Quick qualitative analysis of PCDD/Fs from the iron ore sintering process

Baosteel Technical ResearchVolume4, Number2, June 2010, Page 29Quick qualitative analysis of PCDD/Fs from the iron ore sintering processHE Xiaolei, YU Yongmei and u XianweiEnvironnent & Resources Division ,Research Institute ,Baoshan lron & Steel Co.,Ltd,Shanghai 201900 ,ChinaAbstract: Comprehensive two-dimensional gas chromatography coupled with time of-flight mass spectrometry (GCxGC-TOFMS) was applied to the quick qualitative analysis of 17 types of polychlorinated dibenzo-p-dioxins and furans(PCDD/Fs) from the iron ore sintering process. The GC x GC-TOFMS parameters were evaluated and optimized to yielda complete separation for the 17 types of PCCD/Fs from interference and to the atanment of high sensitivity. Afteroptimization ,all the 17 types of PCCD/Fs were separated within 42.5 min. The critical and most toxic tetrachlorodibenzo-p-dioxin (TCDD) detection can be up to 0.5 ug /L. The method was used to analyze real fly ash samples. Accurateidentification and determination of all the analyses were by GC x GC ,mass spectrum library and ion ratio. The technique isthought to be a substiute for and a supplement to high resolution gas chromatography/ high resolution mass spectrometry(HRGC/HRMS) in the field of dioxin analysis.Key words: GC x GC; TOFMS; dioxin; PCDD/Fsdol: 10 3969/j. issn. 1674 - 3458.2010.02.006data acquisition rates ( e.g. up to 500 spectra per1 Introductionsecond) for full scan spectra, which permits thecharacterization and resolve of the narrow peaksPolychlorinated dibenzo-p-dioxins and furansgenerated by GC x (( PCDD/Fs) are two groups of highly toxic andThis study focused on the optimization of GC Xpersistent organic pollutants. In most countries, theGC - TOFMS conditions for the analysis of PCDD/Fssintering process is now believed to be a significantin the detection and separation of 17 types of PCCD/Fs.source of PCDD/Fs. In order to assess the potentialA quick qualitative method has been established , whichmenace of the release of PCDD/Fs from the process,acan be applied to screen dioxin from the PCDD/Fsquick qualitative method for screening PCDD/Fs has tofrom the iron ore sintering process in a short time.be established.The three current world-wide standard methods forcongener specific dioxin analysis, including EPA2 Experimental methodsmethod 1613 ( US Environmental Protection Agency)EN 1948 ( European method) , and JIS K 0311 and JIS.1 MaterialsK 0312 (Japanese methods) require high-resolution gasSolvents for the trace analysis ( Hexane, Toluene,chromatography/high resolution mass spectrometryDichloromethane) and sodium sulphate were from( HRGC/HRMS). HRMS offers high specificity andTedia, America. Silica gel was obtained from Merck，sensitivity. However , this instrument is very expensiveGermany. Basic alumina was from Sigma Aldrich,and bulky , and requires people of high competence toAmerica. 1 C-labeled isotope dilution standards wereoperate it.obtained from Wellington Laboratories Inc. ( Guelph,A few recent studies have reported the dioxinsOnt. , Canada). The intermal and recovery standardsanalysis using two-dimensional gas chromatographywere EPA-1613LCS and EPA-1613ISS solutions,coupled with time of-fight mass spectrometry(GC xrespectively. The calibration solutions were EPA-GC - TOFMS)(2-4.It is thought to be a promising1613CVS set , in which the concentrations of the nativeanalysis tool for PCDD/Fs. It only costs 60% ofcompounds ranged from 0.5 μg/L to 2000 μg/L, andHRGC/HRMS. In GC x GC, two columns of differentthe labeled compound concentrations were 100 μg/L .types are coupled together , and a special device calledand 200 μg/L.modulator is used to rapidly and frequently compressThe GC x GC-TOFMS instrument was the Pegasusand bring together the solutes from the first column andrelease them onto the second dimension column. Due to4D (LECO Corp. ,St. Joseph,MI,USA). The two GCthe narrower focusing of the peaks, this process cancolumns were connected in series with the onesignifcantly increase the signal/noise ratios (S/N) ofdimen:中国煤化I-0 m Rtx Dioxin 2the solutes ，and the secondary column can well separatecolumA) and the otherthe target compounds from the co-eluting matrixMYHC N M H G m high-tempera-components. The time-of-flight (TOF) has very fastture BPX-50 (SGE Corp. ).Coresponding autbor: HE Xiaoli; E-mail; hexiali@ bostel com30Basteel Technical Research, VoL4, No.2, Jun. 20102.2 Sample preparationto separate 17 target compounds from one another.The dust samples were taken from the electro- staticAccording to the manufacturer , the maximum tempera-precipitator ( ESP) and the emission samples weretures for both the columns were 320亡and 360C ,collected from the main stack of the sinter plant. Therespectively. The maximum temperature setting for thedust samples from ESP were air dried, acidifed ansecondary column in the GC method was criticalfitered. The filtered samples were Soxhlet extractedbecause the modulator temperature offset was limitedwith toluene for 24 h and toluene was exchanged byby the set temperature. Therefore , we had to consider ithexane. The extracts were concentrated to about 1 mLin this experiment.and shaken with 20 mL 98% H2 SO4 to decolour andIn the current HRGC/HRMS methodology, gener-remove alkaline metal complexes. After that, theally,a weakly polarized column, such as DB -5 andextracts were evaporated and subjected to a two-stageDB225 ,is used. Restek Rtx Dioxin 2 was provided incleaning-up involving a multi-layer silica column andthis experiment for its higher maximum workingan alumina column. In the first stage , a multi-layeredtemperature. This column had previously been evalu-silica chromatography column ( activated silica , sulph-ated at the Environmental Ministry with quantitativeuric acid silica, basic silica , silver nitrate silica , sodiumresults achieved with DB -5 and DB225 GC columns.sulphate) was employed. After the concentra-tion , theWe preliminarily utilized the existing DB - 5 tempera-cleaned extracts were loaded onto an alumina columnure program in Rtx Dioxin 2 to define the elutionwhere polychlorinated biphenyls ( PCBs) were elutedorder and temperature of the 17 types of PCCD/Fs. The .with dichloromethane/ n-hexane ( 2: 98，volume ratio,program was as follows :the initial temperature was sethereby brief called v/v) and PCDD/Fs were elutedat 140C ,kept for one minute , and then increased fromwith dichloromethane/ n-hexane (50; 50, v/v). Prior to140C to 180C at 40C/ min, increased from 180C to .the GC x GC-TOFMS analysis, the final PCDD/Fs250C at 2C/ min ,increased to 300C at 5C/ min , andelutes were concentrated to a 20 μL nonane bykept at 300C for 10 min. The total analysis time wasnitrogen flow.80 min. As expected，all the analytes were eluted by theprinciple of boiling point. The elution order was nearly2.3Chromatographythe same. There was still difference in tetra-chloinatedInlet type :unsplit stream samplinghomologue due to the different natures of the twoTemperature program:temperature rising from 130Ccolumns. Under the same condition, 2, 3, 7, 8-to 210C at 40C/ min ,temperature rising from 210C totetrachlorodibenzo-p dibenzofuran ( TCDF) was the320C at 4C/min , and kept at 320C for 10 min;first peak on the DB-5 column, followed by 1,2,3,4-Modulation period:3 s;tetrachlorodibenzo-p-dioxin (TCDD) on Rtx Dioxin 2Hot pulse duration:0. 8 s;as the second peak. It can be observed that the firstOffset modulator temperature :40C;peak 1,2,3 ,4-TCDD eluted at 290C and the last peakOffset secondary oven temperature: + 20C;appeared at 300七and stayed for 25 min.MS transfer line temperature :270C.According to the above elution order , we adjusted thetemperature program. The initial temperature rising rate2.4 Masswas increased. The temperature rose from 130C toIon source temperature :250C;210C at 40C/min. Because there were no PCDD/FsFilament voltage:70 eV;appearing below 210C , a slower temperature programMass scan range :full scan;(rising from 210C to 320 at 4C/min) was thenScanning rate :100 spectra per second;used to separate most target analytes. The faster ovenMass resolution : 10000 utemperature programs might not fully separate PCDD/Fs from other interferences expected in real samples.3 Results and discussionAdditionally，overly narrow peak widths in 'D separa-tion could limit the modulation frequency in the3.1Optimization of 'D GC chromatographyGCxGC mode. In order to elute octa-chlorin-atedChoosing the GC oven temperate programs and thedibenzo-p-dioxin ( OCDD ) and octa-chlorinateddibenzo-p-dibenzofuran (OCDF) as soon as possible ,gas flow rate were crucial for the D GC chromato-the maximum temperature was set at 320C for 10 min.graphy optimization. Before optimizing the GC x GCTherefore, the 17 types of PCCD/Fs analyzing timeconditions, we utilized 'D GC to optimize the twocould be shortened to 42.5 min.parameters to achieve the basic separation of 17 typesFour different constant gas flow rates (1.5 mL/ min,of PCCD/Fs.The 'D column temperature program determines the2.0 mL min, 2.5 mL /min, and 3.0 mL/ min) weretested after setting the oven temderature program. In theretention time of all the analyses. In comparison withGC中国煤化工flow yielded better'D column temperature program, the "D column couldsepartime and broaderbe regarded as a constant temperature process. ThepeaksYHCNMHGpeak heights andGC x GC modulation requires a wider peak on the 'Dsensitivity. Experiments were conducted to find ancolumn to ensure its resolution, so 2- 5C/minioptimal medium. Among the 17 types of 2,3, 7, 8-generally recommendedol . The experiment was aimedPCDD/Fs, two groups of hexa-chlorinated homolo-HE Xiaolei, et al. Quick qualitative analysis of PCDD/Fs from the iron ore sintering process31gues,i e. the group of 1,2,3,4,7,8-hexa-chlorinateddibenzo-p-dibenzofuran (HxCDF) and 1,2,3,6,7,8-6000 r .332HxCDF and the group of 1, 2, 3, 4, 7, 8-hexa-=3.0schlorinated dibenzo-p-dioxin (HxCDD) and 1,2,3,6,50007 ,8-HxCDD, were most difficult to be separated. In thetest，we observed that different gas flow rates did not4 000have obvious regular effects on their chromatography.A flow rate of 2 mL/min might be better for theE 3000former compound group and a flow rate of 1 mL/minmight be better for the latter group. But at a flow rate2 000of 2.5 mL/min, the 17 types of PCCD/Fs could beanalyzed within the above analyzing time of 42.5 min. .1 000Under the final optimized' D condition, StandardEPA1613 CS3 solution was analyzed. The 17 most1sttime: 1 368.75 1371.75 1371.75 1374.75 1377.75important PCDD/F congeners were separated within2nd time: 1.25 0.25 2.25 1 .2:0.2542. 5 min. The resolution of the pair(1,2,3,4,7,8-Retention time/sHxCDF and 1,2,3,6,7, 8-HxCDF ) was 17% , which(a)complied with the EN1948 standard7l (≤25%), asshown in Fig. 1.5 000二二32ar-3.5s40 000Re solution= X100%35 0003 0003000025 000200015 000 t1000000017601780 1800 1820 1840 1860 1880 19001st time: 1401 251404.75 1404.75 1408.25 1411.75 1411.752nd time: 2.75 1.25 3.25 1.75 0.25 2.25Retentioo time/sFig.1 GC chromatograms of 'D separation of 1,2,3,4,7,8-(bHxCDF and 1,2,3,6,7 ,8-HxCDF1=4.0s3.2 Optimization of GC x GC chromatography3.2.1 Optimization of the modulation period4000The longer the modulation period, the greater thepeak height enhancement occurs. However, 'D separa-tion would suffer if the modulation time is too long' 8).3000Because optimizing the modulation frequency of thefirst dimension peak is important for preserving theseparation achieved in 'D GC and improving sensitivitythrough peak focusing. Otherwise, a shorter themodulation period would generate more ^D slices.Therefore, the last slices would not be detected withinthe modulation period and overlapped by the slices1st time: 1404.75 1404.75 1408.75 1408.75 1412.75 1412.75from the next period. Thus ,the influences of the abovefactors have to be taken into consideration beforedeciding the modulation period.c)We tested three different modulation periods (3 s,Fig.2 GC x GC profiles of the TCDD slics of diferent3.5 s and 4 s). The summed 'D peak numbers of themodulation periodsmodulated slices in the three modulation periods werealmost the same. The heights of the maximum peaks ofmnarent influence onthe slices were similar and had lttle effect on thethe中国煤化工he HxCDF andsensitivity. Fig.2 shows the GC x GC profiles of theHxCD_i.e. the pair of 1,TCDD slices in the three different modulation periods.2,3,4HCN.M.HG-HxCDf and theThe heights of the maximum peaks are from 5000 toother pair of 1,2,3,4,7, 8-HxCDD and 1,2,3,6,7,8-6000. Other types of PCDD/Fs have almost the sameHxCDD) , have identical mass spectra, so they must beprofile.chromatographically separated for the congener specifi32Baosteel Technical Research, VoL4, No. 2, Jun. 2010analysis. The 'D separations were preserved in the 3 sthe 4 s modulation period ,as shown in Fig. 3. In termns ofand 3.5 s modulation periods ,and the 3 s setting yieldedseparation, we chose the 3s setting as the opimizeda better separation visibly. But the two pairs co-eluted inmodulation period at this point.HxCDDw-3.0s1-3.5 s1yr4.0s1-3.0s15-3.5sN-4.0sFig.3 'D separation of HxCDD and HxCDF of different modulation periodsBut if setting 3 s as the modulation period , anotherthan the primary oven temperature ，which was lowerobservation has to be made for the OCDF peak , whichthan the maximum temperature 360C of BPX-50.is the most polarized and is the last to be eluted on the"D column. The 0CDF peak appears in the two3.3 Optimization of the TOFMSmodulation periods and generates a tailing peak. In250C is the maximum ion source temperatureorder to elute the OCDF peak earlier on the‘ Drecommended by the instrument. At this temperature ,column, we changed the gas flow rate fromit generates the greatest peak heights for all evaluated2.5 mL/min to 3 mL/ min at the 38th minute.compounds. Different electron energies ( 50 eV,60eV,70 eV ,80 eV, and 90 eV) were compared in3.2.2 Optimization of other GC x GC parametersthe same way as in the previous experiments.70 eV orOnce we had optimized the modulation period80 eV are norm-ally selected for the electron voltage.(3 s) , the hot pulse duration default was 0.8 s in theThe sensitivity is slightly higher at 80 eV than 70 eVoperating software. In a previous research，only smallfor most compounds. But if the electron voltage is setdifferences were observed because peak heights wereat 80eV, a self-built spectrum library using theall within 85% - 110% of each other from 0.4 s tomixture standard ("C-PCDD/Fs to "C-PCDD/Fs at1 s. Thus , we chose a middle value of0.8 s as the hota fixed ratio) has to be established. In real samplepulse duration. The previous research had showedanalyses, the ratio of '2C-PCDD/Fs to "C-PCDD/Fstemperature offset versus the primary oven had somewas different, which would lead to worse spectraleffect on the ratio of 'D slices，but it was not a criticalmatches ( Fig.4). In our experiment, 70 eV wafactor. We decided to use 40C as the offset modula-chosen to be the electron energy , for it was availabletion temperature, which agreed with most previousnot only in the self-built spectrum library but also inresearches. The default secondary column ovenhe National Institute of Standards and Technologytemperature used in the GC x GC was 20C higher(NIST) library.311 000100050018252241283中国煤化工200250300YHCNMHG" 300Mass(a)CS1: 0.5 g /L "C-TCDD+100ug /L "C-TCDD(b)CS1:500ug/L " -TCDD+100ug/L IC -TCDDFig.4 Mass spectrum of TCDD3.HE Xiaolei, et al. Quick qualitative analysis of PCDD/Fs from the iron ore sintering process3.4 Analytical performancethe highest point(CS5). For the 200 μg/L 2,3,7,8-TCDD (CS5), four slices were detected. The S/N3.4.1 Sensitivity testwas 6869. 6, which was one-fifth higher than the 'DMethod sensitivity could be evaluated by deter-chromatography (Fig.5). For 0.5 μg/L 2,3,7,8-mining whether theTOFMS, using the sensitivityTCDD (CS1) ,only two slices were detected. This isenhancement possibilities of GC x GC，could meet thenot surprising considering the low level, but thcalibration requirements for the EPA Method 1613.resulted two slices could be easily detected and theThe most critical low point is the 0.5 μg/L standardS/N was 2.819 ( Fig.6). Comparatively ,it could notfor 2,3,7, 8-TCDD, which is considered to be thebe detected at all on‘ D column. The modulationmost toxic of the 17 types of PCCD/Fs. Under theprocess enhanced the sensitivity of the TOFMS by theoptimized GC x GC-TOFMS condition, we analyzedtwo standard solutions , the lowest point (CS1) andzone compression of the chromatographic peaks.1000 -3201000-3209000TCDD-2 (, TCDD-38000TCDD7 000舍6000穿600旨500040003000TCDD-1TCDD42000ED-CS51 000D-CSS0LoIst time: 1 398.75 1 404.75 1 407.75 1 413.75 1 419.7513901395140014051410141514202nd time: 1250.252.25 1 .25Retention time/sRetention time/ s(E)GC-TOFMS(b) GCXGC-TOFMSFig.5 Sensitivilty of TCDD (m/z" 320) in CS5185- 32320900180, TCDD-1, TCDD-2880175、 170860165160840s58201502D-CS1'D-CS1l当800lst time: 1 404.75 1 407.75 1 407.75 I 410.75 1 413.751392 I 394 13961398 1400 1402 1404 I 4062nd time:” 0254 22525 025Rctention time/sRetention time /s(a)GC-TOFMS(b)GC xGC-TOFMS●:m/z - mass to-charge ratioFig.6 Sensitivity of TCDD (m/z° 320) in CS13.4.2 Selectivity testtheir analysis when a mass spectrometer operating at aHRMS was performed using sector instruments thatnominal mass was used. For example ,the two ions , asroutinely operate above 10000 of mass resolution.nominal masses, used for 2,3,7 ,8-TCDD monitoringHRMS offered the required specificity and sensitivity中国煤化工-MS permitted thedown to the femtogram range. HRMS was operated inmonit321. 8936, whichthe selected ion monitoring ( SIM) mode in pre-substaHC N M H Gof ntrference.defined retention time windows. PCDD/Fs wereCompared with HRMS , the identification and quali-almost always at much lower concentrations thanfication of the 17 types of PCCD/Fs were as follows:compounds that had the potential for interfering with(1) The 1st dimension time;34Basteel Technical Research, VoL4, No.2, Jun.2010(2) The 2nd dimension time;374 and 376) and 6Ci-dioxins (m/z 390 and 392) in3) Comparing the analyte spectra with the referencereal samples. The intensity of the matrix interferencespectra;peaks appears to be similar to the analytes. Most of the(4) Comparing ion abundance ratios with the theore-peaks are baseline -separated in the chromato graphictical ion abundance.plane in 'D ,although significant co-elutions arise in 'D.Using TCDD as an example ,ions 319. 8965/321. 8936These close elutions do not prohibit the measurement ofshould exist at a rate of 0.77 (due to 35Cl and 37Clthe individual 6C1-dioxins and furans. In addition ,M +2/natural abundance) ,and any rate between 0.65 and 0.89M +4 is proved to be at a reasonable rate (1.24% 士was acceptable.15% ). Compared with the reference spectra, the match isFigs. 7 and 8 illustrate the separation of the 17 typesbetween 650 and 1000. The qualitative results of the 17of PCCD/Fs from the sintering fly ash sample andtypes of PCCD/Fs are given in Table 1. .standard sample. In practice,' D was made of a non-polarized stationary phase, so that most compoundswere only retained depending on their volatility while'D was a more polarized stationary phase so that morepolarized compounds were retained longer than lesspolarized ones. Such a set of columns allows separa-tion of sample components according to compoundclasses. The 'D elution order of PCDD/Fs , as given inFigs.7 and 8, was 4CI, 5Cl, 6Cl, 7Cl and 8Cl,depending on their volatility. PCDFs were alwaysprior to PCDDs. The“D column was very efficient inseparating the target analytes from matrix interferencesthat were not eliminated during the sample preparationsteps. The relatively poor mass resolution of theIst dimension time/sTOFMS instrument compared with the HRMSFig.7 GC x GC contour plot of the ny ash sampleinstrument was partly compensated by the GC x GCseparation efficiency(9 . The two-dimensional chroma-tographic separation process reduces the risk of Co-elution with the interfering analytes of similar massesand offers two retention time values for the enhancedselectivity. In addition to the measurements of thetarget compounds, GC x GC-TOFMS offered thecapability to identify other analytes present in thextract and characterized by masses included in the60- 520 m/z mass range, which will help us toevaluate the interferences and eliminate them by thesample preparation.The contour plot of Fig.7 ilustrates the importanceof the separation achieved in D for 6C1-furans (m/zFig.8 GC x GC contour plot of the standard sampleTable 1 GC x GC-TOFMS qualitative results of the ny ash sampleExpected analyte 1st dimension 2nd dimension Expected Calculated lon ratioNameMatchS/NR. T./stime /stime /ion ratio ion ratio result2,3,7 ,8-TCDF778 1380. 75 ,1.5701377. 751.580.77 0. 786062264.42,3,7 ,8-TCDD575 1407.75,1.540 1410.751.60.77 0.77848 Passed 81. 941,2 ,3,7 ,8-PeCDF760 1560.75,1.670 1557.751.661.551. 5638Passed1 88882,3,4,7 ,8-PeCDF827 1617. 75,1.7001614.751.681.551. 60041913.51 ,2 ,3,7 ,8-PeCDD579 1626. 75,1. 6701623. 751. 651. 70671043.81,2,3,4,7 ,8-HxCDF8831776. 75,1.7701773. 751.751.24 1. 24261948.91 ,2,3,6,7 ,8-HxCDF891 1782.75,1.790 1779.751.781.241. 25922790. 62,3 ,4 ,6,7 ,8-HxCDF925 1821.75,1.840 1818.751.811.234108.51 ,2,3,4,7 ,8-HxCDD6181824.75,1.7901821.751.1941 1. 432717. 11 ,2,3,6,7 ,8-HxCDD666 1830.75,1. 7901827.751.771. 244706. 12.1,2,3,7,8 ,9-HxCDD6201851.75,1. 8301 848. 75.81.233704. 131 ,2,3,7 ,8 ,9-HxCDF679 1875. 75,1.9801872.75中国煤化工Passs775. 631,2,3,4 ,6,7 ,8-HpCDF 956 1974. 75 ,2.0301971.759644.31,2,3,4 ,6,7 ,8-HpCDD 9292061.75,2. 1802058. 750HCNM H G Passed1836.41,2,3,4,7,8 ,9-HpCDF768 2109. 75 ,2. 4102106. 75742. 67OCDD940 2352. 75 ,2.5202349. 752.440.89 0.91306 Passed2952. 20CDF9402367. 75 ,2.6702364. 752.580.890. 928332447.4_HE Xiaolei, et al. Quick qualitative amalysls of PCDD/Fs from the iromn ore sintering process354 Conclusions1086(1) :45-60.[5] Xu Guowang. Modem practical chromatography [ M].(1) GC x GC-TOFMS parameters were evaluatedBeijing :Chemistry Industry Press ,2004 :246-28.and optimized to yield a complete separation for the[6] Wu Jianfang. Application of comprehensive two-dimensional gas chromatography/Time-of-flight mass17 types of PCCD/Fs from interference, and to thespectrometry in the analysis of volatile oil of traditionalattainment of high sensitivity.Chinesemedicine and flavor of liquor [ D ]. Nanjing:(2) All the 17 types of PCCD/Fs were separatedNanjing University of Science & Technology，2004: 19-within 42.5 min and the analysis time was shortenedby 1/4 ,compared to HRGC/HRMS.[7] European Union Standard. EN1948-1 : 2006 , Stationary(3) Accurate identification and determination of allsource emissions determination of the mass concentrationanalytes were by GC x GC ,mass spectrum library andof PCDDs/PCDFs and dioxin-like PCBs[S]. 2006: 12-13.ion ratio.(4) GCx GC-TOFMS has been proved to be a[8] Eunha Hoh, Katerina Mastovska, Steven J Lehotay,et al. Optimization of separation and detection conditionssubstitute for and a supplement to HRGC/HRMSfor comprehensive two-dimensional gas chromatography-when used to measure PCDD/Fs in the iron oretime-of-flight mass spectrometry analysis of polych-sintering process.lorinated dibenzo-p-dioxins and dibenzofuransJournal of Chromatography A ,2007 ,1145(1) :210-221.References[9] Focant Jean Francois , Eric J Reiner , Karen MacPherson,et al. Measurement of PCDDs . PCDFs . and non-ortho-[1] Wu Yongning and Zhao Yunfeng. Process of dioxinPCBs by comprehensive two- dimensional gas chromato-analytical technology[ J ] . Chinese Journal of Foodgraphy-isotope dilution time-of-fight mass spectrometryHygiene , 999,11(5):4146.(GC x GC-IDTOFMS)[J]. Talanta ,2004 ,63(1):1231-[2] Houtman Corine J, Petra Booij， Jover Eric， et al.1240.Estrogenic and dioxin-like compounds in sediment fromZierikzee harbour identified with CALUX assay-directedfractionation combined with one and two dimensionalgas chromatography analyses[ J]. Chemosphere ,2006 ,65(11) :2242-2252. .Peter Haglund, Peter Korytar , Peter Danielsson, et al.Environmental analytical chemistry using hyphenated andbioanalytical techniques[ J]. Analytical and BioanalyticalChemistry ,2008 ,390(7) :1815-1827.4] Focant J F, Eppe Gauthier , Marie Louise Scippo,et al.Comprehensive two dimensional gas chromatographyHE XiaoleiYU YongneiLI Xianweiwith isotope dilution time-of-flight mass spectrometry forthe measurement of_ dioxins and polychlorinatedbiphenyls in fodsufs [J]. Chromatography A, 2005,中国煤化工MYHCNMHG