–5 mg/kg body mass. However, long-term, excessive consumption of sodium saccharin is closely associated with cancer risk in humans by cancer epidemi-ological studies. It is necessary to detect and quantify the content of sodium saccharin in food.
The most widely used methods for the determination of sodium saccharin are high performance liquid chromatography (HPLC), thin layer chromatography (TLC), and enzyme-linked immunosorbent assays (ELISA).These methods are reliable and repro-ducible; however, they suffer from specific disadvantages with regards to quantification. For example, the HPLC method is time-consuming and the equipment is expensive. The TLC method needs significant sample pretreatment, and qualitative or quanti-tative detection is difficult without combining with other methods. The ELISA method is susceptible to environmental changes, which limits the stability of real-life application. Surface-enhanced Raman spectroscopy (SERS) is emerging as a powerful technique for detect-ing a wide variety of analytes at very low concentrations. The most critical aspect to perform a SERS experiment is the choice and/or fabrication of the noble-metal substrates because the sensitivity of SERS strongly depends on the electromagnetic and chemical enhancement in the presence of metallic (Cu, Ag, Au, etc.) nanostructures . Thus, to further widen the appli-cations of the SERS technique, it is crucial to fabricate SERS-active metallic nanostructures with both high enhancement and excel-lent reproducibility. For this purpose, many different methods, such as e-beam lithography, colloidal lithography, and vapor deposition have been explored extensively to cre-ate functional noble metal SERS substrates. While these substrates have been shown to exhibit large Raman enhancement and good SERS reproducibility, most are, unfortunately, too laborious and expensive to fabricate in large sizes and quantity. This limits their use in a wide variety of fields, especially in routine laboratory anal-ysis and point-of-care sample analysis in the field.
Oblique angle deposition (OAD) has been rec-ognized as a stable and effective method for the fabrication of nanostructures in recent years. The SERS substrates fabricated by OAD exhibit good sensitivity and high reproducibility. Driskell et al. studied the reproducibility of AgNR array substrates fabri-cated by OAD. They showed that when SERS measurements were collected over multiple locations on the same substrate, the resulting data possessed a relative standard deviation (RSD) of 10%. Additionally, the RSD measured between different substrates was 15%. Moreover, the substrates fabricated by OAD have already been applied to many applications such as environmental moni-toring, food safety and biological testing. For example, Han et al. achieved the sensitive detection of metronidazole and ronidazole from environmental samples by using these AgNR array substrates. Kumar et al. demonstrated a flexible AgNR substrate for rapid detection of thiram pesticide from fruit peels. Wu et al. reported the indirect detection of Pseudomonas aeruginosa on AgNR array substrates by quantitative detection of pyocyanin (PCN) as the biomarker. However, there is no report about the use of AgNR array substrates for the detection of SS.
Herein, we used the highly reproducible AgNR substrates fab-ricated by OAD to detect SS in commercial soft drinks. A density functional theory (DFT) calculation was first performed to under-stand the characteristic peaks and vibrational modes of SS, and then used to determine its corresponding SERS peaks. The SERS inten-sity versus SS concentration was systematically determined and the LOD is obtained. Moreover, a rapid and simple sample pretreatment method was applied to separate SS from soft drinks, and the partial least square discriminant analysis (PLS-DA) was used to analyze and confirm the detection of SS in real soft drink samples.
Materials
Silver (99.999%) and titanium (99.995%) pellets were purchased from Kurt J. Lesker Co., Ltd. (USA). Sodium saccharin (≥99.9%), ethyl acetate (≥99.5%) and ethanol (≥99.7%) were acquired from Sinopharm Chemical Reagent Co., Ltd. (China). Soft drinks were pur-chased from a local supermarket. Ultra-pure water (≥18.2 M ) was used in all experiments.
Fabrication of AgNR array substrates
AgNR substrates were fabricated in a custom-built e-beam deposition system (DE500, DE Technology Inc., Beijing, China). Briefly, glass slides were cut into 1 × 1 cm2 pieces, sonicated in ethanol for 5 min by an ultrasonic cleaning machine at least 3 times, dried with N2, and loaded into the deposition chamber. Depositions were performed once the chamber reached a pressure of 5 × 10−7 Torr. First, a 20 nm film of titanium, and then 100 nm of silver film was deposited onto the glass substrates at rates of 0.2 nm/s and 0.3 nm/s, respectively. These films were deposited at an incident angle = 0◦ relative to the normal of the substrate. Next, the substrate was rotated to an oblique angle of = 86◦ with respect to the vapor incident direction (Fig. 1B), and 2000 nm of silver was deposited at a rate of 0.3 nm/s. The deposition thickness and rate were monitored by a quartz crystal microbalance (QCM) facing directly toward the vapor direction. The substrate surface morphology was examined by a field-emission scanning electron microscope (SEM, SU8010, Hitachi, Tokyo, Japan).
Sample preparation
Different concentrations of SS (CSS = 0.5, 1, 10, 20, 50, 80, 100 mg/L) were prepared using ultrapure water to determine the SERS intensity versus CSS calibration curve and the LOD. In addition, SS was added into the four different soft drinks: Sprite, Cola, Fanta, and Schweppes. SS concentrations of 5, 10, 20, 30, 40, 50, 60, 80, and 100 mg/L were achieved for each soft drink. Before SERS detec-tion, all of the samples were heated for 5 min to release CO2. Then, each sample was mixed with ethyl acetate with a ratio of 1:1 (v/v). The mixture was stirred by a vortex mixer and allowed to separate into layers for 1 min. After stratification, the upper layer liquid was used for SERS detection (Fig. 1C).
SERS measurement
In order to confirm the SERS enhancement of the substrate, 2 L of different concentrations of BPE solution (CBPE = 10−5, 10−6, 10−7, 10−8, 10−9 and 10−10 M, dissolved in methanol) were applied to the AgNR substrate surface, and allowed to dry under ambient con-ditions. 1 mL of 10−2 M BPE (dissolved in methanol) solution was added to a cuvette. Then SERS and Raman spectrum was taken from the substrate with a laser power of 100 mW and an integration time of 10 s. The Raman and SERS intensity at = 1200 cm−1 was selected to assess the enhancement factor (EF) of the substrates. For real sample detection, 2 L drop of soft drink sample (spiked with different concentrations of SS) was placed on the surface of the AgNR substrate, and allowed to dry in air. Then SERS spectra were taken from the substrate with a laser power of 30 mW and an integration time of 10 s. At least nine random sampling loca-tions were measured from each drop. All SERS measurements were performed using a Raman analyzer (ProRaman-L-785A2, Enwave Optronics, Irvine, CA) equipped with a 785 nm diode laser.
Data analysis
All of the SERS spectra were plotted with Origin 8.5 soft-ware (Origin Lab, Northampton, MA). The raw spectra obtained from the Raman analyzer were used without further processing unless otherwise specified. Spectra analysis, including subtracting the baseline and fitting peaks, were performed using GRAMS/AI spectroscopy software suite (Thermo Fisher Scientific, Waltham, MA). Statistical data analysis was conducted with Matlab 2011b (Mathworks, Denver, CO) using the PLS-Tool-box (Eigenvector, Wenatchee, WA).
Density function theory (DFT) calculation
The Gaussian 09 W DFT software package was used to cal-culate the Raman spectra of SS and identify the corresponding vibrational modes. The DFT calculations were based on Becke’s three-parameter exchange function (B3) with the dynamic cor-relation function of Lee, Yang, and Parr (LYP). The molecular structure of SS was optimized using the B3LYP function in conjunc-tion with a modest 6-311g (d) basis set.
Fabrication and characterization of AgNR array substrates
Fig. 2A and B show representative top-view and cross-sectional view SEM images of the AgNR substrate. Both images show that the tilted and aligned Ag nanorods are formed, and many of them are in fact connected with each other (Fig. 2A). The measured nanorod length L = 900 ± 90 nm, the rod diameter D = 150 ± 40 nm, the average rod-to-rod spacing S = 100 ± 30 nm, and the nanorod.
Detection of sodium saccharin in soft drinks
Soft drinks have a sweet flavor, however, due to the expensive price of sugar, some factories add SS into soft drinks to reduce cost. Thus, on-site detection of SS in food is highly necessary to guar-antee food safety and human health. In this study, we used AgNR array substrates and four different, but commonly consumed soft drinks to carry out the SERS-based detection: Cola, Sprite, Fanta and Schweppes. For the rapid testing, all of these soft drinks were mixed with SS (CSS = 100 mg/L), and then introduced onto the substrate surface. Once the solution evaporated, the substrate was subjected to the SERS measurement. Without pretreatment, it was difficult to distinguish the peaks of SS due to the strong background signal of the complicated components in soft drinks. Moreover, sugar and CO2 play an important role in soft drinks and may influence the results. Based on the conventional treatment, the samples were heated to release the CO2 first, and then saccharin was extracted by ethyl acetate. With this facile pretreatment, SS was detected suc-cessfully in these soft drink samples. Fig. 5A shows the SERS spectra of these four different soft drinks. There is a similarity of the spectra between Sprite and Fanta due to their similar formula. Cola exhib-ited a weaker SERS signal compared to the other three kinds of soft drinks. Although the drinks themselves have a considerable SERS signal, SS was still detected successfully after a simple and rapid pretreatment. The result shows that characteristic peaks of SS are easily distinguishable from the background signal of the soft drinks (Fig. 5B).
The SERS spectra of SS in different soft drinks with different concentrations were displayed in Fig. 6, and the calibration curve of detecting SS from soft drink samples was established in Fig. S3.
In soft drink samples, the SERS intensity of SS was increasing with the increased CSS linearly at the range of 5–60 mg/L. However, the rate of intensity increased were getting smaller and shows the non-linear relationship between intensity and CSS when CSS was continuously getting higher. This is due to that there is a competi-tive adsorption process between saccharin and the other additives in soft drink samples when the mixture was added onto the AgNR substrates. As the concentration increased, the number of saccha-rin molecules attached to the AgNR progressively increases up to a limit where all SERS active sites are occupied . PLS-DA is a classification method which emphasizing latent variables between or among classes. It is used to determine the positive or negative detection with its advantages of full-spectrum, multivariate and supervised analysis. In order to distinguish the existence of SS in these four different soft drinks and determine the LOD, PLS-DA was used to analyze the SERS spectra. Red dashed lines indicate the positive/negative threshold generated by the model. The result shows that the sensitivity and specificity between the positive and negative are achieved for SS among the four different soft drinks (Fig. 6) with different concentrations. In PLS-DA analysis, SS is hard to be distinguished at the low concentrations, and the LODs of SS in Sprite, Cola, Fanta, and Schweppes were ultimately determined to be 20, 5, 10 and 20 mg/L, respectively. Furthermore, the LOD can be improved by extracting SS from larger volume of drink sample, i.e., keeping the ethyl acetate volume the same, and increasing the volume of the soft drink during the SS extraction (Fig. S4).
Conclusions
A quick and highly efficient SERS method for the quantitative detection of sodium saccharin in soft drinks has been developed using AgNR array substrates in this study. These SERS substrates possess “hot spots” located at the top and bottom of the AgNR arrays according to the FDTD analysis, and an EF measured to be as high as 108. These AgNR substrates can achieve the sensitive detection of sodium saccharin with a broad linear range of 0.5 mg/L–100 mg/L, a LOD of 0.3 mg/L, and a LOQ of 0.8 mg/L. In addition, with the assistance of a one-step liquid-liquid extraction process, sodium saccharin can be extracted and detected. PLS-DA analysis was performed to determine the LODs of sodium saccharin in four soft drinks, Sprite (20 mg/L), Cola (5 mg/L), Fanta (10 mg/L), and Schweppes (20 mg/L). This SERS-based detection method can be performed within 10 min, which is ideal for multiple applications and on-site detection. In addition, this method could be a con-venient alternative to the standard procedures for the quantitative determination of sodium saccharin in soft drinks.