Talanta 132 (2015) 535–540

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Rapid extraction of melamine in powdered milk for direct electrospray ionization tandem mass spectrometry analysis Elisângela do Carmo Domingo a,n, Aline Auxiliadora Tireli b, Cleiton Antonio Nunes a, Alexandre Vieira Batista a, Mário César Guerreiro b, Sandra Maria Pinto a a b

Department of Food Science, Federal University of Lavras, Lavras, Minas Gerais, Brazil Department of Chemistry, Federal University of Lavras, Lavras, Minas Gerais, Brazil

art ic l e i nf o

a b s t r a c t

Article history: Received 25 March 2014 Received in revised form 25 September 2014 Accepted 1 October 2014

A combination of a simple pretreatment for melamine extraction and direct analysis in electrospray ionization tandem mass spectrometry (ESI–MS/MS) is proposed. Three pretreatments were evaluated. The first was based on suppressing interference using acetonitrile. The second used sulphuric acid and trichloroacetic acid to suppress interference and for melamine extraction, respectively. The third used sulphuric acid to suppress milk interference, trichloroacetic acid for melamine precipitation, and ethyl acetate for melamine extraction. However, only the last pretreatment suppressed milk interference in melamine detection and a good linearity (R2 ¼0.99) was obtained. The presence of MS/MS 85 on melamine fragmentation spectrum showed the selectivity of this method. The limit of detection and limit of quantification were 0.269 mg L  1 and 0.897 mg L  1, respectively. The recoveries and relative standard deviation (RDS) of method were lower than 114% and 7.86%, respectively. Further, the research was extended to elucidate the nature of the melamine in the extract through infrared spectroscopy and microscopy analyses. The precipitate was characterized as melaminium bis(trichloroacetate) dihydrate, which is generated through hydrogen bound formation in an interaction between melamine and trichloroacetic acid. Therefore, a simple, fast, and easy method for melamine extraction and direct ESI–MS/MS analysis was developed. & 2014 Elsevier B.V. All rights reserved.

Keywords: Adulteration Melaminium bis(trichloroacetate) dihydrate Trichloroacetic acid precipitation Ethyl acetate extraction

1. Introduction Milk is a complete food and source of essential nutrients, including carbohydrate, protein, fatty acids, calcium, phosphorus and magnesium. Moreover, the milk constituents can affect dairy products performance and quality; therefore, the industries usually define milk quality based on the nutrient levels, mainly protein and fat. These parameters have been used to calculate the payout to the supplier [1]. Unfortunately, adding nonauthentic nutrients has been used to adulterate milk and, consequently, increase economic gain [2]. This adulteration is defined as the removal or replacement of milk components and addition of substances without a purchaser’s knowledge, including water, whey, sucrose, starch, salt, sodium

List of abbreviations: ESI–MS/MS, electrospray ionization tandem mass spectrometry; FDA, Food and Drug Administration; HB, hydrogen bond; HMF, 5-hydroxymethylfurfural; HTST, high temperature short time; LOQ, limit of quantification; LOD, limit of detection; LC–MS/MS, liquid chromatography tandem mass spectrometry; R2, linearity; USA, United States of America n Corresponding author. Tel.: þ 55 35 38291324. E-mail address: [email protected]fla.br (E.d.C. Domingo). http://dx.doi.org/10.1016/j.talanta.2014.10.003 0039-9140/& 2014 Elsevier B.V. All rights reserved.

hydroxide, and formaldehyde [2,3]. Melamine is a nitrogen-rich compound commonly used to manufacture plastics, laminates, glues and adhesives [4]. Although it is not an ingredient in food, it was reported as present in Chinese powdered milk and infant formula in 2008. It was used to falsely increase the apparent milk protein and was not discovered because the routinely used methods could not distinguish between nitrogen from protein and non-protein sources, which resulted in incorrectly high protein measurements [5]. Consuming tainted products by the Chinese population affected 294,000 infants and children with more than 50,000 hospitalizations and at least 6 deaths [6]. This scandal scared the world and created the need for sensitive, specific, rapid, and reliable screening methods. The Food and Drug Administration (FDA) from the United States of America (USA) established a liquid chromatography–mass spectrometry (LC–MS/MS) method to detect melamine in infant formula. This technique is accurate, but it is time consuming and requires hazardous solvent [7,8]. Many researchers have reported analytical methods for melamine detection and quantification, such as gas chromatography-single quadrupole mass spectrometry [9], a spectrophometric method [10], liquid chromatography tandem mass spectrometry [1,4,11–14], and an enzyme-linked immunosorbent assay [7]. However, these most available methods are time-consuming,

536

E.d.C. Domingo et al. / Talanta 132 (2015) 535–540

labor-intensive and expensive. Therefore, to simplify the sample preparation and improve the methods for detecting melamine, it is important to monitor contamination and prevent damage to consumer health. Recent advances in ambient ionization mass spectrometry facilitate direct, rapid, real-time, and high-throughput analyses with little or no sample pretreatment [19,20]; it has been used to characterize food and identify adulterants, and authors report it as a reliable method [21,22]. In this paper, we proposed a simple and rapid sample preparation to detect melamine in powdered milk using electrospray ionization tandem mass spectrometry (ESI–MS/MS). Further, the research was extended to infrared spectroscopy and microscopy analyses to identify the nature of the precipitate formed during extraction.

2. Material and methods 2.1. Sample and reagents Powdered milk was prepared from pasteurized milk (Grain Lait, Perdões, Brazil; high temperature short time [HTST]). Melamine was obtained from Sigma (St. Louis, USA) with chemical purity of greater than 98% and dissolved in Milli-Q purified water.

by Desmarchelier et al. [12]. The powdered milk (1 g) and melamine aliquots were placed into a falcon polypropylene tube that was mixed and allowed to stand for 10 min. The water (5 mL) and acetonitrile (5 mL) were added successively, and the resulting slurry was thoroughly mixed after each solvent was added. The slurry was further diluted with acetonitrile (30 mL) and water (10 mL) and placed onto an automated shaker for 5 min. The tube was then centrifuged at 4000g at room temperature for 10 min (Centrifuge, Beckman, GS-15R), and the liquid supernatant was used for direct injection into ESI–MS/MS. The second pretreatment was performed by adjust to pH 4.6 with 2 mol L  1 sulphuric acid (pHmeter, Quimis, Q400AS); thereafter, the samples were filtered. The filtered aliquot (5 mL) was added to 24% trichloroacetic acid (5 mL) and left to stand for 15 min. The mixture was filtered and the permeate was divided in two parts. The first aliquot of extract was used for injection into ESI–MS/MS. The remaining portion (4 mL) was added to ethyl acetate (4 mL) followed by stirring (Vortex, Phoenix-AP56); thereafter, the supernatant was used for direct injection into ESI–MS/MS (the third pretreatment sample). The extraction steps are described in Fig. 1. 2.5. ESI–MS/MS analysis

Triplicates of powdered milk were produced from pasteurized milk concentrated in a rotary evaporator (70 1C, TE-0581, TE-210, Tecnal, Piracicaba, Brasil) until the concentration of 52% of total solids; thereafter, the concentrate was dried in mini spray dryer (MSD 1.0, Labmaq, Ribeirão Preto, Brasil). The temperature for the inward and outward drying air was 160 and 100 1C, respectively, and the feed flow rate was 0.79 L h  1 in system with dual fluid nozzle atomization.

An aliquot of formic acid was added to yield 0.1% as the final concentration in the extract. Sample introduction was performed by using a micro syringe at a flow rate of 10 mL min  1. To avoid cross-interference between samples, HPLC-grade methanol (Merck, Brasil) was injected between each infusion. ESI–MS/MS analyses were performed using Agilent 1100 Series LC/MSD trap equipment in positive ion mode to identify the melamine. The spectra were obtained through 50 scans at 0.2 s intervals. The MS parameters were as follows: heat capillary temperature 350 1C, dry gas flow 6 L min  1, nebulizer pressure 15 psi, and capillary voltage 3.5 kV. Mass spectra were acquired and accumulated over 60 s.

2.3. Melamine-added milk standards

2.6. Method performance

2.2. Powdered milk preparation

1

Melamine stock solutions (500 mg L ) were prepared by dissolving melamine in water. The powdered milk was diluted in water and contaminated with melamine at different concentrations (0, 2.5, 5, 10, 15 and 20 mg L  1). 2.4. Melamine extraction Three extraction procedures were evaluated. The first pretreatment was performed in accordance with the methodology described

Milk powder + Milli-Q Water + Melamine Aliquot

Filtered

LOQ ¼

Rð%Þ ¼

TCA addition

Filtration

Protein

Filtered

ESI-MS/MS analysis

EA addition

Supernatant

sb b

ð1Þ

10sb b

ð2Þ

LOD ¼ 3

where sb is the standard deviation of the areas of the blank samples in mass spectrum, and b is the slope of the calibration curve [23]. The recovery and accuracy (%) were calculated according to Eq. (3), using blank matrices spiked with melamine at two fortification levels (8 mg L  1; 12 mg L  1).

pH reduction

Filtration

Fat and protein

The linearity (R2) was evaluated from three calibration curves with six levels of melamine concentrations in powdered milk [23]. The limit of detection (LOD) and limit of quantitation (LOQ) were detected on blank samples (n¼10) and calculated by Eqs. (1) and (2), respectively:

ESI-MS/MS analysis

Fig. 1. The second and third pretreatment steps for powdered milk. TCA—trichloroacetic acid; EA—ethyl acetate; ESI–MS/MS—electrospray ionization tandem mass spectrometry.

ðC 1  C 2 Þ  100 C3

ð3Þ

where C1 is the melamine concentration determined in adulterated milk, C2 is the melamine concentration determined in blank sample, and C3 is the concentration added [24]. The precision was calculated by the relative standard deviation (Eq. (4)). s RSD ¼  100 x

ð4Þ

where s is the standard deviation of replicate, and x is the mean of replicate. The selectivity was evaluated by the presence of melamine fragment MS/MS 85 [24].

E.d.C. Domingo et al. / Talanta 132 (2015) 535–540

2.7. Infrared spectroscopy analysis The infrared spectra were recorded on a Shimadzu FT-IR Prestige 21 spectrometer. The melamine (500 mg L  1) and trichloroacetic acid (24%) solutions were in a 1:1 ratio. The precipitate (2 mg) was dried and mixed with potassium bromide powder (200 mg) and pressed to form a thin, transparent disk. Infrared absorption spectra within the range 4000–400 cm  1 at a 2 cm  1 resolution using 32 scans were recorded at room temperature. 2.8. Optical and electronic microscopy analysis The melamine (500 mg L  1) and trichloroacetic acid (24%) solutions were at a 1:1 ratio, and the crystals images were obtained using optical microscopy (Meiji ML 5000, Meiji Techno América, Santa Clara, USA) coupled with a camera (Cle-Palmer 49901-35, Cole-Palmer, Vernon Hills, USA). The melamine (15 mg L  1) and trichloroacetic acid (24%) solutions were at a 1:1 ratio, and 60 μL of this mixture was added to a blade, to which Poli-L-Lisin was previously added. After 30 min, the Karnovsky solution was added until the blade was covered and stored at 8 1C over 3 h. The sample was washed with cocodilato buffer (0.05 M) three times over 10 min and washed with water three times. Then, the dehydration process began by adding 25, 50, 75, 90% acetone over 10 min then three times in 100% acetone for 10 min each, followed by drying in a critical point dryer. The sample was placed in the gold bath then observed in microscope (LEO Evo 040).

537

To assess melamine in powdered milk through ESI–MS/MS by suppressing milk interference, another pretreatment procedure was evaluated, which was based on eliminating milk interference; this procedure was performed in two steps. The first step consisted of eliminating the insoluble protein and fat. The decrease in pH for milk from 6.6 to 4.6 promoted protein precipitation; and filtering this mixture formed an extract without casein micelles and fat. The fat was eliminated based on the solubility difference from the extract, and the insoluble protein was eliminated based on the characteristic isoelectric points at pH 4.6 [26]. In the second step, the trichloroacetic acid reacted with melamine to form a precipitate; thereafter, it was filtered to eliminate the soluble protein and directly injection into the ESI–MS/MS. The mass spectrum for this extract is shown in Fig. 3. Melamine detection is affected by interference from milk compounds, mainly the ions from lactose and mineral interactions. The m/z 127.1 has a small intensity, which limited the fragmentation and distinction between melamine and HMF, showing that this pretreatment was not efficient to suppress the interference. Many authors have reported interference by the milk matrix during melamine detection. Ibáñez, Sancho and Hernández [27] reported that the matrix interference affected melamine detection in mass spectrometry. Xia et al. [28] reported that the matrix affected melamine determination even using hydrophilic interaction liquid

3. Results and discussion 3.1. Melamine extraction and detection Melamine, 1,3,5-triazine-2,4,6-triamine, has 126.1 Da molecular mass. Detection of this compound in milk can be affected by matrix interference because powdered milk has a compound with similar molecular mass, namely 5-hydroxymethylfurfural (HMF), which has a 126.1 Da molecular mass. HMF is a natural compound formed when milk is heated; its ESI–MS spectrum is characterized by the presence of the m/z 127.1 [21]. However, HMF fragmentation produces the MS/MS fragments 109, 84, and 71, whereas protonated melamine generated ions with m/z 110, 85, and 60, showing that fragmentation can be used to differentiate these compounds [25]. To analyse the melamine in powdered milk samples and suppress matrix interference, extraction procedures were performed. First, the extraction procedure by Desmarchelier et al. [12] was performed for samples spiked with melamine. The mass spectrum for this extract is presented in Fig. 2. The mass spectrum presented a number of m/z signals related to the presence of disaccharide from milk. Lactose is the main carbohydrate in milk and could be identified by two peaks in the spectrum. The signal detected at m/z 381 and 365 were generated from lactose association with sodium and potassium, respectively. However, the low 127.1 signal hampers the fragmentation, which does not show whether this signal is from melamine or HMF, indicating that the high lactose signal disrupts the detection; this occurred at all concentrations evaluated. The additional ions in the spectrum arose from the other powdered milk components. Desmarchelier et al. [12] successfully used this method to detect melamine through liquid chromatography, which separates matrix components from melamine and, consequently, suppresses matrix interferences. On the other hand, ESI–MS/MS does not provide proper separation, which favours matrix effects during melamine detection.

Fig. 2. The mass spectrum for the melamine extract obtained according Desmarchelier et al. [12]. The sample includes 15 mg L  1 of melamine, and the peaks show the m/z 381 [Lactoseþ Na] þ and m/z 365 [Lactoseþ K] þ .

Fig. 3. The mass spectrum for the extract obtained by precipitating melamine using trichloroacetic acid (samples with 15 mg L  1 of melamine). m/z 381 [Lactoseþ Na] þ ; m/z 365 [Lactoseþ K] þ .

E.d.C. Domingo et al. / Talanta 132 (2015) 535–540

chromatography coupled to ESI–MS. He et al. [4] observed that the melamine signal was suppressed by milk compounds. GonzalezAntuña et al. [29] described how the matrix interferes with the mass spectrometry, showing that the competition between matrix analyte compounds and modifying the physical properties of the droplets can affect analysis efficiency. Nevertheless, different solvents have been used to clean the sample for melamine detection, including acetonitrile, methanol, diethyl ether, and dichloromethane [1,9,13,14,30–34]. Thus, an additional step in the extraction procedure was performed to suppress the matrix interference. The sample was prepared over three steps; the first consisted of suppressing the insoluble protein and fat followed by melamine precipitation and ethyl acetate addition to the melamine extract. Thereafter, the supernatant was injected into an ESI–MS/MS, and the results were shown in Figs. 4 and 5. This extract produced a mass spectrum with high intensity for the m/z 127.1 (Fig. 4), which favoured fragmentation (MS/MS) (Fig. 5). As expected for the milk sample spiked with melamine, the fragmentation of m/z 127.1 confirms the presence of melamine through identification of the MS/MS 85.2 peak, which is produced through loss of the N2CH2 group [12]. Moreover, the MS/MS 99 is formed by ionization of C2N5H5þ which is produced by HNC elimination [35]. Many authors have reported melamine detection in milk and its discrimination from HMF in milk. Yang et al. [25] showed interference by HMF in melamine detection and proposed that the fragment MS/MS 85 can be used to exclude false positive signals. Dane and Cody [21] reported that the low levels of melamine may be difficult to identify in the presence of HMF. Chen and Yan [36] used capillary electrophoresis coupled with diode array detection to differentiate melamine from HMF and demonstrated melamine detection. In addition, a calibration curve was obtained using five concentrations of milk contaminated with melamine. The response range was discerned, and the calibration curve is represented by the equation shown in Fig. 6. A linear relationship was observed between the melamine concentration and absolute mass spectral intensity of ion 127, which indicated that the ESI–MS/MS can be efficiently used for melamine detection. The LOD and LOQ were 0.269 mg L  1 and 0.897 mg L  1 respectively, showing that the third extraction can be successfully applied to detect and quantify melamine in powdered milk with a limit below 1 mg L  1. The accuracy of this method was found by melamine addition in milk and quantitative analysis, showing

Fig. 4. The mass spectrum for the extract obtained by melamine precipitation with trichloroacetic acid and solubilized in ethyl acetate (sample with 15 mg L  1 of melamine) as well as m/z 127.1 fragmentation.

Fig. 5. Mass spectrum of m/z 127.1 fragmentation for the extract obtained by the third extraction technique (powdered milk contaminated with 15 mg L  1 of melamine).

70000 60000 50000

Cnts

538

40000 y = 2992.7x + 5560.3 R² = 0.9902

30000 20000 10000 0 0

5

10

15

20

25

Melamine (mg L-1) Fig. 6. The relationship between mass intensity and melamine concentration.

acceptable recovery (109 and 114%). Thus, the satisfactory data was obtained, showing the accuracy of the method developed. Cao et al. [7] reported recovery rates of 72.8–123% for melamine detection using enzyme-linked immunosorbent assay (ELISA). Yang et al. [25] using surface desorption atmospheric pressure chemical ionization to detect melamine in milk products showed 97–113% of recovery. The method precision was lower than 7.86%, showing it can be used to detect melamine in powdered milk. Thus, the extraction followed by ESI–MS/MS is a precise and accurate method to detect melamine in powdered milk. Ibáñez et al. [27] using LC–MS/MS showed good precision (RSDo 5%) for melamine detection and reported that the RSDo20% is acceptable. Cao et al. [7] showed that the precision of ELISA method was lower than 18.9%. Yang et al. [25] reported that the precision of surface desorption atmospheric pressure chemical ionization mass spectrometry to detect melamine was lower than 11.9%. The presence of MS/MS 85 on melamine fragmentation spectrum (Fig. 5) showed that the method can remove the interference of milk compounds [12]. These results show that melamine could be simply and rapidly detected by ESI–MS/MS. It can be concluded that the method had a good precision and accuracy for melamine detection. Many authors reported different methods to detect melamine and showed the sensibility of these, although most of than are timeconsuming, labor-intensive and expensive [4,7,13,14,27,28,32,33]. In this sense, the developed method is a promising technique to melamine detection since it is sensitive, simple and practical. In

E.d.C. Domingo et al. / Talanta 132 (2015) 535–540

539

addition, this method does not require a long extraction procedure and can be applied in routine analysis of powdered milk. 3.2. The nature of the melamine in the extract Melamine precipitation was based on a reaction with trichloroacetic acid, which forms of melamine salt crystals. The interaction between melamine and trichloroacetic acid facilitates crystal formation, and this complexation adds an interesting aspect to the intramolecular hydrogen bond (HB) system formed [37–41]. Thus, to investigate the nature of the precipitate, infrared spectroscopy was employed. The infrared spectrum of melamine and precipitate were shown in Fig. 7. The melamine N–H stretching vibrations absorb in an interferencefree region (3500–3300 cm  1) of the infrared spectrum; thus, they are suitable bands for HB assignment. The presence of HB moves the N–H stretching band to a lower wavenumber due to a weaker of N–H bond. The strong peak at 3419 cm  1 (melamine spectrum) is attributed to NH2 symmetric stretching (νsym), which was shifted to 3414 cm  1 in the melaminium salt, as observed by Kanagathara et al. [42], while asymmetric stretching (νasym ¼3470 cm  1) did not change the wavenumber. In addition, the side chain C–N stretching [39], which occurs at 1439 cm  1 is shifted to 1335 cm  1, which provides insight into the presence of HB during formation of the precipitate, melaminium bis (trichloroacetate) dihydrate, as reported by Perpétuo and Janczak [43] and Kanagathara et al. [42]. The melamine salt crystals are shown in Fig. 8. The melaminium bis(trichloroacetate) dihydrate crystal was detected using different microscopy techniques, which demonstrated the characteristic structure of the salt formed during the second sample pretreatment step, which consists of adding trichloroacetic acid and a reaction with melamine. However, Kanagathara et al. [42] reported a different morphology, which can be explained by the using of trichloroacetic acid at a different concentration.

4. Conclusion The first and second pretreatments were based on melamine extraction using acetonitrile and trichloroacetic acid, respectively. However, these techniques did not efficiently suppress milk interference for melamine detection. For the third pretreatment sulphuric acid was used to suppress protein and fat, and trichloroacetic acid and ethyl acetate were used for melamine precipitation and extraction, respectively. The

Fig. 8. A melaminium bis(trichloroacetate) dihydrate crystal observed using optical microscopy (A) (500 mg L  1 of melamine) and electronic (B) microscopy (15 mg L  1 of melamine).

ESI–MS/MS analysis showed that the matrix interference was removed, and a linear correlation was observed between the melamine concentration and the mass spectrometry results, showing that the adulteration can be detected and quantified. The recovery and precision were accepted. The LOD and LOQ were below than 1 mg L  1. The protocol for melamine extraction was based on melaminium bis(trichloroacetate) dihydrate formation by trichloroacetic acid addition and solubilization with ethyl acetate, which was not previously reported; and we showed that it as a simple, fast, and easy method for melamine extraction from powdered milk and direct ESI–MS/MS analysis. Future studies could use this extract for additional analytical methods to extend the sample pretreatment application.

References

Fig. 7. The near infrared spectrum for melamine and its precipitate using trichloroacetic acid.

[1] G. Abernethy, K. Higgs, J. Chromatogr. A 1288 (2013) 10–20. [2] J.C. Moore, J. Spink, M. Lipp, J. Food Sci. 77 (2012) R118–R126. [3] P.M. Santos, E.R. Pereira-Filho, L.E. Rodriguez-Saona, J. Agric. Food Chem. 61 (2013) 1205. [4] D. He, X. Zhang, B. Gao, L. Wang, Q. Zhao, H. Chen, H. Wang, C. Zhao, Food Control 36 (2014) 36–41. [5] E. Domingo, A.A. Tirelli, C.A. Nunes, M.C. Guerreiro, S.M. Pinto, Food Res. Int. 60 (2014) 131–139. [6] World Health Organization, Toxicological and Health Aspects of Melamine and Cyanuric Acid: Report of a WHO Expert Meeting, in Collaboration with FAO, 2009, from 〈http://goo.gl/ko4Sl〉. [7] B. Cao, H. Yang, J. Song, H. Chang, S. Li, A. Deng, Talanta 116 (2013) 173–180. [8] Y.T. Hsieh, W.T. Chen, I. Tomalová, J. Preisler, H.T. Chang, Rapid Commun. Mass Spectrom. 26 (2012) 1393–1398.

540

E.d.C. Domingo et al. / Talanta 132 (2015) 535–540

[9] P. Lutter, M. Savoy-Perroud, E. Campos-Gumenez, L. Meyer, T. Goldmann, M. Bertholet, P. Mottier, A. Dermarchelier, F. Monard, C. Perrin, F. Robert, T. Delatour, Food Control 22 (2011) 903–913. [10] W. Chansuvarn, S. Panich, A. Imyim, Spectrochim. Acta, A 113 (2013) 154–158. [11] A. Khedr, J. Chromatogr. B 930 (2013) 98–103. [12] A. Desmarchelier, M.G. Cuadra, T. Delatour, P. Mottier, J. Agric. Food Chem. 57 (2009) 7186–7193. [13] X. Deng, D. Guo, S. Zhao, L. Han, Y. Sheng, X. Yi, Y. Zhou, T. Peng, J. Chromatogr. B 878 (2010) 2839–2844. [14] B.N. Tran, R. Okoniewski, R. Storm, R. Jansing, K.M. Aldous, J. Agric. Food Chem 58 (2010) 101–107. [19] M.Z. Huang, S.C. Cheng, Y.T. Cho, J. Shiea, Anal. Chim. Acta 702 (2011) 1–15. [20] G.B. Savindo, J.S. Garcia, Y.E. Corilo, B.G. Vaz, J.J. Zacca, R.G. Cosso, M.N. Eberlin, M.G. Peter, J. Agric. Food Chem. 58 (2010) 9407–9412. [21] A.J. Dane, R.B. Cody, Analyst 135 (2010) 696–699. [22] L. Ming, B. Hu, J. Li, R. Chen, X. Zhang, H. Chen, Anal. Chem. 81 (2009) 7724–7731. [23] G.L. Long, J.D. Winefordner, Anal. Chem. 55 (1983) 712A–724A. [24] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019–3030. [25] S. Yang, J. Ding, J. Zheng, B. Hu, J. Li, H. Chen, Z. Zhou, X. Qiao, Anal. Chem. 81 (2009) 2426–2436. [26] P. Walstra, J.T.M. Wouters, T.J. Geurts, Dairy Science and Technology, second ed., Taylor and Francis, USA, 2006. [27] M. Ibáñez, J.V. Sancho, F. Hernández, Anal. Chim. Acta 649 (2009) 91–97. [28] J. Xia, N. Zhou, C. Zhou, B. Chen, Y. Wu, S. Yao, J. Sep. Sci. 33 (2010) 2688–2697.

[29] A. Gonzalez-Antuña, J.C. Domínguez-Romero, J.F. Garcia-Reyes, P. RodríguezGonzalez, G. Centineo, J.I.G. Alonso, A. Molina-Díaz, J. Chromatogr. A 1288 (2013) 40–47. [30] F.N. Ihunegbo, S. Tesfalidet, W. Jiang, J. Sep. Sci. 33 (2010) 988–995. [31] R.A. Yokley, L.C. Mayer, R. Rezaaiyan, M.E. Manuli, M.W. Cheung, J. Agric. Food Chem. 48 (2000) 3352–3358. [32] X.D. Pan, P.G. Wu, D.J. Yang, L.Y. Wang, X.H. Shen, C.Y. Zhu, Food Control 30 (2013) 545–548. [33] G. Koh, R.S. Chia, Q. Lin, P.S. Cheow, T.L. Teo, T.K. Lee, J. Sep. Sci. 34 (2011) 3043–3052. [34] K. Zeng, Z. Liu, Y. Ning, J. Li, J. Tang, M. Zhang, Chin. J. Chromatogr 31 (2013) 477–480. [35] Shan-Shan Ju, Chau-Chung Han, Chung-Jen Wu, Alexander M. Mebel, YitTsong Chen, J. Phys. Chem. B 103 (1999) 582–596. [36] Z. Chen, X. Yan, J. Agric. Food Chem. 57 (2009) 8742–8747. [37] M.K. Marchewka, Mater. Lett. 58 (2004) 843–848. [38] M.K. Marchewka, Mater. Sci. Eng., B 95 (2002) 214–221. [39] S. Debrus, M.K. Marchewka, M. Drozd, H. Ratajczak, Opt. Mater. 29 (2007) 1058–1062. [40] V. Sangeetha, K. Gayathri, P. Krishnan, N. Sivakumar, N. Kanagathara, G. Anbalagan, J. Cryst. Growth 289 (2014) 30–38. [41] Y. Wang, A.M. Mebel, C. Wu, Y. Chen, C. Lin, J. Jiang, J. Chem. Soc. 93 (1997) 3445–3451. [42] N. Kanagathara, N.G. Renganathan, M.K. Marchewka, N. Sivakumar, K. Gayathri, P. Krishnan, S. Gunasekaran, G. Anbalagan, Spectrochim. Acta, A 101 (2013) 112–118. [43] G.J. Perpétuo, J. Janczak, Acta Crystallogr., Sect. C: Cryst. Struct. Commun. C62 (2006) o372–o375.