Ind. Eng. Chem. Res. 2006, 45, 5211-5216
5211
Microcontaminants in Pentachlorophenol Synthesis. 3. Effect of Temperature and Chlorine Flow Rate at End of Run Jianli Yu, Terry J. Nestrick,† and Phillip E. Savage* Department of Chemical Engineering, UniVersity of Michigan, Ann Arbor, Michigan 48109-2136
Polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs) are undesired byproducts that form as microcontaminants during the synthesis of pentachlorophenol. In temperature-programmed, laboratory-scale synthesis, most of the microcontaminant toxicity forms near the end of the run and while the molten pentachlorophenol product cools and solidifies. Terminating the chlorine feed to the reactor before the end of the run, where some tetrachlorophenol remains, prevented the formation of microcontaminants in high concentrations in the final product. Therefore, it appears that the microcontaminant-forming reactions that occur near the end of the run are inhibited by creating chlorine starvation conditions in the reaction system. In a commercial implementation of this strategy, it may also be beneficial to purge the reactor contents with an inert gas after stopping the chlorine flow to remove the residual chlorine. Introduction Pentachlorophenol is a wood preservative. It is manufactured commercially by the AlCl3-catalyzed chlorination of phenol or chlorophenol mixtures in the liquid phase.1 No solvent is used, so the reaction temperature is maintained above the freezing point of the reactor contents, which increases from 41 °C to about 191 °C as the chlorination progresses. The commercial synthesis of pentachlorophenol is accompanied by the production of parts per million (ppm) levels of polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs) with six or more chlorine substituents.2 Domestic pentachlorophenol manufacturers have, over the years, reduced the level of these microcontaminants in the pentachlorophenol product, but even further reduction is desired from an environmental perspective. Advances in this area have been hampered, in part, by the lack of a convenient and inexpensive means of assessing the toxicity of the microcontaminants in pentachlorophenol. High-resolution gas chromatography with mass spectrometric detection has been the standard method for determining the toxic equivalent (TEQ) concentration of PCDFs and PCDDs in a pentachlorophenol sample. We recently applied a new bioassay developed for dioxins and demonstrated its viability for determining the TEQ concentration in pentachlorophenol samples.3 In this article, we use this new method to determine the effect of starving the system for chlorine near the end of the run on the TEQ concentration in the pentachlorophenol product. We suspect that this variable may have an effect because previous laboratory-scale work4 has shown that the TEQ concentration increases quickly after all of the tetrachlorophenol has been consumed. Additionally, we found that the TEQ concentration continues to increase postreaction as the reactor contents cool and solidify. Additional evidence that the chlorine content may be important is that a major producer stops chlorine flow and purges with nitrogen near the end of their commercial-scale production runs. In the experiments reported herein, we created chlorine starvation conditions near the end of the run by stopping the chlorine flow, by purging with nitrogen, and by adding * To whom correspondence should be addressed. Tel.: (734) 7643386. Fax: (734) 763-0459. E-mail:
[email protected]. † Present address: 4520 Washington St., Midland, MI 48642.
materials that we expected to react with any residual chlorine present in the system. Experimental Section Detailed descriptions of the experimental procedures and analytical methods appear in accounts of our previous work,3,5,6 which include modeling of the reaction kinetics. We simply provide a brief discussion in this article. The semibatch reactor system comprised standard laboratory glassware components. All reagents were purchased commercially and used as received. Chlorine gas was bubbled continuously through a liquid phase containing, initially, about 300 g of molten 2,4,6-trichlorophenol. AlCl3 (0.75 g) served as the catalyst. Pentachlorophenol synthesis was via a temperatureprogrammed reaction, which mimics commercial practice. The reactor temperature was gradually increased from 105 °C, the temperature at which we started the chlorine feed, to about 190 °C during the synthesis to keep the contents in the liquid phase at all times. The freezing point of the mixture increases as chlorination proceeds. Samples were withdrawn periodically by inserting a glass rod into the reactor. The liquid on the rod quickly solidified when removed from the reactor and cooled to room temperature. At the end of the synthesis experiment, the reactor contents were poured into a disposable aluminum pan to cool. A sample of this postreaction solid product was also collected. Samples were analyzed by capillary column gas chromatography to determine the amounts of tri-, tetra-, and pentachlorophenol. The microcontaminant level was determined by an aryl hydrocarbon receptor capture (AhRC) method that used real-time polymerase chain reaction (PCR) for quantification.3 This AhRC PCR bioassay provided the TEQ concentration of PCDDs and PCDFs in the sample. Results and Discussion As noted in the Introduction, previous work revealed that there is a large increase in the TEQ concentration during a narrow window in time near the maximum pentachlorophenol yield. The TEQ concentration typically increases further and is even higher in the cooled and solidified postreaction sample. To obtain a better understanding of how the TEQ concentration develops during pentachlorophenol synthesis, we analyzed all
10.1021/ie060212q CCC: $33.50 © 2006 American Chemical Society Published on Web 06/27/2006
5212
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006
Figure 1. Temporal variation of chlorophenol yields and TEQ Concentrations for a base-case synthesis experiment.
samples from a base-case synthesis experiment. The synthesis was done with 0.25 wt % (0.75 g) AlCl3 catalyst added to molten trichlorophenol at 90 °C, before the chlorine flow began. The chlorine flow rate was 1.5 mol/h at the beginning, cut to 1.1 mol/h after about 2.5 h, and then cut further to 0.9 mol/h after a 30 min temperature ramp. The stirring rate was set to 100-120 rpm. Figure 1 shows that the TEQ concentration is low initially, but takes on higher values as tetrachlorophenol forms. Even so, the TEQ concentration is typically less than 0.5 ppm as long as tetrachlorophenol is still present. After all of the tetrachlorophenol is consumed, however, the TEQ concentration increases rapidly from 0.56 to 2.9 ppm. The TEQ concentration then increases further to 5.5 ppm in the postreaction sample from the aluminum pan as the reactor contents cool and solidify. These observations suggest that product with TEQ concentrations around 0.5 ppm could be obtained by stopping the synthesis while some tetrachlorophenol remains. In this section, we report on several experiments that were conducted to determine whether modifications to the synthesis procedure near the end of the run could produce pentachlorophenol with lower TEQ concentrations. The end-of-run modifications we examined are stopping the chlorine flow rate, adding something to react with residual chlorine, stripping out residual chlorine, and stopping the temperature ramp during the temperature-programmed synthesis. We compare results from these experiments with those appearing in Figure 1, the basecase run. Effect of Stopping Chlorine Flow. Figure 2 shows results from experiments where the chlorine flow was stopped at 189 min, where the reaction mixture had reached 174 °C, and at 266 min, where the reactor temperature was 186 °C. The maximum pentachlorophenol yield in Figure 2a was 80% at 201 min with 0.27 ppm TEQ concentration. The postreaction sample from the aluminum pan had a pentachlorophenol yield of 77% with 0.3 ppm TEQ concentration. The final product contained about 15% tetrachlorophenol. When the chlorine flow was stopped later (Figure 2b), the maximum pentachlorophenol yield was higher and the TEQ concentration was 0.51 ppm. The TEQ concentration increased to 0.81 ppm after 5 min of additional reaction. The postreaction sample had a pentachlorophenol yield of 97% with a 0.81 ppm TEQ concentration. The final product contained no residual tetrachlorophenol. This intermediate product had been consumed completely. The TEQ concentrations in the postreaction samples in both experiments are lower than those from the base-case synthesis
Figure 2. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with chlorine flow stopped near end of run. (a) Cl2 stopped at 174 °C. (a) Cl2 stopped at 186 °C.
run (Figure 1). Thus, stopping the chlorine flow near the end of the run appears to be an effective way to stop the synthesis before higher TEQ concentrations can be reached. Moreover, it is apparent that the TEQ values are lower in Figure 2a, where some tetrachlorophenol remains unconverted, than in Figure 2b, where it had all been consumed. Thus, the precise temperature at which one terminates the chlorine flow may be important. Effect of Adding AlCl3 or Tetrachlorophenol. We next conducted synthesis experiments wherein we stopped the chlorine flow rate at 180 °C, and at the same time added either more (0.75 g) AlCl3 catalyst (to accelerate consumption of any residual chlorine) or 10 g of tetrachlorophenol (to provide more material with which residual chlorine could react). Figure 3 shows the chlorophenol yields and TEQ concentrations from these experiments. With added AlCl3, the postreaction sample had a pentachlorophenol yield of 87% with 0.66 ppm TEQ concentration. With added tetrachlorophenol, the postreaction sample had 91% pentachlorophenol with 0.8 ppm TEQ concentration. The residual tetrachlorophenol yield was 6% when the chlorine flow stopped. The TEQ concentrations in Figure 3 are no lower than those in Figure 2 at equivalent points during the reaction, which indicates that there is no large reduction in microcontaminants accompanying the addition of AlCl3 or tetrachlorophenol near the end of the run. Figures 2 and 3 show that the TEQ concentrations at the maximum tetrachlorophenol yields were typically lower than those of the final pentachlorophenol sample in the aluminum
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006 5213
Figure 3. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with chlorine flow stopped at 180 °C: (a) 0.75 g of AlCl3 added; (a) 10 g of tetrachlorophenol added.
pan. On average, about two-thirds of the final TEQ value has developed by the time the maximum tetrachlorophenol yield is attained. Effect of Purging with Nitrogen and Stopping Temperature Program. As shown in the previous section, a pentachlorophenol product with TEQ concentrations less than 1.0 ppm is obtained when the chlorine flow is stopped while some tetrachlorophenol is still present in the reactor. Adding more catalyst and adding tetrachlorophenol did not affect the TEQ concentrations (within our ability to measure it). We recognize, though, that residual chlorine would nevertheless be present. It is possible that it still affected the TEQ concentrations in subsequent samples. Therefore, we sought to remove as much of the residual chlorine as possible. We accomplished this removal by bubbling an inert gas (N2) through the reaction mixture after stopping the chlorine flow. We examined the effect of this strategy by performing three more syntheses and stopping the chlorine flow at different temperatures near the end of the run (177, 180, and 183 °C). The normal catalyst amount (0.25 wt %) and catalyst addition procedure (added at 90 °C) were used for these experiments. Figure 4 shows the chlorophenol yields and TEQ values for selected samples from experiments at these three different chlorine flow stopping points. The results in Figures 4 are similar to those obtained from analogous experiments that stopped chlorine flow at slightly different temperatures but used no nitrogen purge (see Figure 2). In all cases, the TEQ concentration near the end of the run was around 0.65 ( 0.2 ppm, and it was only occasionally more than 50% higher than the TEQ value that had been reached at the maximum tetra-
Figure 4. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with chlorine flow stopped near end of run and nitrogen purging. (a) Cl2 stopped at 177 °C. (a) Cl2 stopped at 180 °C. (c) Cl2 stopped at 183 °C.
chlorophenol yield. These results suggest that purging with nitrogen provides little additional benefit in terms of reducing the TEQ concentration in the postreaction product from laboratory-scale equipment. One final note before leaving Figure 4 is that the TEQ concentrations in Figure 4b are all higher than any of the TEQ concentrations in Figures 4a and 4c. Because all of these values are high, we suspect that these results are spurious and may not indicate a genuine difference from those of the other experiments. Indeed, we have no reason to expect the TEQ concentrations to be highest when the chlorine flow is stopped at the intermediate temperature. In the experiments just described, the reactor temperature continued to increase to the same final value (188 °C) even
5214
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006
Figure 5. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with chlorine flow and temperature ramp stopped at 183 °C and nitrogen purging.
after the chlorine flow was stopped at some lower temperature. We ran an experiment to determine whether the continued temperature ramp had any effect on the TEQ concentration. Thus, we ran a reaction where both the temperature ramp and chlorine flow were stopped together at 183 °C. Figure 5 presents the results. It is apparent from comparison with Figure 4c that neither the TEQ concentration or the pentachlorophenol yield was changed much by stopping the temperature ramp. Effect of Added Phenol with Nitrogen Purge. As discussed above, bubbling an inert gas (N2) to purge the mixture after chlorine flow stopped did not appreciably reduce the TEQ concentration. Another way to potentially rid the reaction mixture of any residual chlorine is to add phenol with the hope that it will serve as a more effective “chlorine sponge” than tetrachlorophenol at the reaction stopping point. We conducted two synthesis experiments with phenol added (10 g) at different stopping points (180 and 183 °C) for the chlorine flow. Figure 6 presents the results, which can be compared with the results in Figure 4. The experiments are identical except for the addition of phenol at the time when chlorine flow stopped. Comparing Figures 6a and 4b shows that the addition of phenol at 180 °C seems to reduce the TEQ concentration by about a factor of 3-4. This effect may not be real, however, because the TEQ concentrations in Figure 4b may be too high, as discussed earlier. Comparing Figures 6b and 4c shows that the TEQ concentrations are higher in the run with added phenol. This difference is small, though, and within the experimental uncertainty. We suspect that the addition of phenol at 183 °C had no appreciable effect on the TEQ concentration. This lack of an effect at 183 °C also supports the notion that the TEQ concentrations in Figure 4b are high. Given these results, and that purging the mixture with N2 had no measurable effect on TEQ concentration, we do not believe there is much to be gained through the addition of phenol to the reaction mixture near the end of the run. Figure 7 illustrates the TEQ concentrations obtained by stopping the chlorine flow rate before the end of the synthesis experiment. Specifically, Figure 7 shows the average of the TEQ concentrations in the final reactor sample and the postreaction sample for the base-case experiment and for experiments with chlorine flow stopped at different temperatures. All of the data reported in this article for each specific chlorine flow stopping temperature were averaged to compute the TEQ concentrations shown in Figure 7. For example, the data in Figures 3, 4b, and 6a were used to calculate the mean TEQ concentration for stopping chlorine flow at 180 °C. Implicit in this calculation, then, is the notion that the addition of “chlorine sponges”,
Figure 6. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with chlorine flow stopped near end of run, nitrogen purging, and phenol added. (a) Cl2 stopped at 180 °C. (a) Cl2 stopped at 183 °C.
Figure 7. Average end-of-run TEQ concentrations from experiments with different stopping temperatures for chlorine flow.
purging with N2, and stopping the temperature ramp all had much less influence on TEQ than did stopping the chlorine flow rate. Figure 7 shows that terminating the chlorine flow rate at 186 °C prevents the formation of microcontaminants that otherwise would have been present in the product. The TEQ concentration appears to decrease further as the temperature at which chlorine flow is stopped is reduced. Of course, the composition of the final product at the different temperatures is different from the one produced in the base case in that it contains increasing amounts tetrachlorophenol. Synthesis Starting with Phenol. We synthesized pentachlorophenol starting with phenol to determine whether using a
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006 5215
Figure 8. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis starting with phenol and with chlorine flow and temperature ramp stopped at 180 °C and nitrogen purging.
different starting material (phenol or 2,4,6-trichlorophenol) produced any differences in the results. Also, phenol is a starting material often used in commercial practice. Thus, obtaining results for synthesis with phenol wherein the chlorine flow is stopped near the end of the run would be beneficial from a commercial implementation perspective. Since we normally start an experiment with about 300 g of 2,4,6-trichlorophenol, we started this experiment with about 143 g of phenol, which should lead to an equivalent amount of pentachlorophenol product, by moles. We used the same stirring rate (100-120 rpm) and chlorine flow rate (1.5 mol/h) that we used in experiments that started with trichlorophenol. In this run, 0.75 g of AlCl3 was added to the reactor at 60 °C with molten phenol. This is the same mass of catalyst typically used in the experiments that started with trichlorophenol (0.25 wt % based on trichlorophenol). The temperature in the reactor increased quickly to above 100 °C after the chlorine flow began because of the exothermic reaction. Samples were collected about every 30 min. The reaction required more than 10 h to reach the desired end point. We stopped the chlorine flow rate and temperature ramp at 180 °C to leave some tetrachlorophenol in the final product, with the hope that this strategy would help provide a lower TEQ concentration. The chlorinated phenols that formed were 2- and 4-chlorophenol, 2,4- and 2,6-dichlorophenol, 2,4,6-trichlorophenol, tetrachlorophenol, and pentachlorophenol. Their yields were analyzed by gas chromatography with thermal conductivity detection. Chlorophenol yields, TEQ concentrations for selected samples, and the reactor temperature profile appear in Figure 8. We summed the yields of the two monochlorophenols and two dichlorophenols in the plot. The mono-, di-, and trichlorophenol yields increased to maximum values of 84% at 60 min, 83% at 151 min, and 92% at 241 min, respectively, and then decreased to zero at the time the next product reached its maximum yield (i.e., the yield of monochlorophenols was zero at the time the dichlorophenol yields reached a maximum). The yield of tetrachlorophenol increased to its maximum value of 74% at 464 min. The amount of tetrachlorophenol remaining at the end of the run was 8%, and the pentachlorophenol yield was 83% in this sample. The TEQ concentrations were 0.13, 0.11, and 0.14 ppm at the points of the maximum mono-, di-, and trichlorophenol yields. These concentrations should be viewed as estimates because the bioassay we used was calibrated with pentachlorophenol samples. A different chemical matrix existed in these samples with the less chlorinated phenols. Nevertheless, the results indicate that the TEQ values were very
low until tetrachlorophenol was formed. The TEQ concentration increased to 1.98 ppm at the maximum tetrachlorophenol yield, and it was 1.62, 2.44, and 1.37 ppm in the last two pentachlorophenol samples from the reactor and the pentachlorophenol sample in the aluminum pan, respectively. Thus, we see that the TEQ concentration did not change appreciably after the maximum tetrachlorophenol yield appeared. This is the same desirable behavior observed in the experiments with trichlorophenol. The precipitous increase in TEQ concentration shown in Figure 1 was avoided. The end-of-run TEQ concentration of about 1.9 ( 0.5 ppm is higher than those obtained from syntheses starting with trichlorophenol. The reason for this higher TEQ concentration could be that this experiment with phenol required about 200 min longer to go from tri- to pentachlorophenol than did the runs where we started with trichlorophenol. The longer reaction time may have provided greater opportunity for the formation of microcontaminants. Summary and Conclusions 1. Stopping the chlorine flow near the end of the laboratoryscale synthesis is an effective means of producing pentachlorophenol with low TEQ (0.3-0.8 ppm) in the postreaction solidified product. Stopping the chlorine flow rate near the end of the run had a much stronger effect on the TEQ concentration than did any of the catalyst-related variables explored previously.4 Purging with nitrogen after stopping the chlorine flow rate had little additional benefit in terms of TEQ concentration reduction. Purging may be much more important at the commercial scale, however, where volumes are larger and the gas-liquid surface area/reactor volume ratio is much smaller. 2. The reactor temperature at which chlorine flow was stopped appears to influence the TEQ concentration in the final product. The lowest temperature (174 °C) corresponded to product with the lowest TEQ concentration (0.3 ppm). The highest temperature (186 °C) corresponded to the highest TEQ concentration (0.8 ppm). 3. Adding phenol, tetrachlorophenol, or AlCl3 had no consistent and statistically meaningful effect on the TEQ concentration in the product. The addition of “chlorine sponges” does not appear to be a fruitful path for TEQ concentration reduction. 4. When starting the synthesis with phenol, stopping chlorine flow near the end of the run appears to prevent formation of additional microcontaminants. This strategy led to pentachorophenol product with TEQ concentrations about the same as that existing at the maximum tetrachlorophenol yield. 5. Further insights into the behavior of this chlorination reaction system could be obtained by developing a multiphase reactor model consistent with the experimental results reported herein. Previous work on kinetics modeling6 could be useful to this end. Acknowledgment We acknowledge the financial support of the Microcontaminant Reduction Venture, an industry-sponsored consortium committed to reducing microcontaminant levels in technicalgrade pentachlorophenol. Literature Cited (1) Muller, F.; Caillard, L. Chlorophenols. In Ullmann’s Encyclopedia of Industrial Chemistry, on-line edition; posted June 15, 2000, DOI 10.1002/ 14356007/a07_001.
5216
Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006
(2) Buser, H. R.; Bosshardt, H. P. Determination of polychlorinated dibenzo-p-dioxins and dibenzofurans in commercial pentachlorophenols by combined gas chromatography-mass spectrometry. J.sAssoc. Off. Anal. Chem. 1976, 59 (3), 562-569. (3) Yu, J.; Nestrick, T. J.; Allen, R. Savage, P. E. Microcontaminants in Pentachlorophenol Synthesis. 1. New Bioassay for Microcontaminant Quantification. Ind. Eng. Chem. Res. 2006, 45, http://dx.doi.org/10.1021/ ie0602106. (4) Yu, J.; Nestrick, T. J.; Savage, P. E. Microcontaminants in Pentachlorophenol Synthesis. 2. Effect of Catalyst Identity, Concentration, and Addition Strategy. Ind. Eng. Chem. Res. 2006, 45, http://dx.doi.org/ 10.1021/ie0605783.
(5) Yu, J.; Savage, P. E. Reaction Pathways in Pentachlorophenol Synthesis. 1. Temperature Programmed Reaction. Ind. Eng. Chem. Res. 2004, 43, 5021-5026. (6) Yu, J.; Savage, P. E. Reaction Pathways in Pentachlorophenol Synthesis. 2. Isothermal Reaction. Ind. Eng. Chem. Res. 2004, 43, 62926298.
ReceiVed for reView February 20, 2006 ReVised manuscript receiVed May 10, 2006 Accepted May 13, 2006 IE060212Q
