U. S. Food and Drug Administration
Division of Field Science, ORA
Center for Food Safety and Applied Nutrition
Laboratory Information Bulletin No. 4084 Dioxins
July 7, 1997

Quadrupole Ion Storage Mass Spectrometry/Mass Spectrometry Application to the Analysis of all 17 2,3,7,8 substituted Chlorodibenzo-p-dioxins (dioxins) and Chlorodibenzofurans (furans) in Dairy Products and High Fat Foods

NOTE: The laboratory information bulletin is a tool for the rapid dissemination of laboratory methods (or information) which appear to work. It may not report completed scientific work. The user must assure himself/herself by appropriate validation procedures that LIB methods and techniques are reliable and accurate for his/her intended use. Reference to any commercial materials, equipment, or process does not in any way constitute approval, endorsement, or recommendation by the Food and Drug Administration.

The US Food and Drug Administration has monitored dioxins and furans in fish and other foods for over 20 years. The persistent presence of these compounds in the food supply necessitates the maintenance of these efforts into the foreseeable future. Modern analytical approaches for multi-residue organic pesticide measurement prescribed in the FDA PAM I involve a general extraction step and a purification step followed by GLC separation with a specific detector or more recently HPLC when appropriate. These approaches are relatively inexpensive and can easily achieve low parts per billion (ppb) method detection limits (MDLs). Dioxins are found in foods at low to sub parts per trillion (ppt) levels. General and specific interferences at these quantitation limits are severe.

All the earlier FDA methods did not address all of the seventeen 2,3,7,8 substituted dioxins or furans (Firestone 1979, Niemann 1983). Jasinski (1989) tried to modify the existing FDA methodology, but the resulting procedure did not cover all compounds and was more difficult to use than the original method. Gardner published an LIB (3990) (1996) describing an improved approach to the sample preparation. Gardner's approach used a modified blending and extraction technique for dioxins described by Smith (1984) combined with the modified absorbent column clean up of Lamparski (1979, 1980), except without basic silica. The final steps in Gardner's method use carbon and alumina instead of two HPLC steps of the earlier method (Niemann 1983).

The resulting method easily accommodates all the dioxins and furans as well as non ortho PCBs (Smith 1984). The method uses GC/MS exclusively and therefore requires the use of as many ¹³C₁₂ dioxin and furan standards as are available (usually 15). The separation and detection method of choice is HRGC (50 or 60 M narrow bore) and high resolution mass spectrometry. The use of high resolution MS allows a reduction in the ions monitored from ten to five (257, 259, 320, 322, 324 in the case of 2,3,7,8-TCDD). The obvious expense of purchasing and operating a high resolution MS greatly limits the dioxin analysis that can be done by even the most efficient sample preparation procedures.

Quadrupole ion storage (ion trap) devices are the least expensive mass spectrometers on the market today. They can be purchased and operated at a cost similar to that of a GC with autosampling and computer control. The formation and storage of ions above the detector instead of being translated (quadrupole LRMS and sector HRMS) gives ion trap very high sensitivity. Specificity and sensitivity can be improved for targeted analytes by using MS/MS.

The following procedure will describe the quantitation in dairy products of dioxins at 0.2 to 2 ppt using quadrupole ion storage MS/MS. Five parent ions are monitored every 2.5 seconds permitting the use of ¹³C₁₂ labeled isotope dilution quantitation. Isomer specific separation of all the 2,3,7,8-substituted dioxins and furans is achieved in as little as 40 minutes using a 40 M minibore DB-5 ms capillary column.

American cheese samples are prepared according to Gardner (LIB 3990). Cow's milk, peanut oil and butter were prepared by a modification (Hayward 1995) of Smith (1984). In these cases cow's milk was extracted by an AOAC (1993) procedure and cleaned up directly over silica gel, AX21 carbon on glass fibers followed by basic and sulfuric acid silica gel and finally neutral alumina.

All samples were fortified with 15 ¹³C₁₂ labeled 2,3,7,8-substituted dioxins and furans. Methylene chloride fractions from the alumina column prepared as per Gardner or Hayward are collected in a 5 mL reacti-vial containing 5 L recovery standard in nonane and 2 L of tetradecane. The excess solvent is removed under a gentle stream of nitrogen and the sample reconstituted to 10 L with nonane. Three L of extract are subjected to GC/MS/MS analysis. Saturn 5.1 quantitation software is used to calculate the amounts of natives. Recoveries of ¹³C₁₂ labeled standards are calculated manually. The following outlined sections provide the material, reagents and apparatus used in preparing test portions as per Hayward 1995. Most of the materials, reagents and apparatus are identical or very similar to what is found in LIB 3981 and 3990.

Because of the extreme toxicity of these compounds, the analyst must take necessary precautions to prevent exposure to herself/himself, or to others, of materials known or believed to contain PCDDs or PCDFs. Typical infectious waste incinerators are probably not satisfactory devices for disposal of materials highly contaminated with PCDDs or PCDFs. PCDD/Fs are highly toxic, notably causing chloracne, and are presumed to be carcinogenic and teratogenic. Samples containing them are likely to also have PCBs or other toxic compounds. Handle samples and standards with extreme care to avoid skin contact and inhalation of dust or aerosols.

In normal circumstances their non-volatility prevents appreciable vaporization of PCDD/Fs into the atmosphere. Personal protection during the preparation of samples, chemicals, and glassware requires, at minimum, a long-sleeved laboratory coat, complete leg and foot covering, glasses, and gloves of butyl nitrile or another solvent-resistant material not incompatible to the analysis.

Preparation of samples or chemicals emitting potentially toxic or infectious dust, aerosols, or vapors must be entirely conducted in fume hoods. Following skin contact with PCDD/Fs or other toxic or potentially infectious substances, the affected area(s) must be washed at once with cold running water and soap for several minutes. Any work area coming in contact with potentially infectious substances should be wiped with an appropriate disinfectant.

On leaving the lab after work with PCDD/F or other toxic or potentially infectious substances, hands and other significantly exposed skin must be thoroughly washed with soap and water. PCDD/F-contaminated solvent and solid waste is isolated for treatment as "PCB Waste". Further safety information specific to PCDD/Fs and other chemicals is available in Material Safety Data Sheets.

1.0 Apparatus
1.1	Sampling equipment
	1.1.1	Biological samples require containers which withstand freezing, such as Teflon or heavy, round-shouldered, round-bottomed glass jars. The containers should be solvent (hexane) rinsed before use to reduce interference. They may additionally be heated in a muffle furnace at 550°C for 2 hours if they are made of borosilicate glass.
	1.1.2	Screw caps for the sample containers, lined with Teflon. (Solvent-washed foil, used with the shiny side toward the sample, may be substituted for Teflon if the sample is not liquid or corrosive.) Apply tape around cap to completely seal cap to bottom. Sampling equipment must be free of polyvinylchloride, rubber tubing, and other materials which may introduce interferences or absorb analytes. NOTE: All glassware is washed with detergent and rinsed with deionized water followed by acetone and hexane.
1.2	Teflon separatory funnels screw capped, 500 mL and 1 L with Teflon stop cocks.
1.3	Pipets, disposable
	1.3.1	Pasteur, 5 3/4" long x 5 mm ID, controlled drop (VWR Scientific No.14672-608 or equivalent), used for silica and alumina columns, and narrow-tip, 5 3/4" and 9" long, used for sample transfer.
	1.3.1	Serological, 10 and 25 mL (Pyrex, "shorty", VWR Scientific), used for cleanup columns and, with Teflon sleeve, for washing reservoirs.
1.4	Reacti Vials, 5 mL, clear glass (Altech). These vials have the required tapered bottom needed to concentrate samples to 10 l or less and allow the withdrawal of 2 L aliquots for injection.
1.5	Glass vials, 40 mL, 8 mL, and 1.5 mL, clear and amber, with solid Teflon-lined screw caps.
1.6	Ampoules, 2 mL and 5 mL.
1.7	Glass fiber filter sheets, 8"x10" (Nuclepore), and discs, 12.5 cm and 4.7 cm (Nuclepore SN211107); and silanized glass wool.
1.8	Wrist action shaker (VWR No. 57040-049 or equivalent).
1.9	Rotary evaporator (Buchi RE-111A or equivalent) with liquid nitrogen traps for evaporated solvents and a water bath which can be heated to 400C.
1.10	Econo-column racks (Bio-Rad) for Pasteur pipet columns (silica/alumina cleanup).
1.11	300 mm x 48 mm glass chromatographic column with Teflon fittings at both ends (Kontes).
1.12	Polytron homogenizer (Brinkmann Instruments).
1.13	Rotary vane pump (Stokes pump) (capable of 0.01 torr) for the rotary evaporator (Edwards E2M2 or equivalent) or water aspirator.
1.14	Nitrogen blowdown apparatus (N-Evap Analytical Evaporator, Model 112, Organomation Associates Inc., Northborough, Massachusetts or equivalent) with a sand bath which can be heated to 40°C and Teflon tubing connection to traps and gas regulator.
1.15	Low pressure gas regulators with fine control, 0-30 psig, stainless steel diaphragm (J+W Scientific).
1.16	Solvent reservoir (used to introduce solvent to the carbon column see figure 2): 125 mLseparatory funnel, with threaded neck and screw cap and with 1/2" heavy wall glass tubing attached (12.7 mm OD, at the drain stem, to fit 1/2" union; 9.0 mm OD, at the side, to fit 3/8" union). Alternately, a 50 mL or larger Kontes or Ace column that adapts to 1/8" Teflon tubing can be used.
1.17	Stainless steel reducing unions (Swagelok) with Teflon ferrules: 3/8" to 1/8" and 1/2" to 1/8".
1.18	Hamilton HVP four-port four opposing and four-port two way valves with PTFE plugs, 1/8" HPLC tubing adapters and 1/8" clear Teflon tubing.
1.19	Reusable carbon column: An Ace glass column (1.1 cm ID x 6 cm special order) with ends that accept threaded Teflon fittings is generally used. The reusable Ace column is packed with 50 mg AX21 carbon dispersed on 600 mg glass fibers. The fibers are obtained from glass fiber sheets (Gelman Corp.) cut into small pieces (0.5 to 1 cm) and shredded very briefly (do not overdue) in 75 mL 50 dichloromethane with a Polytron homogenizer. Excess solvent is decanted (some fibers may be lost) and the AX21 carbon is mixed with the fibers. The slurry is packed into the column (a tweezers is used to transfer the wet fibers and carbon). The packing requires only about 3 cm of a 1.1 cm ID column; the remaining volume is taken up with 11 mm glass fiber discs. The discs can be cut from an 8 x 11 inch sheet of glass fiber filter using a cork borer. These columns can be reused as long as blanks remain free of interferences.
1.20	Heating tape, flexible, 0.5" wide x 2' (Thermolyne).
1.21	Muffle furnace, heating to 550°C with temperature programming; and oven, heating to 130°C.
1.22	Gas chromatograph/mass spectrometer/data system.
	1.22.1	Gas chromatograph: a temperature-programmable gas chromatograph and all required accessories including syringes, gases, and fused silica capillary columns (DB-5 ms 40 meter x 0.18 mm, 0.18 um film thickness). Temperature programmed for congener separation at 140°C 2 min., 20°/min. to 260°C, 1°/min. to 300°C. temperature program for isomer specific separation of complex mixtures: 140°C 2° min 20°/min to 200°C 5°/min. to 240° 10 min. hold 10°/min. to 280°C 12 min. Septum programmable on column injection or splitless injection is recommended.
	1.22.2	Mass spectrometer: A quadrupole ion storage (ion trap) instrument is required, utilizing 70 volts (nominal) electron energy in the electron impact (EI) ionization mode. Saturn 4D GC/MS with version 5.1 or higher software and MS/MS option. The system must be capable of multiple reaction monitoring (MRM) for 5 ions simultaneously, with a cycle time of 0.5 sec. Maximum integration time for MRM is 125 ms per m/z. A set of several segmented MRM descriptors must be used to eliminate the need for multiple injections per sample extract (see Table IV).
2.0 Reagents and Materials
2.1	Silica gel 60 (70-230 mesh) (EM Science 7754-3), Soxhlet-extracted or otherwise precleaned with dichloromethane/ cyclohexane or equivalent solvents and stored at 130°C.
2.2	Acid silica gel: sulfuric acid (ACS), concentrated (98%) combined 40:60 (by weight) with silica gel. The mixture is homogenized by mechanical shaking 30 minutes.
2.3	Methylene chloride, hexane, benzene, methanol, acetone, toluene, acetonitrile xylene (Burdick and Jackson high purity or equivalent), ethanol (Gold Shield) and ethyl ether (Mallinckrodt preferred), distilled in glass or highest available purity for organic residue analysis. High purity tetradecane, nonane and isooctane.
2.4	Stock standards from concentrated mixtures prepared by Cambridge Isotopes. The concentrates (200 ng/mL 12C native mixed standard 17 congeners), 40 ng/mL TCDD + TCDF 400 ng/mL OCDD, OCDF are stored in the dark at 4°C, preferably in ampules or in 1 mL Reacti-Vials. Primary 13C12 mixed standards (100 ng/mL for spiking samples and final internal calibration are made at 25 ng/mL in a volumetric flask 1 mL. Aliquots of these to mixed stock standards are combined in varying amounts in nonane to provide a appropriate calibration cruve. The 13C12 labeled standards were maintained at 25 ng/mL in all calibration solution. 13C12 labeled analytes are available from Cambridge Isotope Laboratory, Woburn, Massachusetts. EDF, 5999,7999, 8999
2.5	Potassium silicate.(Optional absorbent) (All operations must be as anhydrous) Dissolve 56.6 g of KOH in 300 mL of methanol, transfer to a one-liter boiling flask. Add, in small portions, 100 g silica gel 60 (70-230 mesh) with swirling and break up lumps with a glass rod. Place on a rotary evaporator with the flask immersed in a water bath at 60°C and allow to rotate 105 minutes at atmospheric pressure. The suspension is then poured quickly into a large Chromaflex column fitted with a Nuclepore filter disc and connected to a nitrogen line. Wash the bed with an equal volume of methanol (check pH of methanol eluate) followed by dichloromethane (Caution: this material must be washed throughly with methanol to remove residual KOH otherwise OCDD, HpCDD and certain HxCDFs will be reacted and absorbed irreversibly). Blow the contents dry with nitrogen in a hood. Store the absorbent in an oven at 130°C.
2.6	Preparation of Modified Smith and Stalling / Carbon Column System. A Kontes Chromaflex column, 4.8 cm x 30 cm (or 60 cm) and a washing reservoir (modified separatory funnel, air sample flask or small Ace or Kontes column (100 - 250 mL) are connected by 1/8" heavy walled Teflon tubing to a four-port two-way valve using PTFE plugs. This is connected to a Hamilton four-port four-opposing valve to allow either the larger Chromaflex sample column or washing reservoir to be in line with the four-port four opposing valve. The four opposing valve has a carbon column attached at two of its ports, while the fourth port is left open for collection of eluate, (see Figure 2).
2.7	The Chromaflex column and washing reservoir are equipped to allow connection to low-pressure nitrogen gas by 1/8" Teflon tubing with needle valve control.
2.8	The carbon column is washed with solvents (a previously used column typically with 100 mL toluene, 50 mL methanol, and 50 mL 50:50 dichloromethane/ hexane) in the direction opposite to that of sample loading.
2.9	Two 5 cm glass fiber filter disc are placed on the bottom fitting of the Chromaflex column (see Figure 2, Sample Column), and the fitting is securely slipped into the glass column. Sodium sulfate (anhydrous) is poured into the column to a depth of 2 cm. Silica gel (30 g) is then placed in it, followed by sodium sulfate to a depth of 2 cm (add 5 g potassium silicate and a 1cm layer of sodium sulfate for fish and shellfish). For milk samples only sodium sulfate is used; silica gel is omitted.
2.10	Woelm Activity I neutral alumina. (ICN chemicals) After opening, store Activity I alumina at 130°C.
2.11	Preparation of Silica/Alumina Column. Pack each of two controlled-drop Pasteur pipets with a glass wool plug and a 0.5 cm depth of (anhydrous) Na2SO4. One pipet is then packed with 2 cm (approx. 0.5 g) of Woelm Activity I neutral alumina followed by 0.5 cm of Na2SO4. The other is filled with 2.5 cm of acid silica gel followed by 0.5 cm of Na2SO4, 2.5 cm of potassium silicate (optional) and 0.5 cm of Na2SO4, see Figure 3).
2.12	Attach a 25 mL reservoir ("shorty" disposable pipet, upside down, with the tip excised) to the silica column by a 1/4 inch OD Teflon sleeve and wash the columns together with 10 mL hexane. Make sure air bubbles are excluded and that channeling does not occur. Discard hexane eluate and use the columns promptly.
2.13	Prepurified (99.9995%) nitrogen gas and helium gas (99.9995 or 99.9999%), filtered through an oxygen and hydrocarbon traps.
2.14	Anhydrous sodium sulfate (reagent grade), baked 2 hours in a muffle furnace at 550°C and stored at 130°C.
2.15	Sodium oxalate (analytical reagent).

Two types of calibration procedures are required. Initial calibration is required to establish the linear range of the GC/MS and is required intermittently throughout sample analyses as dictated by results of routine calibration procedures. Routine calibration consists of analyzing the column performance check solution and a calibration solution of 2 ng/mL concentration. A column performance standard is generally not required for the analysis of vertebrate tissues, fluids and secretions. However, a standard fly ash mixture known to contain all isomers is used to assess column performance and to define the elution windows. Response factors are verified from a single calibration standard injected during analysis. The response factors obtained must be within 30% of those obtained from the multi-level calibration.

Initial calibration requires preparation of multi level calibration standards keeping both the recovery standards and the internal standards at a fixed concentration 25 and 400 ng/mL. Proper quantification requires the use of a specific labeled isomer for each congener to be determined. When labeled PCDDs and PCDFs of each homolog are unavailable, use the closest PCDD or PCDF consistent with the technique of isotopic dilution. Each calibration standard should contain the compounds listed in Table I.

Concentration levels for ¹²C standard analytes are 2, 10, 25, 75, ng/mL for electron impact MS². A two-L injection of calibration standards is made. Standards must be in the same solvent (nonane) as final sample extracts. All standards must be stored in an isolated refrigerator at 4°C and protected from light. Acceptable MRM sensitivity is verified by achieving a minimum signal to noise ratio of 10:1 for the m/z 259 ion of 2,3,7,8 TCDD obtained from injection of two L of the 2 ng/mL calibration standard (0.8 pg of 2,3,7,8 TCDD).

From injections of the 4 calibration standards, calculate the relative response factors (RRFs) of analytes vs. the appropriate internal standards. Calculation of relative response factors is described under the calculation section. See Table II for examples of typical responses factors over a calibrated range.

For each analyte calculate the mean relative response factor, the standard deviation, and the coefficient of variation from determinations of relative response factors. The coefficient of variation of the relative response factors for each calibration standard solution should not exceed 30 percent. If this condition is not satisfied, remedial action should be taken (For example check integration of ion responses or check standard concentrations with another instrument or prepare a standard from a different source or check instrument electronics or acquisition data files). At a minimum, calibration criteria should be met for one standard containing all analytes at the concentration(s) corresponding to spiking level of samples.

Routine Calibration

Inject a 2 L aliquot of the column performance check mixture (fly ash, Figure 1). Acquire at least five data points for each GC peak. NOTE: The same data acquisition parameters previously used to analyze concentration calibration solutions during initial calibration should be used for the performance check solution. Determine and document acceptable column performance.

Inject a 2 L aliquot of the calibration standard solution (the appropriate middle concentration of the calibrated range, e.g. 10 ng/mL. Determine and document acceptable calibration as specified above, i.e., MRM sensitivity and relative ion abundance criteria. The measured RRFs of all analytes must be within 30 percent of the mean values established by initial analyses of the calibration standard solutions.

Criteria for acceptable calibration: The % CV for the mean relative Response Factors must not exceed ± 30%. The S/N for the GC signals present in every single ion current profile (SICP) (including the ones for the labeled standards) must be at least 10 (see Tables III and IV for lists of daughter ion detectable at 2 ng/mL).

Milk samples are spiked before extraction begins. A 200 g (or preferably 150 g) milk sample is transferred to a 500 mL narrow mouth Teflon bottle and 1 g sodium oxalate is added, followed by an equal volume of ethanol. An equal volume of 50:50 hexane/ethyl ether is added and the mixture shaken on a mechanical shaker for 15 minutes.The phases are separated and the aqueous phase is extracted with a 1/2 of the original extraction volume of 50:50 hexane/ethyl ether two additional times. Each separated organic phase is applied to a sample30 cm Kontes column (containing 5 cm Na₂SO₄) (see figure 2).

Six chicken egg yolks are separated, blended and 25 g are weighed into a 500 mL Teflon bottle. Fortify with 100 pg/congener ¹³C₁₂ uniformly labeled 2,3,7,8 substituted PCDD/F mixed standard. Add 250 mL of 50:50 dichloromethane and n-hexane and shake or blend 10 seconds with a polytron homogenizer. Add 6 grams anhydrous sodium sulfate per gram yolk and blend with polytron 20 seconds. Decant supernatant into a 30 cm Kontes chromoflex extraction column containing 30 g silica gel 60 with 2 cm sodium sulfate above and below silica layer. Add 50 mL extraction solvent to sample and shake vigorously and decant into extraction column. Egg yolks prepared with this procedure will always contain fine protein particles suspended in the supernatant during the extraction. It is important that the bulk of these fines remain in the extraction bottle otherwise they will greatly slow the column flow. If the bulk of the fines do not settle out (within 5 minutes), then centrifuge the sample bottle five minutes at low speed.

Twenty-five gram test portions of homogenized fish fillets are minced and then homogenized with a Polytron homogenizer (20 mm OD generator) in 50:50 dichloromethane/hexane (10 mL/g tissue) until finely shredded. Sodium sulfate is introduced in small portions and blended into the test portion (10 to 20% at a time) until a ratio of 8g/g tissue wet weight is obtained. The homogenizing continues until the sample becomes a homogenous free-flowing suspension with no lumps. The sample is then fortified with labeled 2,3,7,8 substituted PCDD/F standards four picograms per gram wet weight. Transfer entire sample and extract to a 60 cm Kontes column packed top to bottom with 2 cm sodium sulfate, 5 g potassium silicate, 1 cm sodium sulfate, 30 g silica gel 60 and 2 cm sodium sulfate.

Before the sample is applied to it, the Chromaflex column is eluted with 75 mL of hexane. The valve is closed before the liquid level falls below the top of the column bed. The sample extract is then applied. It passes slowly (2-3 mL/min) through the Chromaflex column, leaving insoluble and polar components bound to the column bed. The extract proceeds to the carbon column, leaving PCDDs, PCDFs and certain polar lipids bound to the AX21 carbon. The eluting solvent flows through to a waste collection flask. The elution is aided with nitrogen gas at 1-5 psig. A 50 mL wash of hexane is added as the last of the sample enters the sample column bed. The two way value is switched to the washing reservoir and 30 mL of methylene chloride is passed through the carbon column.

At this point the four opposing valve is turned 90°C, so that the direction of flow through the carbon column is reversed. The reservoir is loaded with 40 mL toluene, preheated to 40°C by placing it in a sand bath, and the outflow tube placed in a 125 mL boiling flask. With the aid of nitrogen gas (1-5 psig), the toluene elutes the carbon column at 2 mL/min. The toluene eluate is rotary evaporated to "dryness" after the addition of 10 L tetradecane. Proceed to the alumina fractionation (see Figure 3).

Alumina Fractionation

Silica/Alumina Column Operation: The two columns are arranged on the Econo-column rack to allow the effluent from the silica gel column to flow directly into the alumina column (Figure 3). Dissolve the residue in 1 mL of hexane and apply it to the top of the silica gel column. Quantitatively transfer the sample by following with two x 1 mL hexane rinses of the sample container. Rinse entire column system with 1 mL hexane and then discard the silica column. If using regular neutral woelm alumina, wash the column with 2 mL of hexane, collecting the eluate in an 8 mL vial. Save this fraction (Fr# 1).

To collect Fr#2 (the PCDD and PCDF-containing fraction), place a separate 5 mL reacti-vial and elute the column with 2 mL of dichloromethane. (The appropriate volume will depend on the current calibration, as discussed above.) This reacti-vial must already contain 100 pg 1234 TCDD + 123789 HxCDD recovery standard (10 pg/L in nonane) and 2 L tetradecane as a "keeper".

After the alumina step, evaporate the dioxin fraction (avoid allowing the sample to go dry) under a gentle nitrogen stream, 40°C bath temperature. Evaporate the contents of the Reacti-Vial to 2 L under nitrogen and add 8 ul nonane before GC/MS/MS analysis.

After verifying response factors and elution windows with a calibration standard and injecting clean solvent and recovery standard to verify that the GC/MS/MS system is free of contamination, inject 2 or 3 L of the concentrated sample extract onto the capillary column. The column and conditions used are: DB-5 ms (40 meter x 0.18 mm, 0.18 -m film thickness) temperature programmed for congener separation at 140°C 2 min., 20°/min. to 260°C, 1°/min. to 300°C. A segmented MRM acquisition program is used. The different congener groups ( tetrachloro, pentachloro etc.) are monitored sequentially. Certain non-2,3,7,8-substituted PCDD/Fs (left column, Table VI) may not be resolved from 2,3,7,8-substituted congeners (right column) with the DB-5 or DB-5 ms phases. The DB-5 ms column phase (J&W Scientific) will separate 2,3,7,8-TCDF, 1,2,3,7,8-PeCDF, 1,2,3,4,7,8-HxCDF, 1,2,3,7,8,9-HxCDD with the following temperature program: 140°C 2° min 20°/min to 200°C 5°/min. to 240°C 10 min. hold 10°/min. to 280°C 12 min. (See figure 1 for fly ash separation on DB-5 ms). The concentration of the internal standard in the sample extract should be the same or very similar to the calibrat ion standards used to measure the response factors.

Analyze samples with selected reaction monitoring of the ions listed in Table V. The GC/MS run is divided into six selected reaction monitoring time segments. A minimum of the two characteristic daughter or in some cases parent ions (i.e. quantitation and confirmation ions, listed in Tables II, III and IV), must be present in the reconstructed ion chromatogram. These ions should have a relative abundance that is at least within 30% of the expected abundance for the a normal daughter ion spectrum under the reaction conditions used during calibration.

The maximum intensity of each of the specified characteristic ions must coincide within ±2 scans or 2 sec of the labelled internal standard. GC peaks assigned to a given homologous series must have relative retention times within the window established for that series by the column performance solutions (standard fly ash and calibration standard).

Quantitate the PCDD and PCDF peaks from the response relative to the appropriate internal standard. Recovery of each internal standard should be greater than 40 percent. Samples with recoveries significantly below 40 percent or greater than 120 percent must be re-extracted and re-analyzed.

Calculations

Relative response factors (RRFs) of every non-2,3,7,8-substituted congener within a series are assumed to be the same as the 2,3,7,8-substituted congener in that series. All RRF calculations for non-2,3,7,8-substituted isomers in a series are based on a ¹³C₁₂ -2,3,7,8-substituted isomer that is commercially available.

Determine the concentration of the selected 2,3,7,8-substituted isomer for each homolog according to the example equation below. If the appropriate congener cannot be used, then use the closest one and attempt to use only polychlorinated dibenzo-p-dioxins congeners to measure polychlorinated dibenzo-p-dioxins and polychlorodibenzofurans congeners to measure polychlorodibenzofurans.

Concentration (pg/g) of 2,3,7,8-TCDD in sample

=

Qis x As G x Ais x RFF

Where:

Calculate the concentration of each congener using the RRF derived from its response relative to the response of its ¹³C₁₂ labelled isomer whenever possible. Mean relative response factors for selected PCDDs andPCDFs are given in Table II. Relative response factors are calculated using data obtained from the intial calibration curve according to the equation:

Calculate the percent recovery, Ris, for each internal standard in the sample extract, using the equation:

If the concentration in the final extract of any of the 2,3,7,8-substituted PCDD/PCDF compounds exceeds the upper method calibration limits (MCL), the linear range of response versus concentration may have been exceeded; a second analysis of the sample (using an appropriate aliquot) should be undertaken. Calculate total concentration of all isomers within each congener series of PCDDs and PCDFs.

All fortified test portions provided PCDD and PCDF measurements that were within 67 to 125 % of the theoretical value during MS/MS analysis (nominal fortification level) (Tables VII and IX). Cow's milk was successfully measured by MS/MS at the 0.19 parts per trillion (ppt) level (Table VII). MS/MS consistently gave unambiguously results, whereas full scan low resolution MS (EI-MS) did not (Table VII). The results for peanut oil and butter in Table IX demonstrate method QLs of 0.5 pg/g fat for most PCDD/Fs and 0.2 pg/g for 2,3,7,8 TCDD and TCDF. The results for peanut oil and butter in Table IX done using multiple reaction monitoring multiple frequency irradiation using ¹³C₁₂ labeled PCDD/F internal standards (no K⁺ silicate was used). Unfortified analysis of peanut oil, shortening and butter are shown in Table VIII. No PCDD/Fs were identified in either the peanut oil or the shortening (except laboratory background for OCDD/F and HpCDD/F). Nine PCDD/F congeners were confirmed in butter and seven were below 1 pg/g fat.

Table VII provides a direct comparison of fortified and unfortified cheese prepared by Gardner (LIB 3990) and analyzed by EI-MS and MS/MS. The two extracts were split equally and each half was subjected to either EI-MS or MS/MS. Fortified results at 1 ppt are largely the same except the fortified heptachloro- and octachloro- congeners could not be measured by EI-MS. The unfortified values (not shown) obtained by the two methods are quite different. Estimated MDLs or QLs using EI-MS are bet ween 0.3 and 3 ppt. MS/MS detected all but three congeners at levels 10 times lower than EI-MS. During this analysis, both EI-MS and MS/MS were able to reproducibly measure an accidental 2,3,7,8-TCDF and 2,3,4,7,8-PeCDF contaminant in this sample at a sub ppt level of 0.34 /0.38 and 0.42/0.43 ppt. These low QLs in EI-MS are not routinely achieveable, but depend on the column conditions and the matrix background after clean up. Note that the ITEQ measured by MS/MS (0.11 ppt) is 20 times lower than the 2.33 ppt ITEQ yielded by EI-MS of the same cheese extract.

Discussion

Single reaction monitoring provides higher sensitivity and lower noise levels than does multiple reaction monitoring. However, there are several disadvantages. First, closely eluting chemical species will require rapid changes in reaction monitoring. The fastest possible change in single reaction monitoring is six seconds. This results in artifacts in the mass baseline which causes quantification difficulties. Coeluting labeled standards cannot be monitored simultaneously in the same run. This fact preve nts automatic correction for recovery in the calculations and therefore the corrections must be done manually.

Multiple reaction monitoring requires at least two ions (see Table V). Monitoring several ions instead of one can produce higher noise and interferences in certain daughter ion reconstructed ion chromatograms. Therefore, whenever possible chlorodioxins must be monitored separately from chlorofurans. This will require using the slower isomer specific temperature program mentioned earlier (see Table V). Monitoring of several ions, while maintaining high sensitivity, also requires fast scan rates (0.5 sec). A minimum of 125 ms is required for each microscan acquired on a single ion during an analytical scan. This time interval permits a CID time of only 30 ms and a 25 ms ionization time, (maximum ion time) 5 ms isolation time, 10 to 30 ms daughter ion scan time for a 50 to 150 m/z scan range. Since the calibration is done between mass 69 and 131 m/z, the MS/MS automatic calibration is inherently inaccurate for high mass ranges. The Saturn software attempts to correct this difference, but is usually not able to correctly predict the ion's secular frequency. This means that single frequency waveforms calculated at masses as large as PCDDs and PCDFs will be inaccurate causing inefficient deposition of energy at the correct frequency (scanning with a CID bandwidth = zero).

This error results in poor conversion efficiencies of the parent ion to daughter ion and substantially reduces sensitivity. Quadrupole rf modulation can solve this problem (secular frequency modulation), but requires long CID times to completely convert parent ion without excessive ion loss (100 to 150 ms). Energy is only efficiently deposited when the ion is in resonance with the applied waveform (only part of the 100 to 150 ms CID time). Longer scan rates will reduce sensitivity by larger amounts than will poor conversion efficiencies.

Multiple frequency irradiation is used in this method to improve conversion efficiencies while maintaining high scan rates and therefore the highest sensitivity possible with multiple reaction monitoring. This technique uses 3 or 5 waveforms spaced every 500 Hz on either side of the calculated secular frequency of the parent ion. Energy is distributed evenly into all frequencies simultaneously. Table V lists the CID bandwidths used for each PCDD or PCDF along with the excitation amplitudes and rf storage. Emission current is set at 100 amps and the electron multiple is set 100 volts above 10⁵ gain.

1. Firestone D, (1977). Determination of polychlorodibenzo-p-dioxins and polychlorodibenzofurans in commerical gelatins by gas-liquid chromatography. J. agric.Food Chem. 25, 1274-1280.

2. Niemann RA, Brumley WC, Firestone D and Sphon JA (1983). Analysis of fish for 2,3,7,8-tetrachlorodibenzo-p-dioxin by electron capture capillary gas chromatography. Anal.Chem. 55, 1497-1504.

3. Lamparski LL, Nestrick TJ and Stehl RH (1979). Determination of part-per-trillion concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin in fish. Analytical Chemistry 51(9)1453-1458.

4. Langhorst ML and Shadoff LA (1980). Determination of parts-per-trillion concentrations of tetra-, hexa-, and octachlorodibezo-p-dioxins in human milk samples. Analytical Chemistry 52, 2037-2044.

5. Lamparski LL and Nestrick TJ (1980). Determination of tetra-, hexa-, hepta-, and octachlorodibezo-p-dioxin isomers in particulate samples at parts per trillion levels. Analytical Chemistry 52, 2045-2054.

6. Jasinski JS (1989). Multiresidue procedures for the determination of chlorinated dibenzodioxins and dibenzofurans in a variety of foods using capillary gas chromatography-electron capture detection. Journal of Chromatography 478, 349-367.

7. Smith, L.M,. Stalling, D.L and. Johnson, J.L (1984). Determination of part-per-trillion levels of polychlorinated dibenzofurans and dioxins in environmental samples. Analytical Chemistry 56,1830-1842.

8. Fat-containaing foods 970.52H AOAC Official Methods of Analysis. vol 2 (edited by K. Helrich) p.278 the William Byrd press, Richmond Va. (1993).

9. Gardner, A. M. and Adrzejewski, D. (1996) Laboratory Information Bulletin 12, No. 3981.

10. Gardner, A. M., Andrzejewski, D. and Hayward, D. G. (1996) Laboratory Information Bulletin 12, No. 3990.

11. Hayward DG (1995) Polychlorinated dibenzo-p-dioxin and polychlorinated dibenzofuran background determination in milk and cheese using quadrupole ion storage MS/MS. Chemosphere 34, (5-7), 929-939.

Table I. Internal Standards, Target Analytes and Recovery Standard
13C12 2,3,7,8 TCDD		2,3,7,8 TCDD		13C12 1,2,3,4 TCDD
13C12 1,2,3,7,8 PeCDD		1,2,3,7,8 PeCDD		13C12 1,2,3,4 TCDD
13C12 1,2,3,4,7,8 HxCDD		1,2,3,4,7,8 HxCDD		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,6,7,8 HxCDD		1,2,3,6,7,8 HxCDD		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,6,7,8 HxCDD		1,2,3,7,8,9 HxCDD		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,4,6,7,8 HpCDD		1,2,3,4,6,7,8 HpCDD		13C12 1,2,3,7,8,9 HxCDD
13C12 OCDD		OCDD
13C12 2,3,7,8 TCDF		2,3,7,8 TCDF		13C12 1,2,3,4 TCDD
13C121,2,3,7,8 PeCDF		1,2,3,7,8 PeCDF		13C12 1,2,3,4 TCDD
13C12 2,3,4,7,8 PeCDF		2,3,4,7,8 PeCDF		13C12 1,2,3,4 TCDD
13C12 1,2,3,4,7,8 HxCDF		1,2,3,4,7,8 HxCDF		13C12 1,2,3,4 TCDD
13C12 1,2,3,6,7,8 HxCDF		1,2,3,6,7,8 HxCDF		13C12 1,2,3,7,8,9 HxCDD
13C12 2,3,4,6,7,8 HxCDF		2,3,4,6,7,8 HxCDF		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,7,8,9 HxCDF		1,2,3,7,8,9 HxCDF		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,4,6,7,8 HpCDF		1,2,3,4,6,7,8 HpCDF		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,4,7,8,9 HpCDF		1,2,3,4,7,8,9 HpCDF		13C12 1,2,3,7,8,9 HxCDD
13C12 1,2,3,4,7,8,9 HpCDF		OCDF		13C12 1,2,3,7,8,9 HxCDD

7/1/97
cal std amt pg
TCDD/F		0.8	4	10	30
PeCDD/F to HpCDD/F		4	20	50	150
OCDD/F		8	45	150	300
	m/e		RRF	RRF	RRF	avg	sd	RSD

2,3,7,8 TCDF	241	1.01	1.215	1.211	1.215	1.16	0.102	9
2,3,7,8 TCDF	243	0.682	0.639	0.539	0.606	0.62	0.060	10
2,3,7,8 TCDD	257	1.508	1.592	1.527	1.56	1.547	0.04	2.4
2,3,7,8 TCDD	259	1.108	1.371	1.044	1.188	1.18	0.142	12
2,3,7,8 TCDD	287	0.557	0.507	0.374	0.347	0.45	0.102	23
1,2,3,7,8 PeCDF	275	0.827	0.971	1.049	1.047	0.97	0.104	11
1,2,3,7,8 PeCDF	277	1.116	1.115	1.003	0.944	1.04	0.085	8
2,3,4,7,8 PeCDF	275	0.765	0.917	0.986	0.984	0.91	0.104	11
2,3,4,7,8 PeCDF	277	1.038	1.093	0.909	0.839	0.97	0.116	12
1,2,3,7,8 PeCDD	291	1.04	1.165	1.033	1.3	1.135	0.126	11
1,2,3,7,8 PeCDD	293	1.397	1.61	1.489	1.421	1.48	0.095	6
1,2,3,7,8 PeCDD	295	0.478	0.554	0.515	0.445	0.50	0.047	9
1,2,3,4,7,8 HxCDF	309	0.75	0.749	0.915	0.859	0.82	0.083	10
1,2,3,4,7,8 HxCDF	311	1.122	1.202	1.210	1.118	1.16	0.050	4
1,2,3,6,7,8 HxCDF	309	0.868	0.826	0.870	0.866	0.86	0.021	2
1,2,3,6,7,8 HxCDF	311	1.021	1.17	1.109	1.073	1.09	0.063	6
2,3,4,6,7,8 HxCDF	309	0.684	0.719	0.1001	0.729	0.73	0.049	7
2,3,4,6,7,8 HxCDF	311	0.759	0.847	0.956	0.899	0.87	0.084	10
1,2,3,7,8,9 HxCDF	309	0.952	0.892	0.879	0.83	0.888	0.05	5.7
1,2,3,7,8,9 HxCDF	311	1.235	1.256	1.104	1.067	1.17	0.094	8
1,2,3,4,7,8 HxCDD	327	1.453	1.25	1.532	1.004	1.30	0.227	17
1,2,3,4,7,8 HxCDD	329	0.675	0.661	0.583	0.412	0.583	0.12081	21
1,2,3,6,7,8 HxCDD	327	1.165	1.178	1.394	1.274	1.25	0.106	8
1,2,3,6,7,8 HxCDD	329	0.633	0.571	0.49	0.459	0.538	0.07886	15
1,2,3,7,8,9 HxCDD	327	1.109	1.02	1.1	0.957	1.05	0.072	7
1,2,3,7,8,9 HxCDD	329	0.1009	0.646	0.562	0.483	0.625	0.13956	22
1,2,3,4,6,7,8 HpCDF	345	1.167	1.263	1.365	1.375	1.29	0.098	8
1,2,3,4,6,7,8 HpCDF	347	0.604	0.727	0.751	0.783	0.72	0.078	11
1,2,3,4,6,7,8 HpCDF	458	0.734	0.661	1.079	1.142	0.904	0.24167	27
1,2,3,4,6,7,8 HpCDD	359	0.7	0.921	0.882	0.819	0.83	0.097	12
1,2,3,4,6,7,8 HpCDD	361	1.961	2.342	2.464	2.468	2.31	0.239	10
1,2,3,4,6,7,8 HpCDD	363	1.483	1.921	1.919	1.794	1.78	0.206	12
1,2,3,4,7,8,9 HpCDF	345	1.265	1.452	1.468	1.52	1.41	0.110	8
1,2,3,4,7,8,9 HpCDF	347	0.832	0.894	0.937	0.96	0.91	0.056	6
1,2,3,4,7,8,9 HpCDF	458	1.335	0.748	0.777	0.917	0.944	0.27075	29
OCDD	395	0.926	0.989	1.037	0.943	0.97	0.050	5
OCDD	397	1.966	2.429	2.062	1.742	2.05	0.286	14
OCDD	458	2.294	2.217	2.52	2.61	2.41	0.185	8
OCDD	334	0.528	0.679	0.655	0.515	0.59	0.085	14
OCDD	399	0.503	0.423	0.473	0.492	0.473	0.03541	7.5
OCDF	379	1.063	0.873	0.891	0.93	0.94	0.086	9
OCDF	381	0.824	0.854	0.973	1.042	0.92	0.102	11
average RSD = 9.8% Ions in bold not routinely used for quantitation; not included in mean RSD.

*Table III. Quadrupole Ion Storage MS/MS daughter Ions for PCDDs
	M- COCl	M+2- (COCl)	M+4- (COCl)	M+6- (COCl)	M- Cl	M+2- Cl	M+4- Cl	M- (2COCl)	M+2- (2COCl)	M+4- (2COCl)	M	M+2
2,3,7,8 TCDD	257	259	261		285	287
1,2,3,7,8 PeCDD	291	293	295		319	321	323	228	230
1,2,3,4,7,8 HxCDD	325	327	329		353	355	357	262	264
1,2,3,6,7,8 HxCDD	325	327	329		353	355	357	262	264
1,2,3,7,8,9 HxCDD	325	327	329		353	355	357	262	264
1,2,3,4,6,7,8 HpCDD	359	361	363		387	389	391
OCDD	393	395	397	399		423	425		332	334	456	458
Bolded Ions are not detected at the lowest calibration point (0.8 pg TCDD, 4 pg all others, OCDD 8 pg)

ORA/ORO/DFS