Subscriber access provided by - Access paid by the | UCSB Libraries

Article
Investigation of the Abstraction and Dissociation Mechanism in the Nitrogen
Trifluoride Channels: Combined Post-Hartree-Fock and TST Approaches
Daniel Claudino, Ricardo Gargano, Valter Henrique Carvalho
Silva, Geraldo Magela e Silva, and Wiliam Ferreira da Cunha
J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.6b04947 • Publication Date (Web): 29 Jun 2016
Downloaded from http://pubs.acs.org on July 1, 2016

Just Accepted

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry A is published by the American Chemical

Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.
Page 1 of 19 The Journal of Physical Chemistry

1
2 Investigation of the Abstraction and Dissociation Mechanism in the Nitrogen
3 Trifluoride Channels: Combined Post-Hartree-Fock and TST Approaches
4
5 D. Claudino†, R. Gargano‡, Valter H. Carvalho-Silva§, Geraldo M. e Silva‡, W. F. da Cunha* †,‡
6 † Quantum Theory Project, Gainesville, FL 32611-2085, USA
7 ‡ Institute of Physics, University of Brasilia, Brasilia, 70.919-970, Brazil and
§ Grupo de Quı́mica Teórica e Estrutural de Anápolis, Ciências Exatas e Tecnológicas,
8
Universidade Estadual de Goiás, CP 459, 75001-970, Anápolis, Brazil
9
10 *email: wiliam@unb.br
11 phone:+55 61 3107 7718
12
13
14
15 Abstract
16
17 The present paper concludes our series of kinetics studies on the reactions involved in the complex
mechanism of Nitrogen Trifluoride decomposition. Two other related reactions that, along with this
18
mechanism, take part on an efficient boron nitride growth process are also investigated. We report
19 results concerning two abstraction reactions, namely NF2 +N2NF and NF3 +NF2NF2 , and two
20 dissociations, N2 F4 2NF2 and N2 F3 NF2 +NF. State of art electronic structure calculations at
21 the CCSD(T)/cc-pVTZ level of theory were considered to determine geometries and frequencies
22 of reactants, products and transition states. Extrapolation of the energies to the complete basis
23 set limit was used to obtain energies of all the species. We applied the transition state theory to
24 compute thermal rate constants including Wigner, Eckart, Bell and deformed theory corrections in
25 order to take tunneling effects into account. The obtained results are in a good agreement with the
26 experimental data available in the literature and are expected to provide a better phenomenolog-
ical understanding of the NF3 decomposition role in the boron nitride growth for a wide range of
27
temperature values.
28
29
30 I. INTRODUCTION
31
32
The great interest of the scientific community in the physical-chemistry of boron nitride (BN) is justified by the
33
wide range of applications that its allotropes and derived compounds present. Gas sensors1 , semiconducting devices,
34
fuel cells and sources of boron for more effective nuclear fusion reactors2 are just a few of the many possibilities of
35 use for these materials3–8 . BN films can be grown either by chemical or physical vapor decomposition. This growth
36 process takes place on reactions with several different chemical species such as BF3 , N2 , H2 and, most importantly,
37 NF3 . The use of the latter compound was observed to increase the etching rate of BN as well as to improve its
38 stabilization. In order to optimize the growth conditions of BN, an extensive kinetic mechanism was proposed9,10
39 and the fully understanding of the chemistry underlying such process requires the evaluation of the rate constant
40 for several of the elementary reactions present. Among these reactions, those involved in the breaking down of NF3
41 stands out since this process is important per se.
42 The decomposition of Nitrogen Trifluoride (NF3 ) is currently a topic of major concern for environmentalists, sci-
43 entists and members of civil society worried about climate change issues. Besides the aforementioned use on the BN
44 growth process, NF3 is regularly applied in the cleaning of Plasma-enhanced chemical vapor deposition chambers11 ,
45 which are widely used to the production of liquid-crystal displays, thin-film solar cells and several other electronic
46 devices. Although NF3 is usually said to be a less harmful substitute of sulfur hexafluoride or perfluorocarbons12
47 for these tasks, it is still a powerful greenhouse gas. Also, the higher reactivity of NF3 compared with the popular
48 alternative CF4 , for instance, is a point that ought to be carefully studied. As the use of Nitrogen Trifluoride increases
49 due to the development of the electronics industry, a special attention must be payed on how the decomposition of
50 this specie might take place.
51 Although previous studies in the literature have dealt with NF3 decomposition13–16 , a general picture of how such
52 process takes place has yet to be carried out, as these works were concerned with very particular features of the
53 decomposition mechanisms. Albeit of great importance to correctly address each particular event, the consideration
54 of such particularities prevents us from gathering a general understanding of the chemistry of process itself. In
55 this sense, the subset of reactions in the BN growth mechanism that also concerns NF3 decomposition is of major
56 importance for this issue, as it provides a generic path for the transformation of nitrogen trifluoride in other species.
57 Therefore, by studying the reactions involved in the NF3 decomposition mechanism we are simultaneously tackling a
58 problem that concerns two very important fields of study, namely the NF3 decomposition and the BN growth process.
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 2 of 19
2
1
2 For over a decade we have been studying the NF3 mechanism from the transition state theory (TST)
3 viewpoint17 . The collisional processes that pertains both to BN growth and to Nitrogen Trifluoride de-
4 composition are either of unimolecular (NF2 =NF+F, NF3 =NF2 +F, N2 F=N2 +F, N2 F3 =NF2 +NF), abstrac-
5 tion (NF2 +F=NF+F2 , NF3 +F=NF2 +F2 , NF+N=N2 +F, NF3 +N=NF2 +NF) or of exchange (NF+F=NF+F,
6 NF2 +F=NF2 +F,NF3 +F=NF3 +F) nature18 . As we have previously obtained thermal rate constants (TRC) in good
7 agreement with experimental results19,20 (whenever they were available) for all but two of these reactions, we are now
encouraged to conclude the kinetics investigation of these reactions with the computation of TRC for N2 F3  NF2
8
+ NF and NF2 + N  2NF. We also take this opportunity to present results for two other related reactions that are
9
also involved in the BN growth mechanism, namely, N2 F4  2NF2 and NF3 + NF  2NF2 .
10
Due to the wide application of the NF3 system in several fields of technology, it is very important to understand
11
the chemical processes involved in the mechanism of its decomposition. If we are to investigate the correct NF3
12
decomposition path, it becomes indispensable to study and understand the elementary reactions that might play a
13
role of intermediary steps toward clarifying the global mechanism. The calculated thermal rate constants indicate
14
that the NF2 + N  2NF, NF3 + NF  2NF2 , N2 F4  2NF2 , and N2 F3  NF2 + NF collisional processes, indeed,
15 could be part of the global mechanism of the NF3 dissociation. In this sense, our work is taking another important
16 step towards the understanding of the correct decomposition path of nitrogen trifluoride.
17 In the present work, we make use of state of art electronic structure calculations to obtain geometries, energies and
18 frequencies of reactants, products and transition states for the reactions schematised in the Figure 1. These quantities
19 are, then, used to compute the partition functions for each of the species taking part on the different reactions.
20 Following, we obtain thermal rate constants for each reaction, in the scope of TST21 including Wigner22 , Eckart23 ,
21 Bell24,25 and deformed transition state theory26,27 correction in order to take possible tunneling effects into account.
22 As we obtained TRC in good agreement to the experimental results available in the literature, we conclude that the
23 present work is a major contribution in the field of research of both NF3 decomposition and of BN growth. This is
24 because our description provides reaction rates valid over a wide range of temperature, while only a few temperature
25 values were considered in the experiments.
26 This work is organized as follows: section 2 presents a brief summary of the TST as well as of the computational
27 details employed in our calculations; we present and discuss our results in section 3; we conclude the work with a
28 summary of our achievements in section 4.
29
30
31 II. METHODOLOGY
32
33 The TST approach is arguably the most successful procedure to study kinetics of chemical reactions from the
34 theoretical point of view. Its great simplicity allied to a remarkable accuracy makes this theory one of the most
35 reliable approaches to obtain TRC for over eight decades28 . The general idea of the theory lies in considering
36 activated complexes in “quasi-equilibrium” with reactants. The rate of transformation29 is, then, obtained by a mix
37 considerations of thermodynamics, kinetic theory and statistical mechanics.
38 Crucial for TST treatment is obtaining geometries, energies and frequencies of reactants, products and transition
39 states. Considering a general bimolecular reaction such as A + BC → X ‡ → C + AB, these quantities are enough
40 to compute the partition functions QX ‡ , QA , QBC , QC and QAB of the transition state, reactants and products,
41 respectively. Following, according to the theory, the TRC is given by30–33 :
42
43
VaG
 
T ST kB T QX ‡
44 k = exp − , (1)
45 h QA QBC RT
46 where kB is the Boltzmann’s constant; h is the Planck’s constant; T is the temperature; R is the universal gas constant
47 and VaG is the potential barrier:
48
49 VaG = VM EP + εZP E . (2)
50 In this expression, εZP E is the harmonic zero-point energies (ZP E) and VM EP is the Eckart classical potential
51 energy23 point measured from the overall zero energy of the reactants:
52
53 ay by
VM EP = +
54 1 + y (1 + y)2
55
56 where
57
58 y = eα(s−s0 ) ,
59
60

ACS Paragon Plus Environment

Page 3 of 19 The Journal of Physical Chemistry
3
1
2
3
a = ∆H00 = VaG‡ (s = + inf) − VaG‡ (s = − inf),
4
5
6
7 b = (2VaG‡ − a) + 2(VaG‡ (VaG‡ − a))1/2 ,
8
9
10  
11 1 a+b
s0 = − ln ,
12 α b−a
13
14
15 µ(ν ‡ )2 b
α2 = −
16 2V ‡ (V ‡ − a)
17
18 and µ is the reduced mass of the system. a and b depend on the reactants (VaG‡ (s = − inf)), products (VaG‡ (s =
19 + inf)) and TS (VaG‡ ) energies, as well as the imaginary frequency on the TS (ν ‡ )34,35 , and y is related to the reaction
20 coordinate.
21 In order to include quantum tunneling effects along the reaction coordinate into our treatment, it was used the
22 Wigner (κW )30,36 , Eckart (κE )? , Bell-1935 (κBo )24 , Bell-1958 (κB1 )25 , Bell-1958-2T (κB2 )25 , and deformed transition
23 state theory (d-TST)26,27 corrections described by the following equations:
24
25 2
1  ~ν ‡ 

26 κW (T ) = 1 + . (3)
27 24  kB T 
28
29 Z ∞
1 VaG /kB T 0
30 κE (T ) = e PEckart (E 0 )e−E /kB T
dE 0 (4)
31 kB T 0
32
33 " 
VaG VaG
#
VaG VaG −
34 − e
kB T ~ν ‡
~ν ‡ kB T
35
36 κBo (T ) = VaG VaG
(5)
~ν ‡
−
37 kB T
38
39  
~ν ‡
40 2kB T
41 κB1 (T ) =   (6)
~ν ‡
sin
42 2kB T
43
44  
VaG VaG
 
45 ~ν ‡ kB T − ~ν ‡
G
2kB T V e
46 κB2 (T ) =  − a G  (7)
~ν ‡ Va k B T G
47 sin 2kB T ~ν ‡
− V a
48
49
50 kB T QX ‡

VG
 d1
1

~ν ‡
2
51 kdT ST = 1−d a ,d = − (8)
52 h QA QBC kB T 3 2VaG
53
where ν ‡ is the imaginary frequency for crossing the barrier. PEckart is the Eckart quantum mechanical barrier
54 ‡

55 transmission probability and d is deformed parameter26,27 . The crossover temperature Tc = k~ν BT

is the parameter
56 that delimits the degree of tunneling regimes: negligible (T > 4Tc ), moderate (Tc < T < 2Tc ) and deep (T < T c)37 .
57 A definition of a validity temperature, Td = Tc + dVaG /2kB , that delimits the applicability within the negligible or
58 moderate tunneling ranges26 .
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 4 of 19
4
1
2 By adjusting these rates to the modified Arrhenius form, we come up with a suitable way to present our results for
3 means of comparison with experimental data. The modified Arrhenius equation is given by
4 
−Ea

5 k = AT n exp (9)
RT
6
7 where A is the pre-exponential factor, n is the temperature power factor, and Ea is the activation energy.
8 According to Equation 1, the first step in employing the TST is to compute the partition functions of reactants,
9 products and transition states. These partition functions have the contributions of54 :
10 translations
11
12 (2πkB T )3/2 V
Qt = , (10)
13 ~3
14 rotations
15 √
16 8πIa Ib Ic (kB T )3/2
Qr = (11)
17 σ~3
18
and vibrations
19
20 1
Qv = . (12)
21 1− e−~ν/kB T
22
23 In the previous expressions Ia , Ib and Ic are the moments of inertia of each species relative to axis a, b and c,
24 respectively; V is the total volume of each species; σ is the symmetry number and ν is its frequency. Considering
25 the fact that equation 1 explicitly requires the energy of each specie, it becomes clear the necessity of accurately
computing geometries, frequencies and energies of reactants, products and transition states in order to obtain TRC
26
in this formalism.
27
In the present work, the values of the required parameters comes from electronic structure calculations performed
28
by using the CCSD(T)38–40 method, which translates into accounting for all single and double excitations out of
29
the reference determinant and having the contribution from triple excitations being approximated through pertur-
30
bation theory. The correlation consistent basis sets of Dunning41 were employed in the geometry as well as in the
31
evaluation of the corresponding vibrational frequencies and ZPE’s, using the triple-zeta quality set, cc-pVTZ. The
32
geometries obtained in this way were used to carry out single-point energy calculations using the more extended basis
33 sets cc-pVQZ, which allow for the extrapolation to the complete basis set (CBS) limit. Self-consistent field (SCF)
34 and correlation energies are treated separately when extrapolating to the CBS limit. The former is subject to the
35 extrapolation suggested by Parthiban and Martin,42 as shown in Eq. 13, while the correlation energy is extrapolated
36 by using the cubic formula suggested by Helgaker,43 displayed in Eq. 14.
37
38
SCF SCF −5
39 EX = E∞ + ASCF
∞ X (13)
40
41
42 corr corr −3
43 EX = E∞ + Acorr
∞ X (14)
44 All the electronic structure results reported in the present article were carried out by making use of the quantum
45 chemical packages ACESII44 and ACESIII45 . Such calculations were subject to the frozen-core approximation, meaning
46 all the 1s orbitals from both flourine and nitrogen atoms were not included in the computation of the correlation energy.
47 This is an important procedure to obtain results consistently, for the chosen basis set is known to not address core
48 energies adequately. Moreover, the core-correlation is understood to be nearly constant all throughout the potential
49 energy surface46 . That being said a more rigorous treatment of core orbitals, in practical terms, would mean a
50 significantly higher computational cost with very little improvement.
51 The characteristics thermal rate constant, with the tunneling corrections and deformed transition state theory is
52 determined using our own code, which is described in the literature26,27,32,33,47 .
53
54
55 III. RESULTS AND DISCUSSIONS
56
57 We begin our discussion by presenting, in Table I, the geometric parameters obtained for reactants and products.
58 These values, obtained by means of CCSD(T)/cc-pVTZ level of theory calculations, are to be used to compute the
59
60

ACS Paragon Plus Environment

Page 5 of 19 The Journal of Physical Chemistry
5
1
2 moments of inertia of reactants and products. This level is known to provide a suitable compromise between accuracy
3 and computational feasibility48 . Also, frequencies and energies of the stable species are to be computed for these
4 structures. Preliminary tests have shown that the frozen core approximation does not play a fundamental role in
5 defining geometries. Energies and frequencies, on the other hand, are much more sensible to the poor treatment given
6 by cc-pVTZ to the 1s orbitals. Therefore, such orbitals can not be included in the correlation energy for single point
7 and vibrational frequencies calculation. In order to maintain coherence, we chose to follow the same procedure for
geometry optimization as well. A careful comparison of the data presented in Table I with experimental values of
8
literature allows us to conclude that the employed level of theory is suitable for the considered systems. This table
9
contains the structural data in cartesian coordinates just for ease of representation and use of our structures for future
10
research. In the following, however, we make use of internal coordinates to compare our structural parameters to
11
those obtained from the experimental references highlighted. The absolute deviation of the interatomic distances are
12
as small as 3mÅ for NF49,50 , 20mÅ for NF2 51? , 1mÅ for NF3 52 and 7mÅ for N2 F3 52 . As for the angles, we again
13
observe that our calculations yielded accurate predictions, as the absolute deviation is smaller than 1o for all cases,
14
except for the angle between the NNF atoms, whose deviation is around 2o . Overall, we conclude that geometries
15 were obtained with a remarkable agreement with experimental results.
16
17
18 TABLE I: CCSD(T)/cc-PVTZ optimized geometries (Å ) for reactants and products.
19
Specie Atom X Y Z
20
NF N 0.00000000 0.00000000 0.00000000
21
22 F 0.00000000 0.00000000 1.32042458
23
24 NF2 F 0.00000000 0.00000000 0.00000000
25 N 0.00000000 0.00000000 1.34997605
26 F 1.31373762 0.00000000 1.66066686
27
28 NF3 N 0.00000000 0.00000000 1.00000000
29 F 1.23032909 0.00000000 1.60659225
30
F -0.61516455 1.06549625 1.60659225
31
F -0.61516455 -1.06549625 1.60659225
32
33
34 N2 F3 F -2.66594746 0.79631381 -0.24491281
35 F -5.90502325 0.26227956 0.21795906
36 F -4.79430232 2.13238846 0.26742052
37 N -3.75199730 0.13912578 0.21899339
38 N -4.81797771 0.87869115 -0.35580799
39
40 N2 F4 N -4.09606247 3.96454896 0.43341220
41 N -2.85363748 3.99608094 -0.43341210
42
F -4.78564271 5.02537461 -0.10395417
43
F -4.73537156 2.87582905 -0.11027535
44
45 F -2.21432849 5.08480095 0.11027530
46 F -2.16405724 2.93525534 0.10395417
47
48 We now make use of the geometric parameters from Table I to compute single point energies for these species. In
49 this work we decided to be as accurate as one can be on evaluating the energies, since they are known to play a crucial
50 role on determining the TRC. Therefore we first make use of the same CCSD(T)/cc-pVTZ level of theory as used for
51 determining the geometries of the reactants and products. Following we increase the basis set size to cc-pVQZ and
52 perform another single point calculations. By observing the trend of the energies evolution we perform extrapolations
53 of their values to the complete basis set (CBS) limit, according to expressions 13 and 14. This procedure results
54 in the data set of Table II, which presents SCF (on top), correlation (in the middle) and total (bottom) energies
55 given in Hartree. Our energy values are compatible to experimental thermodynamics properties previously obtained
56 in the literature (see Support Information). In fact, one can see that even our starting point CCSD(T)/cc-pVTZ
57 — the very same method used to geometry optimization — yields excellent results as far enthalpies of reactions are
58 concerned . For this level, we obtained a ∆H of -8.13 kcal/mol against the experimental value of -4.1 kcal/mol for the
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 6 of 19
6
1
2 NF2 +N→2NF. For NF3 +NF2NF2 , a good agreement of our -7.23 value against the experimental of -7.70kcal/mol
3 was achieved and for N2 F4 2NF2 we obtained 24.43 kcal/mol, while the experimentally expected was of 20.90.
4 These agreements decidedly show the suitability of the CCSD(T)/cc-pVTZ for determining the geometries and allow
5 us to conclude that the CBS procedure further improves the energies by starting from a very suitable method as far
6 thermodynamics properties are concerned. Inasmuch we employed an accurate coupled cluster method in the limit of
7 the complete basis set, we can consider the values presented in Table II as a new benchmark in the literature as far
as the total electronic energies of these systems is concerned.
8
9
10 TABLE II: SCF (top), correlation (middle) and total (bottom) energies (in a.u.) for the reactants and products of the reactions
11 studied in the present work using cc-pVTZ, cc-pVQZ and extrapolation to the CBS limit.
12 Species cc-pVTZ cc-pVQZ CBS
13 -54.400686 -54.403718 -54.404661
14
N -0.114021 -0.121106 -0.126275
15
-54.514707 -54.524824 -54.530937
16
17
18 -153.837042 -153.848984 -153.852698
19 NF -0.412061 -0.444973 -0.468989
20 -154.249103 -154.293957 -154.321688
21
22 -253.259643 -253.280383 -253.286836
23 NF2 -0.710411 -0.768820 -0.811443
24 -253.970054 -254.049203 -254.098279
25
26
-352.675530 -352.704832 -352.713948
27
28 NF3 -1.004687 -1.088349 -1.149398
29 -353.680218 -353.793181 -353.863347
30
31 -407.076324 -407.109443 -407.119747
32 N2 F3 -1.182977 -1.278055 -1.347436
33 -408.259301 -408.387498 -408.467184
34
35 -506.492404 -506.533878 -506.546782
36 N2 F4 -1.481312 -1.601210 -1.688702
37
-507.973716 -508.135088 -508.235485
38
39
The frequencies for reactants and products are obtained by means of a CCSD(T)/cc-pVTZ level of theory. The
40
frequencies values of the vibrational modes are given in Table III (in cm−1 ) together with the zero point energies (in
41
kcal mol−1 ). One can see that our electronic structure calculations also performs extremely well as far as comparison
42
with available experimental frequency results goes. For NF, we can see that our frequency value of 1154.9 cm−1 is very
43
close to the experimentally obtained value of 1151 cm−1 .? Similar agreement was observed for NF2 , whose absolute
44
deviations for the corresponding vibrational modes are of 13.6, 71.1 and 51.3 cm−1 .51 Finally, the agreement of the
45
vibrational modes of NF3 with experimental results is also as good as 5.5, 11.9, 51.2 and 31.4 cm−1 .53 It is important
46 to note, however, that the inconsistencies in the discrepancies of our frequencies from the experimental data can very
47 reasonably be attributed to the fact that experiments under different conditions were performed for each system. As
48 it was not goal of this work to consider particular peculiarities of experiments, we conclude that our calculations were,
49 indeed, very successful, for in general we obtained a good agreement for all reported frequencies.
50 Now that geometries, energies and frequencies of the equilibrium species (reactants and products) are determined,
51 the next step is to obtain these quantities for the transition states. As usual, we begin by determining the optimal
52 geometry of the four transition states. The corresponding values are presented in Table IV. A careful analysis of
53 these results together with a comparison of the appropriate geometries of Table I suggests that the transformation
54 mechanism is found for each of the reactions studied. For instance, one can see that for the collisional process
55 described by N2 F4  2NF2 the transition state, indeed, presents a nitrogen-nitrogen distance value of 2.9 Å, which is
56 considerably larger than that of the equilibrium reactant (1.515 Å). In other words, from the geometric point of view,
57 N2 F4 is really dissociating into two NF2 molecules, for the N—N distance is greater than that of the reactant and yet
58 not large as to disregard interaction between the two NF2 molecules of product. Similar observations can be made for
59
60

ACS Paragon Plus Environment

Page 7 of 19 The Journal of Physical Chemistry
7
1
2 TABLE III: Harmonic vibrational frequencies (in cm−1 ) and zero-point energies (εZP E ) (in kcal mol−1 ) for the reactants and
3 products of the reactions studied in the present work at the CCSD(T)/cc-pVTZ level of theory.
4
Species ν1 ν2 ν3 ν4 ν5 ν6 ν7 ν8 ν9 ν10 ν11 ν12 εZP E
5
6 NF 1154.9 - - - - - - - - - - - 1.651
7
8 NF2 587.7 1001.8 1120.8 - - - - - - - - - 3.875
9
10 NF3 502.5 659.9 949.2 1058.4 - - - - - - - - 6.607
11
12 N 2 F3 139.4 313.3 448.3 511.0 584.2 838.8 898.6 1042.3 1100.3 - - - 8.400
13
14
N 2 F4 113.3 247.1 364.1 471.1 507.7 547.1 613.0 914.2 975.0 996.4 1042.8 1070.7 11.240
15
16
17 all unimolecular and abstraction channels. These considerations together with the analysis of the normal modes, to
18 be discussed shortly, allow us to conclude that the transition states geometries are qualitatively correct (See scheme
19 of Figure 1).
20
21
22 TABLE IV: CCSD(T)/cc-PVTZ optimized geometries (Å ) for the transition states.
23 Reaction Atom X Y Z
24
NF2 + N  2NF F -0.65477161 0.23083702 0.00036023
25
F 1.60788665 -0.51978165 0.00000996
26
27 N 1.04613299 0.66638957 -0.00025893
28 N -2.33925378 -0.27436941 -0.00024331
29
30 NF3 + NF  2NF2 F -0.33972038 0.40565394 -0.33943676
31 F 1.96021065 0.82722569 0.43880369
32 F 1.47515915 -1.20964925 0.07679894
33 F -2.57464230 -0.54181424 0.21935521
34 N 1.45381672 0.00984004 -0.51746331
35
N -2.16068326 0.69373880 -0.01915243
36
37
38 N2 F4  2NF2 F 1.80887205 0.77652231 -0.29958494
39 F 1.08801047 -0.94677044 0.69826938
40 F -1.80888281 -0.77689838 -0.29931332
41 F -1.08804118 0.94724318 0.69809160
42 N -1.21972319 0.41660911 -0.54115950
43 N 1.21977944 -0.41674028 -0.54078294
44
45 N2 F3  NF2 + NF F 1.56973225 0.84191444 -0.05998237
46 F 0.87792665 -1.15037832 -0.23935588
47
F -2.07942141 -0.18695009 0.14605966
48
N -1.25730591 0.72432083 -0.33118341
49
50 N 0.75770667 -0.05217731 0.53914124
51
52 As far as energies of the transition states are concerned, we repeat the same procedure of obtaining the best possible
53 values as before. In other words, Table V presents SCF, correlation and total energies for cc-pVTZ, cc-pVQZ and the
54 CBS extrapolation in Hartrees. Finally, we compute the frequencies of the transition states. Note that each line in
55 Table VI present one — and only one — imaginary mode, which is a unambiguous signature of a saddle point that
56 characterizes transition states. In order to make sure that the obtained saddle point corresponds to the transition
57 state considered in each reaction we used MOLDEN55 to visualize the stretching and shrinking pattern that represents
58 the imaginary mode. By studying this behavior we were able to see that, for all cases, such imaginary vibrations
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 8 of 19
8
1
2 corresponded to the natural step from turning reactants into products, thus corresponding to the correct transition
3 state. It should be noted that, again, these values arise from a CCSD(T)/cc-pVTZ calculation.
4
5 TABLE V: SCF (top), correlation (middle) and total (bottom) energies (in a.u.) for the transition states of the reactions
6 studied in the present work using cc-pVTZ, cc-pVQZ and extrapolation to the CBS limit. Calculations at CCSD(T) level of
7 theory.
8 Reaction cc-pVTZ cc-pVQZ CBS
9
-307.603826 -307.627170 -307.634434
10
NF2 + N  2NF -0.864204 -0.884089 -0.898601
11
12 -308.468031 -308.511260 -308.533035
13
14 -506.218266 -506.509877 -506.600608
15 NF3 + NF  2NF2 -2.157407 -1.519464 -1.775066
16 -508.375674 -508.029341 -508.375674
17
18 -506.352789 -506.410365 -506.428279
19 N2 F4  2NF2 -1.662333 -1.700755 -1.728791
20 -508.015123 -508.111120 -508.157071
21
22
-407.056659 -407.108321 -407.015400
23
24 N2 F3  NF2 + NF -1.205080 -1.232352 -1.302526
25 -408.261740 -408.340673 -408.372918
26
27
28
TABLE VI: Harmonic vibrational frequencies (in cm−1 ) and zero-point energies (εZP E ) (in kcal mol−1 ) for the transition states
29 of the reactions studied in the present work at the CCSD(T)/cc-pVTZ level of theory.
30
31 Reaction ν1 ν2 ν3 ν4 ν5 ν6 ν7 ν8 ν9 ν10 ν11 ν12 εZP E
32 NF2 + N  2NF 1135.37i - - - - - - 132.85 207.21 318.07 470.45 1173.19 -3.2905
33
34 NF3 + NF  2NF2 923.18i 44.32 112.30 203.14 278.53 453.65 476.34 613.70 641.36 1006.53 1026.000 1089.88 10.1842
35
36 N2 F4  2NF2 167.26i 31.23 80.65 106.77 215.11 303.53 584.22 588.77 965.07 975.82 1062.97 1096.63 -8.5482
37
38 N2 F3  NF2 + NF 570.89i - - - 106.85 172.06 286.34 314.65 564.15 925.00 1030.16 1087.77 -6.4145
39
40
41
As all the quantities (geometries, energies, and frequencies) necessary to compute TRC have been presented, we are
42
in the position of using of Equation 1 (together with the six different tunneling corrections presented in this work) to
43
achieve this goal for the different systems. The total cc-PVQZ energy values were used in TRC calculations, because
44
they provide potential barrier and reaction enthapy in agreement with experimental data (see Support Information),
45
except, for the N2 F4 →2NF2 which the kinetic and thermodynamic parameters were better described by total CBS
46 energies values.
47 Figures 2, 3, 4 show the experimental and theoretical (with and without tunneling corrections) TRC plots against
48 the reciprocal temperature in the range of 200-4000K for the NF2 +N→2NF, NF3 +NF→2NF2 , and N2 F4 →2NF2
49 reactions, respectively. Throughout this work, the TRCs are presented in a logarithmic scale in the vertical axes (in
50 units of cm3 mol−1 s−1 for abstraction reaction and s−1 for dissociation reaction), whereas the reciprocal temperature
51 is scaled as 1000/T and is presented in units of K−1 . These choices allow for a better visualization of the TRC
52 dependence with temperature.
53 We begin by discussing the results for NF2 +N→2NF TRC plots (Figure 2). From these plots one can observe a
54 similar behavior of the TRC with and without tunneling corrections in the range 300-4000K. Nevertheless, the Eckart
55 plot presents a significant deviation at high temperature. Furthermore, tunneling effects become more important
56 as temperatures fall below 300K (close to Tc = 260K). Again, κB0 and κB2 plots deviate strongly from the others
57 methods that take into account tunneling effects. Although tunneling is more significant below the critical temperature
58 Tc , it is important to note that d-TST correction can describe only tunneling effects above of Td = 254K. All the
59
60

ACS Paragon Plus Environment

Page 9 of 19 The Journal of Physical Chemistry
9
1
2 evaluated NF2 +N→2NF TRC (except the Eckart corrections, as mentioned above) are in a good agreement with
3 available experimental data of Cheah and Clyne61 . It should be stressed that the discrepancy between the calculated
4 and the experimentally observed TRC values can be partially attributed to peculiarities of the system as well as the
5 experimental procedure. Cheah and Clyne studied the kinetic of this reaction using atomic resonance absorption in
6 order to account for the electronically-excited species of NF2 once this reaction is an extremely rich source of excited
7 states. However, our study just considered ground state species and our results suggest that the excited states do not
contribute significantly as far as qualitative aspects of the reactive channel described are concerned.
8
For the TRC of the NF3 +NF→2NF2 reaction (Figure 3), it is also noted a similar behavior between the conventional
9
and corrected curves in the range 250-4000K. In this case, however, the Eckart corrections overestimate the tunneling
10
effects for almost all the temperature range considered, except for very high temperatures. For this reaction, tunneling
11
effects are specially important as temperatures fall below 250K, around the cross over temperature (Tc = 211K). In this
12
temperature regime, Bell-1935 (κB0 ) and Bell-1958 (κB2 ) plots deviate significantly from the other corrected curves.
13
The deep tunneling is highlighted and expresses itself by the presence of significant deviation below the Tc . It should
14
be remarked, however, that the d-TST can describe only tunneling above Td = 207K. All theoretical TRC (except
15 the Eckart corrections that overestimated the tunneling effects for almost all the temperature range considered) are
16 in a reasonable agreement with the experimental data of Weiller60 , as can be seen in Figure 3. Considering that the
17 electronic state of NF is formed efficiently in this reaction, neglecting non-adiabatic process in this reactive channel
18 can affect the description of TRC parameters. In fact, experimental TRC measured in the ground (red circles in
19 Figure 3) and a specific excited-state of NF (blue circle in Figure 3) suggest that these states are strongly coupled for
20 this reaction. However, the current potential surface was calculated considering Born-Oppenheimer approximation
21 (adiabatic process). This fact may be a possible reason for the discrepancies found between the theoretical and
22 experimental TRC at high temperatures.
23 From Figure 4, it is possible to note that the N2 F4 →2NF2 conventional TRC has a similar behavior as the TRC with
24 tunneling corrections in the range 250-4000K. This is a fact already expected and can be understood from classical
25 arguments as will be discussed shortly. Tunneling effects start to be important around the cross over temperature
26 (Tc = 38K). These are the cases of Bell-1935 (κB0 ) and Bell-1958 (κB2 ) corrected curves that deviate significantly
27 from the others methods that include the tunneling effects. The magnitude of the d-TST plot (quantified by the
28 parameter d) indicates substantial contribution of the tunneling effects. The deep tunneling is highlighted by a
29 significant deviation below the Tc , even though the d-TST can only properly describe tunneling above of Td . In other
30 words, we can say that the validity temperature for this model is 38K. One can observe a great coincidence of the
31 conventional and corrected curves at high temperature regimes. This is due to the low importance of tunneling in this
32 regime, since thermal excitation prevails in promoting the reaction. In this case it is clear that all calculated TRC
33 are in a reasonable agreement with the experimental data of Baulch and et al.56 . The kinetics of N2 F4 decomposition
34 was measured by shock tube techniques which describes an efficient decomposition process at temperatures above
35 300 K. This reaction presents a strong pressure-dependence with a transition from first to second order reaction at
36 low pressure. Another factor that influences the reaction kinetics is the collision partner (Ar or N2 ) used in the
37 experiment. Based on these facts, we suggest that the discrepancies between the experimental and theoretical TRC
38 can be also attributed to both factors. Furthermore, the obtained TRC for this reaction has been calculated with
39 RRKM approach, and similarly our results, were higher than those determined experimentally57 .
40 So far, we have demonstrated the suitability of the TST formalism for treating this kind of reactive system. There
41 are a limited number of studies about the decomposition mechanism of N2 F3 . A simple consideration regarding this
42 decomposition process is presented in the literature58,59 considering an alternative reactive channel with the formation
43 of N2 F2 and F. Therefore, in order to provide further contribution to this field, we can proceed using the TST formalism
44 in the same fashion adopted so far for the N2 F3 →NF2 +NF reaction, in which experimental evidences are absent in
45 the literature. The goal here is to provide reliable data for future comparison. Figure 5 shows conventional, Wigner,
46 Bell, and deformed transition state theory plots of the thermal rate constants against the reciprocal temperature in
47 the range of 200-4000K. The evaluated TRC, present a similar behavior among conventional and all corrected curves.
48 The exception is, again, the Eckart correction (it is not shown in Figure 5) that overestimated the tunneling effects for
49 almost all the temperature range considered. One can also notice that tunneling effects are negligible at 200-4000K.
50 As expected, the tunneling contribution is recognized to promote the reaction only below the Tc = 131K (this value
51 is not represented in the range plotted in the Figure 5).
52 An interesting way to summarize the TRC values presented in this work is to perform a fitting of their dependence
53 with temperature in a modified Arrhenius form, i.e., equation 9. The fitting coefficients provide an useful way to
54 compare our results with others from the literature. We conclude the present discussion by presenting, in Table VII,
55 the modified Arrhenius coefficients (A, n and Ea ) used to fit the TRC for all reactions in the temperature range
56 200-4000K. The NF3 +NF2NF2 and NF2 +N2NF modified Arrhenius coefficients for the Eckart corrected TRC
57 did not converge. Nevertheless, the importance of such correction is limited as we have shown that it overestimates the
58 tunneling effects. Similary, Bell-modified Arrhenius coefficients did not converge either, because they have a strong
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 10 of 19
10
1
2 curvature in the crossovertemperature (Tc ). The parameters of Table VII allow us to obtain values for TRC at any
3 temperature in a considered range. Therefore, it is of fundamental importance to conduct further studies on these
4 reactions, since by virtue of out description, one does not need to be limited to the temperature values in which the
5 experiments were performed.
6
7 TABLE VII: A (cm3 mol−1 s−1 ), n and Ea (kcal mol−1 ) modified Arrhenius coefficients obtained from the fitting of TRC
8 obtained through TST theory.
9
Reaction Coefficients Conventional Wigner Eckart d-TST
10
NF2 +N2NF A 4.8330x107 1.6384x107 - 6.93062x105
11
12 n 1.6644 1.78600 - 2.16435
13 Ea 38660 37955 - 37037
14
15 NF3 +NF2NF2 A 1.14630x105 4.4114x104 - 7.30300x103
16 n 2.58610 2.69450 - 2.91003
17 Ea 35847 35290 - 34779
18
19 N2 F4 2NF2 A 1.1900x1019 1.777x1019 1.9000x1019 9.05282x1018
20
n 0.29910 0.30680 0.29910 0.39020
21
Ea 47948 47918 47948 47791
22
23
24 N2 F3 NF2 +NF A 4.7091x1018 2.58660x1018 2.1780x1018 9.98308x1017
25 n 0.244280 0.30872 0.24428 0.42977
26 Ea 28352 28068 28352 27848
27
28
29
30 IV. CONCLUSIONS
31
32 In this work we employed the Transition State Theory scheme to calculate thermal rate constants of two abstraction
33 (NF2 +N2NF and NF3 +NF2NF2 ) and two dissociations (N2 F4 2NF2 and N2 F3 NF2 +NF) reactions. In order
34 to do so, accurate electronic structure calculations at the CCSD(T)/cc-pVTZ level were performed to obtain geometries
35 and frequencies of reactants, products and transition states. As for the energies of all the species, we chose to consider
36 an extrapolation to the continuum basis set limit in order to be more precise in treating this sensible quantity. This
37 level was considered whenever its use resulted in consistent barrier height. After computing the thermal rate constants
38 including Wigner, Eckart, Bell and deformed transition state theory tunneling corrections, it was verified that our
39 results were in a good agreement with the experimental data available in the literature. We also presented TRC
40 fitting coefficients of a modified Arrhenius form in order for our results to be useful in future comparisons. With this
41 contribution we conclude a series of works concerning TRC determination for reactions involved in nitrogen trifluoride
42 decomposition and boron nitride growth mechanisms. The kinetic data here reported is expected to be of critical
43 importance in guiding future research concerning the fundamental aspects of such complex and important mechanism.
44
45
46 V. SUPPORTING INFORMATION
47
48 Minimum energy path of the four reactions (with and without ZPE correction) together with thermodynamic
49 properties are provided.
50
51
52 VI. ACKNOWLEDGEMENTS
53
54 The authors gratefully acknowledge the financial support from the Brazilian Research Councils CNPq and FAPDF.
55 W.C. thanks CNPq for supporting his research visit to University of Florida.
56
57
58
59
60

ACS Paragon Plus Environment

Page 11 of 19 The Journal of Physical Chemistry
11
1
2 REFERENCES
3
4
5
6
7 1
Paura, E. N. C.; da Cunha, W. F; Roncaratti, L. F.; Martins, J. B. L.; e Silva, G. M.; Gargano, R. CO Adsorption on
8 Single-walled Boron Nitride Nanotubes Containing Vacancy Defects R. S. C. Advances 2015 5, 27412-27420.
2
9 Labaune, C.; Baccou, C.; Depierreux, S.; Goyon, C.; Loisel, G.; Yahia, V.; Rafelski, J. Fusion Reactions Initiated by
10 Laser-accelerated Particle Beams in a Laser-produced Plasma Nature Communications 2013, 4, 2506-2511.
3
11 Badzian, A.; Badzian, T. Recent Developments in Hard Materials Int. J. Refract. Met. Hard 1997, 15, 3-12.
4
Léger, J. M.; Haines, J. The Search for Superhard Materials Endeavour 1997, 21, 121-124.
12 5
Kester, D. J.; Ailey, A. S.; Lichtenwalner, D. J.; Davis, R. F. Growth and Characterization of Cubic Boron Nitride Thin
13 Films J. Vac. Sci. Technol. 1994 A12, 3074-3081.
14 6
Berns, D. H.; Cappelli, M. A.; Shuh, K. Near-edge X-ray Absorption Fine Structure Spectroscopy of Arcjet-Deposited Cubic
15 Boron Nitride Diamond Relat. Mater. 1997, 6, 1883-1886.
7
16 Ma, X.; Yue, J.; He, D.; Chen, G. Characterization of Cubic Boron Nitride Thin Films Grown on Different Substrates Mater.
17 Lett. 1998 36, 206-209.
8
18 Arya, S. P. S..; DÁmico, A. Study of Sputtered ZnO-Pd Thin Films as Solid State H2 and NH3 Gas SensorsThin Solid
19 Films 1988, 157, 169-174.
9
20 Barreto, P. R. P.; Kull, A. E.; Cappelli, M. A. Kinetic and surface mechanisms to growth of hexagonal boron nitride, in:
Kumar, A.; Meng, W. J.; Cheng, Y. -T.; Zabinski, J.; Doll, G. L.; Veprek, S. (Eds.), Surface Engineering 2002 and Synthesis,
21
Characterization and Applications, number 750, Fall Meeting Proceedings, page Y5.13, Warrendale, 2002 (Materials Research
22 Society).
23 10
Barreto, P. R. R.; Cappelli, M. A.; Matsumoto, S. Growth mechanism for boron nitride, in: International Congress on
24 Plasma Physics, vol. 3, Quebec City, CA 2000, pp. 904.
11
25 Riva, A.; Pittroff, M.; Schwarze, T.; Oshinowo, J.; Wieland, R. Etch Performance of Ar/N2/F2 for CVD/ALD Chamber
26 Clean Solid State Technology 2009, 52, 20-24.
12
27 Reichardta, H.; Frenzela, A.; Schoberb, K. Environmentally Friendly Wafer Production: NF3 Remote Microwave Plasma
28 for Chamber Cleaning Microelectronic Engineering 2001, 56, 73-76.
13
Dillon, T. J.; Vereecken, L.; Horowitz, A.; Khamaganov, V.; Crowleya, J. N.; Lelieveld, J. Removal of the Potent Greenhouse
29
gas NF3 by Reactions with the Atmospheric Oxidants O(1D), OH and O3 Phys. Chem. Chem. Phys. 2011, 13, 18600-18608.
30 14
Yin, H-M.; Yang, B-H.; Han, K-L.; He, G-Z.; Guo, J-Z.; C-P. Liu, C-P.; Gu, Y-S. Theoretical Study of The Reaction H+NF3
31 -¿ NF2+HF Phys. Chem. Chem. Phys. 2000, 2, 5093-5097.
32 15
Chen, H. L.; Lee, H. M.; Chang, M. B. Kinetic Modeling of the NF3 Decomposition via Dielectric Barrier Discharges in
33 N2/NF3 Mixtures Plasma Process. Polym. 2006, 3, 682-691.
16
34 Wang, Y. F.; Wang, L. C.; Shin, M. L.; Tsai, C. H. Effects of Experimental Parameters on NF3 Decomposition Fraction in
35 an Oxygen-based RF Plasma Environment Chemosphere 2004, 57, 1157-1163.
17
36 Barreto, P. R. P.; Vilela, A. F. A.; Gargano, R.; Ramalho, S. S.; Salviano, L. R. NF3+N=NF2+NF Rate Constant Calculated
37 Using TST with Simple Tunneling Correction Theochem 2006, 769, 201-205.
18
Ramalho, S. S.; Barreto, R. P. P.; Martins, J. B. L.; e Silva, G. M.; Gargano, R. A Computational Investigation of the
38
Multiple Channels of the NF2 + F Reaction J. Phys. Chem. A 2009, 113, 14336-14342.
39 19
Ramalho, S. S.; da Cunha, W. F.; Barreto, R. P. P.; de Oliveira Neto, P. H.; Roncaratti, L. F.; e Silva, G. M.; Gargano, R.
40 Thermal Rate Constant Calculation of the NF + F Reactive System Multiple Arrangements J. Phys. Chem. A 2011, 115,
41 8248-8254.
20
42 Ramalho, S. S.; da Cunha, W. F.; Albernaz, A. F.; de Oliveira Neto, P. H.; e Silva, G. M.; Gargano, R. New Journal of
43 Chemistry An extensive investigation of Reactions Involved in the Nitrogen Trifluoride Dissociation 2013, 37, 3244-3251.
21
44 Truhlar, G. G.; Garret, B. C.; Klippenstein, S. J. Current Status of Transition-State Theory J. Phys. Chem. 1996, 100,
45 12771-12800.
22
46 Truhlar, G. G.; Isaacson, A. D.; Garret, B. C. Theory of Chemical Reaction Dynamics, vol. 2. M. C. Baer, Ed. CRC Press:
boca Raton, FL, 1985.
47 23
Eckart, C. The Penetration of a Potential Barrier by Electrons Phys. Rev. 1930, 35, 1303-1309.
48 24
Bell, R. P. Quantum Mechanical Effects in Reactions Involving Hydrogen Proc. R. Soc. London. Ser. A, Math. Phys. 1935,
49 148, 241-248.
50 25
Bell, R. P. The Tunnel Effect Correction for Parabolic Potential Barriers Trans. Faraday Soc. 1959, 55, 1-4.
26
51 Silva, V. H. C.; Aquilanti, V.; de Oliveira, H. C. B.; Mundim, K. C. Uniform Description of non-Arrhenius Temperature
52 Dependence of Reaction Rates, and a Heuristic Criterion for Quantum Tunneling vs Classical non-Extensive Distribution
53 Chem. Phys. Lett. 2013, 59, 201-207.
27
54 Silva, V. H. C.; Aquilanti, V.; Oliveira, H. C. B.; Mundim, K. C. Deformed Transition State Theory: Inclusion of the
55 Tunneling Effect by Euler Exponential, Limit of Validity and Description of Bimolecular Reactions Rev. Process. Qumicos
2015, 9, 226-228.
56 28
Ju, L. P.; Han, K. L.; Zhang, J. Z. H. Global Dynamics and Transition State Theories: Comparative Study of Reaction Rate
57 Constants for Gas-Phase Chemical Reactions J. Comput. Chem. 2009, 30, 305-316.
58 29
Laidler, K. J.; King, M. C. Development of Transition-state TheoryJ. Phys. Chem. 1983, 87, 2657-2664.
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 12 of 19
12
1 30
2 Truhlar, D. G.; Isaacson, A. D.; Garrett, B. C. Theory of Chemical Reaction Dynamics, vol. 2. CRC Press, Inc., Boca Raton,
Florida Oxford Science, Boca Raton, Florida, 1985.
3 31
Pilling, M. J.; Seakins, P. W. Reaction Kinetic, vol. 4. Oxford Science, Oxford, UK, second edition, 1995.
4 32
Barreto, P. R. P.; Vilela, A. F. A.; Gargano, R. A Simple Program to Determine the Reaction Rate and Thermodynamic
5 Properties of Reacting System J. Mol. Struct (Theochem) 2003, 639, 167-176.
6 33
Barreto, P. R. P.; Vilela, A. F. A.; Gargano, R. Theoretical Study of the reactions BF3+BX, Where X=H or N Int. J.
7 Quantum Chem. 2005, 103, 685-694.
34
8 Truong, T. N.; Truhlar, D. G. Ab Initio Transition State Theory Calculations of the Reaction Rate for OH + CH4 -¿ H2O
9 + CH3 J. Chem. Phys. 1990, 93, 1761-1769.
35
10 Pardo, L.; Banfelder, J. R.; Osman, R. Theoretical Studies of the Kinetics, Thermochemistry, and Mechanism of Hydrogen-
11 abstraction from Methanol and Ethanol J. Am. Chem. Soc. 1992, 114, 2382-2390.
36
Henon, E.; Bohr, F. Comparative ab initio MO Investigation on the Reactivity of the Three NH(a ∆), NH(X 3 Σ− ) and
12
NH2(X 2 B1 ) Radical Species in their Bimolecular Abstraction Gas-phase Reaction with the HN3 Molecule J. Mol. Struct
13 (Theochem) 2000, 531, 283-299.
14 37
Christov, S. G. The Characteristic (Crossover) Temperature in the Theory of Thermally Activated Tunneling Processes Mol.
15 Eng. 1997, 7, 109-147.
38
16 Urban, M.; Noga, J.; Cole, S. J.; Bartlett, R. J. Towards a Full CCSDT Model for Electron Correlation J .Chem. Phys.
17 1985, 83, 4041-4046.
39
18 Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. A Fifth-order Perturbation Comparison of Electron
19 Correlation Theories Chem. Phys. Lett. 1989, 157, 479-483.
40
Watts, D.; Gauss, J.; Bartlett, R. J. Coupled-cluster Methods with Noniterative Triple Excitations for Restricted Open-shell
20
HartreeFock and Other General Single Determinant Reference Functions. Energies and Analytical Gradients J. Chem. Phys.
21 1993, 98, 8718-8733.
22 41
Dunning Jr, T. H. Gaussian Basis Sets for Use in Correlated Molecular Calculations. I. The Atoms Boron Through Neon
23 and Hydrogen J. Chem. Phys. 1989, 90, 1007-1023.
24 42
Parthiban, S.; Martin, J. M. L. Assessment of W1 and W2 Theories for the Computation of Electron Affinities, Ionization
25 Potentials, Heats of Formation, and Proton Affinities J. Chem. Phys. 2001, 114, 6014-6029.
43
26 Helgaker, T.; Klopper, W.; Koch, H.; Noga, J. Basis-set Convergence of Correlated Calculations on Water J. Chem. Phys.
27 1997, 106, 9639-9646.
44
28 Stanton, J. F.; Gauss, J.; Perera, A.; Yau, A.; Watts, J. D.; Nooijen, M.; Oliphant, N.; Szalay, P. G.; Lauderdale, W. J.;
Gwaltney, S. R. et al. ACESII is a product of the Quantum Theory Project, University of Florida. Integral packages included
29
are VMOL (Almlöf, J.; Taylor, P. R.), VPROPS (Taylor, P. R.) and ABACUS (Helgaker, T.; Jensen, H. J. A.; Jørgensen,
30 P.; Olsen, J.; Taylor, P. R.).
31 45
Lotrich, V.; Flocke, N.; Ponton, M.; Yau, A.; Perera, A.; Deumens, E.; Bartlett, R.J. Parallel Implementation of Electronic
32 Structure Energy, Gradient, and Hessian Calculations J. Chem. Phys. 2008, 128, 194104-194115.
46
33 T. Helgaker, P. Jrgensen, and J. Olsen, Molecular Electronic-Structure Theory (John Wiley and Sons, Chichester 2000),
34 Chap. 8.
47
35 Ramalho, S. S.; Barreto, P. R. P.; Vilela, A. F. A.; Gargano, R. Theoretical Rate Constants for the Reaction BF2+NF =
36 BF3+N of Importance in Boron Nitride Chemistry Chem. Phys. Lett. 2005, 413, 151-156.
48
37 Cremer, D; Kraka, E; He, Y. Exact Geometries from Quantum Chemical Calculations Jour. Molec. Struct. 2001, 567,
275-293.
38 49
Bettendorff, M.; Peyerimhoff, S. D. Electronic Structure of the Radicals NF and NCI. I. Potential Energy Curves for NF
39 Chem. Phys. 1985, 99, 55-72.
40 50
Radzig, A. A.; Smirnov, B. M. Reference Data on Atoms, Molecules and Ions, Springer-Verlag, Berlin, 1985.
41 51
Harmony, M. D.; Myers, R.; Schoen, L. J.; Lide, D. R.; Mann, D. E. Infrared Spectrum and Structure of the NF2 Radical
42 J. Chem. Phys. 1961, 35, 1129-1130.
52
43 Barreto, P. R. P.; Vilela, A. F. A.; Gargano, R. Thermochemistry of Molecule in the B/F/H/N System Int. J. Quant. Chem.
44 2005, 103, 659-684.
53
45 Lascola, R.; Withnall, R.; Andrews, L. Infrared Spectra of Fluoroimide, Nitrogen Trifluoride, Phosphorus Trifluoride, and
Phosphorus Trichloride and Complexes with Hydrogen Fluoride in Solid Argon J. Phys. Chem. 1988, 92, 2145-2149.
46 54
McQuarrie, D. A. Statistical Mechanics. University Science Books, Second Review Edition, 2000.
47 55
Schaftenaar, G.; Noordik, J. H. Molden: a Pre- and Post-processing Program for Molecular and Electronic StructuresJ.
48 Comput.-Aided Mol. Design 2000, 14, 123-134.
49 56
Baulch, D. L.; Duxbury, J.; Grant, S. J.; Montague, D. C. Evaluated Kinetic Data for High Temperature Reactions J. Phys.
50 Chem. Ref. Data. 1981, 4, 576-635.
57
51 Tsckuikow-Roux, E.; MacFadden, K. O.; Jung, K. H.; Armstrong, D. A. Kinetics of the Thermal Dissociation of Tetrafluo-
52 rohydrazine J. Phys. Chem. 1973, 77, 734-742.
58
53 Brus, L. E.; Lin, M. C. Chemical Hydrogen Fluoride Lasers from Flash Photolysis of Various N2F4+RH SystemsJ. Phys.
54 Chem. 1971, 75, 2546-2550.
59
Padrick, T. D.; Gusinow, M. A. Evidence For Rotational Non-Equilibrium In The CH3I + N2F4 Chemical Laser Chem.
55
Phys. Lett. 1974, 24, 270-274.
56 60
Weiller, B. H.; Heidner, R. F.; Holloway, J. S.; Koffend, J. B. Kinetics of NF: Removal Rate Constants for NF(a1 ∆) and
57 NF(X 3 Σ− ) J. Phys. Chem. 1992, 96, 9321-9328.
58 61
Cheah, C. T.; Clyne, M. A. A.; Whitefield, P. D. Reactions Forming Electronically-excited Free Radicals. Part 1.Ground-state
59
60

ACS Paragon Plus Environment

Page 13 of 19 The Journal of Physical Chemistry
13
1
2 Reactions Involving NF2 and NF Radicals J. Chem. Soc. Faraday Trans. II 1980, 76, 711-728.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 14 of 19
14
1
2 Figures
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56 FIG. 1: Mechanism used to describe the abstraction and dissociation reactions.
57
58
59
60

ACS Paragon Plus Environment

Page 15 of 19 The Journal of Physical Chemistry
15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53 FIG. 2: Conventional (TST), Wigner (κW ), Eckart (κE ), Bell (κB0 , κB1 , and κB2 ), deformed transition state theory (d-TST)
54 and experimental61 plots of TRC against the reciprocal temperature in the range of 200-4000K for the NF2 +N→2NF reaction.
55 The purple and black arrows define the crossover temperature, Tc , and the validity temperature, Td , respectively (see Sec. II).
56
57
58
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 16 of 19
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
FIG. 3: Conventional (TST), Wigner (κW ), Eckart (κE ), Bell (κB0 , κB1 , and κB2 ), deformed transition state theory (d-TST)
26 and experimental60 plots of TRC against the reciprocal temperature in the range of 200-4000K for the NF3 +NF→2NF2 reaction.
27 The purple and black arrows define the crossover temperature, Tc , and the validity temperature, Td , respectively (see Sec. II).
28 The blue circle represent thermal rate constants measured in a specific excited-state of NF.
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

ACS Paragon Plus Environment

Page 17 of 19 The Journal of Physical Chemistry
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53 FIG. 4: Conventional (TST), Wigner (κW ), Eckart (κE ), Bell (κB0 , κB1 , and κB2 ), deformed transition state theory (d-TST)
54 and experimental56 plots of TRC against the reciprocal temperature in the range of 200-4000K for the N2 F4 →2NF2 reaction.
55 The purple and black arrows define the crossover temperature, Tc , and the validity temperature, Td , respectively (see Sec. II).
56
57
58
59
60

ACS Paragon Plus Environment

 The Journal of Physical Chemistry Page 18 of 19
18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53 FIG. 5: Conventional (TST), Wigner (κW ), Bell (κB0 , κB1 , and κB2 ), and deformed transition state theory (d-TST) theory
54 plots of TRC against the reciprocal temperature in the range of 200-4000K for the N2 F3 →NF2 +NF reaction. The purple and
55 black arrows define the crossover temperature, Tc , and the validity temperature, Td , respectively (see Sec. II).
56
57
58
59
60

ACS Paragon Plus Environment

Journal
Page 19of of
Physical
19 Chemi

1
2
3
4
5
6
7
8
9
10
11
12
S 13
Paragon Plus Environme
14
15
16