Rapid, Simultaneous, and Automatic Determination of Lead and Cadmium in Cereals with a New High Performance Composite Hollow Cathode Lamp Coupled to Graphite Furnace Atomic Absorption Spectrometry

A simple, rapid, sensitive, accurate, and automatic graphite furnace atomic absorption spectrometry (GFAAS) method for detecting Cd and Pb in cereals is presented. This method enables the simultaneous determination of Cd and Pb in cereals with a pre-treatment method of diluted acid extraction and a high-performance lead–cadmium composite hollow-cathode lamp (LCC-HCL), and it realizes automatic determination from sample weighing to result output through an automatic diluted acid extraction system. Under the optimization, Pb and Cd in cereals were simultaneously and automatically detected in up to 240 measurements in 8 h. The LOD and LOQ of this method were 0.012 and 0.040 mg·kg−1 for Pb, and 0.0014 and 0.0047 mg·kg−1 for Cd, respectively. The results of the four certified reference materials were satisfied; there was no significant difference compared with the ICP-MS method according to a t-test, and the RSDs were less than 5% for Cd and Pb. The recoveries of naturally contaminated samples compared with the ICP-MS method were favorable, with 80–110% in eight laboratories. The developed method is rapid, low-cost, and highly automated and may be a good choice for grain quality discrimination and rapid analysis of Cd and Pb in different institutions.


Introduction
Cadmium (Cd) and lead (Pb) are toxic heavy metals, and the harmful effects on human health caused by Pb and Cd contamination are well known [1,2]. It has been reported that cereals are the main contributors to Cd and Pb exposure [3]. In developing countries, Pb and Cd pollution are serious [4][5][6]. Considering the high cost of inductively coupled plasma mass spectrometry (ICP-MS), graphite furnace atomic absorption spectrometry (GFAAS) is still the dominant analytical technique used for Pb and Cd analysis in cereals [7][8][9].
GFAAS with single-element analysis is time consuming [10], and the pretreatment operation is complicated. At present, because of the low sensitivity for Pb and the high sensitivity for Cd in GFAAS, the current range of Cd is narrow. To detect cereal samples with consistent contamination levels of Pb and Cd, it is difficult to achieve the simultaneous detection of two heavy metals with the same preprocessor [11][12][13].
Notably, for many years, extensive research has been devoted to multi-element simultaneous measurements based on the GFAAS [14,15]. The application of high-resolution continuous light source technology has been found to be the most successful, and a variety of devices are available [16][17][18]. However, these technologies have good applications in flame atomic absorption but are not applicable to the detection of low contents of Pb and Cd in cereals; therefore, it is necessary to develop the simultaneous determination of lead and cadmium in cereals using GFAAS.
Based on this situation, a new high-performance lead-cadmium composite hollowcathode lamp (LCC-HCL) was designed in this study. High-performance hollow-cathode lamps were first reported by Sullivan and Walsh [19], which could enhance sensitivity by providing a stronger and more stable radiation intensity, thus improving spectral interference. Subsequently, Rensburg and Zeeman [20] developed a multi-element, highintensity hollow-cathode lamp with selective modulation for the detection of gold, platinum, palladium, and rhodium, which has since been widely applied [21,22]. Yingqi and Yongzhang [23] designed a new type of high-performance hollow cathode lamp that used two separate discharges to markedly increase light output and minimize the self-absorption broadening of the spectral line, linearity, and sensitivity. Furthermore, Ke and Yingqi et al. used this lamp to measure Pb in water, and the results indicated that the most sensitive analytical line Pb 217.0 nm exhibited better performance of stability, improved signal-tonoise ratio, and enhanced detection limit compared to the sub-sensitive analytical line Pb 283.3 nm [24,25]. J.B. Willis first reported that the copper (Cu) cathode could be used for the buckle background in the high-performance composite hollow cathode [26]; this method is for the background of non-resonant line buckles because it can provide more non-absorbed lines at a suitable wavelength [27].
In addition, traditional pretreatment methods require tedious steps, a large amount of strong corrosive reagents, and conditions of high pressure and temperature [28][29][30][31][32]. Therefore, they are time consuming and unfriendly to researchers and the environment, and it is difficult to realize automation. To overcome these disadvantages, we developed a rapid and green pretreatment method for diluted nitric acid extraction for the determination of Cd and Pb in cereals in previous studies [33][34][35][36], and it was shown to be a fast, convenient, and reliable pretreatment method that only uses 0.5% HNO 3 as the extraction reagent. Furthermore, the diluted acid extraction technique is conducive to realizing automation.
In this study, a new GFAAS method was established with the development of the LCC-HCL. Through optimization of the conditions for diluted acid extraction, a fully automatic dilute acid extraction system was designed to realize the automated determination of Pb and Cd in cereals. This method is efficient, cost-effective, convenient, and suitable for rapid, simultaneous, and automated detection of Pb and Cd in cereals.

LCC-HCL Development and Evaluation
A hollow-cathode lamp was designed by combining a Pb-Cd composite cathode with a Cu cathode to buckle the background, as shown in Figure 1. The characteristic lines of 217.0 nm for Pb and 228.8 nm for Cd were adopted. The role of the auxiliary cathode is to cause a significant change in the intensity of the elements in the Pb-Cd composite cathode.    A dual-channel simultaneous reception method was adopted to realize the reception and measurement of light at different wavelengths [37,38]. Figure 2 shows that the atomizer (G) was designed with a graphite cup with a large capacity, and the maximum injection volume reached 40 µL. The concave grating (M4) was used as a monochromator, the composite light was illuminated on the concave grating through the incident slit, and the characteristic spectral lines of Pb and Cd were imaged on the Rowland grating after being split. The 217.0 nm wavelength of Pb, 228.8 nm wavelength of Cd, and near wavelength of Cu were projected onto the position of the photomultiplier tubes (P1, P2) through the exit slit, and a reflector (M5) was added to reach the appropriate position. The construction ensured that light with different wavelengths was received and measured, and optical paths were successfully developed for the simultaneous determination of Pb and Cd.
of Pb and Cd.  To ensure that these high-performance composite hollow-cathode lamps could used for the detection of Pb and Cd in cereals, lamp performance was investigated testing the noise, drift, and sensitivity of three random composite hollow cathode lam ( Table 1). The results showed that the noise, drift, and consistency of the sensitivity m the requirements of the application.  To ensure that these high-performance composite hollow-cathode lamps could be used for the detection of Pb and Cd in cereals, lamp performance was investigated by testing the noise, drift, and sensitivity of three random composite hollow cathode lamps ( Table 1). The results showed that the noise, drift, and consistency of the sensitivity met the requirements of the application. The same heating program was used to detect Cd and Pb in cereals, as shown in Table 2. To achieve the best detection, the pyrolysis temperature and atomization temperature were optimized. The pyrolysis temperature was optimized by testing a standard solution (STD solution) and five cereal extractions ( Figure 3). Considering the sensitivity and background, a temperature of 320 • C was selected as the optimum pyrolysis temperature. Similarly, the optimum atomization temperature was 1700 • C.
The same heating program was used to detect Cd and Pb in cereals, as show Table 2. To achieve the best detection, the pyrolysis temperature and atomization perature were optimized. The pyrolysis temperature was optimized by testing a standard solution (STD s tion) and five cereal extractions ( Figure 3). Considering the sensitivity and backgroun temperature of 320 °C was selected as the optimum pyrolysis temperature. Similarly optimum atomization temperature was 1700 °C.

Extraction of Pb and Cd from Cereals
Based on our previous study [33,34], a diluted nitric acid solution was selected to extract Pb and Cd from cereals. To ensure that the Pb and Cd in cereals were extracted completely under the same conditions, the key factors were optimized with three different cereals. The optimized results showed that the following: the concentration of nitric acid solution needed to be above 5%; the extraction temperature needed to be above 20 • C, which was room temperature; the particle diameter size of the samples needed to be less than 0.38 mm; the ratio of liquid to solid needed to be between 1:25 and 1:50; the extraction time needed to be at least 5 min; and it was necessary to separate the liquid and solid under the condition of a free setting time of 10 min. (The details of the optimization are described in the Supplementary Materials).

Automation Diluted Acid Extraction System
To realize the automation of Pb and Cd detection, an automated diluted acid extraction system was designed, as shown in Figure 4, under the optimized conditions of the pretreatment. The system includes a real-time reading part for balance, a part for adding the extraction solution, an oscillation part, and an automatic sample injection part, which can realize the automatic determination from extraction of the sample to the result output. The simultaneous and automated determination of Pb and Cd in cereals is shown in Figure 5. An electronic balance with a communication port is used to read the weighing quality of the sample in real time, and the effective traceability of the measurement results is achieved through software recording. A high-precision syringe pump was used to add the acid solution to the designated sample tube automatically. Magnetic stirring and a constant temperature heating plate were used to realize the oscillation for extraction (5 min). After the setting time (10 min), automatic sample injection and dilution were realized through the sample injection arm and the high-precision injection pump. The sample was then detected by a detection system to obtain the concentration of Pb and Cd in the sample (1.5 min). With this system, up to 240 measurements could be realized within 8 h.

Extraction of Pb and Cd from Cereals
Based on our previous study [33,34], a diluted nitric acid solution was selected to extract Pb and Cd from cereals. To ensure that the Pb and Cd in cereals were extracted completely under the same conditions, the key factors were optimized with three different cereals. The optimized results showed that the following: the concentration of nitric acid solution needed to be above 5%; the extraction temperature needed to be above 20 °C, which was room temperature; the particle diameter size of the samples needed to be less than 0.38 mm; the ratio of liquid to solid needed to be between 1:25 and 1:50; the extraction time needed to be at least 5 min; and it was necessary to separate the liquid and solid under the condition of a free setting time of 10 min. (The details of the optimization are described in the Supplementary Materials).

Automation Diluted Acid Extraction System
To realize the automation of Pb and Cd detection, an automated diluted acid extraction system was designed, as shown in Figure 4, under the optimized conditions of the pre-treatment. The system includes a real-time reading part for balance, a part for adding the extraction solution, an oscillation part, and an automatic sample injection part, which can realize the automatic determination from extraction of the sample to the result output. The simultaneous and automated determination of Pb and Cd in cereals is shown in Figure 5. An electronic balance with a communication port is used to read the weighing quality of the sample in real time, and the effective traceability of the measurement results is achieved through software recording. A high-precision syringe pump was used to add the acid solution to the designated sample tube automatically. Magnetic stirring and a constant temperature heating plate were used to realize the oscillation for extraction (5 min). After the setting time (10 min), automatic sample injection and dilution were realized through the sample injection arm and the high-precision injection pump. The sample was then detected by a detection system to obtain the concentration of Pb and Cd in the sample (1.5 min). With this system, up to 240 measurements could be realized within 8 h.    Table 3 indicate that the relative difference between the measured and theoretical values of each group was not significantly different from the comparison; therefore, no inorganic interference was observed when GFAAS with LCC-HCL was employed to simultaneously detect Cd and Pb in cereals.

Matrix Effect
The most difficult problem to solve is the matrix effect from some soluble proteins, sugars, and fats [39], which can affect the detection of Pb and Cd. To eliminate matrix interference, the chemical matrix modifier, standard curves of different matrices, and their corresponding results were investigated.
After optimization, 0.04% Pd(NO 3 ) 2 was selected to improve the determination of Cd and Pb in cereals. Based on the matrix matching standard curve, three matrix reference materials were used to conduct a comparative study with and without matrix modifiers, and the results are shown in Figure 6. It was shown that a chemical matrix modifier of 0.04% Pd(NO 3 ) 2 was necessary to obtain an accurate value.
To further determine which should be selected for matrix matching, different matrixmatching standard calibration curves were established to eliminate matrix interference and obtain accurate values for Pb and Cd (Figure 7). The standard curves of the different matrices showed no significant differences, and three certified reference materials were applied. The corresponding detected values of these matrix matching standard curves are presented in Table 4, which shows that the detected values were all within the range of the certified reference material assignment. To sum up, it is sufficient to obtain accurate and reliable test results for Pb and Cd by using the matrix matching standard curve of any blank cereal and a matrix modifier of 0.04% Pd(NO 3 ) 2 . different matrices showed no significant differences, and three certified reference mate rials were applied. The corresponding detected values of these matrix matching standard curves are presented in Table 4, which shows that the detected values were all within the range of the certified reference material assignment. To sum up, it is sufficient to obtain accurate and reliable test results for Pb and Cd by using the matrix matching standard curve of any blank cereal and a matrix modifier of 0.04% Pd(NO3)2.   To further determine which should be selected for matrix matching, different matrix-matching standard calibration curves were established to eliminate matrix interference and obtain accurate values for Pb and Cd (Figure 7). The standard curves of the different matrices showed no significant differences, and three certified reference materials were applied. The corresponding detected values of these matrix matching standard curves are presented in Table 4, which shows that the detected values were all within the range of the certified reference material assignment. To sum up, it is sufficient to obtain accurate and reliable test results for Pb and Cd by using the matrix matching standard curve of any blank cereal and a matrix modifier of 0.04% Pd(NO3)2.    The trueness of the method was identified by comparison with the standard method (microwave digestion for pre-treatment and detection by ICP-MS), and four reference materials (maize flour, brown rice flour, wheat flour, and rice flour) as the matrix samples. The results summarized in Table 5 indicate that the value was within the certified value and uncertainty, and the results showed no statistical difference between the two methods at the 95% level according to the t-test, which indicates that the method is accurate enough for the simultaneous determination of Cd and Pb in cereals.

Precision
The precision of the method was studied using a standard solution (Pb 8 µg·L −1 , Cd 4 µg·L −1 ) and three matrix extracted solutions of maize, brown rice, and wheat. The results showed that the RSDs were 1.5% and 1.8% for Pb and Cd in the brown rice extracted solution, 4.4% and 2.9% for Pb and Cd in the wheat extracted solution, 3.2% and 4.8% for Pb and Cd in the maize extracted solution, and 2.2% and 2.7% for the Pb and Cd in the standard solution, respectively, indicating that there was no significant difference between the standard solution and the matrix sample extraction solution, and the RSDs based on seven measurements were always lower than 5.0%.

Calibration Curves, Linearity, and Limit of Detection and Quantification
The calibration curves, linearity, and limit of detection and quantification are shown in Table 6. The limit of detection (LOD) and the limit of quantification (LOQ) were calculated using blank samples (the concentrations of Pb and Cd were both lower than 0.010 mg·kg −1 ) according to ISO 11843-5:2008 [40]. When the weight was 0.2 g and the extraction solution volume was 5 mL, the LOD and LOQ were calculated and expressed in mg·kg −1 . The LOD of this method for Pb decreased in multiples compared with the traditional GFAAS [34], so the reliability and practicability were greatly enhanced.

Naturally Contaminated Sample Analysis
The method was further applied to the naturally contaminated samples analysis. Ten typical samples containing both Pb (0.090 mg·kg −1 -0.40 mg·kg −1 ) and Cd (0.020 mg·kg −1 -0.45 mg·kg −1 ) were chosen as the application samples, which included brown rice, wheat, and maize. The naturally contaminated samples were analyzed using the microwave digestion-ICP-MS method for comparison, and then analyzed using the optimized methods of this study in eight laboratories. The relative recoveries were calculated and are shown in Tables 7  and 8. The relative recoveries were favorable at 80-117%, and the results of the different labs showed good consistency. It was proven that the method developed in this study has good availability, and it could be a satisfactory choice to determine different concentrations of Pb and Cd in various cereal matrices.
An atomic absorption spectrometer (TAS 986, Purkinje, China) with a newly designed optical path (Figure 2) was used for the simultaneous determination of Pb and Cd. The LCC-HCL was manufactured according to our design ( Figure 1) by being entrusted to a company with legal manufacturing qualifications (Baoding Hengming Light Source Electronic Technology Co. Ltd., Baoding, China). The operating conditions and heating program of GFAAS are listed in Table 2.
Inductively coupled plasma mass spectrometry (ICP-MS, Agilent, 7500cx, Santa Clara, CA, USA) was used for comparison. The operating conditions for the optimization of the experiment are shown in Table 9.

Sample Preparation
The samples were ground in a mill until they passed through the sample sieve and then mixed thoroughly.
Samples for microwave digestion pre-treatment were prepared as follows: cereal samples (0.2-0.3 g) were digested with HNO 3 (Guaranteed Reagent, 7 mL) using a microwave digestion system (the microwave digestion program is presented in the Supplementary Materials). The digestion solution was heated (160 • C) until it remained at 1-2 mL. The solution was transferred and fixed at 25 mL using a volumetric flask with ultrapure water; the blank was produced using the same procedure.
Samples for extraction were prepared as follows: Samples (0.2 g) were mixed with 5 mL HNO 3 solution (5%, v/v), shaken for 5 min, and stood for 10 min.
Matrix-matching solution: 4 g blank cereals (brown rice powder, rice powder, wheat powder, flour powder, and maize powder without Pb and Cd) were weighed, and 100 mL HNO 3 solution (5%, v/v) was added, followed by shaking for 5 min at a temperature above 20 • C, and standing for 10 min.
The noise test for LCC-HCL: Three LCC-HCL were randomly selected and preheated for 30 min under lamp working conditions ( Table 2). The noise is the instantaneous noise (peak to peak) recorded in 15 min.
The drift test for LCC-HCL: Three LCC-HCL were randomly selected and preheated for 30 min under lamp working conditions ( Table 2). The drift is the zero-drift recorded in 15 min.
The sensitivity test for LCC-HCL: Three LCC-HCL were randomly selected and preheated for 30 min under lamp working conditions (Table 2), the average absorbance of 7 times determinations of 5% (v/v) HNO 3 solution is A 0 , and the average absorbance of 7 times determinations of 10 µg·L −1 standard solutions of Pb and Cd is A 1 (Cd) or A 1 (Pb), the sensitivity is the difference between A 1 and A 0 .

Conclusions
In this study, a rapid, automatic, and simultaneous method and an automatic diluted acid extraction system for the detection of Pb and Cd in cereals were developed. Under optimal conditions, the method was confirmed to have satisfactory trueness and precision. Moreover, Pb and Cd in cereals could be simultaneously and automatically detected in up to 240 measurements in 8 h when the automatic detection cycle was running, which could greatly reduce both the analysis time and errors caused by manual operation. The developed method is in line with the need for the automatic and efficient detection of Pb and Cd in cereals, and further optimization of the process could make this method applicable to the detection of Pb and Cd in other products, such as milk powder, vegetables, or oils.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27238571/s1, Table S1: Microwave digestion program; Table S2: Effect of extraction time for Cd and Pb; Table S3: The effect of the setting mode and time for Cd and Pb; Figure

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the study population to publish this paper.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.