Influence of Atmosphere and Temperature on Polycyclic Aromatic Hydrocarbon Emissions from Green Anode Paste Baking

Coal tar pitch, a well-known source of polycyclic aromatic hydrocarbons (PAHs), is used as a binder of petroleum coke in prebaked anodes used for electrolysis of aluminum. Anodes are baked up to 1100 °C over a 20-day period, where flue gas containing PAHs and volatile organic compounds (VOCs) are treated using techniques such as regenerative thermal oxidation, quenching, and washing. Conditions during baking facilitate incomplete combustion of PAHs, and due to the various structures and properties of PAHs, the effect of temperature up to 750 °C and various atmospheres during pyrolysis and combustion were tested. PAH emissions from green anode paste (GAP) dominate in the temperature interval of 251–500 °C, where PAH species of 4–6 rings make up the majority of the emission profile. During pyrolysis in argon atmosphere, a total of 1645 μg EPA-16 PAHs are emitted per gram of GAP. Adding 5 and 10% CO2 to the inert atmosphere does not seem to affect the PAH emission level significantly, at 1547 and 1666 μg/g, respectively. When adding oxygen, concentrations decreased to 569 μg/g and 417 μg/g for 5% and 10% O2, respectively, corresponding to a 65% and 75% decrease in emission.


■ INTRODUCTION
Polycyclic aromatic hydrocarbons (PAHs) are a diverse class of organic compounds, containing two or more fused aromatic rings. Many PAHs are classified as mutagenic, carcinogenic, and persistent organic pollutants, and reduced exposure and emissions are therefore recommended by the Council of the European Union. 1,2 Prebaked carbon anodes for aluminum production are a mix of petroleum coke and coal tar pitch, baked in an anode baking furnace up to 1100°C over a 20-day cycle. Coal tar pitch is a well-known source of PAHs. The pitch serves as a binder for the coke particles and as an energy source for the baking through combustion of volatile organic compounds (VOCs). 3 In addition, minor elements including sulfur and trace elements such as vanadium, nickel, and iron will be present in the coke and pitch. 4 Flue gas containing PAHs and VOCs from the baking furnace is often treated using regenerative thermal oxidation (RTO) technology, in addition to quenching and washing towers. 5 In the research by Wittgens et al., 6 a pilot scale combustion chamber with optimized combustion control is presented to facilitate reduced PAH and tar emission in ferroalloy production, finding stoichiometry, gas flow, and temperature to be important conditions.
During anode baking the levels of oxygen in the flue gas have been measured to be between 3 and 11%. 7 Anodes are also reported to be reactive with air and CO 2 during the electrolysis process, at temperatures between 520 and 960°C. This level of oxygen and thermal conditions may however not be enough to facilitate complete combustion of PAH. Chevarin et al. 8 tested the O 2 and CO 2 reactivity of baked anodes and its constituents and found the coke to be most reactive with air, while both coke and anode butts showed high reactivity with CO 2 at 960°C. The coal tar pitch showed the lowest reactivity in all experiments, and this is thought to be influenced by the change in properties of the pitch when baked alone, and with coke particles.
The different structures of PAHs, linear, angular, and cluster, influence the reactivity of the molecules. 9 Angular structures are more stable than linear, as studies show this to be due to better πinteractions, based on Clar's model, and interactions between hydrogen atoms in the bay region. 10 This region is found to increase the molecules' mutagenic and carcinogenic activity. 11 Another influencing factor for PAH stability is the number of rings. Larger PAH molecules have a greater resistance to degradation at ambient conditions, due to increased aromaticity. Degradation mechanisms of large PAH through thermal treatment and chemical oxidation can produce PAH of a smaller size as intermediates if enough energy is not supplied. 12 Sun 13 observed temperature as being a great influence on radical and PAH formation by way of intermolecular reorganization to increase molecular stability and aromaticity, at conditions where complete oxidation is not achieved. Liu et al. 14 investigated PAH emissions from coal combustion in a fluidized bed combustor. In this study, incomplete combustion of PAHs was found at temperatures up to 900°C, being strongly influenced by the level of excess air in the reactor. Pujro et al. 15 investigated catalytic cracking of heavy aromatic molecules at 450°C over a fluidized catalytic cracking zeolite catalyst with and without nickel and vanadium. The condensed polyaromatic compounds were found to have an increased activity with the catalysts over the thermal cracking, and this activity was found to increase with the number of aromatic rings.
The main objective of this study has been to investigate the effect of temperature and varying atmospheres on the EPA-16 PAH emissions from green anode paste (GAP) during initial heating of the anode baking process. Conditions at inert pyrolytic and oxidizing conditions were investigated in a controlled laboratory setup to better understand PAH emission profiles from green anode paste.

■ EXPERIMENTAL METHODOLOGY
GAP for the experiments was supplied by an industrial aluminum production partner. The paste contains calcined petroleum coke, anode butts, and coal tar pitch, which was prepared by milling (Herzog Maschinenfabrik, Osnabruck, Germany). Representative sampling was performed following sample splitting and the spoon method, as described by Petersen et al. 16 PAH composition in the off-gas was measured by performing experiments in a laboratory scale alumina tube resistance furnace (Nabertherm, RHTH 120-300/16-18) (Figure 1), where the off-gas was purged, and the organic content was collected in chilled 2-propanol (≥98% Technical, VWR Chemicals), using ice, and further analyzed by GC-MS (CP7462 Agilent column, Xtr EI 350 ion source). Samples were prepared by direct injection of internal standards containing

ACS Omega
http://pubs.acs.org/journal/acsodf Article deuterated PAH congeners and analyzed using selected ion monitoring (SIM). For each experiment, 5.00 g GAP was placed in a steel holder and heated in a selected atmosphere with a set temperature ramping program (heating rate 5°C/h), allowing off-gas sampling and exchange of bottles ( Figure 2a) through three temperature intervals, 25−250, 251−500, and 501−750°C (Figure 2b). The bottles were changed and ice refilled at the end of a hold period, while the gas stream was paused using a ball valve. This temperature range was chosen to study the evaporation of PAH from the raw material and the initial emission mechanism.
Different atmospheres were tested using a gas flow of 0.300 standard liters per minute. The atmosphere details are presented in Table 1. For experiments with different concentration of O 2 and CO 2 , air (Technical grade, Linde Gas AS, Trondheim, Norway) and CO 2 (≥99.7%) were diluted in argon (Instrument 5.0) to achieve the correct concentration levels.
Experiments were performed in triplicates, except for 100% Ar, which was performed one additional time to test for breakthrough of PAH through the three sampling bottles. This test was performed by adding an analytical thermal desorption (ATD) tube filled with the absorbent Tenax TA and glass wool, to the sampling line after the last sampling bottle (Figure 1). The off-gas from the furnace will pass through the bubbling bottles and the ATD tube, where, if present, light PAH will absorb on Tenax TA. The glass wool is present to filter out water and any particles in the stream and shield the absorbent material. The PAH content in the sampling tube was analyzed using thermal desorption and GC-MS.
PAH content in the GAP material was also analyzed using pyrolysis GC-MS. Pyrolysis was performed using a 4 μg sample placed in a tandem μ-reactor (Frontier Lab 3050TR) for 1 min. Analysis was performed once at 400 and 600°C, and thrice at 750°C. Compounds were separated and detected using GC-MS (Agilent Technologies; GC 7890B and MSD 5977B), and identified by the NIST Library.
■ RESULTS AND DISCUSSION GAP Sample Weight Change. All GAP samples used to perform the PAH emissions tests were weighed before and after each experiment at room temperature to investigate sample mass loss. The average mass loss (%) is presented in Figure 3 for all atmospheres. Inert and CO 2 atmosphere show a similar mass loss of 6.5−7.0%, independent of the concentration of CO 2 , which could be caused by the thermal evaporation of volatile organic components in the sample. The weight loss increased when oxygen was added, to 20.2% and 34.2% for 5% and 10% O 2 , respectively.
Influence of Temperature. All experiments showed a PAH emission profile similar to what is shown in Figure 4, when comparing the effect of temperature. Emission is presented as μg EPA-16 PAH per gram of GAP from experiments in the alumina tube furnace.
Low molecular weight (LMW) PAHs (2 and 3 rings) are the main contributors to the emission at low temperatures, from room temperature to 250°C. Up to 500°C, the high molecular weight (HMW) PAHs (4−6 rings) are emitted, at a higher concentration level. Lastly, PAH emissions decrease significantly     This is also evident in Figure 5, where the total concentration of PAH from each temperature interval is presented. Figure 5 also shows that the total PAH emissions level at low temperatures seem to be independent of the atmospheres tested, as the same level of emission within the standard deviation.
Influence of Atmosphere. The effect of the atmosphere on the PAH emissions is shown in more detail in Figures 6 and 7. Compared to inert atmosphere, adding 5% oxygen reduced the total PAH concentration by 65%, from 1645 μg/g to 569 μg/g. At 10% oxygen, the total PAH concentration was reduced with 75%, to 417 μg/g ( Figure 6). This, seen together with the increased mass loss for the GAP sample (Figure 3), with increasing oxygen levels, indicate increased combustion of the anode paste.
With the temperatures and residence time of the experiment, 10% O 2 is not enough to achieve complete combustion of PAH emitted from GAP. High molecular weight PAHs seem to be more affected by the oxygen level than lower molecular weight PAHs, as the concentration level of four ring PAHs (fluoranthene to chrysene in Figure 6) does not change significantly when the oxygen level increase. As described by Sun 13 and Liu et al., 14 complete combustion of PAH depends on several factors such as temperature and oxygen levels. In this study, the seemingly stable level of four ring PAHs could be the result of combustion products from the decomposition of other HMW PAH compounds, where the insufficient temperature levels and available oxygen could lead to stabilization of the radical combustion products. The majority of the emissions from the experiments occur at temperatures of 500°C or lower (Figure 4), where the HMW PAHs would be in a condensed state. In this state, the molecules could be susceptible to catalytic degradation in a gas−solid phase initiated by the trace elements nickel, vanadium, and iron in the GAP mixture, as described by Pujro et al. 15 Adding CO 2 to the atmosphere did not affect the total PAH concentration significantly compared to the argon atmosphere. The total PAH concentration for atmospheres with 5 and 10% CO 2 were 1547 and 1666 μg/g, respectively (Figure 7).
The weight loss of GAP samples in CO 2 was similar to that in argon, indicating that no significant oxidation of the material occurred. This finding is contradictory to that of Chevarin et al. 8 However, in the current study, the maximum temperature was 750°C as opposed to 960°C, and the differences in temperature would influence the CO 2 reactivity of the anode materials; in addition the change in the levels of PAH emitted at temperatures above 500°C could be too low for an effect to be statistically significant.
Breakthrough Test. Overall, the amount of LMW PAHs present was low for all experiments compared to the HMW species. To confirm this, a repetition of the experiment in an argon atmosphere was therefore performed to test for breakthrough of LMW PAHs from the 2-propanol in the sample collection system. It was done by connecting thermal desorption tubes with absorbent media at the end of the sampling line. Table 2 shows the results of PAHs in the tubes, compared to the average of three experiments of PAHs in 2-propanol.
Results show low amounts of the smallest PAHs (naphthalene and acenaphthene) in the tubes. Fluoranthene is detected in the greatest concentration in the tube out of the five PAHs, corresponding to the relative amounts in the 2-propanol bottles. The noticeable amount of fluoranthene in the tubes could hence be explained by evaporation of 2-propanol, containing PAHs, from sampling bottles to the tube, if the solvent was not kept at sufficiently low temperatures during the experiment. If this occurred, all species present in 2-propanol with lower boiling points than fluoranthene would also be transferred to the tube, as illustrated by data in Table 2.
Pyrolysis GC-MS. Pyrolysis of GAP at 400 and 600°C resulted in similar pyrolytic profiles, both with few identifiable signals. Pyrolysis at 750°C produced a species profile presented in Figure 8, corresponding to the identified species in Table 3.
Isomer PAH can not be identified as references for these compounds were not analyzed. Components are not quantified using this technique, but looking at signal abundance and area, a similar trend is shown when comparing with results from 2propanol samples. The strongest signals originate from components antracene/phenanthrene (signal no. 14, Figure 8) and larger MW PAHs. Together, this confirms the low concentration of LMW PAHs in the 2-propanol samples, and in the tube samples, originating from the green anode paste.
Samples pyrolyzed in the alumina tube furnace produced PAH emissions in the temperature range of 400 to 600°C (Figure 4), in contrast to the pyrolysis GC-MS experiments, and the reason behind the varying results from the two techniques could be the time of heating for the sample, where the pyrolysis  Tripenylen/Benz(a)anthracene 19 Tripenylen/Benz(a)anthracene 20 Benzo(e)pyrene/Benzo(k)fluoranthene 21 Benzo(e)pyrene/Perylene/Benzo(a)pyrene 22 Benzo(e)pyrene/Perylene/Benzo(a)pyrene GC-MS use 1 min, while a tube furnace experiment lasts several hours (Figure 2b).

■ CONCLUSIONS
The effects of temperature and atmosphere on EPA-16 PAH emissions from green anode paste were investigated in a laboratory alumina tube furnace setup with off-line PAH analysis. Atmospheres tested were argon, CO 2 , and O 2 with the temperature interval ranging from room temperature to 750°C . Results show PAH emissions at the level of 1645 μg/g, and 1547 and 1666 μg/g, from inert and CO 2 atmospheres (5% and 10%). With added oxygen, concentrations decreased to 569 μg/ g and 417 μg/g for 5% and 10% O 2 , respectively, corresponding to a 65% and 75% decrease in emission. For the current setup with the given temperatures, available oxygen and species resident time, complete combustion of PAHs was not achieved. A suggestion for future work is expanding the analysis techniques for the off-gas to include gas components such as CO/CO 2 and H 2 and other VOCs to investigate the formation and destruction mechanisms of PAHs in more detail. PAHs are persistent molecules, and sufficient levels of oxygen, temperature, and residence time are needed to achieve complete combustion.