《现代食品工程高新技术》课程教学资源（文献资料）超临界流体萃取技术 Application of ultra-high performance supercritical fluid chromatography for the determination of carotenoids in dietary supplements

Contents lists available at ScienceDirect Journal of Chromatography A ELSEVIER journal homepage:www.elsevier.com/locate/chroma Application of ultra-high performance supercritical fluid chromatography for the determination of carotenoids in dietary supplements Bing Li,Haiyan Zhao*,Jing Liu,Wei Liu,Sai Fan,Guohua Wu,Rong Zhao ARTICLE INFO ABSTRACT rticle histo Aquick and critical fluid ved52015 1 November 2015 ts The influ ed fo 5 of stati bet 2015 nn (150 m 18um by gradie tra-high per ormance supercritical flui the Dietary suppiement ne limits of det he imits of in the 15 Elsevier B.V.All rights reserved 1.Introduction the prevention of prostate can Astaxanthin Carotenoids are natural fat-soluble pigments present in oreceived incre g attention in re ent ye ars.This is becaus vity allow them to pertorm a and dieta equently red UV-lie ms,acute are also obs d wit degenerative diseases3 Furthermore,a-carotene,B-carotene wing to the beneficial effects of these compounds.aninreas nte s of cons detection or dio carotenoid found in and n(DAD)is widely ment o studies h allows much bet separation of carotenoid structural isomers gao@163.com(H.Zhao 6%7o8g9o2mm

Journal of Chromatography A, 1425 (2015) 287–292 Contents lists available at ScienceDirect Journal of Chromatography A jou rn al h om epage: www.elsevier.com/locate/chroma Application of ultra-high performance supercritical fluid chromatography for the determination of carotenoids in dietary supplements Bing Li, Haiyan Zhao∗, Jing Liu, Wei Liu, Sai Fan, Guohua Wu, Rong Zhao Institute of Nutrition and Food Hygiene, Beijing Center for Disease Control and Prevention, Beijing 100013, China a r t i c l e i n f o Article history: Received 15 July 2015 Received in revised form 1 November 2015 Accepted 8 November 2015 Available online 14 November 2015 Keywords: Carotenoids Ultra-high performance supercritical fluid chromatography Dietary supplement a b s t r a c t Aquick andsimpleultra-highperformance supercriticalfluidchromatography-photodiode arraydetector method was developed and validated for the simultaneous determination of 9 carotenoids in dietary supplements. The influences of stationary phase, co-solvent, pressure, temperature and flow rate on the separation of carotenoids were evaluated. The separation of the carotenoids was carried out using an Acquity UPC2 HSS C18 SB column (150 mm × 3.0 mm, 1.8 m) by gradient elution with carbon dioxide and a 1:2 (v:v) methanol/ethanol mixture. The column temperature was setto 35 ◦C and the backpressure was 15.2 MPa. Under these conditions, 9 carotenoids and the internal standard, -apo-8 -carotenal, were successfully separated within 10 min. The correlation coefficients (R2) of the calibration curves were all above 0.997, the limits of detection for the 9 carotenoids were in the range of 0.33–1.08 g/mL, and the limits of quantification were in the range of 1.09–3.58 g/mL. The mean recoveries were from 93.4% to 109.5% at different spiking levels, and the relative standard deviations were between 0.8% and 6.0%. This method was successfully applied to the determination of 9 carotenoids in commercial dietary supplements. © 2015 Elsevier B.V. All rights reserved. 1. Introduction Carotenoids are natural fat-soluble pigments present in plants, algae, and microorganisms. Humans cannot synthesize the carotenoids necessary to maintain normal health, and must acquire them through food and dietary supplements [1,2]. Epi￾demiological studies have found that increased carotenoid intake can reduce the incidence of several chronic conditions, such as cancer, heart disease, vascular disease, retinal diseases, and degenerative diseases [3]. Furthermore, -carotene, -carotene, and -cryptoxanthin are important for humans as they function as vitamin A precursors [4]. The carotenoids lutein and zeaxanthin are found abundantly in human retina macula lutea, and a growing number of studies have indicated that they protect the retina by filtering out blue light, and play a significant role in reducing the risk of visual loss from age-related macular degeneration [5]. Lycopene is a representative carotenoid found in tomatoes, and has attracted increasing research interest over the last decade owing to the link between lycopene-rich foods or supplements and ∗ Corresponding author. E-mail address: gongxiaotougao@163.com (H. Zhao). the prevention or treatment of prostate cancer [6]. Astaxanthin and fucoxanthin are two major marine carotenoids that have also received increasing attention in recent years. This is because their superior antioxidant activity allows them to perform as free radical scavengers and counteract oxidative stress processes, consequently reducing the risk of cardiovascular problems, acute inflammation, and UV-light damage [7,8]. Antioxidant properties are also observed with canthaxanthin, which has been used as a component of tanning pills and as food coloring [9,10]. Owing to the beneficial effects of these compounds, an increas￾ing numbers of carotenoid dietary supplements have appeared on the market in recent years. Therefore, methods for the fast and accurate determination of carotenoids in dietary supplements are urgently required in order to ensure their quality and protect the interests of consumers. HPLC coupled with UV–Vis absorbance detection or diode array detection (DAD) is widely used for the measurement of carotenoids in dietary supplements [11–14]. C18 and C30 columns are those most commonly used. Many studies have reported that the C30 column is more suitable for carotenoid determination and allows much better separation of carotenoid structural isomers, such as lutein and zeaxanthin, and -carotene and -carotene [15–17]. Ultra-high performance liquid chromatography (UHPLC) http://dx.doi.org/10.1016/j.chroma.2015.11.029 0021-9673/© 2015 Elsevier B.V. All rights reserved.

283 B.Liet al/j.Chromatogr.A1425(2015)287-292 B-carotene a-carotene lycopene wwwo 人人次 fucoxanthin canthaxanthin lutein zeaxanthin B-cryptoxanthin Fig 1.Chemical structures of nine c ids in the present study and b 2.Experimental (all9)B-cy arotene fo -B-Ano-8'-ca e988 crdifesivieyre acity nalytical Louis.MO.USA)Il ts were of or better.U tive to the and sho 22.Standardsolution preparation f each ati -sol ents increa the nature and prop o of olvents [261 s oleum ether fo canthaxan in:chloroform for IS) calcu on of s f t concentra 2710 for a-c otene:453nm 259 for B-carotene:40m.210 caro h250 of the ared in DCM-than )ere high performance supercritical fluid chromatography coupled to injected for purity determination in the UHPSFC system.Purity of

288 B. Li et al. / J. Chromatogr. A 1425 (2015) 287–292 Fig. 1. Chemical structures of nine carotenoids in the present study. with sub-2 m particle-size columns is generally considered offering shorter analysis times, smaller peak widths and higher chromatographic resolution compared to conventional HPLC [18]. However, until now, there are no commercially available UHPLC C30 stationary phase columns. Bijttebier et al. compared the separation of complex carotenoid mixtures between the HPLC C30 column and different UHPLC columns. The study found that the overall performance of HPLC C30 column was better than all the tested UHPLC columns in the carotenoids separation. However, the main drawbacks of this method were the long analysis times [18]. Long analysis times not only mean large solvent volume usage which is not environmentally friendly, but also are not suitable for the determination of unstable compounds. Supercritical fluid chromatography (SFC) is a suitable alter￾native method for carotenoids determination. CO2 is the most used mobile phase in SFC. Above its critical pressure (PC = 7.3 MPa) and temperature (TC = 31 ◦C) the supercritical fluid has lower vis￾cosity and higher diffusivity relative to conventional liquids, and higherdensity anddissolving capacity compared withconventional gases. These unique properties make SFC an effective analytical method for thermally unstable and involatile compounds. Rela￾tive to the corresponding HPLC separation, SFC allows higher flow rates and shorter run times. Furthermore, due to less organic sol￾vent consumption, this technology is considered to be a green approach [19–25]. Owing to the low polarity of CO2, SFC is effi- cient for the separation of hydrophobic compounds. In addition, the SFC mobile phase is also very flexible, and organic co-solvents, such as methanol, can be added to change its polarity. However, the addition of these co-solvents increase critical parameters to higher pressure and temperature figures which are depending on the nature and proportion of the co-solvents [26], such as when the mobile phase CO2-MeOH 70:30 was used, critical parameters increased to 135 ◦C and 16.8 MPa [27]. Therefore, the majority of today’s separations are actually not performed under a supercriti￾cal state but rather in subcritical conditions (with P > PC and T < TC) [26,27]. The addition of polar cosolvents enhances the fluid solvent strength, enabling simultaneous determinationof different polarity carotenoids [28,29]. Recent technological advancements that employed sub-2 m particles in UHPLC have been transferred to ultra-high performance supercritical fluid chromatography (UHPSFC) [30]. The aim of this work was to develop a rapid and reliable method using ultra￾high performance supercritical fluid chromatography coupled to photodiode arraydetector (PDA) and sub-2 mparticle columns for the determination of 9 selected carotenoids, including lycopene, - carotene, -carotene, -cryptoxanthin, astaxanthin, fucoxanthin, canthaxanthin, lutein and zeaxanthin in dietary supplements. 2. Experimental 2.1. Chemicals and materials -Carotene, -carotene, lutein, zeaxanthin, fucoxanthin (all ≥95.0%), -cryptoxanthin, astaxanthin (≥97.0%), lycopene (≥90.0%), canthaxanthin (≥98.8%), and the internal standard (IS) trans-Apo-8 -carotenal (≥96.0%) were purchased from Sigma–Aldrich (St. Louis, MO, USA). The chemical structures of the 9 carotenoids are shown in Fig. 1. Methanol and ethanol were purchased from Fisher Scientific (Pittsburgh, PA, USA). Dichloromethane (DCM), and dimethyl sulfoxide (DMSO) were purchased from Dikma Scientific (Lake Forest, CA, USA). Butylated hydroxytoluene (BHT) was purchased from Sigma–Aldrich (St. Louis, MO, USA). All reagents were of HPLC-grade or better. Ultra￾pure water of purity ≥18.0 -cm was prepared using a milli-Q system (Millipore, Millford, MA, USA). 2.2. Standard solution preparation Stock solutions of each carotenoid and IS were prepared in DCM + 0.1% BHT. The concentrations of these solutions were deter￾mined using a Shimadzu – 2600 UV–Vis spectrophotometer. Before measuring absorbance, the stock solutions were diluted in suit￾able solvents (hexane for -carotene, -carotene, lycopene and astaxanthin; ethanol for fucoxanthin, -cryptoxanthin, lutein and zeaxanthin; petroleum ether for canthaxanthin; chloroform for IS). The wavelengths of UV/visible absorbance detection and extinction coefficients used for the calculation of the exact concentrations of the standards are as follows: 471 nm, 3450 for lycopene; 445 nm, 2710 for -carotene; 453 nm, 2592 for -carotene; 470 nm, 2100 for astaxanthin; 466 nm, 2200 for canthaxanthin; 452 nm, 1601 for fucoxanthin; 452 nm, 2350 for -cryptoxanthin; 445 nm, 2550 for lutein; 450 nm, 2540 for zeaxanthin and 461 nm, 2640 for IS. Absorbances of the solutions were between 0.200 and 0.800, and were recorded in triplicate. Individual working solutions of each carotenoid and IS prepared in DCM-ethanol 1:2 (v/v) were injected for purity determination in the UHPSFC system. Purity of

B.Li et al/J.Chromatogr.A1425(2015)2s7-292 289 L-AA 6 BEH CSH FP BEH 2-EP DIOL 人 68 DEA 2-PIC ton:1)ycopene.2】a-carotene.3B-Caro (5)canth hin.(8)lutei d (9)zca a.The oncenua Wa Milford.MA).En y3132 ontrol a employed for th 2.3.Samples inearelutiongradi ent was0min.+B:4min.8A+15 +10%B:15.0min.90%A+ he flow rate was 1. Adiet 2.4.Extraction procedure 3.Results and discussion en, 3.1.The influence of stationary phase noid co nds.On the basis of s in their for 60min oiadsotenoid added. s sonicated fo thus non lar.B lutein,z xanthin.astaxan h layer was transferred into to a vial for analysis. s a-carotene a h 2.5.Instrumental and chromatographic conditions rote noidsdifcult ested for the column oven. an automated arot high strength silica bo

B. Li et al. / J. Chromatogr. A 1425 (2015) 287–292 289 Fig. 2. Chromatograms of nine carotenoids on eight different columns: (A) HSS C18 SB, (B) 1-AA, (C) BEH, (D) CSH FP, (E) BEH 2-EP, (F) DIOL, (G) DEA, (H) 2-PIC; mobile phase: (A) CO2 and (B) methanol/ethanol (1:2, v/v); the gradient: 0 min, 90% A+ 10% B; 4 min, 85% A+ 15% B; 9 min, 75% A+ 25% B; 10 min, 75% A+ 25% B; 10.5 min, 90% A+ 10% B; 15.0 min, 90% A+ 10% B; flow rate: 1.0 mL/min; column temperature: 35 ◦C; backpressure 15.2 MPa. Identification: (1) lycopene, (2) -carotene, (3) -carotene, (4) astaxanthin, (5) canthaxanthin, (6) fucoxanthin, (7) -cryptoxanthin, (8) lutein, and (9) zeaxanthin. a carotenoid was expressed as the peak area of that carotenoid as a percentage of the total area of the chromatogram. The concentra￾tion calculated was corrected accordingly [31,32]. 2.3. Samples Dietary supplements (tablet or capsule) containing a single carotenoid or a mixture were purchased in local drugstores or from internet stores. A dietary supplement used as a blank sample (without analyzed carotenoids) contained vitamin D, as stated on its package. 2.4. Extraction procedure Prior to analysis, test samples were stored at 4 ◦C in the dark. Thirty tablets were ground into a fine powder using a mill (IKA, Staufen, Germany), and ca. 1.5 g of sample was weighed. Three hun￾dred microliters of 1.25 mg/ml of the IS and 5 mL DMSO:water (3:1, v/v; 0.1% BHT, w/v) were added. The samples were mixed thor￾oughly for 60 s using a vortex mixer and placed in an ultrasonic bath for 60 min at 37 ◦C to dissolve the coating. The mixture was shaken vigorously by hand every 5 min. Then, 25 mL of DCM-ethanol (1:2, v/v; 0.1% BHT, w/v) was added, and the mixture was sonicated for further 30 min. The samples were shaken for 1 h using a mechanical shaker at a speed of 200 rpm, after which the samples were cen￾trifuged for 10 min at 10,000 rpm at 4 ◦C. An aliquot of the upper layer was transferred into to a vial for analysis. 2.5. Instrumental and chromatographic conditions UHPSFC-PDA analysis was performed using an Waters Acquity UPC2 system consisting of a binary solvent deliv￾ery pump, an auto-sampler, a column oven, an automated backpressure regulator, and photodiode array detector (all from Waters Corp., Milford, MA). Empower software (version 3) was used for instrument control and data acquisition. An ACQUITY UPC2 HSS C18 SB (150 mm × 3.0 mm, 1.8 m) from Waters was employed for the separation at 35 ◦C. The binary mobile phase was composed of (A) CO2 and (B) methanol/ethanol (1:2, v/v). The linear elution gradient was 0 min, 90% A+ 10% B; 4 min, 85% A+ 15% B; 9 min, 75% A+ 25% B; 10 min, 75% A+ 25% B; 10.5 min, 90% A+ 10% B; 15.0 min, 90% A+ 10% B. The flow rate was 1.0 mL/min. The backpressure was set at 15.2 MPa. Sample injection volume was 1.0 L. The PDA detector was operated from 210 nm to 520 nm with compensation from 540 to 600 nm. Each peak was quantified at the wavelength of maximum absorption. 3. Results and discussion 3.1. The influence of stationary phase Carotenoids are polyisoprenoid compounds. On the basis of the presence of oxygen atoms in their molecules, carotenoids are divided into two groups, i.e., oxygenated carotenoids (named xanthophylls) and hydrocarbon carotenoids (named carotenes) [1–3]. -Carotene, -carotene, and lycopene are carotenes, and are thus non-polar. -Cryptoxanthin, lutein, zeaxanthin, astaxan￾thin, fucoxanthin, and canthaxanthin are xanthophylls, and are thus polar. Carotenoids have a very similar structure. In addition, some are structural isomers such as -carotene and -carotene, and lutein and zeaxanthin, have the same chemical formulae, the only difference between them being the double bond position of the terminal ring. This structural similarity makes the separation of the carotenoids difficult. Eight stationary phases were tested for the separation of nine carotenoids including: (1) Waters ACQUITY UPC2TM HSS C 18 SB (high strength silica bonded with C18), (2) ACQUITY UPC2TM Torus

290 B.Li et aL/1.Chromatogr.A 1425(2015)287-292 4 A METHANOL B METHANOL-FTHANOL C METHANOL-ETHANOL LA 、 oi(: thin (8- Linear rangeμgm) LOD (Hg/mL) LOQ(ug/mL) 470-20 B-Crypto phenyl group).(5)Waters ACQUITY UPC diethylamin bo ding).(ACQUITY mwith the e of The nts of all chromatograms of pure standards of nine tation a. Except on the colums.non-polar carotenoids UHPSFCnd mobile phase in UHPLC made the speed were all improved by UHPSFC.These findings demonstrated Astaxanthin and B-cryptox nthin Table2 and relative standard deviations(RSDs)of four carotenoids in samples added ramt B-Carote 1077 Astaxanthin id Lutein Fucoxanthin

290 B. Li et al. / J. Chromatogr. A 1425 (2015) 287–292 Fig. 3. The influence of co-solvents (methanol, methanol–ethanol 1:1, and methanol–ethanol 1:2) on the separation of carotenoids on HSS C18 SB column. Mobile phase: (A) CO2 and (B) methanol/ethanol(1:2, v/v);the gradient: 0 min, 90% A+ 10% B; 4 min, 85% A+ 15% B; 9 min, 75% A+ 25% B; 10 min, 75% A+ 25% B; 10.5 min, 90% A+ 10% B; 15.0 min, 90% A+ 10% B; flow rate: 1.0 mL/min; column temperature: 35 ◦C; backpressure 15.2 MPa. Identification: (1) lycopene, (2) -carotene, (3) -carotene, (4) astaxanthin, (5) canthaxanthin, (6) fucoxanthin, (7) -cryptoxanthin, (8) lutein, and (9) zeaxanthin. Table 1 Validation parameters for 9 carotenoids. Analyte Linear range (g/mL) Correlation coefficient r2 LOD (g/mL) LOQ (g/mL) Lycopene 4.70–47.0 0.999 0.94 3.15 -Carotene 4.66–46.6 0.999 0.80 2.66 -Carotene 4.32–43.2 0.997 0.95 3.18 Astaxanthin 5.82–174.6 0.998 1.08 3.58 Canthaxanthin 2.70–27.0 0.999 0.77 2.55 Fucoxanthin 3.74–56.1 0.999 0.64 2.15 -Cryptoxanthin 2.82–70.4 0.999 0.66 2.20 Lutein 1.31–39.3 0.998 0.33 1.09 Zeaxanthin 1.63–40.8 0.998 0.42 1.39 1-AA (hybrid silica with 1-aminoanthracene bonding), (3) Waters ACQUITY UPC2TM BEH (hybrid silica without bonding), (4)Waters ACQUITY UPC2TM CSH FP (charged surface hybrid silica bonded with a fluoro phenyl group), (5) Waters ACQUITY UPC2TM BEH 2-Ethylpyridine (hybrid silica with a 2-ethylpyridine bonding), (6) ACQUITY UPC2TM Torus DIOL (hybrid silica with high density diol bonding), (7) ACQUITY UPC2TM Torus DEA (hybrid silica with diethylamine bonding), (8) ACQUITY UPC2TM Torus 2-PIC (hybrid silica with 2-picolylamine bonding). All selected UHPSFC column dimensions were 3.0 mm (inner diameter) and 150 mm (column length), particle sizes were 1.7 m with the exception of 1.8 m for HSS C18 SB column. The same chromatographic conditions described in Section 2.5 were applied to the retention behavior experiments of all eight columns, so any difference observed in separation and retention could be solely ascribed to the stationary phase effects. Fig. 2 shows chromatograms of pure standards of nine carotenoids and IS, respectively, obtained with eight columns. Except on the HSS C18 SB and 1-AA columns, non-polar carotenoids lycopene, -carotene, and -carotene could not be identified on the other six columns as they eluted at the same time. Conversely except the CSH FP column which was needed further optimization (Rs of 1.06 for the lutein and zeaxanthin), the different polarity columns provided the baseline separation of the polar structural isomers lutein and zeaxanthin. On the DIOL column, lutein co￾eluted with fucoxanthin. Astaxanthin and -cryptoxanthin were difficult to resolve when the DEA column was used. On the CSH FP column, no baseline separation of lutein, zeaxanthin and astax￾anthin was achieved. The BEH, BEH 2-EP and 2-PIC column could separate all the polar carotenoids -cryptoxanthin, lutein, zea￾xanthin, astaxanthin, fucoxanthin, and canthaxanthin very well although non-polar carotenoids lycopene, -carotene, and - carotene were not separated from each other. The HSS C18 SB and 1-AA columns separated all carotenoids very well. All carotenoids were baseline-separated and the changes of elution orders between the two columns were observed. These findings demonstrated that the stationary phase in UHPSFC had strong impact on the selectiv￾ity of separation [21]. The HSS C18 SB column was selected in the present study. The 1-AA column also provided promising separa￾tions. Further developments on this column will be described in a future paper. Bijttebier et al. developed a method for the separation of carotenoids by an UHPLC with a ACQUITY UPLC HSS C18 SB column [18]. In the reversed-phase UHPLC, in order to obtain better sepa￾ration, complicated mobile phase (A: 50:22.5:22.5:5 water + 5 mM ammonium acetate:methanol:acetonitrile:ethylacetate, B: 50:50 acetonitrile:ethyl acetate) wasused. Thepolar carotenoids astaxan￾thin, zeaxanthin, lutein, canthaxanthin were eluted firstly and the non-polar carotenoids lycopene, -carotene, and -carotene had stronger retentions. In our study, when the carotenoids separated by UHPSFC with the HSS C18 SB column, the non-polar carotenoids lycopene, -carotene, and -carotene were eluted firstly and later eluting compounds were the polar carotenoids. The same station￾ary phase in conjunction with a subcritical fluid mobile phase in UHPSFC and a aqueous/organic mobile phase in UHPLC made the elution orders of non-polar and polar carotenoids clearly alter. In comparison to the study of Bijttebier et al., resolution and analysis speed were all improved by UHPSFC. These findings demonstrated Table 2 Recoveries and relative standard deviations (RSDs) of four carotenoids in samples (n = 6). Compound Concentration added (g/mL) Recovery (%) %RSD -Carotene 4.80 99.2 2.8 9.60 107.1 3.8 Astaxanthin 19.4 102.9 3.5 38.8 103.9 1.0 Lutein 4.37 109.5 6.0 8.74 100.6 3.2 Fucoxanthin 6.23 93.4 3.4 12.5 99.7 0.8

.山etaL/Chromate0g.A1425(2015)287-29 291 enoids in real dietary supplements Sample Carotenoid content (mean SD)mg/g a-Carotene B-Carotene B-Cryptoxanthin Lutei Lycopen Astaxanthin Fucoxanthin Canthaxanthin 2944990 2614904 0±0.08 887±007 16±0.05 580±0.41 ND 0o000000 0211+a004 deviation from the analysis of three independently prepared sample that UHPSFC could be a powerful technique for tuning the selectiv- ity of UHPLO 3.2.The infuence ofco-solvent B-carotene.astaxanthin. fucoxa thin.and lutein The rec CO2 has low polarity.Thus.in order to separate different polar fourcarote attwo ntration. he recovery result n3 and 109.with SD values lowe 6.0%are used as the Isandadded at the egnn ids lycor o-caroter sample ed.diuiuontoelin ate error during dilution.Onc carotenoids asta axanthin.fucoxanthin.and canthaxanthin were not lyte conce he- 3.5.Determination of carotenoids in real samples was used. nd c 0 and fucoxanthin increased to and34.resp ively.How s in some dietary supplemen vere high.some nTable3.B-Cryptoxanthin and were no n the tested sample 4.Conclusions AUHPSFC method coupled with sub-2 um particle-size column 3.3.The influence ofpressure.temperature and fow rate the separation of dition were .The results otained demonstrate that UHPSFC is ure isomers o-caroteneandcarotand which were dare wel ortcictcminaionotarotenoiding 的embgmicwenientmcswereobanedwh aroten 3.4.Method validation References ibration curves were constructed by plotting the peak area 2RK (each analyte/S)versus each analyte at six concentration levels

B. Li et al. / J. Chromatogr. A 1425 (2015) 287–292 291 Table 3 Results of UHPSFC analysis of carotenoids in real dietary supplements. Sample Carotenoid content (mean ± SD) mg/ga -Carotene -Carotene -Cryptoxanthin Lutein Zeaxanthin Lycopene Astaxanthin Fucoxanthin Canthaxanthin Sample1 ND 0.229 ± 0.020 ND 0.048 ± 0.002 0.650 ± 0.042 1.50 ± 0.08 0.041 ± 0.003 ND ND Sample2 ND ND ND 2.94 ± 0.17 2.61 ± 0.11 ND ND ND ND Sample3 ND ND ND ND ND ND ND 15.0 ± 0.5 ND Sample4 ND ND ND 26.0 ± 0.3 8.87 ± 0.07 ND ND ND ND Sample5 ND ND ND ND ND 14.1 ± 0.9 ND ND ND Sample6 ND ND ND ND ND 7.80 ± 0.07 ND ND ND Sample7 ND 1.66 ± 0.05 ND 5.80 ± 0.41 ND ND ND ND ND Sample8 ND 3.48 ± 0.21 ND ND ND ND ND ND ND Sample9 0.211 ± 0.004 1.37 ± 0.02 ND 0.038 ± 0.002 0.032 ± 0.022 ND ND ND ND Sample10 ND 0.178 ± 0.006 ND 0.565 ± 0.019 0.689 ± 0.029 ND ND ND ND ND, not detected. a SD, standard deviation from the analysis of three independently prepared samples. that UHPSFC could be a powerful technique for tuning the selectiv￾ity of UHPLC. 3.2. The influence of co-solvent CO2 has low polarity. Thus, in order to separate different polar carotenoids, polar organic co-solvents are need to be added. The low viscosity and high polarity of methanol make it the most commonly used organic co-solvent in SFC [29], so methanol was initially chosen as the co-solvent. Fig. 3A shows that HSS C18 SB column separated the non-polar carotenoids lycopene, -carotene, and -carotene very well with methanol. However, the more polar carotenoids astaxanthin, fucoxanthin, and canthaxanthin were not sufficiently resolved. In order to obtain adequate separation of the polar carotenoids, the lower polarity solvent ethanol was added to the co-solvent to improve retention. The result shown in Fig. 3B indicates that when a 1:1 (v/v) mixture of methanol and ethanol was used, all carotenoids show complete baseline separation, the resolutions of astaxanthin and canthaxanthin, and canthaxanthin and fucoxanthin increased to 2.46 and 2.34, respectively. How￾ever, when the amount of ethanol was increased to a ratio of 1:2, the lower polarity of the 1:2 methanol-ethanol enhanced the retention of astaxanthin, canthaxanthin and fucoxanthin, the res￾olutions are slightly higher (Rs of 3.16 and 2.81 for the astaxanthin and canthaxanthin, canthaxanthin and fucoxanthin, respectively), (Fig. 3C). 3.3. The influence of pressure, temperature and flow rate In supercritical conditions,the density of mobile phase is impor￾tant for retention and the density can be strongly influenced by pressure and temperature [26]. But with the addition of co￾solvent (5–45%), the effects of pressure and temperature on eluting strength were dramatically lessened [27]. A variety of different back pressures (11.0, 13.8, 15.2, 17.9 and 19.3 MPa), temperatures (20 ◦C, 25 ◦C, 30 ◦C, 35 ◦C) and flow rates (0.8 mL/min, 1.0 mL/min, 1.5 mL/min) were tested to evaluate the influences on retention times and separation selectivity for the nine carotenoids selected in this study. Under the tested conditions, all carotenoids are well￾separated, no substantial change in selectivity of separation was observed and slightly shorter retention times were obtained with the increase of pressure and flow rate. 3.4. Method validation Method validation was performed by evaluating linearity, limit of detection (LOD), limit of quantification (LOQ), and recovery. Cal￾ibration curves were constructed by plotting the peak area ratio (each analyte/IS) versus each analyte at six concentration levels, and the resulting correlationcoefficient(R2 > 0.997) was considered satisfactory (Table 1). Four carotenoids were selected for the evaluation of recovery: -carotene, astaxanthin, fucoxanthin, and lutein [28]. The recov￾ery was performed by spiking a blank supplement sample with the four carotenoids attwo levels of concentration. The recovery results and the spiking levels are listed in Table 2. Satisfactory recoveries between 93.4% and 109.5% with RSD values lower than 6.0% are achieved. The LODs (s/n = 3) and LOQs (s/n = 10) are listed in Table 1. -Apo-8 -carotenal was used as the IS and added at the beginning of the sample preparation to eliminate error during dilution. Once the IS was added, dilution volume is no longer critical because ana￾lyte concentration is normalized to the IS [33]. For identification of carotenoids, the retention times and spectra were used. 3.5. Determination of carotenoids in real samples The developed method was applied to the determination of 10 commercially available dietary supplements. Because the lev￾els of carotenoids in some dietary supplements were high, some extracts were diluted before analysis. The obtained results are given in Table 3. -Cryptoxanthin and canthaxanthin were not detected in the tested samples. 4. Conclusions A UHPSFC method coupled with sub-2 m particle-size columns to determine 9 carotenoids in dietary supplements was devel￾oped and validated. 8 types of commercially available columns were investigated for the separation of carotenoids. By using 1:2 methanol/ethanol as a co-solvent, the HSS C18 SB and 1-AA columns separated all carotenoids very well. The optimized chro￾matographic method was able to separate 9 carotenoids and IS in just 10 min. The results obtained demonstrate that UHPSFC is powerful for the separation of carotenoids, especially for the struc￾ture isomers -carotene and -carotene, and lutein and zeaxanthin which were difficult to resolve due to the similarities in their struc￾tures. This method was shown to be simple, rapid, and selective for the determination of carotenoids in dietary supplements. We believe that this analytical method will be very useful to the rou￾tine analysis of carotenoids in dietary supplements, and willfurther the quality of these products and the protection of consumers. References [1] J.V. Woodside, A.J. McGrath, N. Lyner, M.C. McKinley, Carotenoids and health in older people, Maturitas 80 (2015) 63–68. [2] R.K. Saini, S.H. Nile, S. Park, Carotenoids from fruits and vegetables: chemistry, analysis, occurrence, bioavailability and biological activities, Food Res. Int. 76 (2015) 735–750.

B.Li et aL/1 Chromatogr.A 1425 (2015)287-292 310P. operties of carotenoids in human health. .T.Uchikata.M.Sh S.Nishi aki, 1mn eot mumcot et 55 coo7 V.J D.W.Hu 013) they.D ltra-hig carotenoids and andarsandcommtiecial5plemeis. 24 0172-1 ayha,R. ,D 126 is,C.West,E Les er.D.G 12]o 2列 e vitamins and care oids in sta aphy. 0g.A1382(2015)2-46 1281 E.H.Wavsek.Method for the deter 29 1301 0222 I16C 132 nd hleCe2 (12) Produce.1 AOAC Int 86 (2003)411 ination of pesti

292 B. Li et al. / J. Chromatogr. A 1425 (2015) 287–292 [3] O.P. Conrad, M.Y. Gan, Functional properties of carotenoids in human health, Int. J. Food Prop. 10 (2007) 201–230. [4] A.V. Rao, L.G. Rao, Carotenoids and human health, Pharmacol. Res. 55 (2007) 207–216. [5] S. Beatty, M. Boulton, D. Henson, H.-H. Koh, I.J. Murray, Macular pig￾ment and age related macular degeneration, Br. J. Ophthalmol. 83 (1999) 867–877. [6] N.P. Holzapfel, B.M. Holzapfel, S. Champ, J. Feldthusen, J. Clements, D.W. Hut￾macher, The potential role of lycopene for the prevention and therapy of prostate cancer: from molecular mechanisms to clinical evidence, Int. J. Mol. Sci. 14 (2013) 14620–14646. [7] G. Riccioni, N. D’Orazio, S. Franceschelli, L. Speranza, Marine carotenoids and cardiovascular risk markers, Mar. Drugs 9 (2011) 1166–1175. [8] C. Wang, D.W. Armstrong, C.-D. Chang, Rapid baseline separation of enantiomers and a mesoform of all-trans-astaxanthin, 13-cis-astaxanthin, adonirubin, and adonixanthin in standards and commercial supplements, J. Chromatogr. A 1194 (2008) 172–177. [9] A. Sujak, J. Gabrielska, J. Milanowska, P. Mazurek, K. Strzaaka, W.I. Gruszecki, Studies on canthaxanthin in lipid membranes, Biochim. Biophys. Acta 1712 (2005) 17–28. [10] M. Garone, J. Howard, J. Fabrikant, A review of common tanning methods, J. Clin. Aesthet. Dermatol. 8 (2015) 43–47. [11] R. Aman, S. Bayha, R. Carle, A. Schieber, Determination of carotenoid stereoisomers in commercial dietary supplements by high-performance liquid chromatography, J. Agric. Food Chem. 52 (2004) 6086–6090. [12] J.B. Thomas, K.E. Sharpless, J.H. Yen, C.A. Rimmer, Determination of fat￾soluble vitamins and carotenoids in standard reference material 3280 multivitamin/multielementtablets by liquid chromatography with absorbance detection, J. AOAC Int. 94 (2011) 815–823. [13] A. Müller, B. Pietsch, N. Faccin, J. Schierle, E.H. Waysek, Method for the deter￾mination of lycopene in supplements and raw material by reversed-phase liquid chromatography: single-laboratory validation, J. AOAC Int. 91 (2008) 1284–1297. [14] I. Brabcová, M. Hlavácková, ˇ D. Satínsk ˇ y, ´ P. Solich, A rapid HPLC column switch￾ing method for sample preparation and determination of -carotene in food supplements, Food Chem. 141 (2013) 1433–1437. [15] J. Szpylka, J.W. DeVries, Determination of -carotene in supplements and raw materials by reversed-phase high pressure liquid chromatography, J. AOAC Int. 88 (2005) 1279–1291. [16] C. Emenhiser, L.C. Sander, S.J. Schwartz, Capability of a polymeric stationary phase to resolve cis–trans carotenoid isomers in reversed-phase liquid chro￾matography, J. Chromatogr. A 707 (1995) 205–216. [17] B. Simonovska, I. Vovk, V. Glavnik, K. Cerneli ˇ c, ˇ Effects of extraction and high￾performance liquid chromatographic conditions on the determination of lutein in spinach, J. Chromatogr. A 1276 (2013) 95–101. [18] S. Bijttebie, E. D’Hondt, B. Notena, N. Hermans, S. Apers, S. Voorspoels, Ultra high performance liquid chromatography versus high performance liquid chro￾matography: stationary phase selectivity for generic carotenoid screening, J. Chromatogr. A 1332 (2014) 46–56. [19] D.R. Gere, R. Board, D. McManlglll, Supercritical fluid chromatography with small particle diameter packed columns, Anal. Chem. 54 (1982) 736–740. [20] A. Matsubara, T. Uchikata, M. Shinohara, S. Nishiumi, M. Yoshida, E. Fukusaki, T. Bamba, Highly sensitive and rapid profiling method for carotenoids and their epoxidized products using supercritical fluid chromatography coupled with electrospray ionization-triple quadrupole mass spectrometry, J. Biosci. Bioeng. 113 (2012) 782–787. [21] L. Nováková, P. Chocholous,ˇ P. Solich, Ultra-fast separation of estrogen steroids using subcritical fluid chromatography on sub-2-micron particles, Talanta 121 (2014) 178–186. [22] C. West, E. Lesellier, A unified classification of stationary phases for packed col￾umn supercritical fluid chromatography, J. Chromatogr. A 1191 (2008) 21–39. [23] A.G.-G. Perrenoud, J.-L. Veuthey, D. Guillarme, Comparison of ultra-high perfor￾mance supercritical fluid chromatography and ultra-high performance liquid chromatography for the analysis of pharmaceutical compounds, J. Chromatogr. A 1266 (2012) 158–167. [24] Y.M. Choo, M.H. Ng, A.N. Ma, C.H. Chuah, M.A. Hashim, Application of super￾critical fluid chromatography in the quantitative analysis of minor components (carotenes, vitamin E, sterols, and squalene) from palm oil, Lipids 40 (2005) 429–432. [25] T. Bamba, N. Shimonishi, A. Matsubara, K. Hirata, Y. Nakazawa, A. Kobayashi, E. Fukusaki, High throughput and exhaustive analysis of diverse lipids by using supercritical fluid chromatography-mass spectrometry for metabolomics, J. Biosci. Bioeng. 105 (2008) 460–469. [26] L. Nováková, A.G.-G. Perrenoud, I. Francois, C. West, E. Lesellier, D. Guillarme, Modern analytical supercritical fluid chromatography using columns packed with sub-2 m particles: a tutorial, Anal. Chim. Acta 824 (2014) 18–35. [27] E. Lesellier, C. West, The many faces of packed column supercritical fluid chro￾matography – a critical review, J. Chromatogr. A 1382 (2015) 2–46. [28] V. Abrahamsson, I. Rodriguez-Meizoso, C. Turner, Determination of carotenoids in microalgae using supercritical fluid extraction and chromatography, J. Chro￾matogr. A 1250 (2012) 63–68. [29] A. Matsubara, T. Bamba, H. Ishida, E. Fukusaki, K. Hirata, Highly sensitive and accurate profiling of carotenoids by supercriticalfluid chromatography coupled with mass spectrometry, J. Sep. Sci. 32 (2009) 1459–1464. [30] M.D. Jones, B. Avula, Y.-H. Wang, L. Lu, J. Zhao, C. Avonto, G. Isaac, L. Meeker, K. Yu, C. Legido-Quigley, N. Smith, I.A. Khan, Investigating sub-2 mm particle stationary phase supercritical fluid chromatography coupled to mass spec￾trometry for chemical profiling of chamomile extracts, Anal. Chim. Acta 847 (2014) 61–72. [31] D.J. Hart, K.J. Scott, Development and evaluation of an HPLC method for the analysis of carotenoids in foods, and the measurement of the carotenoid con￾tent of vegetables and fruits commonly consumed in the UK, Food Chem. 54 (1995) 101–111. [32] J.B. Thomas, D.L. Duewer, I.O. Mugenya, K.W. Phinney, L.C. Sander, K.E. Sharpless, L.T. Sniegoski, S.S. Tai, M.J. Welch, J.H. Yen, Preparation and value assignment of standard reference material 968e fat-soluble vitamins, carotenoids, and cholesterol in human serum, Anal. Bioanal. Chem. 402 (2012) 749–762. [33] M. Anastassiades, S.J. Lehotay, D. Stajnbaher, ˇ F.J. Schenck, Fast and easy multiresidue method employing acetonitrile extraction/partitioning and “dis￾persive solid-phase extraction” for the determination of pesticide residues in produce, J. AOAC Int. 86 (2003) 412–431.