Efficient Polymerization Inhibition Systems for Acrylic Acid Distillation

Mar 5, 2012 - Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, United States. ‡ Matière Moll...
0 downloads 9 Views 1MB Size
Article pubs.acs.org/IECR

Efficient Polymerization Inhibition Systems for Acrylic Acid Distillation: Vapor-Phase Inhibitors Renaud Nicolay,̈ †,‡ Jaroslav Mosnácě k,†,§ Kishore K. Kar,⊥ Stanley O. Fruchey,⊥ Michael D. Cloeter,⊥ Richard S. Harner,⊥ and Krzysztof Matyjaszewski*,† †

Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennsylvania 15213, United States Matière Molle et Chimie, ESPCI-CNRS (UMR 7167), 10 rue Vauquelin, 75005 Paris, France § Polymer Institute, Centre of Excellence GLYCOMED, Slovak Academy of Sciences, Dúbravská cesta 9, 845 41 Bratislava, Slovakia ⊥ The Dow Chemical Company, Midland, Michigan 48674, United States ‡

ABSTRACT: Nitroso compounds, such as 2-methyl-2-nitrosopropane, nitrosobenzene, and 4-nitrosophenol, were tested as volatile inhibitors for vapor-phase inhibition of acrylic acid polymerization during its reflux at 113 °C under reduced pressure. The experimental parameters were set to mimic the conditions employed for the industrial distillation of acrylic acid. Nitrosobenzene was found to be the most efficient vapor-phase inhibitor for the distillation of acrylic acid. The inhibition time was also found to be dependent on the pressure used for the experiments. Decreasing the pressure from 110 to 85 mbar resulted in an almost 2-fold increase in the inhibition time. The combination of nitrosobenzene with a liquid-phase inhibitor, namely, 4hydroxy-2,2,6,6-tetramethylpiperidine 1-oxyl, significantly improved the efficiency of vapor-phase inhibition by decreasing the consumption of nitrosobenzene in the liquid phase.



INTRODUCTION Chemical companies face long-term operational problems at plants distilling very reactive monomers, such as acrylic acid (AA). These very reactive monomers can foul the distillation column over time, requiring the distillation towers to be shut down and cleaned after several months. Effective polymerization inhibitors are essential for the safe and continuous operation of manufacturing plants, as well as for the safe storage of acrylic monomers. While hydroquinone monomethyl ether, p-benzoquinone, or phenothiazine (PTZ) are usually used as commercial liquid-phase inhibitors/stabilizers,1−4 nitric oxide gas is sometimes used as a radical trap in the gaseous phase. However, nitric oxide gas does not efficiently inhibit polymerization at the surfaces of distillation plates or within the tower interior and cannot prevent fouling with poly(acrylic acid). In the present work, we focused on the development of volatile inhibitors, and inhibition systems suitable for both the gaseous and liquid phases, in order to allow long-term operation of the distillation tower without cleaning. Many radical traps can potentially be used as volatile inhibitors. Efficient volatile inhibitors should fulfill at least the following two characteristics: (i) they should have a boiling point (bp) close to the bp of AA; (ii) they should be efficient radical traps. The first characteristic implies that the radical trap should be a relatively low-molar-mass organic molecule. Nitroxides and nitroso compounds can meet both characteristics. On the one hand, the bp depends on the structure of the compound; on the other hand, nitroxides and nitroso compounds are known to be efficient radical traps. Nitroxides and alkoxyamines are formed by single or double addition of radicals to nitroso compounds, respectively (Scheme 1).5,6 Therefore, nitroso compounds present an advantage over nitroxides because they can trap 2 © 2012 American Chemical Society

Scheme 1. Radical Inhibition Mechanism with Nitroso-tertbutane

equiv of radicals compared to nitroxides, which can react with only 1 equiv of radicals.



EXPERIMENTAL PART Materials. Nitroso-tert-butane, nitrosobenzene, 4-nitrosophenol, phenothiazine (PTZ), 4-hydroxy-2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPOL), manganese acetate tetrahydrate, and all other reagents were purchased from Aldrich and used as received. N-(1-Phenylethyl)phenothiazine (PEPTZ) was synthesized according to the procedure previously reported in the literature.7 Acrylic acid (AA), purchased from Aldrich, was purified by two successive distillations.7 Testing Experiments and Analysis. For the volatile inhibition tests, a flask containing various concentrations of both volatile and liquid-phase inhibitors was equipped with a glass column containing stainless steel chips and a condenser on the top (Figure 1). The solution was heated at 113 °C under a reduced pressure of 110 mbar, unless otherwise specified. The time for formation of a solid polymer in the packing was determined visually. Received: Revised: Accepted: Published: 4467

August 2, 2011 February 26, 2012 March 4, 2012 March 5, 2012 dx.doi.org/10.1021/ie201709y | Ind. Eng. Chem. Res. 2012, 51, 4467−4471

Industrial & Engineering Chemistry Research

Article

Figure 3. Liquid-phase inhibitors used for the vapor-phase inhibition tests: PTZ, phenothiazine; PEPTZ, N-(1-phenylethyl)phenothiazine; TEMPOL, 4-hydroxy-2,2,6,6-tetramethylpiperidine 1-oxyl.

Table 1. Vapor-Phase Inhibition of AA at 113 °C and 110 mbar in the Presence of 500 ppm of PTZ refluxing time before the appearance of the solid polymer (min)

Figure 1. Experimental setup for volatile inhibitor tests.



RESULTS AND DISCUSSION Two commercially available nitroso compounds were selected for preliminary studies: 2-methyl-2-nitrosopropane (I; 2methyl-2-nitrosopropane, also called nitroso-tert-butane, exists in equilibrium with its dimer with an equilibrium position dependent on the temperature; the bp is not available) and nitrosobenzene (II; bp = 59 °C at 18 mmHg; Figure 2). In

a

volatile inhibitor

0 ppm

I II

30a 30a

100 ppm

500 ppm

192a

50a 450a (360)b

b

Polymer formed in the packing. Polymer formed in the liquid phase (PAA is not soluble in pure AA).

compared to 30 min without the volatile inhibitor. However, a significant improvement was observed when nitrosobenzene was used as the volatile inhibitor. Inhibition times of 190 and 450 min before the appearance of the solid polymer in the packing were observed when using 100 and 500 ppm of nitrosobenzene, respectively (Table 1). Surprisingly, some poly(acrylic acid) formed in the liquid phase after 6 h when 500 ppm of nitrosobenzene was used in conjunction with 500 ppm of PTZ. Such a result was unexpected for two reasons. First, the inhibition time in the vapor phase was 7.5 h for the same experiment, while the total concentration of the inhibitor in the liquid phase was at least the same, if not higher, than that in the vapor phase, as 500 ppm of nonvolatile PTZ was used in conjunction with the 500 ppm of nitrosobenzene. Second, when 100 ppm of PTZ in AA was tested alone as a liquid-phase inhibitor at 113 °C, the observed time to gelation was around 64 h.7 The shorter induction time obtained in the liquid phase compared to that in the vapor phase could indicate that side reactions between PTZ and nitroso compounds or nitroxides are taking place in acidic media. A simple qualitative model reaction was designed to assess the possibility of side reactions between nitroxides and PTZ. TEMPOL and PTZ were mixed and stirred in anisole at room temperature (RT) in order to observe any change of color of the solution. Solutions of TEMPOL and its derivates are typically orange or red, a color characteristic of nitroxide radicals. If nitroxides are reduced by PTZ, the color of the solution should change. After 2 h of stirring, no color change was observed, and the color of the solution remained orange (Figure 4A). At this point, the reaction media was acidified by introducing acetic acid (in excess compared to TEMPOL and PTZ). Acetic acid was chosen as the saturated equivalent of AA. At this point, the color instantaneously changed to dark green and then progressively to light green (Figure 4B−D).

Figure 2. Nitroso compounds tested here as vapor-phase inhibitors: I, nitroso-tert-butane; II, nitrosobenzene; III, 4-nitrosophenol.

order to prevent polymerization of AA in the liquid phase and/ or to limit consumption of the tested volatile inhibitors in the liquid phase, the nitroso compounds were tested in combination with PTZ, a commonly used liquid-phase inhibitor (Figure 3). PTZ and its derivatives inhibit radical polymerization via a catalytic process in an AA environment.7,8 Monomeric or polymeric carbon-centered radicals are reduced by single-electron transfer to form the corresponding carbanion and a PTZ N-radical cation. Protonated PTZ is then generated by abstracting a hydrogen atom from another monomeric or polymeric radical, thereby forming a C−C double bond. PTZ is finally regenerated by an acid−base reaction between the protonated PTZ and a monomeric or polymeric carboxylate anion. As shown in Table 1, the use of 500 ppm of nitroso-tertbutane as a volatile inhibitor yielded only a moderate improvement compared to the blank experiments. The inhibition time before apparition of the solid polymer in the packing was 50 min with 500 ppm of nitroso-tert-butane, 4468

dx.doi.org/10.1021/ie201709y | Ind. Eng. Chem. Res. 2012, 51, 4467−4471

Industrial & Engineering Chemistry Research

Article

Figure 4. Solution of TEMPOL and PTZ in anisole: (A) after 2 h at RT; (B) immediately after introduction of acetic acid; (C) 1.5 h after introduction of acetic acid; (D) 3.5 h after introduction of acetic acid. Solution of TEMPOL in anisole: (E) after 15 min at RT; (F) 2 h after introduction of acetic acid. Solution of PTZ in anisole: (G) after 15 min at RT; (H) 2 h after introduction of acetic acid.

Scheme 2. Possible Mechanism for the Redox Reaction between TEMPOL and PTZ

Table 2. Volatile Inhibitor Tests at 113 °C entry

a

P (mbar) vial no. inhibition timea (h)

inhibitor system

1

100 ppm of nitrosobenzene + 100 ppm of PEPTZ + 10 ppm of Mn Ac2·4H2O

110

2

200 ppm of nitrosobenzene + 10 ppm of MnIIAc2·4H2O

110

3

500 ppm of nitrosobenzene + 10 ppm of MnIIAc2·4H2O

110

4

200 ppm of nitrosobenzene + 10 ppm of MnIIAc2·4H2O

85

5

100 ppm of nitrosobenzene + 100 ppm of TEMPOL + 10 ppm of MnIIAc2·4H2O

110

6

100 ppm of nitrosobenzene + 100 ppm of TEMPOL

110

7

100 ppm of nitrosophenol + 100 ppm of PEPTZ

110

II

1 2 3 1 2 3 1 2 1 2 3 1 2 3 1 2 3 1 2 3

5 5.5 7 7.25 12 14 18 21.5 19.5 20.5 22 10.5 14 17.5 11.8 15.5 18.1 0.83 1 1.17

average inhibition time (h) 5.83 ± 1.0

11.1 ± 3.5

19.8 ± 2.5 20.7 ± 1.3

14 ± 3.5

15.1 ± 3.2

1 ± 0.17

Refluxing time before the appearance of the solid polymer in the packing.

Two control experiments were run in parallel. Solutions of TEMPOL in anisole and PTZ in anisole were prepared. After

15 min at RT, the solutions were acidified with acetic acid. In both cases, no change in color was observed; the TEMPOL and 4469

dx.doi.org/10.1021/ie201709y | Ind. Eng. Chem. Res. 2012, 51, 4467−4471

Industrial & Engineering Chemistry Research

Article

nitrosobenzene reacts faster with AA radicals in the liquid phase than PEPTZ does, assuming that no side reaction takes place between nitroxide radicals and PEPTZ. A further increase in the concentration of nitrosobenzene, up to 500 ppm, led to an increase in the inhibition time, up to almost 20 h. The influence of the distillation pressure was also investigated (Table 2, entries 2 and 4). The bp of AA is 139 °C, while the bp of nitrosobenzene is 59 °C/18 mmHg, which approximately corresponds to 160 °C at atmospheric pressure. Therefore, one should be able to increase the amount of nitrosobenzene in the gas phase by decreasing the pressure. Indeed, the average inhibition time of 11.1 h at 110 mbar and 113 °C, when using 200 ppm of nitrosobenzene and 10 ppm of manganese acetate tetrahydrate, was increased to 20.7 h when the distillation pressure was decreased to 85 mbar. In addition, the experiments at 85 mbar yielded valuable information on which direction one could eventually follow to further improve the system, i.e., find a commercially available nitroso compound with a lower bp or decrease the pressure during AA distillation. To maximize the consumption of the volatile inhibitor, i.e., nitrosobenzene, in the vapor phase and minimize its consumption in the liquid phase, we investigated the use of nitrosobenzene in conjunction with another liquid-phase inhibitor, e.g., the nitroxide TEMPOL. Nitroxides react with carbon radicals to form alkoxyamines with rate constants on the order of 108−109 M−1 s−1, i.e., with a rate close to diffusion control.11−16 Such a radical trap should therefore allow for a maximum “preservation” of the volatile inhibitor for the vapor phase. The results with 100 ppm of nitrosobenzene, 100 ppm of TEMPOL, and 10 ppm of manganese acetate tetrahydrate are 25% better than the results obtained with 200 ppm of nitrosobenzene and 10 ppm of manganese acetate tetrahydrate, with an average inhibition time of 14 h versus 11.1 h, respectively (Table 2, entry 5). The mechanism responsible for the regeneration of nitroxides from hydroxylamine in the presence of manganese(II) is not yet completely established.10 Therefore, we were also interested in conducting experiments without manganese acetate tetrahydrate. A test was conducted with only 100 ppm of nitrosobenzene and 100 ppm of TEMPOL (Table 2, entry 6). This inhibitor combination gave an average inhibition time of 15.1 h, which is comparable to the 14 h observed when 10 ppm of manganese acetate tetrahydrate was used in addition to the 100 ppm of nitrosobenzene and 100 ppm of TEMPOL. This indicates that the regeneration of nitroxides from hydroxylamines catalyzed by manganese(II) was limited under reduced pressure, i.e., with a limited amount of air. Finally, a test aiming to confirm the importance of the nitroso compound’s structure was realized. This test was designed to exemplify the benefit of choosing the right nitroso compound for a specific monomer. A total of 100 ppm of 4nitrosophenol was used in conjunction with 100 ppm of TEMPOL, instead of 100 ppm of nitrosobenzene (Table 2, entry 7). The only structural difference between these two compounds is the presence of a hydroxy substituent in the para position of the nitroso moiety in 4-nitrosophenol. This hydroxyl group, however, significantly increases the melting point (mp) and the bp of the nitroso compound; nitrosobenzene, bp = 59 °C/18 mmHg → bp = ∼160−165 °C at atmospheric pressure; 4-nitrosophenol, mp 132−144 °C. The very low inhibition period obtained with 4-nitrosophenol, i.e., 1 h, reflects the fact that this compound remains in the liquid phase and cannot act as a vapor-phase inhibitor under the

PTZ solutions remained orange and yellow, respectively (Figure 4E−H). These experiments confirmed the occurrence of a side reaction between nitroxide radicals and PTZ in acidic media. This side reaction could correspond to the reduction of nitroxide radicals to hydroxylamine by PTZ. A possible mechanism is shown in Scheme 2. Such a side reaction should not occur with N-alkylated derivatives of PTZ. In a parallel study, we showed that Nalkylated derivatives of PTZ, such as PEPTZ, are more efficient liquid-phase inhibitors for AA compared to standard PTZ.7 Therefore, a model reaction using PEPTZ and TEMPOL was conducted. TEMPOL and PEPTZ were mixed and stirred in anisole at RT in order to see if there was an eventual change of the color of the solution. After 15 min of stirring, no color change was observed and the solution remained orange. At this point, the medium was acidified by introducing acetic acid (in excess compared to TEMPOL and PEPTZ). The color progressively became slightly darker and finally became red. The reason for this color change is unclear because the red color also is characteristic of nitroxide radicals and one of the PTZ oxidative byproducts is known to have a dark-red-brown color.3,8 However, such a byproduct is not expected to form with N-alkylated derivatives of PTZ. In addition, when PEPTZ was tested in conjunction with TEMPOL as the liquid-phase inhibition system for AA, a synergistic effect was observed and longer inhibition periods were obtained.9 Taking these observations into account, a mixture of inhibitors consisting of 100 ppm of PEPTZ, 100 ppm of nitrosobenzene, and 10 ppm of manganese acetate tetrahydrate was tested as a vaporphase inhibitor system for AA at 113 °C and 110 mbar (Table 2, entry 1). Manganese(II) compounds can be used in combination with nitroxides as liquid-phase inhibition systems. Their role in such systems is to regenerate nitroxides from hydroxylamines.10 As shown in Table 2, an average reflux time of 5.83 h was observed before the appearance of the solid polymer in the packing. This inhibition period is almost twice as long as that obtained with 500 ppm of PTZ/100 ppm of nitrosobenzene, i.e., an inhibition time of 3.2 h (Table 1). This can be ascribed to the suppression of the side reaction between PTZ and nitroxides formed by the single addition of an AA radical to nitrosobenzene (Scheme 2). To improve the vapor-phase inhibition without increasing the concentration of the volatile inhibitor, one should try to have the volatile inhibitor predominately consumed in the vapor phase and as little as possible in the liquid phase. In other words, the liquid-phase inhibitor should ideally react faster with AA radicals in solution than the vapor-phase inhibitor, in order to “save” the volatile inhibitor for the vapor phase. Rate constants of the reaction of AA radicals with PEPTZ and nitrosobenzene are not available in the literature, but it is possible that nitrosobenzene reacts with AA radicals at a similar rate, or even faster than PEPTZ does. Therefore, the efficiency of nitrosobenzene as the volatile inhibitor for AA distillation was investigated alone, i.e., without PTZ derivatives as liquid-phase inhibitors. Tests with various concentrations of nitrosobenzene (200 or 500 ppm) and a constant concentration of manganese acetate tetrahydrate (10 ppm) were run (Table 2; entries 2 and 3). An average inhibition time of 11 h was observed when using 200 ppm of nitrosobenzene alone. This inhibition period is twice as long as the inhibition period obtained with a mixture of 100 ppm of nitrosobenzene and 100 ppm of PEPTZ, even though the total inhibitor concentration is the same. This tends to confirm that 4470

dx.doi.org/10.1021/ie201709y | Ind. Eng. Chem. Res. 2012, 51, 4467−4471

Industrial & Engineering Chemistry Research

Article

(9) Kar, K. K.; Fruchey, O. S.; Cloeter, M. D.; Harner, R. S.; Matyjaszewski, K.; Nicolay, R.; Mosnacek, J. PCT Int. Appl. WO 2010096512 A1 20100826, 2010. (10) Minisci, F.; Recupero, F.; Cecchetto, A.; Gambarotti, C.; Punta, C.; Faletti, R.; Paganelli, R.; Pedulli, G. F. Mechanisms of the aerobic oxidation of alcohols to aldehydes and ketones, catalysed under mild conditions by persistent and non-persistent nitroxyl radicals and transition metal saltsPolar, enthalpic, and captodative effects. Eur. J. Org. Chem. 2004, 1, 109−119. (11) Starnes, W. H. Jr.; Schilling, F. C.; Abbas, K. B.; Cais, R. E.; Bovey, F. A. Mechanism for the formation of chloromethyl branches in poly(vinyl chloride). Macromolecules 1979, 12 (4), 556−562. (12) Park, G. S.; Saleem, M. The mechanism of the formation of chloromethyl side groups in vinyl chloride polymerization. Polym. Bull. 1979, 1 (6), 409−413. (13) Beckwith, A. L. J.; Bowry, V. W.; Moad, G. Kinetics of the coupling reactions of the nitroxyl radical 1,1,3,3-tetramethylisoindoline-2-oxyl with carbon-centered radicals. J. Org. Chem. 1988, 53 (8), 1632−1641. (14) Chateauneuf, J.; Lusztyk, J.; Ingold, K. U. Absolute rate constants for the reactions of some carbon-centered radicals with 2,2,6,6-tetramethyl-1-piperidinoxyl. J. Org. Chem. 1988, 53 (8), 1629− 1632. (15) Bowry, V. W.; Ingold, K. U. Kinetics of nitroxide radical trapping. 2. Structural effects. J. Am. Chem. Soc. 1992, 114 (13), 4992− 4996. (16) Beckwith, A. L. J.; Bowry, V. W.; Ingold, K. U. Kinetics of nitroxide radical trapping. 1. Solvent effects. J. Am. Chem. Soc. 1992, 114 (13), 4983−4992.

conditions of the test. Using nitroso compounds with higher bp's might be suitable for less volatile monomers, e.g., 2ethylhexyl acrylate, 1-vinyl-2-pyrrolidinone, etc., provided that the nitroso compounds do not decompose before they distill.



CONCLUSIONS An efficient vapor-phase inhibition system for the distillation of AA was described. The system is based on the use of a volatile inhibitor, namely, nitrosobenzene, presenting a slightly higher bp than that of AA. This is a valuable characteristic to ensure that (1) the concentration of the volatile inhibitor in the vapor phase is sufficient to provide efficient inhibition and (2) the amount of the volatile inhibitor distilled out of the system is minimized. It was also demonstrated that the inhibition efficiency can be further improved by optimizing the distillation pressure. In order to decrease the consumption of nitrosobenzene in the liquid phase, nitrosobenzene was used in conjunction with efficient liquid-phase inhibitors, such as N-alkylated derivatives of PTZ or TEMPOL, with the latter reacting with radicals with rates close to diffusion control. Such a combination of inhibitors led to a dramatic improvement of the inhibition efficiency of nitrosobenzene in the vapor phase. Although the vapor-phase inhibition system developed here was only tested for AA distillation, the principles can be applied to other distillable monomer systems, by selecting a volatile inhibitor with the appropriate bp and/or by optimizing the conditions employed for distillation, such as the pressure.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (K.M.). Tel.: +1 412 268 3209. Fax: +1 412 268 6897. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Cutie, S. S.; Henton, D. E.; Powell, C.; Reim, R. E.; Smith, P. B.; Staples, T. L. The effects of MEHQ on the polymerization of acrylic acid in the preparation of superabsorbent gels. J. Appl. Polym. Sci. 1997, 64 (3), 577−589. (2) Becker, H.; Vogel, H. The role of hydroquinone monomethyl ether in the stabilization of acrylic acid. Chem. Eng. Technol. 2006, 29 (10), 1227−1231. (3) Becker, H.; Vogel, H. Phenothiazine as stabilizer for acrylic acid. Chem. Eng. Technol. 2006, 29 (8), 931−936. (4) Li, R. J.; Schork, F. J. Modeling of the inhibition mechanism of acrylic acid polymerization. Ind. Eng. Chem. Res. 2006, 45 (9), 3001− 3008. (5) Maschke, A.; Shapiro, B. S.; Lampe, F. W. Mechanism of the lowtemperature scavenging of methyl-d3 radicals by NO. J. Am. Chem. Soc. 1963, 85, 1876−1878. (6) Hoffmann, A. K.; Feldman, A. M.; Gelblum, E.; Hodgson, W. G. Mechanism of the formation of Di-tert-butyl nitroxide from tertnitrobutane and sodium. J. Am. Chem. Soc. 1964, 86 (4), 639−646. (7) Mosnácě k, J.; Nicolay, R.; Kar, K. K.; Fruchey, S. O.; Cloeter, M. D.; Harner, R. S.; Matyjaszewski, K. Efficient Polymerization Inhibition Systems for Acrylic Acid Distillation: New Liquid Phase Inhibitors. Ind. Eng. Chem. Res. 2012, 51 (10), 3910−3915. (8) Levy, L. B. Inhibition of acrylic acid polymerization by phenothiazine and p-methoxyphenol. II. Catalytic inhibition by phenothiazine. J. Polym. Sci., Part A: Polym. Chem. 1992, 30 (4), 569−576. 4471

dx.doi.org/10.1021/ie201709y | Ind. Eng. Chem. Res. 2012, 51, 4467−4471