1962
Ind. Eng. Chem. Res. 2004, 43, 1962-1965
Hydrogen Sulfide Scavenging by 1,3,5-Triazinanes. Comparison of the Rates of Reaction Jan M. Bakke* and Janne B. Buhaug Department of Chemistry, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway
The reaction of 1,3,5-triethyl-1,3,5-triazinane (1b) and 1,3,5-trimethyl-1,3,5-triazinane (1c) with sodium hydrogen sulfide has been studied. The products were the corresponding thiadiazinanes 2 and dithiazinanes 3. The rates of the reaction, together with that of 1,3,5-tris(2-hydroxyethyl)1,3,5-triazinane (1a), were studied by 1H NMR spectroscopy. Their relative rates of reaction with HS- were as follows: 1c, 1; 1b, 19; 1a, 31. The rates of hydrolysis were studied for the same three compounds. In this reaction, the relative rates were as follows: 1c, 1; 1b, 12; 1a, 24. The reactions of 1,3,5-trioxane (4a) and 2,4,6-trimethyl-1,3,5-trioxane were also studied. These two compounds did not react with HS- or were hydrolyzed under the conditions used for the reactions of 1a-c. Introduction Dihydrogen sulfide, H2S, is present in low concentrations in natural gas from many sources. Because of its obnoxious characteristics,1 it is necessary to remove it during processing, before the end use. 1,3,5-Tris(2hydroxyethyl)-1,3,5-triazinane (1a, later triazinane) is an effective scavenger in this process. We have reported a study of the products from the reaction of triazinane with H2S and also its rate of hydrolysis as a function of pH and temperature.2 In 1a, the hydroxyl groups on the side chains increase the hydrophilicity of the molecule, making it eminently suited for reactions in the aqueous phase. However, it might be of interest to remove H2S directly in the well itself, where the medium might have a more hydrophobic character. For this purpose, more lipophilic side chains might be of interest. We have therefore investigated the reactions between H2S and two such substances with methyl or ethyl groups at the nitrogen atoms of the triazinane ring, 1,3,5-triethyl1,3,5-triazinane (1b, later ethyltriazinane) and 1,3,5trimethyl-1,3,5-triazinane (1c, later methyltriazinane). We have also studied the reactions of two oxygen heterocyclic compounds, 1,3,5-trioxane (4a) and 2,4,6trimethyl-1,3,5-trioxane (4b), with H2S. Results and Discussion Products from the Reaction of the Nitrogen Heterocyclic Compounds 1a-c with H2S. The products from the reaction of 1a with H2S were identified as 3,5-bis(2-hydroxyethyl)-1,3,5-thiadiazinane (2a) and 5-(2-hydroxyethyl)-1,3,5-dithiazinane (3a). The identification was carried out by extensive use of NMR spectroscopy.2 We have used the same techniques for the identification of the products from the reactions of 1b and 1c with H2S/HS-. 1b was reacted with 2 equiv of NaHS at pH 10.0. By analogy with the reaction of 1a with HS- under the same conditions, the formation of 3,5-diethyl-1,3,5thiadiazinane (2b) and 5-ethyl-1,3,5-dithiazinane (3b), in addition to ethylamine, was expected. The NMR * To whom correspondence should be addressed. Tel.: +4773594095. Fax: +4773594256. E-mail:
[email protected].
spectroscopy data (Table 3) confirmed this. The 1H NMR spectrum of the reaction mixture showed this to contain unreacted 1b (15%), 2b (8%), and 3b (10%), in addition to ethylamine (67%). In an experiment in which 1b was reacted with 1 equiv of NaHS, only 2b and ethylamine were formed, supporting the previous assignments. In the reaction mixture with 2 equiv of HS-, a nonaqueous phase was observed in the aqueous reaction mixture after 4 weeks at room temperature. A 1H NMR spectrum of the nonaqueous layer showed it to contain compounds 2b and 3b in relative amounts of 1:7. At the same time, a 1H NMR spectrum of the remaining aqueous phase showed that 2b was present but there were only traces of 3b. This observation corresponds well with the assignment of compounds 2b and 3b because substitution of a ring nitrogen with sulfur would make the molecule less polar. 1,3,5-Trithiane has a low solubility in water. In the previous experiments, 1b was reacted with a solution of sodium hydrogen sulfide. Upon H2S removal, an aqueous solution of 1b would react with gaseous H2S. We therefore performed a reaction under these conditions by bubbling H2S through an aqueous solution of 1b. After 10 min, a 1H NMR spectrum showed that 3b and ethylamine had been formed. Neither 1b nor 2b was found, indicating that the reaction was complete under these conditions. The products from the reaction of 1c were investigated by first reacting them with 2 equiv of NaHS at pH 10.0. The reaction was expected to proceed in the same way as those of 1a and 1b, with the formation of 3,5-dimethyl-1,3,5-thiadiazinane (2c) and 5-methyl1,3,5-dithiazinane (3c), in addition to methylamine. The reaction was very slow, but after 3 days, a precipitate of white, needle-shaped crystals was observed. The crystals and the mother liquid were investigated by the same NMR techniques as those used previously. The data are given in the Experimental Section (Table 4). The results were in accordance with the reaction path in Scheme 1, 1c reacted with HS- to give 2c, and this reacted further with HS- to give 3c. The crystals were identified as 3c. Only traces of the signals from 3c were observed in the 1H and 13C NMR spectra of the reaction mixture. The proposed structure of 3c corresponds well with this
10.1021/ie030510c CCC: $27.50 © 2004 American Chemical Society Published on Web 04/02/2004
Ind. Eng. Chem. Res., Vol. 43, No. 9, 2004 1963 Scheme 1
Figure 1. Half-lives of 1a-c at different pHs. Table 1. Rate Constants for the Hydrolysis of 1a-c at Different pHs and Temperatures compd
pH
kobs (10-4 s-1)
1a
10.0 9.0 8.0 10.0 9.0 8.0 10.0 9.0 8.0 7.5
2.0 ( 0.1 15 ( 1 143 ( 1 1.2 ( 0.1 7.9 ( 0.1 70 ( 1 0.05 ( 0.01 0.55 ( 0.01 6.8 ( 0.1 18 ( 1
1b 1c
k1 (10-5 s-1)
k2 (105 M-1 s-1)
7(2
14 ( 1
7(3
6.9 ( 0.1
3(4
0.58 ( 0.03
Chart 1
conditions, and the electrophilicity of 4 would be lower than that of 1H+. Kinetic Investigations. (a) Hydrolysis. To study the reactivity of the triazinanes 1a-c toward H2S/HS-, it was necessary to first determine their rates of hydrolysis, an important side reaction. We have reported the kinetics of the hydrolysis of 1a.1 The rate of the reaction followed the rate law (1), where [T] is the because the dithiazinane compound with two sulfur atoms in the ring is expected to be the least watersoluble of the components in the reaction mixture. Again, it was important to see if the same products were formed when 1c was reacted with gaseous H2S. The gas was bubbled through an aqueous solution of 1c for 23 min. A 1H NMR spectrum was recorded. This showed that all of the starting material had disappeared, while methylamine and dithiazinane 3c had been formed. This was consistent with the results for 1a and 1b, which also formed the corresponding dithiazinane compounds upon reaction with H2S. The results are summarized in Scheme 1. Reactions with the Oxygen Heterocyclic Compounds 4a,b (Chart 1). In 4a and 4b, the nitrogen atoms of 1a-c have been substituted with oxygen atoms. Oxygen is more electronegative than nitrogen, and this might make the carbon atoms of the ring more electrophilic. Compounds 4a and 4b might therefore be more reactive both toward HS- and toward hydrolysis. Accordingly, we reacted the two compounds with HSand monitored the reaction by 1H NMR spectroscopy. However, neither 4a nor 4b reacted with HS-. Because the reaction with the nitrogen heterocycles 1a-c was pH dependent, we varied the pH of the reaction from 10 to 7, but no reaction with HS- was observed. The compounds were also stable for several days toward hydrolysis at this pH range. This difference in reactivity was probably caused by the difference in basicity between compounds 1 and 4. The reacting species in the case of the triazinanes was not 1 but 1H+. The analogous 4H+ would not have been formed under the applied
-d[T]/dt ) k1[T] + k2[T][H+]
(1)
concentration of the triazinane investigated, k1 the rate constant for the uncatalyzed reaction, and k2 that for the catalyzed one. With constant pH
kobs ) k1 + k2[H+]
(2)
The concentration of triazinane was monitored by 1H NMR spectroscopy. The observed first-order rate constants for the hydrolysis reactions of 1b and 1c at three different pHs are given in Table 1. From these and eq 2, k1 and k2 for 1b and 1c were calculated. The results, together with the rate constants for 1a,3 are also given in Table 1. Table 1 shows that the uncatalyzed hydrolysis reaction as expressed by k1 was insignificant compared to the catalyzed one. It is, furthermore, evident that 1c was the most stable and 1a the least stable toward hydrolysis of the three triazinanes investigated. This indicated that 1c might be used at a lower pH than 1b and particularly 1a for H2S scavenging. From k2 in Table 1, it was possible to calculate the half-life of each of the three scavengers at different pHs. This is shown by the graphs in Figure 1. Once the rates of hydrolysis for the three triazinanes had been established, their reactions with H2S/HSwere studied at pH levels where the rates of hydrolysis were low enough to be neglected. It should be noted that these experiments were carried out on rather dilute solutions of triazinane, 0.1 vol %. In concentrated solutions where the activity of water is lower, the rate
1964 Ind. Eng. Chem. Res., Vol. 43, No. 9, 2004
of hydrolysis would be expected to be slower than those reported here. (b) Reactions with HS-. Introductory experiments indicated that the rate of the reaction between sodium NaHS with triazinane 1a is dependent on the pH of the medium, and a reaction mechanism explaining this point was proposed.2 It was, therefore, reasonable to assume that protonated (hydroxyethyl)triazine (TH+) was the reacting species. In the investigated pH region, HS- is completely dominating in the H2S/HS/S2- equilibria.4 From this, the reaction was assumed to be of second order with the rate law (3).
-d[T]/dt ) k[TH+][HS-]
(3)
The hydrolysis reaction can be neglected because of the pH range of the analyses. The concentration of the protonated triazinane is related to pH through Ka:
Ka ) [T][H+]/[TH+]
(4)
Substituting for [TH+] in eq 3 gives
-d[T]/dt ) k′3[T][H+][HS-]
(5)
where k′3 ) kKa-1. If the kinetic experiments are carried out at constant pH, [H+] will be constant. This results in an apparent second-order reaction, with k′2 ) k′3[H+]. The new rate expression is
-d[T]/dt ) k′2[T][HS-]
(6)
This can be simplified further by keeping the concentration of HS- constant. This is done by running the experiment with excess HS- ([HS-] g 10[T]). The change in [HS-] then becomes negligible, and the reaction can be considered a pseudo-first-order reaction. Only one parameter (the concentration of the triazinane) needs to be monitored over time, and the rate expression is
-d[T]/dt ) k′1[T]
(7)
k′1 ) kKa-1[H+][HS-]
(8)
where
By monitoring of [T], k′1 will be obtained from eq 7. The other rate constants k′2 and k′3 can be calculated from k′1, [H+], and [HS-]. It should be noted that an analogous set of equations could be derived in which [T] was kept constant and [HS-] monitored. In principle, this could be done simply by use of a sulfide-sensitive electrode. However, despite considerable effort, we were not able to get reproducible results by that approach. We therefore chose to run the kinetic investigations at constant pH, with a 10-fold excess of NaHS and monitoring of [T] by 1H NMR spectroscopy. Acetonitrile was used as an internal standard. The results from the kinetic investigations of the reactions of NaHS with 1a, 1b, and 1c at different pHs are given in Table 2. From these data, with constant pH and [HS-], the rate constants k′3 ) kKa-1 were obtained and are also given in Table 2. It contains a measure of the basicity of the amino nitrogen of 1 as expressed by Ka. There is no reason to believe that Ka will be very different for the three triazinanes investigated and, furthermore, the
Table 2. Rate Constants (k′3 ) kKa-1) for the Reaction of 1a-c with [HS-] ) 0.48 M at 22 °C compd 1a 1b 1c
pH
k′1 (10-3 s-1)
11.0 10.5 10.0 11.0 10.5 10.0 11.0 10.5 10.0
0.41 ( 0.01 1.3 ( 0.1 4.3 ( 0.1 0.21 ( 0.01 0.61 ( 0.01 2.6 ( 0.1 0.014 ( 0.003 0.056 ( 0.002 0.14 ( 0.01
k′3 (107 M-2 s-1) 9.1 ( 0.2 5.6 ( 0.4 0.29 ( 0.03
basicity of the amino groups will be an integrated part of the reactivity of the triazinanes. We, therefore, consider that the k′3 obtained gives a valid expression of the relative reactivity of the three triazinanes toward H2S/HS-. Conclusion In the previous sections, the stability toward hydrolysis and reactivity toward HS-/H2S of the three triazinanes 1a-c were examined. From Table 2, the stability toward hydrolysis was 1c > 1b > 1a, with the rate of hydrolysis of 1c only ca. 1/20 that of 1a, indicating that 1c might have an advantage for H2S removal under conditions where hydrolysis would be a problem. However, a comparison of the rate constants for the reaction of HS- with 1a-c (Table 2) shows that 1c not only reacted slower in the hydrolysis reaction but also reacted slower in the reaction with HS-; 1c reacted only 1/ 30 as fast as 1a. In addition, the product from the reaction of 1c with 2 equiv of HS-, 3c, was barely soluble in water; this alone would exclude it as an industrial H2S scavenger. Experimental Section Instrumentation and Chemicals. NMR spectra were recorded on Bruker DPX 300 or 400 MHz instruments. Electron impact mass spectrometry (EIMS) spectra were obtained on a Finnigan MAT 95 XL spectrometer. Sulfide detections were performed using an Orion Sure-Flow Combination Silver/Sulfide electrode, model 9616. An Orion Sure-Flow Combination electrode, model 8172, was used for all pH measurements. All buffered solutions were made by adjusting the pH of a 0.50 M Na2HPO4 solution with aqueous HCl or solid Na2CO3. Aqueous solutions of 1b and 1c were provided by Dynea Oil Field Chemicals, Randaberg, Norway. Reaction between 1b and HS-. 1b (8.6 g, 5.0 mmol) and NaHS (0.54 g, 9.7 mmol) were dissolved in a buffered water solution at pH 10.0 (10 mL). The reaction was run at room temperature. The reaction products were identified by NMR spectroscopy, and the NMR data and their assignments are presented in Table 3. In an equimolar reaction, 1b (0.086 g, 0.50 mmol) and Na2S (0.038 g, 0.48 mmol) were dissolved in a buffered water solution at pH 10.0. 2b and ethylamine were identified as reaction products by NMR spectroscopy. The experiment with gaseous H2S was performed by bubbling the concentrated gas through a solution of 1b (0.429 g, 2.50 mmol) in water (50 mL) for 10 min. An NMR analysis of the reaction mixture revealed that only 3b and ethylamine were present. Reaction between 1c and HS-. 1c (0.65 g, 5.0 mmol) and NaHS (0.54 g, 9.7 mmol) were dissolved in
Ind. Eng. Chem. Res., Vol. 43, No. 9, 2004 1965 Table 3. NMR Data for the Products of the Reaction between 1b and NaHS 1H
13C
NMR δ (ppm)
multiplicity
1.04-1.11
t t t t q q q q s, br s s+s s
2.55 2.70 2.82 3.09 3.4 3.76 4.12 4.66
J (Hz)
7.3 7.2-7.3 7.2-7.3 7.3-7.4
NMR δ (ppm)
11.2 11.2 11.7 16.5 46.7 35.6 46.9 42.7 71.7 71.7 32.5 + 54.2 56.1
Table 4. NMR Data for the Products of the Reaction between 1c and NaHS 1H
NMR δ (ppm) 2.27 2.38 2.47 3.25 3.65 4.02
13C
multiplicity s s s s, br s s
NMR δ (ppm) 41.9 28.8 43.1 77.5 77.6 58.5
NOESY coupling to δ (ppm) 3.25 3.65, 4.02 2.27 2.55, 4.02 2.55, 3.65
compd 1c methylamine 2c 1c 2c 2c
a buffered water solution (10 mL) at pH 10.0. The reaction was run at room temperature, and the reaction mixture was analyzed by NMR spectroscopy. The NMR data and their assignments are given in Table 4. After 3 days, white crystals were filtered from the solution and washed with water. The crystals were analyzed by NMR spectroscopy, EIMS, and high-resolution mass spectrometry (HRMS).1H NMR (400 MHz, CDCl3): δ 2.69 (s, N-CH3), 4.09 (s, N-CH2-S), 4.42 (s, S-CH2-S). 13C NMR (400 MHz, CDCl3): δ 60.0 (NCH3), 37.6 (N-CH2-S), 34.4 (S-CH2-S). EIMS (70 eV)m/z (% relative intensity): 137.1 (M + 2, 1.4), 135.1 (M+, 16), 89.0 (9), 70.1 (2), 57.1 (18), 46.0 (4), 45.0 (4), 44.0 (18), 43.0 (3), 42.0 (12), 41.0 (1). HRMS (70 eV) m/z: calcd for 3c, 135.01771; found, 135.01765. In the experiment with gaseous H2S, the concentrated gas was bubbled through a solution of 1c (0.32 g, 0.25 mmol) in water (50 mL) for 23 min. Only 3c and methylamine were found in the reaction mixture, which was analyzed by 1H NMR spectroscopy. Reactions with the Oxygen Heterocyclic Compounds 4a and 4b. 4a (0.23 g, 2.5 mmol) was dissolved in a buffered solution (50 mL) at pH 10.0, and Na2S (0.19 g, 2.4 mmol) was added. The solution was left at room temperature for 3 days, after which no new reaction products were observed in the NMR spectra. In another experiment, 4a (0.23 g, 2.5 mmol) was dissolved in water (50 mL), and NaHS (0.14 g, 2.4 mmol) was added. The pH of the solution was adjusted to 7.0. No change was observed in the NMR spectra. The methyl-substituted trioxane 4b (0.33 g, 2.5 mmol) was dissolved in water (50.0 mL). NaHS (0.14 g, 2.4 mmol) was added, and the pH was adjusted to 10.0. The solution was analyzed by1H NMR spectroscopy, and
1H NMR coupling to δ (ppm)
2.70 3.09 2.82 2.55 1.04-1.11 1.04-1.11 1.04-1.11 1.04-1.11
NOESY coupling to δ (ppm) all other signals
1.04-1.11, 3.40 1.04-1.11 1.04-1.11, 3.76, 4.12 1.04-1.11, 4.66 1.04-1.11, 2.55 1.04-1.11, 2.82, 3.76 1.04-1.11, 2.82, 3.76 1.04-1.11, 3.09
compd ethylamine 3b 2b 1b 1b ethylamine 2b 3b 1b 2b 2b + 3b 3b
after 3 days, the pH was lowered to 7.0. No change was observed in the NMR spectra. Hydrolysis. The rates of hydrolysis of 1a-c were determined by dissolving the compounds in 0.5 M Na2HPO4 buffers of the desired pH in an NMR tube. The concentrations of triazinane in the solutions were 0.10 vol %. The hydrolysis was followed by recording of 1H NMR spectra with fixed intervals (1-3 min), using acetonitrile as an internal standard. From each spectrum, the concentrations of 1a-c were calculated and the rate constants kobs, k1, and k2 determined. Reactions with HS-. Investigation of the reaction rate between triazinanes 1a-c and HS- was performed by mixing known amounts of three solutions in an NMR tube: (a) Solution 1 (0.10 mL). Triazinane 1a, 1b, or 1c (0.25 mmol) and acetonitrile (internal standard, 0.0103 g, 0.25 mmol) were dissolved in D2O (5.0 mL). (b) Solution 2 (0.10 mL). NaHS (0.14 g, 0.25 mmol) was dissolved in D2O (5.0 mL). (c) Solution 3 (0.80 mL). A 0.50 M solution of Na2HPO4 was dissolved in D2O. Before mixing, the pH in all three solutions was adjusted to the desired pH. Solutions 1 and 3 were mixed first, and the time was started when solution 2 was added. 1H NMR spectra were recorded with fixed intervals and the triazinane concentrations determined for each of them. From these data, the observed rate constants k′1, k′2, and k′3 were determined. Literature Cited (1) World Health Organization. Environmental Health Criteria 18: Hydrogen Sulfide, Geneva, Switzerland, 1981. (2) Bakke, J. M.; Buhaug, J.; Riha, J. Hydrolysis of 1,3,5-Tris(2-hydroxyethyl)hexahydro-s-triazine and Its Reaction with H2S. Ind. Eng. Chem. Res. 2001, 40 (26), 6054. (3) Buhaug, J. B. Investigation of the Chemistry of Liquid H2S Scavengers. Ph.D. Thesis, Norwegian University of Science and Technology, Trondheim, Norway, 2002. (4) CRC Handbook of Chemistry and Physics, 78th ed.; Lide, D. J., Editor in Chief; CRC Press: New York, 1997.
Received for review June 17, 2003 Revised manuscript received February 9, 2004 Accepted February 18, 2004 IE030510C
