食品、农产品、饲料、粮油中POPs萃取前处理检测方案(抽提萃取)

Developing a Unified Extraction Technique for Foodstuffs ...Or， let’s try to move toward a simpler， cheaper， faster process -phase I- milk extraction Jeffrey C. Archer and Roy G. Jenkins Jr.2 U.S. Food and Drug Administration， Arkansas Regional Laboratory 3900 NCTR Rd.，Bldg. 26， Jefferson， AR 720792U.S. Food and Drug Administration， Port Everglades Resident Post 1800 Eller Dr.， Suite 200， Ft. Lauderdale， FL 33316 Jeffrey.Archer@fda.hhs.gov；Office： 870-543-4076 Background and Objective： The extraction process for POPs can be manually exhaustive for manyStatus：matrices - so we have been investigating alternate extraction techniques·Completed：to minimize variation between matrices while automating the process.◎ Automate extraction processThe major objectives involve reducing time and cost as shown in 3separated phases：o One analyst successfully extracted milk and QC (60 total)samples in a single (45 hour) work week·Automate extraction processO Feasibility of eggs， animal feeds， tissueO 1 analyst extract two batches in a single day (12 samples)·Incomplete：·Rapid fat cleanup and sample polishingO Rapid fat cleanup to separate analytes from fat◎ Single batch (6 samples) be extracted， cleaned and begininstrumental analyses within a single workdayO Rapid analyses of dioxins / furans， PCBs and PBDEs in a singleanalytical run · Analytical results from 1 instrument in 1 run Experimental： The C. Gerhardt HydrothermTM - SoxthermTM system was designed as an automated Weibull-Stoldt Fat Determination system. Acid hydrolysis of thesample takes place； the fat is allowed to dry on a filter. The fat is extracted from the filtration residue with an organic solvent； the solvent is evaporatedand dried for fat determination. With some method modifications， it is being used as the first step (extraction) of POPs from high fat matrices. Results： Experimental Method： Table 1： Extraction Comparison for Liquid Milk Samples 315 n=152 n=315 n= n=315 npg/g fat ng/g fatum of PCB28.PCB52.CB28.Sum of PCB28.pg/g fatum of DioxinnSSum of Dioxins of DioxinsSuPanCdB DSi(ox2i0n-0L5ike and Dioxin-LikePCB52，PCB1011PCB10P PCB101CBs( LikePCBs (2( PCB138.PCB2005 WHO-PCB138. PCB138， PCE 2005WHoxinsSum of Dioxin： PCBsPCDD/F NHO-PCDD/F WHPERRF· 153 and PCB180 .P(PCB(2005WHO- (2005WHO WITEQ CDD EQ)2005WHea l 153 and PGA1ilPCDDIF-TEQ) TEQ)verage 292 1.34 0.474 0.359 0.45 0.504 0.348 0.4836.20 0.069 0.058 0.069 0.00.4540.3600.454 035?Median 0.321 0.318 0.435 0.3 0.398 ·n=315 previous assignment using liquid-liquidextractions ·n=6 validation study results using automated process ·n=152 current assignment using automated process Results are statistically equivalent between the techniques； however the %RSD is lower for the automated process. Figure 1： Two-Tailed Distribution Test using Total TEQs. · Null hypothesis， HO： The two methods follow the same distribution (reference method is liquid-liquid， testmethod is automated extraction) · Blue curve represents the distribution of 324 TEQ values from liquid-liquid extraction (reference method) ·Red curve represents the distribution of 152 TEQ values from the automated extraction (test method)This distribution shows that the null hypothesis cannot be rejected， thus the test method results are acceptablewhen compared to the standard method. Table 2： Congener Specific Results ComparingMilk ExtractinMethods. 1234 9.5E-05 ).34889 0.05615 0.25131 -0.003763649).03024 -0.033416141 ).13895 0.00474 0.021740.00968 0.09586 0.005593520.03856 0.00544 0.02752 -0.012835646 6.7E-0 0.03988 66.7E-06 O0.000340.0267 -0.000576132 0.0014-0.0033437960.00771 0.000180.09383 .0006300285 -7E-050.0003200.051270.15579 -BE-050.00091-0.00060.00495 -0.0004443760.00164 0383 1.00091 0.051270.155790.001590.00726 7.8E-050.00112 -0.029953614|0.09901-0.0142983870.04516 0.32961 0.002290.00057 0.066770.0303 -0.0122563820.03737-0.0059578270.02502 123478-HxCDF 0.00015 0.02496 -0.0003773380.00228 12367B-HxCOF 0.00179 D.01727 -0.0024625370.00792 0.00029 0.00160.00013 0.00532 -0.0003352620.00079 -2.81528E-056.3E-05 2.49629E-050.0001 -0.001250821 0.0101 Conclusion： The results show that the HydrothermTM/SoxthermTM can be used as the extraction stepfor liquid milk， with a high level of confidence as demonstrated with both basic statisticsand a two-tailed distribution test. The precision and accuracy data are acceptable and“real" sample data compare well with historical milk results extracted by a liquid-liquidtechnique. Other matrices such as egg， tissue and animal feed show extreme promisesuggesting that the matrix type (solid or liquid) does not matter； only that fat is presentin the matrix. References： 1. Kotz A， Malisch R， Wahl K， Haedrich J； Comparison of Extraction Methods forDetermination of PCDD/Fs， PCBs and Lipids in Food of Animal Origin andConsequences for Control of Maximum Levels. Organohalogen Compounds 2012， 74，160-163. 2. Bench BJ， Whittington W， Kranz M， Guerrera F； Validation Study-Total fat contentby automated acid hydrolysis [HYDROTHERM]-ISO8262-1 http：//www.gerhardt.de/fileadmin/Redaktion/Validierungsstudie HYDROTHERM EN.pdf ( The authors w ould like t o acknowledge the F DA POPs team members at the Arkansas Regional Lab o ratory for their i nvolvement in this p roject- i t would not have been accomplished without everyone’s help. Add i tionally， we . would like to thank M arkus Kr a nz of C. Gerhardt for the analytical discussions as the project procee d ed. ) The C. Gerhardt HydrothermTM – SoxthermTM system was designed as an automated Weibull-Stoldt Fat Determination system. Acid hydrolysis of the sample takes place; the fat is allowed to dry on a filter. The fat is extracted from the filtration residue with an organic solvent; the solvent is evaporated and dried for fat determination. With some method modifications, it is being used as the first step (extraction) of POPs from high fat matrices.