Document mBbXpwvLm35Dr78Y5bq9Ymw6O

Detecting sn.all quantities oi residual vinyl chloride monomer S. G. Gilbert, J. R. Giacin, J. R. Morano, & J. D. Rosen Department of Food Science Rutgers University New Brunswick, New jersey Introduction. Attention has recently been focused on the health hazard to the general public from residual vinyl chloride monomer (VCM) in polyvinyl chloride (PVC) based polymers. The need for this scrutiny Ffas arisen from the findings of an unusually high inci dence of angiosarcoma in humans exposed to high concentrations of VCM (ca. 1000 ppm) through occu pational exposure over long periods of time (1). This relationship has been strengthened greatly with direct experimental evidence of angiosarcoma in mice exposed to VCM dosages at levels as low as 50 ppm (inhalation) over an eight month period (2). The concern for the general public arises from the mag nitude of the annual PVC production (about 5 billion pounds in 1973, (3)) and its use in packaging consumer products. There are two distinguishable areas for concern. The first is that of exposure through inhalation of the highly volatile gas, either by occupational exposure through release to the atmosphere or by its use as a propellant. Present regulations by EPA ban the latter (4), while recent OSHA rulings specify a reduction to 1 ppm in the atmosphere of factories involved in PVC production and use (5). Five ppm is allowable for periods shorter than 15 minutes. The second area of concern involves the possibility of exposure by ingestion of VCM as a result of its mi gration to food contacted with PVC in packaging ma terials. We estimate approximately 250 million pounds of PVC may be involved annually in food packaging at present levels (6). The regulatory aspects of this type of contact are within the jurisdiction of FDA under the category of "incidental food additives." These regulations specify that: "A material used in the production of containers and packages is subject to the definition (of food additive) if it may reasonably be expected to become a component.... of food packaged in the container." The regulation further states that: "If there is no migration of packaging component from he package to the food, it does not become a component of the food and thus is not a food ad ditive" (7). .The FDA Guidelines tentatively define Summary: This article summarizes procedures for de tecting small quantities of vinyl chloride monomer in packaging materials. The lower limit of detection by gas chromatography (flame ionization detector) was 2.2 x 10-* gm, which is equivalent to a concentration limit of 5 ppb (w/w) in a polymer sample. Gas chromatography supplemented by a form of mass spectrometry known as mass-fragmentography was also employed, allowing detection of 8.7 x Iff-13 gm VCM. The authors point out that eliminating materials to the limits of detectability may have more serious ill effects than allowing reasonable levels of suspect ma terial to be present. "no" as if computations show that no more than 0.01 ppm of a substance would be added to food assuming it ail migrated. The extracts usually need not be tested for such a substance" (8). In the absence of carcinogenesis, the 1 x 10"8g/g level of additive (migrant) in food has been accepted as the equivalent of "no." Where carcinogens are concerned, the equivalent "no" has not been defined, except in the vague terms of the most sensitive analytical techniques available for quanti fication of the substance in question. The Delaney Clause states that, "no additive shall be deemed to be safe if it is found to induce cancer when ingested by man or animal, or if it is found, after tests which are appropriate for the evaluation of the safety of food additives, to induce cancer in man or animal, .. ." (9). Therefore, the concern with re sidual VCM in PVC packaging material is whether or not VCM migrates from PVC to the food contained in the package. In this paper we deal with the question of VCM mi gration from PVC to foods. As a first step, an analytical procedure for the quantitation of vinyl chloride was developed. For the past ten years, our laboratory has been in volved with migration and desorption studies as a means of evaluating the interaction between low Tht* snide wit adapted from a paper presented at a meeting of the section of Polymer Science. The manuscript is in press by he New Vork Academy of RSV 0012328 molecular weight compounds id polymers. These studies have dealt with various aspects of migration ranging from the practical problem of migrants affecting sensory properties of packaged food (10) to theoretical studies where sorption-desorption thermodynamic parameters were evaluated (11, 12). Caseous migrants have also been used as probes for the study of macromolecuiar structure (13). Studies on sensorially important residues have shown that contact of the polymer with such substances in the normal manufacturing process re sulted in measurable sorption or retention, even under the extremes of commercial conditions used for their removal (14). The method developed for the determination of such residues is now an accepted American Society for Testing and Materials (ASTM) procedure (15). The apparatus is shown schematically in Figure 1. This method involves distillation of vo latile residues from the polymers into a hermetic limited volume at sufficiently elevated time/ temperature conditions so as to provide essentially complete transfer from solid to gaseous phase. Gas chromatographic analysis (GLC) of the gas phase then gives tentatively, both a qualitative identification and quantitative upper limit for the presence of the residue in question. This gas phase partition has been related to liquid phase transfer in certain cases with good agreement between methods (16). The extension of this method to the measurement of residual VCM in PVC film and resin has been es tablished in our laboratory. In addition, the feasibility of using mass fragmentography for the analysis of VCM in headspace gas has been determined. This method of detection was found to be immediately applicable to air sampling (17). Materials and methods. The appratus used con sisted of: (a) Gas chromatograph. Hewlett-Packard Model 1520 gas chromatography, equipped with back flush valve. GLC conditions: 10' x V4" i.d. stainless steel column packed with 10Vo SE30 on 90-100 mesh Anakrom ABS. Nitrogen flow rate 60 ml/min.; temperatures (*C)-- injection port 170*, detector 220*, column main tained at ambient temperature. Vinyl chloride reten tion time, 2.0 min. Inject 10 mi sample with gas tight syringe, equipped with one-way metal stopcock valve. (b) Hot jar. A 250 mi (V: pint) Mason jar fitted with stainless steel lid specially fabricated with Swagelock fittings and septum as described by Wilks and Gilbert v14>. Weigh accurately sample of PVC film or resin and transfer to Mason jar. The area and volume of PVC film, are recorded prior to analysis. The apparatus utilizing the hot jar and syringe are placed in a 90* C oven and maintained for 30 min. The GLC response obtained from three separate 10 ml on-column injec tions of the headspace is measured. No decom- Fifurt i Apparatus used to sample gas. Method involves the distillation of volatile residues from the polymen into a hermetic limited volume under elevated time/temperature conditions. position of the polymer to liberate VCM was detected after prolonged heating (24 hr.) at 90* C. A standard curve of response v$. VCM concen tration was constructed from vinyl chloride mixtures of known concentration. A standard mixture of vinyl chloride diluted with nitrogen was prepared containing 7 ppt (v/v) vinyl chloride and was then used to prepare vinyl chloride mixtures of varying concentrations. (c) Gas chromatography--mass spectrometer. A Du Pont Model 21-490 mass spectrometer, equipped with digital mass marker and dynamic dual ion de tector of the design of Pareles and Rosen (18) was in terfaced to a Varian Model 2740 gas chromatograph via a single-stage jet separator. GLC conditions were similar to those used for GLC detection of vinyl chloride. Vinyl chloride retention time was 1.52 min. Detect vinyl chloride by repeatedly scanning both the m/e 62 and 64 ions. Simultaneously record on 2-pen recorder (17). Inject 10 ml sample with gas tight syringe. Results and discussion. Tabie 1 represents the lev els of VCM detected in freshly prepared PVC homopoiymer and plasticized resin. Also listed is the retained VCM after the two resins were equilibrated in air. As shown in Table I, a marked decrease in residual VCM was found in the resin following plasticizer ad dition. The levels being 254 and 0.59 ppm, respec tively. This decrease in residual VCM may be at tributed to air equilibration as well as losses during processing. In addition, the plasticizer will modify the physical properties of the polymer (e.g., Tg) and, therefore, affect the solubility of VCM in the resin. Table I also shows that the concentration of re sidual VCM in the resins was markedly decreased by equilibration of the polymer samples in air. The level of VCM in the plasticized and unplasticized resins, following equilibration, was 0.02 and 2.0 ppm, respec tively, which represented approximately 7.0 and I.OVo of the initial VCM concentrations. For these studies, polymer samples were transferred from a dosed container to an open crystallizing dish and placed in a fume hood for 48 hrs. This favorable partitioning of VCM between PVC and its environment may provide a method of re moving residual VCM from the resin. Analysis for residual VCM in several representative industrial grade PVC products was carried out and the RSV Q0L2329 TaMtl Residual VCM in PVC resin by hot jar technique Sample* number 1 2 3 4 Resin type Unpiasticued Unplasiicized Plasticized Plasticized Equilibration** time lhours1 0 48 0 48 Average*** response 1.85 x 10* 1.78 x I01 4.39 x 10* 1.83 x 10 Total VCM (gm) 6.35 x JO-4 5.18 x I0'7 1.49 x I0-7 5.33 x 10"* *2.5 gm. sample of powder used "Sample placed in hood Response for 10 ml injection - range x attenuation x peak height (cm) Average for two runs VCM concentration in resin (ppm) 254.0 2.07 0.597 0.0213 results are tabulated in Table II. As shown in Table II, the detected level of VCM was between 5 ppb for PVC film to 71 ppb for PVC coated fabric. The lower limit of detection of VCM was 2.2 x 10"* gm by flame ionization (FID). Therefore, if a 10 gm sample of PVC film or resin is assayed by the hot jar procedure and we assume all the vinyl chloride present is released on heating, a detectable limit-of 5.6 ppb (w/w) is possible. Conventional GIC procedures for analysis of re* sidual VCM in PVC resin require dissolution of the polymer sample in a suitable solvent (e.g., THF) and direct on-coiumn injection of the resultant solution. While this procedure is very useful, the sensitivity of the method is limited by: a. solubility of the polymer sample in the solvent, b. the maximum sample size which can be injected into the gas chromatograph and c. solvent impurities. The sensitivity of the procedure utilizing the hot jar is limited by sample size and reactor (jar) volume and can be easily increased by varying these parameters. This technique, therefore, appears to offer some ad vantages over conventional methodology. That the procedure utilizing the hot jar must be essentially quantitative is supported by the results of Berens (19), who found that above the glass transition temperature (Tg) the solubility of VCM.in PVC resins was negligible. It should be reemphasized, however, (hat this procedure is valid only in a negative sense in that a CLC method can only establish the maximum amount of a suspected material producing a response in the detector under specified operating conditions but cannot provide positive identification. The need for structural conformation of unknown contaminants (e.g., vinyl chloride) is difficult below the ppm level. We, therefore, employed mass fragmentography, a technique where the mass spec trometer is used as a sensitive and selective detector for CLC effluents (17). For VCM analysis, the parent (m/e 62) and the P+2 (m/e 64) fragments were recorded simultaneously. A ratio of 3 to 1 for these peaks is obtained for vinyl chloride as a result of the natural isotopic distribution of chlorine - 37. Peak heights at these ion mass num bers and a comparison of their ratios combined with the CLC retention time afforded a specific and sensitive method for VCM determination at the 8.7 x 10-'* g level, in headspace analysis. The lower FID limit determined on our detector was 2.2 x 10-* g. The mass fragmentography method is, therefore, al most 250 times more sensitive than flame ionization detection with the added advantage of structure confirmation. To date, headspace analysis by this technique has been carried out on standard vinyl chloride mixtures, and the procedure is immediately applicable to air sampling. The application of the procedure using the hot jar with mass fragmentography detection for analysis of residual VCM in polymer samples is presently being investigated. This method when applied to food simulants or to foods themselves is expected to be much more sensitive than the 10-1 g/g level presently considered Table II Residual VCM in PVC film, coated foil and coaled fabric by hot jar technique Sample number 1 2 3 film type Plasticized PVC PVC coated foil PVC coated fabric Sample wr. Igm) 4.5 8.0 6.0 Thickness (mil) 0.75 1.70 -- A verage* response** 2.15 x 10` 1.27 x 10 1.35 x 10J Total VCM igm) 6.7 x 10"* 4.0 x 10"' 4.25 x lO-7 VCM con. in PVC (ppb)*** 15 5 71 Average of two injections 'Response for 10 ml injection range x attenuation x peak height (cm) ***These analyses.were preformed following major servicing of the instrument which increased the detector sensitivity tenfold to a detection level of2.2x 10"logVCM 8 22 RSV 0012330 A as the "no migration" level. consider it quiie probable that this or equally advanced techniques in analytical chemistry could be developed to dem onstrate the presence of VCM in any sample of PVC and by calculation infer its transfer to PVC packaged food. Studies on desorption of low molecular weight molecules from polymers showed that two factors must be considered in relating VCM partition from polymer to food. The first concerns the equilibrium concentration of migrant in the food and the second is the time required to attain equilibrium. In can be assumed that at any concentration of VCM in the polymer, there will be a correspondingly lower concentration in the contacting food with a maximum at equilibrium. It is quite likely that over three orders of magnitude will be found between the concentration of VCM in the polymer to that in the contacting phase. These assumptions are based on considerations of the partition coefficient and-the weight ratio of packaging material to contacting phase. The equilibrium concentration of VCM in a food resulting from contact with a PVC package may be estimated by solution of equation (1). [M], - [M]JWP/(KWP + W.JJ (D When (M] is the concentration of VCM in the food at equilibrium, [M]Po is the concentration of VCM in PVC package at zero time, K is the partition coefficient, and Wp and W, are the weights of the PVC package and food phase, respectively. If the following estimates for these factors are made, -lx 10"* g/g, K - TOO, Wp - 1 gm and W, - TOO gm; then [M], is equal to [1 x 10-"] (1)/[(1> HOO) + (1001] - 0.5 x 10-" g/g If we assume complete transfer of VCM from the package to food, |M],- 1 x 10-* g/g. In ail probability, at sufficiently low levels of VCM in the polymer, the equilibrium concentration of mi grant in the food will be below that estimated for [M], by equation U) because of nonlinear disorption of the VCM from the polymer. Berens (19) reported such a non-ideal behavior for the VCM/PVC system against VCM vapor. While there is still considerable question as to the presence of a defined "no effect" threshold, the proposed OSHA standard of 1 ppm (v/v) will be considered as having a sufficient safety factor. The per day level of VCM inhaled by exposure to 1 ppm vinyl chloride (v/v) for 8 hrs. per day with a relative air in take of 101 liters/day may be estimated by solution of equation (2), after first calculating the density of VCM at STP. g VCM/day - IQ^/day x 1 day/24 hrs. x 8 hr. (work day) x 1 x 10"*/ x 2.54 g/ - 0.85 x 10-a g VCM/day (2) The total expos to VCM from food contacting PVC packaging material may be estimated by solution of equation (3). (E) * [M)t x (G) x (B> (3) Where (E) is the effective total quantity of VCM in gested, (M], is the concentration of VCM in food, (G) is the total weight of food contacted (ingested) in g/ day, and (B) is the biological effectiveness of VCM in terms of its response at that dosage. If the following estimates for these factors are made, (M|, 1 x 10"* g/g, (G) - 1.814 x IP g/day (USDA estimated daily food consumption, with the further assumption that 10% of ail food consumed is packaged in PVC) and (B) - 1, then (E) is equal to 1.8 x 10"* g VCM/day. At this ingestion level, an inges tion dosage equivalent to the yearly inhalation dosage at 1 ppm for forty hours per week for forty-six weeks would require 300 years. If (M], 1 x 10"* g/g (i.e. partition coefficient - 1000) an ingestion period of 3000 years would be required to equal the yearly inhalation dosage. While a number of assumptions have been made in these calculations, we feel that the estimate of (E) is very conservative and represents an order of mag nitude approximation. The safety factors in this calculation are quite high: 1 The evidence for angiosarcoma in humans is for only a few individuals exposed (inhalation) to levels of 1000 ppm VCM for at least 10 years. 2 The biological effectiveness factor could easily be below one but is used at its maximum. 3 The partition coefficient (equationl) does not take into account the strong possibility that nonlinear de sorption at low VCM concentrations will greatly affect this factor, reducing it essentially to zero at some finite VCM level in the polymer. 4 10% of all food consumed is packaged in PVC. 5 No provision has been made for migration of VCM from the polymer into the atmosphere. Clearly this should be taken into consideration as losses of VCM by migration from the polymer to the atmosphere have been found in recent unpublished work. There appears, therefore, to be only a remote possibility of the demonstration of toxicity by inges tion of VCM migrating from properly controlled food grade PVC used in food packaging. Since in such a case we would be dealing with probabilities of risk to the general public rather than with certainties, there should be some assessment of the risks occurring from disruption of thoroughly de veloped procedures for protection of our food sup ply. for example, there is the dearly defined danger from highly toxic microbial toxins, including carcinogens, which could result from improperly protected foods due to a "Delaney" ban of PVC. A list of references follows. tf^oSrreni RSV 0012331 Switch to side-loaded cr on cuts 3-man line to one worker, doubles output capacity Fluidmaster Inc. is a leading producer of toilet waterlevei control valves sold to homeowners through home centers and hardware stores. The previous end-loaded Fluidmaster carton was high cost, both in material and in handling. Their new sideloaded carton costs substantially less... is much faster to form, load and close... and requires only 1 /3 the previous labor. IMotKti t "informal Fact Finding I ig on Possible Hazards of Vinyl Chloride Manufacture and Um." U.S. Department of Labor. Occupational Safety and Health Administration. February IS, 1974. Data contained therein. 2. Manufacturing Chemists Association. Report of April 16, 1974 as reported by Chem. Eng. News. p. 16, May 20,1974. j. Polyvinyl Chloride--H-A- Savretnidt, pg. 1 (1969), Reinheld Book Corp. 4. Federal Register, VoL 39, No. 101--Thursday, May 23, 1974. Consumer Product Safety Commission (16CFR--Part 1SOO). 5. Federal Register, Vol. 39. No. 92--Friday, May 10, 1974. Department of Labor Occupational Safety and Health Administration (29C FR--Pan 1910) (Docket OSH-36). 6. Modern Packaging Encyclopedia. December 1973, pg. 82-46. McGrawHill Publishing Corp. 7. Federal Food. Orug and Cosmetic Act. Subpart F of Section 121. as Amended in 1958. 4. Food and Drug Administration Guidelines (August, 1966) p. 8. 9. Code of Federal Regulations, Section 409C. 3A. 19. Wilks, R.A., and Gilbert, S.G. Journal of Food Science, Vol. 37,72 (1972). 11. Varsano. J.L., Gilbert. S.C. Journal of Pharmaceutical Science, Vol. 62, 67 (1973). 12. Varsano, J.L., Gilbert, S.G. Journal of Pharmaceutical Science, Vol. 62, 92 11973). 13. Lieberman, E.R., Gilbert, S.G., and Shrinivasa, V., New York Academy of Sciences Transactions, Vol. 34, No. 6. December 1972. 14. Wilks, R.A. and Gilbert, S.G. Materials Research and Standards 8, 29 (1968). 15. ASTM Designation M51-72. Standard Method of Test for Residual Solvents in F legible Barrier Materials. 16. Wilks, R.A. Rh.O. Thesis, Rutgers University (1971). 17. Rosen, J.D-, Morano, J.R., Pareies, S.R.. Ciacin, J.R., and Gilbert, 5.G. J. Assoc. Offie. Anal. Chem., July 197S (in press). 18. Pareies, S.R- and Rosen, J.D- Analytical Chemistry, Voi. 46,2056 (1974). 19. Berens, A.R. Polymer Preprints, September 1974 (p. 197], development quiz key to Pluksmastefs cost-cutting peckagmg operavon it tnetr automated 3machine Una. Cartons ara formed ana glued from blanks by Patera PQ-V former . . . loaded by womer as (hey east by on conveyor . . closed and saaiad by Patars CCY'O Oosat. . . and sackad S-wide by S-deap into comjqatad casas by Bemay Packing 4 Closing Machine which also seals and a/eels the casas to holding conveyor. Before: Three workers were required to insert the control valves into the 3" square end of 3 x 3 x 15" cartons . .. very much like stuffing items into close-fitting sleeves. Under pres sure to keep the line moving, workers tore many cartons which had to be scrapped. Now: With side-loaded cartons, a single worker literally drops the valve and fitting into the carton ... filling as many cartons as three workers filled before. i OTHER SAVINGS Lower carton costs: The new flat die-cut carton blanks cost $4.25 per M less than the pre-giued folded blanks used pre viously. Less storage space: Delivered flat. 12.000 to a skid, the new blanks take up much less space than the previous pre-gtued blanks which came packed 200 to a corrugated case. Leas handling time: Peters PG-V holds 3000 die-cut blanks. The previous machine held only 100 pre-glued blanks, which kept a worker busy unloading cases and loading the machine every few minutes. Doubted output capacity: As Fluidmaster sales increase, it will be only a matter of machine adjustment to increase packaging line speed. Peters machines are capable of forming cartons and closing & sealing cartons at speeds of more than 100 per minute... double the present output. End-opening feature retained: Although side-loaded, the car tons are formed so they can be end-opened and reclosed in the store, permitting customers to examine the contents before purchase. Readers are invited to contribute questions and their answers. The answers to these questions are on page 32. 1. What is the principal advantage of fourdrinier board over cylinder board when used in the manu facture of folding cartons? 2. In view of the interest in waste dis posal and recycling, what is the ad vantage of a cylinder board ma chine versus a fourdrinier board machine? 3. When manufacturing folding cartons, in which direction does the grain of the board normally run; vertically or horizontally around the carton? 4. What are the principal methods of forming rigid plastic containers? 5. What are the principal advantages claimed for ultra-violet drying inks used to print packaging materials? 6. Measurement of tinplate is by the base box. What is a base box equivalent to in terms of area? For more information on the Peters machines used in the Fluid- master operation, write to Peters Machinery Company. 4700 N. Ravenswood Avenue, Chicago, Illinois 60640 or phone 312-561-9000. Credit Submitted by Dr. Harold 1. Raphael, Director, Department of Packaging Science, Rochester institute of Technology, Rochester, NY 14623 24 For more information on advertised products check Reader Service Card RSV 0012332