- Email: [email protected]
Development and validation of HPLC method for the determination of ferrocyanide ion in food grade salts
Accepted Manuscript Development and validation of HPLC method for the determination of ferrocyanide ion in food grade salts Ho Soo Lim, Ju Young Hwang, EunA Choi, Gunyoung Lee, Sang Sun Yoon, MeeKyung Kim PII: DOI: Reference:
S0308-8146(17)31223-2 http://dx.doi.org/10.1016/j.foodchem.2017.07.070 FOCH 21454
To appear in:
Food Chemistry
Received Date: Revised Date: Accepted Date:
5 April 2017 20 June 2017 14 July 2017
Please cite this article as: Soo Lim, H., Young Hwang, J., Choi, E., Lee, G., Sun Yoon, S., Kim, M., Development and validation of HPLC method for the determination of ferrocyanide ion in food grade salts, Food Chemistry (2017), doi: http://dx.doi.org/10.1016/j.foodchem.2017.07.070
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Development and validation of HPLC method for the determination of ferrocyanide ion in food grade salts
Ho Soo Lim a, Ju Young Hwang a, EunA Choi a, Gunyoung Lee a, Sang Sun Yoon a and MeeKyung Kim a, *
a
Food additives and Packaging Division, Ministry of Food and Drug Safety, Cheongju,
Chungcheongbuk-do 28159, Korea
Running title: Determination of ferrocyanide ion in food grade salts
*
Corresponding author: MeeKyung Kim, Ph.D Food additives and Packaging Division, National Institute of Food and Drug Safety Evaluation, Ministry of Food and Drug Safety, Cheongju, Chungcheongbuk-do 28159, Korea Tel: +82-43-719-4351 Fax: +82-43-719-4350 E-mail: [email protected]
Abstract
A rapid, simple, and reliable HPLC method was developed and validated to determine the presence of ferrocyanide ions (FeCNs) in food grade salts. An analytical column coupled with a guard column and mobile phase comprised of sodium perchlorate and sodium hydroxide (NaOH) were employed at a detection wavelength of 221 nm. Samples were dissolved in NaOH and filtered. For processed salts including herbs and spices, a C18 cartridge was applied to minimize interference from salt matrices. The method validation was based on linearity, selectivity, accuracy (recovery), precision, LOD, LOQ, and measurement uncertainty. This method exhibits good linearity from 0.1−10 mg/L (r2 = 0.9999). The LOD and LOQ values were determined to be 0.02 and 0.07 mg/kg, respectively. The FeCN recoveries in six salt matrices ranged from 80.3−102.2% (RSD = 0.3−4.4%). These results indicate that the proposed method is suitable for FeCN ion determination in various food grade salts.
Keywords: Ferrocyanide ion; food grade salts; method validation; HPLC
1. Introduction Sodium ferrocyanide, potassium ferrocyanide, and calcium ferrocyanide are complex cyanides that have been used as anticaking agents in food grade salts for many years (Bode et al., 2012; Gupta, Pel, Steiger, & Kopinga, 2015). These food additives are prepared from the reaction of hydrogen ferrocyanide with sodium hydroxide (NaOH), potassium hydroxide, and calcium hydroxide, respectively. They have low toxicity because of the strong chemical bond between iron and the cyanide groups. The additives were evaluated by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) and the Scientific Committee on Food (SCF). These committees established an acceptable daily intake (ADI) of 0−0.025 mg/kg of body weight (JECFA, 1974; European Commission, 1991). The use of ferrocyanides in food grade salts is authorized in the European Union, Japan, and Korea. However, in the United States, only sodium ferrocyanide is permitted in food grade salts. In Korea, these additives are used in food grade salts at levels within the legal limit, in accordance with the Korea Food Additives Code (MFDS, 2015a); the ferrocyanide ion (FeCN) content should not exceed 10 mg/kg for food grade salts, except for solar salts (MFDS, 2015b). In these salts, no FeCN should be detected. According to the food labelling regulations of the Korea Food Sanitation Act (MFDS, 2016), ferrocyanides should be labelled on the products as ‘calcium ferrocyanide, potassium ferrocyanide, or sodium ferrocyanide’. In the European Union, the maximum level of ferrocyanides in food grade salts and their substitutes should not exceed 20 mg/kg (calculated as anhydrous potassium ferrocyanide) individually or combined (European Commission, 2011). In the United States, sodium ferrocyanide should not exceed 13 mg/kg as anhydrous sodium ferrocyanide for food grade salts (US FDA, 2016). In Japan, a usage level of 20 mg/kg has been set for food grade salts (MHLW, 2016), whereas, in CODEX, ferrocyanides are set at <14 mg/kg for food grade salts and
<20 mg/kg for substitutes and sauces (CODEX, 2016). Accordingly, monitoring the ferrocyanide content in products is necessary to manage these standards, as well as to ensure food safety. Several analytical methods that determine the FeCN content in salts have been reported. These include spectrophotometry (Roberts & Wilson, 1969; Kubota, Onishi, Yomota, & Tanamoto, 2004),
infrared spectrometry (Drew, 1973), X-ray fluorometry (Koga & Niino, 2004), on-line redox derivatization high-performance liquid chromatography (HPLC) (Saitoh, Soeta, Minamisawa, & Shibukawa, 2013), and reflective colorimetry (Suzuki, Ishigaki, Oshita, Yamane, & Kawakubo, 2013). Among the cited techniques, HPLC has the advantage of being able to determine the FeCN content in salts without interference from other components. The HPLC method is also popular because of its wide usage, accuracy, and convenience. However, specific precision, limit of detection (LOD), and limit of quantification (LOQ) values from LC methods have not been reported to date. This method has been applied to determine the ferrocyanide content in only one type of food grade salt. To date, it has not been used to investigate different salts, such as processed salts. Recently, the total production and imports of food grade salts in South Korea has increased. The estimation of ferrocyanide intake is necessary because ferrocyanides (sodium ferrocyanide, potassium ferrocyanide, and calcium ferrocyanide) are used to improve the quality of food grade salts. To assess the level of their intake, method development and monitoring data must be undertaken. The official method of the Korea Food Code is problematic in that it has very low sensitivity and is largely unable to detect FeCNs in food grade salts due to the interference of sodium chloride (NaCl). Also, in the case of complex processed salts containing spice and flavour, FeCNs could not be detected. Thus, development of new methods for FeCN detection is necessary to ensure food safety through the improvement of method sensitivity. In this study, we have developed a rapid and quantitative method for the analysis of FeCN in various types of salts by HPLC coupled to a diode array detector. Method validation was performed to determine linearity, selectivity, accuracy (recovery), precision, LODs, and LOQs. In addition, measurement uncertainty was calculated in food grade salts for the first time; six types of commercial salts were analysed.
2. Materials and methods
2.1. Chemicals and reagents
Sodium ferrocyanide standard was purchased from Sigma-Aldrich (St. Louis, MO, USA). NaOH was used as a solvent in sample preparation and was purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Sodium perchlorate (NaClO4) used in the mobile phase preparation was supplied by Sigma-Aldrich. Water for all applications was obtained from a Milli-Q ultra-pure water system (Millipore, Bedford, MA, USA; resistivity ≥ 18.2 MΩ cm).
2.2. Preparation of the standard solution A FeCN stock standard solution was prepared by transferring 50 mg of sodium ferrocyanide in a 100 mL volumetric flask and adding 0.02 M NaOH solution to reach a final volume of 100 mL. The working standard solutions were produced by diluting the stock solution in 0.02 M NaOH solution to form 0.1, 0.5, 1, 5, and 10 mg/L aliquots. The stock solutions were prepared weekly and stored at 4 °C; working standard solutions were prepared daily.
2.3. Samples A total of 801 salt samples were purchased from food markets in big cities including Seoul, Daejeon, Gwangju, Daegu, Busan, Cheongju, and Sejong in South Korea. The samples were categorized into six categories according to the Korea Food Code, namely, reworked salts (42), burnt-molten salts (175), refined salts (30), processed salts (352), solar salts (169), and other salts (33). Reworked salt refers to a product made by dissolving raw salt (100%) in purified water, sea water, or concentrated sea water, and processing it through filtration, precipitation, recrystallization, dehydration, and adjustment of the salt concentration. Burnt-molten salt is a product made by modifying raw salt (100%) through burning, melting, or other methods. However, it excludes those products processed by washing, grinding, or compressing. Refined salt refers to a product of concentrated brine obtained by purifying sea water (including deep ocean water) through an ion exchange membrane, or to salt obtained by drying a solution of raw salt (100%) in water in a vacuum evaporator. Solar sea salt refers to crystals of NaCl made through natural distillation of sea water at a salt field, and to products obtained by grinding, washing, dehydrating, or drying them. Processed salt refers a product made by
adding foods or food additives to reworked salt, burnt-molten salt, refined salt or other salt (not less than 50%) and processing it. The other salts include powdered or crystallized products manufactured into by processing rock-salt or lake-salt into an edible form. Salt products included in the above salt types are excluded from this category. Six FeCN-free salts were also selected to validate the method. All samples were stored at room temperature (20−25°C).
2.4. Sample preparation Each salt sample (2−5 g) was transferred into a 50 mL volumetric flask and dissolved in a 0.02 M NaOH solution to generate a total volume of 50 mL. The sample was incubated in an ultrasonic bath for 5 min. The extract was then filtered through a 0.22 µm polyvinylidene difluoride (PVDF) membrane filter for HPLC analysis. For the processed salts, 2 mL of the extract was subjected to an enrichment procedure using a Sep-Pak C18 cartridge (Waters Co., Milford, MA, USA; 500 mg) at a flow rate of ∼3 mL/min. The cartridge was preconditioned with 5 mL methanol followed by 5 mL water. After loading the sample, the flow-through solution was collected into a 15 mL conical tube. The solution was then filtered with a 0.22 µm PVDF membrane filter.
2.5. HPLC
HPLC analysis was performed on a Shiseido Nanospace SI-2 (Shiseido Co., Tokyo, Japan) coupled to a photodiode array detector (PDA). The HPLC system included an online degasser, a binary pump, an autosampler, and a column oven. An AG11-HC (4 mm × 50 mm, 9 µm) guard column and AS11HC (4 mm × 250 mm, 9µm) analytical column were used for chromatographic separation. Both columns were purchased from DIONEX (Thermo Fisher Scientific, Sunnyvale, CV, USA). The FeCN chromatograms were analysed by Ezchrom Elite software, Shiseido Corporation. All separations were carried out isocratically at 35 °C with a mobile phase comprised of 200 mM NaClO4 and 20 mM NaOH. The flow-rate was maintained at 0.6 mL/min, and a 10 µL sample volume was injected for all experiments. The FeCN eluted from the columns was monitored by a PDA
detector set at 221 nm. The FeCN absorption spectra were recorded at wavelengths ranging from 200 to 600 nm. Peak identification was performed by comparing the retention times and absorption spectra of the samples with those of the standard solutions.
2.6. Method validation
The HPLC-UV method for the determination of FeCN in food grade salts was validated for linearity, selectivity, accuracy (recovery), precision, LOD, and LOQ according to AOAC guideline (2012). The selectivity was tested by examining the chromatograms to verify the absence of interfering peaks in the food grade salts. Matrix-matched calibration was prepared by spiking five concentrations (0.1, 0.5, 1, 5, 10 µg/mL) of FeCN in reagent-grade NaCl. The linearity was calculated using these five concentrations in triplicate. The LOD and LOQ values were calculated from calibration curves attained from each reagent-grade salt spiked with FeCN at five concentrations. These were determined at 3.3 and 10 σ/S, respectively, where σ is the standard deviation of the intercept and S is the slope of the regression line determined from the calibration curve. Accuracy (recovery) and precision (expressed as %RSD) were evaluated by analysing six samples spiked with the stock solution to afford final concentrations of 1, 5, and 10 mg/kg. Intra-day (three repetitions on the same day) and inter-day (three repetitions over three different days) precision tests were performed. Finally, the FeCN measurement uncertainty was calculated by the bottom-up method from the validation data afforded during each step of the analytical procedure (JCGM, 2008; EuraChem, 2000). Four sources of uncertainty were considered for the determination of measurement uncertainty: the calibration curve, matrix (recovery), standards preparation, and sample preparation. Thus, U (expanded uncertainty) was calculated by multiplying the combined standard uncertainty by a coverage factor (k = 2) that yields a confidence level of ∼95%.
2.7. Stability test To study storage stability, the FeCN concentrations in two salt and two solution matrices were examined. In salt matrices, 2 g of reworked salt and reagent grade NaCl were spiked at a level of 5
mg/kg of FeCN. In solution matrices, 2 g of reworked salt and reagent grade NaCl were dissolved in 0.02 M NaOH, and then FeCN was added at a final concentration of 5 mg/kg. The samples were stored at room temperature for 90 days, withdrawn according to a predefined schedule (6, 9, 12 h; 1, 3, 7, 14, 30, 60, 90 d), and stored at 4 °C until the analyses were performed. The tests were repeated in triplicate at each time point.
3. Results and discussion
3.1. Method performance To optimize the FeCN detection conditions, parameters that could influence the analytical performance were investigated. These included LC conditions, mobile phase, sample amounts and sample clarification for FeCN extraction.
3.1.1. Liquid chromatographic conditions The FeCN peak was not detectable in reagent-grade NaCl, although it was detected in the standard solution (5 mg/L FeCN in 20 mM NaOH). To acquire detectable FeCN peaks, different combinations of NaClO4 and NaOH solutions were tested as the mobile phase. A good separation of FeCN from spiked reagent-grade sodium chloride was achieved with a 200 mM NaClO4/20 mM NaOH solution mobile phase under the isocratic elution conditions described in Section 2.5 (Fig. 1). When the FeCN absorption spectra was recorded at a range of 200−400 nm, maximum absorption was observed at 221 nm. We next compared the chromatograms afforded from two ion exchange columns, Shodex IC SI90 4E (4.0 × 250 mm, 9 µm) and Dionex AS11-HC (4.0 × 250 mm, 9 µm) coupled with an AG11-HC (4 mm × 50 mm, 9 µm) guard column, by analysing the extracts of spiked reagent-grade NaCl. When only the Shodex column was used, FeCN was not detected in any of the samples. However, when the Dionex columns were used, FeCN peaks were observed in the chromatograms of spiked reagent-grade NaCl. An AS11-HC column coupled with an AG11-HC column was found to be the most suitable for determining FeCNs in food grade salts (Fig. 2). The flow rate was then increased from 0.6 mL/min to
1.2 mL/min, and a column temperature of 35 °C or 40 °C was applied; the highest FeCN peak area was afforded at a flow rate of 0.6 mL/min and a column temperature of 35 °C. At these conditions, the analysis was completed in <10 min, and the chromatograms displayed FeCN retention times of 7.8 min. Therefore, we chose these HPLC conditions for further study. The chromatograms presented in this study have not been reported in the literature to date. Recently, Saitoh et al. (2013) developed an on-line redox derivatization HPLC method to determine Fe (II) and Fe (III) cyanide complexes in food grade salts. The LC system consisted of two C18 silica columns treated with trimethylstearylammonium chloride and a small column packed with porous graphitic carbon (PGC) placed between them. In this method, the peak of the Fe (II) cyanide complex (ferrocyanide ion) was asymmetric, exhibiting tailing. In addition, since three columns were used in series, this method could be considered relatively expensive and troublesome. In our present research, the developed method, comprised of a widely used HPLC system, provided a simple analysis of FeCN using an ion exchange column with UV detection (wavelength 221 nm), and was found to be suitable for the quantification of FeCN in food grade salt.
3.1.2. FeCN extraction from food grade salts Samples (2−5 g) were employed to examine the effects of NaCl concentration on the peak area. The FeCN peak areas significantly decreased when the sample amount was increased to 5 g; 2 g salt samples provided the greatest FeCN peak areas. In the processed salts containing spices such as pepper and flavours, sample clarification is important to achieve good chromatographic peak resolution. In this study, a Sep-Pak C18 cartridge was employed to improve FeCN resolution from endogenous interfering peaks. To choose the appropriate loading volume on the C18 cartridge, we analysed the received solutions after loading each 2−5 mL sample on the C18 cartridge. The 2 mL loading volume afforded the best peak area and separation (Fig. 3A). To confirm the reliability of purification with the Sep-Pak C18 cartridge, we compared the extracts of a processed salt product containing sodium ferrocyanide and rosemary leaves before and after the cartridge was applied. When the cartridge was not applied, the FeCN peak could not be found due to interference from the other
sample components. However, a detectable peak was afforded after applying the cartridge (Fig. 3B).
3.2. Method validation The analytical method developed in this study was compared with the method of the Korea Food Code, as well as other methods. In the Korean Food Code method, the detection sensitivity of FeCN is very low, due to the influence of copious NaCl present in food grade salt. Yamane et al. (2006) showed that, in a detection method for FeCN with a flow injection system, the correlation coefficient (r2) of the calibration curve was 0.999, recoveries were 95–104%, and precision was 3% as RSD. An LOD of 0.003 ppm was presented but no LOQ was disclosed. Kubota et al. (2004) analysed FeCN in food grade salt using spectrophotometry. The correlation coefficient (r2) of the calibration curve was 0.9996, recoveries were 91.1–105.5% and precision was less than 4%. The LOD and LOQ were not presented. In addition, Koga et al. (2004) determined the presence of FeCN in a 20% NaCl solution using X-ray fluorometry. The correlation coefficient (r2) was 0.99, and the LOD was 0.1 mg/kg. Recovery, precision, and LOQ were not shown. Haba & Wilson (1963) reported a small-scale titrimetric determination of ferrocyanide. The method may be applied to amounts of 1–100 mg of the complex ion. Titration is susceptible to variation according to the amount of titrating solution, and is interfered with chloride and any other substances present in a matrix (Fennema, 1996). In the above papers, the analyses of FeCN were performed in the matrices of some food grade salts or reagent-grade NaCl and were not applied to various complex food grade salts containing processed salts. Some validation results were presented, but since the above methods utilize absorption or fluorescence, the detection of FeCN can be influenced by the matrix. Recently, Saitoh et al. (2013) introduced on-line redox derivatization HPLC; the reported correlation coefficient of the calibration curve (r2) was 0.999, but no validation results were provided. In our current study, the developed method has been established by greatly improving the conditions, including the mobile phase, column, and purification process, from the Korea Food Code method. The correlation coefficient (r2 ) of the calibration curve is 0.9999 or higher, and recoveries were 80.3–103.5% for six kinds of food grade salts, including processed salt. The precision expressed as %RSD was less than 4.5% for intra- and inter-day analyses, the LOD was 0.02 ppm, and the LOQ was 0.07 ppm. The validation results of the
method are considered to be equal to or higher than those of conventional methods. The sample pretreatment time was similar, but the instrument analysis time was relatively short (~15 min) compared to the HPLC method using on-line redox derivatization of Saitoh et al. (2013). Saitoh’s method included a similar sample pretreatment process time to our method. In the analysis of metal cyanide complexes using HPLC, the method required a long time of about 40 min. The accuracy and sensitivity of the method were not determined because the recovery, LOD and LOQ were not presented. This method is very problematic and not widely used because of the inclusion of a small column packed with porous graphitic carbon (PGC) between two C18 silica columns treated with trimethylstearylammonium chloride. Thus, our method was developed to allow for facile FeCN detection at a UV wavelength using a widely used HPLC system.
3.2.1. Linearity To support regulatory action, a method must be proven as accurate, sensitive, and able to identify an analyte with high selectivity. For this purpose, validation of analytical methods should include assessments of linearity, selectivity, accuracy (recovery), precision, LOD, LOQ, and measurement uncertainty. The linearity of the method was examined by analysing five concentrations ranging from 0.1 to 10 mg/L in spiked reagent-grade NaCl (n= 3). A calibration curve of peak area versus concentration was then plotted. The average correlation coefficient (r2) was determined to be 0.9999 (Table 1). A study using a flow injection system with an anion exchange column generated a linear FeCN calibration curve in the range of 0−0.3 ppm for FeCN in the presence of 0.5 M NaCl (Yamane, Isawa, & Osada, 2006). The coefficient value was higher than those reported in the literature (Saitoh et al., 2013; Suzuki et al., 2013).
3.2.2. LOD and LOQ The LOD and LOQ values were calculated from linear curves acquired from three replicate injections of each solution, which were prepared by spiking five known concentrations of FeCN in
reagent-grade NaCl. The LOD and LOQ values were calculated using the equations LOD = 3.3 × σ/S and LOQ=10 × σ/S where σ is the SD of the response, and S is the slope of the corresponding calibration curve. The FeCN LOD and LOQ values were calculated to be 0.02 and 0.07 mg/kg, respectively (Table 1), which were higher than those reported in the literature (Kubota et al., 2004; Koga & Niino, 2004; Suzuki et al., 2013). Two separate studies were previously carried out to determine the FeCN content in food grade salts. One study employed an LC with UV detection at 218 nm. The other consisted of an LC system equipped with a small column packed with porous graphitic carbon (PGC) treated with a redox reagent (Jang, Cho, Bae, & Kim, 2010; Saitoh et al., 2013, respectively). However, neither method reported specific LOD or LOQ values.
3.2.3. Accuracy and precision Accuracy (recovery) was reported as the percentage of FeCN recovered after spiking various known concentrations of FeCN in a blank. Recoveries were calculated by comparing the peak area of the fortified samples with that of the standard. Recovery tests were performed by spiking three known concentrations (1, 5, and 10 mg/kg) of FeCN samples into food grade salts in triplicate. In the reworked salt, burnt-molten salt, refined salt, processed salt, solar salt, and other salt matrices, the average FeCN recoveries were determined to be 85.7−103.5%, 80.3−98.7%, 85.7−102.2%, 92.6−93.9%, 87.1−97.8%, and 86.1−97.7%, respectively. The results are summarized in Table 1. In a study conducted by a portable LED-based 8-channel reflective colorimetric method (Suzuki et al., 2013), an average recovery of 96% was achieved for added FeCN at a 10 mg/kg level. This agrees with the results afforded at the 5 mg/kg level (92.6−97.8%) in this research. Precision, the degree of repeatability of an analytical method, is normally expressed as the relative standard deviation (RSD) for spiked samples. There are two types of precision: intra-day and interday precision. The former was calculated by injecting each of the three different concentrations of mixtures three times on the same day. The latter was evaluated over three different days at the same concentration levels (Thompson, Ellison, & Wood, 2002). The intra-day precision (%RSD) varied
from 0.4 to 4.4% while the inter-day precision ranged from 0.3 to 4.4%. These results demonstrate that the analytical method is sufficiently repeatable. Kubota et al. (2004) reported similar intra-day precision values for FeCN.
3.2.4.Measurement uncertainty To evaluate its suitability, the uncertainty in the results of an analytical method must be quantified to help the analyst make reasonable decisions (Rozet, Marini, Ziemons, Boulanger, & Hubert, 2011). The uncertainty measurements of FeCN in reworked salt were estimated using the proposed method. The uncertainty factors related to the analytical procedure were identified and used to calculate the combined standard uncertainty according to the International Organization for Standardization (ISO., 1993) and Eurachem/Citac (EURACHEM, 2000) guidelines. The five sources of uncertainty considered were associated with the calibration curve, the recovery from salt matrices, the standard solutions, and the sample preparation. The value for U was determined to be 2.5% (0.13 mg/kg) of the measured value (4.9 mg/kg; Table 2) based on a 95% confidence level using a coverage factor of 2. The contribution of each of the individual uncertainty sources is summarized in Table 2. Among the five sources of uncertainty, those associated with the recovery from the salt matrices and the calibration curves are the main contributors to the expanded uncertainty. Therefore, analysts should devote more attention to factors affecting the recovery and calibration curves.
3.2.5. Stability of FeCN in salt and solution matrices The quality of food grade salt depends on the stability of FeCN during its shelf life. To date, no studies on the storage stability of FeCN in food grade salts have been reported. Thus, to the best of our knowledge, this is the first reported examination of the stability of FeCN in food grade salts. The salt matrices included reworked salt and NaCl spiked at a level of 5 mg/kg; the liquid-type matrices were solutions of reworked salt and NaCl dissolved in a 0.02 M NaOH solution. For the analysis of salt matrices, each sample was withdrawn at defined time intervals and extracted in triplicate; each extract was analysed on the HPLC system. In the solution matrices, the sample
solutions were directly injected into the HPLC system after filtration with a 0.22 µm syringe filter. The analysis revealed that no decrease in FeCN concentration was observed over the time investigated (data not shown). Until the end of the studied storage period, similar FeCN concentrations were afforded in all the matrices.
3.3. Application to real samples To assess its reliability, the proposed method was used to analyse the levels of FeCN in various food grade salts purchased from local markets. A total of 801 samples were analysed and classified into six groups according to the types of salts listed in the Korea Food Code. Quantification of FeCN was carried out using standard calibration curves and confirmed by comparing the retention time and absorption spectra of the samples to those of the standards. The FeCN concentrations in the analysed samples are listed in Table 3. The results revealed that reworked salt, burnt-molten salt, refined salt, and processed salt contained FeCN. The highest detection rate was found in refined salt, followed by reworked salt, processed salt, and burnt-molten salt. Among the refined salts, the highest concentration (9.6 mg/kg) was found in a salt from the USA. The concentration levels of the detected samples were below the maximum permissible limit of 10 mg/kg stipulated in Korea. FeCN was not detected in the solar salt and other salts categories.
4. Conclusions A simple, rapid, and reliable HPLC method was developed and validated for the determination of FeCN in food grade salts. Satisfactory method validation results were afforded for FeCN after optimization of the experimental conditions. The measurement uncertainty of the entire procedure was also calculated. The stability of FeCN in the salt and solution matrices was proven under common storage conditions at room temperature (20−25 °C). The application of a C18 cartridge to the processed salts allowed for significant removal of interfering components and a good FeCN resolution towards endogenous interfering peaks. To our knowledge, this is the first study to report the analysis of FeCN in food grade salts using an HPLC-UV method. The method was also verified via the analysis of
different types of salts. Four types of salts were found to contain FeCN; the highest concentration of 9.6 mg/kg was found among the refined salts. However, none of the detected samples exceeded the maximum permissible limit of 10 mg/kg stipulated in Korea. These results demonstrate that this method shows great potential for application in the routine analysis of food grade salts.
Acknowledgements Funding: This work was supported by the Ministry of Food and Drug Safety in 2016 [16161MFDS-011].
References AOAC (Association of Official Analytical Chemistry). Guidelines for single laboratory validation of chemical methods for dietary supplements and botanicals, 2012. URL http://www.eoma.aoac.org/ app_k.pdf. Accessed 04/01/16. Bode, A. A. C., Vonk, V, van den Bruele, F. J., Kok, D. J., Kerkenaar, A. M., Mantilla, M. F., Jiang, S., Meijer, J. A. M., van Enckevort, W. J. P., & Vlieg, E. (2012). Anticaking activity of ferrocyanide on sodium chloride explained by charge mismatch. Crystal Growth & Design, 12, 1919–1924. CODEX (CODEX Alimentarius). Codex General Standard for Food Additives (GSFA), 2016. URL http://www.fao.org/fao-who-codexalimentarius/standards/gsfa/en. Accessed 30.12.16 Drew, D. M. (1973). Simultaneous determination of ferrocyanide and ferricyanide in aqueous solutions using infrared spectrometry. Analytical Chemistry, 45, 2423−2424. European Commission (1991). Reports of the Scientific Committee for Food, 25th series, pp. 18. EURACHEM (2000). Quantifying uncertainty in analytical measurement, 2nd ed., pp. 1−30. European Commission (EC) (2011). Commission Regulation (EU) No 1129/2011 of 11 November 2011 amending Annex II to regulation (EC) No 1333/2008 of the European parliament and of the council by establishing a union list of food additives. Official Journal of the European Union, L295, 1–177. Fennema, O. R. (1996). Food Chemistry. (3rd ed.). New York: Marcel Dekker.
Gupta, S., Pel, L., Steiger, M., & Kopina, K. (2015). The effect of ferrocyanide ions on sodium chloride crystallization in salt mixtures. Journal of Crystal Growth, 410, 7–13. Haba, F. R., Wilson, C. L. (1963). The small-scale titrimetric determination of complex cyanides. Microchimica Acta, 51, 305–312. Jang, M. S., Cho, S. D., Bae, D. H., & Kim, G. H. (2010). Safety and quality assessment of Kimchi made using various salts. Korean Journal of Food Science and Technology, 42, 160–164. JCGM (Joint Committee for Guides in Metrology) (2008). Evaluation of measurement-Guide to the expression of uncertainty in measurements, 1st ed., pp. 1−27. JECFA (1974). Evaluation of certain food additives, 18th Report of the Joint FAO/WHO Expert Committee on Food Additives, WHO Health Organization Technical Report Series, No. 557. Geneva World Health Organization, p. 23−24. Koga, A., & Niino, Y. (2004). Analysis of hexacyanoferrate (II) in salt. Bulletin Society of Sea Water Science, 58, 24−29. Kubota, H., Onishi, Y., Yomota, C., & Tanamoto, K. (2004). Analysis of ferrocyanides in food grade salts. Japanese Journal of Food Chemistry, 11, 32−35. MFDS (Korea Ministry of Food and Drug Safety) (2015a). Korea Food Additives Code. Seoul: Korea Food Industry Association. MFDS (Korea Ministry of Food and Drug Safety) (2015b). Korea Food Code. Seoul: Korea Food Industry Association. MFDS (Korea Ministry of Food and Drug Safety) (2016). Korea Food Sanitation Act. Seoul: Korea Food Industry Association. Miller, E. J. C., & Miller, J. N. (1993). Statistics for analytical chemistry (3rd ed.). New York: Ellis Horwood PTR Prentice Hall. Ministry of Health, Labour and Welfare (MHLW) (2016). URL. http://www.mhlw.go.jp. Accessed 30.12.16. Roberts, R. F., & Wilson, R. H. (1968). The determination of ferrocyanide and related compounds in commercial sodium chloride. Analyst, 93, 237−243.
Rozet, E., Marini, R. D., Ziemons, E., Boulanger, B., & Hubert, P. (2011). Advances in validation, risk and uncertainty assessment of bioanalytical methods. Journal of Pharmaceutical and Biomedical Analysis, 55, 848–858. Saitoh, K., Soeta, N., Minamisawa, H., & Shibukawa, M. (2013). On-line redox derivatization liquid chromatography for selective separation of Fe(II) and Fe(III) cyanide complexes using porous graphitic carbon. Analytical Sciences, 29, 715−721. Suzuki, Y., Ishigaki, M., Oshita, K., Yamane, T., & Kawakubo, S. (2013). Development of a portable LED-based 8-channel reflective colorimeter and its application to simple determination of hexacyanoferrate (II) in common salts. Bulletin of the Society of Sea Water Science, 67, 47−51. Thompson, M., Ellison, S.L.R., Wood, R. (2002). Harmonized guidelines for single laboratory validation of methods of analysis (IUPAC technical report), Pure Applied Chemistry, 74, 835–855. U.S. Food and Drug Administration (USFDA) (2016). Substances Generally Recognized as Safe (From eCFR-Code of Federal Regulations) URL http://www.ecfr.gov. Accessed on 30.12.16. Yamane, T., Isawa, M., & Osada, S. (2006). Simple, rapid and sensitive determination of hexacyanoferrate (II) (ferrocyanide) in salts by flow injection system utilizing anion exchange column as an integrated field for sensitive detection reaction and separation/preconcentration. Bulletin of the Society of Sea Water Science, 60, 352−357.
Legends of Figures Figure 1. Chromatogram of reagent grade NaCl spiked at a level of 5 mg/kg FeCN. The peak was separated by liquid chromatography detected by a photodiode array set at 221 nm; FeCN retention time ∼7.8 min.
Figure 2. Chromatograms of salt samples spiked with 5 mg/kg FeCN; (A) reworked salts; (B) burntmolten salts; (C) refined salts; (D) processed salts; (E) solar salts; and (F) other salts.
Figure 3. Chromatograms of (A) flow-through solution after the application of reagent grade NaCl spiked with 5 mg/kg of FeCN on a C18 cartridge; (B) processed salt product containing FeCN before and after the application of a C18 cartridge.
Table 1. Validation results of the analytical method.
Samples
Added standard
Precision(%RSD) Recovery (%)
(mg/kg)
Reworked salt
Burnt∙molten salt
Refined salt
Processed salt
Solar salt
Other salt
intra-day
inter-day
Linearity
LOD
LOQ
(r )
(mg/kg)
(mg/kg)
0.02
0.07
2
1
85.7±1.49
2.19
4.4
5
97.8±2.67
2.71
0.6
10
103.5±1.42
1.37
0.3
1
80.3±0.71
0.88
3.0
5
94.9±2.14
2.26
1.9
10
98.7±0.39
0.40
1.0
1
85.7±3.20
3.74
4.3
5
96.2±0.51
0.53
0.4
10
102.2±1.30
1.27
3.3
0.9999 (y=214710x10148)
1
93.9±2.97
3.2
1.3
5
92.6±4.03
4.4
1.1
10
92.9±4.00
4.3
1.2
1
87.1±1.49
1.71
3.6
5
97.8±3.56
3.64
0.7
10
92.9±1.97
2.12
1.0
1
86.1±0.74
0.74
3.4
5
95.0±1.09
1.09
0.4
10
97.7±0.88
0.88
0.4
Table 2. Expanded uncertainty of FeCN level using HPLC at 5 mg/kg in reworked salt and contributions of the individual uncertainty sources of the calibration curve (U1), recovery from salt matrix (U2), standard preparation (U3), and sample preparation (U4).
Measurement (mg/kg)
4.9 A
Combined standard uncertainty (mg/kg)
Effective degree of freedom
Expanded uncertainty (mg/kg)A
Contributions (%)
U1
U2
U3
U4
0.06
25.2
0.12
41
45
10
4
k = 2, 95% confidence level.
Table 3. Concentrations (mg/kg) and the range of FeCN in food grade salts which are classified according to the Korea Food Code. Food item
Detected No./Tested No.
Concentration (mg L−1 or kg−1)A Range
Tested
Detected
B
Reworked salt
15/42
n.d. - 7.0
1.9
5.3
Burnt∙molten salt
10/175
n.d.- 5.3
0.3
5.0
Refined salt
16/30
n.d.- 9.6
4.5
8.5
Solar sea salt
0/169
n.d.
-
-
Other salt
0/33
n.d.
-
-
Processed salt
30/352
n.d.- 8.0
0.4
4.2
A B
Detection range and mean values of tested and detected FeCN levels are shown. n.d.: not detected means below limit of detection (0.02 mg/kg).
Research Highlights
An HPLC method was developed to determine the presence of FeCN in food grade salts. The method exhibits good linearity over a range of 0.1−10 mg/L (r2 = 0.9999).
FeCN recoveries in six salt matrices ranged from 80.3 to 102.2% (RSD = 0.3−4.4%).
Validation also included measurements of selectivity, accuracy, and measurement uncertainty.
This method was successfully applied to the analysis of FeCN in commercial food grade salts.
S0308-8146(17)31223-2 http://dx.doi.org/10.1016/j.foodchem.2017.07.070 FOCH 21454
To appear in:
Food Chemistry
Received Date: Revised Date: Accepted Date:
5 April 2017 20 June 2017 14 July 2017
Please cite this article as: Soo Lim, H., Young Hwang, J., Choi, E., Lee, G., Sun Yoon, S., Kim, M., Development and validation of HPLC method for the determination of ferrocyanide ion in food grade salts, Food Chemistry (2017), doi: http://dx.doi.org/10.1016/j.foodchem.2017.07.070
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Development and validation of HPLC method for the determination of ferrocyanide ion in food grade salts
Ho Soo Lim a, Ju Young Hwang a, EunA Choi a, Gunyoung Lee a, Sang Sun Yoon a and MeeKyung Kim a, *
a
Food additives and Packaging Division, Ministry of Food and Drug Safety, Cheongju,
Chungcheongbuk-do 28159, Korea
Running title: Determination of ferrocyanide ion in food grade salts
*
Corresponding author: MeeKyung Kim, Ph.D Food additives and Packaging Division, National Institute of Food and Drug Safety Evaluation, Ministry of Food and Drug Safety, Cheongju, Chungcheongbuk-do 28159, Korea Tel: +82-43-719-4351 Fax: +82-43-719-4350 E-mail: [email protected]
Abstract
A rapid, simple, and reliable HPLC method was developed and validated to determine the presence of ferrocyanide ions (FeCNs) in food grade salts. An analytical column coupled with a guard column and mobile phase comprised of sodium perchlorate and sodium hydroxide (NaOH) were employed at a detection wavelength of 221 nm. Samples were dissolved in NaOH and filtered. For processed salts including herbs and spices, a C18 cartridge was applied to minimize interference from salt matrices. The method validation was based on linearity, selectivity, accuracy (recovery), precision, LOD, LOQ, and measurement uncertainty. This method exhibits good linearity from 0.1−10 mg/L (r2 = 0.9999). The LOD and LOQ values were determined to be 0.02 and 0.07 mg/kg, respectively. The FeCN recoveries in six salt matrices ranged from 80.3−102.2% (RSD = 0.3−4.4%). These results indicate that the proposed method is suitable for FeCN ion determination in various food grade salts.
Keywords: Ferrocyanide ion; food grade salts; method validation; HPLC
1. Introduction Sodium ferrocyanide, potassium ferrocyanide, and calcium ferrocyanide are complex cyanides that have been used as anticaking agents in food grade salts for many years (Bode et al., 2012; Gupta, Pel, Steiger, & Kopinga, 2015). These food additives are prepared from the reaction of hydrogen ferrocyanide with sodium hydroxide (NaOH), potassium hydroxide, and calcium hydroxide, respectively. They have low toxicity because of the strong chemical bond between iron and the cyanide groups. The additives were evaluated by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) and the Scientific Committee on Food (SCF). These committees established an acceptable daily intake (ADI) of 0−0.025 mg/kg of body weight (JECFA, 1974; European Commission, 1991). The use of ferrocyanides in food grade salts is authorized in the European Union, Japan, and Korea. However, in the United States, only sodium ferrocyanide is permitted in food grade salts. In Korea, these additives are used in food grade salts at levels within the legal limit, in accordance with the Korea Food Additives Code (MFDS, 2015a); the ferrocyanide ion (FeCN) content should not exceed 10 mg/kg for food grade salts, except for solar salts (MFDS, 2015b). In these salts, no FeCN should be detected. According to the food labelling regulations of the Korea Food Sanitation Act (MFDS, 2016), ferrocyanides should be labelled on the products as ‘calcium ferrocyanide, potassium ferrocyanide, or sodium ferrocyanide’. In the European Union, the maximum level of ferrocyanides in food grade salts and their substitutes should not exceed 20 mg/kg (calculated as anhydrous potassium ferrocyanide) individually or combined (European Commission, 2011). In the United States, sodium ferrocyanide should not exceed 13 mg/kg as anhydrous sodium ferrocyanide for food grade salts (US FDA, 2016). In Japan, a usage level of 20 mg/kg has been set for food grade salts (MHLW, 2016), whereas, in CODEX, ferrocyanides are set at <14 mg/kg for food grade salts and
<20 mg/kg for substitutes and sauces (CODEX, 2016). Accordingly, monitoring the ferrocyanide content in products is necessary to manage these standards, as well as to ensure food safety. Several analytical methods that determine the FeCN content in salts have been reported. These include spectrophotometry (Roberts & Wilson, 1969; Kubota, Onishi, Yomota, & Tanamoto, 2004),
infrared spectrometry (Drew, 1973), X-ray fluorometry (Koga & Niino, 2004), on-line redox derivatization high-performance liquid chromatography (HPLC) (Saitoh, Soeta, Minamisawa, & Shibukawa, 2013), and reflective colorimetry (Suzuki, Ishigaki, Oshita, Yamane, & Kawakubo, 2013). Among the cited techniques, HPLC has the advantage of being able to determine the FeCN content in salts without interference from other components. The HPLC method is also popular because of its wide usage, accuracy, and convenience. However, specific precision, limit of detection (LOD), and limit of quantification (LOQ) values from LC methods have not been reported to date. This method has been applied to determine the ferrocyanide content in only one type of food grade salt. To date, it has not been used to investigate different salts, such as processed salts. Recently, the total production and imports of food grade salts in South Korea has increased. The estimation of ferrocyanide intake is necessary because ferrocyanides (sodium ferrocyanide, potassium ferrocyanide, and calcium ferrocyanide) are used to improve the quality of food grade salts. To assess the level of their intake, method development and monitoring data must be undertaken. The official method of the Korea Food Code is problematic in that it has very low sensitivity and is largely unable to detect FeCNs in food grade salts due to the interference of sodium chloride (NaCl). Also, in the case of complex processed salts containing spice and flavour, FeCNs could not be detected. Thus, development of new methods for FeCN detection is necessary to ensure food safety through the improvement of method sensitivity. In this study, we have developed a rapid and quantitative method for the analysis of FeCN in various types of salts by HPLC coupled to a diode array detector. Method validation was performed to determine linearity, selectivity, accuracy (recovery), precision, LODs, and LOQs. In addition, measurement uncertainty was calculated in food grade salts for the first time; six types of commercial salts were analysed.
2. Materials and methods
2.1. Chemicals and reagents
Sodium ferrocyanide standard was purchased from Sigma-Aldrich (St. Louis, MO, USA). NaOH was used as a solvent in sample preparation and was purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Sodium perchlorate (NaClO4) used in the mobile phase preparation was supplied by Sigma-Aldrich. Water for all applications was obtained from a Milli-Q ultra-pure water system (Millipore, Bedford, MA, USA; resistivity ≥ 18.2 MΩ cm).
2.2. Preparation of the standard solution A FeCN stock standard solution was prepared by transferring 50 mg of sodium ferrocyanide in a 100 mL volumetric flask and adding 0.02 M NaOH solution to reach a final volume of 100 mL. The working standard solutions were produced by diluting the stock solution in 0.02 M NaOH solution to form 0.1, 0.5, 1, 5, and 10 mg/L aliquots. The stock solutions were prepared weekly and stored at 4 °C; working standard solutions were prepared daily.
2.3. Samples A total of 801 salt samples were purchased from food markets in big cities including Seoul, Daejeon, Gwangju, Daegu, Busan, Cheongju, and Sejong in South Korea. The samples were categorized into six categories according to the Korea Food Code, namely, reworked salts (42), burnt-molten salts (175), refined salts (30), processed salts (352), solar salts (169), and other salts (33). Reworked salt refers to a product made by dissolving raw salt (100%) in purified water, sea water, or concentrated sea water, and processing it through filtration, precipitation, recrystallization, dehydration, and adjustment of the salt concentration. Burnt-molten salt is a product made by modifying raw salt (100%) through burning, melting, or other methods. However, it excludes those products processed by washing, grinding, or compressing. Refined salt refers to a product of concentrated brine obtained by purifying sea water (including deep ocean water) through an ion exchange membrane, or to salt obtained by drying a solution of raw salt (100%) in water in a vacuum evaporator. Solar sea salt refers to crystals of NaCl made through natural distillation of sea water at a salt field, and to products obtained by grinding, washing, dehydrating, or drying them. Processed salt refers a product made by
adding foods or food additives to reworked salt, burnt-molten salt, refined salt or other salt (not less than 50%) and processing it. The other salts include powdered or crystallized products manufactured into by processing rock-salt or lake-salt into an edible form. Salt products included in the above salt types are excluded from this category. Six FeCN-free salts were also selected to validate the method. All samples were stored at room temperature (20−25°C).
2.4. Sample preparation Each salt sample (2−5 g) was transferred into a 50 mL volumetric flask and dissolved in a 0.02 M NaOH solution to generate a total volume of 50 mL. The sample was incubated in an ultrasonic bath for 5 min. The extract was then filtered through a 0.22 µm polyvinylidene difluoride (PVDF) membrane filter for HPLC analysis. For the processed salts, 2 mL of the extract was subjected to an enrichment procedure using a Sep-Pak C18 cartridge (Waters Co., Milford, MA, USA; 500 mg) at a flow rate of ∼3 mL/min. The cartridge was preconditioned with 5 mL methanol followed by 5 mL water. After loading the sample, the flow-through solution was collected into a 15 mL conical tube. The solution was then filtered with a 0.22 µm PVDF membrane filter.
2.5. HPLC
HPLC analysis was performed on a Shiseido Nanospace SI-2 (Shiseido Co., Tokyo, Japan) coupled to a photodiode array detector (PDA). The HPLC system included an online degasser, a binary pump, an autosampler, and a column oven. An AG11-HC (4 mm × 50 mm, 9 µm) guard column and AS11HC (4 mm × 250 mm, 9µm) analytical column were used for chromatographic separation. Both columns were purchased from DIONEX (Thermo Fisher Scientific, Sunnyvale, CV, USA). The FeCN chromatograms were analysed by Ezchrom Elite software, Shiseido Corporation. All separations were carried out isocratically at 35 °C with a mobile phase comprised of 200 mM NaClO4 and 20 mM NaOH. The flow-rate was maintained at 0.6 mL/min, and a 10 µL sample volume was injected for all experiments. The FeCN eluted from the columns was monitored by a PDA
detector set at 221 nm. The FeCN absorption spectra were recorded at wavelengths ranging from 200 to 600 nm. Peak identification was performed by comparing the retention times and absorption spectra of the samples with those of the standard solutions.
2.6. Method validation
The HPLC-UV method for the determination of FeCN in food grade salts was validated for linearity, selectivity, accuracy (recovery), precision, LOD, and LOQ according to AOAC guideline (2012). The selectivity was tested by examining the chromatograms to verify the absence of interfering peaks in the food grade salts. Matrix-matched calibration was prepared by spiking five concentrations (0.1, 0.5, 1, 5, 10 µg/mL) of FeCN in reagent-grade NaCl. The linearity was calculated using these five concentrations in triplicate. The LOD and LOQ values were calculated from calibration curves attained from each reagent-grade salt spiked with FeCN at five concentrations. These were determined at 3.3 and 10 σ/S, respectively, where σ is the standard deviation of the intercept and S is the slope of the regression line determined from the calibration curve. Accuracy (recovery) and precision (expressed as %RSD) were evaluated by analysing six samples spiked with the stock solution to afford final concentrations of 1, 5, and 10 mg/kg. Intra-day (three repetitions on the same day) and inter-day (three repetitions over three different days) precision tests were performed. Finally, the FeCN measurement uncertainty was calculated by the bottom-up method from the validation data afforded during each step of the analytical procedure (JCGM, 2008; EuraChem, 2000). Four sources of uncertainty were considered for the determination of measurement uncertainty: the calibration curve, matrix (recovery), standards preparation, and sample preparation. Thus, U (expanded uncertainty) was calculated by multiplying the combined standard uncertainty by a coverage factor (k = 2) that yields a confidence level of ∼95%.
2.7. Stability test To study storage stability, the FeCN concentrations in two salt and two solution matrices were examined. In salt matrices, 2 g of reworked salt and reagent grade NaCl were spiked at a level of 5
mg/kg of FeCN. In solution matrices, 2 g of reworked salt and reagent grade NaCl were dissolved in 0.02 M NaOH, and then FeCN was added at a final concentration of 5 mg/kg. The samples were stored at room temperature for 90 days, withdrawn according to a predefined schedule (6, 9, 12 h; 1, 3, 7, 14, 30, 60, 90 d), and stored at 4 °C until the analyses were performed. The tests were repeated in triplicate at each time point.
3. Results and discussion
3.1. Method performance To optimize the FeCN detection conditions, parameters that could influence the analytical performance were investigated. These included LC conditions, mobile phase, sample amounts and sample clarification for FeCN extraction.
3.1.1. Liquid chromatographic conditions The FeCN peak was not detectable in reagent-grade NaCl, although it was detected in the standard solution (5 mg/L FeCN in 20 mM NaOH). To acquire detectable FeCN peaks, different combinations of NaClO4 and NaOH solutions were tested as the mobile phase. A good separation of FeCN from spiked reagent-grade sodium chloride was achieved with a 200 mM NaClO4/20 mM NaOH solution mobile phase under the isocratic elution conditions described in Section 2.5 (Fig. 1). When the FeCN absorption spectra was recorded at a range of 200−400 nm, maximum absorption was observed at 221 nm. We next compared the chromatograms afforded from two ion exchange columns, Shodex IC SI90 4E (4.0 × 250 mm, 9 µm) and Dionex AS11-HC (4.0 × 250 mm, 9 µm) coupled with an AG11-HC (4 mm × 50 mm, 9 µm) guard column, by analysing the extracts of spiked reagent-grade NaCl. When only the Shodex column was used, FeCN was not detected in any of the samples. However, when the Dionex columns were used, FeCN peaks were observed in the chromatograms of spiked reagent-grade NaCl. An AS11-HC column coupled with an AG11-HC column was found to be the most suitable for determining FeCNs in food grade salts (Fig. 2). The flow rate was then increased from 0.6 mL/min to
1.2 mL/min, and a column temperature of 35 °C or 40 °C was applied; the highest FeCN peak area was afforded at a flow rate of 0.6 mL/min and a column temperature of 35 °C. At these conditions, the analysis was completed in <10 min, and the chromatograms displayed FeCN retention times of 7.8 min. Therefore, we chose these HPLC conditions for further study. The chromatograms presented in this study have not been reported in the literature to date. Recently, Saitoh et al. (2013) developed an on-line redox derivatization HPLC method to determine Fe (II) and Fe (III) cyanide complexes in food grade salts. The LC system consisted of two C18 silica columns treated with trimethylstearylammonium chloride and a small column packed with porous graphitic carbon (PGC) placed between them. In this method, the peak of the Fe (II) cyanide complex (ferrocyanide ion) was asymmetric, exhibiting tailing. In addition, since three columns were used in series, this method could be considered relatively expensive and troublesome. In our present research, the developed method, comprised of a widely used HPLC system, provided a simple analysis of FeCN using an ion exchange column with UV detection (wavelength 221 nm), and was found to be suitable for the quantification of FeCN in food grade salt.
3.1.2. FeCN extraction from food grade salts Samples (2−5 g) were employed to examine the effects of NaCl concentration on the peak area. The FeCN peak areas significantly decreased when the sample amount was increased to 5 g; 2 g salt samples provided the greatest FeCN peak areas. In the processed salts containing spices such as pepper and flavours, sample clarification is important to achieve good chromatographic peak resolution. In this study, a Sep-Pak C18 cartridge was employed to improve FeCN resolution from endogenous interfering peaks. To choose the appropriate loading volume on the C18 cartridge, we analysed the received solutions after loading each 2−5 mL sample on the C18 cartridge. The 2 mL loading volume afforded the best peak area and separation (Fig. 3A). To confirm the reliability of purification with the Sep-Pak C18 cartridge, we compared the extracts of a processed salt product containing sodium ferrocyanide and rosemary leaves before and after the cartridge was applied. When the cartridge was not applied, the FeCN peak could not be found due to interference from the other
sample components. However, a detectable peak was afforded after applying the cartridge (Fig. 3B).
3.2. Method validation The analytical method developed in this study was compared with the method of the Korea Food Code, as well as other methods. In the Korean Food Code method, the detection sensitivity of FeCN is very low, due to the influence of copious NaCl present in food grade salt. Yamane et al. (2006) showed that, in a detection method for FeCN with a flow injection system, the correlation coefficient (r2) of the calibration curve was 0.999, recoveries were 95–104%, and precision was 3% as RSD. An LOD of 0.003 ppm was presented but no LOQ was disclosed. Kubota et al. (2004) analysed FeCN in food grade salt using spectrophotometry. The correlation coefficient (r2) of the calibration curve was 0.9996, recoveries were 91.1–105.5% and precision was less than 4%. The LOD and LOQ were not presented. In addition, Koga et al. (2004) determined the presence of FeCN in a 20% NaCl solution using X-ray fluorometry. The correlation coefficient (r2) was 0.99, and the LOD was 0.1 mg/kg. Recovery, precision, and LOQ were not shown. Haba & Wilson (1963) reported a small-scale titrimetric determination of ferrocyanide. The method may be applied to amounts of 1–100 mg of the complex ion. Titration is susceptible to variation according to the amount of titrating solution, and is interfered with chloride and any other substances present in a matrix (Fennema, 1996). In the above papers, the analyses of FeCN were performed in the matrices of some food grade salts or reagent-grade NaCl and were not applied to various complex food grade salts containing processed salts. Some validation results were presented, but since the above methods utilize absorption or fluorescence, the detection of FeCN can be influenced by the matrix. Recently, Saitoh et al. (2013) introduced on-line redox derivatization HPLC; the reported correlation coefficient of the calibration curve (r2) was 0.999, but no validation results were provided. In our current study, the developed method has been established by greatly improving the conditions, including the mobile phase, column, and purification process, from the Korea Food Code method. The correlation coefficient (r2 ) of the calibration curve is 0.9999 or higher, and recoveries were 80.3–103.5% for six kinds of food grade salts, including processed salt. The precision expressed as %RSD was less than 4.5% for intra- and inter-day analyses, the LOD was 0.02 ppm, and the LOQ was 0.07 ppm. The validation results of the
method are considered to be equal to or higher than those of conventional methods. The sample pretreatment time was similar, but the instrument analysis time was relatively short (~15 min) compared to the HPLC method using on-line redox derivatization of Saitoh et al. (2013). Saitoh’s method included a similar sample pretreatment process time to our method. In the analysis of metal cyanide complexes using HPLC, the method required a long time of about 40 min. The accuracy and sensitivity of the method were not determined because the recovery, LOD and LOQ were not presented. This method is very problematic and not widely used because of the inclusion of a small column packed with porous graphitic carbon (PGC) between two C18 silica columns treated with trimethylstearylammonium chloride. Thus, our method was developed to allow for facile FeCN detection at a UV wavelength using a widely used HPLC system.
3.2.1. Linearity To support regulatory action, a method must be proven as accurate, sensitive, and able to identify an analyte with high selectivity. For this purpose, validation of analytical methods should include assessments of linearity, selectivity, accuracy (recovery), precision, LOD, LOQ, and measurement uncertainty. The linearity of the method was examined by analysing five concentrations ranging from 0.1 to 10 mg/L in spiked reagent-grade NaCl (n= 3). A calibration curve of peak area versus concentration was then plotted. The average correlation coefficient (r2) was determined to be 0.9999 (Table 1). A study using a flow injection system with an anion exchange column generated a linear FeCN calibration curve in the range of 0−0.3 ppm for FeCN in the presence of 0.5 M NaCl (Yamane, Isawa, & Osada, 2006). The coefficient value was higher than those reported in the literature (Saitoh et al., 2013; Suzuki et al., 2013).
3.2.2. LOD and LOQ The LOD and LOQ values were calculated from linear curves acquired from three replicate injections of each solution, which were prepared by spiking five known concentrations of FeCN in
reagent-grade NaCl. The LOD and LOQ values were calculated using the equations LOD = 3.3 × σ/S and LOQ=10 × σ/S where σ is the SD of the response, and S is the slope of the corresponding calibration curve. The FeCN LOD and LOQ values were calculated to be 0.02 and 0.07 mg/kg, respectively (Table 1), which were higher than those reported in the literature (Kubota et al., 2004; Koga & Niino, 2004; Suzuki et al., 2013). Two separate studies were previously carried out to determine the FeCN content in food grade salts. One study employed an LC with UV detection at 218 nm. The other consisted of an LC system equipped with a small column packed with porous graphitic carbon (PGC) treated with a redox reagent (Jang, Cho, Bae, & Kim, 2010; Saitoh et al., 2013, respectively). However, neither method reported specific LOD or LOQ values.
3.2.3. Accuracy and precision Accuracy (recovery) was reported as the percentage of FeCN recovered after spiking various known concentrations of FeCN in a blank. Recoveries were calculated by comparing the peak area of the fortified samples with that of the standard. Recovery tests were performed by spiking three known concentrations (1, 5, and 10 mg/kg) of FeCN samples into food grade salts in triplicate. In the reworked salt, burnt-molten salt, refined salt, processed salt, solar salt, and other salt matrices, the average FeCN recoveries were determined to be 85.7−103.5%, 80.3−98.7%, 85.7−102.2%, 92.6−93.9%, 87.1−97.8%, and 86.1−97.7%, respectively. The results are summarized in Table 1. In a study conducted by a portable LED-based 8-channel reflective colorimetric method (Suzuki et al., 2013), an average recovery of 96% was achieved for added FeCN at a 10 mg/kg level. This agrees with the results afforded at the 5 mg/kg level (92.6−97.8%) in this research. Precision, the degree of repeatability of an analytical method, is normally expressed as the relative standard deviation (RSD) for spiked samples. There are two types of precision: intra-day and interday precision. The former was calculated by injecting each of the three different concentrations of mixtures three times on the same day. The latter was evaluated over three different days at the same concentration levels (Thompson, Ellison, & Wood, 2002). The intra-day precision (%RSD) varied
from 0.4 to 4.4% while the inter-day precision ranged from 0.3 to 4.4%. These results demonstrate that the analytical method is sufficiently repeatable. Kubota et al. (2004) reported similar intra-day precision values for FeCN.
3.2.4.Measurement uncertainty To evaluate its suitability, the uncertainty in the results of an analytical method must be quantified to help the analyst make reasonable decisions (Rozet, Marini, Ziemons, Boulanger, & Hubert, 2011). The uncertainty measurements of FeCN in reworked salt were estimated using the proposed method. The uncertainty factors related to the analytical procedure were identified and used to calculate the combined standard uncertainty according to the International Organization for Standardization (ISO., 1993) and Eurachem/Citac (EURACHEM, 2000) guidelines. The five sources of uncertainty considered were associated with the calibration curve, the recovery from salt matrices, the standard solutions, and the sample preparation. The value for U was determined to be 2.5% (0.13 mg/kg) of the measured value (4.9 mg/kg; Table 2) based on a 95% confidence level using a coverage factor of 2. The contribution of each of the individual uncertainty sources is summarized in Table 2. Among the five sources of uncertainty, those associated with the recovery from the salt matrices and the calibration curves are the main contributors to the expanded uncertainty. Therefore, analysts should devote more attention to factors affecting the recovery and calibration curves.
3.2.5. Stability of FeCN in salt and solution matrices The quality of food grade salt depends on the stability of FeCN during its shelf life. To date, no studies on the storage stability of FeCN in food grade salts have been reported. Thus, to the best of our knowledge, this is the first reported examination of the stability of FeCN in food grade salts. The salt matrices included reworked salt and NaCl spiked at a level of 5 mg/kg; the liquid-type matrices were solutions of reworked salt and NaCl dissolved in a 0.02 M NaOH solution. For the analysis of salt matrices, each sample was withdrawn at defined time intervals and extracted in triplicate; each extract was analysed on the HPLC system. In the solution matrices, the sample
solutions were directly injected into the HPLC system after filtration with a 0.22 µm syringe filter. The analysis revealed that no decrease in FeCN concentration was observed over the time investigated (data not shown). Until the end of the studied storage period, similar FeCN concentrations were afforded in all the matrices.
3.3. Application to real samples To assess its reliability, the proposed method was used to analyse the levels of FeCN in various food grade salts purchased from local markets. A total of 801 samples were analysed and classified into six groups according to the types of salts listed in the Korea Food Code. Quantification of FeCN was carried out using standard calibration curves and confirmed by comparing the retention time and absorption spectra of the samples to those of the standards. The FeCN concentrations in the analysed samples are listed in Table 3. The results revealed that reworked salt, burnt-molten salt, refined salt, and processed salt contained FeCN. The highest detection rate was found in refined salt, followed by reworked salt, processed salt, and burnt-molten salt. Among the refined salts, the highest concentration (9.6 mg/kg) was found in a salt from the USA. The concentration levels of the detected samples were below the maximum permissible limit of 10 mg/kg stipulated in Korea. FeCN was not detected in the solar salt and other salts categories.
4. Conclusions A simple, rapid, and reliable HPLC method was developed and validated for the determination of FeCN in food grade salts. Satisfactory method validation results were afforded for FeCN after optimization of the experimental conditions. The measurement uncertainty of the entire procedure was also calculated. The stability of FeCN in the salt and solution matrices was proven under common storage conditions at room temperature (20−25 °C). The application of a C18 cartridge to the processed salts allowed for significant removal of interfering components and a good FeCN resolution towards endogenous interfering peaks. To our knowledge, this is the first study to report the analysis of FeCN in food grade salts using an HPLC-UV method. The method was also verified via the analysis of
different types of salts. Four types of salts were found to contain FeCN; the highest concentration of 9.6 mg/kg was found among the refined salts. However, none of the detected samples exceeded the maximum permissible limit of 10 mg/kg stipulated in Korea. These results demonstrate that this method shows great potential for application in the routine analysis of food grade salts.
Acknowledgements Funding: This work was supported by the Ministry of Food and Drug Safety in 2016 [16161MFDS-011].
References AOAC (Association of Official Analytical Chemistry). Guidelines for single laboratory validation of chemical methods for dietary supplements and botanicals, 2012. URL http://www.eoma.aoac.org/ app_k.pdf. Accessed 04/01/16. Bode, A. A. C., Vonk, V, van den Bruele, F. J., Kok, D. J., Kerkenaar, A. M., Mantilla, M. F., Jiang, S., Meijer, J. A. M., van Enckevort, W. J. P., & Vlieg, E. (2012). Anticaking activity of ferrocyanide on sodium chloride explained by charge mismatch. Crystal Growth & Design, 12, 1919–1924. CODEX (CODEX Alimentarius). Codex General Standard for Food Additives (GSFA), 2016. URL http://www.fao.org/fao-who-codexalimentarius/standards/gsfa/en. Accessed 30.12.16 Drew, D. M. (1973). Simultaneous determination of ferrocyanide and ferricyanide in aqueous solutions using infrared spectrometry. Analytical Chemistry, 45, 2423−2424. European Commission (1991). Reports of the Scientific Committee for Food, 25th series, pp. 18. EURACHEM (2000). Quantifying uncertainty in analytical measurement, 2nd ed., pp. 1−30. European Commission (EC) (2011). Commission Regulation (EU) No 1129/2011 of 11 November 2011 amending Annex II to regulation (EC) No 1333/2008 of the European parliament and of the council by establishing a union list of food additives. Official Journal of the European Union, L295, 1–177. Fennema, O. R. (1996). Food Chemistry. (3rd ed.). New York: Marcel Dekker.
Gupta, S., Pel, L., Steiger, M., & Kopina, K. (2015). The effect of ferrocyanide ions on sodium chloride crystallization in salt mixtures. Journal of Crystal Growth, 410, 7–13. Haba, F. R., Wilson, C. L. (1963). The small-scale titrimetric determination of complex cyanides. Microchimica Acta, 51, 305–312. Jang, M. S., Cho, S. D., Bae, D. H., & Kim, G. H. (2010). Safety and quality assessment of Kimchi made using various salts. Korean Journal of Food Science and Technology, 42, 160–164. JCGM (Joint Committee for Guides in Metrology) (2008). Evaluation of measurement-Guide to the expression of uncertainty in measurements, 1st ed., pp. 1−27. JECFA (1974). Evaluation of certain food additives, 18th Report of the Joint FAO/WHO Expert Committee on Food Additives, WHO Health Organization Technical Report Series, No. 557. Geneva World Health Organization, p. 23−24. Koga, A., & Niino, Y. (2004). Analysis of hexacyanoferrate (II) in salt. Bulletin Society of Sea Water Science, 58, 24−29. Kubota, H., Onishi, Y., Yomota, C., & Tanamoto, K. (2004). Analysis of ferrocyanides in food grade salts. Japanese Journal of Food Chemistry, 11, 32−35. MFDS (Korea Ministry of Food and Drug Safety) (2015a). Korea Food Additives Code. Seoul: Korea Food Industry Association. MFDS (Korea Ministry of Food and Drug Safety) (2015b). Korea Food Code. Seoul: Korea Food Industry Association. MFDS (Korea Ministry of Food and Drug Safety) (2016). Korea Food Sanitation Act. Seoul: Korea Food Industry Association. Miller, E. J. C., & Miller, J. N. (1993). Statistics for analytical chemistry (3rd ed.). New York: Ellis Horwood PTR Prentice Hall. Ministry of Health, Labour and Welfare (MHLW) (2016). URL. http://www.mhlw.go.jp. Accessed 30.12.16. Roberts, R. F., & Wilson, R. H. (1968). The determination of ferrocyanide and related compounds in commercial sodium chloride. Analyst, 93, 237−243.
Rozet, E., Marini, R. D., Ziemons, E., Boulanger, B., & Hubert, P. (2011). Advances in validation, risk and uncertainty assessment of bioanalytical methods. Journal of Pharmaceutical and Biomedical Analysis, 55, 848–858. Saitoh, K., Soeta, N., Minamisawa, H., & Shibukawa, M. (2013). On-line redox derivatization liquid chromatography for selective separation of Fe(II) and Fe(III) cyanide complexes using porous graphitic carbon. Analytical Sciences, 29, 715−721. Suzuki, Y., Ishigaki, M., Oshita, K., Yamane, T., & Kawakubo, S. (2013). Development of a portable LED-based 8-channel reflective colorimeter and its application to simple determination of hexacyanoferrate (II) in common salts. Bulletin of the Society of Sea Water Science, 67, 47−51. Thompson, M., Ellison, S.L.R., Wood, R. (2002). Harmonized guidelines for single laboratory validation of methods of analysis (IUPAC technical report), Pure Applied Chemistry, 74, 835–855. U.S. Food and Drug Administration (USFDA) (2016). Substances Generally Recognized as Safe (From eCFR-Code of Federal Regulations) URL http://www.ecfr.gov. Accessed on 30.12.16. Yamane, T., Isawa, M., & Osada, S. (2006). Simple, rapid and sensitive determination of hexacyanoferrate (II) (ferrocyanide) in salts by flow injection system utilizing anion exchange column as an integrated field for sensitive detection reaction and separation/preconcentration. Bulletin of the Society of Sea Water Science, 60, 352−357.
Legends of Figures Figure 1. Chromatogram of reagent grade NaCl spiked at a level of 5 mg/kg FeCN. The peak was separated by liquid chromatography detected by a photodiode array set at 221 nm; FeCN retention time ∼7.8 min.
Figure 2. Chromatograms of salt samples spiked with 5 mg/kg FeCN; (A) reworked salts; (B) burntmolten salts; (C) refined salts; (D) processed salts; (E) solar salts; and (F) other salts.
Figure 3. Chromatograms of (A) flow-through solution after the application of reagent grade NaCl spiked with 5 mg/kg of FeCN on a C18 cartridge; (B) processed salt product containing FeCN before and after the application of a C18 cartridge.
Table 1. Validation results of the analytical method.
Samples
Added standard
Precision(%RSD) Recovery (%)
(mg/kg)
Reworked salt
Burnt∙molten salt
Refined salt
Processed salt
Solar salt
Other salt
intra-day
inter-day
Linearity
LOD
LOQ
(r )
(mg/kg)
(mg/kg)
0.02
0.07
2
1
85.7±1.49
2.19
4.4
5
97.8±2.67
2.71
0.6
10
103.5±1.42
1.37
0.3
1
80.3±0.71
0.88
3.0
5
94.9±2.14
2.26
1.9
10
98.7±0.39
0.40
1.0
1
85.7±3.20
3.74
4.3
5
96.2±0.51
0.53
0.4
10
102.2±1.30
1.27
3.3
0.9999 (y=214710x10148)
1
93.9±2.97
3.2
1.3
5
92.6±4.03
4.4
1.1
10
92.9±4.00
4.3
1.2
1
87.1±1.49
1.71
3.6
5
97.8±3.56
3.64
0.7
10
92.9±1.97
2.12
1.0
1
86.1±0.74
0.74
3.4
5
95.0±1.09
1.09
0.4
10
97.7±0.88
0.88
0.4
Table 2. Expanded uncertainty of FeCN level using HPLC at 5 mg/kg in reworked salt and contributions of the individual uncertainty sources of the calibration curve (U1), recovery from salt matrix (U2), standard preparation (U3), and sample preparation (U4).
Measurement (mg/kg)
4.9 A
Combined standard uncertainty (mg/kg)
Effective degree of freedom
Expanded uncertainty (mg/kg)A
Contributions (%)
U1
U2
U3
U4
0.06
25.2
0.12
41
45
10
4
k = 2, 95% confidence level.
Table 3. Concentrations (mg/kg) and the range of FeCN in food grade salts which are classified according to the Korea Food Code. Food item
Detected No./Tested No.
Concentration (mg L−1 or kg−1)A Range
Tested
Detected
B
Reworked salt
15/42
n.d. - 7.0
1.9
5.3
Burnt∙molten salt
10/175
n.d.- 5.3
0.3
5.0
Refined salt
16/30
n.d.- 9.6
4.5
8.5
Solar sea salt
0/169
n.d.
-
-
Other salt
0/33
n.d.
-
-
Processed salt
30/352
n.d.- 8.0
0.4
4.2
A B
Detection range and mean values of tested and detected FeCN levels are shown. n.d.: not detected means below limit of detection (0.02 mg/kg).
Research Highlights
An HPLC method was developed to determine the presence of FeCN in food grade salts. The method exhibits good linearity over a range of 0.1−10 mg/L (r2 = 0.9999).
FeCN recoveries in six salt matrices ranged from 80.3 to 102.2% (RSD = 0.3−4.4%).
Validation also included measurements of selectivity, accuracy, and measurement uncertainty.
This method was successfully applied to the analysis of FeCN in commercial food grade salts.