Cite this:
Phys. Chem. Chem. Phys
., 2011, 13, 8515–8525
Exploring the dynamics of reaction N(
2
D)+C
2
H
4
with crossed
molecular-beam experiments and quantum-chemical calculations
Shih-Huang Lee,*
abChih-Hao Chin,
aWei-Kan Chen,
aWen-Jian Huang
aand
Chu-Chun Hsieh
abReceived 7th November 2010, Accepted 4th March 2011 DOI: 10.1039/c0cp02439b
We conducted the title reaction using a crossed molecular-beam apparatus, quantum-chemical calculations, and RRKM calculations. Synchrotron radiation from an undulator served to ionize selectively reaction products by advantage of negligibly small dissociative ionization. We observed two products with gross formula C2H3N and C2H2N associated with loss of one and two
hydrogen atoms, respectively. Measurements of kinetic-energy distributions, angular distributions, low-resolution photoionization spectra, and branching ratios of the two products were carried out. Furthermore, we evaluated total branching ratios of various exit channels using RRKM calculations based on the potential-energy surface of reaction N(2D)+C2H4established with the
method CCSD(T)/6-311+G(3df,2p)//B3LYP/6-311G(d,p)+ZPE[B3LYP/6-311G(d,p)]. The combination of experimental and computational results allows us to reveal the reaction dynamics. The N(2D) atom adds to the CQC p-bond of ethene (C2H4) to form a cyclic complex
c-CH2(N)CH2that directly ejects a hydrogen atom or rearranges to other intermediates followed
by elimination of a hydrogen atom to produce C2H3N; c-CH2(N)CH+H is the dominant product
channel. Subsequently, most C2H3N radicals, notably c-CH2(N)CH, further decompose to
CH2CN+H. This work provides results and explanations different from the previous work of
Balucani et al. [J. Phys. Chem. A, 2000, 104, 5655], indicating that selective photoionization with synchrotron radiation as an ionization source is a good choice in chemical dynamics research.
I.
Introduction
Dinitrogen (N2) is the most abundant molecule in the
atmosphere but too inert to react with other atmospheric molecules due to the strong NRN triple bond. In contrast, atomic nitrogen, particularly in an electronic excited state, could react with some atmospheric molecules. The chemistry of nitrogen atoms plays an important role in the atmosphere because nitrogen atoms can be produced from nitrogen-containing molecules, notably N2, by cosmic radiation and
lightning. In laboratories nitrogen atoms are producible typically from N2by various discharge approaches.
Atomic nitrogen with an electronic configuration of 1s22s22p3 has three lowest-lying electronic states4S, 2D and
2
P. State2D lies 2.38 eV and 2P 3.576 eV above the ground state4S.1Nitrogen atoms in state2D are believed to be more
reactive than in states 4S and2P based on measurements of
rate coefficients of reactions of nitrogen atoms with various
molecules.2 A nitrogen atom in state 2D has a radiative lifetime of 17 h,1which makes feasible the reaction of atomic N(2D) with other molecules in experiments. Thus, most studies on reaction dynamics of nitrogen atoms were devoted to the reactions of2D nitrogen atoms with H2/D2,3–5H2O,6CH4,7–9
SiH4,10,11C2H2,12,13and C2H4.14–16Reactions started with an
insertion mechanism of N(2D) atoms into bonds H–H, CH3–H
and SiH3–H to form reaction complexes H–N–H, CH3–N–H
and SiH3–N–H, as well as an addition mechanism to the
p-bond of CHRCH and CH2QCH2to form cyclic complexes c-CH(N)CH and c-CH2(N)CH2, respectively. The complexes
underwent either direct decomposition or various isomerization processes followed by decomposition. Two product channels SiNH2+2H and SiNH+H2+H were reported for the
N(2D)+SiH
4reaction whereas only single channels with the
elimination of a hydrogen atom were reported for the reactions of N(2D) with H2, CH4, C2H2and C2H4.
As for the title reaction N(2D)+C2H4, a part of the doublet
potential-energy surface (PES) was established by Takayanagi et al. using the method PMP4(SDTQ)/cc-pVTZ//MP2/ cc-pVDZ.14The calculations predicted that the main reaction
mechanism is addition of atomic N(2D) onto the CQC p-bond of C2H4 to form a three-membered cyclic complex
c-CH2(N)CH2that undergoes various pathways for elimination
a
National Synchrotron Radiation Research Center (NSRRC), 101 Hsin-Ann Road, Hsinchu Science Park, Hsinchu 30076, Taiwan
b
Department of Applied Chemistry, National Chiao-Tung University, 1001 University Road, Hsinchu 30010, Taiwan.
E-mail: shlee@nsrrc.org.tw; Fax: +886-3-578-3813; Tel: +886-3-578-0281
www.rsc.org/pccp
PAPER
of hydrogen atoms to form various C2H3N isomers. The
authors calculated product branching ratios of c-CH2(N)CH2
with internal energy 104.3 kcal mol1, i.e., zero collision energy (Ec) for reactant N(2D)+C2H4, using Rice–Ramsperger–
Kassel–Marcus (RRKM) theory based on the established ab initioPES.14The calculations predicted that c-CH2(N)CH+H
is the dominant channel with a branching ratio of 0.848. Thermal rate coefficients of reaction N(2D)+C2H4 in the
temperature range 225–292 K were measured and expressible as 2.3 1010eEa/kT cm3molecule1s1; activation energy
Eaequals 1.0 0.1 kcal mol1, indicating a small barrier in the
entrance channel.15
Balucani et al. conducted the title reaction at Ec = 8 kcal mol1in a crossed molecular-beam apparatus
equipped with an electron-impact ionizer.16Product ions were observed at mass-to-charge ratios (m/z) 38–41 u. Because the signals at those m/z values have similar TOF spectra and laboratory angular distributions, the ions with m/z = 41 u were assigned to products with gross formula C2H3N and
those ions with m/z = 38–40 u to daughter ions of C2H3N due
to severe dissociative ionization. The authors employed the data recorded at m/z = 40 u for analysis because of large ion signals. The authors, however, might have ignored the possibility that the ions with m/z = 41 u are the isotopic variants of product C2H2N with loss of two hydrogen atoms,
since the signal of m/z = 40 u is much larger than that of m/z = 41 u. If so, the signal of C2H2N isotopmers might overlap
with the signal of product C2H3N at m/z = 41 u. Selective
photoionization17can give a solution to this ambiguity. The merits of photoionization with synchrotron vacuum-ultraviolet (VUV) radiation as an ionization source in studies of unimolecular photodissociation17–19 and atom–molecule chemical reactions11,20,21have been described elsewhere. The advantage of negligibly small dissociative photoionization enabled the direct detection of reaction products C2H3N and
C2H2N which were indistinguishable using electron-impact
ionization. The present work corrected the previous experi-mental results16 conducted with electron-impact ionization and explored more deeply the dynamics of the reaction of N(2D) with ethene (C2H4). To support the experimental
observations, a comprehensive PES was established and RRKM calculations were also performed in the present work.
II.
Experiments
The experimental apparatus and procedure have been described elsewhere;11,20,21 thus, only a brief description is given here. The crossed molecular-beam apparatus comprised two source chambers, a reaction chamber and a detection chamber. One source chamber equipped with an Even–Lavie valve and a discharge device22 served to generate a pulsed beam of nitrogen atoms. A mixture of 3% N2and 97% He was
supersonically expanded from the Even–Lavie valve with a backing pressure of 104 psi. A discharge was synchronously ignited while a molecular nitrogen pulse passed through the discharge device driven with a high-voltage pulse of 1 kV and 10 ms. The N-atom beam was collimated with two successive skimmers. In the other source chamber, an Even–Lavie valve heated up to 110 1C served to expand
supersonically neat ethene at a stagnation pressure of 54 psi. The pulsed molecular ethene beam was collimated with a skimmer. An L-shaped copper panel installed near the reaction region was chilled to 14 K to diminish the background pressure in the reaction region. The cold panel had a hole of diameter 3 mm in each side for penetration of both atomic and molecular beams that intercepted at 901 to each other in the reaction chamber. Nitrogen atoms and ethene had mean velocities of 1750 and 890 m s1, respectively, giving a collision energy of 4.3 kcal mol1. Reaction products were scattered into a whole solid angle but only the products flying towards an ion detector were ionized with undulator radiation after a free flight along a path of length 100.5 mm. The undulator radiation had photons with harmonic frequencies in addition to the desired photons with the fundamental frequency. A windowless gas cell installed on the beam line was filled with noble gas to absorb high harmonic photons. An additional optical filter of MgF2 served to effectively suppress high
harmonic photons when the energy of fundamental-frequency photons was less than 10 eV. The filtered VUV beam was focused into a size of diameter B1 mm in the ionization region. The flux wasB1 1016photons s1and the energy resolution (DE/E) wasB4%. The photon energy was tunable on adjusting the gap of the undulator for the purpose of selective ionization. Ion optics extracted product cations into a quadrupole mass filter for selection of species with a specific m/z ratio. An ion detector of Daly type counted the selected cations and a multi-channel scaler (MCS) sampled ions into 4000 bins of 1 ms. Two trigger-pulse generators operating at 200 Hz synchronized the experimental components. A time-of-flight (TOF) spectrum of a product can be obtained after subtracting the ion flight interval from the total flight duration and a background from the raw signal if necessary. To obtain product TOF spectra at various laboratory scattering angles (Y) between the N-atom beam and the detection axis, the source-chamber assembly was rotated from Y =201 to 1101.
III.
Computations
The computational details have been described elsewhere;11,20,23 thus, only a brief description is given here. We established a comprehensive potential-energy surface for the N(2D)+C2H4
reaction using the method CCSD(T)/6-311+G(3df,2p)// B3LYP/6-311G(d,p)+ZPE[B3LYP/6-311G(d,p)]. Computations were conducted using a suite of programs Gaussian-03 in a computer equipped with four CPUs and 16 GB of memory. Structures of molecules at stationary and transition states (TS) were optimized at the level of B3LYP/6-311G(d,p); harmonic vibrational frequencies and zero-point energies (ZPE) of molecules at optimized structures were computed also at the same level of theory. In addition to the examination of the number of imaginary vibrational frequency, the calculation of intrinsic reaction coordinate (IRC) served to verify the connection of a transition structure with its reactant and product. Total energies of atomic and molecular species were computed at the level of CCSD(T)/6-311+G(3df,2p). The same computational method was applied also to molecular cations for the calculations of adiabatic ionization energy (IE). We calculated rate coefficients of individual reaction steps
for the multi-channel dissociation of reaction complex c-CH2(N)CH2with Ec= 5 kcal mol1using RRKM theory24–26
and variational transition state theory (VTST).26,27 RRKM theory was applicable to a reaction path with a transition state whereas VTST theory to a barrierless reaction path. The microcanonical rate coefficient k(E) of a reaction step can be expressed as kðEÞ ¼s
h
WaðEEaÞ
rðEÞ using RRKM theory; s is a
symmetry factor (or reaction path degeneracy), h is the Plank constant, Wa(E Ea) represents the number of states accessible in a transition structure with a barrier height Ea, and r(E) denotes the density of states of a reactant with internal energy E in a reaction step. The values of Wa(E Ea) and r(E) were estimated using a direct-count method based on the computed harmonic vibrational frequencies. In the VTST calculations, we calculated k(E) as a function of reaction coordinate qa assuming each molecular structure optimized along qa is a pseudo transition state. A minimal rate coefficient could be found along qa according to the equation dkðE;qdqaaÞ¼ 0. Kinetic master equations d½Cdti¼
P kn½Cj
P
km½Ci were constructed and solved using
stationary-state approximation, where [C]i and [C]j are
concentrations of species i and j at time t as well as knand
km are microcanonical rate coefficients computed with the
RRKM or VTST approach. At t - N, product branching ratios can be yielded.
IV.
Results and analysis
Different from the work of Balucani et al.,16we observed two
reaction products with gross formula C2H3N at m/z = 41 u
and C2H2N at 40 u. To verify our assignments, Fig. 1 presents
the TOF spectra of both species with m/z = 41 and 40 u recorded at Y = 441 with photoionization energies at 10.8, 11.7 and 15.1 eV. The species of m/z = 41 u has two features peaking around 90 and 160 ms but the species of m/z = 40 u has a single feature peaking around 100 ms in TOF distributions. The distinct TOF distributions indicate the presence of two reaction products; the two features of m/z = 41 u are ascribed to product C2H3N and the single feature of m/z = 40 u to
product C2H2N. The species with m/z = 40 u (C2H2N) has a
TOF distribution insensitive to photoionization energy in the range of 10.8–15.1 eV, indicating that the dissociative ionization from product C2H3N to C2H2N+is negligible compared with
the ion signal of product C2H2N in this energy range. In
contrast, the TOF distribution of species with m/z = 41 u varies with photoionization energy; the TOF distribution has two apparent features with photon energies at 10.8 and 11.7 eV but becomes similar to that of m/z = 40 u at 15.1 eV. The variation of TOF distributions is attributed to the contribution of isotopic variants of product C2H2N to the signal of m/z =
41 u. Based on the natural abundances 1.1 102 of 13
C, 3.66 103of15N and 1.5 104of2H atoms,28the isotopic ratio of 41 u to 40 u of C2H2N is expected to be about 0.026
because C2H2N has two carbon, two hydrogen and one
nitrogen atoms. Fig. 1 shows that the peak signals of m/z = 40 u (C2H2N) recorded with photoionization energies 10.8,
11.7 and 15.1 eV are about 5, 7.4 and 20 counts per 103pulses, respectively. Therefore, isotopic variants of C2H2N will have
about 0.13, 0.19 and 0.52 counts per 103pulses contributing to the peak signals of m/z = 41 u recorded with photons at 10.8, 11.7 and 15.1 eV, respectively. Because product C2H2N has a
yield much greater and an ionization threshold higher than that of product C2H3N (vide infra), the isotopic contribution
of C2H2N to m/z = 41 u becomes more significant when
photoionization energy is much larger than the ionization thresholdB10 eV of C2H2N.29The TOF distributions of both
species of m/z = 40 and 41 u become similar in appearance with photoionization energy larger than 15 eV, which accounts for the observation of Balucani et al. using electron-impact ionization.
Fig. 2 exhibits Newton diagrams associated with two-dimensional product velocity contours for the reactions of N(2D)+C2H4 to C2H3N+H and to C2H2N+2H. In the
laboratory frame, the flight direction of nitrogen atoms is defined as Y = 01 and of ethene as Y = 901. YcmD 461 is the
flight direction of the center of masses (c.m.) of the reaction system. In the c.m. frame, the incident direction of nitrogen atoms is defined as y = 01 and the opposite direction, i.e., the incident direction of ethene, as y = 1801. The scattering directions of reaction products C2H3N and C2H2N are called
forward, sideways and backward for y scanning from 01 through 901 to 1801. A simulation program served to mimic TOF spectra of products using forward convolution of initial guess c.m. kinetic-energy and angular distributions with experimental parameters. The transformation of signals from the laboratory frame to the c.m. frame was detailed in ref. 11. From the best fit to the experimental TOF spectra detected at various laboratory angles, angle-specific kinetic-energy distributions P(Et; y) and
an angular distribution P(y) of products in the c.m. frame are derivable. Et denotes the total kinetic energy of two Fig. 1 TOF spectra of products with m/z = 41 u and 40 u recorded at
Y = 44owith photoionization energy 10.8, 11.7, and 15.1 eV. Left (right) ordinates show the relative ion signals of products with m/z = 41 (40) u normalized to 103 pulses. Each panel shows the
photoionization energy employed.
momentum-matched products. The kinetic-energy distribution at any y value is derivable by interpolation.
Fig. 3 presents the TOF spectra of species with m/z = 41 u from Y = 151 to 701 using photoionization energies 9.6 eV and 11.7 eV which are below and above, respectively, the ionization thresholdB10 eV of product C2H2N.
29
The data with Y larger than 701 incur interference of impurities in the ethene beam and thus are omitted here. With photoionization energy 9.6 eV, the feature along with red-line simulation is assigned to product C2H3N with loss of a hydrogen atom.
Since C2H3N has a small velocity in the c.m. frame, the C2H3N
products recoiled forward and backward with respect to the c.m. flying direction are observable in the laboratory frame and thus the TOF distributions of C2H3N near Ycm are
bimodal. The isotopic variants of product C2H2N (blue line)
that has an ionization threshold around 10 eV appear in the case of 11.7 eV. The blue components were simulated with the same kinetic-energy and angular distributions derived from the TOF spectra of product C2H2N (vide infra). The c.m. ratio
of the red to the blue component was determined to be 0.89 : 0.11 from the best global fit to the experimental TOF spectra recorded at 11.7 eV. Fig. 4 exhibits P(Et) distributions
at seven y angles and the P(y) distribution derived from the best simulation as shown in Fig. 3. Fig. 5 presents the low-resolution photoionization spectrum of species recorded at m/z = 41 u; ion signals were detected at Y = 441 and integrated from 50 ms to 250 ms in the flight time. The photoionization
spectrum was divided into two curves; the blue line denotes the photoionization spectrum of isotopic variants of product C2H2N
(vide infra) and the red line denotes the photoionization spectrum of product C2H3N. The ratio of the red to blue curve is 0.77 : 0.23
at 11.7 eV determined from the partitioning of TOF spectra of m/z = 41 u recorded with photoionization energy 11.7 eV. Arrows indicate the adiabatic ionization energies of isomers CH2NCH and c-CH2(N)CH.
Fig. 6 presents the TOF spectra of species with m/z = 40 u from Y = 151 to 701 using photoionization energy 11.7 eV. The feature along with simulation is assigned to product C2H2N with loss of two hydrogen atoms. Because hydrogen
atoms are too elusive to be detected in the present work, we assume a two-body dissociation process with a product mass ratio of 40 : 2 to simulate the TOF spectra of product C2H2N.
Fig. 7 exhibits P(Et) distributions at seven y angles and the
P(y) distribution derived from the best simulation as shown in Fig. 6; the P(Et) and P(y) distributions were employed also to
simulate the blue components of m/z = 40 u shown in Fig. 3. Fig. 8 presents the low-resolution photoionization spectrum of product C2H2N; ion signals were detected at Y = 441 and
integrated from 60 ms to 240 ms in the flight time. Arrows indicate the adiabatic ionization energies of isomers CH2NC
and CH2CN. The combination of product TOF spectra
(Fig. 1, 3 and 6) and product photoionization spectra (Fig. 5 and 8) clearly indicates that there are two distinct products with gross formula C2H3N and C2H2N observed in the present
work. Fig. 9 presents the doublet PES established for the N(2D)+C2H4 reaction including various exit channels.
The green and red paths are not calculated in the work of Takayanagi et al. Thick solid lines denote some likely pathways leading to products C2H3N+H and C2H2N+2H.
CH3NCH, CH2CHNH and CH3CNH have symmetry Csand
thus have cis- and trans-conformers according to the geometry of skeletons CNCH, CCNH, and CCNH, respectively, as shown in Fig. 9.
We obtained average kinetic-energy releasehEti and fraction
ft of available energy (Eava) in translation based on
hEti =
R R
Et P(Et; y) dEtd(cos y)/
R R
P(Et;y)dEt d(cos y) and
ft=hEti/Eava. Table 1 lists the values of Eava,hEti, and ftfor
product channels C2H3N+H and C2H2N+2H. The Eavavalues
of the most-probable isomeric channels c-CH2(N)CH+H and
CH2CN+2H were adopted for the calculations ofhEti and ft.
Table 2 summarizes the adiabatic ionization energies of six isomers of C2H3N and three isomers of C2H2N. The difference
between calculated and experimental IE values is within0.3 eV for CH3CN, CH3NC, c-CH2(N)CH and CH2CN; thus, the other
calculated IE values might be reliable. Table 3 summarizes total branching ratios of various two-body dissociation channels predicted by RRKM and VTST calculations with Ec = 5 kcal mol1; branching of secondary dissociation is
beyond the present calculations.
V.
Discussion
Ignoring the initial angular momenta of reactants supersonically cooled, the total angular momentum J* of a reaction system equals approximately the orbital angular momentum L* of a two-particle collision system, i.e., J*ffi L*¼ mb* u*rel.30 m is a Fig. 2 Newton diagrams and two-dimensional product velocity contours
for the reactions of N(2D)+C
2H4- C2H3N+H and N(2D)+C2H4
-C2H2N+2H. Dashed lines denote the detection axes at laboratory angles
151–701.
reduced mass, b*is an impact parameter, and u*relis the relative
velocity between two colliding reactants. Parameter b* is perpendicular to and randomly polarized about axis u*rel.
The opacity function P(b), a reaction probability as a function of b, governs the dynamics of a two-particle collision.30The
spatial distribution of reaction products has a symmetric axis along vector u*rel in the c.m. frame. For a direct reaction
with a transient complex, the product angular distribution is
dynamically controlled and typically forward–backward asymmetric.30 In contrast, for a reaction with a persisting (long-lived) complex the original memory in dynamics becomes obscure and the angular distribution approaches forward-backward symmetric due to rotation of the reaction complex before decomposition.30,31
Fig. 3 Angle-specific TOF spectra of product C2H3N recorded at m/z = 41 u with photoionization energies 9.6 eV and 11.7 eV which are below
and above, respectively, the ionization thresholdB10 eV of product C2H2N. Open circles denote the experimental data and solid curves denote the
simulations. Each panel shows the corresponding laboratory angle Y. Only C2H3N (red curve) is observable with photon energy at 9.6 eV whereas
the isotopic variant of C2H2N (blue curve) additionally appears with photon energy at 11.7 eV. Black curves in the case of 11.7 eV are the sums of
red and blue curves. Fig. 4 and 7 present the corresponding kinetic-energy and angular distributions for the red and blue components, separately.
Fig. 4 Angle-specific distributions of kinetic energy and the angular distribution of product C2H3N.
Fig. 5 Low-resolution photoionization spectrum of product C2H3N
recorded at m/z = 41 u. Ion signals were detected at Y = 441 and integrated from 50 ms to 250 ms in the flight time. The step increment is ca. 0.2 eV. Black open circles denote the experimental data. Blue circles represent the photoionization spectrum of product C2H2N
same as that shown in Fig. 8 with a weighting according to the isotopic natural abundance. Red circles, the differences between the black and blue circles, represent the photoionization spectrum of product C2H3N. The ratio of the red to blue circle at 11.7 eV is 0.77 : 0.23
in the laboratory frame. Arrows indicate the calculated adiabatic ionization energies of CH2NCH and c-CH2(N)CH.
Since the reactivity of nitrogen atoms in states4S and2P are typically much smaller than in state2D and the population of nitrogen atoms in state2P is less than in state2D,11reactions of N(4S) and N(2P) atoms with ethene are neglected here.
Although the rate coefficient of N(2P) was close to N(2D)
reactions with C2H4, the deactivation process of N(2P) was
determined to be the spin-allowed quenching process N(2P)+C2H4(S0) - N(
4
S)+C2H4(T1). 15
The N(2D)+C2H4
reaction starts with the addition of atomic N(2D) to the CQC p-bond of ethene to form c-CH2(N)CH2 through a small
barrier at the entrance channel. Sato et al. reported an activation energy Ea= 1.0 0.1 kcal mol1for this reaction.15Takayanagi
et al.reported a classical barrier height 3.2 kcal mol1for this addition process using a computational method CASSCF/ cc-pVDZ; the barrier height was reduced to 1.6 kcal mol1 with a correction of CASMP2 calculations but the location of the potential maximum shifted toward the reactant side.14The
reaction complex c-CH2(N)CH2 subsequently undergoes
either direct dissociation or a series of isomerization processes followed by decomposition mainly to C2H3N+H and
C2H2N+2H at exit channels (vide infra). The mechanism of
atomic N(2D) inserting into bond C–H of ethene is negligible because of a significant entrance barrier that is 13 kcal mol1 at the CASSCF(5,5)/cc-pVDZ level of theory.14
A Reaction N(2D)+C2H4- C2H3N+H
Based on the PES shown in Fig. 9, the following lists some likely product channels for elimination of a hydrogen atom. N(2D) + C2H4- CH3CN + H DH1 =107.6 kcal mol1 (1) - CH3NC + H DH1 =84.2 kcal mol1 (2) - CH2CNH + H DH o =79.8 kcal mol1 (3) - c-CH2(N)CH + H DHo=60.2 kcal mol1 (4) - CH2NCH + H DHo=51.4 kcal mol1 (5) - c-CH(NH)CH + H DH1 =27.0 kcal mol1 (6) where, DH1 denotes a reaction enthalpy computed at 0 K. The numbering of reaction paths (or product channels) is based on the order of reaction enthalpy. Available energy Eava of a
reaction can be calculated based on Eava= Ec DH o
. The RRKM and VTST calculations predicted that reactions (1)–(6) have branching ratiosB0, 0.042, 0.023, 0.869, 0.044, andB0, respectively. Although reaction (1) produces the most stable isomer CH3CN (acetonitrile), the multiple isomerization
processes as shown in Fig. 9 results in almost no yield to this
Fig. 6 Angle-specific TOF spectra of product C2H2N recorded at
m/z = 40 u with photoionization energy 11.7 eV. Open circles denote the experimental data and solid curves denote the simulations. Each panel shows the corresponding laboratory angle Y.
Fig. 7 Angle-specific distributions of kinetic energy and the angular distribution of product C2H2N.
Fig. 8 Low-resolution photoionization spectrum of product C2H2N
recorded at m/z = 40 u. Ion signals were detected at Y = 441 and integrated from 60 ms to 240 ms in the flight time. The step increment is ca.0.2 eV. Arrows indicate the calculated adiabatic ionization energies of CH2NC and CH2CN.
product. Fig. 9 indicates that reaction complex c-CH2(N)CH2
might directly eject a hydrogen atom through TS10 to produce c-CH2(N)CH (2H-azirine) (reaction 4a). In addition, one of
hydrogen atoms of c-CH2(N)CH2might migrate to the nitrogen
atom to form c-CH2(NH)CH that subsequently ejects the
hydrogen atom through TS9 to produce c-CH2(N)CH
(reaction 4b). Reactions (4a) and (4b) were predicted to have branching ratios 0.819 and 0.05, respectively, by RRKM calculations. c-CH2(NH)CH can rupture the CH2–NH bond
to form cis-CH2CHNH that subsequently decomposes to
CH2CNH (ketene imine)+H through TS18 (reaction 3).
Because of large dissociation energy, the decomposition from
c-CH2(NH)CH to c-CH(NH)CH+H (reaction 6) was predicted
to be negligible by RRKM calculations. c-CH2(N)CH2can break
the C–C bond through TS8 to form planar intermediate CH2NCH2 that subsequently breaks one of bonds C–H to
produce CH2NCH+H (reaction 5). c-CH2(N)CH2can rearrange
also to cis-CH3NCH through TS12 followed by decomposition to
CH3NC (methyl isocyanide)+H (reaction 2). Takayanagi et al.
ignored reactions (2) and (5) that were predicted to have non-negligible branching ratios by RRKM calculations.
Fig. 9 Potential-energy surface for the reaction N(2D)+C2H4. The green and red paths are not calculated in the work of Takayanagi et al. Thick
solid lines denote some likely pathways leading to products C2H3N+H and C2H2N+2H.
Table 1 Available energy Eava, a
average kinetic-energy releasehEti,
and fractions of translational energy ftfor two product channels of
reaction N(2D)+C2H4at Ec= 4.3 kcal mol1
Product channels C2H3N+H C2H2N+2H
Eava/kcal mol1 64.5 18.2
hEti/kcal mol1 23.1 7.3b
ft(%) 35.8 40.1
a
The Eava values of the most-probable isomeric channels
c-CH2(N)CH+H and CH2CN+2H were adopted for the calculations
ofhEti and ft. b
ThehEti value of the channel C2H2N+2H was derived
based on an assumption of a two-body dissociation process with a product mass ratio of 40 : 2.
Table 2 Adiabatic ionization energies (IE) of various isomers of C2H3N and C2H2N
Species Theoretical IEa/eV Experimental IEb/eV CH3CN 12.5 12.20 0.01 CH3NC 11.2 11.53 0.04 c-CH2(N)CH 9.94 10.1 c-CH(NH)CH 8.12 N/A CH2CNH 8.64 N/A CH2NCH 7.84 N/A CH2CN 10.2 9.9 0.1 CH2NC 9.28 N/A CHNCH 8.66 N/A
aThe present values calculated at the level of CCSD(T)/
6-311+G(3df,2p)//B3LYP/6-311G(d,p)+ZPE[B3LYP/6-311G(d,P)].
b
Quoted from NIST Chemistry WebBook: http://webbook.nist.gov/ chemistry/.
The maximal kinetic-energy release is in good agreement with the energetic limit of reaction (4) and the ionization threshold of C2H3N is in good agreement with the IE of
CH2NCH from reaction (5). Thus, isomers c-CH2(N)CH
and CH2NCH might have contributions to product C2H3N
although the TOF distributions of product C2H3N shown as
the red curves in Fig. 3 have no significant change with the increase of photon energy from 9.6 eV to 11.7 eV. Since atomic hydrogen fragment carries no internal energy, the distribution of internal energy (Eint) of product C2H3N can
be derived directly based on P(Eint) = P(Eava Et).
c-CH2(N)CH, if produced alone, will have average internal
energy 41.4 kcal mol1 (1.80 eV), which might cause a red shift in the ionization threshold to some extent. c-CH2(N)CH2 was calculated to have moments of inertia
IA= 19.130 uA˚2, IB= 23.107 uA˚2, and IC= 35.506 uA˚2.
The present RRKM calculations indicate that reaction complex c-CH2(N)CH2has a decay rateB4.52 10
12
s1 corresponding to a lifetime of 0.22 ps that is shorter than rotational periods 13.4, 16.2 and 24.8 ps of c-CH2(N)CH2
with L = 1 along principal axes a, b and c, respectively. In the theoretical study of reaction N(2D)+H2- NH+H, the
angular distribution of NH was nearly forward–backward symmetric with a slight bias towards backward scattering at low collision energy.4 Deviations from symmetry arose
when small impact parameters (small L values) were favored, as then reactions favored backward scattering in hydrogen-atom elimination.4 This argument is applicable also to the present result of backward-biased angular distribution. TS10 in the dominant reaction path (4a) has the leaving hydrogen atom to be recoiled preferentially into the backward hemisphere. Rotational periods become 0.40, 0.48 and 0.74 ps for L = 47 along principal axes a, b and c, respectively; this L value is calculated with b = 1.636 A˚ estimated from the equilibrium structure of c-CH2(N)CH2.
Thus, products could be scattered into the forward and backward hemispheres for large L values. The angular distribution behaves forward-backward (or sideways) peaking when products are recoiled preferentially perpendicular
(or parallel) to L* based on a statistical model.31The angular
distribution behaves nearly isotropic when products are recoiled into a wide angular range with respect to vector L*.31 B Reaction N(2D)+C2H4- C2H2N+2H
The contribution from internally-excited C2H3N (e.g.,
c-CH2(N)CH) to C2H2N +
by dissociative photoionization is negligible herein based on the following reasons. First, the enthalpy of reaction c-CH2(N)CH- CH2CN++H+ewas
calculated to be 12.2 eV; thus, product c-CH2(N)CH requires
internal energy at least 50.7 kcal mol1 (2.2 eV), which is larger than the barrier height 43.6–43.9 kcal mol1of isomer-ization and the enthalpy 46.3 kcal mol1 of dissociation c-CH2(N)CH - CH2CN+H, to undergo this dissociative
ionization process with photon energy 10 eV. Second, Fig. 1 indicates that the species with m/z = 40 u has a TOF distribution insensitive to the photoionization energy from 15.1 to 10.8 eV (even down to 10.2 eV but not shown therein) and quite different from that of the species with m/z = 41 u. Third, the bimodal feature of m/z = 41 u in TOF distributions at scattering angles near Ycmremains as the photoionization
energy is below the appearance threshold 10 eV of species with m/z = 40 u; Fig. 3 presents the TOF spectra of m/z = 41 u recorded at 9.6 eV. Fourth, Fig. 8 indicates that the appearance threshold of C2H2N+is in good agreement with the ionization
energy of CH2CN. Accordingly, the signal observed at
m/z = 40 u is attributed mainly to product C2H2N rather
than a daughter ion of product C2H3N.
To account for the production of C2H2N+2H, we calculated
the following five isomeric channels as shown in Fig. 9. N(2D) + C2H4- CH2CN+2H DH1 =13.9 kcal mol1
(7)
- CH2NC + 2H DH1 = 8.9 kcal mol1 (8)
- HCCNH + 2H DH1 = 18.9 kcal mol1 (9)
- bent-CH2CN + 2H DH1 = 36.6 kcal mol1 (10)
- c-CH(N)CH + 2H DH1 = 40.5 kcal mol1 (11) bent-CH2CN and cyclic-CH(N)CH (c-CH(N)CH) can be
produced from successive elimination of two hydrogen atoms from the same and from different carbon atoms, respectively, of reaction complex c-CH2(N)CH2. Only reaction (7) is
energetically accessible in the present work. C2H2N can also
be produced by elimination of a hydrogen molecule as the following three reactions.
N(2D) + C
2H4- CH2CN + H2 DH1 =115.2 kcal mol1
(12) - CH2NC + H2 DH1 =92.4 kcal mol1 (13)
- CHNCH + H2 DH1 =53.0 kcal mol1 (14)
The production of C2H2N by elimination of a hydrogen
molecule cannot satisfactorily account for the experimental observations based on the following reasons. First, a reaction like (12)–(14) having a large exothermicity and a large exit barrier typically recoils products into large translational energy; however, Fig. 7 indicates that the maximal kinetic-energy
Table 3 Calculated product branching ratios for the decomposition of complex c-CH2(N)CH2from the N(
2
D)+C2H4reaction
Labels Products Branching ratiosa(%)
Branching ratiosb(%) 1 CH3CN+H 0.03 0.8 2 CH3NC+H 4.18 N/C 3 CH2CNH+H 2.28 13.2 4 c-CH2(N)CH+H 86.91 (= 81.93 c + 4.98d) 84.8 5 CH2NCH+H 4.40 N/C 6 c-CH(NH)CH+H B0 B0 12 CH2CN+H2 B0 B0 13 CH2NC+H2 0.21 N/C 14 CHNCH+H2 1.17 N/C 15 CH3+NCH 0.39 N/C 16 CH3+HNC 0.09 1.2 17 CH2N+CH2 0.34 B0 18 C2H3+NH B0 B0 a
Calculated at Ec= 5 kcal mol1in the present work. b
Calculated at Ec= 0 kcal mol1by Takayanagi et al. (ref. 14); N/C denotes ‘‘Not
Considered’’.cFor reaction (4a).dFor reaction (4b).
release is near the energetic limit of reaction (7). Second, if products from reactions (12)–(14) have small kinetic energy release like that in Fig. 7, C2H2N will be predicted to have a
large internal energy and have a considerable red shift in the ionization threshold; however, Fig. 8 indicates that the ionization threshold of C2H2N is in good agreement with the ionization
energy of CH2CN (cyanomethyl). Product CH2CN from
reaction (7) has an average internal energy merely 10.9 kcal mol1 (0.47 eV). Third, reactions (12)–(14) have negligible branching ratios based on RRKM calculations. Thus, reaction (7) is the most likely channel for the production of C2H2N.
Fig. 9 presents four pathways for the production of CH2CN+2H from the spontaneous dissociation of products
CH3CN, CH2CNH, and c-CH2(N)CH. CH3CN- CH2CN + H DH1 = 93.7 kcal mol1 (10) CH2CNH- CH2CN + H DH1 = 65.9 kcal mol1 (30) c-CH2(N)CH- CH3CN- CH2CN + H DH1 = 46.3 kcal mol1 (40) c-CH2(N)CH- CH2CNH- CH2CN + H DH1 = 46.3 kcal mol1 (40 0)
CH3CN and CH2CNH can decompose to CH2CN+H
with-out barriers. In contrast, c-CH2(N)CH needs isomerization to
CH3CN and CH2CNH through TS22 and TS23, respectively,
before decomposition to CH2CN+H. TS22 lies 2.4 kcal mol1
and TS23 2.7 kcal mol1below asymptote CH2CN+H, which
makes reactions (40) and (40 0), possible. We suggest that reactions (40) and (40 0) are more important than reactions
(10) and (30) since reaction (4) is much more dominant than
reactions (1) and (3) by RRKM calculations. This suggestion coincides with the experimental finding that the channel C2H2N+2H is about 5.7 times the branching ratio of the
channel C2H3N+H. CH2NCH can rearrange to c-CH2(N)CH
through TS24 and subsequently undergoes reactions (40) and
(40 0); this reaction path has a possible contribution only to the
products with Et less than 10.9 kcal mol1 because of the
large isomerization barrier (TS24) lying 7.3 kcal mol1above asymptote CH2CN+H.
C. Branching ratios
Branching ratios for channel C2H3N+H to channel
C2H2N+2H was estimated to be 15 : 85 from the c.m. ion
signals of products C2H3N and C2H2N recorded with
photo-ionizing energy 11.7 eV on assuming both products have the same detection efficiency. The contribution of isotopomers of C2H2N to the signal of m/z = 41 u was removed. It might not
be quantitatively accurate enough in the present analysis without a calibration of detection efficiency (e.g. ionization cross sections) but the channel C2H2N+2H should be
doubtless greater than the channel C2H3N+H in branching.
RRKM calculations predict that reaction (4) is over-whelmingly dominant among two-body dissociation channels. Thus, most of nascent C2H3N products, notably c-CH2(N)CH,
are suggested to further decompose to CH2CN+H in order to
rationalize the experimental branching ratios. Other energetically accessible product channels unobserved in experiments are listed below. N(2D) + C2H4- CH3+ HCN DH1 =111.6 kcal mol1 (15) - CH3+ HNC DH1 =98.1 kcal mol1 (16) - CH2N + CH2 DH1 =29.0 kcal mol1 (17) - C2H3+ NH DH1 =22.4 kcal mol1 (18)
Although reactions (15) and (16) have large exothermicities, Fig. 9 indicates that the multi-isomerization processes with large barriers diminish the dissociation probability. Reactions (17) and (18) have small exothermicities and thus lose competition based on RRKM calculations.
D Comparison with previous works
In the experimental part, we explicitly observed two products C2H3N and C2H2N using selective photoionization. In contrast,
Balucani et al. claimed that they observed product C2H3N at
m/z = 40 u using electron-impact ionization.16Fig. 1 and 3 indicate that the contribution of isotopic variants of C2H2N to
m/z = 41 u increases and the TOF distribution of m/z = 41 u becomes similar to that of m/z = 40 u with the increase of photoionization energy, which accounts for the reason why Balucani et al. misrecognized the species of m/z = 40 u as daughter ions of the species with m/z = 41 u. In another words, what they reported are the results of product C2H2N
rather than C2H3N. On the other hand, the forward-biased
angular distribution reported by Balucani et al.16 at Ec =
8 kcal mol1differs from the present result having a slight bias to the backward part at Ec = 4.3 kcal mol1. The
collision-energy dependence of product angular distributions was observed also in the reaction of N(2D)+C2H2
-HCCN+H;13the angular distribution of HCCN was isotropic at Ec= 3.1 kcal mol1and became forward preferred at Ec=
9.5 kcal mol1. Balucani et al.13indicated that at low collision energy the lifetime of a persistent reaction complex (or inter-mediate) is longer than its rotational period but at high collision energy the lifetime of an osculating complex is comparable to its rotational period and an osculating model30,32 is applicable. Accordingly, the collision-energy effect can account for the difference between the present work and Balucani’s et al. in product angular distributions.
In the computational part, we established a more complete potential-energy surface than the work of Takayanagi et al.14 for the N(2D)+C2H4reaction. Takayanagi et al. ignored the
cis–trans isomerization processes of intermediates CH2CHNH
and CH3CNH, the dissociation pathways leading to
CH3NC+H (reaction 2), CH2NCH+H (reaction 5),
CH2NC+H2 (reaction 13), CHNCH+H2 (reaction 14) and
CH3+NCH (reaction 15), and the secondary dissociation
pathways leading to various isomers of C2H2N+2H. The
complements (color lines in Fig. 9) make a satisfactory explanation possible for the experimental observations as addressed in the above sections. For instance, the ionization threshold of product C2H3N is in good agreement with the IE
of CH2NCH, indicating that reaction (5), in addition to
reaction (4), might have a contribution to product C2H3N.
Furthermore, channel C2H2N+2H is about 5.7 times
the branching of channel C2H3N+H, indicating that most
of primary product c-CH2(N)CH from reaction (4) further
decomposes to CH2CN+H, consistent with
quantum-chemical calculations. In the part (black lines in Fig. 9) similar to the work of Takayanagi et al.,14 the energies of inter-mediates and products relative to reactant N(2D)+C2H4
differ from the present computational values within 3 kcal mol1except for intermediate CH
2CHNH (vide infra).
Some of the molecular structures computed in the present work differ from those reported by Takayanagi et al. mainly on molecular conformations. Fig. 10 presents the optimized structures of molecules situating on reaction path (3) leading to product CH2CNH+H computed with two
methods of B3LYP/6-311G(d,p) (upper path) and MP2/ cc-pVDZ (lower path). The molecular structures of B3LYP/ 6-311G(d,p) correspond to the stationary and transition states of reaction path (3) shown in Fig. 9. The connection of a transition structure with its reactant and product was verified with the calculation of IRC. On the other hand, the lower path of Fig. 10 shows the results of method MP2/cc-pVDZ same as those of Takayanagi et al. but with some corrections; the same notations as those of Takayanagi et al. were adopted in the lower path. For instance, c-CH2(NH)CH presented in the upper path of Fig. 10 has
a conformation with dihedral angle +HCNH = 127.71 but Takayanagi et al. reported another conformer M2 that has +HCNH = 15.01. M2 cannot correlate directly with TS1 without the two additional structures shown in a bracket. M30 and TS110 differ from M3 and TS11 reported
by Takayanagi et al. M30 (TS110) has a reflection operator bisecting the methyl group and has a point group Cs but
M3 (TS11) has a point group C1. Moreover, M3 is likely
an electronic excited state lying B17 kcal mol1 above cis-CH2CHNH.
VI.
Conclusions
This work explored the dynamics of reaction of atomic N in state 2D with ethene (C2H4) at collision energy
Ec = 4.3 kcal mol1 by interrogating product channels
C2H3N+H and C2H2N+2H. We measured kinetic-energy
distributions, angular distributions, and photoionization spectra of products C2H3N and C2H2N. Furthermore,
we established a comprehensive doublet PES for the N(2D)+C2H4 reaction and calculated branching ratios of
various product channels at Ec= 5 kcal mol1using RRKM
and VTST approaches. The combination of experimental and theoretical results unveils the reaction mechanisms. The N(2D) atom adds to the CQC p-bond of ethene to form a cyclic complex c-CH2(N)CH2 that subsequently undergoes various
dissociation pathways. c-CH2(N)CH2mainly ejects one of the
hydrogen atoms to produce c-CH2(N)CH+H with a branching
ratio 0.87 predicted by RRKM calculations. Most of c-CH2(N)CH further decompose to CH2CN+H following
isomerization to CH3CN or CH2CNH, which rationalizes
the experimental branching ratios of 15 : 85 for channel C2H3N+H to channel C2H2N+2H. The angular distribution
of C2H3N has a propensity to the backward part, implying
that c-CH2(N)CH2is a short-lived complex compared with its
rotational period. Measurements of kinetic-energy release and ionization thresholds allowed us to identify product C2H3N as
a mixture of c-CH2(N)CH and CH2NCH as well as product
C2H2N as CH2CN. The present work corrects the previous
work of Balucani et al., elucidating the merits of selective photoionization with synchrotron radiation as an ionization source.
Fig. 10 The stationary and transition structures of reaction complexes and intermediates situating on the reaction path (3) computed by two methods of B3LYP/6-311G(d,p) (upper) and MP2/cc-pVDZ (lower). The present notations are employed in the upper path but the notations of Takayanagi et al. (ref. 14) are employed in the lower path. Two structures in a bracket are ignored in the work of Takayanagi et al. M30(TS110) differs from M3 (TS11) computed by Takayanagi et al.
Acknowledgements
We thank National Synchrotron Radiation Research Center, Academia Sinica, and the National Science Council of Taiwan (Grants NSC97-2113-M-213-003-MY3 and NSC99-2811-M-213-003) for supports.
References
1 W. L. Wiese, M. W. Smith, and B. M. Glennon, Atomic Transition Probabilities, Natl. Bur. Stand. Ref. Data Ser., Natl. Bur. Stand. (U.S.), Circ. No. 4 (U.S. GPO, Washington, D. C., 1966). 2 Chemical kinetics data from NIST web page: http://kinetics.nist.
gov/kinetics/index.jsp.
3 H. Umemtot, T. Asai and Y. Kimura, J. Chem. Phys., 1997, 106, 4985.
4 L. A. Pederson, G. C. Schatz, T.-S. Ho, T. Hollebeek, H. Rabitz, L. B. Harding and G. Lendvay, J. Chem. Phys., 1999, 110, 9091.
5 N. Balucani, M. Alagia, L. Cartechini, P. Casavecchia, G. G. Volpi, L. A. Pederson and G. C. Schatz, J. Phys. Chem. A, 2001, 105, 2414.
6 H. Umemoto, T. Asai, H. Hashimoto and T. Nakae, J. Phys. Chem. A, 1999, 103, 700.
7 H. Umemoto, Y. Kimura and T. Asai, Chem. Phys. Lett., 1997, 264, 215.
8 H. Umemoto, T. Nakae, H. Hashimoto, K. Kongo and M. Kawasaki, J. Chem. Phys., 1998, 109, 5844.
9 Y. Kurosaki, T. Takayanagi, K. Sato and S. Tsunashima, J. Phys. Chem. A, 1998, 102, 254.
10 W.-K. Chen, I.-C. Lu, C. Chaudhuri, W.-J. Huang and S.-H. Lee, J. Phys. Chem. A, 2008, 112, 8479.
11 I.-C. Lu, W.-K. Chen, C. Chaudhuri, W.-J. Huang, J. J. Lin and S.-H. Lee, J. Chem. Phys., 2008, 129, 174304.
12 T. Takayanagi, Y. Kurosaki, K. Misawa, M. Sugiura, Y. Kobayashi, K. Sato and S. Tsunashima, J. Phys. Chem. A, 1998, 102, 6251.
13 N. Balucani, M. Alagia, L. Cartechini, P. Casavecchia, G. G. Volpi, K. Sato, T. Takayanagi and Y. Kurosaki, J. Am. Chem. Soc., 2000, 122, 4443.
14 T. Takayanagi, Y. Kurosaki, K. Sato and S. Tsunashima, J. Phys. Chem. A, 1998, 102, 10391.
15 K. Sato, K. Misawa, Y. Kobayashi, M. Matsui, S. Tsunashima, Y. Kurosaki and T. Takayanagi, J. Phys. Chem. A, 1999, 103, 8650.
16 N. Balucani, L. Cartechini, M. Alagia, P. Casavecchia and G. G. Volpi, J. Phys. Chem. A, 2000, 104, 5655.
17 S.-H. Lee, J. J. Lin and Y. T. Lee, J. Electron Spectrosc. Relat. Phenom., 2005, 144–147, 135.
18 S.-H. Lee, J. Chem. Phys., 2009, 131, 174312. 19 S.-H. Lee, Phys. Chem. Chem. Phys., 2010, 12, 2655.
20 I.-C. Lu, W.-K. Chen, W.-J. Huang and S.-H. Lee, J. Chem. Phys., 2008, 129, 164304.
21 S.-H. Lee, W.-K. Chen and W.-J. Huang, J. Chem. Phys., 2009, 130, 054301.
22 I.-C. Lu, W.-J. Huang, C. Chaudhuri, W.-K. Chen and S.-H. Lee, Rev. Sci. Instrum., 2007, 78, 083103.
23 C.-H. Chin and S.-H. Lee, J. Chem. Phys., 2011, 134, 044309. 24 P. J. Robinson and K. A. Holbrook, Unimolecular Reactions,
Wiley, New York, 1972.
25 H. Eyring, S. H. Lin and S. M. Lin, Basic Chemical Kinetics, Wiley, New York, 1980.
26 J. I. Steinfield, J. S. Francisco and W. L. Hase, Chemical Kinetics and Dynamics, Prentice-Hall, Engelwood Cliffs, NJ, 1999. 27 A. M. Mebel, V. V. Kislov and R. I. Kaiser, J. Phys. Chem., 2006,
110, 2421.
28 D. R. Lide, Handbook of Chemistry and Physics, CRC, Boca Raton, FL, 78th edn, 1997.
29 From NIST Chemistry WebBook: http://webbook.nist.gov/chem istry/.
30 R. D. Levine and R. B. Bernstein, Molecular Reaction Dynamics and Chemical Reactivity, Oxford University Press, New York, 1987.
31 R. Grice, Int. Rev. Phys. Chem., 1995, 14, 315, and references therein.
32 W. B. Miller, S. A. Safron and D. R. Herschbach, Discuss. Faraday Soc., 1967, 44, 108.