J. Phys. Chem. 1994, 98, 12958-12963
12958
Kinetics of the High-Temperature H2S Decomposition D. Woiki and P. Roth* Institut f i r Verbrennung und Gasdynamik, Universitat Duisburg, 47048 Duisburg, Germany Received: July 13, 1994; In Final Form: September 27, 1994@
The thermal decomposition of 5-100 ppm H2S diluted in Ar was studied behind reflected shock waves at temperatures 1887 K IT I2891 K and pressures around 1.3 bar by applying atomic resonance absorption spectroscopy (ARAS) for time-resolved concentration measurements of H and S atoms. Both the S and H concentration profiles showed almost linear increases at early reaction times with the S atoms exceeding the H atoms by a factor of 10-20. Therefore reaction R1, H2S Ar H2 S Ar (rate coefficient k l ) , was regarded as the initial step in the HzS decomposition. The rate coefficient kl was determined from the slope of the early S concentration profiles to be kl = 1.9 x 1014exp(-32860 WT)cm3 mol-' s-'. The subsequent reaction between H2S and S atoms (reaction R2), H2S S products (rate coefficient kz), was investigated in two different manners: first by evaluating the quasi-stationary S concentrations observed at longer reaction times in pyrolysis experiments of 100 ppm H2S and second by monitoring the decay of photolytically generated S atoms in laser flash photolysis-shock wave experiments with 30 ppm CS2 and 50-150 ppm HzS. Both groups of experiments covered the temperature range 1340 K IT I2120 K and result in a rate coefficient k2 = 5.7 x 1014exp(-7600 WT)cm3 mol-' s-'. H concentration profiles measured during H2S/Ar pyrolysis were analyzed using a simplified reaction mechanism, which was able to predict the experimental findings. In that case it was necessary to introduce a reaction channel (R2a), forming the reaction products HS2 and H, with an efficiency of 35-57% of the overall reaction R2.
+
+
Introduction The pyrolysis of H2S behind shock waves was studied by Bowman and Dodge' in the temperature range 2700 K 5 T 5 3800 K at relative concentrations of 7000-25000 ppm H2S diluted in Ar by calibrated molecular UV absorption measurements of H2S. Concentration measurements of S atoms were indirectly performed by monitoring the UV emission of electronically excited S2 molecules formed by the recombination of S atoms. The authors assumed the abstraction of H atoms to be the primary decomposition reaction H2S
+ Ar -SH + H + Ar k,*
AH298= 377.8 kJ/mol
and determined the corresponding rate coefficient to be kl* = 2.0 x 1014exp(-37300 WT)cm3 mol-' s-l by comparing their experimental results with computer simulations. Bowman and Dodge' also compared their measured S concentration profiles with the results of their computer simulations and obtained agreement within the experimental error limits. In the examples given in ref 1 the calculated S profiles appear to have a more pronounced induction period than observed in the experiments. This might indicate an incomplete description of the early steps in the H2S decomposition by the reaction mechanism reported in ref 1. In another shock tube study of Higashihara et al.2 the formation of S2 during the pyrolysis of 50 000 ppm H2S diluted in Ar was measured by W-absorption in the temperature range 2380 K 5 T 5 3000 K. The authors observed no induction periods of the S2 formation in their experiments and summarized their results by [Sz](t)= [S2]-(1 - exp(-kt)) with k = 1.6 x lo9 exp(-36240 WT) s-l. Although the authors could not determine the mechanism of the S2 formation, they assumed reaction R1* to be the primary step and estimated a rate @
Abstract published in Advance ACS Abstracts, November 1, 1994.
0022-3654/94/2098-12958$04.50/0
-
+ +
-
coefficient of kl* = 1.3 x l O I 3 exp(-46300 WT) cm3 mol-' s-'. However, it is difficult to verify the observation of no induction period for the S2 formation by computer simulations using reaction R1* as the only initial decomposition step. Atomic resonance absorption spectroscopy (ARAS) measurements of H atom concentrations during the pyrolysis of H2S behind shock waves in the temperature range 1965 K IT 5 2560 K were performed by Roth et aL3 The authors used very low initial concentrations of 25-200 ppm H2S diluted in Ar and observed almost linear increases of the H concentrations at early reaction times which they interpreted by a direct H formation via reaction R1*. The corresponding rate coefficient was reported to be kl* = 4.6 x 1014 exp(-41500 WT) cm3 mol-l s-'. In some experiments at comparably low temperatures Roth et aL3 found quasi-stationary H concentrations at longer reaction times which were explained by the competition between formation and consumption of the H atoms via reactions R1* and R3, respectively: H,S
+ H -.SH + H, k3
(R3)
From a steady-state analysis of the H atoms the authors estimated a rate coefficient of k3 = 1.1 x 1013exp(- 1500 WT) cm3 mol-' s-1. Yoshimura et aL4 performed laser flash photolysis experiments in highly diluted, shock heated H2S/Ar reaction systems and monitored the decay of photolytically generated H atoms in the temperature range 1053 K 5 T 5 2237 K. These experiments provide pseudo-first-order conditions for reaction R3, and a rate coefficient of k3 = 1.9 x loL4exp(-2491 WT) cm3 mol-' s-' was directly determined. This k3 value is a factor of about 11 higher than the value given by Roth et aL3 It must be pointed out that the quasi-stationary H concentrations measured in ref 3 cannot be verified by computer simulation using the initial reaction R1* and the k3 value of ref 4. It is the aim of the present study to provide additional information on H2S pyrolysis and to solve the contradiction 0 1994 American Chemical Society
J. Phys. Chem., Vol. 98, No. 49, 1994 12959
Kinetics of the High-Temperature H2S Decomposition between the results of Roth et aL3 and Yoshimura et al.4 Therefore, both S and H atom concentration profiles were measured during the high-temperature, low initial concentration thermal decomposition of hydrogen sulfide. A simplified reaction mechanism for the formation and consumption of S is proposed including reactions R1 and R2:
H,S
+
H2S
H,
+ s + ~r
+ S -products
(R1)
.
O
E
; Y
.-0
0.5-
E
T-2482K
E
20 ppm H,S in Ar
B 9 0.0-
-
k2
(R2) 0
The H concentration profiles obtained are sensitive to the HS2 H product channel of reaction R2. The rate coefficient of the overall reaction R2 was determined in laser flash photolysisshock wave experiments of CSz/H2S/Ar gas mixtures.
200
3.0,
1
n
'
1
The experiments were carried out behind reflected shock waves in two different shock tubes, both made of stainless steel and specially prepared for ultra-high-vacuum (UHV) purposes?-8 They can be baked out and pumped down to pressures below 5 x mbar by turbo molecular pumps. Both shock tubes are equipped with diagnostics for ARAS consisting of a microwave excited discharge lamp, the optical absorption path in the shock tube, a vacuum UV monochromator, and a solar blind photomultiplier. Gas mixtures were prepared manometrically in stainless steel UHV storage cylinders, which were also baked out and were evacuated using separate turbo molecular pumping units. The residual gases in all UHV devices were analyzed by quadrupole mass spectrometers and were found to be practically free of hydrocarbons. All gas mixtures used in the present study were of the highest commercially available purity, i.e., Ar 2 99.9999%, H2S 2 99.9%, and CS:! I 99.9%. The experiments on the H2S pyrolysis were performed in a 79 mm internal diameter shock tube consisting of a 3.5 m long driver section and of a 5.7 m long driven section. The spectral lines of the H and S atoms were separated by a McPherson 1 m vacuum W monochromator with a spectral resolution of 0.25 nm; for details, see refs 7 and 8. The shock tube used for the laser flash photolysis (LFP) experiments has an internal diameter of 80 mm and consists of a 3.5 m long driver section and of a 6 m long driven section. The spectral line of the S atoms was separated with a spectral resolution of 0.5 nm by a McPherson 0.5 m vacuum W monochromator. The laser light (11 = 193 nm) of a Lambda Physik ArF excimer laser was coupled into the measurement plane of the shock tube through an end plate made of quartz glass. The originally rectangular laser beam was expanded by a cylindrical lens allowing the illumination of the whole ARAS diagnostic path, for details see ref 6. The laser was operated in single pulse mode. It has a pulse width of 13 ns and an average energy between 25 and 50 mJ per pulse, measured behind the quartz glass window of the shock tube. The calibration of the A R A S techniques of S and H atoms is based on high-temperature COS or H2 decomposition described e l s e ~ h e r e . ~Possible ~ ~ . ' ~ absorption of the resonance radiation of the S atoms (SI) and of the H atoms (La) by the molecular reactants H2S, CS2, and H2 was also taken into account. The measured absorption cross sections of the species for the SI or the La radiation were determined to be I5.4 x 1 0 - l ~cm2
os,(H,) I2.8 x
cm2
800
'
1
'
1
'
d[S]/dtlo= 1.28 x 10" cm.)/i3
5
*
z.2.0
Experimental Section
600
400 Time I ps
+
u,,(H,s)
1
w
W
*
C 'P 0
e
*
5 1.0 C
8 v)
0.0
0
400
200
600
800
Time I p s
Figure 1. Example of measured SI light extinction (upper part) and S concentration profile (lower part) obtained in a pyrolysis system of 20 ppm H2S diluted in Ar showing an almost linear S increase.
as,(CS2) = 2.7 x oLa(H2S)= 2.5(f1.5)x
cm2 (ref 8) cm2 (refs 3 and 4)
For most of our experiments the interference absorptions of molecular species are negligibly low because of the very low concentrations.
Results H2S Pyrolysis. The pyrolysis of H2S was studied behind reflected shock waves in the temperature range 1887 K IT I 2891 K at pressures around 1.3 bar. The relative initial concentrations of H2S were between 5 and 100 ppm. S or H atom resonance absorption was measured in the reacting gas mixtures. A representative absorption profile for S atoms monitored in a reaction system with comparably low initial concentrations of 20 ppm H2S is shown in the upper part of Figure 1. The corresponding S concentration profile, obtained by applying the calibration of the SI absorption, is given in the lower part of Figure 1. It is obvious that the S atom increase is nearly linear at early reaction times. A second example of S absorption profiles, typical for the experiments performed at higher initial concentrations of 100 ppm H2S, is given in Figure 2. The corresponding concentration profile is plotted in the lower part of Figure 2. In four further experiments also performed with 100 ppm H2S no distinct linear S atom increases were obtained, but curved profiles were obtained, resulting in quasi-stationary S concentrations. The measured H atom profiles were practically identical to those reported in the H2S decomposition study of Roth et aL3 Except the very early reaction time around t = 0-10 ps (time resolution of ARAS), the H atom profiles also showed an almost linear increase at early reaction times leveling off for longer times. Examples of measured H absorption and concentration profiles are given
Woiki and Roth
12960 J. Phys. Chem., Vol. 98, No. 49, I994 1
-
E
s
F
'
1
'
1
'
1
'
'oio71--l-T
T=2116K p = 1.41 bar 100 ppm H,S
l
0
'
l
200
'
l
400 Time I ps
'
l
600
'
l
800
2.0
10611
,
!H
V
1os
3.5
c,o; t, I
l5wm 2Oppm 50PPM
(3 20wm
25PPM
4.0
:r\,I
4.5
~
~
5.0
lo4KIT 1 - 0 . 0 0
200
400 Time I p s
600
800
Figure 2. Example of measured SI light extinction (upper part) and S concentration profile (lower part) obtained in a pyrolysis system of 100 ppm H2S diluted in Ar showing a quasi-stationary S concentration for longer reaction times. in ref 3. All experimental conditions together with observed S or H atom characteristics during HzS pyrolysis experiments are summarized in Table 1. For a first data reduction apparent rate coefficients were determined by dividing the measured slopes of the early S and H increase by the initial reactant concentrations [HzS] and [Ar]. The values obtained are summarized in the Arrhenius diagram of Figure 3. The data points scatter around two straight lines which can be expressed by apparent rate coefficients determined from least square fits: klS,app - 1.9 x 10'4exp(-32860
WT)cm3 mol-' s-' (1)
klH,app = 4.6 x 1014exp(-41500
WT)cm3 mol-' s-l (2)
The diagram also contains the earlier result of Roth et which is identical with the present k l ~value. , ~ ~ ~ HzS/CSz Photolysis. LFP experiments were performed behind reflected shock waves in the temperature range 1337 K IT I 1879 K at pressures around 1.1 bar in gas mixtures of 30 ppm CS2 and 50-150 ppm HzS, and the resonance absorption of S was monitored. CS2 was added to the reaction systems because the laser photolysis generation of S was not observed in H2S/Ar mixtures in similar shock tube experiments. The CS2 addition yields S atom concentrations between 2.5 and 10 ppm after the laser pulse, depending on the laser energy (see also ref 11). The high excess of H2S favors a consumption of the S atoms mainly through reaction E.The absorption profile of a typical experiment is shown in Figure 4. The arrival of the reflected shock wave in the measurement plane is associated with a weak light extinction caused by molecular absorption of HzS. After the excimer laser pulse at time zero, the signal increases step-like to a peak value followed by a decay of the signal for the remaining observation time. For a f i s t data
Figure 3. Amhenius plot for the apparent rate coefficients ~ I H and , ~ k l ~determined , ~ in the present study from H2S pyrolysis experiments compared with the results of the H formation rate of ref 3. TABLE 1: Experimental Conditions and S and H Atom Characteristics, Obtained from High Temperature HzS/Ar Pyrolysis Reaction Systems [H~SIO, (d[XlWmd[HzSl [ h l ,
T,K
P,bar
2449 2564 2232 2239 2423 2300 2100 2127 2174 2276 2282 2362 2501
1.15 1.17 1.48 1.38 1.43 1.38 1.98 1.31 1.34 1.33 1.87 1.41 1.32
ppm cm3 mol-' s-l H measurements; X = H 7.1 x 1.1 x 4.4 x 5.0 x 1.5 x 7.6 x 4.5 x 2.7 x
1015 10'6 1015 1015 10l6 1015 1015 1015 4.0x 1015 9.5 x 1015 1.4 x 10l6 1.0 x 1016 2.1 1015 S measurements;X = S 2406 1.31 5 3.2 x loL6 2646 1.18 5 7.1 x 10l6 2891 1.22 5 1.8 x 1017 2331 1.38 10 3.6 x 10l6 2421 1.31 10 7.5 x 1016 2650 1.27 10 1.4 x 1017 2721 1.21 10 1.4 x 1017 2255 1.19 15 2.8 x 10l6 2374 1.07 15 5.3 x 1016 2410 1.20 15 7.1 x 10l6 1.5 x 1017 2528 1.13 15 15 2.2 x 1017 2667 1.17 2162 1.40 20 3.5 x 1016 2229 1.38 20 4.2 x loL6 2257 1.31 20 5.1 x 10l6 2393 1.27 20 1.1 x 1017 20 2482 1.18 1.3 x 1017 1887 1.48 100 2.0 x 1012 2021 1.54 100 5.0 x 1OI2 2085 1.43 100 1.3 x 1013 2116 1.41 100 1.4 x 1013 2123 1.56 100 1.2 x 1013 reduction a value [h(-h(l- AN))] was plotted against reaction
time, and the slope
20 20 25 25 25 25 50 50 50 50 50 50 50
[XI,,,
cm-3
~
Kinetics of the High-Temperature H2S Decomposition
J. Phys. Chem., Vol. 98, No. 49, 1994 12961
E
0)
3
30 ppm CS, + 150 ppm H,S
I
'
I
'
0
I
200
'
I
400
'
I
600
800
Time I p s
Figure 4. Example of measured light extinction of SI resonance radiation in a shock heated CSfl2SIAr gas mixture photolyzed by an excimer laser flash at 1 = 193 nm. d dt
t = -[ln(
-In( 1 - AN))]
(3)
was determined for reaction times t 5 300 p s from each experiment, see insert of Figure 4. In eq 3 AN represents the normalized absorption, which was introduced instead of the measured absorption because of a small nonresonant light fraction in the discharge lamp spectrum; for details see ref 6. A summary of the LFP-shock waves experiments together with the z values obtained is given in Table 2. Discussion
A comparison of the formation rates of S and H atoms during the high-temperature, low initial concentration pyrolysis of HzS makes it evident that the apparent rate coefficient klS,app is between 10 and 20 times higher than kl~,,,, in the temperature range of the present study. Due to these experimental facts the direct formation of S atoms via reaction R1 H2S
+ Ar !!.H2 + S + Ar
AH298= 297.5 kJ/mol (R1)
has to be regarded as the major reaction channel of the H2S decomposition at high temperatures. The contribution of other reactions to the S concentrations was examined by computer simulations with sensitivity analysis12 based on the simplified H2S pyrolysis scheme of Table 3. A typical result showing the normalized sensitivity of reactions R1 -R7 to the S atom profiles is plotted in Figure 5 . As a result of the kinetic simulations it can be assumed that the initial slope of the S concentration profiles is determined practically only by reaction R1 and that secondary reactions are of minor importance in experiments with initial H2S concentrations of 5-20 ppm. Therefore the kls,app value obtained must be regarded as the rate coefficient kl of the primary decomposition step of H2S via reaction R1. In the pyrolysis experiments with higher initial concentrations of 100 ppm HzS the measured steady-state S concentrations can be interpreted by assuming only reactions R1 and R2 to be responsible for the formation and consumption of the S atoms. H,S
+ S -products k2
400 600 800 Time / ps Figure 5. Normalized sensitivity of the reaction mechanism of Table 3 with respect to S atoms.
0
'
(R2)
This assumption was verified by computer simulations using the reaction mechanism of Table 3. From both reactions it
200
TABLE 2: Experimental Conditions, S Atom Characteristics, and Time Constants G of SI Absorption Profiles Obtained from High-Temperature CS2/HZS/Ar Photolysis Reaction Systems T, K P,bar [CSZIO, ppm 1725 1.13 30 1879 1.04 30 1578 1.08 30 1589 1.09 30 1652 1.16 30 1337 1.19 30 30 1404 1.18 1451 1.14 30 1505 1.15 30 1595 1.13 30
[H~SIO, ppm 50 50 100 100 100 150 150 150 150 150
[s]photd[cs2lO~96
r, S-l
8.2 11.0 17.3 17.0 19.0 33.0 18.0 19.0 25.0 19.9
3939 4348 3226 4144 5415 4950 5000
4000 6465 5310
TABLE 3: Simplified Reaction Mechanism of the H S System Highly Diluted in Ar
ki= A exp(-T,/q
cm3 mol-' s-l rate coefficient
reaction 1 2 3 4 5 6 7
+ +
--
+ +
H2S Ar H2 S Ar HIS S products HzS+H-SH+Hz H2+S4SH+H S+SH-Sz+H &+Ar-S+S+Ar H2+Ar-H+H+Ar
A
ref
TA
1.9 x 1014 32860 7600 5.7 x 1014 2491 1.9 x 1014 6.0 x 1014 12070 2.0 x 1013 0 4 . 8 ~ 1 0 ' ~38750 2 . 2 ~1014 48350
this study this study 4 13
est 16 14
follows
The k~ values obtained from the steady-state analysis are given in the Arrhenius diagram of Figure 6 by the closed symbols. The LFP-shock wave experiments in CS2/H2S/Ar reaction systems were designed to study reaction R2 under almost firstorder conditions, and the time constant z was determined from the S absorption decay. For further data interpretation the knowledge of the behavior of H2S during the 193 nm photolysis is required. A photodissociation of H2S forming S atoms was not observed under our experimental conditions. The formation of H atoms from the 193 nm photolysis of HzS was determined by Yoshimura et al! to be 4% of the initial H2S for laser energies approximately 2 times higher than in the present experiments. Therefore, it can be concluded that in the present experiments the H2S initial concentration is almost not affected by the laser photolysis. The influence of the background
Woiki and Roth
12962 J. Phys. Chem., Vol. 98,No. 49,1994
1 oi4
I
'
I
'
I
'
I
H,S + S + products
3
--
T=2239K,p=1.38bar 25 ppm H,S in Ar
c?
5 4.0 '0
...**'
.
I
r
c 3.0-
.-0
8 ' # ,'
......."
uk,=0.52
...' ...*'
..."
Pyrolysis Experiments 100 ppm H,S in Ar
0
-
Photolysis Experiments 0
wr"+wppmH$
A
30ppmC$+!soppmk@
-
13 30wmCS,+l00wmH$
10"
I
5
4
7
6
8
0
i o 4WT Figure 6. Arrhenius plot for the rate coefficient kz determined by
pyrolysis and photolysis experiments. reaction
S
+ CS2-products
on the observed t values was also considered, resulting in an expression for the time constant of (5) The corresponding rate coefficient ks2 was measured in ref 11. The k2 values determined from the measured time constants t by eq 5 are summarized in the Arrhenius diagram of Figure 6, see open symbols. The data points scatter around a straight line which can be represented by the Arrhenius expression
(6)
The higher temperature side this line fits well to the k2 values obtained from the steady-state analysis of the S measurements during the pyrolysis of H2S. The error limit of k2 is suggested to be a factor of 2 including the error of the time constant z and the combined errors of post-shock temperature and concentration. The obvious deviation of the S absorption decay from the fust-order behavior at times t 1 300 pus must be the result of further secondary reactions like, e.g., the HS2 decomposition or bimolecular reactions of H and SH with HS2 radicals. These reactions are sensitive to the S concentrations at longer reaction times but have practically no effects on the observed early S atom profiles. The remaining open question is the formation of H atoms during the H2S pyrolysis. This problem can be solved by considering possible product channels of reaction R2: H2S
-
+ S h a HS2 + H
- + -S2+ k2b
kzc
SH
SH
H2
400
AH298
= -38.4 kJ/mol
(R2a)
AH298
= +22.2 kJ/mol
(R2b)
AH298= -128.1 kJ/mol (R2c)
The different reaction channels of reaction R2 were included into the reaction mechanism of Table 3. The branching ratio of the H atom forming reaction R2a was determined in computer simulations by fitting kza values to the experimentally observed H profiles. In this fitting procedure only the rate coefficient
600
800
Time / p Figure 7. Measured and calculated concentration profies of H atoms with variations of the branching ratio kdk2.
+
kcs2
k2 = 5.7 x 1014exp(-7600 WT) cm3 mol-' s-l
200
k%was varied and the sum of kb k2b 4-k2c was always chosen to be equal to the measured overall rate coefficient k2 of eq 6. The sensitivity of the calculated H concentrations to variations of kza is shown for one experiment in Figure 7. The best fit to the measured H profile (points) was obtained with a branching ratio of k2$k2 = 0.52. The dashed and dotted lines show the calculated result for different k2& ratios of 0.8 and 0.2, respectively. In this way all H profiles obtained during the pyrolysis of H2S could be well fitted using a branching ratio of k24k2 between 0.35 and 0.57. The calculated H profiles were found to be almost insensitive to variations of the branching ratios k2dk2 or k2Jk2. Even the elimination of one reaction channel (R2b) or (R2c) led only to small deviations between calculated and measured H profiles. The consumption of H atoms is mainly given through reaction R3. The corresponding rate coefficient was directly measured in ref 4. The k3 value reported by Roth et aL3 was not regarded in the present reaction scheme because it was estimated under the assumption of a direct H formation via reaction R1*. Reaction R4 was investigated in our laboratory by pyrolysis and photolysis experiments very similar to the present ones.I3 Reaction R5 contributes significantly to the H formation at reaction times t L 400 p s . The rate coefficient of reaction R5 was estimated with regard to the corresponding reactions of 0 atoms and OH radicals.14 Variations of the estimated k5 value by multiplying or dividing it by a factor of 2 lead to a small deterioration of the kinetic model for longer reaction times. It should be pointed out that possible secondary reactions of HS2, which were temporarily added to the reaction scheme considering the analogy to the Hd02 reaction system, have sigmficant influences on the H profiles and lead to stronger alterations of the branching ratio of reaction R2. The proposed mechanism of Table 3 should therefore not be understood as the "last word" in H2S decomposition kinetics but as a scheme, which satisfactorily describes all details of our experimental findings. It demonstrates that the formation of atoms via the secondary reaction R2a is fast enough to provide the almost linear increases observed in the present study and also by Roth et aL3 The short induction period of the H formation which is seen in the calculated profiles of Figure 7 cannot be measured in the experiments because of too small absorption signals at the very early reaction time. The present result of the HIS decomposition was confirmed by a recent shock tube study of Olschewski et al.15 The authors monitored the H2S decay by UV absorption at 1 = 215 nm and
Kinetics of the High-Temperature HzS Decomposition
J. Phys. Chem., Vol. 98, No. 49, 1994 12963
found a rate coefficient for the disappearance of HzS to be
d[H,S]ldt 2WZSl [AI
= 2.0 x 1014exp(-33000 WT) cm3 mol-' s-' (7)
which is practically identical to our kl value. This excellent agreement between these two independent experiments supports the present findings that the S atom elimination dominates the H2S decomposition. An analysis of the rate coefficient kl in terms of unimolecular rate theory is also given in ref 15 and identifies the activation energy and the absolute value of kl to be consistent with theory. The comparison of the present results with the earlier HzS decomposition studies in terms of deduced rate coefficients appears to be not useful because of the different initial reactions R1 or R1*, respectively. Therefore the experimental findings of the other studies were compared with computer simulation by applying the present reaction mechanism to the different experimental conditions. The H2S decay monitored in the experiments of Bowman and Dodge' could not be fitted by our reaction scheme. The calculation leads to 'HzS decay rates approximately 5 times faster than observed in ref 1. This disagreement must be caused by further secondary reactions of radicals like, e.g., HS2 or H ~ S Zwhich , are not important for our very low concentration conditions but which must be considered in the experiments of ref 1. However, in the absence of kinetic data for these species, we did not try to improve our reaction scheme to fit the results of ref 1. The SZ formation rates calculated from our mechanism are close to the observations of Higashihara et In the present simplified reaction scheme S2 is formed mainly via reactions R2c and R5. Further possible SZ formation reactions like, e.g., the recombination of SH SH S2 H2, were not considered. The results of Yoshimura et al.? concerning reaction R3, are not affected by our results because the pseudo-first-order study of the reaction H H2S SH HZof ref 4 appears to be independent of the correct knowledge of the H2S pyrolysis.
+
+
+
+
Acknowledgment. The authors thank Professor A. Burcat, Faculty of Aerospace Engineering, Technion, Haifa, Israel, for
his help in estimating thermodynamic data for the HS2 and H2S2 radicals and Professor J. Troe, Universitiit Gottingen, for helpful discussions. We further thank Mrs. C. Krniecik, Mrs. N. Schlosser, and Mr. L. Jerig for their help in conducting the experiments. The financial support of the Deutsche Forschungsgemeinschaft is gratefully acknowledged. References and Notes (1) Bowman, C. T.; Dodge, L. G. Kinetics of the Thermal Decomposition of Hydrogen Sulfide Behind Shock Waves. Symp. (Int.) Combust. [Proc.] 1976, 16, 971-982. (2) Higashihara, T.; Saito, K.; Yamamura, H. SZ Formation During the Pyrolysis of H2S in Shock Waves. Bull. Chem. SOC.Jpn. 1976,49 (4), 965-968. (3) Roth, P.; h h r , R.; Barner, U.Thermal Decomposition of Hydrogen Sulfide at Low Concentrations. Combust. Flame 1982,45, 273-285. (4) Yoshimura, M.; Koshi, M.; Matsui, H. Non-Arrhenius Temperature Dependence of the Rate Constant for the Reaction of H+H& Chem. Phys. Len. 1992, 189, 199-204. (5) Markus, M. W.; Woiki, D.; Roth, P. Two-Channel Thermal Decomposition of CH3. Symp. (Int.) Combust. [Proc.] 1992, 24, 581588. (6) Woiki, D.; Markus, M. W.; Roth, P. A Shock Tube-Laser Flash Photolysis Study of the Reaction COS S CO S2. J . Phys. Chem. 1993, 97 (38), 9682-9685. (7) Thielen, K.;Roth, P. Resonance Absorption Measurements of N and 0 Atoms in High Temperature NO Dissociation and Formation Kinetics. Symp. (In?.) Combust. [Proc.] 1984, 20, 685-693. (8) Woiki, D.; Roth, P. A Shock Tube Study on the Thermal Decomposition of CS2 Based on S(3P) and S(lD) Concentration Measurements. Shock Wave J., in press. (9) Woiki, D.; Roth, P. Shock Tube Measurements on the Thermal Decomposition of COS. Ber. Bunsen-Ges. Phys. Chem. 1992, 96 (lo), 1347-1352. (10) Thielen, K.;Roth, P. Stosswellenuntersuchungenzum Start der Reaktion CO 0 2 . Ber. Bunsen-Ges. Phys. Chem. 1983, 87, 920-925. (11) Woiki, D.; Roth, P. Oxidation of S and SO by 02 in HighTemperaturePyrolysis and Photolysis Reaction Systems. Submitted to Int. J . Chem. Kinet. (12) Lutz, A. E.;Kee, R. J.; Miller, J. A. SENKIN: A FORTRAN Program for Predicting Homogeneous Gas Phase Chemical Kinetics with Sensitivity Analysis. SANDIA REPORT, SAND87-8248.UC-4, 1988. (13) Woiki, D.; Roth, P. On the Reaction S Hz SH H Studied in Pyrolysis and Photolysis Systems Behind Reflected Shock Waves. To be published. (14) Warnatz, J. Rate Coeficients in the C/H/O System; Gardiner, W. C., Jr., Ed.; Springer-Verlag: New York, 1984. (15) Olschewski, H. A.; Troe, J.; Wagner, H. Gg. UV Absorption Study of the Thermal Decomposition Reaction H2S H2 S(3P). J . Phys. Chem., following paper in this issue. (16) Higashihara, T.; Saito, K.; Murakami, I. The Dissociation Rate of SZ Produced from COS Pyrolysis. Bull. Chem. SOC.Jpn. 1980, 53, 1518.
+
+
f.
+
+
-.
+
-
+