The pyrolysis of azetidine: Shock-tube kinetics similarities and contrasts with two analogs

Article abstract of DOI:10.1002/(SICI)1097-4601(1998)30:3<185::AID-KIN3>3.0.CO;2-O

The kinetics of decomposition of azetidine {(CH₂)₃N - H)} was measured using single-puise shock-tube techniques, over the temperature range 855-1100 K, in high argon dilution. These data confirm and extend an earlier investigation that utilized the very lowpressure pyrolysis method. A brief survey of many reports regarding the interesting features of azetidine is presented. In two appendices the thermodynamic and kinetic data on trimethylene sulfide, oxide, and immine are intercompared. New ab-initio calculations are cited for the parent species and their fragmentation products.

Full text of DOI:10.1002/(SICI)1097-4601(1998)30:3<185::AID-KIN3>3.0.CO;2-O

The Pyrolysis of Azetidine:
Shock-Tube Kinetics
Similarities and Contrasts
with Two Analogs
Y-X. ZHANG, C-L. YU, S. H. BAUER
Department of Chemistry, Baker Chemical Laboratory, Cornell University, Ithaca, New York 14853
Received 1 May 1997; accepted 30 June 1997
ABSTRACT: The kinetics of decomposition of azetidine {(CH2)3N9H)} was measured using
single-pulse shock-tube techniques, over the temperature range 855–1100 K, in high argon
dilution. These data confirm and extend an earlier investigation that utilized the very low-
pressure pyrolysis method. A brief survey of many reports regarding the interesting features
of azetidine is presented. In two appendices the thermodynamic and kinetic data on tri-
methylene sulfide, oxide, and immine are intercompared. New ab-initio calculations are cited
for the parent species and their fragmentation products. ᭧ 1998 John Wiley & Sons, Inc. Int J Chem
Kinet 30: 185–191, 1998.
INTRODUCTION
tional data for the three listed species, to obtain esti-
mates of how well ab-initio calculations agree with
experiments for such structures. Our shock-tube re-
sults served to confirm and extend the kinetics of py-
rolysis reported by Kamo et al. [1(a)] who used the
VLPP technique.
The structure, thermochemistry, and molecular dy-
namics of fragmentation of four-member rings with
the composition (H2C)3X, (X ϭ S,O,NH) have re-
ceived extensive study during the past decades. While
our current interest is directed to azetidine (X ϭ NH,
because it is the core compounds for the tri-nitro sub-
stituted species {TNAZ}), we found it instructive to
assemble the available experimental and computa-
For the parent species the published data are assem-
bled in Appendix A. The most plausible route for dis-
sociation passes through biradical intermediates, gen-
erated when one bond in the ring is disrupted. That is
the major activation step. The overall activation en-
ergies range from 54.8 to 65 kcal/mol. The overall en-
Þ
tropy of activation (⌬S ) arises primarily from the
Correspondence to: S. H. Bauer
Contract grant Sponsor: Army Research Office
Contract grant number: Grant # DAAH 04-95-1
large amplitude motion about the bond opposite to the
broken one and the increased loosening of the second
bond prior to breaking.
᭧ 1998 John Wiley & Sons, Inc.
CCC 0538-8066/98/030185-07

186
ZHANG, YU, AND BAUER
X
и
и
и
X
и
X
H2C"CH2 ϩ H2C"X
X
X
Scheme I
Since the first step in the pyrolysis of the ring gener-
ates H2C ϭ CH2 and H2C ϭ X, we focused on estab-
lishing the fate of the H2C ϭ X species, to determine
whether these products, at about 1000 K, remain intact
or fragment to produce reactive radicals that could at-
tack the parent compound or the ethylene. In Appendix
B we summarized published experimental magnitudes
and recently derived ab-initio estimates (via CBS-4).
It appears that the activation energies for fragmenta-
tion of H2C ϭ X range from ca. 80 to 90 kcal/mol; for
AZ only a rough estimate is available. The derived
much higher barrier. A recent calculation at the CBS-
4 level [8] for the enthalpy difference between the
ground state conformation of AZ and the planar tran-
sition structure {for the heavy atoms and H(N)} found
5.57 kcal/mol. The low-frequency IR and Raman
spectra were accounted for by an unsymmetrical po-
tential well for the flapping motion of the ring atoms,
with no second well for an axial H(N) [9]. The shoul-
der in the potential energy function appears between
the flapping vibrational levels 3 and 4. Indeed, the
CBS-4 calculation [8] for a structure initially con-
strained near the axial position, at an excitation of
1.65 kcal, showed no local minimum. However, later
microwave spectra suggest that there may be a very
shallow minimum at the location of the shulder [10].
The vacuum UV absorption spectrum of AZ was
recorded by Clark and Pickett [11]. Also, Clary and
Henshaw [12(a)] suggested, on the basis of a theoret-
ical analysis, that inversion at the N atom in AZ would
be induced by intense laser radiation. Electron impact
spectra of azetidine were recorded by Gallegor and
Kiser [12(b)].
1
kuni at 1000 K is ca. 4 ϫ 10Ϫ5 sϪ , so that the corre-
sponding half-times are several orders of magnitude
longer than the residence time in our reactor. Volkova
et al. [1(b)] reported that H2C ϭ NH did decompose
in
a
low pressure pyrolysis unit. (ca. 3 ϫ
10Ϫ3 torr, 400–700ЊC).
COMMENTS ON AZETIDINE
The gas-phase structure was determined by electron
diffraction and microwave spectroscopy [2]. The
equatorial position of H(N) was confirmed as the dom-
inate conformation. In the gas phase the interatomic
distances in the ring of AZ differ somewhat from those
in the crystalline solid and particularly from solid tri-
nitro-azetidine. So does the flap angle [3]. The vibra-
tional frequencies were remeasured, and with suitable
scaling were reproduced by ab-initio calculations at
the 6–31G* level [4]. The inversion rate at the N atom
was estimated from temperature-dependent NMR line-
SHOCK-TUBE EXPERIMENTS
The shock tube is of stainless steel with a 1 inch inner
diameter. The lengths of the driver and driven sections
are 120 cm and 170 cm, respectively. A damping tank
is attached to the driven section, next to the diaphragm
holder. Shock waves were generated by increasing the
pressure of the He driver until the mylar diaphragm
broke. Typical pressures on the driver and driven sides
are 80 psig and 300–400 torr, respectively. Two
piezo-electric pressure sensors are stationed 10 cm
apart at the end of the driven section. Their summed
signal was recorded and digitized through Biomation
8100 and Northern Tracor Signal Analyzer, and then
Þ
Þ
width measurement; ⌬H ϭ 5.0 kcal/mol and ⌬S ϭ
Ϫ10.8 eu [5]. Ab-initio calculations by these investi-
gators, at the MP3 level indicated a barrier height of
6.48 kcal. (At the same level, the derived inversion
barrier for ammonia was found to be 5.33 kcal, com-
pared with the microwave value of 5.93 [6]). An ear-
lier theoretical analysis [7] using STO-3G led to a

PYROLYSIS OF AZETIDINE
187
stored in an IBM AT computer. The accuracy of the
timed interval is Ϯ1 ␮s, which corresponds to Ϯ
8 K. The temperature in the reflected shock region was
evaluated according to the computational procedure
recommended by Gardiner et al [13] for very dilute
reaction mixtures, and the shock-tube code developed
by Mithell and Kee (Sandia National Laboratory).
Identical results were obtained from the above two al-
gorithms for the mixtures used in experiments. The
effective heating time (ca. 1.3 ms) was defined as the
time elapsed from the beginning of the reflected shock
wave to the point where the pressure signal was 80%
of that in the reflected shock region. The storage tank,
gas handling line, and the shock-tube were maintained
at ca. 80ЊC throughout the experiments. Azetidine
(Ͼ98% purity) from Aldrich was used without further
purification. High purity Helium and Argon were used
for the driving and carrier gases, respectively. Re-
search grade ethylene, cyclopropane, and propene for
GC calibration, are all from Matheson. (Caution: AZ
rapidly attacks neoprene oh rings. All exposed plastic
material was replaced with Teflon).
Figure 1 Unimolecular rate constants observed: (ٗ)
0.02% (mole fraction percentage); (᭞) 0.03%, and (O)
0.19% azetidine in Ar. P5 ϭ 5.0–6.3 atm. The line was cal-
culated via RRKM using the parameters listed in Table I.
P5 ϭ 5.0 atm.
Immediately after each shock, a 16 ml of sample
was collected through the sampling valve at the end
of the driven section and analyzed in a Nicolet GC/
9630 unit, using FID. To separate AZ from its reaction
products a packed GC Carbograph I column (from All-
tech) was used. Injector and detector temperature were
typically set at 220ЊC and 250ЊC, respectively. The
column temperature was set at 80ЊC for the first six
minutes at then increased at a rate of 10ЊC/min until
it reached the temperature of 275ЊC. The carrier gas flow
rate was typically ca. 25 ml/min. The chromatograms
were recorded with a Hewlett-Packard 3396A Integrator.
Azetidine, ethylene, and acetylene(above 1200 K)
were identified and well separated by the above col-
umn under the specified operational conditions. Typ-
ical retention times for these compounds 14 min, 4
min, and 2.8 min, respectively. CH2"NH is an ex-
pected product but it was not observed in the above
TABLE I Parameters Used in RRKM Calculations
E0 ϭ 52.0, kcal/mol
Reaction path degeneracy 2
Vibrational frequencies in molecule(cm
3357.7
Ϫ1
)
3003.2 2960.0 2931.6 2928.1 2861.9 2859.4
1498.7 1473.7 1450.1 1361.1 1320.5 1256.0
1250.9 1195.9 1174.8 1145.5 1082.8
1028.2 990.5 948.8 928.1
910.4 802.1 736.4 647.9
217.0
Ϫ1
Vibrational frequencies in complex(cm
3357.7
)
3003.2 2960.0 2931.6 2928.1 2861.9 2859.4
1498.7 1473.7 1450.1 1361.1 1320.5 1256.0
1250.9 1195.9 1174.8 1145.5 1082.8
928.1 802.1 736.4 527.57
486.83 467.13 332.44
111.34
Collision efficiency 1
Collision cross section 5.0 A
˚
Rotational constants A,B,C: 11.4487, 11.3366, 6.6090
(GHz)
Figure 2 The computed fall-off curve at 930 K.

188
ZHANG, YU, AND BAUER
analysis. There are indications that it is destroyed or
absorbed on the GC packing material.
through the measured rate constants is illustrated in
Figure 1. That led to
First-order rate constants of the reaction were es-
timated from measured GC peak areas of azetidine and
ethylene, respectively, with the following formula:
ϱ
1
ku ϭ 1.03 ϫ 1015 exp(Ϫ54800/RT), sϪ .
ϱ
ϱ
kuni ϭ ln{[AZ]0/[AZ]t}/⌬t ϭ ln{[C2H4] /([C2H4] —
[C2H4]t)}/⌬t, where [AZ]t and {[C2H4]t were obtained
from the GC peak areas of azetidine and ethylene in
the postshock mixture, [AZ]0 from the GC peak area
The fall-off curve(at 930 K) is shown in Figure 2. An
extended analysis of this aspect in the conversion of
cyclic hydrocarbons was presented by Lin and Laidler
[15]. In our analysis we assumed that ring opening
occurs primarily by breaking a C—N bond. This is
justified because the excitation energy required to
break a C—C bond in a four-member ring, as in the
cyclobutane, is about 10kcal/mol higher than that ob-
served for AZ (refer to Appendix A).
ϱ
of the preshocked mixture, and [C2H4] from the GC
peak area of postshock mixture at high temperatures
where all azetidine had decomposed. ⌬t is the heating
time. Although the two methods gave almost identical
results, it was found that rate constants obtained from
azetidine peaks had a larger scatter, presumably due
to variable losses of azetidine by wall absorption dur-
ing sample transfers. Therefore, the final rate constants
were determined from peak areas of C2H4.
CONCLUSION
Shock-tube pyrolysis experiments confirmed and ex-
tended kinetic data on the fragmentation of azetidine.
Tabular summaries are presented of the available ther-
mochemical parameters for AZ and its analogs, tri-
methylene oxide and sulfide, supplemented with ab-
initio calculations (at CBS-4 level). The decomposi-
tion pathways for the products generated upon the ini-
tial fragmentation of the ring structures were also sum-
marized. We found no correlation between the
magnitudes of the large amplitude flapping motions of
the rings and the corresponding activation energies for
ring fission.
RESULTS
The temperature range covered was 855–1300 K. Be-
low 1200 K C2H4 was the only reaction product. At
higher temperatures, C2H2 was formed, presumably
from subsequent decomposition of ethylene. The AZ
fragmentation appears to be a clean unimolecular pro-
cess. A least-squares fit to the experimental data up to
ϱ
60% conversion (850–970 K) gave ku ϭ 1.20 ϫ
1015 exp (Ϫ54800 Ϯ 4200)/RT), sϪ . For conversions
1
greater than 60% the experimental points fall below
the least-squares line, a “bend-over” typically ob-
served for similar pyrolyses (cf. pyrrolidine [14]).
An RRKM calculation was performed with the pa-
rameters listed in Table I. The resulting line, drawn
This investigation was supported by the Army Research Of-
fice under Grant # DAAH 04-95-1, which we gratefully ac-
knowledge.

PYROLYSIS OF AZETIDINE
189
APPENDIX A
APPENDIX A Table and References
Barrier
to
X
:
⌬HЊf(298)
kcal/mol S0(298) eu
C2H4 ϩ H2CX
␦Sr
Flap
Angle
Inversion
kcal/mol
Ϫ1
dDHr
Species
ku(T), s
14.481
67.542
25.43(298)
24.87(900)
40.23
39.74
2 ϫ 1015 exp(Ϫ57000/RT)6a
1 ϫ 1014 exp(Ϫ52000/RT)6b
26Њ7
0.788
S
H2C"S
H2C"CH2
27.413
12.55
55.224
52.405
Ϫ19.255
65.55
5.80(298)
5.97(900)
39.21
40.23
2.63 ϫ 1015 exp(Ϫ62000/RT)9a large amp 0.0448,10
5.13 ϫ 1015 exp(Ϫ63000/RT)9b motion
about 0
O
H2C"O
Ϫ25.955
6.795
52.35
63.25
3.98 ϫ 1015 exp(Ϫ62500/RT)11 27.912
1.4612
5.014a
CH2
18.21(298)
18.24(900)
41.58
42.26
23.4713
22.95,15
63.914
52.965
11.93(298)
11.48(900)
41.50
41.28
2.51 ϫ 1015 exp(Ϫ55600/RT)13 29.714b
9.54 ϫ 1014 exp(Ϫ54800/RT)16
H
H
N
H2C ϭ N
1 J. B. Pedley, R. D. Naylor, and S. P. Kirby, Thermochemical Data of Organic Compounds, Chapman and Hall, London, 1986, 2nd ed.
2 Calculated from molecular structure and vibrational frequencies. (a) C. T. Nielser, Acta Chem. Scand. A31, 791 (1997), (b) R. A. Shaw,
C. Castro, N. Ibrahim, and H. Wieser, J. Phys. Chem., 92, 6528 (1988).
3 B. Ruscic and J. Berkowitz, J. Chem. Phys., 98, 2568 (1993).
4 Calculated from molecular constants, cited by M. Jacox, J. Phys. Chem. Ref. Data. 17, 269 (1988).
5 Polynomial sets listed in Tables of Thermofunctions—A. Burcat, B. McBride, Technion-Areospace Eng. Report (1996).
6 (a) M. Yamada, et al. Nippon Kagaku Kaishi 59, 870 (1987). (b) P. Jeffers, C. Dasch, S. H. Bauer, Int. J. Chem. Kinet. 5, 545 (1973).
7 K. Karakida, K. Kuchitsu, Bull Chem. Soc. Japan 48, 1691 (1975).
8 L. A. Carreira, R. C. Lord, T. B. Malloy, Jr. in Topics in Current Chem. Vol. 82. Springer-Verlag, Berlin, 1979.
9 (a) L. Zalotai, et al. Int. J. Chem. Kinet. 15, 505 (1983). (b) K. A. Holbrook, R. A. Scott, J. Chem. Soc. Farady Trans. 71, 1849 (1975).
10 S. Chan, et al. J. Chem. Phys. 44, 1103 (1966).
11 D. R. Lewis, et al. J. Phys. Chem. 88, 4112 (1984).
12 T. Egawa, et al. J. Chem. Phys. 86, 6018 (1987).
13 T. Kamo, M. Yamada, J. Tang, Y Ohshima, Nippon Kagaku Kaishi 8, 1560 (1987).
14 Calculated from molecular constants, cited by (a), R. Dutler, A. Rauk, R. A. Shaw, J. Phys. Chem. 94, 118 (1990). (b) H. Gunther, G.
Schrem, H. Oherhammor, J. Mol. Spec. 104, 152 (1984).
15 ⌬Hf0(H2C"NH) is controversial; proposed values range from 16-26 kcal/mol.
16 This report.

190
ZHANG, YU, AND BAUER
TABLE B1
APPENDIX B
⌬HЊf
kcal/mol
(298.15 K)
SЊ
e.u.
(298.15 K)
In the attached tables we summarized the thermo-
chemical constraints that control the fragmentation of
the species H2C ϭ X (X ϭ S, O, NH) that are gen-
erated upon dissociation of the corresponding four-
member rings. For eleven of the 18 species that are
included, NASA format polynomials are available:
(Burcat and McBride, Technion report, 1996). For the
remaining seven {C3H6S(c); C3H6NH(c); H2CS; HCS;
H2CN; HCNH(tr); HCNH(cis)} the enthalpies of for-
mation, structures, and vibrational frequencies were
assembled from various reports, and their thermo-
chemical functions were computed for a wide range of
temperatures. These were then fitted to standard poly-
nomials (available upon request). In Table B1 we
listed the derived enthalpies of formation (kcal/mol)
and standard entropies (eu, ideal gas phase) for the
seven species. A wide range of values were reported
for the enthalpy of formation of H2CNH, ranging from
16.5 to 26.4; we chose 22.9 kcal, as accepted by Bur-
cat and McBride. Values for the other species in that
family are consistent with that magnitude. Table B2 is
a compilation of enthalpy and entropy increments for
a sequence of interconversions at 298 K and 900 K.
The cited activation energies were estimated on the
basis of analogy with the two measured rate constants.
Cp
Species
cal/mol-K
C3H6S(c)
H2C"S
HCS
C3H6NH(c)
H2C"NH
H2CN
14.482
27.413
71.824
23.475
22.91
57.406
71.406
76.406
67.34
55.22
56.43
63.84
52.961
53.58
54.71
54.82
16.78
9.12
8.86
15.98
9.101
9.02
9.09
9.29
HCNH(tr)
HCNH(cis)
1 A. Burcat, B. McBride, Technion-Areospace Eng. Report.
(1996).
2 J. B. Pedley, R. D. Naylon, and S. P. Kirby, Thermochemical
Data of Organic Compounds, 2nd ed. Chapman and Hall, London,
1986, AND—Calculations based on molecular structure and vibra-
tional frequencies reported by Nielser and Shaw (see ref. 2, APP.A)
3 B. Ruscic and J. Berkowitz, J. Chem. Phys., 98, 2568 (1993),
AND—Calculations based on data complied by M. Jacox, J. Phys.
Chem. Ref. Data. 17, 269 (1988).
4 Ref. 3 above, and J. Senekowitch, S. Carter, P. Rosmur, H-J.
Werner, Chem. Phys. 147, 281 (1990).
5 T. Kamo, et al. Ref. 13; Dutler, et al., Ref. 14, APP.A
6 Calculated from data cited in: (a) K. L. Nesbit, et al. J. Phys.
Chem. 95, 7613 (1991). (b) R. A. Bair, T. H. Dunning, J. Chem.
Phys. 82, 2280 (1985).
Table B2 Thermodynamical Increments Upon Reaction. Values in kcal/mol and e.u. at 298.15 K and 900 K
Reaction
⌬HrЊ
⌬SrЊ
Ea(estm)
R1. H2CS
R2. H2CS
R3. HCS
!:
H2
H
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
ϩ
CS
39.59(298)
41.10(900)
96.51(298)
98.38(900)
47.28(298)
48.68(900)
Ϫ0.47(298)
1.29(900)
88.09(298)
90.08(900)
15.65(298)
17.17(900)
23.56(298)
26.08(900)
9.37(298)
26.31
29.51
28.63
32.37
21.29
23.98
26.19
29.19
28.75
32.69
21.04
24.06
27.40
32.69
26.51
31.14
28.04
32.04
29.17
33.35
97
!:
!:
!:
!:
!:
!:
!:
!:
!:
HCS
CS
99
48
H
R4. H2CO
R5. H2CO
R6. HCO
H2
H
CO
ca. 77 (slightly
less than R5)
kbi ϭ 1.26 ϫ 1016 exp(77890/RT)
cm3/mol.s (ref. 1)
18
HCO
CO
H
R7. H2CNH
R8. H2CNH
R9. H2CNH
R10. H2CNH
H2
H2
H
HNC
HCN
H2CN
HCNH(tr)
84
85
11.58(900)
86.60(298)
88.61(900)
100.60(298)
102.71(900)
86
H
100
(Continued)

PYROLYSIS OF AZETIDINE
191
Table B2 (Continued)
R11. H2CNH
R12. HCN
!:
!:
!:
!:
!:
H
ϩ
HCNH(cis)
105.60(298)
107.34(900)
14.19(298)
14.5(900)
26.97(298)
28.93(900)
27.16(298)
29.34(900)
12.97(298)
14.83(900)
29.28
33.71
0.88
106
HNC
H
ku ϭ 3.5 ϫ 1013 exp(Ϫ47190/RT),
Ϫ
1
1.55
s
(ref. 2)
30
R13. H2CN
ϩ
ϩ
ϩ
HCN
HNC
HCN
22.07
25.94
21.83
26.18
20.95
24.63
R14. HCNH(tr)
R15. HCNH(tr)
H
30
30
H
1 D. L. Baulch, J. Phys. Chem. Ref. Data. 21, 411 (1992).
2 M. C. Lin, Y. He, and C. F. Melius, Int. J. Chem. Kinet., 24, 1003 (1992).
8. C. F. Wilcox, Jr. Cornell University (unpublished).
9. (a) L. A. Carreira, R. O. Carter, J. R. Durig, J. Chem.
Phys. 57, 3384 (1972); (b) A. G. Robiette, T. R. Bargers,
H. L. Strauss, Molecular Phys. 42, 1519 (1981).
10. Ref. 2, p163.
11. P. A. Clark, L. W. Pickett, J. Chem. Phys. 64, 2060
(1976).
12. (a) D. C. Clary, and J. P. Henshaw, J. Phys. Chem., 87,
3025 (1983); (b) E. J. Gallegos, R. W. Kiser, J. Phys.
Chem., 66, 136 (1962).
13. W. C. Gardiner, Jr., and B. F. Walker, and C. B. Wake-
field, in Shock Waves in Chemistry, A. Lifshitz, Ed.,
Marcel Dekker, New York, 1981, Chap. 7.
14. A. Lifshitz, M. Bidani, A. Agranat, A. Suslensky, J.
Phys. Chem. 91, 6043 (1987).
BIBLIOGRAPHY
1. (a) T. Kamo, M. Yamada, J. Tang, Y. Ohshima, and A.
Amano, Nippon Kagaku Kaishi, 8, 1560 (1987); (b) V.
V. Volkova, et al. Tetrahedron Lett., 577 (1978).
2. H. Gu¨nther, G. Schrem, H. Oherhammer, J. Mol. Spec.
104, 152 (1984).
3. (a) T. C. Archibald, R. Gilardi, K. Baum, J. Org. Chem.
55, 2920 (1990). (b) See also crystal structures by: P.
Huszthy, et al. Heterocyclic Chem. 30, 1197 (1993).
4. R. Dutler, A. Rauk, R. A. Shaw, J. Phys. Chem. 94, 118
(1990).
5. R. Dutler, A. Rauk, T. S. Sorensen, J. Am. Chem. Soc.
109, 3224 (1987).
6. M. F. Maming, J. Chem. Phys. 3, 136 (1935).
7. J. Catala`n, O. Mo, M. Ya`neˆz, J. Mol. Str. 43, 251
(1978).
15. M. C. Lin, K. J. Laidler, Trans. Farady Soc. 64, 927
(1968).