Document Jnxvo9kN2ypXvmEOZYO4y0Ee

Pilot-Scale Fluoropolymer Incineration Study: Thermal Treatment of a Mixture of Fluoropolymers under Representative European Municipal Waste Combustor Conditions Dr. Gehrmann, Hans-Joachim1; Dr. habil. Bologa, Andrei1; Dr. Aleksandrov, Krasimir1; Bergdolt, Philipp1; Dr. Taylor, Philip2; Dr. Schlipf, Michael3; Dr. Ameduri, Bruno4; Gunasekar, Priyanga5; Kapoor, Deepak5 1 Institute for Technical Chemistry (ITC) at Karlsruhe Institute of Technology (KIT); 2 P Taylor & Associates, LLC, USA; 3 Pro-K, Germany; 4 ICGM, University of Montpellier, France; 5 Gujarat Fluorochemicals Significance and Motivation A recent study by Conversio, a consultancy based in Germany, has shown that at its end-of-life approximately 85% of all fluoropolymers end up in waste-to-energy recovery incinerators. A subsequent question of regulators was: Do fluoropolymers get fully incinerated without any formation of short chain or long chain PFAS? A recent project executed by the Karlsruhe Institute of Technology (KIT) in cooperation with Socit Gnrale de Surveillance (SGS) was conducted to assess the same. Experimental Parameters Main applications of the four highest volume fluoropolymers (PTFE, PVDF, PFA and FKM) representing more than 80% of commercial fluoropolymer production based on data from Pro-K (German association of polymers processors) were considered. Post-use samples from these applications were incinerated as a mixture under standard operating conditions for municipal and industrial waste incineration. Figure 1 presents the experimental conditions. Experiments were conducted under two sets of conditions over a period of 9 days. The first experiments were conducted at a process setting of 860C and 2.0 s residence time. These experiments were conducted in three stages. Initially, background tests were performed using natural gas and 100 kg/h wood chips. This was followed by the same fuel conditions with the addition of 320 g/h of fluoropolymer. The final test involved switching back to background conditions. The duration of each of these tests ranged from 9 - 13 hrs. A second set of experiments was conducted at a process setting of 1100C and 2.0 s residence time. These tests were conducted in the same sequence as the first set of tests. The feed rates for the wood chips and the fluoropolymer mixture were identical to the tests at 860C and 2.0 s residence time. The test duration for this second set of tests also ranged from 9 - 13 hrs. 1|Page Figure 1: Experimental setup The fluoropolymers were fed as a mixture at relative proportions that correspond to the mass fractions sold in the European marketplace. These data are also shown in Figure 1. Suspension and emulsion polymerized PTFE application samples represented about 70 mass percent of the fluoropolymer feed rate. The main operational parameters for the two sets of tests are summarized in Figure 2. The temperature of the flue gas outlet exiting the rotary kiln was in the range of 800-900C. The temperature of the flue gas post-combustion chamber outlet was very close to the targets for these tests (860and 1100C in the combustion chamber for setting 1 and 2, respectively). The O2 and CO measurements for setting 1 and 2 varied somewhat. For setting 1, the values were 11.2 vol % dry and 0.2 mg/m3, respectively, while for setting 2 the O2 measurements were somewhat lower (7.0 % with an increase in the CO concentration (1.2 mg/m3). The water vapor concentration as measured in the boiler exit ranged from 6.2% in setting 1 to 8.49% in setting 2. 2|Page Rotary kiln combustion chamber mass flow wood chips main air mass flow heating oil volume flow natural gas volume flow combustion air inclination rotation speed temperature flue gas outlet thermal power volume flow natural gas to burner D4.1 sum of volume flow combustion air to burner D4.1 volume flow natural gas to burner D4.2 sum of volume flow combustion air to burner D4.2 residence time temperature flue gas post-combustion chamber outlet (with control) CO (level E2) O2 (level E2) thermal power total thermal power rotary kiln and post combustion chamber volume flow O2 CO water vapour unit kg/h mN3/h kg/h mN3/h mN3/h rev p.m. C MW setting S1 RUN 1, 2, 3 98 setting S2 RUN 4, 5, 6 98 418 423 61 46 4 4 872 753 2 0.2 0.4 800 - 900 1.1 0.9 mN3/h mN3/h mN3/h mN3/h s C mg/m3 Vol.-% dry MW MW mN3/h Vol.-% dry mg/m3 Vol.-% wet 22 35 671 429 22 35 671 428 2 860 1095 0.2 1.2 11.2 7.0 0.46 0.72 1.59 1.67 3958 11.9 1.35 6.20 3238 9.0 1.64 8.49 boiler / fluegas Figure 2: Main operational parameters at two experiments There were multiple sampling locations for this study. Flue gas was sampled near the exit of the combustion chamber (location 1), at the exit of the boiler (location 2), and at the entrance to the stack (location 3), while liquids and residues were also sampled and analyzed after each RUN (see Figure 3, Test facility sampling locations). The test facility BRENDA comprises a rotary kiln with a post-combustion chamber, a boiler for heat recovery and a flue gas cleaning system, which complies with German emission regulations (17 BImschV). The thermal power of the rotary kiln is of maximum 1.5 MW, while that of the post-combustion chamber is about 1 MW, which results in a total thermal output of BRENDA of maximum 2.5 MW. The fluoropolymers mixture after blending with wood chips and consequent weighing was delivered to the rotary kiln. To secure optimal combustion conditions, natural gas and heating oil were supplied additionally to the rotary kiln, while the post combustion chamber was supplied with natural gas only. The mass flow of the fluoropolymers mixture was set at 320 g/h, which corresponds to a pure Fluorine mass flow of 230 g/h. This level increases the fluoropolymer ratio to fuel, while at the same time keeps the Fluor-concentration below the total halogen limit of 1%, as set by the legislature. The combustion gases of the rotary kiln enter the post combustion chamber (PCC). It contains two natural gas burners staggered in an antiparallel manner, with a slight shift to each other. The temperature and the residence time in PCC were adjusted mainly with the help of the above mentioned burners, supported by a slight shift of about 200 kW into the post combustion chamber. 3|Page Figure 3: Test facility-BRENDA at KIT The minimum residence time is calculated according the methodology of the German Technical Supervision Agency ("TV") from 2007. The data which were published in the report were re-calculated and then adapted to the operational conditions in this study (Setting 1 and Setting 2). Figure 4 presents the layout of the post combustion chamber with the geometry relevant for the determination of the residence time. Fig. 4: BRENDA layout with details relevant for the residence time 4|Page Table 1 shows the detailed values for the design of the settings. The volume flow of the required flue gas amount to reach the two seconds was calculated with a target value search. Table 1. Parameters calculated for the residence time in the PCC PFAS Project, Level E1b setting 1 Start post combustion zone [m] 1 meter above 7.65 the burners Temperature in the post combustion chamber 860 (PCC) [C] setting 2 7.65 1100 Volume flow VPCC [mN3/h wet] after boiler 3947 3257 Cross section PCC [m2] Volume flow VPCC [m3/h] Height h [m] level E1b Residence time from start PCC zone to level E1b [s] 2.82 16,382 10.88 2.00 2.82 16,382 10.88 2.00 The two seconds are the residence time of flue gas from start of post-combustion zone until PFAS sampling point E1b, calculated with calibrated temperature measurements on the top of post combustion chamber (PCC). The flue gas was sampled for both short-chain and long-chain PFAS in addition to organic and inorganic fluoride. Volatile organic C1-C4 fluorocarbons were also sampled using a tedlar bag at all three sampling locations. At location 2, gas-phase HF was measured in near real-time using a tunable diode laser (TDL). The purpose of the three gas-phase sampling locations was to assess the potential emissions of PFAS at different locations in the system and to use this data to assess potential sources of PFAS in this system. PFAS sampling of residues and liquids is also shown in Figure 3. In addition to these three sampling points, flue gas scrubber water upstream of the SCR catalyst was collected and analyzed for PFAS. Table 2 provides a list of analytes measured in this study and the Limit of Quantification (LOQ). In addition to PFAS and fluoride ion, volatile C1-C4 fluorocarbons and trifluoroacetic acid (TFA) were also measured. The C1-C4 fluorocarbons were measured by gas chromatography coupled to mass spectrometry (GC-MS). Adsorbable organic fluoride (AOF) was measured using Combustion Ion Chromatography (CIC) and inorganic fluorine in impinger samples were measured by Ion Selective Electrode. TFA was measured using Ion chromatography (IC) and long chain PFAS from impinger samples were measured using UltrahighPerformance Liquid Chromatography coupled to tandem Mass Spectrometry (UPLC-MS/MS). HF was also measured at the post-combustion zone location using TDL spectroscopy. Appendix 1 presents a list of long-chain PFAS measured in this study. 5|Page Table 2. Analytes and reporting limits Analyte Volatile C1-C4 Compounds (CF4, CHF3, C2F6, C2HF5, CF2=CF-CF3, cy-C4F8) Adsorbable Organic Fluorine Inorganic Fluorine Trifluoroacetic Acid PFAS (see Appendix for list of compounds measured) LOQ 5-30 ug/m3 2 ug/L 0.1 ug/L 0.02 ug/L 0.02 ug/L Note: LOQ for AOF, Inorganic fluorine, TFA, and PFAS are for aqueous samples. Experimental Results Fluorine Recoveries Fluorine recoveries ranged from 69 to 84% using the TDL (at sample location 2). The variability in these data from run to run was low. In contrast, the impinger data analyzed at the same sample location showed about 10 to 20% lower fluorine recoveries. The data are summarized in Table 3. The TDL data provide strong evidence for complete mineralization of fluoropolymer feed mixture. Run Settings Table 3: Fluorine Recovery (TDL Measurement) HF (TDL) volume flow @standard wet conditions volume flow @270 C mg/mB3 wet Gas [mN3/h] [mB3/h] Fluorine g/h Fluorine Recovery % 860C, > 2s, oil + 2 nat. gas + wood chips + 230 g/h F 1100C, > 2s, oil + 5 nat. gas + wood chips + 230 g/h F 23.50 23.93 25.80 25.44 26.58 26.93 3,956 3,952 3,943 3,299 3,231 3,217 7,866 175.64 76% 7,859 178.62 78% 7,841 192.16 84% 6,560 158.53 69% 6,424 162.23 71% 6,397 163.64 71% Long-chain PFAS A large majority of the PFAS measured in impinger samples were near or below reporting limits (>98% of data collected at 860C and >96% of data collected at 1100C). Table 3 presents PFAS data for 4 compounds where measurements exceeded reporting limits in several cases. Of particular note is a HFPO-DA measurement which exceeded reporting limits by a factor of 47. Maximum PFBA, PFBS, and 6:2 FTS measurements exceeded reporting limits by much lower factors, ranging from 9 - 12. These data was re-analyzed to assess the veracity of data. The results are also presented in Table 4. The results indicate that the high measurement values for HPFO-DA could not be reproduced. The results for PFBA and PFBS were also lower when re-analyzed. The lack of reproducibility of data and the lower 6|Page measurement values upon re-analysis suggests that cross-contamination is a possible reason for high measurement values for HPDO-DA, PFBA, and PFBS in the initial analysis. PFAS analyses of wastewater and ash residue samples indicated a large majority of the samples were below reporting limits. One notable exception was a deslagger water bath sample where HFPO-DA was a factor of 16 above the report limit. Initial Analysis PFAS Compound PFBA PFBS 6:2 FTS HFPO-DA Table 4. PFAS Analysis of Impinger Samples RL (ng/m3) 2.8 1.4 1.4 1.4 # > RL 5 22 17 31 ng/m3 (max) 35.8 19.5 12.5 66.3 Re-Analysis PFAS Compound PFBA PFBS 6:2 FTS HFPO-DA RL (ng/m3) 2.8 1.4 1.4 1.4 # > RL 0 7 11 16 ng/m3 (max) 2.8 10.7 16.2 25.2 Note: For each data set, the total number of measurements equal 54: 27 for each combustion condition. Short-chain PFAS TFA was non-detect for all 76 impinger samples analyzed, at a reporting limit of 14 g/m3 (ppb). Volatile Fluorocarbons (FC) Tetrafluoromethane (CF4) was the only volatile FC detected in the GC-MS analysis. Values of CF4 at stack were near detection limits (20-27 g/m3) and detected in 2 of 14 samples. The results are considered questionable because CF4 was only detected in one post-combustion sample. There is no plausible reason for larger CF4 values downstream of the combustion unit unless a non-combustion source is considered. Discussions There is one prior published pilot-scale study of the combustion of PTFE (Aleksandrov et al. 2019). Combustion tests were performed at two conditions: 870C and 4 s residence time and 1020C and 2.7 s residence time and wood chips were used as the supplemental fuel. The prior study burned 0.3 wt % PTFE. Sampling was performed at a single location, downstream of the waste heat boiler. Thirty-one PFAS compounds were sampled and analyzed (see Table 1 of Aleksandrov et al. for a list of PFAS measured). 7|Page Fluorine recoveries were determined indirectly via IR water vapor measurements. The fluorine recoveries ranged from 56 to 78%, with three of the four tests yielding recoveries less than 70%. Eleven PFAS compounds were detected from the combustion and/or control samples and each at a level above 100 ng/m3 in at least one sample. PFOA was detected in all but one sample and at values as high as 2.7 g/m3 (see Table 3 of Aleksandrov et al.). The current study differs from the prior test in two important ways. The fluorine recoveries in this study were determined from direct spectroscopic measurements and were above 70% in five of the six tests. Secondly, PFAS reporting limits were on the order of 1 ng/m3 or less and a large majority of samples (>98%) were at or below reporting limits. The current study provides strong evidence that incinerating a mixture of fluoropolymers under representative municipal waste combustion conditions leads to complete mineralization of the C-F bonds, no significant emissions of long-chain PFAS, and no significant emissions of TFA or light fluorocarbons such as CF4 or C2F6. The prior study did not provide evidence that the PFAS detected were from sources other than the combustion of PTFE. Conclusions The study clearly demonstrated that fluoropolymers are converted to inorganic fluorides and carbon dioxide. The inorganic fluorides detected were hydrogen fluoride. A large majority of samples indicated that long-chain PFAS were below levels of 1 ng/m3 (> 99% of samples associated with 860C condition and > 98% of samples associated with 1100C condition). There were no short chain PFAS detected post incineration. TFA was non-detectable in all samples with a reporting limit of 14 g/m3. The results confirm that fluoropolymers at their end of life when incinerated under representative European municipal incinerators conditions do not generate any measurable levels of PFAS emissions and therefore pose no risk to human health and the environment. The main reason to include fluoropolymers in the EU PFAS restriction proposal was persistence (resistance to degradation in the environment) in the environment. The absence of organic fluorides and more specifically PFAS in tests representative of municipal waste incineration confirms complete mineralization of fluoropolymers and provides critical data in support for exempting Fluoropolymers from the EU REACH PFAS restriction proposal. References TV report from 19th of January 2007: Expert opinion on compliance with and monitoring of the combustion conditions (residence time, temperature) in the afterburning zone of the THERESA test facility at the Forschungszentrum Karlsruhe GmbH Aleksandrow, K, Gehrmann, H-J, Hauser, M., Matzing, H., Pigeon, D., Stapf, D., and Wexler, M., Waste incineration of Polytetrafluoroethylene (PTFE) to evaluate potential formation of per- and Poly-Fluorinated Alkyl Substances (PFAS) in flue gas, Chemosphere, 2019, 226, 898-906. 8|Page Appendices 1. List of long-chain PFAS analytes analyzed in this study 9|Page Fluorinated Replacement Chemicals 9-Chbrohexadecalluoro-3-oxanonane-I -sulfonic acid 9C1-PF3ONS (F-53B Major)' 1I-Chloroeicosafluoro-3-oxaundecane-1-sulfonic acid 1ICI-PF3OUdS OR (F-530 Minor)' 1I-Chloroeicosafluoro-3-oxaundecane-1-sulfonate 756426-58-I 763051-92-9 83329-89-9 ' 3C4-PFOS ' 3C4-PFOS Perfluorinated sulfonamide ethanols (FOSEs) 2-(N-methylperfittoro-I -octanesulfonamido)-ethanol 5 N-McFOSE 24448-09-7 2-O1-ethylperfluoro-1-octanesulfonamido)-ethanol 5 N-EtFOSE Additional Targets Nonafluoro-3.6-dioxaheptanoic acidh5 NFDIIA Perfitioro(2-edwxyethane)sulfonic acid" PFEESA Sodium perfluoro-1-dodecanesulfonates PFDoS Perfluoro-4-methoxybutanoic acid" PEMBA Perfluoro-3-medmx oic acid" PFMPA 3:3 Fluorotelomer carboxylic acid' 3:3 FTCA 5:3 Fluorotelomer carboxylic acids 5:3 FICA 7:3 Fluorotelomer carboxylic acid or 3-perfluoropheptyl 7:3 FICA or pmpanoic acid& 5 FHpPA 211-perfluoro-2-decenoic acid' 2-perfluorodecyl ethanoic acid' 2-perfluorooctyl ethanoic acid' 211 fluoro-2-octenoic acid' 2-perfluorohexyl Minnow acid' 8:2 FTUCA or FOUEA 10:2 FDEA 8:2 FTA or FOEA 6:2 FHUEA 6:2FTCA or 6:2 IAEA 1691-99-2 151772-58-6 113507-S2-7 1260224-54-1 863090-89-5 377-73-1 356-02-5 914637-49-3 812-70-4 70887-84-2 53826-13-4 27854-31-5 70887-88-6 53826-12-3 d7-N- MeFOSE d9-N-EtFOSE "Cs-PFFIxA "Cs-PFEIS "G-PFOS "Cs-PFPeA "Ca-PFBA "C2-FlI EA '3C-2FlI EA "CrFOEA I3CrFOUEA "C2-FDEA "CrFOEA "C--Fl IUEA "Cr Fl!EA 10 | P a g e