Thermal decomposition of trimethylamine borane as a precursor to nanocrystalline CVD BC x N y films

Article abstract of DOI:10.1134/S0020168511100128

We have studied the kinetics of BC _x N _y chemical vapor deposition through trimethylamine borane decomposition at atmospheric pressure. The rate constant of the heterogeneous interaction between trimethylamine borane and an adsorption center has been determined to be k _s ⁰ = 2.7 × 10⁷exp(-10560/T) cm/s. The obtained kinetic parameters of the reaction fully determine the growth rate of nanocrystalline carbonitride films under kinetic control. The film thickness has been determined as a function of time, temperature, reactant concentration, and reactor dimensions.

Full text of DOI:10.1134/S0020168511100128

ISSN 0020ꢀ1685, Inorganic Materials, 2011, Vol. 47, No. 11, pp. 1199–1204. © Pleiades Publishing, Ltd., 2011.
Original Russian Text © O.P. Korobeinichev, A.G. Shmakov, A.A. Chernov, M.L. Kosinova,V.S. Sulyaeva, F.A. Kuznetsov, 2011, published in Neorganicheskie Materialy, 2011,
Vol. 47, No. 11, pp. 1317–1322.
Thermal Decomposition of Trimethylamine Borane As a Precursor
to Nanocrystalline CVD BCxNy Films
O. P. Korobeinicheva, A. G. Shmakova, A. A. Chernova, M. L. Kosinovab,
V. S. Sulyaevab, and F. A. Kuznetsovb
a Institute of Chemical Kinetics and Combustion, Siberian Branch, Russian Academy of Sciences,
Institutskaya ul. 3, Novosibirsk, 630090 Russia
b Nikolaev Institute of Inorganic Chemistry, Siberian Branch, Russian Academy of Sciences,
pr. Akademika Lavrent’eva 3, Novosibirsk, 630090 Russia
eꢀmail: chernov@kinetics.nsc.ru
Received February 14, 2011
Abstract—We have studied the kinetics of BCxNy chemical vapor deposition through trimethylamine borane
decomposition at atmospheric pressure. The rate constant of the heterogeneous interaction between trimeꢀ
thylamine borane and an adsorption center has been determined to be 0 = 2.7 107exp(–10560/
× T) cm/s. The
ks
obtained kinetic parameters of the reaction fully determine the growth rate of nanocrystalline carbonitride
films under kinetic control. The film thickness has been determined as a function of time, temperature, reacꢀ
tant concentration, and reactor dimensions.
DOI: 10.1134/S0020168511100128
INTRODUCTION
The purpose of this work was to study the products
and kinetics of TMAB decomposition in the CVD of
nanocrystalline films.
Although B–C–N ceramic materials have long
been studied and have found industrial applications,
the deposition and the physicochemical and mechanꢀ
ical properties of thin B–C–N films have attracted
researchers' interest only in the past fifteen years.
EXPERIMENTAL
Trimethylamine borane was decomposed at atmoꢀ
spheric pressure in flowing nitrogen in a flow reactor
(Fig. 1). We used boron hydride–trimethylamine
manufactured by Merck (catalog no. 8.21915.0010).
The furnace temperature was varied from 200 to
Hexagonal (graphiteꢀlike) boron carbonitride
ꢀBCxNy) may be intermediate in properties between
(h
graphite (a good conductor) and ꢀBN (dielectric) [1,
h
2]. BCxNy films are potentially attractive for basic
technologies of nextꢀgeneration memory devices and
as protective coatings in semiconductor devices, e.g.,
in nextꢀgeneration microprocessors [3].
800 C
, the heated volume in the reactor was 7.7 cm3,
°
and the gas flow rate was 3 cm3/s. The TMAB content
was 0.45%. The residence time of the substance in the
flow reactor was 1–2 s.
Golubenko et al. [2] presented a theoretical analyꢀ
sis of the feasibility of boron carbonitride chemical
vapor deposition (CVD) from trimethylamine borane
(TMAB), (СН3)3N : BH3. They used thermodynamic
modeling to determine the vapor composition and
identify the condensed phases in equilibrium with the
vapor phase. According to their results, the set of conꢀ
densed phases in equilibrium with the vapor phase
depends to a significant degree on the initial composiꢀ
tion of the vapor mixture and the deposition parameꢀ
ters. CVD from TMAB + nitrogen mixtures leads to
the formation of BN + C mixtures over the entire
range of conditions studied. Thus, in the case of
TMAB + nitrogen mixtures, carbon (graphite) is
deposited together with boron nitride.
These experimental conditions were chosen
because TMAB actively decomposes not only in the
vapor phase but also on the reactor wall. The products
of the two TMAB decomposition paths are, most
likely, identical. Based on general ideas of thermal uniꢀ
molecular reactions [4], it is reasonable to assume
that the transition pressure (p1/2) for TMAB is below
10–5 atm. Therefore, TMAB decomposition in the
vapor phase at a pressure of 1 atm is a unimolecular
process [5].
The mean square mixing length in a flow reactor is
2
At a flow rate of 5–6 cm3/s and a residence
z = 2Dt.
time in the reactor from 1 to 2 s, complete mixing takes
0.5–1 s. = 0.09 is much less than
2/(8
Data on the chemistry of trimethylamine borane and ranges from 2 to 200 depending on temperature
decomposition are not available in the literature. and is the radius of the reactor, 0.35
R
D)
2R ξ*v
(
)
ξ
(R
D =
1199

1200
KOROBEINICHEV et al.
Probe
Furnace, 100–800°C
MKS flow
controller
Thermocouple
Thermostat
N2 cylinder
Bubbler
Substance
Fig. 1. Schematic of the flow reactor.
0.05 cm2/s is the diffusion coefficient of TMAB in
nitrogen, is adsorption probability, and is the averꢀ
products were also characterized by gas chromatograꢀ
phy/mass spectrometry (GC/MS). To this end, a samꢀ
ple was taken at the reactor outlet and was then anaꢀ
lyzed on an HP 5973 gas chromatograph/mass specꢀ
trometer equipped with a Valco sixꢀport valve (0.25ꢀml
sample loop). For chromatographic separation and
analysis, we used an HPꢀ5MS capillary column (30 m
0.25 mm). Samples were taken by freezing decomposiꢀ
tion products from the reactor outlet in a nitrogen trap.
ξ
v
age velocity of molecules). Therefore, the decomposiꢀ
tion on the reactor wall is kinetically controlled.
Accordingly, the decomposition rate on the wall will
be determined by the rate of interaction with an
adsorption center. The rate constant of the elementary
reaction
×
ТМАB + s → ТМАBads
To obtain signal intensities for individual subꢀ
stances, one should know the contribution of each
substance to the signal at a given ionization energy. We
relied on the 70ꢀeV contributions reported in the literꢀ
ature and calibrations at a reactor temperature of
373 K. To obtain signal intensities for individual subꢀ
stances, we subtracted contributions from the other
substances. Some of the peaks were separated by
another technique: by lowering the ionizing electron
energy to the level where fragment peaks disapꢀ
peared.
is
k = S V v 4 exp −E RT
,
(
)(
)
(
(
)
)
s
s
where Es is the activation energy,
wall, and is the reactor volume.
Thus, TMAB decomposes as a result of concurrent
processes. If the amount of adsorption centers far
exceeds the initial amount of the substance, the rate
equation for the decomposition process can be written
in the form
S
is the area of the
V
d ТМАB dt = −k ТМАB − k ТМАB .
[
]
[
]
[
]
s
m
RESULTS AND DISCUSSION
Samples for analysis of decomposition products
were taken 8 mm from the end of the reactor on its axis
Figures 2 and 3 present primary data on the intenꢀ
using a system that incorporated an MS7302 softꢀionꢀ sity of mass peaks versus reactor temperature for a
ization quadrupole mass spectrometer and molecular sample taken at the reactor outlet. Starting at 488 K,
beam sampler [6]. The system has high sensitivity and the intensity of the main peak (72 amu) in the mass
enables one to measure low concentrations (down to spectrum of TMAB (parent peak at 73 amu) drops
10 ppm) of not only stable but also labile substances rather rapidly. At the same time, the peak at 59 amu
(such as atoms and free radicals). The decomposition rises rapidly with increasing reactor temperature. This
INORGANIC MATERIALS Vol. 47
No. 11
2011

THERMAL DECOMPOSITION OF TRIMETHYLAMINE BORANE AS A PRECURSOR
1201
2000
1500
1000
500
3000
5
2600
5
4
3
2
1
4
3
2200
1800
1400
1000
600
2
1
0
200
300
400
500
600
С
700
200
0
–200
–500
–1000
Reactor temperature
,
°
200
300
400
500
600
700
Reactor temperature, °С
Fig. 3. Intensity of mass peaks at (
) 28 (13.5 eV), and ( ) 25 amu as a function of reactor
temperature.
1) 56, (2) 30, (3) 29,
Fig. 2. Intensity of mass peaks at (
1
) 17, (
2
) 16, (
3
) 15,
(
4
5
(4
) 14, and (
5
) 13 amu as a function of reactor temperature.
peak is the main one in the spectrum of trimethyꢀ temperature, we detect dimethylamine (DMA),
lamine (TMA), (CH3)3N. Independent GC/MS analꢀ
yses confirmed that TMA was the main TMAB
decomposition product.
(CH3)2NH, whose concentration peaks at 723 K. At
873 K, we observe a maximum in the signal intensity of
another TMAB decomposition product: monomethyꢀ
lamine (MMA), (CH3)NH2. Starting at 633 K, the
methane (CH4) concentration at the reactor outlet
increases steadily. There are also trace levels of BH3
and ammonia. After each experiment, we find a carꢀ
bon black deposit on the reactor wall. In mass spectra,
carbon black or solid carbon cannot be detected. At
the same time, the percentage of solid carbon can be
determined using the carbon balance. This is quite jusꢀ
tified and, since no carbonꢀcontaining products other
Figure 4 presents GC/MS analysis data for a samꢀ
ple taken at the reactor outlet.
Figure 5 shows the TMAB and decomposition
product concentrations at the reactor outlet as funcꢀ
tions of temperature. As seen, the TMAB concentraꢀ
tion at the reactor outlet decreases rapidly with
increasing temperature. In the range 488–650 K,
TMAB decomposes almost completely, releasing
TMA vapor. Its concentration peaks near 633,
approaching the initial TMAB concentration. At this than those mentioned above were detected, the perꢀ
ТМА
ТМАB
ТEА
1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2
Fig. 4. GC/MS analysis data:
T = 523 K, p = 1 atm.
r
INORGANIC MATERIALS Vol. 47
No. 11 2011

1202
1.4
KOROBEINICHEV et al.
reaction is firstꢀorder, the variation in TMAB concenꢀ
tration is given by
Carbon balance
1.2
1.0
0.8
0.6
0.4
0.2
0
A A = exp −kτ ,
(
)
0
where
A
is the mole fraction of TMAB at the reactor
is
outlet, А0 is the initial TMAB concentration, and
τ
С
the residence time in the reactor. If the initial step of
TMAB decomposition is BH3 removal and trimethyꢀ
lamine formation, we can evaluate the energetics of
the reaction using earlier data [7–12]:
ТМАB
ТМА
СН4
CH N ⋅ BH → CH N+BH3,
(
)
(
)
3
3
3
3
3
DМА
ММА
ΔrH =–34.08 kcal/mol.
400
500
600
700
800
900 1000
According to Issoire and Long [8], the enthalpy of
TMAB formation is –20 kcal/mol. Tel’noi and
Rabinovich [9] also calculated the enthalpy of TMAB
formation: –14.28 kcal/mol. We used these data. As a
result, TMAB decomposition was shown to be an
endothermic reaction. Accordingly, the activation
energy for the unimolecular decomposition reaction
in the vapor phase should exceed 34 kcal/mol and
should be ~40 kcal/mol.
Temperature, K
Fig. 5. TMAB and decomposition product concentrations
at the reactor outlet as functions of temperature.
centage of solid carbon can be determined using the
carbon balance.
Figure 7 compares the TMAB concentration proꢀ
files calculated for different activation energies and
preexponential factors. As seen, the dominant step of
the process at temperatures below 650 K is TMAB
decomposition on the reactor wall. The unimolecular
decomposition reaction in the vapor phase can comꢀ
pete with this process only at higher temperatures.
Figure 8 shows Arrhenius plots of the TMAB
decomposition rate constant. At temperatures above
675 K, the rate constant of the unimolecular decomꢀ
position in the vapor phase increases markedly.
Clearly, 675 K is the optimal temperature for chemical
vapor deposition.
To determine the order of the reaction, we meaꢀ
sured the degree of TMAB decomposition as a funcꢀ
tion of initial concentration at a reactor temperature
of 573 7 K. The degree of decomposition in this temꢀ
perature range was found to be ~50 2%. The initial
trimethylamine borane concentration was varied by
changing the bubbler temperature. The carrier gas
flow rate was 3.27 cm3/s (at 293 K).
Figure 6 plots the degree of TMAB decomposition
under these conditions versus initial TMAB concenꢀ
tration. Experimental data demonstrate that TMAB
decomposition is a firstꢀorder reaction. Using these
data, we can find parameters of an effective TMAB
From the above results, we estimated the enthalpy
thermolysis reaction. Under the assumption that the of TMAB adsorption on quartz glass: H = 21.12 –
Δ
0.8
7
–1
0.5
0.4
0.3
0.2
0.1
0
A
= 9
×
10
s
,
E
= 21120 cal/mol
14 –1
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
A
= 1
×
10
s
,
E
= 40000 cal/mol
0.5 0.02
0.1 0.2
0.3 0.4
0.5 0.6
0.7
0.8
300
400
500
600
700
800
Reactor temperature, K
[ТМАB]0, %
Fig. 7. Degree of TMAB decomposition as a function of
reactor temperature. The points represent experimental
data and the curves represent calculation results.
Fig. 6. Degree of TMAB decomposition as a function of
initial TMAB concentration.
INORGANIC MATERIALS Vol. 47
No. 11
2011

THERMAL DECOMPOSITION OF TRIMETHYLAMINE BORANE AS A PRECURSOR
1203
10
5
14
A
= 1
×
10
,
E
= 40000 cal/mol
7
A
= 9 × 10 , E = 21120 cal/mol
0
Spectrum 2
–5
–10
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2
103/ , K–1
60 µm
T
Fig. 8. Arrhenius plots of the decomposition rate constant.
The points represent experimental data.
Fig. 9. Electron micrograph of a condensed phase deposꢀ
ited onto a quartz substrate.
40 = –18.88 kcal/mol. We can also determine the rate energy dispersive Xꢀray spectroscopy, the B : C : N
constant of the elementary reaction between TMAB atomic ratio in the fibers is 3 : 1 : 10. The B : C : N
and an adsorption center. To this end, the effective atomic ratio in the dense zones is 7.8 : 2.6 : 1.
decomposition rate constant, ks, should be divided by
the ratio of the reactor area to its volume (3.35). In this
The set of condensed phases in films depends to a
significant degree on the initial vapor mixture compoꢀ
way we find that the rate constant of the heterogeneous
sition and temperature [2]. According to thermodyꢀ
reaction is
namic modeling results, the phase forming at low temꢀ
peratures (<500 K) is boron nitride. At temperatures
k0 = 2.7 ×107 exp −21120 RT сm s.
(
)
s
above 1055 K, a threeꢀphase mixture, BN + C + B4C
forms. Its elemental composition indicates that the
fibers consist mainly of C + B4C
,
When heterogeneous TMAB decomposition preꢀ
vails, the thickness of the film as a function of time and
temperature is proportional to the rate of interaction
with adsorption centers:
.
It can be seen from Fig. 5 that the fraction of carꢀ
bon in the solid phase varies considerably with temꢀ
perature. At 723 K, it reaches 20% of the total amount
of carbon in TMAB. A similar situation holds in the
case of nitrogen: more than 80% of the nitrogen atoms
leave the reactor in the form of TMA vapor, without
depositing on the substrate. The film forming under
such conditions consists of three phases in dense
zones, with a BN : C : B4C ratio of 12 : 10 : 21.
L t = V * ТМАB
V
] [
Sr 1 − exp S k0 t V
.
)
)
( )
[
]
(
(
r
r
s
r
Here,
V*
=
10–21 cm3 is the average “packing” size;
[TMAB] is the TMAB concentration (molecule/cm3);
S
r and
and is the deposition time.
At = 3.9 Pa, = 670 K, a TMAB content of 5%
Vr are the reactor area and volume, respectively;
t
p
Т
in the mixture with nitrogen, and an average molecuꢀ
lar size in the film of 10–21 cm3, the film growth rate is
10 nm/min, in good agreement with earlier data [13].
Figure 9 is a scanning electron microscopy image
of a film grown on a quartz glass substrate. The substrate
was mounted at the reactor outlet, and the reactor and
substrate temperatures were maintained at 723 K. Analꢀ
ysis showed that the film was inhomogeneous and conꢀ
sisted of thin “fibers” and dense zones, lending supꢀ
port to our assumption that there were two TMAB
decomposition paths. Elemental analyses of the film
showed that the fibers contained no nitrogen. It seems,
therefore, likely that the dense zones result from
decomposition in reactions with adsorption centers on
CONCLUSIONS
We have studied the products and kinetics of
TMAB decomposition in flowing nitrogen in a flow
reactor.
The kinetic parameters of the rate constant of
TMAB adsorption on quartz have been determined as
functions of temperature. These parameters fully
determine the growth rate of nanocrystalline carboniꢀ
tride films under kinetic control.
TMAB decomposition is a firstꢀorder reaction.
The first step of the process is TMAB adsorption on
the substrate surface and that the fibers result from the the reactor wall. The next step is BH3 removal and the
formation of BH3 and carbon black in the vapor phase, formation of trimethylamine at temperatures below
which then adsorb on the substrate. According to 650 K or a threeꢀphase film at higher temperatures.
INORGANIC MATERIALS Vol. 47
No. 11 2011

1204
KOROBEINICHEV et al.
Pittsburgh, PA: Combustion Institute, 1996, pp. 1035–
REFERENCES
1042.
1. Kuznetsov, F.A., et al., Materials and Basic Technoloꢀ
gies of NextꢀGeneration Electronic Devices: Dielectric
Layers, Izv. Vyssh. Uchebn. Zaved., Mater. Elektron. Tekh.,
2007, vol. 4, pp. 54–62.
7. Chase, M.W., NIST–JANAF Thermochemical Tables,
Fourth Edition, J. Phys. Chem. Ref. Data, 1998, monoꢀ
graph 9, p. 1ꢀ1951.
8. Issoire, J. and Long, C., Etude de la thermodynamique
chimique de la reaction de formation des methyꢀ
lamines, Bull. Soc. Chim. Fr., 1960, pp. 2004–2012.
9. Tel’noi, V.I. and Rabinovich, I.B., Thermal Chemistry
of Main Group Element Organic Compounds, Usp.
Khim., 1980, vol. 49, no. 7, pp. 1137–1173.
10. Sana, M., Leroy, G., and Wilante, C., Enthalpies of
Formation and Bond Energies in Lithium, Beryllium,
and Boron Derivatives: 2. Dative, Single, and Triple
Bonds, Organometallics, 1992, vol. 11, no. 2, pp. 781–
787.
11. Leroy, G., Sana, M., and Wilante, C., Evaluation of the
Bond Energy Terms for the Various Types of Boron–
Nitrogen Bonds, Theor. Chem. Acta, 1993, no. 85,
pp. 155–166.
12. Grant, J., et al., Revised Heat of Formation for Gasꢀ
eous Boron: Basis Set Limit Ab Initio Binding Energies
of BF3 and BF, Phys. Chem. A, 2009, vol. 113,
pp. 6121–6132.
13. Kosinova, M.L., Rumyantsev, Yu.M., Golubenko, A.N.,
et al., Chemical Composition of Boron Carbonitride
Films Grown by PlasmaꢀEnhanced Chemical Vapor
Deposition from Trimethylamineborane, Inorg. Mater.,
2003, vol. 39, no. 4, pp. 366–373.
2. Golubenko, A.N., Kosinova, M.L., Titov, A.A., and
Kuznetsov, F.A., Thermodynamic Modeling of BCxNy
Chemical Vapor Deposition in the B–C–N–H Sysꢀ
tem, Inorg. Mater., 2003, vol. 39, no. 4, pp. 362–365.
3. Baake, O., Hoffmann, P.S., Klein, A., et al., Chemical
Character of BCxNy Layers Grown by CVD with Trimꢀ
ethylamine Borane, www.interscience.wiley.com/jourꢀ
nalxrs, Wiley, 2008.
4. Robinson, P.J. and Holbrook, K.A., Unimolecular
Reactions, London: Wiley, 1972. Translated under the
title Monomolekulyarnye reaktsii, Moscow: Mir, 1975,
pp. 125–127.
5. Kosinova, E.A., Sulyaeva, V.S., and Sysoev, S.V., Temꢀ
peratureꢀDependent Triethylamine Borane Vapor Presꢀ
sure, XLI Mezhdunarodnaya nauchnaya studencheskaya
konferentsiya “Student i nauchnoꢀtekhnicheskii progress”
(Khimiya) (XLI Int. Scientific Student Conf.: Student
and Progress in Science and Technology (Chemistry),
Novosibirsk, 2003, pp. 145–146.
6. Chernov, A.A., Korobeinichev, O.P., Shvartsberg, V.M.,
and Mokrushin, V.V., Destruction of Trimethyl Phosphate
in H2/O2/Ar Flame, 26th Int. Symp. on Combustion,
INORGANIC MATERIALS Vol. 47
No. 11
2011