Microcontaminants in Pentachlorophenol Synthesis. 4. Effect of Nickel

Jun 20, 2006 - Polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs) are ... synthesis strategies on the concentration of polychlorinated...
0 downloads 0 Views 128KB Size
Download PDF
Ind. Eng. Chem. Res. 2006, 45, 5217-5222

5217

Microcontaminants in Pentachlorophenol Synthesis. 4. Effect of Nickel and Other Metal Powders 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. Adding small amounts (at least 100 ppm) of Ni powder to the reactor reduces the toxic equivalent (TEQ)) concentration in the product to about 70% of what it would be without Ni. By adding Ni powder at the start of the run and terminating the chlorine flow near the end of the run, pentachlorophenol with a TEQ concentration of about 0.4 ppm can be repeatedly synthesized in the laboratory. Co and Mn are about as effective as Ni in inhibiting the formation of toxic microcontaminants, but Mo, W, Ti, Cu, Fe, and Zn have no measurable effect on the TEQ concentration. Introduction This article is the fourth in a series on the effect of alternative synthesis strategies on the concentration of polychlorinated dibenzofuran (PCDF) and dibenzodioxin (PCDD) microcontaminants in pentachlorophenol, a wood preservative. The first article1 described a new analysis method (a bioassay) for measuring more quickly and inexpensively the toxic equivalent (TEQ) concentration of PCDDs and PCDFs in pentachlorophenol. The second2 and third3 articles examined the effects of catalyst-related and end-of-run reaction variables, respectively, on the TEQ concentration. Increasing the AlCl3 catalyst concentration led to a modest reduction in TEQ, but stopping the chlorine flow near the end of the run was more important as it prevented the formation of high concentrations of microcontaminants. In this article, we explore the effect of added Ni (and other metals) powder on the microcontaminant level in pentachlorophenol. Specifically, we focus attention on the impact of the time during the reaction at which Ni is added, the amount of Ni added, and the effectiveness of Ni relative to other transition metals. Experimental Section The materials, reactor system, and chromatographic analyses have been described in detail in earlier articles.4,5 Likewise, the bioassay method we use for measuring the TEQ concentration has been the topic of an earlier article.1 Since extensive experimental details appear elsewhere, we give just a short overview here. We synthesize pentachlorophenol by the AlCl3-catalyzed (0.25 wt %) chlorination of trichlorophenol in a round-bottom flask. Chlorine gas continuously bubbles through the molten reaction medium. As chlorination occurs, the melting point of the mixture increases, so the reactor temperature is increased during the course of the experiment to keep it a few degrees above the melting point. Samples of the reaction mixture are withdrawn periodically and analyzed for chlorophenols (by gas chromatography) and for microcontaminants (by bioassay). At the end of the experiment, the reactor contents are poured into a disposable aluminum pan to cool and solidify. We routinely analyzed a sample of this postreaction product from the * 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.

aluminum pan. Acetone, aqueous HNO3 solution, and methanol were used successively to clean the reactor after each experiment. All synthesis experiments reported in this article were done by stopping the chlorine gas flow when the reactor temperature reached 180 °C. This mode of operation leaves a small amount of tetrachlorophenol unconverted in the reactor, but previous work3 showed that it gives lower TEQ values than those obtained by driving the tetrachlorophenol to complete conversion. This article focuses on the effect of Ni and other metal powders added to the reaction mixture. Other than the addition of Ni, the experiments are identical to those described previously. Results and Discussion This section provides results from several pentachlorophenol synthesis experiments. We first report on the effect of the time during the reaction at which Ni is added. We then report on the influence of the amount of Ni added and the influence of other transition metals. In addition, we provide some results from experiments similar to those reported previously, with the difference that Ni is now added to the reaction mixture. These results deal with synthesis starting with phenol and experiments aimed at determining the effect of the amount of catalyst used. Time of Ni Addition. We did three different experiments in which the only variable was the time at which Ni powder (0.75 g ≈ 2500 ppm) was added. In all cases, the chlorine flow to the reactor was replaced with N2 when the reactor temperature reached 180 °C. Figure 1 provides the results. The results in this figure can be compared with those reported previously3 from experiments that are similar except for the addition of Ni powder at 180 °C. We used previous results3 from seven different experiments that stopped chlorine flow at 180 ( 3 °C to calculate mean TEQ concentrations of 0.38 ( 0.12 ppm for the sample at the maximum tetrachlorophenol yield, 0.53 ( 0.14 ppm for the last sample taken from the reactor, and 0.59 ( 0.18 ppm for the postreaction sample from the aluminum pan. The uncertainties shown here are the standard deviations about the mean. The TEQ concentrations in Figure 1a are slightly higher than the means given above, but they are within the range of TEQ values seen experimentally from experiments with chlorine flow stopping at about 180 °C but with no added Ni. Thus, adding Ni near the end of the run does

10.1021/ie060213i CCC: $33.50 © 2006 American Chemical Society Published on Web 06/20/2006

5218

Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006

Figure 1. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with 2500 ppm Ni powder. (a) Ni added at 180 °C. (b) Ni added at 160 °C. (c) Ni added at 90 °C.

not appear to have an effect on the TEQ concentrations. Adding Ni earlier does seem to have benefits, though, as shown in Figures 1b and 1c. In Figure 1b, the Ni was added when the reactor reached 160 °C. In Figure 1c, Ni was added (along with the catalyst) at 90 °C. In both cases, the TEQ concentrations were around 0.3 ppm in all samples, and there was no increase in TEQ concentration after the maximum tetrachlorophenol yield was reached. Taken collectively, these results suggest that Ni powder does not alter the TEQ concentration present at the maximum tetrachlorophenol yield and that it does not reduce the TEQ concentration much when added near the end of the run. Rather, Ni seems to play the role of inhibiting the TEQ concentration increase that had occurred in other runs between the maximum tetrachlorophenol yield and the final pentachlorophenol product. These results are encouraging, as they hint at a path to reducing the TEQ concentration formed during pentachlorophenol synthesis.

Figure 2. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with 2500 ppm Ni powder. (a) Replicate of Figure 1c. (b) Replicate of Figure 1c. (c) Replicate of Figure 1c except temperature ramp also stopped at 180 °C.

We did three additional experiments to check the reproducibility of these results. Figure 2 presents the results. Figures 2a and 2b shows the chlorophenol yields and TEQ concentrations obtained at the same experimental conditions used to generate Figure 1c. The TEQ values were about 0.38 ppm at the maximum tetrachlorophenol yield and about 0.5 ppm for the last sample withdrawn from the reactor. The TEQ concentrations for the postreaction pentachlorophenol samples taken from the aluminum pan into which the reactor contents were dumped at the end of the experiment were 0.46 and 0.65 ppm. Figure 2c gives the results from an experiment at the same conditions but with both the chlorine flow and the temperature ramp stopped at 180 °C. The TEQ concentration was 0.25 ppm for the last sample from the reactor and 0.34 ppm for the postreaction pentachlorophenol product. These TEQ concentrations are very close to the ones shown in Figure 1c, and they confirm that the

Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006 5219

Figure 3. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with 2500 ppm Ni powder added at 90 °C: (a) 0.75 wt % AlCl3 catalyst; (b) 1.5 wt % AlCl3 catalyst.

TEQ reductions obtained by adding Ni at the start of the synthesis reaction are reproducible. We verified that the cleaning procedure removed residual Ni between runs by periodically doing experiments with no added Ni and verifying that pentachlorophenol with higher TEQ was produced. For example, one such run with no added Ni led to a TEQ concentration at the maximum tetrachlorophenol yield of 0.79 ppm and a TEQ concentration in the final postreaction sample of 2.4 ppm. These values are higher than those that appear for runs with added Ni. Effect of Catalyst Amount. We previously reported2 that using catalyst concentrations greater than 0.25 wt % appeared to have a modest effect on reducing the TEQ concentration in the pentachlorophenol product. We desired to discover whether this behavior also occurred during synthesis with 2500 ppm of added Ni. To investigate whether adding more catalyst affects the TEQ concentrations, we ran experiments with 0.75 and 1.50 wt % AlCl3 catalyst. Results from a run with the normal catalyst concentration (0.25 wt %) appear in Figure 2c. In each of these experiments, the temperature ramp was terminated at 180 °C, where the chlorine flow was stopped. Figure 3 shows the results for the experiments with higher catalyst concentrations. Regardless of the amount of catalyst added, the final samples taken from the reactor had similar TEQ concentrations (0.25, 0.13, and 0.23 ppm for 0.25, 0.75 and 1.5 wt % catalyst, respectively). The TEQ concentrations in the postreaction samples taken from the aluminum pan were also similar (0.34, 0.25, and 0.38 ppm for 0.25, 0.75 and 1.5 wt % catalyst, respectively). It appears that running at catalyst concentrations higher than 0.25 wt %

has no appreciable effect on the TEQ concentrations when Ni is also added to the reactor. Effect of Ni Concentration. Having discovered that adding 2500 ppm of Ni during the synthesis reduced the TEQ concentration in the product, we next sought to determine the effect of using higher or lower concentrations of Ni. We first tested the effect of adding more Ni. Figure 4a shows the results from a synthesis experiment with 5000 ppm. The TEQ concentrations for the last sample from the reactor and for the postreaction sample were 0.26 and 0.44 ppm. These values do not differ appreciably from the TEQ levels obtained with less Ni, which indicates that the higher Ni level did not appear to reduce the TEQ level any further. Finding that adding more Ni did not have a beneficial effect on the TEQ concentration, we next sought to determine how little Ni was required to obtain a reduction in the TEQ concentration in the pentachlorophenol product. Figure 4b shows the results from synthesis using 1250 ppm. At 225 min, 84% pentachlorophenol was produced with 12% tetrachlorophenol and 0.33 ppm TEQ. At 234 min and 180 °C, 90% pentachlorophenol was obtained with 5% tetrachlorophenol residual and 0.71 ppm TEQ. The pentachlorophenol yield in the postreaction sample was 89% with 5% tetrachlorophenol and 0.37 ppm TEQ. We next used 833 ppm Ni, which is one-third of the base-case amount, in a synthesis experiment. Figure 4c gives the results. In the last sample from the reactor, at 180 °C and 255 min, the pentachlorophenol yield was 90% with 7% tetrachlorophenol remaining and a 0.37 ppm TEQ concentration. The TEQ concentration in the postreaction pentachlorophenol sample, where the pentachlorophenol yield was 87% with 6% tetrachlorophenol, was 0.39 ppm. Finally, Figure 4d shows results from synthesis with 100 ppm Ni. Again, the TEQ concentrations were about 0.4 ppm, even with this small amount of Ni added. The results in Figure 4b-d indicate that using 100, 833, or 1250 ppm of Ni provides about the same TEQ concentrations in the final pentachlorophenol product as using 2500 ppm Ni. This TEQ concentration is about 0.3-0.4 ppm. Synthesis Starting with Phenol. We synthesized pentachlorophenol starting with phenol to determine whether using a different starting material (phenol or 2,4,6-trichlorophenol) produced any differences in the results. Phenol is often used as the starting material in industrial practice. Since we normally start an experiment with about 300 g of trichlorophenol, we started these experiments with about 143 g of phenol, which should lead to an equivalent amount of 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. We added 0.75 g of Ni powder to the reactor at 105 °C (300 min). A 1.5 g sample of AlCl3 was used in this run, but it was added as 0.75 g at 60 °C and the second 0.75 g at 105 °C. Samples were collected about every 30 min. The reaction required about 10 h to reach the desired end point. We stopped the chlorine flow and temperature ramp at 180 °C to leave some tetrachlorophenol in the final product, knowing that this strategy provides a lower TEQ concentration in the product when starting with trichlorophenol.3 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, and the reactor temperature appear in Figure 5. We summed the yields of the two monochlorophenols and two dichlorophenols in the

5220

Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006

Figure 4. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with Ni powder added at 90 °C: (a) 5000 ppm Ni; (b) 1250 ppm Ni; (c) 833 ppm Ni; (d) 100 ppm Ni.

Figure 5. Temporal variation of chlorophenol yields and TEQ concentrations for synthesis with 2500 ppm Ni starting with phenol.

plot. The mono-, di-, and trichlorophenol yields increased to maximum values of 87% at 60 min, 81% at 121 min, and 92% at 246 min, respectively, and then decreased to zero at the time the next product reached its maximum yield (i.e., the monochlorophenol yield was zero at the time the dichlorophenol yield was at its maximum). The yield of tetrachlorophenol increased to its maximum of 79% at 434 min. At 589 min 12% tetrachlorophenol remained, where the pentachlorophenol yield was 79%. The TEQ concentrations were 0.14, 0.11, and 0.13

ppm for the samples with the maximum mono-, di-, and trichlorophenol yields. These quantitative results should be viewed as estimates, because the chemical matrix in these samples differed from that in the more highly chlorinated pentachlorophenol standards used to calibrate the bioassay. Nevertheless, the results indicate that the TEQ concentrations were very low (∼0.1 ppm) until tetrachlorophenol was formed. The TEQ concentration increased to 0.43 ppm at the maximum tetrachlorophenol yield. It subsequently increased further to 0.79

Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006 5221 Table 1. TEQ Concentrations (ppm) from Synthesis in the Presence of Different Metal Powders sample

metal

max tetrachlorophenol yield

next to last reactor sample

final reactor sample

postreaction sample

none3 Ni (mean) Ni (Figure 1c) Ni (Figure 2a) Ni (Figure 2b) Ni (Figure 2c) Ni (Figure 3a) Ni (Figure 3b) Ni (Figure 4a) Ni (Figure 4b) Ni (Figure 4c) Ni (Figure 4d) Mn Co Mo W V Ti Cu Fe Zn

0.38 ( 0.12 0.38 ( 0.14 0.34 0.39 0.37 0.28 0.25 0.33 0.23 0.58 0.36 0.35 0.33 0.18 0.40 0.31 0.74 0.49 1.01 0.34 0.15

0.32 ( 0.13 0.33 0.07 0.38 0.51 0.14 0.42 0.28 0.33 0.32 0.30 0.34 0.25 0.20 0.24 0.45 0.60 0.67 0.51 0.43

0.53 ( 0.14 0.39 ( 0.16 0.29 0.53 0.50 0.25 0.13 0.23 0.26 0.71 0.37 0.43 0.35 0.29 0.64 0.55 0.69 0.62 0.88 0.90 0.46

0.59 ( 0.18 0.42 ( 0.11 0.34 0.46 0.65 0.34 0.25 0.38 0.44 0.37 0.39 0.40 0.57 0.35 0.42 0.90 0.84 1.13 0.79 0.93 0.80

and 1.10 ppm for the last pentachlorophenol sample from the reactor and the postreaction pentachlorophenol sample, respectively. One can compare the TEQ values in Figure 5 with results reported previously3 from a similar phenol chlorination experiment without added Ni. The previous experiments without added Ni led to a TEQ concentration near the end of the run of about 1.9 ppm. The end-of-run TEQ concentrations in Figure 5 are only about half as large. This result shows again that adding Ni reduces the TEQ concentration in the pentachlorophenol product, in this instance for chlorination starting with phenol. Note, too, that these two experiments that started with phenol were done in the same reactor vessel and the results were sensitive to the presence or absence of Ni powder. This outcome implies that the reactor cleaning procedure used between runs is adequate to prevent Ni carryover from one run to the next. Effect of Other Metals. Since Ni powder has an effect on the TEQ concentration in pentachlorophenol, we reasoned that other metals might also have an effect. Therefore, we synthesized pentachlorophenol in the presence (2500 ppmw) of some other transition metals. These metals were added to the reactor as powders along with the AlCl3 catalyst. The metals tested were Mn, Co, Mo, W, V, Ti, Cu, Fe, and Zn. The temporal variations of the yields of chlorophenols were nearly identical in these experiments. The only noticeable difference was for synthesis in the presence of Cu. This reaction took longer than did the other experiments with added metal. It appears that copper inhibits the chlorination reaction. Table 1 summarizes the results for TEQ analyses from samples taken at comparable reaction points with these different added metals. The first row shows the mean values ((standard deviation) from experiments that stopped the chlorine flow between 177 and 183 °C but added no Ni. The second row shows the mean values ((standard deviation) for all runs with Ni added at the start of the reaction. The remaining rows show results from single analyses of single samples from an individual experimental run. The data in the first two rows show that adding Ni reduces the TEQ concentration in the postreaction sample by about one-third. The TEQ concentrations in row 2, with added Ni, are all about the same, whereas the TEQ

concentrations in row 1, without added Ni, appear to increase during the reaction. Thus, it appears that Ni inhibits the formation of toxic microcontaminants that would otherwise form after the maximum tetrachlorophenol yield has been reached. In the experiments with added metal, the TEQ concentrations were below 1 ppm in all but two samples. Some of the metals tested appear to be modestly more effective at lowering TEQ than others. If one lists the four metals that gave the lowest TEQ values for the last sample withdrawn from the reactor and the four metals that gave the lowest TEQ values for the sample taken from the postreaction solidified product in the aluminum pan, one finds that Ni, Co, and Mn appear on both lists. The mean TEQ concentration in these samples for these three metals was 0.39 ppm. Likewise, if one lists the four metals that gave the highest TEQ concentrations for both samples, Fe and V are appear on both lists. The mean TEQ concentrations for these samples with these two metals was 0.84 ppm. Seeing that Ni and Co powders were both effective in reducing the TEQ concentration in synthesized pentachlorophenol, we desired to discover how the presence of these metals influenced the relative amounts of the different microcontaminants. Therefore, we submitted three samples for analysis by high-resolution gas chromatography-mass spectrometry (GCMS). These were samples of pentachlorophenol synthesized from trichlorophenol with no added metal, with 100 ppm added Ni, and with 100 ppm added Co. These analyses revealed that the samples with added Ni and with added Co had roughly the same concentrations of the different polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs). In all three samples, the compound responsible for most (>70%) of the TEQ concentration was 1,2,3,4,6,7,8-heptaCDD. The concentration of this compound was about the same (55 ( 4 ppm) in all three samples. There were only two compounds for which there was a significant difference in the samples produced with and without metal. The concentration of hexaCDDs was roughly halved by adding Ni or Co. The concentration of octaCDD, on the other hand, was about 50% higher in the samples with added Ni and Co. These results indicate that the main effect of Ni or Co was to reduce the TEQ concentration in the samples by reducing the amount of 2,3,7,8-substituted hexaCDDs. Summary and Conclusions 1. With the conditions used in the present experiments, adding Ni powder to the reactor at the start of a synthesis run reduces the TEQ concentration in the final product by about one-third when starting the synthesis with trichlorophenol. The mean TEQ concentration in the postreaction samples was 0.4 ( 0.1 ppm. When starting with phenol, the TEQ reduction obtained by adding Ni powder was from about 1.9 to 1.1 ppm. 2. Nickel needs to be added before the chlorine flow is terminated for it to reduce the TEQ concentration in the product. This finding suggests that nickel inhibits the PCDF- and PCDDforming reactions occurring during the chlorination reaction rather than inhibiting their formation postreaction. 3. Nickel levels as low as 100 ppm were effective in reducing the formation of PCDFs and PCDDs. 4. Co and Mn appear to offer TEQ concentration reductions during pentachlorophenol synthesis that are comparable to those obtained with added Ni. Acknowledgment We are grateful for the assistance and GC-MS expertise provided by Dr. Yves Tondeur of Alta Analytical Perspectives. We acknowledge the financial support of the Microcontaminant

5222

Ind. Eng. Chem. Res., Vol. 45, No. 15, 2006

Reduction Venture, an industry-sponsored consortium committed to reducing microcontaminant levels in technical-grade pentachlorophenol.

(4) Yu, J.; Savage, P. E. Reaction Pathways in Pentachlorophenol Synthesis. 1. Temperature Programmed Reaction. Ind. Eng. Chem. Res. 2004, 43, 5021-5026.

Literature Cited

(5) Yu, J.; Savage, P. E. Reaction Pathways in Pentachlorophenol Synthesis. 2. Isothermal Reaction. Ind. Eng. Chem. Res. 2004, 43, 62926298.

(1) 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, 5199-5204. (2) 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, 5205-5210. (3) Yu, J.; Nestrick, T. J.; Savage, P. E. Microcontaminants in Pentachlorophenol Synthesis. 3. Effect of Temperature and Chlorine Flow Rate at End of Run. Ind. Eng. Chem. Res. 2006, 45, 5211-5216.

ReceiVed for reView February 20, 2006 ReVised manuscript receiVed May 15, 2006 Accepted May 20, 2006 IE060213I