Electron attachment to PCl3 and POCl3, 296-552 K

Article abstract of DOI:10.1063/1.476594

Rate constants for electron attachment to PCl₃ and POCl₃ have been measured over the temperature range 296-552 K in 135 Pa of helium gas, using a flowing-afterglow Langmuir-probe apparatus. Electron attachment to PCl₃ is dissociative, producing only Cl^- ions in this temperature range. The rate constant is 6.4±1.6×10^-8cm³s^-1 at 296 K and increases with temperature in a way that may be described by an activation energy of 43±10meV. Electron attachment to POCl₃ is a richer process in which a nondissociative channel (POCl₃^-) competes with two dissociative ones (POCl₂^- and Cl^-). The rate constant for electron attachment to POCl₃ is 1.8±0.4×10^-7cm³s^-1 at 296 K and is relatively temperature independent in our temperature range. POCl₂^- is the major product over the entire temperature range. Ab initio MP2 and MP4 calculations have been carried out on ground-state neutral and anionic PCl_n and POCl_n for n=1-3. The calculated adiabatic electron affinities agree with experimental estimates where available. The calculations yield C_3v structural symmetries for PCl₃ and POCl₃, and C_s symmetries for PCl₃^- and POCl₃^-. The degree of distortion between the respective neutrals and anions is explored in the calculations, and the implications for electron attachment reactions are outlined.

Full text of DOI:10.1063/1.476594

Electron attachment to PCl 3 and POCl 3 , 296–552 K
Thomas M. Miller, John V. Seeley, W. B. Knighton, Roger F. Meads, A. A. Viggiano, Robert A. Morris, Jane M.
Van Doren, Jiande Gu, and Henry F. Schaefer III
Citation: The Journal of Chemical Physics 109, 578 (1998); doi: 10.1063/1.476594
View online: http://dx.doi.org/10.1063/1.476594
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/109/2?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Electron attachment to CF3 and CF3Br at temperatures up to 890 K: Experimental test of the kinetic modeling
approach
J. Chem. Phys. 138, 204316 (2013); 10.1063/1.4807606
Kinetics of electron attachment to OH and HNO3 and mutual neutralization of Ar+ with NO2 − and NO3 − at 300
and 500 K
J. Chem. Phys. 136, 124307 (2012); 10.1063/1.3694876
Electron attachment to P O Cl 3 : Measurement and theoretical analysis of rate constants and branching ratios
as a function of gas pressure and temperature, electron temperature, and electron energy
J. Chem. Phys. 124, 124322 (2006); 10.1063/1.2176613
Electron attachment to 5-chloro uracil
J. Chem. Phys. 118, 4107 (2003); 10.1063/1.1540108
Electron attachment to SF 5 CF 3 (296–563 K) and calculations of the neutral and anion thermochemistry
J. Chem. Phys. 116, 6021 (2002); 10.1063/1.1457442
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

JOURNAL OF CHEMICAL PHYSICS
VOLUME 109, NUMBER 2
8 JULY 1998
Electron attachment to PCl3 and POCl3, 296–552 K
Thomas M. Miller,a) John V. Seeley,b) W. B. Knighton,c) Roger F. Meads,d)
A. A. Viggiano, and Robert A. Morris
Air Force Research Laboratory, Space Vehicles Directorate/VSBP, 29 Randolph Road,
Hanscom Air Force Base, Massachusetts 01731-3010
Jane M. Van Doren
Department of Chemistry, College of the Holy Cross, 1 College Street, Worcester,
Massachusetts 01610-2395
Jiande Gu and Henry F. Schaefer III
Center for Computational Quantum Chemistry, University of Georgia, Athens, Georgia 30602
͑Received 2 February 1998; accepted 2 April 1998͒
Rate constants for electron attachment to PCl3 and POCl3 have been measured over the temperature
range 296–552 K in 135 Pa of helium gas, using a flowing-afterglow Langmuir-probe apparatus.
Electron attachment to PCl3 is dissociative, producing only ClϪ ions in this temperature range. The
rate constant is 6.4Ϯ1.6ϫ10Ϫ8 cm3 sϪ1 at 296 K and increases with temperature in a way that may
be described by an activation energy of 43Ϯ10 meV. Electron attachment to POCl3 is a richer
process in which a nondissociative channel (POClϪ3 ) competes with two dissociative ones ͑POClϪ2
and ClϪ͒. The rate constant for electron attachment to POCl3 is 1.8Ϯ0.4ϫ10Ϫ7 cm3 sϪ1 at 296 K
and is relatively temperature independent in our temperature range. POClϪ2 is the major product over
the entire temperature range. Ab initio MP2 and MP4 calculations have been carried out on
ground-state neutral and anionic PCln and POCln for nϭ1–3. The calculated adiabatic electron
affinities agree with experimental estimates where available. The calculations yield C3v structural
symmetries for PCl3 and POCl3, and Cs symmetries for PClϪ3 and POCl3Ϫ. The degree of distortion
between the respective neutrals and anions is explored in the calculations, and the implications for
electron attachment reactions are outlined. © 1998 American Institute of Physics.
͓S0021-9606͑98͒02726-3͔
INTRODUCTION
able to determine the electron affinities of PCl3 ͑0.82 eV͒,
POCl3 ͑1.41 eV͒, and POCl2 ͑3.83 eV͒ within an uncertainty
of about 0.15 eV. More recently, our laboratory undertook a
study of the reactivity of POCl3 with a variety of negative
ions.7 In the present work, we have investigated the reactiv-
ity of PCl3 and POCl3 with thermal electrons, measuring rate
constants and product branching ratios over the temperature
range 296–552 K. We have also calculated the structures and
energies of neutral and anionic PCln and POCln , seeking
information on the energetics of dissociative electron attach-
ment and clues about the role played in the attachment pro-
cess by geometry changes induced by the added electron.
The solution phase chemistry of phosphorous trichloride
and phosphoryl chloride is well characterized as these mol-
ecules are used in a number of industrial processes including
the synthesis of important organophosphorous compounds,1
such as flame retardants, plasticizers, insecticides, and fuel
additives.2 In contrast, the gas phase chemistry of PCl3 and
POCl3, including the ion–molecule and electron–molecule
reactions, has not been well studied despite the fact that the
pure substances have relatively high vapor pressures. The
gas phase structures of the neutral molecules are known:
PCl3 is pyramidal while gas phase POCl3 is a distorted tet-
rahedron in which the PCl3 group is pyramidal.3,4 The for-
mation of positive and negative ions by electron bombard-
ment on PCl3 and POCl3 was studied in the 1960s by
Halmann and Klein.5 A decade later, Mathur et al.6 carried
out a crossed beam experiment in which PCl3 or POCl3 mol-
ecules collided with energetic Cs atoms. From the thresholds
for production of various negative ions, Mathur et al. were
EXPERIMENT
The present measurements utilized a flowing-afterglow
Langmuir-probe ͑FALP͒ apparatus. The apparatus and tech-
nique have been discussed in detail in the literature.8,9
A
weak electron-Heϩ, -Arϩ plasma (109 cmϪ3) is generated in
a fast flow of He buffer gas ͑a plasma velocity of
104 cm sϪ1, measured by timing the propagation along the
flow tube axis of a pulse disturbance of the electron density͒.
The decay in electron density was measured along the axis of
the flow tube with a movable Langmuir probe before and
after introduction of an electron-attaching gas at a level of
200–500 ppb. The data were fit to the solution of the rate
equation governing the coupled effects of ambipolar diffu-
a͒
Under contract ͑F19628-95-C-0020͒ to Visidyne, Inc., Burlington, MA.
Air Force Geophysics Scholar, 1995–1997. Present address: Department
b͒
of Chemistry, Oakland University, Rochester, MI 48309-4401.
AFOSR University Resident Research Program participant, 1996. Perma-
c͒
nent address: Department of Chemistry, Montana State University, Boze-
man, MT 59717.
d͒
Under contract to Wentworth Institute of Technology, Boston, MA.
0021-9606/98/109(2)/578/7/$15.00
578
© 1998 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

J. Chem. Phys., Vol. 109, No. 2, 8 July 1998
Miller et al.
579
TABLE I. Measured rate constants ka for electron attachment to PCl3 and POCl3, accurate to Ϯ25%. The
branching fractions for the ion products of attachment to POCl3 are also given in the table. Attachment to PCl3
yields solely ClϪ ions in the temperature range indicated.
Branching fractions
for eϪϩPOCl3 products
PCl3
ka
POCl3
ka
ClϪ
T ͑K͒
(10Ϫ7 cm3 sϪ1
)
(10Ϫ7 cm3 sϪ1
)
POClϪ3
POClϪ2
296
372
457
552
0.64
0.91
1.1
1.8
2.1
1.8
1.5
0.29
0.14
0.06
0.01
0.70
0.85
0.91
0.94
0.006
0.014
0.03
1.4
0.05
sion and electron attachment8,9 to give the electron attach-
ment rate constant. The use of PCl3 and POCl3 offered no
special problems. PCl3 and POCl3 were used as obtained
from the manufacturer10 without further purification aside
from several freeze-pump-thaw cycles to outgas the samples.
Because very low concentrations of reactant are needed in
the experiment, mixtures of typically 0.26% PCl3 or POCl3
in liquid-nitrogen-purified He gas were made immediately
prior to use. No ‘‘stickiness’’ or passivation problems were
observed. These mixtures allow the reactant flowmeters11 to
be operated in a range where they are most accurate. An
additional benefit of the dilute mixtures comes about because
the flowmeters are calibrated for He gas, and no correction is
needed for such low fractions of trace reactant gas. The
FALP apparatus was operated at 135 Pa ͑1 Torr͒ helium
buffer pressure at all temperatures.
The electron attachment rate constants for PCl3 are plot-
ted versus 1/kT in Fig. 1. Over the temperature range 296–
552 K the data may be described by an activation energy of
43Ϯ10 meV. The fitted line in Fig. 1 is given by
ka PCl ͒ϭ3.4ϫ10Ϫ7 cm3 sϪ1 exp Ϫ0.043 eV/kT͒. ͑2͒
͑
͑
3
The prefactor in this expression is essentially identical to the
collisional rate constant calculated from a Maxwell–
Boltzmann average at 296 K of the theoretical s-wave cap-
ture cross section of Klots:14
2
0
1/2
␴
ϭ ␲a /2E͒ 1Ϫexp Ϫ 32␣E͒ ,
͔
͖
͑3͒
͑
͓
͑
͕
cap
though there is no reason, in principle, that it must be. In Eq.
͑3͒ a0 is the Bohr radius, E is the electron energy ͑in har-
trees͒, and ␣ is the molecular polarizability ͑in a30, from Ref.
15͒. A comparison of the measured and collisional rate con-
stants implies that attachment to PCl3 occurs on one out of
every six collisions, on average, at 296 K. The dipole mo-
The rate constants are considered accurate to Ϯ25%.8,9
The relative uncertainty is approximately Ϯ15%. ͑Only rela-
tive values of the electron densities are needed for the deter-
mination of the electron attachment rate constant, provided
that the electron densities are low enough that electron–ion
recombination may be ignored.͒ Periodically, measurements
were made for electron attachment to SF6 for comparison
with accurately known rate constants.9,12
16
ment ͑0.56 D for PCl3 and 2.54 D for POCl3͒ affects the
collisional rate constant for PCl3 by less than 1%, and at the
level of several percent for POCl3.17
Electron attachment to POCl3 gives three ionic products
in the same temperature range:
EXPERIMENTAL RESULTS
Table I gives the rate constants measured for electron
attachment to PCl3 and POCl3. Attachment to PCl3 is disso-
ciative, yielding ClϪ ionic product over the temperature
range 296–552 K:
eϪϩPCl3→ClϪϩPCl2ϩ0.33 eV.
͑1͒
Formation of the parent ion PClϪ3 is exothermic by 0.8 eV,
but this channel is not observed in the present work. At 296
K, the attachment mass spectrum shows additionally PClϪ4 as
a result of ClϪ clustering with PCl3. This ion–molecule re-
action is almost negligible at 383 K with the very low PCl3
concentrations used in the electron attachment work, and im-
perceptible at still higher temperatures. The mass spectra
also show small ͑1%͒ amounts of POClϪ2 and POCl3Ϫ which
we take to indicate the presence of trace oxygen in the He
buffer gas in the flow tube, despite liquid-nitrogen trapping
of the He gas. The attachment results for PCl3 contrast with
those for PF3,13 for which thermal electron attachment is
endothermic.
FIG. 1. Arrhenius plot for electron attachment to PCl3. The slope of the
fitted line is Ϫ42.9Ϯ0.8 meV, but considering the statistical uncertainty in
the measurements, an activation energy of 43Ϯ10 meV is appropriate.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

580
J. Chem. Phys., Vol. 109, No. 2, 8 July 1998
Miller et al.
gies in expressions ͑5͒ and ͑6͒ may be taken as apparent
activation energies for the two product channels. Thus, there
is a small activation energy (16Ϯ5 meV) for process ͑4b͒,
production of POClϪ2 . This process is exothermic by 0.32 eV
according to the results of Ref. 6, with an uncertainty of
ϩ0.10/Ϫ0.25 eV. Similarly, process ͑4c͒, production of
ClϪ, is exothermic by 0.10 eV according to the results of
Ref. 6, with an uncertainty of ϩ0.10/Ϫ0.30 eV. Ab initio
calculations to be described below indicate that reaction ͑4c͒
is endothermic by 0.09 eV; unfortunately, the calculations
are only accurate to about 0.5 eV. The uncertainties in either
determination of the reaction enthalpy are such that the 120
Ϯ40 meV activation energy ͓Eq. ͑6͔͒ observed in the present
work may represent an endothermicity, although activation
energies in electron attachment are common even for exo-
thermic processes.
There have been no previous measurements of electron
attachment rate constants for PCl3 and POCl3, but the forma-
tion of negative ions upon electron bombardment of these
gases was studied by Halmann and Klein5 in the 1960s. They
found ClϪ ionic product from attachment to both PCl3 and
POCl3, peaked at 50Ϯ50 meV electron energy. PClϪ2 from
POCl3 was observed at 400Ϯ400 meV electron energy.
Their observations for PCl3 are in agreement with the present
work, but that for POCl3 is not. We note, however, that Hal-
mann and Klein’s electron energies exceeded ours, and their
gas pressures were far lower than ours, so a direct compari-
son of ion products may be misleading. On the other hand,
Halmann and Klein themselves noted that finding PClϪ2 from
POCl3 is an unlikely turn of events.
FIG. 2. Arrhenius plot for the two dissociative channels observed in electron
attachment to POCl3. The slopes of the fitted lines are Ϫ16Ϯ2 meV
(POClϪ2 ) and Ϫ118Ϯ11 meV (ClϪ). Considering the statistical uncertain-
ties in the data, activation energies of 16Ϯ5 meV and 120Ϯ40 meV, re-
spectively, are appropriate.
→POClϪ3 ϩ1.41 eV
eϪϩPOCl3→POClϪ2 ϩClϩ0.32 eV
→ClϪϩPOCl2ϩ0.10 eV.
͑4a͒
͑4b͒
͑4c͒
The reaction enthalpies for the electron attachment product
channels for PCl3 and POCl3 were calculated from results
reported in Ref. 6 and do not include the thermal energy of
the electron. The uncertainties in the exothermicities are
typically 0.15 eV.
The measured rate constants for electron attachment to
POCl3 show little dependence on temperature within experi-
mental uncertainty. The decrease observed at our highest
temperature is likely that predicted by Eq. ͑3͒. Attachment to
POCl3 is seen to occur in one of every two collisions at 296
K.
AB INITIO CALCULATIONS
Ab initio calculations were carried out on neutrals and
anions relevant to the present work in order to provide an
understanding of the structural and energetic changes which
occur upon addition of an electron to PCl3 and POCl3. The
results of these calculations provide information about rela-
tive positioning of the potential energy surfaces for PCl3 and
POCl3 including adiabatic and vertical electron attachment
energies, vertical detachment energies of PClϪ3 and POCl3Ϫ,
and the energy of dissociation products relative to reactants.
These results are compared to experimental data ͑Ref. 6͒
where possible. For completeness, the calculations included
smaller molecules ͑e.g., PCl͒ uninvolved in the present ex-
periments except for estimating reaction enthalpies for ener-
getically closed channels. Below is given a general descrip-
tion of the calculations. In two cases, convergence problems
required a higher level of theory to be used, and these ex-
ceptions are noted in the tables ͑footnote a of Table IV and
footnote b of Table V͒. The calculations were performed on
several computers: a 200 MHz Micron-PC based on the
Pentium-Pro processor at AFRL, an IBM SP-2 at the AFRL
High Performance Computing Center at Maui, and IBM RS/
6000 machines at the CCQC of the University of Georgia.
The GAUSSIAN 94 package18 of molecular structure pro-
grams was used to calculate the geometries of neutral and
anionic PCln and POCln at the MP2͑Full͒/6-311ϩG(d) level
of theory. These results are given in Table II, and in Fig. 3
Branching fractions for reaction ͑4͒ are given in Table I
for the temperatures at which electron attachment rate con-
stants were measured. Branching fractions were obtained at
68 and 135 Pa buffer gas pressures, and the results are shown
in Fig. 2. A possible pressure dependence in the ClϪ branch-
ing fraction is evident in the figure, but the branching frac-
tions are so small that we hesitate to make much of the
difference between the 68 and 135 Pa results. Branching
fractions for the two dissociative attachment channels, ͑4b͒
and ͑4c͒, were found to increase with temperature, while the
branching fraction for the nondissociative channel decreased
with temperature. Over our entire temperature range the
POClϪ2 ion product was dominant. The branching fractions
for the two dissociative channels ͑4b͒ and ͑4c͒ were modeled
as Arrhenius-type processes as shown in Fig. 2, combining
data at 68 and 135 Pa, with the results ͑296–552 K͒
fraction ͑POClϪ͒ϭ1.4 exp Ϫ0.016 eV/kT͒,
͑5͒
͑6͒
͑
2
fraction ͑ClϪ͒ϭ0.6 exp Ϫ0.120 eV/kT͒.
͑
Since the total rate constant for reaction ͑4͒ is temperature
independent ͑within our experimental uncertainty͒, the ener-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

J. Chem. Phys., Vol. 109, No. 2, 8 July 1998
Miller et al.
581
correction to the energy differences needed in the present
work. The results of the energy calculations are given in
Table IV. Finally, Table V presents adiabatic electron affini-
ties ͑EAs͒, vertical electron attachment energies, and vertical
electron detachment energies calculated at the MP4/6-311
ϩG(d) level of theory. The vertical electron attachment en-
ergy is defined as the energy of the neutral molecule at its
equilibrium geometry minus the energy of the corresponding
anion evaluated at the equilibrium geometry of the neutral.
The vertical detachment energy is defined as the energy of
the neutral molecule evaluated at the equilibrium geometry
of the anion, minus the energy of the anion at its equilibrium
geometry. Figure 4 shows estimated potential energy curves
for the POCl3 system to illustrate the vertical energies. The
vertical attachment and detachment energies in Table IV in-
clude ZPE correction for the neutral molecule in the case of
attachment, and the anion in the case of detachment. The
vertical attachment energies were calculated to give us an
idea of where the anion potential surface intersects that of the
neutral, if at all. The vertical detachment energy was calcu-
lated to give us a feeling for the degree of structural distor-
tion between neutral and anion, in energy units.
FIG. 3. Sketches of the neutral and anionic MP2͑Full͒/6-311ϩG(d) struc-
tures of PCl3 and POCl3. The Cl atoms are labeled with superscripts for
identification with the bond lengths and angles given in Table II and charges
given in Table III.
The calculated adiabatic EAs ͑Table V͒ agree with those
for PCl3, POCl3, and POCl2 determined from collisional ion-
ization measurements6 within experimental and estimated
calculational uncertainty. Interestingly, the calculated EAs
for the PCln do not show the odd/even oscillation in magni-
tude noted strikingly for the SFn series,9 and somewhat for
the PFn series.13,20 The NPA data ͑Table III͒ on where the
‘‘extra’’ electron is placed to form the ground state anion
show that in both PClϪ3 and POClϪ3 80% of the extra negative
charge is localized on the Cl atoms. In contrast, the extra
negative charge is localized more on the P atom ͑53–73%͒
than on the Cl atoms in PClϪ2 , PClϪ, and POClϪ1. In all of
the POCln species the O atom is largely a spectator to the
attachment process: its charge and bond length ͑to P͒ hardly
for the PCl3 and POCl3 systems. In these calculations the
molecules were not confined to any particular symmetry, but
if bond lengths or angles came out to be the same within the
precision given in Table II, they are grouped together to
simplify the table. Table III presents results from a natural
population analysis18,19 ͑NPA͒ of the MP2 electron densities
to tell, within the NPA model, where the ‘‘extra’’ electron on
the anion resides. Energies of these same neutral and anionic
species were calculated at the MP4/6-311ϩG(d) level using
the MP2 geometries. Hartree–Fock calculations HF/6-311
͓
ϩG(d) were performed to determine zero-point energy
͔
͑ZPE͒ corrections. It is common to multiply the HF ZPEs by
an empirical factor19 of approximately 0.9, but we chose not
to do so because such scaling only amounted to about 3 meV
TABLE II. Equilibrium geometries for PCln and POCln neutrals and anions calculated at the MP2͑Full͒/6-311ϩG(d) level of theory. The Cl atoms are
labeled with a superscript where needed ͑to cross reference bond lengths and angles͒. No symmetry was assumed in the calculations, but if bond lengths or
angles came out equal within the precision quoted below, they are combined to simplify the table. Dihedral angles are viewed along the P–Cl1 axis in PCl3
and along the O–P axis in POCl3. The bond lengths are in Å and bond angles are in degrees.
Molecule
PCl3
Calculated structure
r͑P–Cl1͒ϭ2.0579, r͑P–Cl2,3)ϭ2.0573,angle͑Cl1 –P–Cl2,3͒ϭ100.82
dihedral͑Cl2,Cl3͒ϭ103.37
r(P–Cl1)ϭ2.0960, r(P–Cl2,3)ϭ2.3842, angle͑Cl–P–Cl2,3͒ϭ93.81,
dihedral͑Cl2,Cl3͒ϭ162.19
PClϪ3
PCl2
PClϪ2
PCl
r(P–Cl)ϭ2.0497, angle͑Cl–P–Cl͒ϭ101.99
r(P–Cl)ϭ2.1753, angle͑Cl–P–Cl͒ϭ98.84
rϭ2.0166
rϭ2.1936
r(O–P)ϭ1.4642, r(P–Cl1,3)ϭ2.0105, r(P–Cl2)ϭ2.0104, angle͑O–P–Cl1͒ϭ114.89,
angle͑O–P–Cl2͒ϭ114.85, angle͑O–P–Cl3͒ϭ114.93,
dihedral͑Cl1,Cl2͒ϭ119.97, dihedral͑Cl1,Cl3͒ϭϪ120.03
r(O–P)ϭ1.4811, r(P–Cl1)ϭ2.1633, r(P–Cl2,3)ϭ2.3081, angle͑O–P–Cl1͒ϭ109.63,
angle͑O–P–Cl2,3͒ϭ108.26, dihedral͑Cl1,Cl2͒ϭϪdihedral͑Cl1,Cl3͒ϭ109.95
r(O–P)ϭ1.4732, r(P–Cl)ϭ2.0417, angle͑O–P–Cl͒ϭ116.00, dihedral͑Cl,Cl͒ϭ122.35
r(O–P)ϭ1.494, r(P–Cl)ϭ2.2800, angle͑O–P–Cl͒ϭ106.28, dihedral͑Cl,Cl͒ϭ101.22
r(O–P)ϭ1.4830, r(P–Cl)ϭ2.0869, angle͑O–P–Cl͒ϭ109.87
r(O–P)ϭ1.5258, r(P–Cl)ϭ2.3528, angle͑O–P–Cl͒ϭ108.29
PClϪ
POCl3
POClϪ3
POCl2
POClϪ2
POCl
POClϪ
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

582
J. Chem. Phys., Vol. 109, No. 2, 8 July 1998
Miller et al.
TABLE III. Natural population analysis of the MP2 charge density on neutral and anionic PCln and POCln. The
‘‘extra electron’’ number is the difference in the neutral and anionic charge. The Cl atoms are labeled with
superscripts so that they may be identified with particular bond lengths and angles given in Table II.
Charge associated with atom
Cl2
Cl3
O
P
Cl1
System
0.760
0.553
0.21
Ϫ0.253
Ϫ0.351
0.10
Ϫ0.253
Ϫ0.601
0.35
Ϫ0.253
Ϫ0.601
0.35
PCl3
PClϪ3
Extra electron
PCl2
0.552
Ϫ0.100
0.65
Ϫ0.276
Ϫ0.450
0.17
Ϫ0.276
Ϫ0.450
0.17
PClϪ2
Extra electron
PCl
0.236
Ϫ0.496
0.73
Ϫ0.236
Ϫ0.504
0.27
PClϪ
Extra electron
POCl3
1.547
1.399
0.15
Ϫ0.199
Ϫ0.404
0.21
Ϫ0.199
Ϫ0.492
0.29
Ϫ0.199
Ϫ0.492
0.29
Ϫ0.950
Ϫ1.0120
0.06
POClϪ3
Extra electron
POCl2
1.405
1.117
0.29
Ϫ0.238
Ϫ0.547
0.31
Ϫ0.238
Ϫ0.547
0.31
Ϫ0.929
Ϫ1.022
0.09
POClϪ2
Extra electron
POCl
Ϫ0.840
Ϫ1.028
0.19
1.119
0.655
0.53
Ϫ0.349
Ϫ0.628
0.28
POClϪ
Extra electron
change as an electron is added to the system.21
P–Cl bond length change is the greatest ͑0.32 Å͒ for PCl3,
The vertical electron detachment energies ͑Table V͒ are
too large to have any relevance to the thermal attachment
results but are interesting from the point of view of stability
of the anions, especially against photodetachment. It is seen
that POCl2 not only has a large EA ͑3.7 eV͒, but its photo-
detachment threshold is also large ͑4.6 eV͒. The vertical de-
tachment energy for PClϪ3 is surprisingly large ͑3.3 eV͒
given the moderate adiabatic EA ͑1.2 eV͒, but the difference
in these two energies is consistent with the large difference
in the P–Cl bond length between the neutral and anion. The
among the molecules studied in this work.
DISCUSSION
First, let us consider reaction ͑1͒ for dissociative electron
attachment to PCl3. Table V shows that the PClϪ3 surface lies
slightly above that of the neutral ͑a small negative value for
the vertical attachment energy͒ at the equilibrium geometry
of the neutral molecule. Thus, the most likely explanation for
TABLE V. Adiabatic electron affinities, vertical attachment energies, and
vertical detachment energies, in eV, calculated at the MP4/6-311
ϩG(d)//MP2͑Full͒/6-311ϩG(d) level of theory except as noted. The ZPE
corrections are included ͑see text and Fig. 4͒. The adiabatic EAs are ex-
pected to be accurate to Ϯ0.5 eV, and the vertical energies are probably
good to within Ϯ0.6 eV.
TABLE IV. Total energies for netural and anionic PCln and POCln evalu-
ated for geometries calculated at the MP4/6-311ϩG(d)//MP2͑Full͒/6-311
ϩG(d)ϩZPE HF/6-311ϩG(d)//HF/6-311ϩG(d) level. The zero-point
͓
͔
energies ͑ZPE͒ have not been scaled in any way. The ZPE-corrected ener-
gies were used to evaluate EAs ͑Table V͒ and dissociation energies ͑text͒.
Vertical
electron
attachment
energy
Vertical
electron
detachment
energy
Molecule
ZPE
MP4 energyϩZPE
Adiabatic
electron
affinitya
0.005 270
0.002 991a
0.002 971
0.002 270
0.001 284
0.000 857
0.010 609
0.007 707
0.007 542
0.006 386
0.005 268
0.004 060
Ϫ1719.887 785
Ϫ1719.933 503
Ϫ1260.183 752
Ϫ1260.234 410
Ϫ800.432 140
Ϫ800.506 558
Ϫ1794.996 462
Ϫ1795.057 555
Ϫ1335.275 044
Ϫ1335.409 581
Ϫ875.637 465
Ϫ875.676 534
PCl3
Molecule
PClϪ3
PCl2
PCl3
PCl2
PCl
POCl3
POCl2
POCl
1.24
1.38
2.03
1.66
3.66
1.06
Ϫ0.50b
1.16
1.81
0.15
2.60
3.26
1.48
2.17
2.95
4.57
1.96
PClϪ2
PCl
PClϪ
POCl3
POClϪ3
POCl2
POClϪ2
POCl
POClϪ
0.81
aCollisional ionization experiments ͑Ref. 6͒ yield EAs of 0.82
Ϯ0.10 eV͑PCl3͒,
Ϫ0.10 eV͑POCl2͒.
1.41ϩ0.20/Ϫ0.10 eV͑POCl3͒,
and
3.83ϩ0.25/
bA larger basis set 6-311ϩG(2df ) was needed for this energy difference
in order to obtain convergence of the SCF series for the energy of the anion
at the geometry of the neutral.
͓
͔
aNo stable HF structure could be found for PClϪ3 , so this ZPE value was
calculated at the MP2͑Full͒/6-311ϩG(d) level.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

J. Chem. Phys., Vol. 109, No. 2, 8 July 1998
Miller et al.
583
mospheric pressure drift tube. The weak ClϪ product channel
is nearly thermoneutral according to available data6 and re-
sults of the present calculations, so the observed 120 meV
activation energy may be due to a genuine endothermicity or
may be a result of the relative rates of dissociation and sta-
bilization of POClϪ3 . The dominance of the POClϪ2 product
channel over the ClϪ product channel, differing only in the
location of the electron, is likely due to energetics but should
also be favored on the basis of structure if the temporary
POClϪ3 can access its equilibrium geometry ͑POClϪ2 is struc-
turally similar to POClϪ3 while POCl2 is not͒.
As noted in the previous section, the calculations imply
that the O atom in POCl3 is unimportant from structural and
charge density points of view in forming the anion. How-
ever, the presence of the additional atom ͑over PCl3͒ pro-
vides more vibrational modes in which the energy gained
upon addition of an electron can be stored. As a result, the
lifetime of the initially formed, internally excited POClϪ3 is
sufficiently long that collisional stabilization occurs and the
parent anion is observed—in contrast to the PCl3 case, where
the parent anion is not observed in these experiments.
The calculations verify that POCl2 has a large EA ͑3.8
eV experimentally, and 3.7 eV calculated͒. The structure
͑Table II͒ of POClϪ3 is similar to that of POClϪ2 joined to an
additional Cl atom. The only significant difference is that the
dihedral angle ͑looking down along the O–P bond͒ between
the two identical Cl atoms in POClϪ3 is 101°, while in POClϪ2
these same two Cl atoms quite reasonably open up to 140°
͑but note, not to 180°, the planar configuration͒. The NPA
data in Table III show that these two Cl atoms carry about
the same amount of negative charge in both POClϪ3 and
POClϪ2 ͑the P atom in POClϪ2 carries most of the negative
charge that a third Cl atom would claim͒. The structural
similarity between POClϪ3 and POClϪ2 may contribute to the
dominance of the POClϪ2 ion product in electron attachment
to POCl3.
FIG. 4. A cut through the potential energy surfaces of POCl3 ͑upper curve͒
and POClϪ3 ͑lower curve͒ along an arbitrary reaction coordinate connecting
the POCl3 and POClϪ3 minima. The shapes of the potential energy curves are
crude estimates based on the energies given in Tables IV and V. The vibra-
tional levels shown are the zero point energies. VAE denotes the vertical
attachment energy, and VDE denotes the vertical detachment energy. Ex-
perimental data place the POCl2ϩClϪ and POCl2ϪϩCl asymptotes closer
together ͓see Eq. ͑4͔͒.
the 43 meV activation energy that we observe for electron
attachment to PCl3 is that the incoming electron must give
the molecule enough internal energy to access the anion po-
tential surface. Once accessed, the PClϪ3 system does not
stabilize ͑under our conditions͒ but dissociates into
ClϪϩPCl2. The geometry calculations for PCl3 and PClϪ3
͑Table II͒ show that while one of the P–Cl bonds is little
changed by the attachment process, the other two elongate
by 0.33 Å, and therefore the attachment process leaves the
anion well on the way to dissociation of one P–Cl bond if
the parent complex is not stabilized collisionally or radia-
tively.
Reaction ͑4͒ for electron attachment to POCl3 is more
complex since the parent anion is partially stabilized at the
lower temperatures of the present work, in 135 Pa of helium
gas. At all temperatures studied, the formation of POClϪ2 is
favored, and a small amount of ClϪ is observed. From Table
V and Fig. 4 it is seen that the POClϪ3 surface intersects the
neutral surface in the vicinity of the POCl3 minimum. Since
the attachment rate coefficient is large, we surmise that the
anion potential surface is readily accessible to the electron-
molecule system, and the question of whether POClϪ3 or
POClϪ2 is the major final product depends solely on the time
required for collisional or radiative stabilization of the parent
anion. Thus the 16 meV activation energy for the POCl2Ϫ
product channel may be quite different if measured in an ion
cyclotron resonance experiment ͑no collisions͒ or in an at-
CONCLUSIONS
Rate constants have been measured for electron attach-
ment to PCl3 and POCl3 over the temperature range 296–552
K. Attachment to PCl3 yields only ClϪ ionic product in this
temperature range, with an activation energy of 43
Ϯ10 meV. Attachment to POCl3 results in a competition be-
tween a nondissociative channel ͑yielding POClϪ3 ionic prod-
uct͒ and two dissociative channels ͑POClϪ2 and ClϪ͒. POCl2Ϫ
is the major ionic product observed at all temperatures stud-
ied. Branching fractions were determined, as well as activa-
tion energies for the two dissociative channels.
Ab initio structure and energy calculations were carried
out for neutral and anionic PCln and POCln , for nϭ1–3.
The calculated adiabatic EAs agree with those for PCl3,
POCl3, and POCl2 determined from collisional ionization
measurements6 within experimental and estimated calcula-
tional uncertainty. In particular, the EA of POCl2 is large, 3.8
eV experimentally or 3.7 eV calculated. The calculations
show that both PCl3Ϫ and POCl3Ϫ have Cs symmetry. POCl3Ϫ
has a structure very similar to that of POClϪ2 , with an extra
Cl atom appended, taking negative charge from the P atom.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14

584
J. Chem. Phys., Vol. 109, No. 2, 8 July 1998
Miller et al.
¨
by W. Lindinger, T. D. Mark, and F. Howorka ͑Springer-Verlag, Heidel-
berg, 1984͒, pp. 284–306.
The O atom in the POCln series appears to be unimportant to
the initial electron attachment process, but the additional vi-
brational modes related to this O atom are crucial to the
survival of the POClϪ3 ion product of attachment to POCl3.
9 T. M. Miller, A. E. S. Miller, J. F. Paulson, and X. Liu, J. Chem. Phys.
100, 8841 ͑1994͒.
10 Aldrich Chemical Co., Inc., Milwaukee, WI. Both PCl3 and POCl3 were
sold as 99.999% pure. The manufacturer’s analysis showed no detectable
impurities in the POCl3 lot, and 4 ppm impurity in the PCl3 lot.
11 MKS Instruments, Inc., Burlington, MA.
ACKNOWLEDGMENTS
12 R. W. Crompton and G. N. Haddad, Aust. J. Phys. 36, 15 ͑1983͒; Z. L.
Petrovic and R. W. Crompton, J. Phys. B 17, 2777 ͑1986͒.
13 T. M. Miller, J. F. Friedman, A. E. S. Miller, and J. F. Paulson, Int. J.
Mass Spectrom. Ion Processes 149/150, 111 ͑1995͒.
We appreciate continued support for this Laboratory by
the Air Force Office of Scientific Research. Some of the
calculations reported here were carried out at the Maui High
Performance Computing Center of the Air Force Research
Laboratory. JMVD would like to acknowledge support from
the Camille and Henry Dreyfus Foundation. We also wish to
thank Professor Keiji Morokuma for some helpful discus-
sions.
14 C. E. Klots, Chem. Phys. Lett. 38, 61 ͑1976͒. For PCl3 ͑␣ϭ12.8
ϫ10Ϫ24 cm3, Ref. 15͒ the capture rate constant is 3.49ϫ10Ϫ7 cm3 sϪ1 at
296 K. For POCl3 ͑␣ϭ15ϫ10Ϫ24 cm3, estimated͒ the capture rate con-
stant is 3.6ϫ10Ϫ7 cm3 sϪ1 at 296 K.
15 T. M. Miller, in Handbook of Chemistry and Physics, 78th Ed., edited by
D. R. Lide ͑CRC, Boca Raton, 1997͒, Sect. 10, p. 192.
16 D. R. Lide, in Handbook of Chemistry and Physics, 78th Ed., edited by D.
R. Lide ͑CRC, Boca Raton, 1997͒, Sect. 9, p. 43.
1 A. G. Massey, Main Group Chemistry ͑Horwood, New York, 1990͒, pp.
298–300.
17 D. C. Clary and J. P. Henshaw, Int. J. Mass Spectrom. Ion Processes 80,
31 ͑1987͒.
2 R. T. Sanderson, Simple Inorganic Substances: A New Approach ͑Drieger,
Malabar, FL, 1989͒, pp. 336–343.
18 M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, G. A. Peters-
son, J. A. Montgomery, K. Raghavachari, M. A. Al-Lahan, V. G. Zakrze-
wski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nan-
ayakkara, M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W.
Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox,
M. Head-Gordon, C. Gonzalez, and J. A. Pople, GAUSSIAN 94, Revision
C.3, Gaussian, Inc., Pittsburgh, PA, 1995.
3 R. A. Cotton, Advanced Inorganic Chemistry, 5th ed. ͑Wiley, New York,
1988͒, pp. 390–393 and 404–405.
4 A. F. Wells, Structural Inorganic Chemistry, 4th ed. ͑Clarendon, Oxford,
1975͒, pp. 678–680.
5 M. Halmann and Y. Klein, Adv. Mass Spectrom. 3, 267 ͑1966͒; J. Chem.
Soc. ͑London͒ 1964, 4324 ͑1964͒.
6 B. P. Mathur, E. W. Rothe, S. Y. Tang, and G. P. Pack, J. Chem. Phys. 65,
565 ͑1976͒.
19 J. B. Foresman and Æleen Frisch, Exploring Chemistry with Electronic
Structure Methods, 2nd Ed. ͑Gaussian, Pittsburgh, 1996͒, pp. 63–64.
20 G. S. Tschumper, J. T. Fermann, and H. F. Schaefer, J. Chem. Phys. 104,
3676 ͑1996͒; G. L. Gutsev, J. Chem. Phys. 98, 444 ͑1993͒.
21 This statement is obviously not quite so true for POCl, where the O atom
gains 0.19 negative charges upon electron attachment ͑Table III͒.
7 R. A. Morris and A. A. Viggiano, Int. J. Mass Spectrom. Ion Processes
164, 35 ͑1997͒.
8
ˇ
ˇ
D. Smith and P. Spanel, Adv. At., Mol., Opt. Phys. 32, 307 ͑1994͒; D.
Smith and N. G. Adams, in Swarms of Ions and Electrons in Gases, edited
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.252.200.218 On: Wed, 26 Nov 2014 18:22:14