The latest version of this guidance issued on April 10, 2002. Below is an earlier version.

Guidance for Industry

Preparation of Premarket Notifications for Food Contact Substances:

Chemistry Recommendations

U. S. Department of Health and Human Services
Food and Drug Administration
Center for Food Safety and Applied Nutrition (CFSAN)
September, 1999 (revision 1.1, May, 2000)

This document addresses FDA's recommendations for chemistry information that should be submitted in a premarket notification (PMN) for a food-contact substance (FCS). These recommendations are based on an adaptation and revision of the June, 1995 document for indirect food additive petitions entitled "Recommendations for Chemistry Data for Indirect Food Additive Petitions." Highlights of this document include:

• Alternate approaches to estimating migration to food, such as migration modeling, are presented.

• Consumption factors (CFs) for several specific polymer packaging categories, such as polyolefins, cellophane, and nylons, have been updated.

• Testing for "wet-end" additives used in the manufacture of paper and paperboard is discussed.

• FDA Form No. 3480 "Notification for New Use of a Food Contact Substance" is available (PDF).

• CF for poly(ethylene terephthalate) (PET) updated to 0.21 (5/2000).

Note: If your browser does not display the table of contents and sections using Roman numerals and alphabetic characters, you may request a hard copy of the document as described above.

1. INTRODUCTION

2. CHEMISTRY INFORMATION FOR PREMARKET NOTIFICATIONS
   1. Identity
   2. Use
   3. Intended Technical Effect
   4. Migration Testing & Analytical Methods
      1. Design of the Migration Experiment
         1. Migration Cell
         2. Test Sample
         3. Food Simulants
         4. Temperature and Time of Test
         5. End Tests
      2. Characterization of Test Solutions & Data Reporting
      3. Analytical Methods
         1. Description of the Method
         2. Standard Curves
         3. Examples of Spectra or Chromatograms
         4. Example Calculations
         5. Validation of Analytical Methods
      4. Migration Database
      5. Migration Modeling
   5. Consumer Exposure
      1. Calculation of Exposure
         1. Consumption Factor
         2. Food-type Distribution Factor
         3. Concentration in the Daily Diet and EDI
         4. Cumulative Exposure
      2. Exposure Refinement

APPENDIX II. SELECTED MIGRATION TESTING PROTOCOLS

1. General Protocols (Single-Use Applications) Corresponding to Condition of Use
2. Adjuvants for Polyolefins
3. Adjuvants for Polymers (other than Polyolefins)
   Adjuvants for More than One Polymer
4. Articles Intended for Repeated Use
5. Coatings for Cans
6. Uncoated & Clay-Coated Papers with Latex Binders
7. Specially Treated Papers
8. Adhesives (Room temperature or below)
9. Laminates & Coextrusions
10. Boil-In-Bags
11. Special High-Temperature Applications
    1. Dual-Ovenable Trays
    2. Microwaveable Containers
    3. Microwave Heat-Susceptor Packaging
12. Colorants for Plastics
13. Dry Foods with Surface Containing No Free Fat or Oil
14. Wet-End Additives used in the Manufacture of Paper and Paperboard

1. INTRODUCTION

   Section 309 of the Food and Drug Administration Modernization Act of 1997 (FDAMA), amends Section 409 of the Federal Food, Drug, and Cosmetic Act (FFDCA) (21 U.S.C. 348) to establish a premarket notification (PMN) procedure as the primary method by which the Food and Drug Administration (FDA) regulates food additives that are food contact substances (FCS). A food contact substance is any substance that is intended for use as a component of materials used in manufacturing, packing, packaging, transporting, or holding food if the use is not intended to have any technical effect in the food.

   Notifications for food contact substances must contain sufficient scientific information to demonstrate that the substance that is the subject of the notification is safe for the intended use (21 U.S.C 348(h)(1)). Because the safety standard is the same for all food additives, whether subject to the petition process or the PMN process, information in a PMN should be comparable to that recommended for inclusion in a food additive petition.

   This guidance has been prepared by the Office of Premarket Approval of the Center for Food Safety and Applied Nutrition (CFSAN) at the Food and Drug Administration in accordance with FDA's "Good Guidance Practices" (62 FR 8961; Feb. 27, 1997). The purpose of this document is to provide general guidance for the chemistry information that should be included in a PMN for an FCS. This guidance represents FDA's current thinking on the chemistry information for a PMN. It does not create or confer any rights for or on any person and does not operate to bind the Agency or the public. An alternative approach may be used if such approach satisfies the requirements of the applicable statute and regulations. For situations not addressed in this guidance, notifiers are advised to consult FDA. Periodically, FDA will update this guidance in light of new information.

2. CHEMISTRY INFORMATION FOR PREMARKET NOTIFICATIONS

   The chemistry information that FDA recommends to support the proposed use of a FCS is described in Section III., items C. through F., of the guidance document entitled "Preparation of Premarket Notifications for Food-contact Substances: Administrative." This information is reproduced below in italics, followed by a detailed discussion on each recommendation. A clear and concise presentation of the information in the format described below will facilitate review of the PMN.

   1. Identity

      Item C: Notifications should include detailed information on the chemical identity of the FCS and the impurities and residual reactants from the production of the FCS including, where possible, the chemical and structural formula and the Chemical Abstract Service (CAS) Registry Number.

      Identity information is used to describe the FCS that is the subject of a PMN and to identify substances that may migrate into food from use of the FCS. Migrating substances may include not only the FCS itself, but also degradation products and impurities in the FCS that may have no functional effect.

      Information identifying the FCS should be as complete as possible with respect to its name, composition, and method of manufacture. These items include:

      1. Chemical Name. The IUPAC or Chemical Abstracts name is acceptable.

      2. Common or Trade Names. These should not be the only means of identification. FDA does not maintain a compilation of common or trade names.

      3. Chemical Abstracts Service. Registry Number⁽¹⁾.

      4. Composition. A full description of the composition of the FCS is used to compile a list of potential migrants to food. This should include: chemical formulae, structures, molecular or formula weights for single compounds or components of commercial mixtures. For polymers, notifiers should also submit the weight average (M_w) and number average (M_n) molecular weight, the molecular weight distribution, and the methods used for their determination. If the molecular weight is not readily obtainable, the notifier should furnish other properties of the polymer that are functions of the molecular weight, These as intrinsic or relative viscosity or melt flow index.

         In addition, the notifier should provide the following information:

         1. A complete description of the manufacturing process, including purification procedures, and the chemical equations for all steps of the synthesis.

         2. A list of reagents, solvents, catalysts, purification aids, etc., used in the manufacturing process, the amounts or concentrations used, their specifications, and their CAS Reg. Nos.

         3. Chemical equations for known or likely side reactions occurring during manufacture of the FCS, including catalyst degradation reactions, if known.

         4. Concentrations of all major impurities (e.g., residual starting materials, including all reactants, solvents, and catalysts, in addition to byproducts and degradation products) together with supporting analytical data and calculations. In the case of polymers, concentrations of residual monomers should be included.

         5. Spectroscopic data to characterize the FCS. In some cases an infrared (IR) spectrum is sufficient, but occasionally other information, such as visible and ultraviolet absorption spectra or nuclear magnetic resonance (NMR) spectra, are more useful.

            Those data and information not intended for public disclosure, such as trade secret or confidential commercial information, should be so identified.

      5. Properties. Notifiers should submit the physical and chemical specifications of the FCS (e.g., melting point, impurity specifications). Notifiers should report properties that can affect migration potential, such as solubilities in food simulants. In the case of new polymers, notifiers should provide glass transition temperatures, ranges for densities and melt flow indices, and information on morphology (e.g., degree of crystallinity) and stereochemistry. For new adjuvants in regulated polymers, notifiers should submit information on the properties of the polymer (e.g., T_g) used in migration testing.

      6. Analyses. If the FCS is a component of an otherwise regulated material (e.g., an antioxidant in a regulated polymer), notifiers should provide analytical methods for determining the FCS in the material. Notifiers should submit supporting analytical data (refer to Section D.3.).

   2. Use

      Item D: Notifications should include detailed information on the intended conditions of use of the food contact material manufactured with the food contact substance (e.g., maximum use temperature, type of food with which the substance is intended to come into contact, duration of the contact, and whether the food contact material is intended for repeat or single use application.)

      Notifiers should examine general use limitations in notifications and regulations for similar food contact substances and propose a comprehensive set of limitations on the notified use. Certain of these limitations may be the basis for assumptions made in deriving exposure estimates for the FCS. Any applicable limitations should be included in the description of the notified use. In the absence of appropriate limitations, FDA may be required to use assumptions in estimating exposure that would result in more conservative values for certain classes of FCS.

      1. Notifiers should provide the maximum use level of the FCS and the types of food-contact articles in which it may be used. "Use level" refers to the concentration of a substance in the food-contact article, not in the food. Notifiers should state the range of possible uses, such as films, molded articles, coatings, etc., and report the anticipated maximum thickness and/or weight per unit area of these articles.

      2. Notifiers should list the types of food (with examples) expected to be used in contact with the FCS and the maximum temperature and time conditions of food contact⁽²⁾. Classifications that may be helpful are given in 21 CFR 176.170(c), Table 1 (Types of Raw and Processed Foods) and Table 2, which lists various conditions of use. These tables are not all-inclusive, however.

   3. Intended Technical Effect

      Item E: Notifications should include a statement of the intended technical effect of the food contact substance, and data that demonstrate the minimum amount of the substance that will achieve the intended technical effect.

      Notifiers should present data to show that the FCS will achieve the intended technical effect and that the proposed use level is the minimum level required to accomplish the intended technical effect. In these circumstances, "technical effect" refers to the effect on the food-contact article, not on the food. An example would be the effect of an antioxidant in preventing oxidative degradation of a particular polymer. In the case of a new polymer, notifiers should present data that demonstrate the specific properties of the polymer that make it useful for food-contact applications. This information is frequently available in product technical bulletins.

      In cases where the use level of an FCS is self-limiting, notifiers should provide supporting data.

   4. Migration Testing & Analytical Methods

      Item F: Notifications should include sufficient data to enable FDA to calculate the estimated daily intake (EDI) resulting from the intended use of the substance, including levels of residual reactants and impurities.

      A notifier should provide information sufficient to permit estimation of the daily dietary concentration of the FCS, i.e., consumer exposure. FDA will calculate the concentration of the FCS expected in the daily diet from analyzed or estimated levels of an FCS in food or food simulants. A more complete discussion of this topic is given in Section II.E. and Appendix IV.

      The concentration of an FCS in the daily diet may be determined from measured levels in food or in food simulants, or estimated using information on formulation or residual levels of the substance in the food contact article and the assumption of 100% migration of the FCS to food. Although FDA has always accepted reliable analyses of FCS in real foods, in practice, many analytes are difficult to measure in food. As an alternative, notifiers may submit migration data obtained with food simulants that can reproduce the nature and amount of migration of the FCS into food. Because an FCS may be used in contact with many foods with different processing conditions and shelf lives, the submitted migration data should reflect the most severe temperature/time conditions to which the food-contact article containing the FCS will be exposed.

      Before undertaking migration studies, the notifier should carefully consider the potential uses of the FCS. If, for example, use at temperatures no higher than room temperature is anticipated, it makes little sense to conduct migration experiments that simulate high temperature food contact. Such experiments would lead to elevated levels of the FCS in the food simulants that might require a more extensive toxicological data package to support the exaggerated exposure. In some cases where the use level of the FCS is low, it may be possible to dispense with migration studies altogether by assuming 100% migration of the FCS to food. The following example illustrates this approach:

      Consider an adjuvant added prior to the sheet-forming operation in the manufacture of paper. If analysis or calculation shows that the final adjuvant concentration in paper cannot exceed 1 mg/kg and the basis weight of the finished paper is 50 pounds/3000 ft², or 50 mg/in², then the maximum weight of adjuvant per unit area of paper is 1 x 10^-6 g adjuvant/g paper x 50 mg/in² = 0.000050 mg/in². If all the adjuvant migrates into food and 10 grams of food contacts 1 square inch of paper (FDA's default assumption), the maximum concentration in food would be 5 g/kg. It may be expected that this low concentration in food would lead to a commensurately low EDI for the FCS. Therefore, although migration studies which could result in further lowering of the estimate of daily intake, such studies might be unnecessary.

      Levels in food should be based on the results of migration testing or other methods as applicable, so as to reflect as closely as possible the actual use conditions of the food contact article containing the FCS. In general, migration values determined using the assumption of 100% migration to food should be avoided to reduce conservatisms to the greatest extent possible.

      1. Design of the Migration Experiment
         1. MIGRATION CELL. When use of an FCS is anticipated with one particular type of package, such as a beverage bottle, packages may be filled with food simulants and tested. For more general uses or when the package surface area does not produce sufficient extractives for adequate characterization, a migration cell should be used in which a specimen of known surface area is extracted by a known volume of simulant. The two-sided migration cell described in an article by Snyder and Breder (Snyder and Breder, 1985) is recommended. Although this specific cell may not be universally applicable, FDA recommends that two of its essential features be incorporated in modified designs. These are:

            (1)   Polymer plaques of known surface area and thickness (see Section II.D.1.b. for further discussion) are separated by inert spacers (such as glass beads) so that simulant flows freely around each plaque. Migration from the plaque is considered to be two-sided.

            (2)  The headspace is minimized, and gas-tight and liquid-tight seals are maintained. (Minimum headspace and gas tightness are of lesser importance if the migrant of interest is non-volatile.)

            Additionally, and importantly, the cell should be subjected to mild agitation to minimize any localized solubility limitation that might result in mass-transfer resistance in the food simulant.

            For applications in which a two-sided cell design is not suitable, such as laminate constructions, the notifier may refer to the references in Appendix V. for applications describing other cell designs. The notifier may also devise an alternative cell. FDA is willing to comment on any such design prior to performance of the migration experiment.

         2. TEST SAMPLE. Some important considerations are the following:

            (1)  Formulation:   Use the highest proposed concentration of the FCS in the food-contact article in preparing samples for migration testing. Notifiers should provide information that characterizes resin samples used in testing, including the concentrations and identities of other components that may be present, the chemical composition of the resin (including co-monomer content where appropriate), molecular weight range, density, and melt flow index. If the formulation is plasticized, the most highly plasticized formulation should be used for testing.

            (2)  Sample Thickness & Surface Area:   Report both the thickness of the test plaque and surface area of the sample tested. If a plaque is tested by immersion and is of sufficient thickness to ensure that the initial FCS concentration at its center is unaltered by migration that occurs from both sides during the test period, the surface area of both sides can be used to calculate migration (units of mg/in²).

            Migration may be considered to be independent from both sides of the sample if the sample plaque thickness is at least 0.05 cm (20 mil or 0.020 in) and not more than 25 percent of the FCS has migrated by the end of the experiment. If these conditions are not met, the surface area of only one side should be used in the calculation and consideration should be given in the PMN to proposing a limitation on film thickness.

            (3)  Polymer properties:   If the FCS is a polymer adjuvant, notifiers should perform migration testing on the polymer with the lowest average molecular weight. If the FCS is a new polymer, notifiers should test the polymer that would be expected to give the highest levels of extractives, i.e., the polymer with the lowest average molecular weight, percent crystallinity, and degree of cross-linking.

         3. FOOD SIMULANTS. The following food simulants are recommended. Additional discussion on this subject is found in Appendix I.

            Food-Type as defined in 21 CFR 176.170(c) Table 1	Recommended Simulant
            Aqueous & Acidic Foods (Food Types I, II, IVB, VIB, and VIIB)	10% Ethanol(a)
            Low- and High-alcoholic Foods (Food Types VIA, VIC)	10 or 50% Ethanol(b)
            Fatty Foods (Food Types III, IVA, V, VIIA, IX).	Food oil (e.g., corn oil), HB307, or Miglyol 812(c)
            afor exceptions, see main text. bactual ethanol concentration may be substituted (see main text and Appendix II.). cHB307 is a mixture of synthetic triglycerides, primarily C10, C12, and C14. Miglyol 812 is derived from coconut oil (see main text and Appendix I.).

            When food acidity is expected to lead to significantly higher levels of migration than with 10% ethanol, or if the polymer or adjuvant is acid-sensitive, or if trans-esterification occurs in ethanol solutions, separate extractions in water and 3% acetic acid in lieu of 10% ethanol should be conducted⁽³⁾.

            10% Ethanol is intermediate in alcohol concentration between wine and beer. Migration levels to wine and beer are not expected to be very different from 10% ethanol values. Therefore, test results developed with 10% ethanol can generally be used to evaluate exposures and support clearances for contact with alcoholic beverages with up to 15 volume % ethanol.

            Unsaturated food oils (like corn and olive oils) can at times be difficult matrices for the analysis of a migrant because these oils are susceptible to oxidation, especially at high temperature. Miglyol 812, a fractionated coconut oil having a boiling point range of 240° to 270°C and composed of saturated C₈ (50-65%) and C₁₀ (30-45%) triglycerides, is an acceptable alternative fatty-food simulant for migration testing.⁽⁴⁾ HB 307, a mixture of synthetic triglycerides, primarily C₁₀, C₁₂, and C₁₄, is also useful as a fatty-food simulant.⁽⁵⁾

            In some cases, analysis of a migrant in a food oil will not be practical and a simple solvent must be used. There does not appear to be one solvent that will effectively simulate a food oil for all polymers. A list of various polymers and their recommended fatty-food simulants appears in Appendix I. For other polymers, a notifier should consult with the FDA concerning use of an appropriate fatty-food simulant before performing migration experiments.

            The simulant volume should ideally reflect the volume-to-specimen surface area ratio expected to be encountered in actual food packaging. A ratio of 10 mL/in² is acceptable. Other ratios may be acceptable if migration levels do not approach concentrations reflecting the partition limit (i.e., the solubility of the FCS in the food simulant). Precipitation of the FCS from solution or a cloudy solution is an indication that this limit has been reached. The volume-to-surface area ratio should be reported.

         4. TEMPERATURE AND TIME OF TEST. Notifiers should conduct migration testing under the most severe conditions of temperature and time anticipated for the proposed use. If the intended use of the food contact article involves contact with food at temperatures higher than room temperature, tests should be conducted at the highest use temperature for the maximum time period. In many instances, short time periods of elevated temperature-food contact are immediately followed by extended periods of storage at ambient temperatures. For such applications, FDA's recommended migration protocols call for short-term accelerated testing designed to simulate FCS migration that may occur during the entire food-contact scenario. Recommended protocols for selected situations are given in Appendix II.; however, depending on the particular food-contact application, a specific protocol may be devised.

            For room-temperature applications, a test temperature of 40°C (104°F) for 10 days is recommended. This accelerated testing protocol is based on studies showing that experimental migration levels were roughly equivalent to levels obtained after extended storage (6-12 months) at 20°C (68°F) ⁽⁶⁾.

            For refrigerated or frozen food applications, the recommended test temperature is 20°C (68°F).

            For polymers, such as polyolefins, that are used with food at temperatures above their glass transition temperatures (i.e., the polymer is in the rubbery state), the highest migration values (typically, but not always, the ten day values) are used by FDA to calculate the concentration of migrants in food.

            Polymers such as polyethylene terephthalate (PET) and polystyrene (PS), however, are used with food at temperatures below their glass transition temperatures (i.e., the polymer is in the glassy state). At a fixed temperature, the rate of diffusion of migrants through a polymer in the glassy state is lower than if the polymer were in the rubbery state. For this reason, accelerated testing for 10 days at 40°C might underestimate migration that would occur during the entire food-contact scenario. Therefore, migration data obtained over ten days at 40°C should be extrapolated to 30 days in order to better approximate migration levels expected after extended time periods at ambient conditions. The notifier may carry out testing for 30 days to avoid uncertainties in extrapolation. If a notifier provides data that demonstrate that a different extrapolation period is more appropriate for a given adjuvant/polymer combination, such information would be used for evaluating exposure.

            For restricted uses where the maximum shelf life and food-contact temperature of an article are known, the notifier is encouraged to carry out migration studies for the maximum shelf life under temperature conditions approximating expected use. Notifiers may want to consult FDA before undertaking such tests.

            For each migration experiment, FDA recommends that portions of the test solutions should be analyzed during at least four time intervals. Recommended sampling times for a ten-day test are 2, 24, 96, and 240 hours. FDA recommends that notifiers analyze a blank or control using a test cell identical to that used for the test article.

         5. END TESTS. It is important for potential notifiers to realize that the appropriate migration test conditions for a new FCS are not those described in §175.300, §176.170 or other sections in 21 CFR. These published "end-test" extractions are quality control test methods that are used to verify whether a particular product is equivalent to the material that served as a basis for the regulation. End tests bear no relation to the migration testing recommended for evaluating probable exposure to a new FCS and may not be used to support notification of an FCS.

      2. Characterization of Test Solutions & Data Reporting

         Notifiers should perform migration studies in triplicate and analyze the test solutions for the migrants. If the PMN is for a polymer, the notifier should determine the amount and nature of total nonvolatile extractives (TNEs). Ordinarily, the TNEs are determined gravimetrically. The nature of the extractives, which may include monomers, oligomers, adjuvants, and catalyst residues, is determined by suitable chemical or physical tests, such as NMR, UV-visible, and atomic absorption spectroscopy, mass spectrometry, and gas or liquid chromatography. The notifier should indicate the limit of quantitation and selectivity of the methods used. If quantitation of individual migrants is not possible, the notifier should determine the distribution of the extractives between organic and inorganic fractions by solvent fractionation (i.e., the fraction of the TNE residue that is soluble in chloroform). This serves, as a first step, to focus on the migrants of interest (e.g., organic components) in determining exposure estimates. In these instances, FDA generally will estimate exposure to TNEs from the use of the FCS assuming that the TNEs (or chloroform-soluble TNEs) consist solely of low molecular weight oligomers that are chemically equivalent. Because the degree of toxicological testing depends on the magnitude of the exposure estimate, it should be to the notifier's advantage to quantitate the components in the TNEs that are not chemically equivalent (e.g., differentiate between low molecular weight oligomers and polymer adjuvants).

         Test solutions from polymers that are the subject of a PMN should also be analyzed for constituent monomers. Alternatively, the known residual monomer level in the polymer may be used to calculate monomer dietary concentrations by using the density of the polymer, the maximum anticipated thickness of the food contact article, and by assuming that all of the residual monomer migrates into food and that ten grams of food contact one square inch of food-contact article.

         If the PMN is for a polymer adjuvant, FDA normally recommends that the test solutions be analyzed only for the adjuvant. Occasionally, however, it may be necessary to quantitate, in the test solutions, impurities or decomposition products present in the adjuvant if they might be expected to become components of the daily diet in toxicologically significant quantities. A common example would be the presence of carcinogenic impurities in the adjuvant.

         It may also be necessary to quantitate, in the test solutions, decomposition products produced (a) as a result of the FCS exhibiting its intended technical effect in the food contact article, or (b) in the test solutions after migration of the FCS. An example would be the use of a new antioxidant for polyolefins. Polymer antioxidants, by their very nature, would be expected to partially decompose during thermal processing of the resin or food-contact article containing the substance. Frequently, decomposition also occurs after migration of the FCS into food or food simulant, where temperatures may reach 120°C with fatty-food simulants. Information on decomposition in the food simulants may be obtained by conducting stability studies on the FCS in parallel with the migration studies.

         Notifiers should report results in terms of milligrams of substance extracted per square inch (mg/in²) of surface area. Migration amounts are often expressed in terms of mg/dm². The mixed unit mg/in² is preferred, however, to facilitate conversion to concentration in food. If ten grams of food are in contact with one square inch of packaging surface, a migration of 0.01 mg/in² corresponds to a concentration in food of 1 mg/kg. For specialized food-contact applications where an assumed ratio of 10 g food per in² is not appropriate, such as in dual-ovenable trays and microwave heat-susceptor applications, notifiers should use the lowest ratio from the actual food-contact applications and provide justification for the ratio selected.

      3. Analytical Methods

         Notifiers should submit the following for each method:

         1. DESCRIPTION OF THE METHOD.   A detailed description that can be followed by an experienced analytical chemist. The description should include discussions on the procedure's accuracy, precision, selectivity, limit of quantitation (LOQ), and limit of detection (LOD)⁽⁷⁾. If a literature reference is available, a copy should be included in the PMN.

         2. STANDARD CURVES.   Standard curves or calibration curves obtained by analyzing a prepared medium fortified with several known amounts of analyte to obtain concentrations both greater than, and less than, the concentration of migrant in the test solutions. The prepared medium may be the pure solvent, a solution of known ionic strength, etc. The data points from which the standard curve is derived should bracket the concentration of the migrant in the test solution. An analyte concentration of 1 mg/kg determined from a standard curve obtained from concentrations of 10, 15 and 20 mg/kg would be unacceptable. The correlation coefficient and standard errors of the Y intercept and the slope should be reported with the standard curve.

         3. EXAMPLES OF SPECTRA OR CHROMATOGRAMS.   Sample spectra and chromatograms, clearly identifying and labeling all major peaks to avoid ambiguities in interpretation.

         4. EXAMPLE CALCULATIONS.   Example calculations relating the data obtained from instrumental methods to the reported levels (preferably in milligrams migrants per square inch of sample surface area) provide the reviewer with an internal check on the reported method.

         5. VALIDATION OF ANAYLTICAL METHODS.   Validation of a method's intended use, the determination of accuracy and precision, usually involves: 1) replicate analyses of appropriate matrices fortified with known amounts of the analyte, at concentrations similar to those encountered in the migration studies, and 2) determination of the percentage recovery of the fortified analyte. In cases where a polymer adjuvant is the subject of interest, test solutions of the polymer formulated without the adjuvant may serve as the matrix for fortification and recovery measurements. Recovery is defined as the difference between measured analyte levels in the fortified and unfortified matrices. Percent recovery is the recovery divided by the fortified level times 100, i.e., if "a" is the measured level in the unfortified solution, "b" is the measured level in the fortified solution and "c" is the fortification level, then percent recovery equals (b-a)/c x 100.

            If migration test solutions are fortified, they should be fortified before analytical workup but after the prescribed test time, e.g., 240 hours. FDA recommends that the actual test solutions be fortified and not the pure food simulants. Fortification of pure simulants instead of the test simulants is probably the most common deficiency in the validation section of an analytical method.

            The notifier should perform fortification and recovery experiments using three (3) sets of triplicate samples of the test simulants with each set fortified at a separate level. The fortification levels should be one-half (½), one (1), and two (2) times the measured concentration of the analyte in the food simulant. In the event that the FCS is not detected, the notifier should determine the LOD for the method. For quantifiable levels of the analyte, acceptable recoveries should meet the following criteria:

            Levels in food or food simulants(a)	Acceptable average recovery	Acceptable relative standard deviation
            <0.1 mg/kg	60-110%	<20%
            >0.1 ppm	80-110%	<10%
            (a)If 0.001 mg of a substance is extracted from one square inch of packaging material into 10 grams of food or food simulant, the estimated concentration in food is 0.1 mg/kg.

            In evaluating the precision of the analytical method, the variability arising from analyses of individual samples can be eliminated by performing triplicate analyses on a homogeneous composite (a blend of the triplicate samples) where practicable.

            Other validation procedures may be appropriate depending on the particular analysis. For example, analysis of the same test solution by two independent analytical methods would be acceptable validation. Similarly, the method of standard additions is an acceptable alternative in certain cases, such as metal analysis by atomic absorption spectroscopy. In this case, the notifier should fortify the matrix at two separate concentrations (at least) in addition to the unfortified concentration, and verify the linearity of the standard addition curve by calculation of the least squares correlation coefficient (r should be >0.995).

            The notifier should submit representative spectra or chromatograms from validation analyses of fortified and blank samples. Spectra or chromatograms of the "blank" should be submitted to facilitate the verification of the absence of interferences. An illustrative example appears in Appendix III.

      4. Migration Database

         Migration data for specific migrant/polymer/food simulant systems at given temperatures that exhibit a predictable migration-time behavior, e.g., Fickian diffusion, may be used to predict migration at other temperatures. Thus, the need for migration studies for new applications, which in certain cases such as high temperature applications may be difficult to perform, may be reduced.

         For example, migration data obtained over 10 days (240 h) at 40°C that exhibits Fickian behavior, in combination with migration data obtained at other temperatures (e.g., 60°C and 80°C), may be extrapolated by means of an Arrhenius plot to predict migration under retort conditions (121°C/2 h and 40°C/238 h), if no apparent change in polymer morphology, such as glass transition or polymer melting, is expected between 30°C and 130°C. Apparent diffusion coefficients, D, at 121°C for each migrant/polymer/food simulant can be obtained from a plot of ln D vs 1/T(K). Thus, migration for 2 hours at 121°C can be estimated and added to migration after 238 hours at 40°C to obtain total migration expected for retort and ambient storage conditions. The density and thickness of the polymer sample and initial concentration of the migrant in the polymer are also necessary for the calculations.

         The FDA migration database is intended as a resource for migration data, including diffusion coefficients and relevant polymer/additive properties. FDA continues to compile migration data from various sources for use in estimating migration levels for FCSs. Reliable migration data, e.g., data that follow Fickian diffusion, submitted in support of a PMN would be added to the database. In addition, only migration levels that have been measured at three or more time intervals for a given temperature will be considered for inclusion in the migration database. Notifiers may submit suitable data for inclusion into the database in the form of a letter, as part of a notification, or in a Food Additive Master File. The FDA migration database (hardcopy) is available through the Freedom of Information Act (FOIA) (also see the CFSAN website at http://vm.cfsan.fda.gov/~dms/foia.html).

      5. Migration Modeling

         FDA recommends that a notifier submit relevant and reliable data on an FCS for use in estimating migration levels in food and, in turn, potential consumer exposure. As discussed above, migration levels in food are typically estimated based on the results of migration testing under the anticipated conditions of use or, in certain cases, under the assumption of 100% migration of the FCS to food. These two approaches are adequate in most instances.

         A third alternative involves migration modeling. One simple approach to modeling migration for specific migrant/polymer/food simulant systems, based on select experimental data, was discussed above in Section II.D.4. However, the mathematical modeling of migration using the basic principles of diffusion has not found widespread application, largely due to the lack of necessary material constants such as diffusion coefficients. Diffusion coefficients may often be found in the open literature or the FDA migration database. In all cases, the source of any material constants used in migration modeling should be appropriately referenced.

         Recently, however, semi-empirical methods have been developed to determine migration levels with limited or, in certain cases, no migration data (see, e.g., (Limm and Hollifield, 1996) and (Baner, et al, 1996)). These diffusion models rely on estimation of diffusion coefficients based on the nature of the migrant and the physical properties of the polymer. They may be useful substitutes for, or additions to, experimental data under limited circumstances. Several caveats should be considered in the application of such diffusion models. First, distribution of the migrant in the polymer is considered isotropic. Non-isotropic distribution, whether intentional or unintentional, would be expected to result in non-Fickian migration. Two, other aspects of migration, such as partitioning, mass transfer, polymer morphology, shape/polarity of the migrant, and plastization of the polymer are not considered in these models. These factors should be carefully considered when deriving migration levels to food using modeling techniques.

   5. CONSUMER EXPOSURE

      Migration data developed using the procedures outlined in Section II.D. are intended to provide estimates of the highest level of migration to food that might result from the anticipated use of the FCS. FDA estimates probable exposure to the FCS by combining the migration data with information on uses of food-contact articles that may contain the FCS (i.e., on the fraction of a person's diet likely to contact packaging materials containing the FCS).

      From a given concentration of the FCS in the daily diet, the estimated daily intake (EDI) is calculated as the product of that concentration and the total food intake, assumed to be 3000 grams per person per day (solids and liquids). A concentration in the daily diet of 1 ppm corresponds to an EDI of 1 x 10^-6 g FCS/g food x 3000 g food/person/day or 3 mg/person/day.

      Both the concentration in the daily diet and the EDI from the subject PMN and the cumulative EDI (CEDI) from all regulated uses and effective PMNs are used by FDA in the safety evaluation of an FCS. The CEDI of the FCS should be used in determining the types of toxicity studies used to establish safety under the proposed conditions of use. Toxicological data recommendations for several tiers of CEDIs resulting from all proposed and permitted uses of the FCS, including regulated uses, uses that were the subject of previous PMNs, and the use in the subject PMN, are described in the document entitled "Preparation of Premarket Notifications for Food Contact Substances: Toxicology Recommendations".

      1. Calculation of Exposure
         1. CONSUMPTION FACTOR.   The term "Consumption Factor" (CF) is used to describe the fraction of the daily diet expected to contact specific packaging materials. The CF represents the ratio of the weight of all food contacting a specific packaging material to the weight of all food packaged. CF values for both packaging categories (e.g., metal, glass, polymer and paper) and specific food-contact polymers are summarized in Table I of Appendix IV. These values were derived using information on the types of food consumed, the types of food contacting each packaging surface, the number of food packaging units in each food packaging category, the distribution of container sizes, and the ratio of the weight of food packaged to the weight of the package. These values may, however, be modified as new packaging information is received. Several of the values contained in Table I and Table II of Appendix IV have been updated since the 1995 document entitled "Recommendations for Chemistry Data for Indirect Food Additive Petitions".

            When FDA computes exposure to an FCS, the Agency assumes that the FCS will capture the entire market for which it is intended for use. This approach reflects both uncertainties about likely market penetration as well as limitations in the data surveyed. Thus, if a company proposes the use of an antioxidant in polystyrene, it is assumed that the antioxidant will be used in all polystyrene manufactured for food contact. In certain cases where an adjuvant is intended for use in only a part of a packaging or resin category, a lower CF representing the coverage that is sought may be used. For example, if a stabilizer is intended for use only in rigid and semirigid (polyvinyl chloride) (PVC), a CF of 0.05 rather than 0.1 could be used in estimating exposure since only about 50% of all food-contact PVC could contain the stabilizer. Another example is the division of polystyrene into impact and non-impact categories (see Table I, Appendix IV.). In order to reduce conservatisms, notifiers are encouraged to submit as detailed information as possible on the anticipated resin or packaging market(s) that may be captured by articles manufactured from the FCS.

            When new products are introduced, they will initially be treated as replacement items for existing technology. FDA generally makes estimates based on the assumption that the new product will capture the entire market. For example, the retortable pouch was initially treated as a replacement for coated metal cans and was assigned a CF of 0.17. As additional information on actual use of the retortable pouch became available, the CF was lowered to 0.05. In certain cases, the submission of resin or packaging market data may lead to the use of a lower CF.

         2. FOOD-TYPE DISTRIBUTION FACTOR.   Before migration levels can be combined with CF values to derive estimates of probable consumption, the nature of the food that will likely contact the packaging material must be known. Migration into a fatty-food simulant, for example, will be of little use in estimating probable exposure if the packaging material is used exclusively to package aqueous food. To account for the variable nature of food contacting each packaging material, "food-type distribution factors" (f_T) have been calculated for each packaging material to reflect the fraction of all food contacting each material that is aqueous, acidic, alcoholic and fatty. Appropriate f_T values for both packaging categories and polymer types appear in Table II of Appendix IV.

         3. CONCENTRATION IN THE DAILY DIET AND EDI.    FDA recommends the following approach for calculating the concentration of the FCS in the daily diet. The concentration of the FCS in food contacting the packaging material, <M>, is derived by multiplying the appropriate f_T values by the migration values, M_i, for simulants representing the four food types. This, in effect, scales the migration value from each simulant according to the actual fraction of food of each type that will contact the packaging material.

            <M> = f_{aqueous and acidic}(M _{10% Ethanol})+f_alcohol (M _{50% Ethanol})+f_fatty(M_fatty)

            where M_fatty refers to migration into a food oil or other appropriate fatty-food simulant.

            The concentration of the FCS in the diet is obtained by multiplying <M> by CF. The EDI is then determined by multiplying the dietary concentration by the total weight of food consumed by an individual per day. FDA assumes that an individual consumes 3kg of food (solid and liquid) per day (see Appendix IV. for sample calculations):

            EDI = 3 kg food/person/day x <M> x CF

         4.    CUMULATIVE EXPOSURE.   If the FCS that is the subject of a PMN is already regulated for other uses in 21 CFR 170-199, or has been the subject of previous effective PMNs, FDA generally estimates the cumulative exposure to the FCS from the proposed and permitted uses (see the example in Appendix IV.). Information on the regulatory status of an FCS may be obtained by inspection of 21 CFR 170-199, searching the CFR on the Government Printing Office (GPO) World Wide Website at http://www.access.gpo.gov/nara/cfr/index.html, or contacting FDA directly. Information on effective PMNs for an FCS may be obtained through the FDA website or by contacting FDA directly. An estimate of cumulative exposure for the regulated and notified uses of an FCS can be obtained by contacting FDA. FDA also maintains a database of cumulative EDIs for food contact substances on the agency's internet site.

            The approach outlined above is designed to deal with the majority of FCSs intended for single-use. For estimating dietary exposures to components of repeat-use items and articles used in or with food processing equipment, exposure estimates will also consider estimates of the amount of food to be contacted during the service life of the food-contact article (see Appendix II. Section 4.).

      2. Exposure Refinement

         Exposure estimates will, in general, be made using the aforementioned procedures. More refined exposure estimates may be possible, however, with additional information provided in a PMN. For instance, subdividing packaging or resin categories could reduce the calculated exposure by lowering the CF for the category. The division of PVC into rigid and plasticized categories and PS into impact and non-impact categories, cited above, are two examples. Another example is the division of polymer coatings for paper into subcategories, such as poly(vinyl acetate) coatings, styrene-butadiene coatings, etc. If an FCS is to be used solely in styrene-butadiene coatings for paper, use of the CF for polymer-coated paper (0.2, Appendix IV. Table 1), would be a gross exaggeration. As noted above, FDA encourages the submission of marketing information that may be used to subdivide the packaging market(s) anticipated for articles manufactured from the FCS.

         In those cases where the nature of the coverage requested may necessitate more detailed information or where a notifier believes that exposure will be overstated by simply selecting CF and f_T values presented in Appendix IV., data of the following type may be submitted to facilitate calculations of CF and f_T values for materials likely to contain the FCS:

         1. Estimates of the total amount of food in contact with the packaging material determined using either:

            (1)  package unit data (number of units and their size distribution), or

            (2)  pounds of packaging material produced for food contact, container size distribution, and ratios of weight of food packaged to weight of package.

         2. Characterization of the foods that might contact the food package, along with supporting documentation, and the likely f_T values.

         3. Information that would demonstrate that only a fraction of a packaging or resin category would be affected by the coverage sought.

         4. Technological limitations that could affect the type of food contacted or the fraction of the diet that might be contacted.

APPENDIX I.

FATTY-FOOD SIMULANTS FOR SPECIFIC POLYMERS

A food oil is the most extreme example of a fatty food. If contact with fatty foods is anticipated, FDA recommends that a notifier conduct migration studies using a food oil as the food simulant. In addition to food oils such as corn and olive oil for which extensive migration data already exist, the use of HB307 (a mixture of synthetic triglycerides, primarily C₁₀, C₁₂, and C₁₄) as a fatty-food simulant has previously been recommended. Studies in FDA laboratories have shown that Miglyol 812, a fractionated coconut oil having a boiling range of 240-270°C and composed of saturated C₈ (50-65%) and C₁₀ (30-45%) triglycerides, is also an acceptable alternative. Since use of these oils for FCS migration may not always be practicable, the use of aqueous-based solvents that simulate the action of these liquid fats is sometimes necessary. While it seems unlikely that one solvent will be found that simulates the action of a food oil for all food-contact polymers, the following list presents polymers for which adequate data exist to support the use of aqueous-based solvents as fatty-food simulants. The recommendation of these solvents is based upon studies done at FDA, at the National Institute of Standards and Technology (formerly The National Bureau of Standards), and by Arthur D. Little, Inc. under contract to FDA (a list of general references pertaining to these studies is shown in Appendix V). For polymers other than those listed below, notifiers should consult FDA before undertaking any migration experiments.

1. Polyolefins complying with §177.1520 and ethylene - vinyl acetate copolymers complying with §177.1350	95% or absolute ethanol
2. Rigid (polyvinyl chloride)	50% ethanol
3. Polystyrene and rubber-modified polystyrene	50% ethanol
4. Poly(ethylene terephthalate)	50% ethanol

Absolute or 95% ethanol has been found to be an effective fatty-food simulant for polyolefins; however, it appears to exaggerate migration for other food-contact polymers.

Previous test protocols (prior to 1988) recommended the use of heptane as a fatty-food simulant. To account for the aggressive nature of heptane relative to a food oil, division of migration values by a factor of five was permitted. Studies have shown, however, that the exaggerative effect of heptane relative to a food oil varies over orders of magnitude depending on the polymer extracted. Thus, heptane is no longer recommended as a fatty-food simulant. However, we recognize that in cases where very low migration is anticipated, such as for inorganic adjuvants or certain highly cross-linked polymers, heptane can be useful due to the ease of analytical workup. Because of the known variance in the exaggerative effect of heptane relative to food oil, if heptane is used, migration values will generally not be divided by any factor unless there is adequate justification.

APPENDIX II.

SELECTED MIGRATION TESTING PROTOCOLS

The following migration testing protocols are intended to simulate most anticipated end-use conditions of food-contact articles. These protocols are based on the premise that migration to aqueous- and fatty-based foods is typically diffusion-controlled within the polymer, strongly affected by the temperatures encountered during food contact, and further modified by the solubility of the FCS in the foods. Therefore, migration testing with food simulants at the highest temperatures to be experienced by the package during food contact is recommended. Testing with actual fatty foods is also an option, although determination of the analytes of interest is often very difficult. In those instances where the expected use conditions are not adequately simulated by these protocols or testing with food simulants at the highest anticipated food-contact temperature is not practical, alternatives to those protocols presented below should be developed in consultation with FDA.

1. General Protocols (Single-Use Applications) Corresponding to Condition of Use

   As noted in Appendix I., migration to fatty foods is evaluated using a fatty food, a pure liquid fat, or, alternatively, aqueous ethanol solutions when analytical limitations preclude sensitive analyses. As noted in Section II.D.1.c., migration to aqueous, acidic, and low-alcoholic foods is generally evaluated using 10% ethanol and migration to high-alcohol foods is generally evaluated using 50% ethanol.

   The migration protocols given below are intended to model thermal treatment and extended storage conditions for polymers, such as polyolefins, used with food at temperatures above their glass transition temperatures. The extended storage period generally involves testing at 40°C for 240 hours (10 days). As discussed in Section II.D.1.d., migration data obtained at 10 days for polymers used below their glass transitions temperature should be extrapolated to 30 days to better approximate migration levels expected after extended storage at ambient conditions.

   FDA recommends the following approaches:

   1. High temperature, heat sterilized or retorted above 100° C (212°F)

      10% Ethanol(a)	121°C (250°F) for two hours
      50% Ethanol	71°C (160°F) for two hours
      Food Oil (e.g., corn oil) or HB307 or Miglyol 812	121°C (250°F) for two hours
      50% or 95% Ethanol(a),(b)	121°C (250°F) for two hours
      (a)Requires a pressure cell or autoclave, see Appendix V. Appropriate safety precautions should be exercised when using equipment generating pressures above 1 atmosphere. (b)Depends on food-contact layer, see Appendix I.

      After two hours at elevated temperatures, continue the tests at 40°C (104°F) for 238 hours to a total of 240 hours (10 days). Analyze the test solutions at the end of the initial two hour period, and after 24, 96 and 240 hours.

   2. Boiling water sterilized.   The protocol remains the same as for Condition of Use A. except that the highest test temperature is 100°C (212°F).

   3. Hot filled or pasteurized above 66°C (150°F).    Add solvents to the test samples at 100°C (212°F), hold for 30 minutes, and then allow to cool to 40°C (104°F). Maintain the test cells at 40°C (104°F) for ten days with samples taken for analysis after the intervals indicated for the previous protocols. If the maximum hot fill temperature will be lower than 100°C (212°F), test solvents may be added at this lower temperature. Alternatively, perform migration studies for 2 hours at 66°C (150°F) followed by 238 hours at 40°C (104°F). For the alternative method, the longer time at the lower temperature (2 hours at 66°C vs 30 minutes at 100°C) compensates for the shorter time at 100°C.

   4. Hot filled or pasteurized below 66°C (150°F).    The recommended protocol is analogous to that for C. except that all test solvents are added to the test samples at 66°C (150°F) and held for 30 minutes before cooling to 40°C (104°F).

   5. Room temperature filled and stored (no thermal treatment in the container).    The notifier should conduct migration studies for 240 hours at 40°C (104°F). Analyze the test solutions after 24, 48, 120 and 240 hours.

   6. Refrigerated storage (no thermal treatment in the container)    The recommended protocol is identical to that for E. except that the test temperature is 20°C (68°F).

   7. Frozen storage (no thermal treatment in the container) The recommended protocol is identical to F. except that the test time is five (5) days.

   8. Frozen or refrigerated storage; ready-prepared foods intended to be reheated in container at time of use:

      10% Ethanol(a)	100°C (212°F) for two hours
      Food Oil (e.g., corn oil) or HB307 or Miglyol 812TM	100°C (212°F) for two hours
      50% or 95% Ethanol(a),(b)	100°C (212°F) for two hours
      (a)Requires a pressure cell or autoclave, see Appendix V. (b)Depends on food-contact layer, see Appendix I.

      Applications involving the heating and cooking of food at temperatures exceeding 121°C (250°F) are not included under conditions of use A.-H. Migration testing protocols for these applications are discussed in Section 11. of this Appendix.

2. Adjuvants for Polyolefins
     

   In general, under identical testing conditions, levels of migrants from low-density polyethylene (LDPE) are higher than from high-density polyethylene (HDPE) or polypropylene (PP). Migration studies done solely on LDPE (complying with §177.1520(a)(2)) at 100°C (approximately the highest temperature at which LDPE remains functional) are, therefore, generally sufficient to provide coverage for all polyolefins including PP, which may be used for retort applications. In such a case, the CF for all polyolefins (CF = 0.33) generally will be used instead of the individual CF for LDPE (see Appendix IV. Table I). Levels of migrants from PP are generally higher than from HDPE. Thus, migration studies conducted with PP are generally adequate to permit coverage in HDPE under the same conditions of use. In such a case, the combined CF of PP and HDPE will be used in deriving exposure estimates. However, when seeking coverage for all polyolefins it is usually advantageous to perform migration testing on HDPE, PP and LLDPE, complying with §177.1520, as well as LDPE. By doing this, actual migration values for these polyolefins, which will likely be lower that those obtained from LDPE, may be used to calculate the EDI.

3. Adjuvants for Polymers (other than Polyolefins)
   Adjuvants for More than One Polymer

   The migration testing protocols for polymers other than polyolefins are the same as those in Section 1. of this Appendix. Consult Appendix I. for the recommended fatty-food simulant.

   If use of an FCS is sought without limitation to specific polymers, the notifier may obtain this broad coverage by testing with an unoriented LDPE sample complying with §177.1520(a)(2). The test protocol depends on the anticipated conditions of use (refer to Section 1. of this Appendix). If the most rigorous applications correspond to Condition of Use A. (Section 1.A.), the test temperature should be the highest temperature at which the polymer remains functional (ca. 100°C for LDPE). The CF for all polymers (Appendix IV. Table 1, CF = 0.8) should be used with the migration data to calculate the concentration of the FCS in the daily diet. In general, a lower calculated concentration in the daily diet will result if a series of representative polymers are separately tested and individual consumption factors are applied (refer to the examples in Appendix IV.). A notifier should consult with FDA to determine which representative polymers should be tested.

4. Articles Intended for Repeated Use

   A notifier should test the article with 10% and 50% ethanol and a food oil (e.g., corn oil) or other fatty-food simulant (e.g., HB307 or Miglyol 812) for 240 hours at the highest intended temperature of use. The test solutions should be analyzed for migration of the FCS after 8, 72, and 240 hours. Notifiers should provide estimates of the weight of food contacting a known area of repeat-use article in a given time period as well as an estimate of the average lifetime of the article. Together with the migration data, this will allow calculation of migration to all the food processed over the service life of the article.

   In the case of an adjuvant in a repeat-use article, FDA strongly recommends an initial calculation of a "worst case" level in food by assuming 100% migration of the adjuvant over the service life of the article and dividing that value by the quantity of food processed. If this calculated concentration is sufficiently low, migration studies will be unnecessary.

5. Coatings for Cans

   The migration testing protocol is usually that outlined in Section 1.A. of this Appendix for high temperature, heat sterilized or retorted products. If broad coverage is sought for all types of coatings, the notifier should consult with FDA to determine which coatings should be tested. For use conditions less severe than retort sterilization at 121°C, notifiers should follow the migration test protocols outlined in Sections 1.B.-G. of this Appendix which most closely approximate the most severe expected use conditions.

6. Uncoated & Clay-Coated Papers with Latex Binders

   These papers are intended for contact with food at temperatures less than 40°C for short periods of time. The FDA recommended protocol is the following:

   10% Ethanol	40°C (104°F) for 24 hours
   50% Ethanol	40°C (104°F) for 24 hours
   Food Oil (e.g., corn oil) or HB307 or Miglyol 812	40°C (104°F) for 24 hours

   Migration studies conducted on uncoated or clay-coated papers typically result in a high level of extractives due to the large number of low-molecular weight, soluble components in both paper and paper coatings. Therefore, when total nonvolatile or chloroform-soluble total nonvolatile extractives are determined for a paper coating, FDA recommends that the corresponding extractives should not be subtracted from uncoated paper as a blank correction. Rather than using paper as a support for the coating, it is often useful to apply the coating to a suitable inert substrate, such as glass or metal, for use in migration testing. For a new adjuvant in paper coatings, FDA recommends the analysis of the test solutions for the unregulated adjuvant. For a new polymer used in paper coatings, FDA recommends the analysis of the test solutions for constituent oligomers and monomers.

7. Specially Treated Papers

   This class includes such types as fluoropolymer- and silicone-treated papers that have oil-resisting and heat-resisting properties. The specific protocol depends on the particular uses anticipated. It is recommended that the notifier either devise a protocol and submit it to FDA for comment or request comment from FDA about appropriate test conditions.

8. Adhesives (Room temperature or below)

   In previous chemistry documents for indirect additives, migration tests were not recommended for adhesives intended for use at room temperature or below and in accordance with §175.105. (High temperature applications are discussed in Section 9.). This recommendation was based on consideration of subparagraph (a)(2) of §175.105 which specifies that the adhesive is either separated from food by a functional barrier, or the quantity of adhesive that contacts aqueous and fatty food is limited to the trace amount at seams and edges.

   If a notifier proposes to use the adhesive or adhesive component in concordance with the limitations of §175.105, migration levels for the notified substances will generally be assumed to be no greater than 50 ppb, as is the case for petitioned adhesives components. Applying a CF of 0.14 for adhesives gives a dietary concentration of 7 ppb. If the assumptions of §175.105 cannot be supported, notifiers should submit data or calculations to model the intended use of any adhesive component. If a notifier wishes to perform migration testing, multilaminate samples should be fabricated with the maximum anticipated amount of the adhesive component and with the minimum thickness of the food contact layer. The migration protocol corresponds to condition of use E. Alternatively, migration levels in food can be estimated based on migration modeling (see Section II.D.5.).

9. Laminates & Coextrusions

   Components of multilayer structures used above room temperature are the subject of two regulations. One covers laminates used in the temperature range 120°F (49°C)-250°F (121°C) (21 CFR 177.1395) and the other covers laminate structures used at temperatures of 250°F (121°C) and above (21 CFR 177.1390). Layers not separated from food by barriers preventing migration during expected use must be listed in these regulations, or be the subject of an effective PMN, unless they are authorized elsewhere for the intended use conditions as specified in 21 CFR 177.1395(b)(2) and 21 CFR 177.1390(c)(1). Test protocols presented in Sections 1.A.-H. may be appropriate for evaluating the level of migration from non-food-contact layers of some laminate structures. End uses that differ considerably from those considered in these Guidelines, however, should be the subject of special protocol development in consultation with FDA.

10. Boil-In-Bags

    FDA recommends the protocol employed in Condition of Use C.

11. Special High-Temperature Applications

    Advances in packaging technology have led to the development of food packaging materials that can withstand temperatures substantially exceeding 121°C (250°F) for short periods of time for the purposes of heating and cooking of ready-prepared food. FDA recommends use of the following protocols for migration testing of dual-ovenable containers and microwave heat susceptor materials.

    1. DUAL-OVENABLE TRAYS

       For high temperature oven use (conventional and microwave), FDA recommends migration testing at the maximum intended conventional oven cooking temperature for the longest intended cooking time, using a food oil, or a fatty-food simulant such as Miglyol 812.

    2. MICROWAVABLE CONTAINERS

       The temperature ultimately experienced by a food-contact material when cooking foods in a microwave oven is dependent on many factors. Some of these are food composition, heating time, mass and shape of the food, and shape of the container. For example, food with mass in excess of 5 g/in² container surface area and having a thick shape will require longer cooking times to achieve the desired degree of interior cooking than if it had a lower mass-to-surface area ratio and were thinner. Because the ultimate temperature of the container will depend on many factors and is, therefore, not readily predicted, it is recommended that notifiers consult with FDA on any planned testing protocol prior to initiating migration testing.

    3. MICROWAVE HEAT-SUSCEPTOR PACKAGING

       The high temperatures attained by packaging using susceptor technology may result in (a) the formation of significant numbers of volatile chemicals from the susceptor components and (b) loss of barrier properties of food-contact materials leading to rapid transfer of nonvolatile adjuvants to foods. Studies by FDA, with hot vegetable oil in contact with a susceptor, have shown that the susceptor materials liberate volatile chemicals that may be retained in the oil at parts-per-billion (ppb) levels. FDA recommends the use of the protocol outlined in an article by McNeal and Hollifield (McNeal and Hollifield, 1993) for the identification and quantification of volatiles from susceptors.

       To isolate and identify the total available nonvolatile extractives, notifiers should perform Soxhlet extractions on finely shredded portions of laminated susceptor materials using polar and nonpolar solvents as outlined in Appendix X1 of ASTM method F1349-91. Migration protocols for UV-absorbing nonvolatiles are also outlined in ASTM method F1349-91 and in an article by Begley and Hollifield (Begley and Hollifield, 1991). The ASTM method relies on the determination of a time-temperature profile based on cooking a food product according to label directions, for the maximum cooking time. The temperature reached by a microwave heat susceptor, however, is dependent on the amount and characteristics of the food product. Testing methods should involve a standard set of conditions that represent the maximum anticipated use conditions. Therefore, FDA recommends that migration studies be conducted in a manner similar to that outlined in the article by Begley and Hollifield. The recommended standard test conditions are as follows:

         1)   use laminated susceptor stock representative of the proposed application(s);

         2)   use a microwave oven with an output wattage on the order of 700 watts;

         3)   use a maximum microwave time of 5 minutes;

         4)   use an oil mass-to-susceptor surface area on the order of 5 g/in²; and

         5)   use a water load on the order of 5 g/in².

       Exposure estimates may be based, in the absence of validated migration studies, on the assumption of 100% migration of the total nonvolatile extractives to food, as determined by Soxhlet extractions.

       Validated migration protocols for the direct determination of aliphatic migrants are not available at this time. However, the amount of aliphatic migrants may be estimated by subtracting the UV-absorbing nonvolatiles and inert materials from the total nonvolatiles obtained by Soxhlet extraction (see Appendix X1 in ASTM method F1349-91). Exposure estimates for aliphatic migrants should be based on the assumption of 100% migration to food.

12. Colorants for Plastics

    Some colorants, pigments in particular, may be quite insoluble in the food simulants 10%- and 95%-ethanol. In such cases, solubility information may provide a basis for an alternative to migration testing for evaluating worst-case exposure since migration levels would not be expected to exceed the limits of solubility of the colorant at the proposed use temperature. If the colorant is to be used in all plastic packaging, for which a CF = 0.05 would be used, a solubility below ca. 100 µg/kg at 40°C would lead to a dietary concentration no greater than 5 ppb under conditions as severe as condition of use E. A solubility less than 10 µg/kg would lead to an exposure below the threshold level of 0.5 ppb dietary concentration (see 21 CFR 170.39).

13. Dry Foods with Surface Containing No Free Fat or Oil
    (21 CFR 176.170(c), Table 1, Food Type VIII)

    Although studies have shown migration of certain adjuvants into dry foods (e.g., low molecular weight adjuvants in contact with porous or powdered foods), at the present time no migration testing is recommended.

14. Wet-End Additives used in the Manufacture of Paper and Paperboard

    Paper additives used in the wet-end of papermaking include those designed to improve the papermaking process, such as processing aids, and those designed to modify the properties of the paper, such as functional aids. Functional aids, mostly organic resins or inorganic fillers, are designed to bond to the paper fibers and, thus, are substantive to paper. For those FCSs that are substantive to paper, migration studies should be conducted and the test solutions analyzed for constituents of the substance. For example, in the case of a polymeric retention aid, the test solutions should be analyzed for constituent oligomers and monomers. On the other hand, processing aids are intended to remain with the process water slurry and, thus, are generally not substantive to paper. Exposure estimates for non-substantive additives may be based on migration studies, or alternatively, on scenarios involving partitioning of the additive between paper fibers and slurry water. The following example illustrates this approach:

    Consider an adjuvant added prior to the sheet-forming operation in the manufacture of paper. The intended use level is reported to be 10 mg/kg in the slurry. Since the additive is not substantive to paper, the mass of water (containing the additive) in contact with the pulp at the point in the papermaking process where the slurry enters the drier determines the level of the adjuvant retained in paper. Prior to entering the driers, the slurry is concentrated to contain approximately 33% pulp and 67% water. This corresponds to an adjuvant level of 20 mg/kg relative to the pulp. Assuming that finished paper contains 92% pulp, a paper basis weight of 50 mg/in², 100% migration of the adjuvant to food, and that 10 g of food contacts 1 in² paper, the results in an adjuvant concentration in food of 0.09 mg/kg, or 90 µg/kg. Applying a CF of 0.1 for uncoated and clay-coated paper gives a dietary concentration of 9 ppb.

Polyethylene film containing a new antioxidant was subjected to migration testing with 10% ethanol. The test solutions were analyzed for antioxidant migration. Tests were carried out in separate cells each containing 100 in² of film. Four sets of test solutions (in triplicate) were analyzed at 2, 24, 96 and 240 hours for a total of 12 test solutions. After each time interval, each solution from one set was evaporated to dryness, the residue dissolved in an appropriate organic solvent, and a known aliquot injected into a gas chromatograph.

Validation experiments are carried out with the set of test simulants exhibiting the highest level of antioxidant migration. To validate the analytical method, an additional three sets (in triplicate) using 10% ethanol can be run for 240 hours. Each set of these test solutions can then be fortified with the antioxidant at levels corresponding to one-half (1/2), one (1) and two (2) times, respectively, the average migration value determined for the regular (unfortified) 240 hour test solutions.

Instead, the notifier decided to carry out one large test using enough film and solvent for twelve analyses (three at each of the four time intervals). After 240 hours, the test solution was divided into twelve (12) equal solutions (i.e., four sets of triplicate samples). One set (three solutions) was found to contain antioxidant at an average level of 0.00080 mg/in². This value corresponds to 0.080 mg/kg in food if it is assumed that 10 grams of food contacts 1 in² of film. Of the remaining nine solutions (three sets), three solutions were fortified at concentrations corresponding to 0.00040 mg/in², three were fortified at 0.00080 mg/in² and three were fortified at 0.00160 mg/in². Each solution was worked up and analyzed as described above. To illustrate the recovery calculations, the results for the set of three solutions fortified at one-half times the average migration (0.00040 mg/in²) are summarized in the following table:

The average percent recovery is 74.2%, and the relative standard deviation is 15.2%. These are within the limits specified (see Section II.D.3.e.) for a concentration in food of 0.080 mg/kg (percent recovery 60-110%, relative standard deviation not exceeding 20%). If the corresponding percentages for the other two fortification levels are also within these limits, the validation for the 10% ethanol migration studies would be acceptable. The actual validation procedure used will, of course, depend on the particular type of analysis.

This appendix summarizes packaging data recommended by FDA for evaluating exposure to FCS. An example of how these data are combined with levels of an FCS in food is also presented. A more complete discussion of the source of these data and their use in exposure calculations is presented in Section II.E.

Examples of Exposure Estimate Calculations

The following hypothetical examples are intended to illustrate the calculation of the concentration of an FCS in the daily diet (CF x <M>, i.e., the fraction of food in the diet contacting the packaging material times the average concentration of the FCS in food) and its EDI.

Example 1

A PMN is received that describes the use of a new antioxidant at a maximum level of 0.25% w/w in polyolefins contacting food at or below room temperature (see Appendix II. Sections 1.E. through 1.G.). Migration values from LDPE reported to FDA for the three food simulants are given below:

The notifier used a solvent volume-to-exposed surface area ratio of 10 mL/in². Therefore, solution concentrations are essentially equivalent to food concentrations (under the assumption that 10 g food contacts 1 in² of surface area). The CF and f_Ts for polyolefins are given in Tables I and II, respectively. The <M> for the antioxidant would be calculated as follows:

The concentration of the antioxidant in the daily diet resulting from the proposed use would be:

If there were no other permitted uses, then the EDI would be calculated using the above value:

Example 2

In a subsequent notification, expanded use of the same antioxidant in polycarbonate and polystyrene food contact articles is described. Each polymer would contact food at or below room temperature. Migration levels are given below:

The concentration of the antioxidant in the daily diet resulting from each of the proposed uses is calculated below. A CF of 0.04 for impact polystyrene and a CF of 0.06 for all other polystyrenes was used in the calculation.

Polycarbonates

Polystyrene

Impact Polystyrene

The total concentration of the antioxidant in the daily diet resulting from the additional uses in polycarbonate and polystyrene is approximately 0.032 mg/kg or 0.032 parts per million (ppm).

The contribution to the EDI is:

The cumulative exposure from the previously regulated use (Example 1, 2.5 mg/person/day) and the additional proposed uses would be 2.6 mg/person/day.

APPENDIX V.

REFERENCES

General References

American Society for Testing and Materials (ASTM), E 1303-95, Standard Practices for Refractive Index Detectors used in Liquid Chromatography. ASTM, West Conshohocken, PA 19428-2959.

Arthur D. Little, Inc., July 1983: A Study of Indirect Food Additive Migration. Final Summary Report. 223-77-2360.

Arthur D. Little, Inc., September 30, 1988: High Temperature Migration Testing of Indirect Food Additives. Final Report. FDA Contact No. 223-87-2162.

Arthur D. Little, Inc., August 1990: High Temperature Migration Testing of Indirect Food Additives to Food. Final Report. FDA Contract No. 223-89-2202.

ASTM E 1511-95, Standard Practice for Testing Conductivity Detectors Used in Liquid or Ion Chromatography. ASTM, West Conshohocken, PA 19428-2959.

Baner, A., Brandsch, J., Franz, R. and Piringer, O. , 1996, The Application of a predictive migration model for evaluating the compliance of plastic materials with European food regulations. Food Additives and Contaminants, 13 (5), 587-601.

Begley, T. H. and Hollifield, H. C., 1991, Application of a polytetrafluoroethylene single-sided migration cell for measuring migration through microwave susceptor films. American Chemical Society Symposium Series 473: Food and Packaging Interactions II, Chapter 5, 53-66.

Chang, S., 1984, Migration of low molecular weight components from polymers: 1. Methodology and diffusion of straight-chain octadecane in polyolefins. Polymer, 25, 209-217.

Currie, L. A., 1968, Limit of qualitative detection and quantitative determination, application to radiochemistry. Analytical Chemistry, 40(3), 586-593.

Goydan, R., Schwope, A., Reid, R., and Cramer, G., 1990, High temperature migration of antioxidants from polyolefins. Food Additives and Contaminants, 7(3), 323-337.

Keith, L. H., Crummett, W., Deegan, Jr., J., Libby, R. A., Taylor, J. K., and Wentler, G., 1980, Principles of environmental analysis. Analytical Chemistry, 55, 2210-2218.

Limm, W. and Hollifield, H. C., 1995, Effects of temperature and mixing on polymer adjuvant migration to corn oil and water. Food Additives and Contaminants, 12 (4), 609-624.

Limm, W. and Hollifield, H., 1996, Modelling additive diffusion in polyolefins. Food Additives and Contaminants, 13(8), 949-967.

McNeal, T. P. and Hollifield, H. C., 1993, Determination of volatile chemicals released from microwave-heat-susceptor food packaging. J. AOAC International, 76(6), 1268-1275.

National Bureau of Standards, March 1982: Migration of Low Molecular Weight Additives in Polyolefins and Copolymers. Final Project Report, NBSIR 82-2472. NTIS PB 82-196403, National Technical Information Services, Springfield, VA.

Schwope, A. D. and Reid, R. C., 1988, Migration to dry foods. Food Additives and Contaminants, 5 (Suppl. 1), 445-454.

Schwope, A. D., Till, D. E., Ehntholt, D. J., Sidman, K. R., Whelan, R. H., Schwartz, P. S., and Reid, R. C., 1986, Migration of an organo-tin stabilizer from polyvinyl chloride film to food and food simulating liquids. Deutsche Lebensmittel Rundschau, 82(9), 277-282.

Schwope, A. D., Till, D. E., Ehntholt, D. J., Sidman, K. R., Whelan, R. H., Schwartz, P. S., and Reid, R. C., 1987, Migration of Irganox 1010 from ethylene-vinyl acetate films to foods and food-simulating liquids. Food and Chemical Toxicology, 25(4), 327-330.

Schwope, A. D., Till, D. E., Ehntholt, D. J., Sidman, K. R., Whelan, R. H., Schwartz, P. S., and Reid, R. C., 1987, Migration of BHT and Irganox 1010 from low-density polyethylene (LDPE) to foods and food-simulating liquids. Food and Chemical Toxicology, 25(4), 317-326.

Snyder, R.C. and Breder, C.V., 1985, New FDA migration cell used to study migration of styrene from polystyrene into various solvents. Journal of Association Official Analytical Chemist, 68(4), 770-775.

Till, D., Schwope A. D., Ehntholt, D. J., Sidman, K. R., Whelan, R. H., Schwartz, P. S., and Reid R. C., 1987, Indirect food additive migration from polymeric food packaging materials. CRC Critical Reviews in Toxicology, 18(3), 215-243.

Till, D. E., Ehntholt, D. J., Reid, R. C., Schwartz, P. S., Sidman, K. R., Schwope, A. D., and Whelan, R. H., 1982, Migration of BHT antioxidant from high density polyethylene to foods and food simulants. Industrial & Engineering Chemistry, Product Research and Development, 21(1), 106-113.

Till, D. E., Ehntholt, D. J., Reid, R. C., Schwartz, P. S., Schwope, A. D.; Sidman, K. R., and Whelan, R. H., 1982, Migration of styrene monomer from crystal polystyrene to foods and food simulating liquids. Industrial & Engineering Chemistry, Fundamentals, 21(2), 161-168.

Till, D. E., Reid, R. C., Schwartz, P. S., Sidman, K. R., Valentine, J. R., and Whelan, R. H., 1982, Plasticizer migration from polyvinyl chloride film to solvents and foods. Food and Chemical Toxicology, 20(1), 95-104.

The following are lists of references that contain descriptions, photos, or drawings of migration cells for conducting migration testing for different packaging applications.

Cells for Migration Testing

Conventional Applications

ASTM F34-98, Standard Practice for Construction of Test Cell for Liquid Extraction of Barrier Materials. ASTM, West Conshohocken, PA 19428-2959.

Dow Chemical, Inc., A single-sided migration cell, known as the Dow cell, has been used with food oil at 175°C. The cell is available from: Kayeness, Inc., 115 Thousand Oaks Blvd., Suite 101, P.O. Box 709, Morgantown, PA 19543 (610-286-7555). Model no. D9030.

Figge, K. and Koch, J., 1973, Effect of some variables on the migration of additives from plastics into edible fats. Food Cosmetics Toxicology, 11, 975-988. The cell used was a single-sided cell in contact with food oil at 80°C.

Goydan, R., Schwope, A. D., Reid, R. C., and Cramer, G., 1990. The cell used was a double-sided (immersion), stainless steel cell, with water, 95% ethanol, and oil at 130°C.

Limm, W. and Hollifield, H., 1995. The cell used was a single-sided glass cell with water, food oil, and food at 135°C.

Snyder, R.C. and Breder, C.V., 1985. The cell used was a double-sided (immersion) glass cell with water, 3% acetic acid, 95% ethanol, and oil at 40°C and 50% aqueous ethanol at 70°C. This cell is also specified in ASTM D4754-87 "Standard Test Method for the Two-Sided Liquid Extraction of Plastic Materials Using FDA Migration Cell." ASTM, West Conshohocken, PA 19428-2959.

Till, D.E., Ehntholt, D. J., Reid, R. C., Schwartz, P. S., Sidman, K. R., Schwope, A. D., and Whelan, R. H., 1982. The cells used were glass, single-sided and double-sided (immersion) cells, with water, 3% acetic acid, 95% ethanol, and oil at 40°C.

Microwave Applications

ASTM F1349-91, Standard Test Method for Nonvolatile Ultraviolet (UV) Absorbing Extractables from Microwave Susceptors. ASTM, West Conshohocken, PA 19428-2959.

Begley, T. and Hollifield, H., 1991. The cell was used with food oil at temperatures up to 240°C.

Rijk, R. and De Kruijf, N., 1993, Migration testing with olive oil in a microwave oven. Food Additives and Contaminants, 10(6), 631-645.

1. CAS Registry Numbers for new compounds and assistance with nomenclature can be obtained by writing to Chemical Abstracts Service (CAS) Client Services, 2540 Olentangy River Road, P.O. Box 3343, Columbus, OH 43210, or by visiting their website at http://www.cas.org.

2. Migration into food depends on the chemical structure of the FCS, the nature of the food matrix contacting the FCS, the type of food with which it is in contact, and the temperature and duration of food contact. Prior to the submission of a PMN, a potential notifier may wish to meet or correspond with FDA to discuss appropriate migration testing protocols (see Section III.).

3. In the past, FDA recommended 8% ethanol as an aqueous food simulant. Increasing the ethanol concentration from 8% to 10% will have a minimal impact on migration studies conducted on adjuvant/polymer systems. This change also harmonizes more closely FDA's migration protocols with those of other nations. See the reference list at the end of Appendix II. relating to FDA's development of the use of food simulants.

4. Miglyol 812, a product of Dynamit Nobel Chemicals, is available from HULS America, Inc., 80 Centennial Ave., P. O. Box 456, Piscataway, NJ 08855-0456.

5. HB307 is available from NATEC, Behringstrasse 154, Postfach 501568, 2000 Hamburg 50, Germany.

6. Previous test protocols (prior to 1995) recommended a test temperature of 49°F for 10 days. Recent studies by FDA, however, have shown little difference in migration levels at 49°C and 40°C (104°F). Furthermore, the differences in migration levels between 49°C and 40°C are of even less significance for migration studies requiring elevated temperatures (e.g., 100°C or 121°C) for the first two hours. Up to 80% of the total migration observed over the 10 day period is usually completed within this two hour period at the higher temperature. Therefore, 40°C is regarded as acceptable for migration studies for room-temperature applications and for the portion of the migration test for elevated-temperature applications intended to reflect long term ambient storage.

7. The LOD is the lowest concentration of analyte that the analytical method can reliably detect above a blank (or control). It is preferable that the LOD be determined from analyses of five blank samples. The blank signal (i.e., the analyte response for the blank sample or the width of the baseline close to the actual or expected analyte peak) is measured, and the average signal and standard deviation for the blank are calculated. The signal corresponding to the LOD is located three standard deviations above the average blank signal. The blank signal for the LOD is usually determined from the peak-to-peak noise measured on the baseline close to the actual or expected analyte signal. See American Society for Testing and Materials (ASTM), E 1303-95 or ASTM E 1511-95.

The region for quantitation of the analyte should clearly be above the LOD. The signal corresponding to the LOQ is located ten standard deviations above the average blank signal. See (Currie, 1968) and (Keith., et al, 1980).