A computational study on the decomposition of NH4ClO4: Comparison of the gas-phase and condensed-phase results

(1)

A computational study on the decomposition of NH

4 ClO

4 :

Comparison of the gas-phase and condensed-phase results

R.S. Zhu, M.C. Lin

*

Department of Chemistry, Emory University, M.C. Lin’s Chemical Kinetics Research Atlanta, GA 30322, USA Received 1 August 2006; in ﬁnal form 3 October 2006

Available online 10 October 2006

Abstract

The decomposition of NH4ClO4(AP) in both gaseous and condensed phases has been studied at the CCSD(T)/6-311 + G(3df,2p)//

B3LYP/6-311 + G(3df,2p) level. Formation of NH3+ HClO3is the main channel in the gas phase with 14.2 kcal/mol enthalpy change.

In solution, the enthalpy change for this channel is 29.5 kcal/mol, which is close to the sublimation activation energy,30 kcal/mol in crystalline AP. Formation of H3N–HO +ClO3has higher enthalpy changes, 56.3 kcal/mol and 77.1 kcal/mol in the gas phase and in

1. Introduction

Ammonium perchlorate (AP), NH4ClO4, has been

widely employed as an oxidizer in composite propellants for rocket propulsion because it is cheap and contains a large amount of oxygen which, in combustion, is converted entirely into stable gaseous reaction products[1–3]. AP is a crystalline material composed of a network of ammonium cations (NHþ₄Þ and perchlorate anions (ClO₄Þ [4]. Each perchlorate ion is surrounded by seven ammonium ions with the O–N distances ranging from 2.9 to 3.25 A˚ ; simi-larly, each ammonium ion is surrounded by seven perchlo-rate ions with H–O hydrogen-bond lengths ranging from 1.891 to 2.077 A˚ . The ClO₄ ion has an essentially ideal tet-rahedral structure with Cl–O bond length of 1.44 ± 0.01 A˚ and the bond angle of 109.5 ± 1° [4]; the NHþ₄ ion has a nearly undistorted tetrahedral structure with the average N–H bond length of 1.03 A˚ .

Despite extensive studies on the mechanism for the decomposition and combustion of AP experimentally over

the past few decades, as summarized in detail in the over-views by Jacobs and Whitehead[3]and Tanaka and Beck-stead [5], many questions still remain regarding the key controlling initiation processes within and/or near the burning surface. Tanaka and Beckstead reviewed various assumed initiation reactions; they also put forth a three-phase combustion model (unpublished) to account for the observation of Brill et al.[6]acquired from a time-resolved FTIR measurement at high pressure and that of Ermolin et al.[7]for species detected by mass-spectrometry near a burning AP surface. The mechanism includes the following initiation processes in the solid, liquid and gas phases:

APðsÞ! NH3þ HClO4

APðsÞ! APðlÞ ! NH3þ HClO4

APðlÞ! H2Oþ O2þ HCl þ HNO

APðlÞ! 2H2Oþ Cl þ NO2

APðlÞ! NH3þ OH þ ClO3

To fully substantiate the three-phase combustion model requires a systematic study of the chemistry occurring in each of the three phases by high-level quantum-chemical calculations for the key processes involved. For the decom-position of AP in solid (crystal), it was believed that the ini-tial step in the low-temperature range (<300° C) was one of

*

Corresponding author. Present address. NSC Distinguished Visiting Professor at National Chiao Tung University, Hsinchu, Taiwan. Fax: +1 404 727 6586.

E-mail address:chemmcl@emory.edu(M.C. Lin).

(2)

the following[8]: proton transfer from the ammonium cat-ion to the perchlorate ancat-ion to form ammonia (NH3) and

perchloric acid (HClO4), electron transfer from the

per-chlorate anion to the ammonium cation, and Cl–O bond breakage. Proton transfer was believed to be the rate-con-trolling step with an activation energy of 30.0 kcal/mol

[9]. In the past several years, over 60 elementary reactions related to AP decomposition and the formation of its early products from NHxto ClOy(x = 2, 3; y = 0–4) have been

investigated in great details in our laboratory, many of which are directly relevant to the chemistry of the Freon-polluted stratosphere. The key results have been reported in a review chapter on propellant chemistry[10].

The decomposition of AP in the gas phase or in solution has not been experimentally or computationally studied. In this work, we have investigated the mechanism for the decomposition of AP in the gas phase and in solution formed by the presence of a small amount of water which may include the moisture in the sample and H2O formed

by redox reactions of NH3and HOClO3.

2. Computational methods

The geometric parameters of the reactants, products, intermediates and transition states on the potential energy surfaces of the systems studied in this work were optimized at the B3LYP level of theory [11,12] (i.e., Becke’s three-parameter nonlocal exchange functional with the non-local correlation functional of Lee, Yang, and Parr[13]) with 6-311+G(3df,2p) basis set. All the stationary points have been identiﬁed for local minima and transition states by vibrational analysis. Intrinsic reaction coordinate analyses

[14] have been performed to conﬁrm the connection between transition states and designated reactants, prod-ucts or intermediates. Higher level single-point energy cal-culations of the stationary points were reﬁned at the CCSD(T)/6-311+G(3df,2p) level, based on the optimized geometries at the B3LYP/6-311+G(3df,2p) level. For the condensed-phase calculations, the polarizable continuum model (PCM)[15,16] as implemented in GAUSSIAN03 was

used to account for the continuum solvation eﬀects. The united atom for Hartree–Fock (UAHF) model was used to build the cavity in PCM, denoted as PCM/UAHF. Water with a dielectric constant of 78 was selected to rep-resent a highly polar condensed-phase medium. All calcula-tions have been carried out using the GAUSSIAN03 program

package[17].

3. Results and discussion 3.1. Decomposition of NHþ

4/ClO 4

3.1.1. Equilibrium geometries and frequencies

The geometries of the intermediates, transition states and products optimized at the B3LYP/6-311+G(3df,2p) are shown in Fig. 1. The values in the parenthesis were obtained by the PCM method in water at the same level.

As shown in Fig. 1, the isolated ClO₄ and NHþ₄ both have a tetrahedral structure. For ClO4, the Cl–O bond

lengths in the gas phase and in solution are 1.457 and 1.455 A˚ , respectively, which are close to the experimental Cl–O value 1.44 ± 0.01 A˚ [4]in NH4ClO4crystal; The

cal-culated bond angle, 109.5° is in excellent agreement with the experimental value, 109.5 ± 1° [4]. For an isolated NHþ₄, the calculated N–H bond lengths are 1.024 and 1.029 A˚ , respectively, in the gas phase and in solution, which are in good agreement with the experimental values, 1.03 A˚ [4]and 1.021 ± 0.002 A˚ [18].

In the gas phase, an ammonium cation and a perchlo-rate anion cannot co-exist; they may form a meta-stable NHþ₄/ClO₄ complex (LM1) as shown in Figs. 1 and 2. Vibrational analysis shows that all of its frequencies are positive (seeTable 1), which indicates that it is a molecular complex. The structure of LM1 shows that the NH4 and

ClO4groups in the complex are very distorted from the

tet-rahedral structure. The ClO1 bond length in LM1 is 0.06 A˚ longer than that in the isolated ClO₄. The hydrogen bond H(1)–O(1), 1.411 A˚ is much shorter than the experimental value, 1.891 077 A˚ [4]in crystal. LM1 can easily convert to the other complex H3N–HOClO3(LM2) via a transition

state TS1. In LM2, the sharing hydrogen has almost com-pletely transferred from the ammonium group to the per-chlorate to form a complex combined by perchloric acid and ammonia. The ClO(1) bond was further increased by 0.138 A˚ comparing with that in the isolated ClO₄. TS1 connects LM1 and LM2, its structure is closer to LM2 (seeFig. 1).

In solution, the complex (LM3) predicted by the PCM method has a looser structure. Two of the four Cl–O bonds are identical with 1.447 A˚ separation, which is about 0.008 A˚ shorter than those in the isolated ClO₄ion, the ClO1 bond connecting to the NH4group increases by

around 0.02 A˚ ; the hydrogen bond H(1)–O(1), 1.784 A˚, is much closer to the experimental value, 1. 891 2.077 A˚

[4]in crystal. Three of the four hydrogens are bound iden-tically to N with the N–H bond length of 1.028 A˚ which is only 0.001 A˚ shorter than that in the isolated NHþ₄, the fourth one is bound more weakly. The H3N–HOClO3

complex in solution could not be located, no matter what initial structure was given, it was converged directly to give LM3; the result implies that in solution, NHþ₄ and ClO₄ prefer to be separated, reflecting the effect of solva-tion as confirmed by the energy diagram discussed in the following section.

There is a structure (TS2) where two hydrogens of the ammonium ion are bonded to two oxygen atoms of the perchlorate to form two identical hydrogen bonds. The bond lengths in the gas phase and solution are 1.672 and 2.078 A˚ , respectively. Again, the hydrogen bond in solution (2.078 A˚ ) is close to that in crystal (1.891 2.077 A˚ ) [4]. Vibrational analysis indicates that this structure represents the transition state for the ammo-nium cation ‘swinging’ between two of the oxygen atoms in ClO₄.

(3)

Table 1shows the calculated frequencies and the avail-able experimental values for NH3[19], NHþ4 [20,21], ClO3

[22] and HOClO3 [23]. Comparing the values in Table 1,

one can see that the maximum diﬀerences between the pre-dicted gas-phase frequencies for NH3, NHþ4, ClO3 and

HOClO3 and the experimental values are only 5.3%,

4.0%, 1.7% and 4.6%, respectively. In solution, experimen-tal frequencies for those species listed in Table 1 are not available.

3.1.2. Relative energies

The ZEP-corrected potential energy diagrams obtained at the CCSD(T)/6-311+G(3df,2p)// B3LYP/6-311+G-(3df,2p) level for the gaseous and condensed phases are presented in Figs. 2a and b, respectively. As shown in

Fig. 2a the activation energy for the transfer of a proton from the ammonium cation to the perchlorate anion in the gas phase at TS1 is only 0.5 kcal/mol, it is consider-ably lower than the expected value (30.0 kcal/mol) for

ClO₄ -1.457 (1.455) 109.5 (109.5) 1.024 (1.029) 109.5 (109.5) NH4+ 1.784 1.039 179.0 1.474 1.452 1.447 110.3 110.2 1.028 1.028 LM3 (l) 1.048 1.582 178.8 1.595 1.416 1.428 1.014 114.1 LM2 ( g) TS2 1.672 (2.078) 1.672 (2.078) 1.489 (1.463) 1.428 (1.447) 106.8 (108.6) 109.3 (109.5) 112.6 (110.2) N H Cl O O1 Cl H1 N Cl O1 H1 N O O H H 1.447 LM1 (g) 1.428 1.463 108.2 106.1 110.9 1.517 1.411 1.131 168.0 1.015 1.024 N H1 O1 Cl TS1 (g) 1.525 1.064 179.0 1.590 1.420 1.435 113.3 1.014 1.015 N O1 Cl H1

Fig. 1. Optimized geometries for the species involved in the decomposition of NH4ClO4computed at the B3LYP/6-311+G(3df,2p) level. The values in

(4)

the proton transfer process in the crystal [9]. This is expected since the surrounding ionic coulombic attrac-tions and hydrogen-bonding in crystal are not present in the gas phase. As alluded to above, in the crystal the NHþ₄ and ClO₄ ions are surrounded by each other, they are linked together by the N–H–O type hydrogen bonds, one for each hydrogen, to form a three-dimensional net-work[4]. To create a neutral molecule by proton transfer, it would likely take a much greater energy penalty for dis-rupting the ionic matrix. Thus, as expected, the existence of ion pairs in the gas phase is not favored energetically (with 111.7 kcal/mol endothermicity) and the production of perchloric acid and ammonia is preferred with 14.2 kcal/mol endothermicity from the LM1 complex. The second favored channel is the production of ClO3

and H3N–HO complex with 56.3 kcal/mol heat of

reaction.

In solution, the dissociation energy of the NHþ₄=ClO₄ complex (LM3) to NHþ₄ + ClO₄ is only 3.1 kcal/mol, reﬂecting the eﬀect of solvation as mentioned above.

How-ever, formation of NH3+ HClO4 has to overcome more

energy than that in the gas phase by 15.0 kcal/mol; simi-larly, breaking the Cl–O(1) bond in solution to form H3N–HO + ClO3also costs 20 kcal/mol more energy than

it does in the gas phase. The result suggests that AP decom-position may be enhanced by a strong solvent eﬀect. 3.1.3. The enthalpies of hydration of NHþ₄ and ClO₄

In the preceding section, the predicted structural param-eters and frequencies for several species (NH3, NHþ4, ClO3

and HOClO3) involved in this reaction have been

com-pared with those available experimental values, the agree-ment is quite satisfactory. To further conﬁrm the reliability of our calculations, the enthalpies of hydration for NHþ4 and ClO

4 are compared with available values.

For NHþ₄, the reported data are 76.5 ± 1.4, 77.9 ± 1.4 and 78.9 ± 4.2 kcal/mol [24] and 82.3 kcal/mol [25]; for ClO₄, the reported values are 55.4 ± 1.4, 54.0 ± 1.4 and 53.1 ± 4.2 kcal/mol [24]. These results are in agreement with the predicted values for NHþ₄ and ClO₄,

-1.1 LM2 H3N-HO-ClO3 0.5 TS1 14.2 NH3 + HClO4 56.3 H3N--HOClO3 111.7 NH4+ + ClO4 -0.0 LM3 NH4 + /ClO4 -29.5 NH3 + HClO4 77.1 H3N--HO + ClO3 3.1 NH4+ + ClO4

-(a) Gas phase

(b) PCM (in water) 0.0 LM1 NH₄+/ClO₄ -Kcal/mol Kcal/mol

~

Fig. 2. The schematic diagrams for NH4ClO4dissociation obtained at the CCSD(T)/6-311+G(3df,2p)//B3/6-311+G(3df,2p) level: (a) relative gas phase

(5)

80.1 and 57.2 kcal/mol, respectively, within the uncer-tainties of the reported values.

4. Conclusions

The activation energy for the transfer of a proton from the ammonium cation to the perchlorate anion to form the H3N–HOClO3 complex in the gas phase is only 0.5 kcal/

mol; it is much lower than the experimentally expected value (30.0 kcal/mol ) for the proton transfer mechanism within the crystal. The formation of separated ions in the gas phase is not favored energetically (with 111.7 kcal/ mol endothermicity) and the production of perchloric acid and ammonia is preferred with 14.2 kcal/mol endothermic-ity. On the other hand, in solution the dissociation energy of the NHþ₄/ClO₄ ionic complex to the separated NHþ₄ + ClO₄ is only 3.1 kcal/mol, clearly reﬂecting the important eﬀect of solvation. However, the formation of NH3+

HO-ClO3needs to overcome more energy than that in the gas

phase by 15.0 kcal/mol; similarly the formation of H3N–

HO + ClO3also costs more energy (20 kcal/mol) than that

does in the gas phase. The result suggests a possible strong solvent eﬀect on the AP dissociation kinetics in solution. Acknowledgements

This work was supported by the Oﬃce of Naval Re-search under Grant No. N00014-02-1-0133. MCL grate-fully acknowledges the support from Taiwan’s National

Science Council for a distinguished visiting professorship at the Center for Interdisciplinary Molecular Science, National Chiao Tung University, Hsinchu, Taiwan.

References

[1] T.B. Brill, B.T. Budenz, in Progr. Astronaut. Aeronaut. on Solid Propellant Chemistry, Combustion, and Motor Interio Ballistics, V. Yang, T.B. Brill and W.H. Ren, eds., AIAA, Washington, DC (2000) p. 3.

[2] P. Politzer, P. Lane, J. Mol. Strut. 454 (1998) 229.

[3] P.W.M. Jacobs, H.M. Whitehead, Chem. Rev. 69 (1969) 551. [4] C.S. Choi, H.J. Prask, J. Chem. Phys. 61 (1974) 3523.

[5] M. Tanaka, M.W. Beckstead, Kayaku Gakkaishi 58 (1997) 245. [6] T.B. Brill, P.J. Brush, D.G. Patil, Combust. Flame 94 (1993) 70. [7] N.E. Ermolin, O.P. Korobeinichev, A.G. Tereshchenko, V.M.

Fomin, Fizika Goreniyai Vzryva 18 (1982) 46.

[8] L .L. Bircumshaw, B.H. Newman, Proc. Roy. Soc. A227 (1954) 115; L .L. Bircumshaw, B.H. Newman, Proceedings Royal Society A227 (1954) 228.

[9] P.W.M. Jacobs, A. Russell-Jones, AIAA Journal 5 (1967) 829. [10] R.S. Zhu, M.C. Lin, in: P. Politzer, J.S. Murray (Eds.), A Chapter of

Book on Energetic Materials, Part 2, Detonation and Combustion, Elsevier Science Pub, 2003, p. 373 (chapter 11).

[11] A.D. Becke, J. Chem. Phys. 98 (1993) 5648. [12] A.D. Becke, J. Chem. Phys. 96 (1992) 2155; A.D. Becke, J. Chem. Phys. 97 (1992) 9173.

[13] C. Lee, W. Yang, R.G. Parr, Phys. Rev. 37B (1988) 785. [14] C. Gonzalez, H.B. Schlegel, J. Phys. Chem. 90 (1989) 2154. [15] R. Cammi, B. Mennucci, J. Tomasi, J. Phys. Chem. A 104 (2000)

5631.

[16] M. Cossi, G. Scalmani, N. Rega, V. Barone, J. Chem. Phys. 117 (2002) 43.

Table 1

Vibrational frequencies (cm1) for the reactants, transition states and products in the gas phase and in aqueous solution, computed at the B3LYP/6-311+G(3df,2p) level of theorya

Species Gas Solution

NHþ₄ 1484 (1447), 1723, 3383, 3483 (3343) 1422, 1668, 3274, 3344 ClO4 446, 616, 918, 1093 446, 608, 926, 1065 NHþ 4/ClO4 33, 114, 136, 330, 385, 427, 438, 506, 599, 622, 29, 42, 56, 189, 284, 311, 642, 813, 956, 1142, 1219, 1263, 1521, 1557, 440, 460, 606, 613, 620, 917, 1645, 1684, 1950, 3405, 3536, 3589, 1022,1093,1102,1412,1459, 1469, 1692, 1697, 3124, 3308, 3360, 3361 H3N–HOClO3 22, 45, 79, 252, 359, 399, 426, 469, 583, 586, 589, 769, 1029, 1129, 1172, 1220, 1258, 1565, 1659, 1668, 2351, 3477, 3586, 3593 NH3 1029 (974),1665 (1639) , 3479(3345), 3597 (3447) 1045,1628, 3408, 3514 H3N–HO 212, 222, 222, 612, 764, 1096, 1630, 1664, 265, 285, 293, 702, 934, 1118 3365, 3476, 3590, 3591 1622, 1638, 2900, 3404, 3505, 3510 HOClO3 200, 404 (421), 415 (421) , 545 (555), 568 250, 414, 416, 555, 574, 575, 702, (582), 575 (582), 705 (726), 1046 (1048), 1208 1035, 1201, 1219, 1295, 3167 (1201), 1256 (1265), 1328 (1326), 3726 (3553) ClO3 472 (476), 472 (476), 564 (566), 921 (905), 468, 468, 558, 922, 1062, 1062 1078 (1081), 1078 (1081) TS1 35i, 57, 117, 245, 350, 404, 410, 481, 581, 584, 599, 779, 1020, 1159, 1173, 1213, 1254, 1542, 1664, 1664, 2105, 3477, 3588, 3595 TS2 142i, 19, 241,283,319, 373, 430, 499 254i, 32, 108, 123, 151, 162, 441, 560, 625, 626, 886, 944, 1123, 1217, 459, 606, 615, 617, 922, 1032, 1073, 1278, 1491, 1581, 1672, 1686, 2829, 1102, 1390, 1432, 1467, 1668, 1681, 2994, 3523, 3596 3299, 3357, 3391, 3392

The values in parentheses are the experimental gas-phase data.

a

(6)

[17] M.J. Frisch et al., Gaussian 03, Revision D.01, Gaussian Inc., Pittsburgh, PA, 2003.

[18] M.W. Crofton, T. Oka, J. Chem. Phys. 86 (1987) 5983. [19] S. Su¨zer, L. Andrews, J. Am. Chem. Soc. 109 (1987) 300.

[20] M. Polak, M. Gruebele, B.W. DeCock, R.J. Saykally, Mol. Pys. 66 (1989) 1193.

[21] E. Scha¨fer, R.J. Saykally, A.G. Robiette, J. Chem. Phys. 80 (1984) 3969.

[22] H. Grothe, H. Willner, Angew. Chem. Int. Ed. Engl. 33 (1994) 1482.

[23] A.I. Karelin, Z.I. Grigorovich, V. Ya. Rosolovski, Spectrochim. Acta 31A (1975) 765.

[24] Y. Marcus, Ion Solvation, John Wiley and Sons Ltd., 1985, P. 111 and references therein.

[25] G. Klopman, P. Andreozzi, Theoret. Chim. Acta (Berl.) 55 (1980) 77.