9262
J. Phys. Chem. A 2010, 114, 9262–9269
Thermal Decomposition Pathways of Hydroxylamine: Theoretical Investigation on the Initial Steps Qingsheng Wang,† Chunyang Wei,†,§ Lisa M. Pe´rez,‡ William J. Rogers,† Michael B. Hall,‡ and M. Sam Mannan*,† Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering, Texas A&M UniVersity, College Station, Texas 77843-3122, and Laboratory for Molecular Simulation, Texas A&M UniVersity, College Station, Texas 77842-3012 ReceiVed: May 6, 2010; ReVised Manuscript ReceiVed: July 9, 2010
Hydroxylamine (NH2OH) is an unstable compound at room temperature, and it has been involved in two tragic industrial incidents. Although experimental studies have been carried out to study the thermal stability of hydroxylamine, the detailed decomposition mechanism is still in debate. In this work, several density functional and ab initio methods were used in conjunction with several basis sets to investigate the initial thermal decomposition steps of hydroxylamine, including both unimolecular and bimolecular reaction pathways. The theoretical investigation shows that simple bond dissociations and unimolecular reactions are unlikely to occur. The energetically favorable initial step of decomposition pathways was determined as a bimolecular isomerization of hydroxylamine into ammonia oxide with an activation barrier of ∼25 kcal/mol at the MPW1K level of theory. Because hydroxylamine is available only in aqueous solutions, solvent effects on the initial decomposition pathways were also studied using water cluster methods and the polarizable continuum model (PCM). In water, the activation barrier of the bimolecular isomerization reaction decreases to ∼16 kcal/mol. The results indicate that the bimolecular isomerization pathway of hydroxylamine is more favorable in aqueous solutions. However, the bimolecular nature of this reaction means that more dilute aqueous solution will be more stable. I. Introduction
SCHEME 1: Proposed Initiation Step in Ref 12
Hydroxylamine plays an important role in the semiconductor, chemical, and pharmaceutical industries.1 It is used as a solvent in microchip production to remove organic and inorganic impurities from wafers and also as an important feedstock for dyes, rust inhibitors, and products such as painkillers, antibiotics, and tranquillizers.2 However, it is challenging to handle hydroxylamine free base, and nominally pure hydroxylamine is known to decompose at room temperature.3 Its chemical instability has led to two tragic incidents in the industry.4 Since then, thermal decomposition hazards of hydroxylamine/water solutions have been investigated using calorimeters.5-10 Hydroxylamine decomposition has been found to be sensitive to metals,10 metal ions,6,8 and pH of the solutions.5 The final decomposition products were analyzed as NH3, H2O, N2, N2O, and small amount of NO and H2, and the proportions depend on the experimental conditions.7,11 Previous experimental tests have provided only an overall decomposition behavior of hydroxylamine. The decomposition mechanism of hydroxylamine is still poorly understood, and the proposed mechanisms in the literature are controversial. Nitroxyl (HNO) was proposed to be an intermediate for the decomposition of hydroxylamine by Nast et al.,12 and a reaction (Scheme 1) was developed in which the decomposition is controlled by disproportionation of hydroxylamine to water, ammonia, and nitroxyl. The presence of nitroxyl as an inter* To whom correspondence should be addressed. E-mail: mannan@ tamu.edu. Tel: 1-979-862-3985. Fax: 1-979-845-6446. † Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering. ‡ Laboratory for Molecular Simulation. § Currently an employee of BASF Corporation.
mediate was verified by the appearance of the violet tricyanonitrosonickel(II) on the addition of tetracyanonickel(II).12 However, Lunak et al.13 disagreed with this conclusion and demonstrated that the formation of tricyanonitrosonickel(II) was not due to nitroxyl but to formation and rearrangement of a tricyano-hydroxylammonium nickel complex, resulting from replacement of a cyanide group in tetracyanonickel(II) by a molecule of hydroxylamine. Holzapfel14 studied the kinetics of hydroxylamine decomposition in strong alkaline solutions, and it was assumed that OH-NH-OH and OH-NH-NH-OH were formed during the decomposition. In a previous study, it was found that acid or base can initiate different decomposition pathways of hydroxylamine.5 Hughes et al.15,16 reported the oxidation of hydroxylamine in alkaline solutions, where nitrite, peroxonitrite, and hydrogen peroxide were detected as intermediates in significant quantities together with some nitrate. In addition, the catalytic effects of metal ions on the decomposition of hydroxylamine were studied under different conditions.13,17 In this study, initial decomposition pathways of hydroxylamine were investigated in the gas phase and in aqueous solutions. As part of our ongoing interest in evaluating chemical reactivity for different reactive systems,18 here we sought to elucidate the initial decomposition steps of hydroxylamine using quantum mechanical calculations. Several possible initial decomposition pathways were proposed, and of those, the most favorable pathway was determined. Density functional theory (DFT) and ab initio wave function theory (WFT) were employed to yield a quantitative description of the thermal decomposition
10.1021/jp104144x 2010 American Chemical Society Published on Web 08/02/2010
Thermal Decomposition Pathways of Hydroxylamine
J. Phys. Chem. A, Vol. 114, No. 34, 2010 9263
Figure 1. Molecular structures of hydroxylamine, transition states, and products involved in unimolecular decomposition pathways (pathways 1-4) at the level of B3LYP/BSI. P5 has the similar structure as P4 with bond length 1.06 Å.
mechanism. The results will provide a better understanding of the stability of hydroxylamine and guidance on the design of effective inhibitors to control the decomposition behavior. II. Computational Details Density functional and ab initio calculations were performed using the Gaussian 03 suite of programs.19 All geometries of reactants, products, various intermediates, and transition states were fully optimized in the gas phase using several density functional methods. The Becke3-Lee-Yang-Parr (B3LYP)20,21 and modified Perdew-Wang one-parameter model for kinetics (MPW1K)22 were used with Dunning’s correlation consistent polarized valence double-ζ basis set (cc-pVDZ)23 and a Poplestyle basis set24 including diffuse25 and polarization26 functions (6-31+G(d,p)), respectively. Previous theoretical work27,28 shows that the MPW1K method can provide more accurate energy barriers than B3LYP. The Becke88-Becke95 one-parameter model for kinetics (BB1K) is another hybrid Hartree-Fock density functional model developed by Truhlar and coworkers and has been shown to give excellent saddle point geometries and barrier heights.29 To validate the DFT methods and achieve more accurate energies, single-point energies were calculated with coupled cluster singles and doubles with triples correction CCSD(T)30 using a 6-311+G(3df,2p) basis set based on the optimized geometries obtained from the Møller-Plesset secondorder perturbation theory (MP2)31 with the same basis set. A composite complete basis set (CBS-Q)32 method was also employed to calculate the decomposition pathways of unimolecular reactions. The basis sets used for geometry optimizations and energy calculations are denoted as follows
cc-pVDZ f BSI 6-31+G(d,p) f BSII 6-311+G(3df,2p) f BSIII Frequency calculations at the B3LYP, MPW1K, BB1K, and MP2 levels of theory were performed to obtain zero-point energies and frequencies for all species in the reaction pathways.
The calculated structures were characterized as either a local minimum with no imaginary frequency or transition states with only one imaginary frequency. Some intrinsic reaction path (IRC) calculations33 were conducted to follow the reaction paths and validate that the transition states connect two minima of interest on the potential energy surface. Solvent effects on the initial decomposition pathway were studied using both cluster methods and the polarizable continuum model (PCM).34 Single-point energy calculations using the optimized gas-phase geometry were conducted with PCM because of geometry optimization convergence problems. Relative energies are reported in kilocalories per mole. Representations of the reactants, products, intermediates, and transition states, shown in Figures 1 and 2, were created using the JIMP 2 software program.35 III. Results and Discussion A. Molecular Structures of Hydroxylamine. Experimental data on the molecular geometry for hydroxylamine are available,36 and some theoretical calculations37,38 on the heat of formation, equilibrium geometries, rotational barriers, and vibrational analysis have also been reported. Hydroxylamine (HA, 1) is in Cs symmetry and has two conformations (“trans” and “cis”), as shown in Figure 1 (1 and 2, respectively). The optimized geometries of 1 and 2 at the density functional, MP2, and CCSD levels of theory are compared with experimental and previous theoretical results in Table 1. B3LYP/BSI provides relatively accurate bond lengths within 0.01 Å of the experimental values but a poor H-N-H bond angle (103.9°) compared with the experimental value (107.1°). MPW1K/BSII and BB1K/BSII results are very similar to each other. The calculated N-O bond lengths are both 0.04 Å shorter than the experimental value, but the calculated H-N-H bond angles (106.8 and 106.6°, respectively) are closer to the experimental value (107.1°) than the B3LYP/BSI value (103.9°). From the deviations in Table 1, CCSD/BSII provides slightly better results than MP2/BSIII, especially for the bond lengths. (Deviation is less than (0.004 Å for CCSD/BSII compared with the experimental value.) To obtain highly accurate results for
9264
J. Phys. Chem. A, Vol. 114, No. 34, 2010
Wang et al.
Figure 2. Molecular structures of hydroxylamine, transition states, and products involved in bimolecular decomposition pathways 5 and 6 at the level of B3LYP/BSI.
TABLE 1: Comparison of Optimized Geometries of trans- and cis-NH2OH (1 and 2, respectively) at Different Levels of Theory with Experimental Data (Bond Lengths Are in angstroms and Angles Are in degrees)a
RN-H RN-O RO-H ∠HNH′ ∠HNO ∠HON RN-H RN-O RO-H ∠HNH′ ∠HNO ∠HON
experimentalb
B3LYP/BSI
MPW1K/BSII
CCSD/BSII
MP2/6-311++G(d,p)c
MP2/BSIII
1.016 1.453 0.962 107.1 103.3 101.4
1.027 1.445 0.969 103.9 103.3 101.7
1.010 1.408 0.954 106.8 105.0 103.7
trans-NH2OH (1) 1.011 1.411 0.955 106.6 104.8 103.6
1.017 1.449 0.964 106.1 103.5 102.4
1.017 1.434 0.960 105.9 104.3 101.8
1.014 1.434 0.960 105.9 104.0 102.3
1.028 1.428 0.974 105.7 107.2 107.9
1.010 1.397 0.958 109.3 108.7 109.2
cis-NH2OH (2) 1.010 1.400 0.959 109.2 108.5 109.2
1.018 1.437 0.968 108.1 107.0 108.1
1.017 1.423 0.964 107.9 107.9 107.5
1.015 1.424 0.964 107.7 107.5 107.7
BB1K/BSII
a H and H′ are symmetry equivalent. b Tsunekawa, S. J. Phys. Soc. Jpn. 1972, 33, 167. In this work, the molecular structure of trans-NH2OH (1) was determined by a least-squares method using 13 moments of inertia based on the microwave spectrum of NH2OH observed at room temperature with a conventional 100 kHz Stark-modulated spectrometer. c Boulet, P.; et al. Chem. Phys. 1999, 244, 163.
the geometry of 1, the CCSD level of theory with a moderately sized basis set is necessary. The rotational barrier for the conversion from trans conformation (1) to cis conformation (2) is not available from experiment but can be obtained from previous theoretical studies.38,39 The geometry of the transition state (TS1, Figure 1) and the calculated rotational barriers at different levels of theory are presented in Table 2. In previous work, it was shown that the calculated rotation barrier and energy of reaction decreases with an increase in the size of the basis sets and the correlation corrections.38 In our work, the highest level of theory used is the CCSD(T)/BSIII//CCSD/BSII level of theory with the inclusion of zero-point energy corrections at the CCSD/ BSII level. The trans conformation is found to be 4.10 kcal/ mol more stable than the cis conformation, and the trans-to-cis rotation barrier at the CCSD(T)/BSIII//CCSD/BSII level is 6.19 kcal/mol. CCSD(T)/BSIII single point energies (SPEs) were also calculated at the B3LYP/BSI, MPW1K/BSII, BB1K/BSII, and MP2/BSIII optimized geometries, and in all cases, the ∆Eo,rxn was within 0.08 kcal/mol of the CCSD(T)/BSIII//CCSD/BSII result and within 0.12 kcal/mol for the ∆Eo‡. The lower levels of theory have calculated barriers that are higher than the CCSD(T)/BSIII SPE, consistent with previous theoretical work.38 The lower level calculations also had a much larger
range in their values for the barrier (1.42 kcal/mol) compared with the CCSD(T)/BSIII SPE (0.12 kcal/mol). Because the results from the MPW1K and BB1K calculations are so close, only MPW1K was employed for the pathway analysis. The rotation barrier and relative energy at the B3LYP/BSI level (6.58 and 4.44 kcal/mol, respectively) are close to the results at the CCSD(T)/BSIII//CCSD/BSII level (6.19 and 4.10 kcal/mol, respectively). The B3LYP/BSI geometry for the transition state is in good agreement with the CCSD/BSII results, except for the H-N-H bond angle (102.7°) compared with that of CCSD/BSII (105.1°). The CCSD(T)/BSIII//MP2/BSIII calculation yielded a slightly lower barrier (6.29 kcal/mol) than the MP2/BSIII calculation (6.46 kcal/mol) because of the higher correlation corrections at the CCSD(T) level, which is consistent with the previous work.38 B. Bond Dissociation of Hydroxylamine. Bond strengths and bond dissociation energies (BDEs) are fundamental to chemical reactions, and they can provide deep insight into the stability of chemical compounds. Experimental data on the BDE of 1 are not available, but O-H bond strengths of unhindered dialkylhydroxylamines were determined to be in the range of 72-74 kcal/mol using calorimetric measurements.40 The calculated BDEs of 1 are shown in Table 3. The N-O bond is calculated to be the weakest bond with a BDE at 0 K
Thermal Decomposition Pathways of Hydroxylamine
J. Phys. Chem. A, Vol. 114, No. 34, 2010 9265
TABLE 2: Calculated Relative Energies, ∆Eo, Barrier Heights, ∆Eo‡ (in kilocalories per mole), and Optimized Geometry of Transition State (TS1) between trans and cis Conformations of NH2OH (1 and 2, respectively) at Different Levels of Theory (Bond Lengths in angstroms and angles in degrees)a molecule
B3LYP/BSI
MPW1K/BSII
BB1K/BSII
CCSD/BSII
MP2/BSIII
trans (1) TS1 cis (2)
0.00 6.58 4.44
0.00 7.88 5.33
0.00 7.82 5.34
0.00 7.59 5.56
0.00 6.46 4.16
1.012 1.012 1.425 0.955 105.5 103.4 108.1 107.8
1.019 1.019 1.461 0.964 105.1 102.4 106.9 106.7
1.016 1.017 1.448 0.961 104.3 102.6 107.1 106.2
Geometry of TS1 1.028 1.028 1.457 0.969 102.7 102.3 106.5 105.9
RN-H RN-H′ RN-O RO-H ∠HNH′ ∠HNO ∠H’NO ∠HON
a
1.011 1.011 1.422 0.954 105.6 103.5 108.2 107.8
molecule
CCSD(T)/BSIII// B3LYP/BSI
CCSD(T)/BSIII// MPW1K/BSII
CCSD(T)/BSIII// BB1K/BSII
CCSD(T)/BSIII// CCSD/BSII
CCSD(T)/BSIII// MP2/BSIII
trans (1) TS1 cis (2)
0.00 6.31 4.13
0.00 6.26 4.17
0.00 6.26 4.18
0.00 6.19 4.10
0.00 6.29 4.14
Zero-point energy (ZPE) included.
TABLE 3: Calculated Bond Dissociation Energies and Enthalpies (in kilocalories per mole) of 1 at Various Levels of Theory B3LYP/BSI
MPW1K/BSII
MP2/BSIII
CCSD/BSII
dissociated bond
E0a
H298b
E0
H298
E0
H298
E0
H298
N-O O-H N-H
60.20 67.93 76.60
62.02 67.84 76.43
54.73 69.81 79.45
56.61 69.90 79.33
67.22 75.45 83.09
69.06 75.52 82.94
52.34 72.59 78.45
54.20 72.54 78.32
CCSD(T)/BSIII// B3LYP/BSI
CCSD(T)/BSIII// MPW1K/BSII
CCSD(T)/BSIII// MP2/BSIII
CCSD(T)/BSIII// CCSD/BSII
dissociated bond
E0a
H298b
E0
H298
E0
H298
E0
H298
N-O O-H N-H
59.57 74.05 81.20
61.38 73.96 81.02
58.60 73.07 80.59
60.48 73.16 80.47
59.56 74.05 81.14
61.40 74.13 80.99
59.16 73.96 80.81
61.02 73.91 80.68
a E0 ) ∑electronic and zero point energies of products - ∑electronic and zero point energies of reactants. b H298 ) ∑electronic and thermal correction to enthalpy of products - ∑electronic and thermal correction to enthalpy of reactants.
of 59.16 kcal/mol at the CCSD(T)/BSIII//CCSD/BSII level. The lower level calculations have a large range for the N-O BDE at 0 K with the B3LYP/BSI result (60.20 kcal/mol) being closest to the CCSD(T)/BSIII//CCSD/BSII result (59.16 kcal/mol). The calculated N-O BDEs for all CCSD(T)/BSIII single point calculations at the lower level optimized geometries are within 0.56 kcal/mol of the CCSD(T)/BSIII/CCSD/BSII N-O BDE. The O-H and N-H CCSD(T)/BSIII//CCSD/BSII BDEs (73.96 and 80.81 kcal/mol, respectively) are more than 14 kcal/mol higher in energy than the N-O BDE. The CCSD(T)//BSIII SPE calculated O-H BDEs are within the experimental range (72-74 kcal/mol) for the O-H bond strengths of unhindered dialkylhydroxylamines, indicating that the level of theory is appropriate for this type of system.40 The relatively high energy that is required to break the weakest bond (N-O) suggests that simple bond dissociation reactions of 1 are unlikely to be significant at room temperature. These simple bond breaking reactions cannot explain the highly reactive nature of 1 on its initiation. Therefore, other decomposition reaction pathways must be explored to evaluate its chemical reactivity. C. Unimolecular Pathways. As represented in Scheme 2, six initial steps of decomposition pathways of hydroxylamine were proposed, including unimolecular and bimolecular reac-
SCHEME 2: Proposed Initial Decomposition Pathways of Hydroxylaminea
a
NH2OH without specification is trans-hydroxylamine.
tions. The optimized structures of the species involved in the pathways at the B3LYP/BSI level of theory are shown in Figures 1 and 2, and their free energies are shown in Tables 4 and 5. Pathway 1 is the isomerization of cis-hydroxylamine (2) into trans-hydroxylamine (1), which has been described in the previous section. Pathway 2 involves the isomerization of 1 into ammonia oxide (NH3O, 3, Figure 1) via a 1,2-hydrogen shift from the O to the N atom. The N-O bond length of NH3O (1.34 Å) is shorter than that of trans-hydroxylamine (1.45 Å)
9266
J. Phys. Chem. A, Vol. 114, No. 34, 2010
Wang et al.
TABLE 4: Relative Free Energies (∆Go in kilocalories per mole at 298 K) for Species Involved in Unimolecular Decomposition Pathways at Various Levels of Theory structure
B3LYP/BSI
MPW1K/BSII
1 TS2 3
0.00 50.45 28.56
0.00 51.19 24.95
3 TS3 barrier P1+P2
28.56 67.59 39.03 24.03
24.95 85.03 60.08 35.29
1 TS4 4 P3+P4 P3+P5
0.00 62.92 62.94 85.63 31.77
0.00 68.00 64.16 82.16 24.01
MP2/BSIII
CCSD(T)/BSIII// MP2/BSIII
CBS-Q
0.00 50.01 25.24
0.00 50.22 25.45
0.00 48.53 20.43
25.24 76.60 51.36 27.19
25.45 74.04 48.59 25.05
20.43 72.84 52.41 26.07
0.00 68.08 64.59 83.81 29.86
0.00 64.40 61.47 70.78 25.99
0.00 62.22 59.23 70.52 23.42
Pathway 2
Pathway 3
Pathway 4
TABLE 5: Relative Free Energies (∆Go in kilocalories per mole at 298 K) for Species Involved in Bimolecular Decomposition Pathways at Various Levels of Theory
a
structure
B3LYP/BSI
1+1 (5) TS5 1+3 (6)
0.00 (0.00)a 22.44 (22.34) 28.58 (16.10)
1+1 (7) TS6 2+3 (8)
0.00 (0.00) 31.16 (29.74) 33.06 (24.69)
CCSD(T)/BSII// B3LYP/BSI
CCSD(T)/BSII// MPW1K/BSII
Pathway 5 0.00 (0.00) 27.93 (24.92) 24.59 (14.89)
(0.00) (26.17) (15.47)
(0.00) (26.94) (16.36)
Pathway 6 0.00 (0.00) 34.76 (31.73) 30.28 (23.20)
(0.00) (35.74) (25.04)
(0.00) (36.05) (23.20)
MPW1K/BSII
Species in parentheses are reactant wells, transition-state wells, or product wells including the hydrogen bond effect.
because NH3O has more zwitterionic character (H3N+-O-). Therefore, the electrostatic force between the N and the O atoms in 3 brings the two atoms closer and shortens the N-O bond length, whereas in 1, the repulsion of the lone pairs pushes the N and O atoms away. The free energy of 3 is 25.45 kcal/mol higher than that of 1 at the CCSD(T)/BSIII//MP2/BSIII level (Table 4). The optimized transition-state structure (TS2) for pathway 2 is also shown in Figure 1. This is a central transition state, as evidenced by the O-N-H angle of 60° that is about half of the corresponding angle of ammonia oxide (115°). The shifting hydrogen atom is between the N and O atoms with an N-H bond length of 1.09 Å and an O-H bond length of 1.36 Å. A frequency analysis at the B3LYP/BSI level reveals only one imaginary frequency (1442i cm-1), which corresponds to the normal mode of hydrogen shifting from the O to the N atom. The free energies of the species involved in this pathway are presented in Table 4, and the free energy of activation for this reaction is 50.22 kcal/mol at the CCSD(T)/BSIII//MP2/BSIII. The ∆G‡ values at the B3LYP/BSI, MPW1K/BSII, and MP2/ BSIII levels of theory are within 0.97 kcal/mol of the CCSD(T)/ BSIII//MP2/BSII result. Pathway 3 results in hydrogen elimination of NH3O to form hydrogen gas and HNO. The hydrogen elimination step involves the simultaneous breaking of two N-H σ bonds and the formation of an H-H σ bond and an O-N π bond. The optimized geometry of the transition state (TS3) for this hydrogen elimination step is represented in Figure 1 with two short bonds, H-H (1.27 Å) and N-O (1.23 Å). The rest of the molecule becomes more planar with the formation of the N-O double bond. This is a late transition state based on the N-O
bond length (1.23 Å), which is closer to the product, nitroxyl (1.20 Å). A frequency calculation on TS3 had only one imaginary frequency (1141i cm-1) at the B3LYP/BSI level, whose normal mode is consistent with the forming H-H and breaking H-O bonds. Table 4 lists the calculated free-energy activation barriers of the hydrogen elimination reaction. The activation barrier at MP2/BSIII is 51.36 kcal/mol and is lower than the MPW1K/BSII result (60.08 kcal/mol). The CCSD(T)/ BSIII//MP2/BSIII calculation results in a barrier that is 2.56 kcal/mol lower than the MP2/BSIII barrier. The activation barrier at the CBS-Q level is 52.41 kcal/mol, which is close to the result of CCSD(T)/BSIII//MP2/BSIII. Pathway 4 involves a hydrogen transfer from the N to the O atom, forming water and an NH radical that is in either its singlet or triplet state. In this work, both the singlet and triplet pathways are considered. The optimized geometries for singlet and triplet NH (P4 and P5, respectively) are shown in Figure 1, and their corresponding energies are given in Table 4. Triplet NH is 47.10 kcal/mol more stable than the singlet state at the CBS-Q level and 44.79 kcal/mol more stable at the CCSD(T)/BSIII//MP2/ BSIII level. A singlet transition state (TS4) was located with a calculated barrier of 62.22 kcal/mol above 1. TS4 is a late transition state, because the N-O bond is very long (1.77 Å) and essentially broken, and H2O and NH are almost fully formed. A frequency calculation on TS4 had one imaginary frequency (543i cm-1) at the B3LYP/BSI level, whose normal mode is consistent with the hydrogen atom transferring from the N to the O atom. The product of the singlet pathway is a singlet imidogen-water complex (4), the energy of which is slightly lower than that of TS4.
Thermal Decomposition Pathways of Hydroxylamine
Figure 3. Unimolecular decomposition pathways of hydroxylamine. Free energies were calculated at the CCSD(T)/BSIII//MP2/BSIII level.
All of the unimolecular decomposition pathways that were considered so far have high activation barriers, as shown in Figure 3. The lowest barrier on the potential energy surfaces is pathway 2, and the calculated barrier for isomerization of hydroxylamine into ammonia oxide is ∼50 kcal/mol. This result is consistent with the experimental and theoretical results proposed by Bro¨nstrup et al.,41 although they reported some other hydrogen shift pathways with even higher activation energies. Therefore, a unimolecular route is also unlikely to be the major decomposition mechanism. D. Bimolecular Pathways. Bimolecular decomposition pathways were explored in search of lower barriers on the potential energy surface. The bimolecular path,12 as shown in Scheme 1, was proposed to initiate the decomposition, but a transition state could not be located for this reaction, implying that it may not be an elementary reaction. This bimolecular pathway may involve multiple steps, such as
2NH2OH f NH3O + NH2OH NH3O f HNO + H2 NH3O f NH2O + H Then, radicals NH2O and H react with NH2OH to propagate the chain reactions. In pathways 5 and 6, two H atoms shift between two hydroxylamine molecules, forming ammonia oxide. Two transition states, TS5 and TS6, optimized at the B3LYP level of theory (with only one imaginary frequency, 1166i cm-1 for TS5 and 1430i cm-1 for TS6), are shown in Figure 2. The transition state in pathway 5 (TS5) involves the H atom in the OH group transferring to the N atom of another hydroxylamine molecule; meanwhile, the electron repulsion weakens the O-H bond in TS5, forming ammonia oxide. The H atom in the weakened O-H bond transfers to another O atom, forming a new hydroxylamine molecule. The difference between pathways 5 and 6 is the conformation of the formed hydroxylamine molecule, and the H-contributing groups in pathway 6 are NH2 and OH, instead of two OH groups in pathway 5. As shown in Figure 4 and Table 5, the activation barrier of pathway 5 is 22.44 kcal/mol at the B3LYP/BSI level, and 26.17 kcal/mol at the CCSD(T)/BSII//B3LYP/BSI level, which is significantly lower than that of the unimolecular reaction pathway 2 (50 kcal/mol). It should be noted that the barrier for
J. Phys. Chem. A, Vol. 114, No. 34, 2010 9267
Figure 4. Bimolecular decomposition pathways of hydroxylamine. Free energies were calculated at the CCSD(T)/BSII//B3LYP/BSI level.
TABLE 6: Solvent Effects on Free Energies (in kilocalories per mole at 298 K) for Species Involved in Pathway 5 at the MPW1K/BSII Level of Theory structurea
5
TS5
6
no solvent NH2OH H2O 2H2O 3H2O PCM (H2O)
0.00 0.00 0.00 0.00 0.00 0.00
24.92 20.06 21.48 18.62 16.10 16.81
14.89 10.14 13.43 6.99 4.23 4.32
a For optimized structures of 5, TS5, and 6 hydrogen-bonded with one hydroxylamine molecule and one, two, or three water molecules, see the Supporting Information.
the reverse reaction (NH3O f NH2OH) is very low in the gas phase, which may contribute to the difficulty of detecting ammonia oxide by experimental methods because ammonia oxide might easily isomerize into hydroxylamine.42 E. Solvent Effect. It is important to consider solvent effects on the hydroxylamine decomposition pathways because hydroxylamine is manufactured and used in aqueous solutions. The gas-phase reaction pathway analysis shows that pathway 5 is most likely to occur because of its lowest activation barrier of 26.17 kcal/mol, as shown in Figure 4. This bimolecular reaction involves two hydrogen shifts and produces ammonia oxide, which is less stable than hydroxylamine. Clusters of hydroxylamine and water were used to simulate solvation effects. All of the cluster structures in pathway 5 were fully optimized at the MPW1K/BSII level (Supporting Information), and the free energies are shown in Table 6. The inclusion of solvent molecules or hydroxylamine molecule reduces the activation barriers. For example, one water molecule can stabilize the transition state (21.48 kcal/mol) as well as one hydroxylamine molecule (20.06 kcal/mol). With two water molecules (18.62 kcal/mol) included, the activation barrier becomes lower than the ones with one (21.48 kcal/mol). With three water molecules, the activation barrier is even lower (16.10 kcal/mol), which means the activation energy decreases with an increasing number of water molecules. From Table 6, the free energy of reaction also decreases with an increasing number of water molecules. With three water molecules as solvent, the free energy of reaction becomes 4.23 kcal/mol. The results show that the solvent effect of water is evident in the small clusters containing only a few water molecules.
9268
J. Phys. Chem. A, Vol. 114, No. 34, 2010
Continuum models describing solute-solvent interactions were also used to study the initial decomposition pathway. SPE calculations were conducted using the PCM method at the MPW1K/BSII level based on the optimized structures at the same level of theory. The free energy of activation at 298 K including the thermal correction for TS5 is 16.81 kcal/ mol using the PCM method, which is very close to the result of 16.10 kcal/mol using a cluster containing three water molecules. However, the solvent effect of a small water cluster converges more quickly than that of bulk water. IV. Conclusions The initial hydroxylamine decomposition pathways were investigated using density functional and ab initio methods. In this work, both unimolecular and bimolecular reactions were analyzed to locate a pathway with a low activation barrier. Simple bond dissociations and unimolecular hydrogen shift or elimination reactions require high activation energies. Two bimolecular pathways were found to have lower activation barriers than the unimolecular reactions: Pathways 5 and 6 consist of two hydrogen shifts that can facilitate the isomerization of hydroxylamine to ammonia oxide. Accurate calculations show that the lowest gas-phase activation barrier for the bimolecular isomerization step is 26.17 kcal/mol at the CCSD(T)/BSII//B3LYP/BSI level. Although accurate gas-phase decomposition pathway analysis can provide a good reference, solvent effects on the potential energy surface were investigated using cluster and continuum methods. Water solvent can stabilize the transition states and lower the activation barriers and the free energies of reactions. In aqueous solutions, the bimolecular isomerization step is the most favorable pathway with an energy barrier of only about 16 kcal/mol. The theoretical study shows the potential formation of ammonia oxide in solutions. The N-O bond dissociation enthalpy of NH3O (NH3O f NH3 + 3 O) at 298 K was calculated to be 35 kcal/mol at the G2 level of theory.43 Research on the detection of NH3O and possible decomposition reactions induced by NH3O will be conducted in the future. Acknowledgment. This research was supported by Mary Kay O’Connor Process Safety Center (MKOPSC). We would like to thank the supercomputing facility at Texas A&M University for computer time and software. We also thank Dr. Maria Papadaki at the University of Leeds for helpful discussions. Supporting Information Available: Structures of 5, TS5, and 6 hydrogen-bonded with one hydroxylamine molecule and one, two, or three water molecules in pathway 5. All structures are fully optimized at the MPW1K/BSII level to study the solvent effects for species involved in pathway 5. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Jencks, W. P. Catalysis in Chemistry and Enzymology; McGrawHill, New York, 1969. (2) Hydroxylamine and Its Salts, Manufacturing Chemist and Aerosol News 1964, 35, 29. (3) The Explosion at Concept Sciences: Hazards of Hydroxylamine; U.S. Chemical Safety and Hazard Investigation Board: Washington, DC, 2002. (4) (a) Reisch, M. Chem. Eng. News 1999, 77, 11. (b) Business Concentrates, Chem. Eng. News 2000, 78, 15.
Wang et al. (5) Wei, C.; Saraf, S. R.; Rogers, W. J.; Mannan, M. S. Thermochim. Acta 2004, 421, 1. (6) Iwata, Y.; Koseki, H. J. Hazard. Mater. 2003, 104, 39. (7) (a) Papadaki, M. I.; Pontiki, E.; Liu, L.; Rogers, W. J.; Mannan, M. S. J. Chem. Eng. Data 2009, 54, 2616. (b) Cisneros, L. O.; Wu, X.; Rogers, W. J.; Mannan, M. S.; Park, J.; North, S. W. Process Saf. EnViron. Prot. 2003, 81, 121. (8) Cisneros, L. O.; Rogers, W. J.; Mannan, M. S.; Li, X.; Koseki, H. J. Chem. Eng. Data 2003, 48, 1164. (9) Wang, Q.; Rogers, W. J.; Mannan, M. S. J. Therm. Anal. Calorim. 2009, 98, 225. (10) (a) Cisneros, L. O.; Rogers, W. J.; Mannan, M. S. J. Hazard. Mater. 2001, 82, 13. (b) Cisneros, L. O.; Rogers, W. J.; Mannan, M. S. J. Hazard. Mater. 2002, 95, 13. (11) Hofmann, K. A.; Kroll, F. Ber. 1924, 57B, 937. (12) Nast, R.; Foppl, I. Z. Anorg. Allg. Chem. 1950, 263, 310. (13) Lunak, S.; Veprek-Siska, J. Collect. Czech. Chem. Commun. 1972, 39, 391. (14) Holzapfel, H. Wiss. Z. - Karl-Marx-UniV. Leipzig, Math.-Naturwiss. Reihe 1960, 1959-1960, 17. (15) Hughes, M. N.; Nicklin, H. G. Chem. Ind. 1967, 52, 2176. (16) Hughes, M. N.; Nicklin, H. G. J. Chem. Soc. A 1971, 1, 164. (17) (a) Alluisetti, G. E.; Almaraz, A. E.; Amorebieta, V. T.; Doctorovich, F.; Olabe, J. A. J. Am. Chem. Soc. 2004, 126, 13432. (b) Bengtsson, G.; Fronaus, S.; Bengtsson-Kloo, L. J. Chem. Soc., Dalton Trans. 2002, 2548. (18) (a) Wang, Q.; Ng, D.; Mannan, M. S. Ind. Eng. Chem. Res. 2009, 48, 8745. (b) Liu, L.; Wei, C.; Guo, Y.; Rogers, W. J.; Mannan, M. S. J. Hazard. Mater. 2009, 162, 1217. (c) Wang, Q.; Zhang, Y.; Rogers, W. J.; Mannan, M. S. J. Hazard. Mater. 2009, 165, 141. (19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision B.05; Gaussian, Inc.: Pittsburgh, PA, 2003. (20) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (21) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (22) Lynch, B. J.; Fast, P. L.; Harris, M.; Truhlar, D. G. J. Phys. Chem. A 2000, 104, 4811. (23) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007. (24) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257. (25) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W. J. Comput. Chem. 1983, 4, 294. (26) Frisch, M. J.; Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80, 3265. (27) Li, Q. S.; Xu, X. D.; Zhang, S. Chem. Phys. Lett. 2004, 384, 20. (28) Coote, M. L. J. Phys. Chem. A 2004, 108, 3865. (29) Zhao, Y.; Lynch, B. J.; Truhlar, D. G. J. Phys. Chem. A 2004, 108, 2715. (30) Pople, J. A.; Head-Gordon, M.; Raghavachari, K. J. Chem. Phys. 1987, 87, 5968. (31) Head-Gordon, M.; Pople, J. A.; Frisch, M. J. Chem. Phys. Lett. 1988, 153, 503. (32) Petersson, G. A.; Bennett, A.; Tensfeldt, T. G.; Al-Laham, M. A.; Shirley, W. A.; Mantzaris, J. J. Chem. Phys. 1988, 89, 2193. (33) (a) Gonzalez, C.; Schlegel, H. B. J. Chem. Phys. 1989, 90, 2154. (b) Gonzalez, C.; Schlegel, H. B. J. Phys. Chem. 1990, 94, 5523. (34) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem. 2003, 24, 669. (35) (a) Hall, M. B.; Fenske, R. F. Inorg. Chem. 1972, 11, 768. (b) Manson, J.; Webster, C. E.; Pe´rez, L. M.; Hall, M. B. http://www.chem. tamu.edu/rgroup/hall/laboratories/Programs.html. (36) Tsunekawa, S. J. Phys. Soc. Jpn. 1972, 33, 167. (37) (a) Saraf, S. R.; Rogers, W. J.; Mannan, M. S.; Hall, M. B.; Thomson, L. M. J. Phys. Chem. A 2003, 107, 1077. (b) Tyrrell, J.; LewisBevan, W.; Kristiansen, D. J. Phys. Chem. 1993, 97, 12768. (38) Chung-Phillips, A.; Jebber, K. A. J. Chem. Phys. 1995, 102, 7080. (39) Boulet, P.; Gilardoni, F.; Weber, J.; Chermette, H.; Ellinger, Y. Chem. Phys. 1999, 244, 163.
Thermal Decomposition Pathways of Hydroxylamine (40) Mahoney, L. R.; Mendenhall, G. D.; Ingold, K. U. J. Am. Chem. Soc. 1973, 95, 8610. (41) Bro¨nstrup, M.; Schro¨der, D.; Kretzschmar, I.; Schalley, C. A.; Schwarz, H. Eur. J. Inorg. Chem. 1998, 1529. (42) (a) Kirby, A. J.; Davies, J. E.; Brandao, T. A. S.; da Silva, P. F.; Rocha, W. R.; Nome, F. J. Am. Chem. Soc. 2006, 128, 12374. (b) Kirby,
J. Phys. Chem. A, Vol. 114, No. 34, 2010 9269 A. J.; Davies, J. E.; Fox, D. J.; Hodgson, D. R. W.; Goeta, A. E.; Lima, M. F.; Priebe, J. P.; Santaballa, J. A.; Nome, F. Chem. Commun. 2010, 46, 1302. (43) Curtiss, L. A.; Raghavachari, K.; Trucks, G. W.; Pople, J. A. J. Chem. Phys. 1991, 94, 7221.
JP104144X
