Document wgZy99kY17rMJgdRZ0qemmVbD

Report 2457, June 1990 **e a u - / 6 5 ? RJC000594 ANALYSIS OF FUMES RESULTING FROM THE CURING OPERATION IN THE MANUFACTURE OF TEFLON-COATED ARTICLES (Additional Samplings, March and April 1990) Du Pont registered trademark BaiteneEurope Geneva Research Centres 000132 EID902116 REPORT ANALYSIS OF FUMES RESULTING FROM THE CURING OPERATION IN THE MANUFACTURE OF TEFLON-COATED ARTICLES (Additional Samplings, March and April 1990) for DU PONT DE NEMOURS (Belgium) Mercure Centre Raketstraat 100-rue de la Fuse 100 B - 1130 BRUSSELS June 1990 BATTELLE EUROPE Geneva Research Centres 7, route de Drize 1227 Carouge/Geneva Switzerland Du Pont registered trademark 000113 EID9021I7 ^'OOiOfX TABLE OF CONTENTS Page SUMMARY 1. INTRODUCTION 1 2. SAMPLE COLLECTION 2.1 OUTLINE OF SAMPLING PROCEDURES 2.2 SAMPLING OF PARTICULATES 2.3 COLLECTION OF HYDROGEN FLUORIDEAND HYDROLYSABLE FLUOROCOMPOUNDS 3 3 4 6 3- ANALYTICAL METHODS 3.1 PARTICULATE MATTER (Aerosols) 3.1.1 GRAVIMETRIC DETERMINATIONOF PARTICULATES 3.1.2 SCANNING ELECTRON MICROSCOPY 3.1.3 DETERMINATION OF BISPHENOL A 3.2 DETERMINATION OF HYDROGEN FLUORIDE AND HYDROLYSABLE FLUOROCOMPOUNDS 8 8 8 8 8 10 4- RESULTS 4.1 PARTICULATE MATTER 4.2 ANALYSIS OF HYDROGEN FLUORIDE AND HYDROLYSABLE FLUOROCOMPOUNDS 5. PigCVSSiQE QT-BBSia iS 11 11 13 ^ BIBLIOGRAPHIC REFERENCES FIGURES 1-8 TABLES 1-6 ANNEXES 1-3 20 %$;o()().')r*i 00015 4 EID902118 SUMMARY In a previous study {Reference 2: Battelle Report 2329, October 1989), the composition of the fumes resulting from the curing of different types of coating materials was analysed. Samples were taken from the fumes released during the curing of the most representative Teflon cookware and industrial cooking products, during the routine fabrication procedures in a Teflon coating plant, and analysed in the Geneva Laboratories in Battelle-Europe. The results showed that large variations occurred in the concen- 3 tration of particulate emissions (ranging from less than 0.1 mg/m up to 104 mg/m^) in the frame of SilverStone and SilverStone Supra coatings. Large variations also occurred in the concentration of fluorides in the fumes (ranging from 0.11 mg/m F up to 1.2 mg/m F" ) resulting from the curing of SilverStone coating materials. In order to better interpret the data, further analytical study was undertaken. The objective of the present study was to analyse the particulates and fluorides in the fumes resulting from the curing of one SilverStone coating material, by using different sampling procedures. This was needed because the concentrations are dependent upon the sampling conditions. Samples of fumes from the stacks of one conveyor oven were collected during the usual curing operation of SilverStone pewter coating, in a Teflon coating plant. The following sampling procedures were used: A. Isokinetic sampling according to Italian Standard. The particu lates were collected on glass wool placed in a stainless steel container, or on glass fiber membranes with a nominal pore diameter of 0.45 micrometer. Both types of filters were dried to constant weight at 105C before and after sampling. Gaseous fluorides were collected in impingers containing alkali solution, placed in line after the filter. 00015.5 RJC000597 EID902119 B. Sampling techniques used in the previous study. Polycarbanate membranes with pores of 0.4 micrometer diameter were used for the collection of particulates (aerosols). One set of samples were collected on filters which were conditioned at room temperature in controlled humidity before and after sampling. The other set of membranes were dried to constant weight at 105C before and after collecting the particulates. Fluorides were collected on a separate sampling line, using alkali solutions to absorb hydrogen fluoride and hydrolysable fluorocompounds. The concentration of particulates, collected on different filters and by different procedures, was determined by gravimetry. Some of the filters were examined by scanning electron microscopy in order to obtain information on the morphology of particulates. All the filters, used for the collection of particulates, were extracted by carbon disulfide and bisphenol A was determined in the extracts. The analyses of particulates and fluorides resulting from the curing of SilverStone coating materials are summarized as follows: l- Particulates The average concentration of particulates collected by isokinetic sampling according to Italian Standard was 2.35 mg/m3 in the entry stack and 0.69 mg/m3 in the exit stack. About the same concen trations were found using glass fiber membrane or glass wool filters. The average concentration of particulates collected by Battelle on polycarbonate (PC) membranes and dried to constant weight at 105aC was 2.45 mg/m3J on the entry and 3.87 3 mg/m in the exit stack. The average concentration of particulates collected on PC membranes which were conditioned in 65% relative humidity at room temperature was 18.6 mg/m in the entry stack and 5.68 mg/m in the exit stack. X6m )o o .)r>i 0001:5 6 EID90212G These results, together with the electron microscopic examinations of filters, seem to indicate that a big part of the substances retained on the membrane filters consists of liquid material. The proportion of non volatile to slightly volatile substances of the fumes seems to be higher in the exit than in the entry of the conveyor oven, and these particulates seem to be retained in a different manner by different types of filters. The bisphenol A concentration of all samples was lower than 0.003 3 mg/m . 2. Fluorides The total fluoride concentration in the samples collected in the ( frame of the previous study during the curing, of SilverStone 3 coating products varied between 0.11 and 1.2 mg/m , the average concentration was 0.5 mg/m . In the present study, using the same sampling procedures, similar results were obtained. The concentration varied between 0.07 mg/m and 1.7 mg/m and the average concentration was 0.4 mg/m'*. The average concentration of gaseous fluorides, collected according to the Italian Standard procedure, was only 0.06 mg/m3 . The concentration ranged from 0.01 mg/m up to 0.19 mg/m . The con centration of particulated fluorides which are retained by the filters is not included in these values. 66C()00/)f>J 000117 EID902121 ANALYSIS OF FUMES RESULTING FROM THE CURING OPERATION IN THE MANUFACTURE OF TEFLON-COATED ARTICLES (Additional Samplings, March and April 1990) 1. INTRODUCTION A previous study (Reference 2: Battelle Report 2329, October 1989) designed to cover the analysis of emissions which results from the curing of the most representative Du Pont TEFLON coating products showed that large variations occurred in the concentration of particulate emissions (ranging from less than 0.1 mg/m up to 104 mg/m^) in the frame of SilverStone and SilverStone Supra coatings. These particulates contained mainly decomposition products of ethoxylated alkylphenol type surfactants and Bisphenol A, as well as fatty acids in the case of SilverStone Supra coatings. Some filters used for the collection of particulates, stored at room temperature and reweighed after nine months, showed a considerable weight loss (up to 37% of the weight of particulates measured some days after sample collection). This weight loss indicated the presence of slowly evaporating material. Large variations also occurred in the concentration of fluorides in the fumes (ranging from 0.11 mg/m F up to 1.2 mg/m F ) resulting from the curing of SilverStone coating materials. Du Pont registered trademark 000158 EID902122 RJC'000600 2 In order to interprte better those data, further analytical data on particulates and fluorides were needed. The concentration of fluorides and even more the concentration of particulates may depend on the sampling conditions, as linear gas velocity in probe and stack, nature of the filter material, temperatures for conditioning of the filter, etc. Therefore the present study was designed to analyze the particulates and the fluorides in the fumes resulting from the curing of one SilverStone coating product under different sampling conditions. v RJC00060I 0 0 0 13 & EID902123 3 2. SAMPLE COLLECTION The sampling of fumes has been performed at an industrial plant during the curing of Du Pont SilverStone pewter coating materials, with following three code system: 459-516, 456-236, 456-300. The conveyor oven CTM2 was used for the curing. This oven consists of approximately 16 m long heated tunnel, in which the articles sprayed with dispersion were introduced onto a moving grid and moved slowly through the heated zones at controlled temperatures. The schematic drawing and a typical temperature profile for this conveyor oven are given in the annexes Nos. 1 and 2. The maximum temperature reached in the oven during the sample collection was around 430C. The fumes of the oven are removed through vertical stacks of square cross-sections of about 28 cm x 30 cm. Small (about 50 mm 4>) holes were cut in the stacks at about 2 m height above the upper level of ovens; the sampling probes were introduced into the stacks through these holes. Fume samples were collected simultaneously from the entry and exit stacks, using different sampling procedures. The identification of the samples collected during the curing operation and the sampling conditions are presented in Table 1. 2.1 OUTLINE OF SAMPLING PROCEDURES Two types of sampling procedures were used: A. Isokinetic sampling, according to Italian standards UNICHIM (Reference 1) for the measurement of gaseous emissions. The samples were collected by AMBIO S.C.n.l. B. Sampling techniques described in Battelle Report (Reference 2); Teflon FEP tubes (6 mm O.D., 4 mm I.D.), bent at the end to RJC000602 000120 EID902124 4 90 with a radius of 50 nun were used as sampling probes. The samples were collected by Battelle. Figure 1 shows the diagram of the equipment used by AMBIO for collecting the fume samples. It consisted of two sampling lines: line Q for the collection of particulates on glass wool kept in a stainless steel container (cestello) and of gaseous fluorides in two Drechsel type impingers containing alkali solutions; line A for the collection of particulates on glass fiber membranes and of gaseous fluorides in alkali solutions. The only difference between the two sampling lines is the nature of filter, which is glass wool in the case of line Q and glass fiber membrane in the case of line A. The Italian Standard Methods Unichim No. 58B (Reference 1) recommends the use of glass or quartz wool for the collection of aerosols above 100C and at high concentrations (above 10 mg/m3) and the use of filter membranes below 100C and at low concentrations (below 10 mg/m3). Figure 2 shows the diagram of the equipment used by Battelle for the collection of fume samples. It consisted of two sampling lines: line 1 for the collection of particulates (aerosols) on membrane filters, line 2 for the collection of hydrogen fluoride and hydrolysable fluorocompounds in alkali solution. 2.2 SAMPLING OF PARTICULATES Procedure A AMBIO collected particulates either on glass wool placed in a stainless steel container (Line Q), or on glass fiber membranes of 47 mm diameter, with a nominal pore diameter of 0.45 micrometer 000121 EID902125 RJC000603 5 (Line A) . Both types of filters were dried to constant weight at 105C before use. During sampling the flow rate on both lines was regulated in order to obtain a linear velocity of gas in the probe identical to the linear velocity of gas in the stack. The volume, temperature and pressure of the gas was measured and the normalized sample volume was calculated for 0C and 1013 mbar. The sample volume ranged from 0.0235 to 0.0687 Nm3 . Procedure B Battelle used polycarbonate membranes of 47 mm diameter and 0.01 mm thickness with pores of 0.4 micrometer diameter (Nuclepore Corpor ation, Pleasanton, Ca, U.S.A., Ref. No. SN 111 130), for the collection of particulates (aerosols). One set of these filters were dried to constant weight at 105C; another set was conditioned at 25C and 65% relative humidity during 96 hours before weighing. The filters were then placed in Gelman Stainless Steel filter holders of 47 mm diameter. The flow-rate in the sampling line 1, measured by the rotameter, was 4 to 8 L/min during the collection of particulates. This corresponds to a linear velocity of about 1.4 to 2.8 m/s in the probe. This was lower than the calculated linear gas velocity in the stacks, which was 5.3 m/s in the entry and 6.1 m/s in the exit stack. The pressure-drop across the filter was about 100 mbar at the start of sampling and increased with time. The sample volume for the particulates ranged from 0.067 to 0.228 Nm3 (refer to Table 2). The following parameters were measured during the sampling oper ations: flow rate, gas volume, atmospheric pressure, pressure drop across the sampling device, ambient temperature. These are needed 000122 EID902126 6 in order to calculate the sample volumes for normalised conditions (273K, 760 mmHg), according to the equation: 3 V sample volume in Nm (normalized cubic meters) V * sample in m^ measured in the gas meter T = ambient temperature K, and P = atmospheric pressure - pressure drop in mm Hg. 2-3 COLLECTION OF HYDROGEN FLUORIDE AND HYDROLYSABLE FLUOROCOMPOUNDS Procedure A Gaseous fluorides (HF and hydrolysable fluorocompounds) were sampled in two Drechsel-impingers in series, filled with 250 ml and 150 ml 0.05N aqueous potassium hydroxide solution. The flow rates through the impingers was of the order of 6 to 8 L/min, corresponding to the isokinetic gas velocity through the probe. After sampling, the content of each impinger was transferred into separate polyethylene bottles. Procedure B Fluorides {HF and hydrolysable fluorocompounds) were sampled in three Midget Impingers in series, each filled with 10 ml 0.1 N aqueous sodium hydroxide solution. The impingers were cooled in a bath of ice-water during the sampling, in order to prevent evapor ation of the NaOH solutions due to bubbling the warm fumes through them. The flow rate through the impingers was of the order of 2 L/min, and the sampling lasted in most cases 100 to 120 minutes, which resulted in fume samples of about 0.2 Nm . EID902127 RJt'OOOW)? 7 After sampling the content of the three impingers was transferred into a 50 ml-polyethylene bottle. The impingers were washed with small portions (2-3 ml) of distilled water and the washings were collected in the same bottle. The Teflon FEP tubing leading to the impingers was also rinsed with 10 ml 0.1 N NaOH and small portions (3-5 ml) of distilled water and these washings were collected in a separate polyethylene bottle. RJC000606 000124 EID902128 8 3- ANALYTICAL METHODS The gravimetric determination of particulates collected byprocedure A was done in the laboratory of AMBIO. The other analyt ical determinations were made in the Battelle-Europe laboratories, Geneva. 3-1 PARTICULATE MATTER fAerosols^ 3.1.1 GRAVIMETRIC DETERMINATION OF PARTICULATES After sampling, the filters were removed from the filter holder, transported in closed polycarbonate containers in the laboratory and reweighed under the same conditions as before sampling. One set of filters was dried to constant weight at 105C and the other set was conditioned at 25C and 65% relative humidity before weighing. 3-1.2 SCANNING ELECTRON MICROSCOPY A piece of about 10 mm diameter was cut from the filters, covered with a thin gold layer under vacuum and examined by a scanning electron microscope JEOL JSM-840. Photographs of representative areas were made using magnification x 1000 and x 5000. 3.1.3 DETERMINATION OF BISPHENOL A The filter membranes were placed in 15 ml screw-capped glass vials and extracted with 2.0 ml carbon disulfide (Merck, Spectroscopic grade) in an ultrasonic bath, for 15 minutes at 25C. 3.0 ml carbon disulfide was used for the extraction of glass wool filters. An aliquot of the extract was analyzed by GC-MS. The following instruments were used: Hewlett-Packard model 5995C Gas Chromatograph/Mass Spectrometer HP7673 Automatic Liquid Sampler OOOIPS EID902129 /.(wootora 9 Hewlett Packard Series 300 computer based workstation, with HP9153C Disc Drive and 40 Megabyte Disc HP Think Jet Printer and HP Color Pro Plotter. We used the following instrumental conditions* Gas chromatography column Carrier gas Injector GC/MS interface Ionization Electron energy Temperatures Acquisition mode Ion source pressure Sample volume 15 m x 0.25 mm I.d. fused silica capillary, coated with 1.0 p SPB-1 (chemically bonded, polydimethylsiloxane, Supelco Cat. No. 47300) helium; column head pressure: 7 psi; flow rate: 1 ml/min split/plitless; splitless mode during 0.75 min direct capillary interface electron impact 70 eV Oven: initial: 60C initial time: 1.5 min final: 250C final time: 5.0 min rate: 5C/min Injector port: 220C Transfer line: 280C Ion source: 200C Mass Analyzer: 180C Scan; 40 to 550 m/Z; 0.84 scans/second 1-2 .10"5 torr 1 nl, with HP7673 ALS. The low resolution (M/ M ~ 1000) mass spectra obtained by each GC/MS run were stored on magnetic disc. The peak of bisphenol A in the total ion chromatogram (TIC) was identified by the retention time (35.2 minutes) and by the charact eristic mass spectrum, shown in annex 3. 000126 EID902130 XiWOOiDl'M 10 External standard calibration method, by peak area measurements in the TIC, was used for the quantitation of bisphenol A. Calibration curve was prepared with standard solutions, containing 0, 10, 50 and 100 mg/ul bisphenol A in carbon disulfide. 3.2 d e t e r m i n a t i o n o f h y d r o g e n f l u o r i d e a n d h y d r o l y s a b l e f l u o r o COMPOUNDS We performed the determination of total F~ according to the NIOSH P & CAM 117 method (Reference 3), using a fluoride specific electrode and TISAB buffer solution. Following instruments were used: - Model 407A Specific Ion Meter (Orion Research AG, Kiisnacht, Switzerland) - Model 96-09 Fluoride Combination Electrode (Orion) - Model E536 Potentiograph (Metrohm AG, Herisau, Switzerland) - Model 6.0216.000 Combined pH Glass Electrode (Metrohm). The sample solutions were transferred with small volumes of distilled water into a 150 ml beaker. The pH of the alkaline solution was adjusted with some drops of concentrated acetic acid (Merck, analytical grade) to 7.0 and the solution was evaporated on a hot plate to a residual volume of 3 to 5 ml. 1.0 ml of TISAB III buffer concentrate was added and the solution was transferred into a 10 ml volumetric flask. The volume was adjusted with distilled water to 10 ml and the solution was mixed. The pH of the solution after mixing was about 5.5. Calibration of the instrument was performed with two concentrations of F-standard solutions (e.g. 10- M3 an-d 140 M) which covered the range of the sample concentration. 75 000127 EID902131 11 4. RESULTS The samples of fumes analysed in the frame of this study resulted from the curing of SilverStone, which is one of the most represent ative Du Pont Teflon kitchenware coating. This three code system, with pewter color has the following references: PRIMER INTERMEDIATE COAT TOPCOAT 469-516 456-236 456-300 All samples were collected during routine curing of the coating under fabrication conditions. The maximum temperature during the curing operation in the conveyor oven was around 430C. As can be seen by the temperature profile (Annex 2) the working conditions are typical of those used by the Teflon coating industry. Table 1 presents, in addition to the identification of samples and the sample volumes, the weight applied coated wet dispersion cured in the oven per hour (g/hour), and the flow rates (m /hour) in the stacks where the samples were collected. 4.1 PARTICULATE MATTER The results of the gravimetric determination of the airborn particulate matter (aerosols) collected on filters under different sampling conditions are presented in Table 2. The average concentration of particulates collected b y isokinetic sampling was 2.35 mg/Nm 3 in the entry stacks and 0.69 mg/Nm 3 in the exit stack. About the same concentrations were found by using glass fiber membrane or glass wool filters. The average concentration of particulates collected by Battelle on polycarbonate (PC) membranes and dried to constant weight at 105C RJC000610 000128 EID902132 12 was 2.45 mg/Nm 3 in the entry and 3.87 mg/Nm3 in the exit stack. The average concentration of particulate matter (aerosols ) collected on PC membranes which were conditioned in 65% relative humidity at room temperature was 18.6 mg/Nm"* in the entry stack and 5.68 mg/nm3 in the exit stack. Isokinetic sampling according to Italian standards and Battelle's procedure using PC membranes dried at 105C, both gave approx imately the same result on the gravimetric determination of particulates in the entry stack. In the case of exit stack, the result obtained by Battelle's procedure was higher. The results for the gravimetric determination of particulates obtained by the Battelle-procedure, using PC membranes conditioned in 65% humidity, were considerably higher, than the corresponding results obtained according to the Italian standard procedure. The morphology of particulates, collected by different sampling procedures is shown in Figures 3 to 7. Figure 3 shows the photo graphs obtained by scanning electron microscopy of one glass fiber membrane filter used for the collection of particulates according to the Italian standard procedure. Only very few irregular shaped dust particles were found on this type of filter. The diameter of particles ranged from 5 to 10 microns. Liquid films at the crossing of some glass fibres can also be observed on the photograph with 5000 fold magnification. The photographs of polycarbonate membranes used for the sampling of particulates can be seen in Figures 4 and 5. These filters were dried to constant weight at 105C. The irregular shaped particles of 0.5-10 um diameter are included in an amorphous film, resulting from the drying of liquid droplets. I 19000, )f>| 000123 EID902133 13 v*' Figures 6 and 7 show polycarbonate membranes used for the collection of particulates and conditioned in 65% relative humidity at room temperature. Some areas on both membranes are covered by liquid films. Irregular shaped particles and on sample No. 50 (entry stack, Figure 6 ) also some big conglomerates of 10 to 25 |im diameter can be seen. Figure 8 shows the electron microscopie photograph of a blank polycarbonate membrane, in which regular, round pores can be observed. The results obtained by scanning electron microscopy of the filter membranes are summarized in Table 3. These results show that much of the particulate matter consisted of a liquid material. The surface of the polycarbonate membranes used for particulates collection had mostly an oily appearance. The results on the determination of bisphenol A in the carbon disulfide extract of particulated matter are presented in Table 4. The concentrations of bisphenol A are expressed in micrograms per cubic meter of fumes as well as nanograms per microliter of CS2extract. The concentration of bisphenol A was very low in all samples. The concentration in the fumes of the entry stack were higher than in the exit stack. Very low amounts of bisphenol A (0.5-2 nanograms) were found in blank Nuclepore membranes, these values would correspond to 0.003-0.01 micrograms per cubic meter as 3 blank values, using for the calculation 0.15 Nm sample volumes. 4.2 ANALYSIS OF HYDROGEN FLUORIDE AND HYDROLYSABLE FLUOROCOMPOUNDS The results are presented in Table 5. The fluoride concentration is expressed in terms of mg fluoride ion (F' ) per cubic meter. The sample solutions were analysed separately and the F" concentrations of a pair of solutions belonging to the same sample wre summarised to obtain the Total F~ concentration. PQ0.CL.3i> EID902134 14 The values obtained with the isokinetic sampling method are generally lower (with the exception of sample 3) than the values obtained by the Battelle sampling procedure, based on the NIOSH method. It must be considered that only the gaseous fluorides, passing through the glass fiber membranes of glass wool, were analysed in the case of the samples collected by AMBIo, while the particulated fluorides, deposited in the FEP sampling tubing leading to the impingers, were also analysed in the case of samples collected by Battelle. It is also possible that the gaseous hydrogen fluoride reacted with the glass fibers or glass wool and the resulting particulated fluorosilicates were retained by the filters used by AMBIO. This also may be an explication for the low values obtained with the isokinetic sampling method. RJC00061: oooiai EID902135 15 RJC000614 5. DISCUSSION OF RESULTS Particulates Particulate matter was collected under different sampling procedures in the fumes resulting from the curing of SilverStone pewter coating material. The following filter matrices and procedures were used in the frame of the present study: Filter material Procedure Glass fiber membrane Glass wool Polycarbonate membranes with 0.4 um pores isolation sampling, according to Italian standard; filters dried at 105C dried at 105C filters dried at 105C filters conditioned at room temperature/controlled humidity The concentration of the particulate matter on the fumes resulting from the curing of SilverStone pewter (samples 13 and 15) and of SilverStone black coatings (samples 28 and 29) were already determined in the frame of a previous study (Reference 2). These samples were collected on polycarbonate membranes, conditioned at room temperature and controlled humidity before weighing. Membranes with pore diameter of 0.4 pm were used for the collection of samples 13 and 15, while samples 28 and 29 were collected on membranes with 0.2 um pore diameter. All these samples of the present and the previous study were collected in the entry or exit stacks of the same coveyer oven, during the curing of the same or practically the same coating material. However, the quantity of wet products on the pieces per hour, the flow rates on the stacks, sample volumes, etc. were not always the same during the sample collection, as can be seen in Tables 1 and 2 of the reports of the previous and present studies. 0 0 0 l3 i> EID902I36 16 The results of the determination of particulates in the fumes resulting from the curing of SilverStone coating materials obtained in the present study, as well as the variables of sampling, are compared to those from the previous study in Table 6 . In the case of samples collected by the isokinetic sampling procedure, the ratio of the linear velocities in the probe and in the stack are assumed to be 1. In the other cases, the average linear velocity (cm/s) in the probe was calculated from the measured gas volume (L), the duration of sampling (minutes) and the internal cross-section of the tubing used as probe. The linear velocity in the stacks was calculated from the flow rate in stack (see Table 1 of the previous and present reports) and the cross-section of the stacks (28 cm x 30 cm). In the case of sample 29 (entry stack), the PC membrane with 0.2 nm diameter pores, used for sampling, was rapidly clogged by an oily deposit, the flow rate and linear velocity on the probe diminished rapidly and the pressure drop across the filter increased. Therefore only a small sample volume was collected, which explains that this result is not representative (concentration of particulates (65.6 mg/m3 ). In the case of samples 15 and 28, both collected in the exit stack, the average linear gas velocity in the probe was about three times higher, then the calculated linear velocity of gases in the chimney. The concentration of particulates in these samples was very low, less than 0.1 and 0.15 mg/m3 respectively. The reason for these low values is, that the flow rate on the exit stack was ten times smaller than in the entry stack and the biggest part of fumes was extracted through the entry stack. Discarding these extreme cases and analysing the remaining data, following conclusions can be made: ^19<)oo,)m 000133 EID902137 17 - No correlation was found between the ratio of linear gas veloci ties in probe/stack and the concentration of particulates. - The concentration of "wet" particulate matter, collected on PC membranes and conditioned at room temperature in controlled humidity before weighing, was considerably higher, than the concentration of "dry" particulates, collected either on PC membrane or on glass fiber filters and dried to constant weight at 105C. The concentration of "wet" particulate matter ranged 33 from 3.9 to 20.7 mg/m with an average of 10.5 mg/m and the concentration of "dry" particulate matter ranged from 0.4 8 to 4.6 6 mg/m^ with an average of 2.35 mg/m^. - The concentration of particulates depends on the nature of filter material used for collection, Using PC membranes, the concentration range was between 0.48 and 4.66 mg/m 3 with an average of 3 3.0 mg/m , while in the case of glass fiber filters the range was between 0.58 and 2.77 mg/m^, with an average of 1.7 mg/m^. The concentration of "wet" particulates in the fumes of the exit stack was only about the half of the concentration found in the fumes of the entry stack of the same conveyor oven. The collection of particulate samples 46 and 47, as well as 49 and 50 in the two stacks of the conveyor oven were done simultaneously. The concentration of "dry" particulates, collected on glass fiber membranes or glass wool filters was also lower in the fumes of the exit stack (average 0.69 rng/m^), than in the fumes of the entry stack (average: 2.35 mg/m^). The concentration of "dry" particu lates, collected on polycarbonate membranes, was higher in the fumes of the exit stack (average: 3.87 mg/m ), than in the fumes of the 3 entry stack (average: 2.48 mg/m ). Our explanation for the difference in particulate concentration found in the fumes of the exit and entry stacks, using different filter materials is that the composition of fumes resulting from the curing of SilverStone coating materials is not the same in the R J C 0 0 0 6 16 000134 EID902138 18 entry and exit stacks. From the wet coating entering into the conveyor oven, the most volatile constituents are released first; while the coated articles are approaching the exit, less volatile substances are released and the fumes contain a higher proportion of non-volatile particles. These non volatile substances were retained on the polycarbonate membranes having 0.4 (im pores with a higher efficiency than on the glass fibers. The filter holders of glass fiber membranes or glass wool were heated in the stack in order to prevent condensation of vapors, while the filter holders of the polycarbonate membranes were not heated and permitted thus the condensation of slightly volatile substances. This also makes a difference in the nature and amount of particulated matter retained by the different sampling pro cedures . The Italian standard procedure, using isokinetic sampling and glass fiber membranes below 100C or glass wool above 100C, is designed to measure the concentration of dust in the stack gases, in other words only non-volatile particulated matter. The procedure used by Battelle was aimed at the collection of all the aerosols, both liquids and solids dispersed in the stack gases, in order to get information on their composition. Therefore the concentration of particulates measured by the Battelle procedure is higher than the dust concentration measured according to the Italian standard. The scanning electron microscopy of filter membranes, used for sampling of particulates, showed that a considerable proportion of particulated matter resulting from the curing of SilverStone coating consists of liquid material. The solid particles had irregular shapes and their size varied from 0.5 urn to 10 urn in the case of samples collected in the entry stack and from about 0 .5 to 6 urn in the case of samples collected in the exit stack. Some big (10-25 um) conglomerates, consisting of small solid particles cemented together by a liquid mass, have also been observed in samples from the entry stack. oooias EID902139 /.iwo<).)r>! 19 The bisphenol A concentration, determined in the carbon disulfide extract of particulates, was lower than 0.003 mg/m in all of the samples of the present study. In the frame of the previous study relatively high concentration of bisphenol A (4.2 mg/m ) was found in sample 29 (see Table 9, cont.) of the previous report). Hvdroaen fluoride and hvdrolvsable fluorocomoounds The total fluoride concentration in samples 13, 14, 28 and 29, collected in the fumes resulting from the curing of SilverStone& coatings in the frame of the previous study, varied between 0.11 and 1.2 mg/m , with an average of 0.5 mg/m (refer to Table 5 of the previous report). The present study, using the same sampling procedures, gave similar resultsi the concentration varied between 0.07 and 1.7 mg/m and the average concentration was 0.4 mg/m . The average concentration of gaseous fluorides, collected in alkali solutions, according to the Italian standard procedure, was lower (0.06 mg/m^). The concentrations ranged between 0.01 and 0.19 mg/m"*. The particulated fluorides, retained by the glass fiber filters or glass wool were not determined. KJC0006I8 000136 EID902140 20 BIBLIOGRAPHIC REFERENCES 1. Metodo Unichixn N.588 (Edizione 1982) Misure alle emissioni con vogliati. Determinazione dei fluoruri gassosi e dei fluoruri particellari. Metodo potenziometrico (EM/14). 2. Battelle Report 2329, October 1989. Analysis of Fumes resulting from the Curing Operation in the Manufacture of Teflon-coated Articles. 3. NIOSH Manual of Analytical Methods, Second Edition, Volume 1. U.S. Department of Health, Education ahd Welfare, Cincinati, Ohio (1977). Fluoride and Hydrogen Fluoride in Air, Method No. P & CAM 117. <>l90()0.)f}J 0001,17 EID902141 21 LIST OF FIGURES FIGURE 1 - Diagram of sampling equipment, according to Italian Standards, for the measurement of gaseous emissions. FIGURE 2 - Diagram of sampling equipment used by Battelle. FIGURE 3 - Photographs of glass fiber membrane 1/1 No. 18. FIGURE 4 - Photographs of polycarbonate membrane filter No. 43. FIGURE 5 - Photographs of polycarbonate membrane filter No. 44. FIGURE 6 - Photographs of polycarbonate membrane filter No. 49. FIGURE 7 - Photographs of polycarbonate membrane filter No. 50. FIGURE 8 Photographs of a blank polycarbonate membrane filter. <)c9000Jf>I 000138 EID902142 22 LIST OF TABLES Table 1 - Samples collected in Smaltiriva-plant, Monte Marenzo, 22/390 - 4/4/90. Table 2 - Particulate matter collected under different sampling conditions. Table 3 - Scanning electron microscopy of filter membranes. Table 4 - Determination of bisphenol A in CSj-extract of partic ulates . Table 5 - Determination of hydrogen fluoride and hydrolysable fluorocompounds. Table 6 - Particulate concentration in function of sampling conditions. RJC000621 GOOi^ EID902143 Line Q c><>-- L A T J o 1 o o vi' c > ^ * C L in e __ h ! s M <r=4 y y.. s \ l. 1 > ilr- ilr n x ix j ilr ilr em u 1 Probe 2 Heated filter holder 3 Valve 4 Impingers 5 Drying column, filled with Silicagel 6 Pump 7 Flow controller 8 Thermometer 9 Gasmeter ZZWOOJfX FIGURE 1 Diagram of sampling equipment, according to Italian Standards, for the measurement of gaseous emissions. Line 1 for the collection of particulates Line 2 for the collection of HF and hydrolysable fluorocompounds 1 Probe (FEP tubing) 2 Filter holder 3 Flowmeter (regulated) 4 Impingers; containing alkali solution, copied in ice-water 5 Drying tube, filled with sil icanel 6 Manometer 7 Gas meter 8 Pump Lc90()0Jf>i FIGURE 2 Diagram of sampling equipment used by Battelle. t-rwoo/iRi FIGURE 3 - photographs of glass fiber membrane 1/1 No. 18. 000142 EID902146 RJC000625 FIGURE 4 Photographs of polycarbonate membrane filter No. 43. 000143 EID902147 9c 9000.)f>l FIGURE 5 Photographs of polycarbonate membrane filter No. 44. 000144 EID902148 c9ooo;)r>i FIGURE 6 Photographs of polycarbonate membrane filter No. 49. 0001-^S EID902149 RJC00062X FIGURE 7 Photographs of polycarbonate membrane filter No. 50. 0001-6 EID902150 RJC000629 FIGURE 8 Photographs of a blank polycarbonate membrane f i I t e r . EID902151 TABLE 1 - SAMPLES COLLECTED IN SMALTIRIVA-PLANT/MONTE MARENZO 22/3/90-4/4/90 Sample number 1/1 No.18 Stadt Entry Quantity of Area wet product oaeiad o/h m2/h 8500 58 Flow-rate Date of Collected by In stack sampling m3/h AMBO 1600 3 / 2 2 / 9 0 glass fiber membrane Volume Temp, collected of probe Nm3 dC 0.0253 95 1/2 I A HF 1/2 II A HF Entry 8500 58 1600 3 / 2 2 / 9 0 0.05 N KOH solution 3 / 2 2 / 9 0 0.05 N KOH solution 0.0253 95 1/2 CESTELLO B Entry 8500 58 1600 3 / 2 2 / 9 0 glass wool 0.0425 95 1/1 IQ FLUORURI Entry 1/1 IIQ FLUORURI 8500 58 1600 3 / 2 2 / 9 0 0.05 N KOH solution 3 / 2 2 / 9 0 0.05 N KOH solution 0.0425 95 2/1 No 19 Exit 8500 58 1850 3 / 2 2 / 9 0 glass fiber membrane 0 . 0506 81 2/2 I A HF 2/2 IIA HF Exit 8500 56 1850 3 / 2 2 / 9 0 0.05 N KOH solution 3 / 2 2 / 9 0 0.05 N KOH solution 0. 0506 81 2/2 CESTELLO N Exit 8500 58 1850 3 / 2 2 / 9 0 glass wool 0 . 0687 81 2/1 IO FLUORURI Exit 2/1 IIQ FLUORURI 8500 58 1850 3 / 2 2 / 9 0 0.05 N KOH solution 3 / 2 2 / 9 0 0.05 N KOH solution 0 . 0 6 8 7 81 3/1 No 20 Entry 8500 58 1600 3 / 2 2 / 9 0 glass fiber membrane 0.0235 95 3 I HF 2o 3 II HF 2o Entry 8500 58 1600 3 / 2 2 / 9 0 0.05 N KOH solution 3 / 2 2 / 9 0 0.05 N KOH solution 0.0235 95 BATTEULE 42 Entry 6360 42F 42T 1600 4 / 2 / 9 0 0.4 urn PC membr.105 oC 4 / 2 / 9 0 0.1 N NaOH Impingers 4 / 2 / 9 0 washings ol tubing 0.171 0.1715 0.1715 70 70 70 43 Entry 6360 1600 4 / 3 / 9 0 0.4 urn PC membr.105 oC 0 . 07 0 8 70 4S Entry 6360 45F 45T 1600 4 / 4 / 9 0 0.4 urn PC membr.105 oC 4 / 4 / 9 0 0.1 N NaOH impingers *** 4 / 4 / 9 0 washings of tubing *** 0.228 0.21 3 0.21 3 63 63 63 4 1 Exit 41F 41 T 6360 1850 4 / 2 / 9 0 0.4 urn PC membr.t05 oC 4 / 2 / 9 0 0.1 N NaOH Impingers 4 / 2 / 9 0 washings of tubing 0.0961 0.266 0.266 1 20 1 20 1 20 44 Exit 6 3 6 0 1850 4 / 3 / 9 0 0.4 urn PC membr.105 oC 0 . 0673 121 46 Entry 6360 46F 46T 50 Entry 6360 1600 1600 4 / 3 / 9 0 0.4 urn PC membr. 65% r.h. 0. 17 8 4 / 3 / 9 0 0.1 N NaOH impingers 0.3053 4 / 3 / 9 0 washings of tubing 0.3053 4 / 3 / 9 0 0.4 urn PC membr. 65% r.h. 0. 14 2 62 62 62 62 10OF 1OOT 47 47F 47T 49 Entry Exit Exit 6360 6360 6360 1600 1850 1850 4 / 4 / 9 0 0.1 N NaOH impingers " washings of tubing ** 0.161 0.161 4 / 3 / 9 0 0.4 urn PC membr. 65% r.h. 0 . 1423 4 / 3 / 9 0 0.1 N NaOH impingers * 0.263 4 / 3 / 6 0 washings of tubing 0.263 4 / 3 / 6 0 0.4 urn PC membr. 65% r.h. 0. 1 43 62 62 127 127 127 Notes: 'One Impinger broket "No supply during 7 minutes " *No supply during 22 min *'*No supply during 22 mir RJC000630 0001^8 EID902152 TABLE 2 - PARTICULATE MATTER collected under different sampling conditions SAMPLE FILTER 1/1 No.18 glass fiber membrane 1/2 CESTELLO B glass wool 2/1 No 19 glass fiber membrane 2/2 CESTELLO N glass wool 3/1 No 20 glass fiber membrane STACK VOLUME WEIGHT CONCENTRATION REMARKS of sample of partie. of partie. Nm3 ma ma/Nm3 Entry Entry 0.025 0.0425 0.07 0.11 2.77 2.59 AMBIO isokinetic sampling Exit 0.0506 0.04 0.79 Exit 0.0687 0.04 0.58 Entry 0.0235 0.04 1.70 42 0.4 urn PC membr.105 oC Entry 0.171 0.38 2.21 BATTELLE filters dried 43 0.4 urn PC membr.105 oC Entry 0.0708 0.33 4.66 at 105 oC 45 0.4 urn PC membr.105 oC Entry 0.228 0.11 0.48 4 1 0.4 um PC membr.105 oC Exit 0.0961 0.33 3.43 44 0.4 um PC membr.105 oC Exit 0.0673 0.29 4.31 46 0.4 um PC membr. 65% r.h. Entry 0.178 2.95 16.57 BATTELLE filters 50 0.4 um PC membr. 65% r.h. Entry 0.142 2.94 20.70 conditioned at 65% r.h. 47 0.4 um PC membr. 65% r.h. Exit 0.1423 0.82 5.76 49 0.4 um PC membr. 65% r.h. Exit 0.143 0.80 5.59 KJC'00063 000149 EID902153 TABLE 3 - SCANNING ELECTRON MICROSCOPY OF FILTER MEMBRANES SAMPLE FILTER DESCRIPTION OF PARTICLES FIGURE No. N o .18 43 44 50 49 BLANK glass fiber membrane few unregular shaped particles 5-10 um 3 0.4 um PC membr.105 oC unregular shaped particles 0.5- 10 um plus solidified liquid droplets 4 0.4 um PC membr.105 oC unregular shaped particles 0.5-8 um 5 Included in solidified liquid droplets 0.4 um PC membr. 65% r.h . liquid films,unregular shaped particles 0.5-10 um.conglomerates 10-25 um 0.4 um PC membr. 65% r.h. liquid films,unregular shaped particles 1-6 um (few) 6 7 0.4 um PC membr. 65% r.h. round pores in film 8 000150 EID902154 TABLE 4- DETERMINATION OF BISPHENOL A IN CS2 EXTRACT OF PARTICULATE MATTER SAMPLE FILTER 1/1 No.18 glass fiber membrane 1/2 CESTELLO B glass wool 2/1 No 19 glass fiber membrane 2/2 CESTELLO N glass wool 3/1 No 20 glass fiber membrane STACK Blsphenol A no/ul Volume collected Nm3 Bisphenol A ua/Nm3 Entry 6.08 0.0253 0.48 Entry 16.77 0.0425 1.18 Exit 0.00 0.0 5 0 6 0.00 Exit 1.97 0.0 6 8 7 0.09 Entry 0.24 0.0235 0.02 42 0.4 urn PC membr.105 oC Entry 9 8.40 0.171 1.15 43 0.4 urn PC membr.105 oC Entry 83.71 0.0708 2.36 45 0.4 urn PC membr.105 oC Entry 2 2 .0 7 0.228 0.19 41 0.4 urn PC membr.105 oC Exit 4.47 0.0961 0.09 44 0.4 urn PC membr.105 oC Exit 4.04 0.0673 0.12 46 50 47 49 BLANK BLANK BLANK 0.4 urn PC membr. 65% r.h. Entry 0.4 urn PC membr. 65% r.h. Entry 0.4 urn PC membr. 65% r.h. Exit 0.4 urn PC membr. 65% r.h. Exit 34.81 35.34 6.17 5.82 glass fiber membrane (Gelman) 0.4 urn PC membrane (lot 81A1C13) 0.4 urn PC membrane (lot 81B6D3) 0.00 0.23 1.05 0.178 0.142 0.1423 0.143 0.39 0.50 0.09 0.08 ng/memb 0.00 0.46 2.10 RJC0006. 00l5l EID902155 TABLE 5 - DETERMINATION OF HYDROGEN FLUORIDE AND HYDROLYSABLE FLUOROCOMPOUNDS Sample Stack F - F - Sample F - Total F- Note Mol/L mg/samp Volume mg/Nm3 mg/Nm3 Nm3 1/2 I A HF 1/2 II A HF Entry 4E -06 0.0008 0.0253 0.0319 1E-06 0.0003 0.0253 0.0111 0.04 1/1 IQ FLUORURI Entry 1/1 IIQ FLUORURI 2E -06 0.0004 0.0425 0.0092 1 E-06 0.0002 0.0425 0.0049 0.01 2/2 I A HF 2/2 IIA HF Exit 2 E -0 6 0 .0 0 0 4 0 .0 5 0 6 0 .0 0 7 9 0.03 6E -06 0.0011 0.0506 0.0208 2/1 IQ FLUORURI Exit 2/1 IIQ FLUORURI 2 E-06 0.0005 0.0687 0.0066 2E -06 0.0004 0.0687 0.006 0.01 3 I HF 2o 3 II HF 2o Entry 2E -05 0.0033 0.0235 0.1415 5E -06 0.001 0.0235 0.0437 0.19 42F 42T 45F 45T 46F 46T 10OF 100T 41 F 41T 47F 47T Entry Entry Entry Entry Exit Exit 0.0012 0.2185 0.1715 1.2741 0.0004 0.0732 0.1715 0.4265 0.0001 0.0238 0.213 0.1115 5E -05 0.0087 0.213 0.041 0.0002 0.0405 0.3053 0.1326 5E -05 0.0103 0.3053 0.0336 9E -05 0.0163 0.161 0.1015 2E -05 0.0034 0.161 0.021 4E -05 0.0076 0.266 0.0286 7E -05 0.0137 0.266 0.0514 4E -05 0.0074 0.263 0.0282 5E-05 0.0102 0.263 0.0387 1.70 0.15 0.17 0.12 0.08 0.07 1 Note:1) one impinger broken o o o i:^ EID902156 TABLE 6 - PARTICULATE CONCENTRATION IN FUNCTION OF SAMPLING CONDITIONS o h* '/I Cv w O IosJ SAMPLE 41 42 43 44 45 VOLUME (L ) 144 203 84 107 446 DURATION FLOW-RATE in probe (m in) (L/m in) (average) 20 7.20 24 8.46 22 3.82 22 4.86 90 4.96 Linear velocity Linear velocity Linear velocity P a rtic u la te s STACK Filter Nature in probe in slack probe/stack ;o n c e n tra tlo n conditioned of (cm /s) (cm /s) RATIO ( m g /N m 3) at filte r (average) 239 612 0.39 3.43 Exit 105oC PC membrane 280 529 0.53 2.21 Entry to 0.4 um 127 529 0.24 4.66 Entry costant pores 161 612 0.26 4 . 3 1 Exit weight 164 529 0.31 0.48 Entry 46 249 35 7.11 236 529 0.45 1 6 . 5 7 Entry 25 oC PC membrane 47 187 30 6.23 207 612 0.34 5.76 Exit in 0.4 um 49 186 30 6.20 206 612 0.34 5.59 Exit 65 % rel. pores 50 168 40 4.20 139 529 0.26 20.7 Entry humidity 13 15** 28** 29* 515 120 860 105 920 125 63 203 4.29 8.19 7.36 0.31 142 272 244 10 8 6 8 0 .1 6 3 . 9 Entry 25 oC PC membrane 87 3.12 <0.1 Exit in 0.4 um pores 87 2.81 0.15 Exit 60 % rel. PC membrane 868 0.01 65.6 Entry humidity 0.2 um pores 1/1 No 18 2/1 No 19 3/1 No 20 25 50.6 23.5 1/2 Cestello B 4 2 .5 2/2 Cestello N 6 8 .7 Notes: Sample volume too small, filter clogged "Fumes were extracted mainly through entry stack 1.00 1.00 1.00 1.00 1.00 2.77 0.79 1 .7 2.59 0.58 Entry Exit Entry Entry Exit 105oC to costant weight 105oC to costant weight glass fiber membrane glass wool ^TQOOOJf}!  V o S . . . Sampling lo cation s ANNEX 1 Sketch of the conveyor oven and the scrubber 000154 EID902J 58 OOOIF-S -vi EID902159 RJC00063 <SL9000Jni ANNEX 3 MASS SPECTRUM OF BISPHENOL A REFERENCE S U B S T A N C E 000156 EID902160