Shock tube study of the reaction of H atoms with TiCl4

Article abstract of DOI:10.1021/jp9718414

The reaction of H atoms with TiCl₄ was studied behind reflected shock waves at temperatures between 1190 and 1500 K and pressures around 1.5 bar by applying atomic resonance absorption spectrometry for time-resolved measurements of H atoms at the L_α line. The thermal decomposition of ethyl iodide (C₂H₅I) served as the H atom source. By using a high excess of TiCl₄, the consumption of H could be modeled by pseudo first-order kinetics. For the reaction TiCl₄ + H → TiCl₃ + HCl a temperature-independent rate coefficient of k₁ = 3.5 ± 0.9 × 10¹³ cm³ mol^-1 s^-1 was found for the temperature range of the present experiments.

Full text of DOI:10.1021/jp9718414

J. Phys. Chem. A 1997, 101, 9341-9343
Shock Tube Study of the Reaction of H Atoms with TiCl4
J. Herzler and P. Roth*
9341
Institut f u¨ r Verbrennung und Gasdynamik, Gerhard-Mercator-UniVersit a¨ t Duisburg, 47048 Duisburg, Germany
X
ReceiVed: June 5, 1997; In Final Form: September 17, 1997
The reaction of H atoms with TiCl
and 1500 K and pressures around 1.5 bar by applying atomic resonance absorption spectrometry for time-
resolved measurements of H atoms at the L line. The thermal decomposition of ethyl iodide (C I) served
as the H atom source. By using a high excess of TiCl , the consumption of H could be modeled by pseudo
first-order kinetics. For the reaction TiCl + H f TiCl
4
was studied behind reflected shock waves at temperatures between 1190
R
2 5
H
4
4
3
+ HCl a temperature-independent rate coefficient
13
3
-1 -1
of k
1
) 3.5 ( 0.9 × 10 cm mol
s
was found for the temperature range of the present experiments.
1. Introduction
TiCl4 is the most commonly used precursor for the production
1,2
of industrially very important products such as titania and TiN.
High-temperature gas-phase processes, like chemical vapor
deposition (CVD), high-temperature reactors, and flames are
often used for the production of these substances. Titania is
synthesized on a large scale in flame reactors by the “chloride”
3
process using TiCl4, oxygen, and hydrocarbons along with
various additives as educts. The reaction of TiCl4 + H is one
of the important starting reactions during the oxidation of TiCl4
in the flame. It is also part of the chain reactions by which
TiCl4 decomposes (TiCl4 + H f TiCl3 + HCl, TiCl3 + H2 f
TiCl3H + H) under lower temperatures CVD conditions.
For the further development of these high-temperature gas-
phase production processes, the knowledge of the elementary
kinetic steps is necessary. Despite the industrial relevance of
the titanium compounds, there are no kinetic high-temperature
data on reactions of TiCl4.
Figure 1. H atom calibration curve. Solid line: fit, A ) 1 -
H
-8
0.65
0.55
exp(-7.9 cm × 1.17 × 10 cm × [H] ).
-
8
down to pressures below 1 × 10 mbar by a turbo molecular
pump. Gas mixtures were prepared manometrically in a
stainless steel UHV storage cylinder, which also could be baked
out and evacuated using a separate turbo molecular pumping
unit. The residual gases in all UHV devices were analyzed by
quadrupole mass-spectrometers, and were found to be practically
free of hydrocarbons. More details of the setup are given
In the present study the rate coefficient of the reaction
TiCl + H f TiCl + HCl
(R1)
4
3
6,7
was directly measured by using the shock tube as a high-
temperature wave reactor, detecting H by atomic resonance
absorption spectrometry (ARAS). The well-known two-step
thermal decomposition of ethyl iodide (C2H5I)
elsewhere. Argon used in the present study was of the highest
commercially available purity (g99.9999%). Ethyl iodide
(g99%) and titanium tetrachloride (g99.9%), which are liquid
at normal conditions, were injected and evaporated in separated
stainless steel vessels.
C H I f C H + I
2
5
2
5
The shock tube is equipped with diagnostics for ARAS
consisting of a microwave excited discharge lamp, the optical
absorption path in the shock tube, a 1 m vuv monochromator,
and a solar blind photomultiplier. The lamp was operated with
a flowing gas mixture of 1% H2 in He maintained at a pressure
of about 5 mbar and a microwave power of about 50 W. The
spectral shape of the LR line (λ ) 121.6 nm) emitted by the
resonance lamp is not precisely known due to the influence of
self-absorption and self-reversal in the lamp gas. Hence precise
calibration measurements were made to relate measured absorp-
tion to the corresponding H atom concentration by using the
reaction of O atoms with H2 to produce H atoms.8 This
procedure is suitable because the thermal decomposition of H2
is very slow at temperatures below 2100 K. The source of O
atoms was the dissociation of N2O, which is relatively fast.
Therefore, a mixture of 1 ppm N2O and 200 ppm H2 is suitable
to cover the temperature range from 1500 to 2100 K.10
The results of the calibration procedure shown in Figure 1
can be expressed by a modified Lambert-Beer law:
C H f C H + H
2
5
2
4
was used to produce H atoms.4 At temperatures T g 1190 K
and low initial concentrations, the H formation is very fast
(
>97% within 30 µs) and no subsequent reactions of H, I, and
C2H4 were observed so that ethyl iodide is a well-defined H
atom source at these conditions.5
2
. Experimental Section
,9
The experiments were carried out in a stainless steel pressure
driven shock tube with an internal diameter of 79 mm. It is
divided by a thin aluminum diaphragm into a driver section of
.5 m and a driven section of 5.7 m in length. The internal
surface has been specially prepared for ultrahigh-vacuum (UHV)
purposes. The driven section can be baked out and pumped
3
X
Abstract published in AdVance ACS Abstracts, November 15, 1997.
S1089-5639(97)01841-0 CCC: $14.00 © 1997 American Chemical Society

9342 J. Phys. Chem. A, Vol. 101, No. 49, 1997
Herzler and Roth
Figure 2. Typical example of L
experiment. Mixture: 235 ppm TiCl
kPa.
R
absorption of TiCl
4
in a shock tube
Figure 4. Typical example of L
Ar gas mixture behind a reflected shock wave. Mixture: 32.5 ppm
TiCl and 2 ppm C I in Ar. T ) 1191 K, p ) 166 kPa.
R 4 2 5
absorption obtained in a TiCl /C H I/
4
in Ar. T ) 1130 K, p
5
5
) 151
4
2
H
5
5
5
Figure 5. First-order plot deduced from the absorption signal given
in Figure 4.
Figure 3. Measured absorption cross section of TiCl
temperature range 1130-1500 K.
4
at L
R
in the
(1)
0.55
In mixtures containing C2H5I, the interference absorption of
ethyl iodide and its molecular decomposition products were
assumed to be negligible because of the very low concentration
of these species in the present reaction systems. The temper-
A ) 1 - exp(-lσ [H]
)
H
L
R
with l being the absorption path length. The absorption cross
-
8
0.65
section was found to be σLR ) 1.17 × 10 cm
and the
ature range of our TiCl + H study was 1190 to 1500 K at
4
concentration exponent of 0.55 was introduced to fit the
Lambert-Beer expression of eq 1 to the calibration results. The
slight temperature dependence of the calibration curve does not
influence the first-order analysis of our data because we need
only the concentration exponent which is temperature indepen-
dent within the measurement uncertainty from 1200 to 2100
K.
pressures of about 1.5 bar. The lower limit was determined by
the requirement of fast ethyl iodide decomposition, the upper
11,12
limit by the stability of TiCl4.
Mixtures of 15-40 ppm TiCl4
and 2 ppm C2H5I in Ar were used. This high TiCl4 excess
reduces the importance of subsequent reactions so that the rate
coefficients could be determined by a pseudo first-order analysis.
A typical example of an absorption profile is given in Figure
4
. The signal is composed of two parts: the interference
3
. Results and Discussion
absorption of TiCl4 (dashed line) and the absorption by H atoms.
The measured absorption shows a fast increase at time zero and
reaches a maximum at about 50 µs, followed by a fast decay to
the interference absorption. All measured absorption profiles
were transferred into first-order plots by applying the calibration
For all the results presented in this paper, the absorption of
the LR radiation by the initial molecular reactant TiCl4 was taken
into account. In a first series of experiments the molecular
absorption cross section σLR was determined by performing
experiments with mixtures containing 100-235 ppm TiCl4 in
Ar. A typical example of TiCl4 absorption of LR radiation is
shown in Figure 2. The incident (IS) and reflected shock (RS)
cause a steplike increase of the absorption to a level of about
relation of eq 1. For the example of Figure 4, this plot is given
in Figure 5, where the value ln[(-ln(1 - AH))(
1/0.55)
] is plotted
against the reaction time. The slope is equal to the inverse time
-
1
constant τ .
6
3%. The results obtained from a total of 12 individual shock
-
1
d
dt
(1/0.55)
tube experiments are summarized in Figure 3. The temperature-
dependent absorption cross section of TiCl4 for LR radiation
can be approximated in the temperature range from 1120 to
τ
) - ln[(-ln(1 - A ))
]
(2)
H
where AH is the absorption being related to H atoms. The rate
coefficients of reaction R1 can directly be related to the
measured time constants τ :
1
500 K by
-
1
σ (TiCl ) )
L
4
R
.9 × 10 cm-2 - 3.7 × 10 cm-2 K-1 × T
-
17
-20
-1
9
k ) τ /[TiCl ]
(3)
1
4 0
The absorption cross section at room temperature is about
The individual k1 values obtained from all experiments are
summarized in the Arrhenius diagram of Figure 6. The rate
-
17
-2
7
.5 ( 0.8 × 10
cm .

Reaction of H Atoms with TiCl4
J. Phys. Chem. A, Vol. 101, No. 49, 1997 9343
TABLE 1: Rate Coefficient Data for the Reactions of Chlorinated Compounds + H to HCl + Products
A (cm mol s-1
3
-1
)
E
/R (K)
T-range (K)
authors
this work
Catoire et al. [13]
Exton et al. [14]
Jourdain et al. [15]
Sung et al. [16]
Jourdain et al. [17]
Bradley et al. [18]
Combourieu et al. [19]
Manion et al. [20]
Jones, Ma [21]
reaction
+ H f TiCl
+ H f SiCl
+ H f NCl
a
1
3
TiCl
SiCl
NCl
4
3
+ HCl
+ HCl
+ HCl
+ HCl
+ HCl
+ HCl
3.5 × 10
1.4 × 10
1190-1500
1530-1730
300
298
298
295-807
538-676
298-460
998
1
3
4
3
4800
1
1
2
2
3
2
2.41 × 10
1.45 × 10
PCl
PCl
3
+ H f PCl
+ H f PCl
2
3
2
0.05 × kCl +H
2
1
1
1
1
8
1
1
1
3
3
1
BCl
3
+ H f BCl
2
1.02 × 10
1.70 × 10
1.08 × 10
5.01 × 10
4.82 × 10
3.70 × 10
1.10 × 10
1232
4426
3069
CF Cl + H f CF
3
3
+ HCl
Cl + HCl
CH
CH
CH
CH
CH
2
3
3
3
3
Cl
2
+ H f CH
2
Cl + H f CH
Cl + H f CH
Cl + H f CH
Cl + H f CH
3
3
3
3
+ HCl
+ HCl
+ HCl
+ HCl
298
500-800
298-652
3
3
4680
2315
Westenberg, DeHaas [22]
Aders et al. [23]
References and Notes
(
1) Akhtar, M. K.; Vemury, S.; Pratsinis, S. E. AIChE J. 1994, 40,
1
183.
(
2) Bouteville, A; Imhoff, L; Remy, J. C. J. Electrochem. Soc. 1996,
1
43, 3251.
(3) Clark, H. B. Titanium Dioxide Pigments. In Pigments; Myers, R.
R., Long, J. S., Eds.; Marcel Dekker Inc.: New York, 1975; Part 1.
(
(
(
(
4) Herzler, J.; Frank, P. Ber. Bunsen-Ges. Phys. Chem. 1992, 96, 1333.
5) Herzler, J. Ph.D. Thesis, Stuttgart, 1994.
6) Woiki, D.; Roth, P. Shock WaVes 1994, 4, 95.
7) Thielen, K.; Roth, P. 20th Symp. (Int.) on Comb. 1984, The
Combustion Institute: Pittsburgh, PA, 685.
8) Appel, D; Appelton, J. P. 15th Symp. (Int.) on Comb. 1975, The
Combustion Institute: Pittsburgh, PA, 701.
9) Just, Th. Atomic Resonance Absorption Sperctrometry in Shock
Tubes. In Shock WaVes in Chemistry; Lifshitz, A., Ed.; Marcel Dekker
Inc.: New York, 1981.
(
(
Figure 6. Arrhenius plot for the rate coefficient k
1
.
(
(
(
(
10) Frank, P.; Just, Th. Ber. Bunsen-Ges. Phys. Chem. 1985, 89, 181.
11) Allendorf, M. D.; et al. Electrochem. Soc. Proc. 1995, 95-4, 393.
12) Herzler, J. Unpublished results.
coefficient is temperature-independent within the measurement
uncertainty. The mean value is
13) Catoire, L.; Woiki, D.; Roth, P. Int. J. Chem. Kinet. 1997, 29, 469.
k ) 3.5 ( 0.9 × 10 cm mol s-1
13
3
-1
(14) Exton, D. B.; Gilbert, J. V.; Coombe, R. D. J. Phys. Chem. 1991,
5, 2692.
1
9
(
15) Jourdain, J.-L.; Laverdet, G; Le Bras, G.; Combourieu, J. J. Chim.
Phys. 1980, 77, 809.
16) Sung, J. P.; Setser, D. W.; Heydtmann, H. Ber. Bunsen-Ges. Phys.
where the uncertainty is given at the 1σ level.
(
A comparison of k1 with literature values for this reaction is
not possible because there are no data available. The only rate
coefficients available for the reaction of fully chlorinated
compounds with H and similar reactions of partially chlorinated
organic molecules + H producing HCl are summarized in Table
Chem. 1979, 83, 3, 1272
(17) Jourdain, J.-L.; Laverdet, G.; Le Bras, G.; Combourieu, J. J. Chim.
Phys. 1981, 78, 253.
(
18) Bradley, J. N.; Whytock, D. A.; Zaleski, T. A. J. Chem. Soc.,
Faraday Trans. 1 1976, 72, 2284.
19) Combourieu, J.; Le Bras, G.; Paty, C. 14th Symp. (Int.) on Comb.
(
1
3
1
. It can be seen that an A factor of 3.5 × 10 , which was
1973, The Combustion Institute: Pittsburgh, PA, 485.
(20) Manion, J. A.; Dijks, J. H. H.; Mulder, P.; Louw, R. Recl. TraV.
Chim. Pays-Bas 1988, 107, 434.
determined in this work, is characteristic for this kind of reaction.
All the reactions show only a small temperature dependence
(
(
(
21) Jones, W. E.; Ma, J. L. Can. J. Chem. 1986, 64, 2192.
22) Westenberg, A. A.; Dehaas, N. J. Chem. Phys. 1975, 62, 3321.
23) Aders, W.-K.; Pangritz, D.; Wagner, H. Gg. Ber. Bunsen-Ges. Phys.
13
3
-1 -1
(
assuming an A factor of 3 × 10 cm mol s , the activation
energies of NCl3 + H and PCl3 + H are less than 2 kcal/mol)
but our results are the only ones showing no activation energy.
Chem. 1975, 79, 9, 90