Homogeneous decomposition mechanisms of diethylzinc by Raman spectroscopy and quantum chemical calculations

Article abstract of DOI:10.1021/jp7103787

The gas-phase decomposition pathways of diethylzinc (DEZn), a common precursor for deposition of Zn-VI compounds, were investigated in detail. The homogeneous thermal decomposition of DEZn in N₂ carrier was followed in an impinging-jet, up-flow reactor by Raman scattering. Density Functional Theory calculations were performed to describe the bond dissociation behavior using the model chemistry B3LYP/6-311G(d) to estimate optimal geometries and Raman active vibrational frequencies of DEZn, as well as anticipated intermediates and products. Comparison of the measured DEZn decomposition profile to that predicted by a 2-D hydrodynamic simulation revealed that simple bond dissociation between zinc and carbon atoms is the dominant homogeneous thermal decomposition pathway. The calculations suggest several reactions involving intermediates and Raman scattering experiments confirming the formation of the dimer (ZnC₂H₅)₂. In a different set of experiments, photolysis of DEZn gave evidence for decomposition by β-hydride elimination. The results suggest that β-hydride elimination is a minor pathway for the gas-phase homogeneous pyrolysis of diethylzinc. A reasonable transition state during β-hydride elimination was identified, and the calculated energies and thermodynamic properties support the likelihood of these reaction steps.

Full text of DOI:10.1021/jp7103787

4246
J. Phys. Chem. A 2008, 112, 4246-4253
Homogeneous Decomposition Mechanisms of Diethylzinc by Raman Spectroscopy and
Quantum Chemical Calculations
Young Seok Kim,† Yong Sun Won,† Helena Hagelin-Weaver,† Nicolo´ Omenetto,‡ and
Tim Anderson*,†
Department of Chemical Engineering and Department of Chemistry, UniVersity of Florida,
GainesVille, Florida 32611
ReceiVed: October 26, 2007; In Final Form: December 19, 2007
The gas-phase decomposition pathways of diethylzinc (DEZn), a common precursor for deposition of Zn-
VI compounds, were investigated in detail. The homogeneous thermal decomposition of DEZn in N2 carrier
was followed in an impinging-jet, up-flow reactor by Raman scattering. Density Functional Theory calculations
were performed to describe the bond dissociation behavior using the model chemistry B3LYP/6-311G(d) to
estimate optimal geometries and Raman active vibrational frequencies of DEZn, as well as anticipated
intermediates and products. Comparison of the measured DEZn decomposition profile to that predicted by a
2-D hydrodynamic simulation revealed that simple bond dissociation between zinc and carbon atoms is the
dominant homogeneous thermal decomposition pathway. The calculations suggest several reactions involving
intermediates and Raman scattering experiments confirming the formation of the dimer (ZnC2H5)2. In a different
set of experiments, photolysis of DEZn gave evidence for decomposition by â-hydride elimination. The results
suggest that â-hydride elimination is a minor pathway for the gas-phase homogeneous pyrolysis of diethylzinc.
A reasonable transition state during â-hydride elimination was identified, and the calculated energies and
thermodynamic properties support the likelihood of these reaction steps.
1. Introduction
that sequential homolytic fission of the metal-carbon bond is
the dominant dissociation mechanism (reactions 1a and 1b) on
the basis of the observation. Jackson subsequently investigated
the onset of pyrolysis of diethylzinc in H2 and He flows by
optical monitoring of the rate of condensation of Zn.12 More
recently, Dumont et al. used mass spectroscopic analysis with
capillary sampling at a single point near the heated susceptor
of a pedestal reactor to study the decomposition of Zn(C2H5)2
in both He and H2.13,14 These researchers also observed ethane
and n-butane as the predominant decomposition products. On
the basis of their results as well as those of previous studies,
Dumont and co-workers also concluded that decomposition was
mainly homolytic in nature, with the resultant ethyl radicals
undergoing either recombination to yield butane (reaction 1c),
disproportionation to yield ethane and ethene (reaction 1d), or
decomposition to yield ethane (reaction 1e). Subsequent inves-
tigations by others have supported the following decomposition
mechanism:
Several applications motivate chemical vapor deposition
(CVD) of Zn-containing thin films. In particular, the Zn-VI
compound semiconductors and alloys with other group II-VI
compounds are used as the host semiconductor material in a
variety of optoelectronic devices including light emitting
devices, photodetectors, and lasers.1-3 Among these compound
semiconductors, zinc oxide is commonly used as a transparent
conducting material in emissive device applications as well as
solar cells.4-8 ZnO can be made highly conductive by adjusting
the oxygen vacancy concentration or doping (e.g., Al), whereas
recent reports of p-type conductivity are encouraging for
transparent electronics.9,10 Thin films of Zn-containing com-
pound semiconductors have been deposited by CVD from a
variety of organo-Zn precursors, but most commonly from Zn-
(C2H5)2. Because deposition takes place at an elevated temper-
ature, typically in the range 300-600 °C, homogeneous thermal
decomposition can be important in the reaction pathway. The
multiple early studies of Et2Zn homogeneous decomposition
have suggested Zn-C bond homolysis as the dominant initial
decomposition step,11-14 primarily on the basis of the indirect
evidence of n-C4H10 detection. A later laser pyrolysis investiga-
tion by Linney and Russel,15 however, did not observe n-C4H10
but did identify ZnH2 as a wall deposit, which is a direct product
of double â-hydride elimination of Et2Zn. The aim of this study
to quantitatively probe the rate of disappearance of Et2Zn as
well as search for direct reaction products in an inverted,
impinging-jet test reactor coupled to a Raman spectrometer.
The first detailed study of Et2Zn pyrolysis was reported by
Koski et al.11 using the toluene carrier method. They suggested
Zn(C2H5)2 f •ZnC2H5 + •C2H5
•ZnC2H5 f Zn + •C2H5
2•C2H5 f C4H10
(1a)
(1b)
(1c)
(1d)
(1e)
2•C2H5 f C2H4 + C2H6
•C2H5 f C2H4 + •H
Table 1 summarizes the estimated apparent activation energy
for the first order disappearance of Et2Zn, presumably by
homolytic fission of the first Zn-C bond of DEZn.11-14 It is
noted that the measured activation energy in each study is based
* Corresponding author. E-mail: tim@ufl.edu.
† Department of Chemical Engineering.
‡ Department of Chemistry.
10.1021/jp7103787 CCC: $40.75 © 2008 American Chemical Society
Published on Web 04/12/2008

Decomposition Mechanisms of Diethylzinc
J. Phys. Chem. A, Vol. 112, No. 18, 2008 4247
TABLE 1: Reported Activation Energy for the
Homogeneous Dissociation of the First Zn-C Bond of
Diethylzinc
activation energy frequency
temp
range [°C]
anal.
method
[kcal/mol]
factor [s-1
109-10
]
ref
50
396-489 gas chromatography 11
270a
mass spectroscopy 12
280-520 mass spectroscopy 13
230-470 mass spectroscopy 14
using D2
52.4
56
52
2 × 1016
a Range not reported, only average measurement temperature.
on direct measurement of the concentration decrease of dieth-
ylzinc and indirectly the increase of the •C2H5 radical concentra-
tion. Experimental detection of the primary reaction interme-
•
diate, ZnC2H5, however, has not been reported, presumably
because the lifetimes are too short. Molecular simulations of
•ZnC2H5 decomposition reported in this work support the
presumption that •ZnC2H5 has a short lifetime. As listed in Table
1 the reported activation energy for homolysis of the first Zn-C
bond is in the range 50-56 kcal/mol, which is considered
relatively good agreement among the different measurements.
The only evidence supporting sequential homolysis is the
detection of ethane and n-butane in the outlet of a MOCVD
reactor, neither of which is direct reaction intermediate (i.e.,
•ZnC2H5) and thus potentially misleading.
In contrast to the bond homolysis radical mechanism sug-
gested above, Linney and Russel15 observed a different reaction
pathway using IR laser-powered homogeneous pyrolysis. They
detected H2Zn by FTIR and 1H NMR spectroscopy as the major
Zn product of Et2Zn homogeneous pyrolysis in the temperature
range 280 to 520 °C. This observation and no detection of
n-butane led these researchers to conclude double â-hydride
elimination as the most probable route (reactions 2a and 2b).
Figure 1. Up-flow, impinging-jet reactor for in-situ Raman scattering
measurements.
Zn(C2H5)2 f HZnC2H5 + C2H4
HZnC2H5 f H2Zn + C2H4
(2a)
(2b)
carrier gas is introduced through a center-lined tube, while a
N2 sweeping gas flow envelops the organometallic gas to
minimize wall deposition of precursors or reaction products.
The reactor body is constructed of four optically flat quartz
plates fused on edge to produce a square base. A quartz tube
with four vertical slits is inserted into this rectangular chamber
to provide a better defined flow pattern to simplify modeling.
The vertical slits in the tube permit access of the probe beam
to the reactor centerline and detection of the scattered light.
The up-flowing gas stream impinges on a resistance-heated
pedestal susceptor equipped with a control and measurement
thermocouple.
Before measuring the concentration and temperature profiles
along the centerline of the reactor, a steady-state flow pattern
is first established. To ensure that no closed streamlines exist
in the reactor, a CH4 tracer flow study was performed to
delineate suitable conditions.16 This impinging jet reactor was
designed to eliminate wall reactions and produce a well defined
and easily modeled flow region. The temperature and species
concentration gradients that develop along the centerline allow
study of reactions over a wide range of conditions in a single
steady-state experiment. By adjusting the temperature and gas
velocity, the reaction mechanism could be programmed for a
homogeneous decomposition study. A more detail description
of the reactor as well as the data analysis is given elsewhere.16
A U-1000 Raman spectrometer (Jobin Yvon) was used to
collect and analyze the scattered light. The 488 nm wavelength
of an Ar-ion laser or the frequency doubled 532 nm wavelength
of a Nd/YAG solid-state laser was used as the excitation source
The possibility of heterogeneous reaction contributions and
uncertainty in reaction temperature further complicates the
comparison of previous work. To better understand the homo-
geneous pyrolysis mechanism of Zn(C2H5)2, its decomposition
was studied by Raman spectroscopy in an translatable imping-
ing-jet reactor. The advantages of Raman scattering are its high
spatial resolution and in situ temperature measurement; however,
the relatively low sensitivity for species detection is a drawback.
Furthermore, scattering cross-sections are required for quantita-
tive analysis. The reactor design used in this study includes a
sweep flow to eliminate reactions with walls and limit them
only to the heated pedestal. Quantum chemical calculations were
used to estimate Raman frequencies to assist in the interpretation
of the experimental data. Based on experimental observations
and computational results reported here, the following additional
reaction step of dimer formation is suggested.
2•ZnC2H5 f (ZnC2H5)2
(2c)
2. Experimental and Calculation Details
Except for preliminary experiments using a commercial flow
cell reactor (56 cm3 volume) to measure the Raman scattering
cross-section of DEZn, a custom impinging jet reactor was used
to study the gas-phase decomposition of Zn(C2H5)2 as schemati-
cally shown in Figure 1. The organometallic precursor in a

4248 J. Phys. Chem. A, Vol. 112, No. 18, 2008
Kim et al.
TABLE 2: Calculated Raman Shift (cm-1) of ν[Zn-C] and
ν[Zn-H] for Selected Zn-Containing Moleculesa
and a photomultiplier tube was used as a detector. A maximal
laser power of 1-5 W was used, although the actual excitation
power was lower due to losses from mirrors and other optical
components. A charge-coupled device (CCD) was also available
in system but the analytical measurements were mainly per-
formed using the PMT.
The theory of Raman scattering based on fundamental
molecular motions, including electronic, rotational, and vibra-
tional motions is very well developed.17-19 The Raman intensity
is inversely proportional to the fourth power of the light source
wavelength and to the differential Raman cross-section. To
estimate the mole fraction of a specific molecule from the
measured intensity, the ratio of the Raman intensity of the
molecular species of interest and the carrier gas (N2) is used as
shown in
ν[Zn-C]
Zn(C2H5)2 •ZnC2H5 HZnC2H5
δ[Zn-C & C-H] ν[Zn-C]
HZnC2H5
H2Zn
486.0
313.1
590.5
666.5
1918.9
a The frequency values of first three molecules are Zn-C vibrational
stretching and the other frequency involved Zn-H rocking motion.
ates and thus suggestion of a decomposition mechanism. The
density functional theory (DFT) calculations were performed
using the Gaussian 03 package.20 Bond dissociation energies
were calculated using Becke’s three-parameter hybrid exchange
functional combined with the Lee-Yang-Parr gradient-cor-
rected correlation functional (B3LYP) and the 6-311G(d) basis
set. This method and basis set specify the model chemistry used
to calculate molecular optimized geometries, and atomic or
molecular properties for the reaction species, i.e., Zn-containing
molecules.21-23 This model chemistry was also used to calculate
the Raman active frequencies along with the symmetrical
vibrational motions between zinc and carbon or zinc and
hydrogen atoms in selected intermediate species. Calculated
values of the rotational or vibrational frequencies allowed their
partition functions to be evaluated to yield Gibbs energies of
possible intermediates to judge which reactions are thermody-
namically favored. In identifying reaction intermediates, multiple
combinations of different fragments, i.e., reaction intermediates,
were tested.
Nj
xj )
(3)
(4)
Nj + NQ
NQ (ν˜0 - ν˜j)4 1 - exp(-hcν˜Q/kBT) IQ
)
∑
(ν˜0 - ν˜Q)4
Nj
Ij
1 - exp(-hcν˜j/kBT)
j
where Nj and Ij are the number density and measured intensity
of component j, and the subscript “Q” denotes the Q-branch of
N2 carrier gas. ν˜0, ν˜j and ν˜Q are the wavenumbers of the incident
light, molecular vibration and Q-branch of N2 carrier gas. The
term Σj is the relative Raman cross-section for vibration mode
j.
The Raman intensity of the molecular species of interest is
compared to that of N2 because N2 is at high concentration and
known to have stable rotational and vibrational motions at
reaction conditions. In addition to the vibrational frequency, the
relative Raman cross-section, Σk, is required. In practice, cross-
sections are usually reported relative to that of the Q-branch of
N2 because N2 has sufficient stability to be an internal standard.20
3. Results and Discussion
The two gas-phase decomposition reaction pathways reported
in the literature can be distinguished by their Zn intermediates.
•
The species Zn and ZnC2H5 are produced from homolytic
fission, and HZnC2H5 and ZnH2 result from â-hydride elimina-
tion. The first set of experiments was designed to promote the
decomposition of Zn(C2H5)2 to a sufficient extent that the Zn
intermediates could be detected by Raman scattering. To this
end photolysis of Zn(C2H5)2 was carried out using the incident
laser at high power in the flow cell. To assist in the peak
assignment, DFT calculations were performed to identify
vibrational frequencies of candidate reaction intermediates.
Ckν˜Q (∂σ/∂Ω)k (ν˜0 - ν˜Q)4 1 - exp(-hcν˜k/kBT)
)
)
(5)
∑
(ν˜0 - ν˜k)4
CQν˜k
(∂σ/∂Ω)Q
1 - exp(-hcν˜Q/kBT)
k
where Ck is a constant, which contains specific information on
the kth vibration mode. The relative Raman cross-section shown
above is calculated using the Q-branch of the N2 vibrational
motion of ν˜Q, which appears at 2331 cm-1. The Raman cross-
sections for metalorganic precursors, however, are generally not
known.
Several computational approaches were first tested to repre-
sent species in the DEZn system. Among them, the B3LYP level
calculation with 6-311G(d) basis set was chosen to describe
the nature of bond rupture and analyze the thermodynamic
properties based on geometry optimization calculations. The
products from carbon-carbon dissociation, i.e., Zn(C2H5)2 f
(C2H5)Zn(CH2) + CH3, were not included because the carbon-
carbon bond strength is considerably greater than the zinc-
carbon bond strength. This assertion is supported by an analysis
using the Wiberg bond index in the natural bond orbital
(NBO): Zn-C: 0.4990, C-C: 1.0600, C(next to Zn)-H:
0.9362, and C(next to C)-H: 0.9403. On the basis of the
optimal geometries, the symmetry point groups were identified
and Raman-active vibrational frequencies in normal modes along
with the symmetry groups were obtained.24 Table 2 summarizes
the calculated results.
In a preliminary experiment, the Raman cross-section of
DEZn was measured in a commercial flow-cell reactor. In this
measurement a steady stream of DEZn (10 mol % in N2),
normally at atmospheric pressure, was introduced into the flow-
cell reactor at 3.0 cm/s velocity and sufficient time was allowed
for the reactor to reach steady state. The contents of the cell
were then excited with the 532.08 nm line at 1-3 W and the
scattered intensity of the Zn-C stretch at 480 cm-1 was
recorded. With repeated measurements at room temperature, the
value of the relative Raman cross-section of DEZn was
estimated to be 4.2. Although the cross-section of this molecule
is lower than that of the group III alkyls (e.g., ΣTMIn ) 22.3),
the signal-to-noise ratio at this concentration was sufficient to
quantify the Zn(C2H5)2 concentration from the measured
intensity of the 480 cm-1 line.
Atomic Zn was not expected to be detected by Raman
spectroscopy because the most probable emission radiates from
the excited state (1P1 ) to the ground state (1S0) with just 1.6 ns
o
lifetime, which is too short to be detected.25 Therefore, Raman
scattering experiments were used to search for the possible
Coupling the experimental observations with the computa-
tional calculations enabled identification of reaction intermedi-
•
decomposition fragments ZnC2H5, HZnC2H5, and ZnH2.

Decomposition Mechanisms of Diethylzinc
J. Phys. Chem. A, Vol. 112, No. 18, 2008 4249
Figure 2. Laser-induced Raman shifts of DEZn at room temperature. The lower spectrum was obtained at 1.3 W incident laser power (the laser
power setting was 2 W; however, the intensity decreased to 1.3 W incident onto the reactor due to several mirrors placed in the light pathway) and
the upper one was obtained at 3.3 W incident laser power.
3.1. Laser-Induced Decomposition of Zn(C2H5)2 in a Flow
Cell. As presented in the experimental section, room-temperature
scattering experiments were performed in the flow cell on DEZn
in N2 carrier gas to determine the scattering cross-section of
DEZn. The literature has suggested both homolysis of the Zn-C
bond and â-hydride elimination as the homogeneous decom-
position pathway. No reaction intermediates, however, have been
experimentally detected for Zn species. The only products
detected have been ZnH2 and hydrocarbons (C2H4 and C2H6).
In an attempt to evaluate the use of Raman scattering to detect
intermediates or products, a series of experiments at high laser
power was performed to promote decomposition of DEZn,
presumably via photolysis at the 532 nm wavelength of the
incident beam. Although the preferred photolyisis pathway is
not expected to be the same as that for pyrolysis, identification
of photolysis decomposition products will assist in the inter-
pretation of the subsequent experiments.
Laser-induced photolysis of a stream of 10 mol % Zn(C2H5)2
in N2 nominally at room temperature was studied in the
commercial flow cell operating at steady state. No decomposi-
tion of Zn(C2H5)2 is expected at low power but as the power is
increased photolytic decomposition of DEZn should activate
decomposition sufficient to detect reaction intermediates. Be-
cause all anticipated fragments reside in the Raman shift ranges
250-700 and 1850-1950 cm-1, only these ranges were
scanned. The Raman spectra obtained for laser powers of 2 and
5 W are shown in Figure 2. The lower spectrum, which was
taken at a laser power of 2 W nominal (1.3 W incident) shows
a broad vibration peak around 480 cm-1. Minimal photolysis
is anticipated at this incident power and the observed peak is
attributed the Zn-C stretch in Zn(C2H5)2. The main peak
assignment was carried out on the basis of the literature as well
as computational calculations. It should be noted that all the
calculation results in Raman frequencies were also confirmed
from values in the literature.26,30
compared to the first â-hydride elimination. It is noted that the
peak at 441.7 cm-1 agrees well with the calculated value for
Zn-C stretching of the transition-state species in the product
of the first â-hydride elimination, HZnC2H5. In contrast, no peak
was detected in the spectrum at the 5 W power level at the
•
frequency calculated for the homolysis intermediate ZnC2H5.
•
It is possible, however, that ZnC2H5 radical Raman cross-
section is too small for detection or the radical was depleted by
subsequent reaction. Identifying hydrocarbon decomposition
products would assist in differentiating between the two reaction
pathways. The expected hydrocarbon gas molecules, however,
have small Raman cross-sectional areas rendering their detection
challenging. The detection of ZnH2 by Raman scattering in these
preliminary experiments suggest that the â-hydride elimination
pathway is active at least by photolysis. The results also lend
support to the accuracy of the DFT calculations and encouraged
the more quantitative study of the pyrolytic decomposition of
DEZn using the inverted impinging-jet reactor presented below.
3.2. Pyrolytic Decomposition of Zn(C2H5)2 in an Inverted
Impinging-Jet Reactor. Given the previously reviewed evi-
dence for both homolysis and â-hydride elimination thermal
decomposition pathways, a set of experiments was performed
to measure the rate of pyrolysis of DEZn. In these studies, Zn-
(C2H5)2 was introduced into the center line of the probe reactor
depicted in Figure 1 at an inlet concentration 0.5 mol % in N2.
The purpose of this experiment was to measure the Zn(C2H5)2
profile along the centerline as well as any other detected reaction
products. It is noted that the high temperature reduced the signal-
to-noise level, and transverse and axial diffusion diluted the
precursor as well as reaction products. The existence of a
validated reactor model25 and rate constants from the compu-
tational studies allowed comparison of the experimental profiles
to those predicted by the reaction simulation.
The concentration of DEZn was calculated from measured
Raman scattering intensities using eqs 3 and 4 and the relative
Raman cross-section obtained from the preliminary flow-cell
experiments. The intensity ratio of DEZn relative to nitrogen
was calculated from eq 4 and mole fraction of DEZn in nitrogen
carrier gas was obtained using eq 3 along with the measured
temperature gradient. Figure 3 plots the measured centerline
DEZn concentration at selected positions below the heated
susceptor (set point 600 °C). This figure also shows the
temperature profile calculated using a detailed 2-D asymmetrical
reactor model. This model is described in greater detail
elsewhere, including its validation.16,25 Briefly, the simulation
uses the finite element method (FEM) to simultaneously solve
the steady-state momentum, heat, and mass balance equations
in a reacting system. It is noted that a ∼160 °C difference exists
between the set-point temperature and the gas-phase temperature
Increasing the laser power to 5 W (3.3 W incident) yields an
increase in the intensity of the 480 cm-1 band and introduces
an additional sharp feature at 1917.9 cm-1 26
This feature is
.
attributed to the Zn-H vibrational stretch of ZnH2, consistent
with the calculated frequency of 1918.9 cm-1 listed in Table 2.
Incomplete decomposition of DEZn and the higher incident
power resulted in a strong peak for DEZn. The higher local
temperature also produced a peak shift of ∼1 cm-1 to lower
wavelength. A weak scattering signal is detected around at 666.7
cm-1 assigned to the Zn-H symmetrical vibration of HZnC2H5.
Both of these fragments are expected from decomposition via
â-hydride elimination.
The low intensity of the peak corresponding to HZnC2H5
suggests the second â-hydride elimination is relatively fast

4250 J. Phys. Chem. A, Vol. 112, No. 18, 2008
Kim et al.
Figure 3. Comparison of measured DEZn centerline profile (symbol) to simulated profile assuming only â-hydride elimination (uppermost long
dash line) or only homolysis (second uppermost dash line) using the rate constants shown in Table 4. The solid line includes both reactions. The
computed profiles of ZnC2H5 and HZnC2H5 are also shown with both pathways included.
TABLE 3: Calculated Reaction Rate Constants for
extrapolated to the heater surface. This difference is a result of
Homolytic Fission and â-Hydride Eliminationa
thermal resistances between the central thermocouple embedded
Ea [kcal/mol]
A [s-1
]
2.40 × 1017
in the ceramic heater, the glass envelope, and into the gas phase.
It is also noted that the thermal boundary layer thickness is on
the order of 6 mm at these conditions.
homolytic dissociation
â-hydride elimination
52.0
47.2
7.90 × 1014
a Basis set superposition error (BSSE) correction was considered.
As previously indicated, measurement of the mole fraction
of DEZn is challenging because the relative Raman cross-section
of the Zn-C stretch is small. It was possible, however, to
quantitatively measure the DEZn concentration above a mole
fraction of ∼0.5 mol % as shown in the Figure 3. The onset of
DEZn thermal decomposition is seen to occur at ∼4 mm below
the heated susceptor at a gas-phase temperature of ∼150 °C.
The measured profile was compared to that predicted by the
reactor model assuming the disappearance of DEZn occurs by
homolytic fission using values of Ea and A listed in Table 1 by
either Koski et al.11 or Dumont et al.14
The comparisons depicted in Figure 3 show that the rate
constants suggested by Koski et al.11 give a lower decomposition
extent than using the constants reported by Dumont et al.14 It
is noted that the DEZn profile using the Koski et al. constants
is almost identical to that assuming no decomposition occurs.
This is because transverse diffusion occurs along the axis to
decrease the centerline DEZn concentration and the magnitude
of the decrease is similar to the extent homogeneous decom-
position. It is also obvious from the comparison that the reported
homolysis kinetic data do not fully account for the experimen-
tally observed decomposition extent. It supports the possibility
of additional reaction pathways that would lead to increased
DEZn decomposition including â-hydride elimination and
surface reaction at the quartz plate. If the experimental values
are correct, then either the rate constants are incorrect or another
decomposition pathway is significant. Table 3 summarizes the
calculated reaction rate parameters for both the dissociation and
the â-hydride elimination reactions. In rate constant estimation,
the basis set super position error (BSSE) correction was
considered27-29 and the calculated rates were well matched with
reported rates in Table 1. Figure 3 shows the concentration
profiles predicted with these values. As shown in this figure,
the agreement between experimental concentrations and the
predicted profile that includes both pathways is very good. This
suggests both â-hydride elimination and homolytic fission occur,
with homolysis being the dominant pathway in the temperature
range of study.
3.3. Reaction Intermediates. The evidence supporting simple
dissociation as the preferred pathway includes the observation
of the expected final hydrocarbon products (i.e., butane, ethane,
and ethane) and a measured apparent activation energy (see
Table 1) that is consistent among several investigators and
similar to the value suggested by calculations (12). The primary
reaction intermediates, •ZnC2H5 and •C2H5, however, have not
been detected experimentally, perhaps due to short lifetimes or
small scattering cross-sections. In contrast, the observation of
solid ZnH2 on the cell window during IR laser-powered
homogeneous pyrolysis as well as the expected hydrocarbon
end products supports the route of double â-hydride elimination.
Although the photolysis pathway can be different than the
pyrolysis route, it is noted that the single and double â-hydride
elimination intermediates HZnC2H5 and H2Zn were both
observed upon photolysis of DEZn. Thus a series of additional
reactions were postulated and the likelihood of their existence
probed computationally. The same model chemistry of B3LYP/
6-311G(d) was used for consistency to calculate Gibbs energies.
In particular, reactions that included •ZnC2H5 or ZnH (possible
product from Zn-C bond dissociation of HZnC2H5) were
considered.
Table 4 lists four reactions that involve these intermediates
that gave negative calculated values of Gibbs energy changes
at a reaction temperature of 798.15 °C. Moreover these reactions
are exothermic so that the reactions may be facile in the gas

Decomposition Mechanisms of Diethylzinc
J. Phys. Chem. A, Vol. 112, No. 18, 2008 4251
TABLE 4: Favorable Reactions Involving Intermediates Calculated from B3LYP/6-311G(d) Model Chemistrya
a Thermodynamic properties are derived from the same theory level.
TABLE 5: Symmetrical Raman-Active Zn-C or Zn-H
Stretching of Zn-Containing Intermediates and Estimated
Raman Shifts (cm-1) with a Scaling Factor Applied to
examined, a scaling factor was introduced on the basis of
comparison between the experimental and calculated Zn-C
vibrational frequencies for DEZn.30 DFT calculations gave
accurate frequency values for the Zn-H stretch, especially for
motions containing hydrogen with its symmetrical s orbital.
a
HZn-ZnC2H5 and (ZnC2H5)2
HZn-ZnC2H5
Zn(C2H5)2 (ZnH)2
(ZnC2H5)2
A second experiment was performed to detect products of
reactions between intermediates. In this experiment the susceptor
set point temperature was 600 °C to give sufficient decomposi-
tion and thus a relatively high concentration of intermediates.
The Ar-ion laser power was 1 W and the concentration of DEZn
was 0.5 mol % in N2 carrier gas. On the basis of the calculated
Raman shifts shown Table 5, scans were taken over the
expanded range 200-1300 cm-1 (Figure 4). The lower scan
was obtained 1 mm above the inlet and the upper scan at 2
mm. The multiple scans increased the signal-to-noise ratio to
more confidently extract a value for Raman shift of the
intermediate, which likely has a low Raman cross-section.
ν[Zn-C] δ[Zn-H] δ[Zn-C] ν[Zn-C] ν[Zn-C]
calculation
experiment
scaling factor
corrected
486.0
480.0
0.986 none
480.0 428.2
428.2
434.5
548.81
1203.7
1185.0
0.986
none
434.5
0.986
541.1
1186.3
a For Zn-H motions, the correction factor was not adapted.
•
phase. These reactions include 2•Zn-H T Zn2H2, HZn +
•ZnC2H5 T HZnZnC2H5, and 2•ZnC2H5 T (ZnC2H5)2 in both
cis- and trans-conformations. Other reactions examined included
formation of CH3ZnC2H5, HZnCH3, and (ZnH)3 but the
calculated energy changes were positive.
On the basis of the results in Table 4, the most favorable
reactions are the monoethylzinc dimerization reactions to form
the cis- and trans-conformations. The distance between the two
ethyl groups is sufficiently long (6.27 Å) that there should not
be a difference in their stability. Therefore, no energy rotation
barrier was evident between the cis- and trans-conformations
and the Raman-active vibrational frequencies are identical. Table
5 shows the calculated Raman frequencies for each stable
reaction intermediate along with the Zn-C symmetrical motions.
Given similar Zn-C motions in DEZn and the intermediates
Comparison of the experimental value of the shift at 1185
cm-1 to those listed in Table 5 suggests the presence of
(ZnC2H5)2 as a reaction product of the ZnC2H5 intermediate.
In this assignment, the Raman shift of 1185 cm-1 is not directly
attributed to Zn-C vibrational stretching. At first glance, the
calculated Raman shift values of Zn-C stretch for DEZn (480
cm-1) and (ZnC2H5)2 (1185 cm-1) does not appear correct. The
observed shift at 1185 cm-1 was detected in a careful scan of
the full range. The broad peak shape suggested that the shift

4252 J. Phys. Chem. A, Vol. 112, No. 18, 2008
Kim et al.
Figure 4. Raman spectra at two positions in the reactor (1 and 2 mm above the inlet) indicating presence (ZnC2H5)2.
in an up-flow impinging-jet reactor using Raman scattering and
following the 480 cm-1 line attributed to symmetrical stretch
of the Zn-C bond. The Raman shift for the diethylzinc species
was confirmed using the B3LYP/6-311G(d) model chemistry.
The measured concentration profile was then compared to those
predicted by a 2-D hydrodynamic simulation including various
rate equations. This comparison indicates that homolytic fission
of the Zn-C bond is the dominant initial reaction, although
â-hydride elimination is also active. The reaction intermediate
•ZnC2H5 does not have a sufficient lifetime to be detected and
secondary products such as ethane, ethane, and butane have very
low cross-sections and are thus also difficult to detect using
Raman spectroscopy. The dimerization of the reaction inter-
mediate ZnC2H5 was confirmed both by experiment and by DFT
calculations. Consequently, the dissociation of DEZn and a
second minor â-hydride elimination reaction were identified as
likely pathways in the presented research. The agreement
between the calculated Raman shifts and the experimentally
observed shifts is remarkable and provides evidence for the
benefits of combining DFT calculations and experiments.
Figure 5. Schematic diagram of two suggested competing reactions
with reaction enthalpies [kcal/mol] listed.
involved Zn. In an attempt to associate a particular intermediate
with this shift, seven intermediate species were screened by
DFT. The first screen was on the basis of ∆G values, which
suggested that ZnH2, HZn-ZnC2H5, and (ZnC2H5)2 were
thermodynamically likely. Examination of the Zn-C stretching
of (ZnC2H5)2 using the DEZn scaling factor gave a surprisingly
close result. This was rationalized on the basis that the Zn-
C-C angle is calculated as 116° rather than 90°. Because the
Zn–C stretching was induced by the C–C stretching motion,
the relative Raman cross-section was expected to be small.
Therefore, the actual detection of this peak was challenging and
was only possible by using multiple scans and long integration
times. On the basis of experimental observations and support
from DFT calculations, the thermal decomposition of DEZn has
energies as depicted in Figure 5.
References and Notes
(1) Waag, A.; Gruber, Th.; Thonke, K.; Sauer, T.; Kling, R.; Kirchner,
C.; Ress, H. J. Alloy Compd. 2004, 371, 77.
(2) Wang, L.; Pu, Y.; Chen, Y. F.; Mo, C. L.; Fang, W. Q.; Xiong, C.
B.; Dai, J. N.; Jiang, F. Y. J. Cryst. Growth 2005, 284, 459.
(3) Bachari, E. M.; Baud, G.; Ben Amor, S.; Jacquet, M. Thin Solid
Films 1999, 348, 165.
(4) Cho, S.; Ma, J.; Ki, Y.; Sun, Y.; Wong, G. K. L.; Ketterson, J. B.
Appl. Phys. Lett. 75 (18), 2761.
(5) Haga, K.; Katahira, F.; Watanabe, H. Thin Solid Films 1999, 343,
145.
4. Conclusions
The decomposition pathways of Zn(C2H5)2 were investigated
by both Raman spectroscopy and DFT calculations at the
B3LYP/6-311G(d) chemistry level. With the aid of the DFT
calculations the species formed during laser-induced photolysis
could be identified as HZnC2H5 (minor species) and H2Zn
(major species). This gives support for including â-hydride
elimination as a photolytic reaction pathway. During pyrolysis,
however, these species were not detected, suggesting â-hydride
elimination is not dominant in this mode. As summarized below,
the dimer (ZnC2H5)2 was observed during pyrolysis, and
although not a direct decomposition product, its detection
supports homolysis as the dominant DEZn thermal decomposi-
tion pathway. The thermal decomposition of DEZn was followed
(6) Mycielski, A.; Kowalczyk, L.; Szadkowski, A.; Chwalisz, B.;
Wysmolek, A.; Stepniewski, R.; Baranowski, J. M.; Petemski, M.; Witowski,
A.; Jakiela, R.; Barcz, A.; Witkowska, B.; Kaliszek, W.; Jedrzejcak, A.;
Suchocki, A.; Lusakowska, E.; Kaminska, E. J. Alloy Compd. 2004, 371,
150.
(7) Li, X.; Yan, Y.; Gessert, T. A.; DeHart, C.; Perkins, C. L.; Young,
D.; Coutts, T. J. Electrochem. Solid State Lett. 2003, 6 (4), C56.
(8) Ghandhi, S.; Field, R.; Shealy, J. Appl. Phys. Lett. 1980, 37 (5),
449.
(9) Look, D. C.; Reynold, D. C.; Litton, C. W.; Jones, R. L.; Eason,
D. B.; Cantwell, G. Appl. Phys. Lett. 2002, 81 (10), 1830.
(10) Minegishi, K.; Koiwai, Y.; Kokuchi, Y.; Yano, K. Jpn. J. Appl.
Phys. 1997, 36, L1453.
(11) Koski, A. A.; Price, S.; Trudell, B. C. Can. J. Chem. 1976, 54,
482.
(12) Jackson, D. A., Jr. J. Cryst. Growth. 1989, 94, 459.

Decomposition Mechanisms of Diethylzinc
J. Phys. Chem. A, Vol. 112, No. 18, 2008 4253
(13) Dumont, H.; Marbeuf, A.; Bouree, J. E.; Gorochov, O. J. Mater.
Chem. 1992, 2 (9), 923.
(14) Dumont, H.; Marbeuf, A.; Bouree, J. E.; Gorochov, O. J. Mater.
Chem. 1993, 3 (10), 1075.
(15) Linney, R. E.; Russell, D. K. J. Mater. Chem. 1993, 3 (6), 1587.
(16) Hwang, J. Y.; Park, C.; Huang, M.; Anderson, T. J. Cryst. Growth
2005, 279, 521.
(17) Long, D. A. Raman Spectroscopy; McGraw-Hill: New York, 1977.
(18) Schro¨tter, H. W.; Klo¨ckner, H. W. In Raman Spectroscopy of Gases
and Liquids; Weber, A., Ed.; Springer-Verlag: New York, 1979; Chapter
4.
D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A.
G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, revision C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(21) Simka, H.; Willis, B. G.; Lengyel, I.; Jensen, K. F. Prog. Cryst.
Growth Charact. 1997, 35, 117.
(22) Smith, S. M.; Schlegel, H. B. Chem. Mater. 2003, 15, 162.
(23) Nagel, V. B.; Bruser, W. Z. Anorg. Allg. Chem. 1980, 468, 148.
(24) Corliss, C. H.; Bozman, W. R. Experimental Transition Prob-
abilities for Spectral Lines of SeVenty Elements; National Bureau of
Standards monograph; NBS: Washington, DC, 1962; p 53.
(25) Hwang, J. Y.; Park, C.; Huang, M.; Anderson, T. J. Electrochem.
Soc. 2005, 152 (5), C334.
(26) Wang, X.; Andrews, L. J. Phys. Chem. A 2004, 108, 11006.
(27) Kim, Y. S. Ph.D. dissertation, Chemical Engineering, University
of Florida, 2007.
(28) Pelekh, A.; Carr, R. W. J. Phys. Chem. A 2001, 105, 4697.
(29) Duijneveldt, F. B.; Duijneveldt-van Rijdt, J. G. C. M.; Lenthe, J.
H. Chem. ReV. 1994, 97 (7), 1873.
(30) Park, C. H.; Zhang, S. B.; Wei, S. H. Phys. ReV. B 2002, 66,
073202.
(19) Jackson, R. L. J. Chem. Phys. 1992, 96 (8), 5938.
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K.
N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A.