Food Chemistry 138 (2013) 1742–1748

Contents lists available at SciVerse ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Analytical Methods

Determination of seven synthetic dyes in animal feeds and meat by highperformance liquid chromatography with diode array and tandem mass detectors

Tingting Zou a, Pingli He a, Amangul Yasen a, Zhen Li b,⇑a State Key Laboratory of Animal Nutrition, China Agricultural University, Beijing 100193, PR Chinab State Key Laboratory of Plant Physiology and Biochemistry, College of Biological Sciences, China Agricultural University, Beijing 100193, PR China

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 September 2012Received in revised form 12 November 2012Accepted 17 November 2012Available online 29 November 2012

Keywords:Synthetic food dyesAnimal feedMeatHPLC-DADHPLC–MS/MS

0308-8146/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.foodchem.2012.11.084

⇑ Corresponding author. Tel./fax: +8610 62731128.E-mail address: lizhenchem@gmail.com (Z. Li).

An efficient method was developed for the simultaneous determination of seven commonly used syn-thetic sulfonate dyes (Ponceau 4RC, Sunset yellow, Allura red, Azophloxine, Ponceau xylidine, Erythrosineand Orange II) in animal feed and meat using high performance liquid chromatography (HPLC-DAD) andtandem mass spectrometry (HPLC–MS/MS). Ethanol–ammonia–water (80:1:19, V/V/V) solution was usedas extract solution, which can extract target species while reducing interference from the sample matri-ces. The recoveries of these 7 dyes in animal feed and chicken meat were between 71% and 97% with rel-ative standard deviations less than 14.8%. HPLC–MS/MS was employed as a further means ofconfirmation to assure accuracy of the results. Limits of detection for these dyes were in the range of0.02–21.83 ng mL�1. The proposed method can be applied to confirmative screening of seven commonlyused food colorants in feed and meat samples.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Food colorants are an important class of food additives. They arewidely used in drinks, juices, meat products and sweets to preserveor restore the natural color of food products and enhance appeal(Vidotti, Cancino, Oliveira, & Rollemberg, 2005). Natural food dyeshave been used more and more for consumer preference, however,synthetic food dyes are still widely used in food and feed industrybecause of their low cost and high stability. Most of the syntheticdyes show good resistance against degradation and pose littlethreat to human and animal health. But some of these substancesand their metabolites pose potential health risk to human beingsand may even be carcinogenic, especially when they are consumedin excessive amounts (Robens et al., 1980). These adverse healtheffects include allergy and asthmatic reaction (Ibero, Eseverri,Barroso, & Botey, 1982; Miller, 1982; Settipane et al., 1976), DNAdamage (Combes & Haveland-Smith, 1982; Sasaki et al., 2002),hyperactivity (McCann et al., 2007; Rowe & Rowe, 1994) and carci-nogenesis (JECFA., 1975) etc. Therefore, the use of synthetic dyes infoodstuff is strictly controlled by legislation throughout the world(EC., 1994; GB2760-2011, 2011). In Japan, all ingredients includingfood colorants are required to be listed on the package label(Yoshioka & Ichihashi, 2008). In China, the maximum amountallowed for most colorants is no more than 100 mg kg�1

(GB2760-2007, 2007). In Europe, the maximum level allowed for

ll rights reserved.

Allura red used in luncheon meat and breakfast sausage with aminimum cereal content of 6% is 25 mg kg�1, the maximum levelallowed for Ponceau 4RC used in jam and jellies is 100 mg kg�1

(EC, 1994). Thus, it is necessary to develop accurate and reliableanalytical methods for the confirmative determination of syntheticfood dyes in foodstuffs of various matrices to ensure food safetyand consumer health.

Various analytical methods have been reported for determiningsynthetic food dyes from soft drinks, juices, fruit jellies, candies,edible animal tissues and spices. Such methods include: capillaryelectrophoresis (Huang, Shih, & Chen, 2002; Prado, Boas, Bronze,& Godoy, 2006), thin layer chromatography (Oka, Ikaia, Kawamura,Yamada, & Inoue, 1987; Oka et al., 1994), electrochemistry (Com-beau, Chatelut, & Vittori, 2002; Ni, Bai, & Jin, 1996), spectropho-tometry (Dos Santos, Demiate, & Nagata, 2010; Soylak, Unsal, &Tuzen, 2011; Unsal, Soylak, & Tuzen, 2012) and ion-pair chroma-tography (Fuh & Chia, 2002). With the fast development of detec-tion technology, high performance liquid chromatography (HPLC)with ultraviolet/visible (UV/Vis) and diode-array detectors (DAD)(Yoshioka & Ichihashi, 2008; Minioti, Sakellariou, & Thomaidis,2007) and liquid chromatography–mass spectrometry (LC–MS)(Feng et al., 2011) also have been applied to the detection of syn-thetic dyes in foodstuffs. Feng et al. established a sensitive screen-ing method for 40 dyes in soft drinks by liquid chromatography–electrospray tandem mass spectrometry. LC–MS based methodsshowed significant improvement in sensitivity and specificity com-pared with traditional methods, but it has strict requirements insample pretreatment due to interference and signal suppression

T. Zou et al. / Food Chemistry 138 (2013) 1742–1748 1743

from the sample matrices, also the instrument is very expensive. Incontrast, HPLC-DAD is very popular for qualitative and quantitativedetermination with excellent precision, accuracy and lower cost,which can be much more practical and economical in detectingnon-illict additives such as food colorants.

Most of the reported LC based methods focused on the determi-nation of food dyes in liquid samples, like soft drink and fruitdrinks, or water soluble samples like fruit jelly, jam and candies.However, these methods are not suitable for determining dyes insolid food matrix. This is due to the high protein content in sam-ples like meat, poultry products and animal feed. The high proteinand salt content of such samples may induce severe interferencefor LC–UV–Vis or LC–MS measurements, and lead to low sensitivityand specificity of the developed methods. In this paper, wedeveloped an HPLC-DAD method for the determination of sevencommonly used food dyes in high protein content matrices (feedand chicken meat), HPLC–MS was used as a confirmational stepto assure accuracy of the results.

The ethanol–ammonia extraction method developed here pro-vides an efficient and convenient way of removing protein, starchand other interferences from the sample. The extract solution canbe analyzed directly with LC-DAD and LC–MS, thus, eliminate thecomplicated SPE cleanup step, which is required for most LC–MSmethods. Detection sensitivity in the low ppb range can beachieved for both detection methods. With the developedmethod, we analyzed dye-spiked complete feed and poultry meatsamples, recovery and accuracy of the method were alsoevaluated.

2. Experimental

2.1. Reagents and chemicals

Orange II (94%), Ponceau 4RC (75%) and Allura red (80%) werepurchased from the Dr. Ehrenstorfer GmbH Company (Augsburg,Germany). Azophloxine (P96%), Ponceau xylidine (P96%), Eryth-rosine (P96%) and Sunset yellow (P96%) were purchased fromFluka (Buchs, Switzerland) (Table 1). HPLC grade methanol, ethanoland acetonitrile were obtained from Fisher Scientific International(Hampton, NH). HPLC grade formic acid and ammonium acetatewere purchased from Dikma Technology (Richmond Hill, Canada).All other reagents were analytical grade. Ultra-pure water with aresistance of 18.2 MX cm�1 was purified using a Milli-Q system(Millipore, Bedford, USA). Nylon membrane filters (0.22 lm) wereobtained from Whatman (Maidstone, UK).

2.2. Apparatus

The HPLC-DAD method was developed using an Agilent 1200HPLC system with binary gradient pump, auto-sampler, tempera-ture controlled column oven and DAD detector (Agilent Technolo-gies, Fermont, CA). HPLC–MS/MS analysis was performed on anAgilent 1260 HPLC system coupled with an Agilent 6460 triplequadrupole mass spectrometer.

2.3. Preparation of standard solutions

Standard stock solutions of each dye (1 mg mL�1) were preparedby dissolving the dyes in pure water with the exception of sunsetyellow which was dissolved in 20 mM ammonium acetate–metha-nol (1:1, V/V). The standard solutions were stored in darkness at�20 �C until use. Mixed standard stock solution was prepared bymixing and diluting the standard stock solution of each dye withpure water to a final concentration of 100 lg mL�1 and stored at4 �C. Standard working solution was prepared daily by diluting

the mixed standard stock solution with ethanol–ammonia–water(80:1:19, V/V/V) to appropriate concentrations. The pH of standardworking solutions was adjusted to 5.0 with formic acid. For HPLC-DAD analysis, the concentrations of standard working solutionranged from 100 to 5000 ng mL�1; for HPLC–MS/MS analysis, theconcentrations ranged from 0.1 to 2000 ng mL�1. All the workingsolutions were prepared fresh before use.

2.4. Sample collection and preparation

Complete feed samples were prepared at the pilot mill of theMinistry of Agricultural Feed Industry Center (Beijing, China). Poul-try meat (chicken) was purchased from local supermarket. Beforeextraction, the feed samples were pulverized by a grinder, andsieved through a No. 60 mesh; the chicken meat samples werehomogenized in a homogenizer for 5 min (Da Kang, Tianjin, China).

2.5. Sample extraction

Dye-spiked samples were prepared in a 50 mL conical flask bymixing 2 g of complete feed sample or homogenized chicken meatsample with a series of the 7-dye mixed standard solutions at var-ious concentrations. Sample extraction was achieved by adding10 mL ethanol–ammonia–water (80:1:19, V/V/V) extraction solu-tion and stirring for 30 min. The extraction solution was centri-fuged at 12,000 rpm for 10 min at 4 �C to remove protein. ForHPLC-DAD analysis, 2.5 mL of the supernatant was evaporated todryness in 50 �C water bath under nitrogen stream. The residuewas reconstituted with 0.5 mL ethanol–ammonia–water (80:1:19,V/V/V). The solution was adjusted to pH 5.0 with formic acid andpassed through a membrane filter (0.22 lm). Twenty micro litersof the solution was injected into the HPLC-DAD system. By doingso, the dyes were concentrated 5 times in the final injected samplecompared with the extraction supernatants. For HPLC–MS/MSanalysis, the extraction supernatant was adjusted to pH 5.0 withformic acid and passed through a 0.22 lm membrane, then 10 lLof the solution was injected for detection. Blank samples were pre-treated in the same manner.

2.6. HPLC conditions

Chromatographic separation of the dyes mixture was achievedon a Waters Atlantis T3 column (2.1 mm � 150 mm, 3 lm). Thecolumn temperature was set at 30 �C. Mobile phase consisted A:20 mM ammonium acetate and B: acetonitrile. A gradient programwas used for elution: 5% solvent B (initial), 5–40% solvent B (from 0to 4 min), 40–90% solvent B (from 4 to 9 min) and 90% solvent B(from 9 to 11 min). After 11 min, the ratio was reduced to 5% sol-vent B. A 5-min equilibration was necessary before the next injec-tion, so the total run time was 16 min. The mobile phase wasdelivered at a flow rate of 0.2 mL min�1. The diode-array detectorwas set to monitor the 7 dyes at two pre-selected wavelengths:484 and 526 nm.

2.7. HPLC–MS/MS conditions

Waters Atlantis dC18 column (2.1 � 150 mm, 5 lm) was usedfor LC–MS/MS confirmative analysis. The gradient elution programwas a little different from LC-DAD experiments: 5% solvent B (ini-tial), 5–40% solvent B (from 0 to 5 min), 40–90% solvent B (from 5to 9 min) and 90% solvent B (from 9 to 11 min). After 11 min, theratio was reduced to 5% solvent B, the flow rate was 0.3 mL min�1.The triple quadrupole tandem mass spectrometer operated undermultiple reaction monitor mode (MRM) for quantitative and qual-itative analysis. Negatively charged ion species from the 7 dyeswere monitored. The optimized electrospray ionization conditions

Table 1Chemical structures and tandem mass spectrometry parameters of the 7 dyes tested.

Compound Chemical structure Retention time (min) Molecular weight (g/mol) Quantitative transition m/z Qualitative transition m/z Fragmentor voltage (V) Collision energy (V)

Ponceau 4RC 4.9 604.5 536.7/301.9 536.7/301.9 130 10536.7/428.9 130 20

Sunset yellow 5.2 452.0 406.8/206.9 406.8/206.9 170 30406.8/326.9 170 25

Allura red 5.6 496.4 472.8/194.8 472.8/194.8 190 35472.8/199.8 190 35

Azophloxine 5.7 509.0 463.9/343.8 463.9/344.1 120 30463.9/359.1 120 25

1744T.Zou

etal./Food

Chemistry

138(2013)

1742–1748

Ponceauxylidine

6.3 480.4 434.9/355.0 434.9/355.0 120 30434.9/274.1 120 25

Erythrosine 7.1 879.6 834.1/662.8 834.1/662.8. 150 35834.1/536.8 150 35

Orange II 7.6 350.3 327.0/171.1 327.0/171.7 150 25327.0/155.8 150 35

T.Zouet

al./FoodChem

istry138

(2013)1742–

17481745

1746 T. Zou et al. / Food Chemistry 138 (2013) 1742–1748

were: gas temperature 350 �C, gas flow 10 L min�1, sheath gastemperature 400 �C, sheath gas flow 12 L min�1 and capillary volt-age 3500 V. Detailed MRM settings were listed in Table 1.

3. Results and discussion

3.1. Optimization of sample pretreatment

The development of an efficient sample extraction procedure iscritical for accurate determination of dyes in animal feed andmeat samples. The complex matrices of the samples pose chal-lenges for the sample extraction procedure, which has to be effec-tive in removing interfering matrices while retaining the highestsensitivity and specificity. Four extract solvents were evaluatedin this experiment based on chemical properties of the 7 dyes(Table 1). The extract solvents include: ethanol–ammonia–water80:1:19, (V/V/V), 80:5:15, (V/V/V), 80:10:10, (V/V/V) and metha-nol–ammonia–water 80:1:19, (V/V/V). Ethanol–ammonia–water80:1:19, (V/V/V) and 80:10:10, (V/V/V) showed similar extractionefficiency in terms of recovery and reproducibility for the 7 dyes(Fig. 1). Most of the interference proteins present in feed and meatsamples were precipitated in the presence of high percentage eth-anol, and they can be removed efficiently by the subsequentrefrigerated centrifugation procedure. Afterwards, pH of thesupernatant solutions was adjusted to approximate 5.0 with for-mic acid, so that it is compatible with the HPLC mobile phase. Eth-anol–ammonia–water 80:1:19 (V/V/V) was chosen as the finalextract solvent for the 7 dyes because it’s much easier to adjustthe pH of supernatant to pH 5.0 with less ammonia in thesolution.

In this experiment, further purification of the extract solutionwith solid-phase extraction was also considered, Waters OasisHLB (hydrophilic–lipophilic-balanced reversed-phase, 3 cc,60 mg) and Waters Oasis MAX (mixed-mode anion exchange sor-bent, 3 cc, 60 mg) cartridge were tested. With the extra SPE purifi-cation step, the chromatograms showed little improvement,impurities observed during the first 2 min of LC elution cannotbe removed with SPE. Since all analytes of interest eluted after4 min, the existence of these low retention interferences did not af-fect the detection. Moreover, these 7 dyes possess different chem-ical characteristics, especially ER, which has distinct chemicalstructure from the other dyes, no single SPE column can achievesatisfactory retention and recovery performance for each singledye. For these reasons, in the developed method, SPE was notadopted as it is not necessary, and without SPE, the whole pretreat-ment process is very convenient and effective.

Fig. 1. Recoveries of each dye from complete feed using four different extractsolutions: (a) ethanol–ammonia–water (80:1:19, V/V/V); (b) ethanol–ammonia–water (80:5:15, V/V/V); (c) ethanol–ammonia–water (80:10:10, V/V/V); (d) meth-anol–ammonia–water (80:1:19, V/V/V). 4RC, Ponceau 4RC; SY, Sunset yellow; AR,Allura red; AZ, Azophloxine; PX, Ponceau xylidine; ER, Erythrosine; OG, Orange II.

3.2. HPLC-DAD method development

The chemical structures of the 7 dyes studied are shown inTable 1. Here we chose Atlantis T3 column due to its excellentresolving power for middle to high polar compounds. The flow ratewas set at 0.2 mL/min, as a higher flow rate resulted in poor sepa-ration between Ponceau 4RC and Sunset yellow. Optimal UV–Vissignals were obtained using scan mode of the DAD detector, mostof the dyes showed good sensitivity at 526 nm, but for Sunset yel-low, the wavelength was set at 484 nm for enhancing signal inten-sity. The chromatogram of mixed standard solution showed clearbaseline separation of the 7 dyes in a single 16-min LC run(Fig. 2A).

Linearity, limit of detection, precision and recovery were deter-mined to evaluate the validity of the method. Linearity was studiedby analyzing mixed standard working solutions of the 7 dyes atseveral concentrations ranging from 100 to 5000 ng mL�1 in HPLC.Most of the 7 dyes showed satisfactory linearity within the concen-tration range of 500–5000 ng mL�1 (R2 > 0.996). Limit of detection(LOD), which was defined as the concentration at 3 times the signalintensity of noise, was determined by examining HPLC-DAD chro-matograms of the extract solution of the dye-spiked complete feedsamples. The LOD values were in the range of 6.43–74.81 ng mL�1

(Table 2).The precision of the method was evaluated for intra- and inter-

day precision. The intra-day RSD for 7 dyes in spiked feed sampleranged from 3.0% to 19.0%, and the inter-day RSD ranged from0.7% to 8.6%. The intra-day RSD was higher than inter-day RSD be-cause some of the dyes showed degradation at room temperatureduring the extended period of analysis of 5 replicates. The overallprecision and accuracy of the method were sufficient for the quan-tification of the 7 dyes in real samples.

The established method was applied to recovery analysis ofdye-spiked blank feed and meat samples to validate this methodfor routine monitoring of animal feed and meat samples. Dyeswere spiked in complete feed at 1 and 5 mg kg�1, and poultry meatat 0.5 and 1 mg kg�1. Six replicates were tested for each concentra-tion. Chromatograms for complete feed sample spiked with 7 dyemixture (1 mg kg�1) and a blank feed sample were shown inFig. 2B. The clean baseline of the chromatogram indicates thedeveloped sample extraction method can extract dyes from thesample effectively while bringing in minimum interference fromthe sample matrix. Calibration curves obtained from external stan-dards were used to calculate recovery of the developed method.The recoveries for spiked feed samples ranged from 71% to 94%

Fig. 2. Chromatograms of HPLC-DAD: (A) the 7-dye mixed standard; (B) completefeed sample spiked with 7-dye mixed standard (1 mg kg�1) and blank feed sample.(1) Ponceau 4RC; (2) Sunset yellow; (3) Allura red; (4) Azophloxine; (5) Ponceauxylidine; (6) Erythrosine; (7) Orange II.

Table 2Linear relationships, sensitivities and LODs for the detection of 7 dyes using HPLC-DAD and HPLC–MS/MS.

Compound HPLC-DAD HPLC–MS/MS

Linear Range(ng mL�1)

Linear equation R2 LOD(ng mL�1)

Linear range(ng mL�1)

Linear equation R2 LOD(ng mL�1)

Ponceau 4RC 500–5000 Y = 76.018x + 35.811 0.993 22.01 5.0–1000 y = 49.073x � 537.22 0.997 6.49Sunset yellow 100–1000 Y = 13.146x + 1.67 0.999 74.81 1.0–1000 y = 78.37x � 400.03 0.999 2.18Allura red 500–5000 Y = 68.8x � 12.9 0.998 68.02 2.0–100 y = 95.469x + 114.81 0.999 14.49Azophloxine 500–5000 Y = 128.12x + 21.238 0.996 19.97 3.0–1500 y = 72.931x + 740.79 0.999 19.87Ponceau xylidine 500–5000 Y = 222.85x � 32.005 0.998 14.74 2.0–2000 y = 105.49x + 4900.3 0.980 21.83Erythrosine 500–5000 Y = 482.67x � 9.8918 0.999 6.43 0.2–200 y = 936.78 + 364.14 0.999 0.10Orange II 500–5000 Y = 353.62x � 2.8836 0.999 8.57 0.1–100 y = 1550.7x + 1391.9 0.999 0.02

Table 3Percentage recoveries (%) and relative standard deviations (%) of the 7 dyes in different matrices (n = 6).

Compound HPLC-DAD HPLC–MS/MS

Feed Poultry meat Feed Poultry meat

1 mg kg�1 5 mg kg�1 0.5 mg kg�1 1 mg kg�1 1 mg kg�1 5 mg kg�1 0.5 mg kg�1 1 mg kg�1

Ponceau 4RC 83(7.8) 91(10.6) 87(10.5) 92(14.8) 53(9.4) 63(10.8) 108(8.8) 59(3.6)Sunset yellow 90(2.5) 72(7.0) 94(6.3) 94(2.0) 57(4.3) 51(3.6) 94(15.8) 74(7.2)Allura red 94(3.2) 90(4.9) 97(4.7) 85(1.2) 106(9.5) 85(9.4) 113(19.3) 76(9.9)Azophloxine 88(2.5) 81(4.4) 94(3.1) 84(0.7) 86(14.5) 83(5.8) 92(13.9) 78(6.9)Ponceau xylidine 87(2.2) 71(1.6) 91(1.8) 83(0.8) 102(9.7) 65(4.0) 83(6.4) 82(6.8)Erythrosine 84(4.1) 83(3.8) 84(4.0) 78(4.0) 73(13.8) 88(8.9) 81(11.5) 79(5.6)Orange II 90(3.4) 86(2.7) 93(0.9) 89(0.2) 103(9.1) 87(4.7) 100(7.3) 89(7.0)

T. Zou et al. / Food Chemistry 138 (2013) 1742–1748 1747

and RSDs were less than 10.6%. The recoveries for spiked chickensamples ranged from 78% to 97%, and RSDs were less than 14.8%(Table 3).

3.3. HPLC–MS/MS analysis

3.3.1. Optimization of HPLC–MS/MS conditionsThe complexity of the sample matrix may induce interference in

HPLC-DAD measurements, so HPLC-DAD alone is not adequate forconfirmative determination of low-level additives in food and feedsamples. HPLC–MS/MS was used as a confirmative step to furtherconfirm the existence of the dyes in the sample matrices.

Fig. 3. Chromatograms of HPLC–MS: (A) the 7-dye mixed standard; (B) chickensample spiked with 7-dye mixed standard (0.5 mg kg�1) and blank chicken meatsample; (Inset) quantitative product ion chromatograms of the 7 dyes. (1) Ponceau4RC; (2) Sunset yellow; (3) Allura red; (4) Azophloxine; (5) Ponceau xylidine; (6)Erythrosine; (7) Orange II.

Most of the dyes contain a sulfonate group in their molecularstructure, and the most abundant precursor ions of the dyes weretheir negatively charged sodium adduct ions [M�Na]� or[M�2Na]2�, these ions were chosen as the precursor ions forMRM. Characteristic product ions were obtained by ramping thecollision energy; the two most abundant product ions of the spe-cies were selected for MRM confirmation with the most abundantproduct ion used for quantitation. Optimal MS parameters werelisted in Table 1 and quantitative product ion chromatograms ofthe 7 dyes were shown in Fig. 3 (Inset). Since, the colorants showeddegradation at room temperature, in order to obtain comparabledata from LC-DAD and LC–MS/MS experiments, samples must berun simultaneously on both instruments. A Waters dC18 columnwas used here in addition to the T3 column used in LC-DAD anal-ysis. The relatively old dC18 column showed slightly lower separa-tion efficiency, so Allura red and Azophloxine were co-eluted fromthe column, but with different MRM transition, these two com-pounds could still be distinguished. Chromatograms for the 7-dyemixed standard was shown in Fig. 3A, extract ion chromatogramfor chicken meat spiked with the 7-dye mixture (0.5 mg kg�1)and a blank chicken sample were shown in Fig. 3B.

3.3.2. HPLC–MS/MS method validationLC–MS based method showed better sensitivity, most of the 7

dyes showed satisfactory linearity within the concentration rangeof 0.1–1000 ng mL�1 (R2 > 0.996), and limit of quantitation (LOD)was in the range of 0.02–21.83 ng mL�1 (Table 2). Intra-day RSDsfor 7-dye mixed standard ranged from 2.0% to 16.2%, and the in-ter-day RSDs ranged from 0.1% to 2.6%. The intra-day RSDs for se-ven colorants in spiked samples ranged from 5.7% to 17.7%, and theinter-day precision ranged from 4.3% to 14.5%.

The actual spiked quantity and the measured concentration infeed and meat matrices showed good consistency (Table 3). Therecoveries for spiked feed samples ranged from 51% to 106% andrelative standard deviations were less than 14.5%. The recoveriesfor spiked chicken samples ranged from 59% to 113% and relativestandard deviations were less than 15.8%. The HPLC–MS/MS results

1748 T. Zou et al. / Food Chemistry 138 (2013) 1742–1748

were in good accordance with HPLC-DAD, which proved the accu-racy of the method developed.

4. Conclusion

In summary, the very simple and effective HPLC-DAD methodwith ammonia–ethanol extraction provides satisfactory specificityand sensitivity for the determination of seven commonly used sul-fonate dyes in feed and meat samples. Meanwhile, HPLC–MS/MS,which served as a further confirmative step, showed good sensitiv-ity under complex sample matrix situation and assured accuracy ofthe method developed. Analysis of dye-spiked samples with thesimple extraction proctol developed here showed that the recover-ies ranged from 71% to 97%, which can meet regular analyticalrequirements. With miniature and portability of new LC instru-ments, this method could provide a way for onsite monitoring ofcolorants in feed and animal products.

Acknowledgments

Financial supports from the Key Projects in The National Sci-ence & Technology Pillar Program during the eleventh five-yearplan (2009BADB7B07), Agro scientific Research in the Public Inter-est (201203088), and the National Natural Science Foundation ofChina (31170343) are gratefully acknowledged.

References

Combeau, S., Chatelut, M., & Vittori, O. (2002). Identification and simultaneousdetermination of Azorubine, Allura red and Ponceau 4R by differential pulsepolarography: Application to soft drinks. Talanta, 56, 115–122.

Combes, R. D., & Haveland-Smith, R. B. (1982). A review of the genotoxicity of food,drug and cosmetic colours and other azo, triphenylmethane and xanthene dyes.Mutation Research, 98, 101–243.

Dos Santos, M. E., Demiate, I. M., & Nagata, N. (2010). Simultaneous determinationof tartrazine and sunset yellow in food by spectrophotometry UV–VIS andmultivariate calibration methodology. Ciencia E Tecnologia De Alimentos, 30,903–909.

EC. (1994). Directive of the European Parliament and of the council 94/36/EC of June30, 1994 on colours for use in foodstuffs. Official Journal, L237, 13 (10/9/1994).

Feng, F., Zhao, Y. S., Yong, W., Sun, L., Jiang, G. B., & Chu, X. G. (2011). Highly sensitiveand accurate screening of 40 dyes in soft drinks by liquid chromatography–electrospray tandem mass spectrometry. Journal of Chromatography B, 879,1813–1818.

Fuh, M. R., & Chia, K. J. (2002). Determination of sulphonated azo dyes in food byion-pair liquid chromatography with photodiode array and electrospray massspectrometry detection. Talanta, 56, 663–671.

GB2760-2007 (2007). Hygienic standards for the use of food additives. Ministry ofHealth of the People’s Republic of China.

GB2760-2011 (2011). Hygienic standards for the use of food additives. Ministry ofHealth of the People’s Republic of China.

Huang, H. Y., Shih, Y. C., & Chen, Y. C. (2002). Determining eight colorants in milkbeverages by capillary Electrophoresis. Journal of Chromatography A, 959,317–325.

Ibero, M., Eseverri, J. L., Barroso, C., & Botey, J. (1982). Dyes, preservatives andsalicylates in the induction of food intolerance and/or hypersensitivity inchildren. Allergologia et immunopathologia (Madrid), 10, 263–268.

JECFA (1975). Amaranth. In WHO food additives series (vol. 8). Geneva: World HealthOrganization.

McCann, D., Barrett, A., Cooper, A., Crumpler, D., Dalen, L., Grimshaw, K., et al.(2007). Food additives and hyperactive behaviour in 3-year-old and 8/9-year-old children in the community: A randomised, doubleblinded, placebo-controlled trial. Lancet, 370, 1560–1567.

Miller, K. (1982). Sensitivity to Tartrazine. British Medical Journal, 285, 1597–1598.Minioti, K. S., Sakellariou, C. F., & Thomaidis, N. S. (2007). Determination of 13

synthetic food colorants in water-soluble foods by reversed-phase high-performance liquid chromatography coupled with diode-array detector.Analytica Chimica Acta, 583, 103–110.

Ni, Y., Bai, J., & Jin, L. (1996). Simultaneous adsorptive voltammetric analysis ofmixed colorants by multivariate calibration approach. Analytica Chimica Acta,329, 65–72.

Oka, H., Ikaia, Y., Kawamura, N., Yamada, M., & Inoue, H. (1987). Simple method forthe analysis of food dyes on reversed-phase thin-layer plates. Journal ofChromatography, 411, 437–444.

Oka, H., Ikaia, Y., Ohno, T., Kawamura, N., Hayakawa, J., Harada, K., et al. (1994).Identification of unlawful food dyes by thin-layer chromatography-fast atombombardment mass spectrometry. Journal of Chromatography A, 674, 301–307.

Prado, M. A., Boas, L. F. V., Bronze, M. R., & Godoy, H. T. (2006). Validation ofmethodology for simultaneous determination of synthetic dyes in alcoholicbeverages by capillary electrophoresis. Journal of Chromatography A, 1136,231–236.

Robens, J. F., Dill, G. S., Ward, J. M., Joiner, J. R., Griesemer, R. A., & Douglas, J. F.(1980). Thirteen-week subchronic toxicity studies of Direct Blue 6, Direct Black38, and Direct Brown 95 dyes. Toxicology and Applied Pharmacology, 54,431–442.

Rowe, K. S., & Rowe, K. J. (1994). Synthetic food coloring and behavior: A doseresponse effect in a double-blind, placebo controlled, repeated-measures study.Journal of Pediatrics, 125, 691–698.

Sasaki, Y. F., Kawaguchi, S., Kamaya, A., Ohshita, M., Kabasawa, K., Iwama, K., et al.(2002). The comet assay with 8 mouse organs: results with 39 currently usedfood additives. Mutation Research, 519, 103–119.

Settipane, G. A., Chafee, F. H., Postman, I. M., Levine, M. I., Saker, J. H., Barrick, R. H.,et al. (1976). Significance of Tartrazine sensitivity in chronic urticaria ofunknown etiology. Journal of Allergy and Clinical Immunology, 57, 541–546.

Soylak, M., Unsal, Y. E., & Tuzen, M. (2011). Spectrophotometric determination oftrace levels of allura red in water samples after separation andpreconcentration. Food and Chemical Toxicology, 49, 1183–1187.

Unsal, Y. E., Soylak, M., & Tuzen, M. (2012). Column solid-phase extraction of sunsetyellow and spectrophotometric determination of its use in powdered beverageand confectionery products. International Journal of Food Science & Technology,47, 1253–1258.

Vidotti, E. C., Cancino, J. C., Oliveira, C. C., & Rollemberg, M. C. E. (2005).Simultaneous determination of food dyes by first derivativespectrophotometry with sorption onto polyurethane foam. Analytical Sciences,21, 149–153.

Yoshioka, N., & Ichihashi, K. (2008). Determination of 40 synthetic food colors indrinks and candies by high-performance liquid chromatography using a shortcolumn with photodiode array detection. Talanta, 74, 1408–1413.