Multiple dynamical pathways in the O+SiH4 reaction studied by the crossed molecular beam method

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

155.33.16.124 On: Mon, 01 Dec 2014 02:40:23

JOURNAL OF CHEMICAL PHYSICS

VOLUME 113, NUMBER 5

1 AUGUST 2000

Multiple dynamical pathways in the O¿SiH reaction studied

4

by the crossed molecular beam method

Jim J. Lin

Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei, Taiwan, Republic of China

Yuan T. Lee

Institute of Atomic and Molecular Sciences, Academia Sinica and Department of Chemistry,

National Taiwan University, Taipei, Taiwan, Republic of China

Xueming Yang^a⁾

Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei, Taiwan,

and Department of Chemistry, National Tsing-Hua University, Hsinchu, Taiwan, Republic of China

͑

Received 28 January 2000; accepted 3 May 2000͒

In this report, the OϩSiH reaction is investigated using crossed molecular beam techniques.

4

Multiple pathways have been observed for this reaction. Angular resolved time-of-ﬂight spectra

have been measured for all reaction products in a single set of experiments. Different product

angular distributions were measured for various product channels, indicating that multiple reaction

pathways with distinctive dynamics are present. An intriguing triple product channel, SiOϩ2H ,

2

3

1

has also been clearly identiﬁed. However, the role of O( P) vs O( D) in the reactions remains to

be clariﬁed in further experimental investigations. This study provides a good example of multiple

dynamical pathways in a complicated reaction, which opens enormous opportunities in investigating

the dynamics of chemical reactions that are important in combustion and atmospheric chemistry.

©

2000 American Institute of Physics. ͓S0021-9606͑00͒01129-6͔

I. INTRODUCTION

new apparatus was set up in our laboratory with a much

improved detection efﬁciency in order to investigate more

complicated chemical reactions. A custom designed, larger

sized quadrupole mass ﬁlter was installed in an effort to in-

Modern experimental studies of reaction dynamics have

mainly focused on elementary chemical reactions during the

last few decades. Chemical dynamics studies of complicated

reactions of energetic species with multiple reaction channels

have received little attention, even though a signiﬁcant frac-

tion of chemical reactions in atmospheric and combustion

processes involve multiple reaction pathways. This is largely

due to the lack of sensitive and low background universal

detection techniques which can detect all chemical reaction

products. Reliable full-ﬂedged quantum mechanical dynami-

cal calculations for complicated reactions are still lacking

even though overall theoretical studies on chemical reactions

are becoming more and more reliable. Therefore, experimen-

tal investigation still remains to be the most viable way to

learn the detailed dynamics of complicated reactions with

multiple reaction pathways.

The universal crossed molecular beam method based on

electron impact ionization of chemical products has been es-

sential in investigating the dynamics of bimolecular chemical

reactions during the last few decades. Product angular and

translational energy distributions can be measured using this

powerful method. Electron impact ionization with quadru-

pole mass selection is, in principle, an ideal universal detec-

tion method for all reaction products. The detection efﬁ-

ciency of this method is, however, limited by the low

ionization efﬁciency, the limited quadrupole transmission ef-

ﬁciency, and the high background in the detector. Recently, a

crease the transmission efﬁciency for product ions. A much

Ϫ12

lower background in the detector (Ͻ10

Torr) was also

achieved using multiple high vacuum pumps and cleaner

1

vacuum conditions. With these efforts, the detection efﬁ-

ciency is much improved for all reaction products. This

much improved instrument opens an exciting new avenue for

investigating the dynamics of more complex reactions with

multiple pathways under single collision conditions. In this

3

1

report, the O( P, D)ϩSiH reaction using this new appara-

4

tus will be presented, providing a good example of extremely

complicated reactions with a clear dynamical picture.

3

1

The reactivity of O( P) and O( D) is very important

because of their roles in combustion processes and chemical

vapor deposition ͑CVD͒ processes. One of the methods to

produce the SiO2 layer over the Si is CVD using oxygen gas

and silane gas.²A number of relevant chemical processes

have been investigated theoretically.⁷However, very little

is known about the detailed mechanism of this oxidation

process even in the gas phase. In addition, this reaction is

–6

–17

very important in the combustion and the explosion of silane.

3

The O( P) reaction with SiH has been quite exten-

4

sively investigated in the gas phase. A few measurements

have been carried out on the total reaction rate at room tem-

perature for this reaction. Atkinson and Pitts¹⁸used ﬂash-

photolysis with O/NO chemiluminescence detection of the O

Ϫ13

Ϫ3 Ϫ1

a͒_A_u_t_h_o_r_t_o_w_h_o_m_c_o_r_r_e_s_p_o_n_d_e_n_c_e_s_h_o_u_l_d_b_e_a_d_d_r_e_s_s_e_d_;_e_l_e_c_t_r_o_n_i_c_m_a_i_l_:

xmyang@po.iams.sinica.edu.tw

atoms and obtained kϭ4.8ϫ10

cm

s

.

Mkryan

cm s

et al.¹⁹measured the rate constant to be 2.8ϫ10

Ϫ13

3 Ϫ1

0

021-9606/2000/113(5)/1831/12/$17.00

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1831

1

1832

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Lin, Lee, and Yang

using a discharge-ﬂow method with ESP detection of O at-

oms. A recent measurement by Ding and Marshall²⁰deter-

mined the rate constant at the room temperature to be 3.5

Ϫ13

3 Ϫ1

ϫ10

consistent with each other. At a temperature of 1000 K, the

cm s . These measured rate constants are quite

total reaction rate was measured to be about

4

Ϫ11

3

Ϫ1 21

ϫ10

cm s

.

Clearly, the reaction rate increases as the

temperature rises. A critical review is also provided for this

reaction in that paper. At even higher temperature, no experi-

mental data are available yet for this reaction. The dynamics

of the OHϩSiH channel from this reaction has been inves-

3

tigated using the laser induced ﬂuorescence ͑LIF͒ technique

2

by Park et al. This work reveals that the OH vibrational

population at vϭ0 and 1 is strongly inverted, indicating that

the abstraction mechanism is likely to play a signiﬁcant role

in the reaction. The rotational distributions for the vϭ0 and

1

states are also quite different from each other. The only

3

rate constant measurement for the OH channel for the O( P)

reaction with silane was done by Setser et al.²³The rate con-

Ϫ12

3 Ϫ1

stant k_O_Hwas determined to be 1.2Ϯ0.4ϫ10

cm s at

room temperature, which is actually a few times larger than

that of the total rate constant measured at the room tempera-

3

1

FIG. 1. Energy diagram for the possible channels of the O( P, D)ϩSiH

4

ture. Obviously, there is no consensus yet on k from pre-

reaction, based on previous ab initio calculations from Refs. 35, 36, and 37.

OH

2

2–24

vious studies. From the previous studies,

the main reac-

3

tion channel for the O( P) reaction with silane is determined

to be the OH formation process likely through an abstraction

channel and the insertion mechanism through the triplet sur-

face is likely spin-forbidden. Therefore reaction channels oc-

cur through the insertion mechanism are likely not important

the possible pathways and the energy diagram. There are a

total of ten different product channels that are energetically

1

accessible for the O( D)ϩSiH reaction, and eight different

4

3

product pathways for the O( P)ϩSiH reaction. For

O( D)ϩSiH , the following channels are accessible:

4

1

3

in the O( P) reaction with silane.

4

1

The O( D) reaction with SiH has also been investi-

4

O͑ D͒ϩSiH →SiH ϩOH

͑I͒

͑IIa/IIb͒

͑IIIa/IIIb͒

͑IV͒

4

3

2

5–29

gated extensively.

The total reaction rate constant for

Ϫ10

3

Ϫ1

→H SiOϩH/H SiOHϩH

this reaction was determined to be 3.0Ϯ0.3ϫ10

cm s

3

2

at room temperature by Okuda et al.²⁹From the same study,

the relative yield of the OH channel was measured to be 36%

with an inverted vibrational state distributions for the OH v

ϭ0 and 1 products. An interesting triple product channel

→

H SiOϩH /HSiOHϩH

2 2 2

→SiOϩ2H2

→

HSiOϩH ϩH/SiOHϩH ϩH ͑Va/Vb͒

2 2

SiOϩ2H has also been observed by direct laser induced

2

ﬂuorescence of SiO, with the relative yield determined to be

about 6%–13%. The H atom product channel has also been

measured, with the yield measured to be 21%. The infrared

diode laser spectroscopy study³⁰of the same reaction mea-

→H SiOϩ2H/HSiOHϩ2H.

͑VIa/VIb͒

2

3

For the O( P)ϩSiH reaction, all channels listed above, ex-

4

cept the H SiOϩ2H/HSiOHϩ2H channels, are energetically

2

accessible.

sured the OH production rate k_O_Hto be

5

Ϫ11

3

Ϫ1 Ϫ1

ϫ10

cm molecule

s

, which is about half the value

2

9

31

II. EXPERIMENT

obtained by Okuda et al. Many theoretical investigations

suggest that the O( D)ϩSiH reaction goes via an insertion

1

The apparatus used in this experiment is a recently built

crossed molecular beam machine which has been described

4

3

2–34

mechanism similar to the analogous CH reaction.

Even

4

1

though much experimental data are available, a clear and

complete dynamical picture has still not been obtained. The

in detail elsewhere. Brieﬂy, the O atomic beam, generated

using the photodissociation of O at 157.6 nm in a skimmed

2

3

1

relative importance of the O( P) and O( D) reactions with

O pulsed beam, was crossed with a skimmed SiH beam at

2

4

SiH is also not exactly clear at the moment. Since the reac-

a ﬁxed angle of 90 deg. The O beam was obtained by ex-

4

2

1

tion rate of O( D)ϩSiH is signiﬁcantly larger than that of

panding a neat O ͑99.99%͒ sample through a commercial

4

2

3

O( P)ϩSiH at room temperature, it is conceivable that

even at high collision energies, the O( D) reaction with SiH

might still be more important than the O( P) reaction with

pulsed valve ͑General Valve͒ with a rise time of about 50 ␮s,

at a stagnation pressure of about 5 atm. The expanded beam

was then skimmed by a sharp edged skimmer ͑Beam Dy-

4

1

4

3

silane.

namics͒ with 2 mm diameter oriﬁce. The O beam was then

2

The heat formations for all the reactants and products

intercepted by an unpolarized 157.6 nm laser beam, gener-

1

3

involved in the pathways for both the O( D) and O( P)

ated by a Lambda Physik LPX210I F laser, with a pulse

2

reactions have already been determined.³

5–37

Figure 1 shows

energy of about 30 mJ at a repetition rate of 50 Hz. The O

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

155.33.16.124 On: Mon, 01 Dec 2014 02:40:23

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Dynamical pathways in the OϩSiH reaction

1833

4

atomic beam was then skimmed once before entering the

and error. Starting from the initial CM energy distributions

and the initial CM angular distribution, the laboratory TOF

spectra and the angular distribution for this channel can be

simulated using the transformation from the CM frame to the

LAB frame. By adjusting these CM translational energy dis-

tributions and the CM angular distribution, satisfactory ﬁts to

the measured TOF spectra and angular distribution are ob-

tained. In this way, the CM translational energy distributions

and the CM angular distribution can be obtained. Simulation

of multiple channel processes can be done in a similar way

by simply adding a few different channels together. Both the

velocity spread (⌬v) and the angular divergence of the two

molecular beams have been incorporated in the simulation.

Therefore the simulated TOF spectra reported in this work

should have included the effects of the beam conditions,

which were measured in situ experimentally.

main chamber. In order to reduce the O background, both in

2

the main chamber and in the detector, the O beam was

2

rotated to an angle between 60° and 70° from the detector

rotating plan The F laser beam was focused on a spot of 4

2

mm͑w͒ϫ2 mm͑h͒ in the interaction region by a spherical-

cylindrical MgF shaping lens. Using the above focusing

2

condition and laser power, the O transition ͑cross-section

2

Ϫ18

2 38

␴

ϭ6.8ϫ10

cm ͒ at 157.6 nm can be easily saturated.

The O molecule breaks into one O( P) atom and one

O( D) atom at 157.6 nm photolysis, therefore the O atom

beam contains both O( P) and O( D) with a ratio of

3

2

1

3

1

3

9

5

0:50. Because both types of O atoms react with silane, it

will not be easy to differentiate the products from the two

3

reactions. However, since the energetics of the O( P) and

1

O( D) reactions are signiﬁcantly different, it is possible to

make some conclusions in certain cases. The SiH molecular

4

III. RESULTS AND DISCUSSIONS

beam was generated by expanding a neat SiH ͑99.99%͒

4

sample at a stagnation pressure of 5 atm through a carefully

adjusted pulsed valve ͑General Valve͒ with a rise time

ϩ

In

this

work,

signals

at

m/eϭ17 (OH ),

ϩ

4

7(H SiO /H SiOH ),

46(H SiO /HSiOH ),

3

2

͑10%–90%͒ of about 60 ␮s, and then skimmed once by a 1.5

ϩ

5(HSiO /SiOH ), and 44(SiO ) were detected from the

OϩSiH reaction. TOF spectra at different laboratory angles

and the total product angular distributions were measured for

all above products. From detailed analysis of these results, a

number of reaction channels have been identiﬁed. While it is

clear that the m/eϭ17 (OH ) product is coming from

the OHϩSiH₃channel ͑channel I͒ and the m/e

ϭ47 (H SiO /H SiOH ) products are from reaction chan-

nels IIa and IIb, signals at other masses ͑mass 46, mass 45,

and mass 44͒ products obviously contain original reaction

products at these masses and various contributions cracking

from higher mass products. Therefore analyses of these ob-

served TOF spectra become quite complicated. Systematic

analyses on these data are carried out, a reasonable picture is

obtained from these analyses. In the following paragraphs,

detailed analyses and results are described.

mm oriﬁce skimmer before entering the main chamber. The

4

3

1

O( P, D) beam, the SiH molecular beam, and the detection

4

3

1

axis are all in the same plane. The speed of the O( P, D)

beam was 2367 m/s with less than Ϯ3% velocity spread. The

angular divergence of the O( P, D) beam was about Ϯ2.5°.

3

1

ϩ

The speed of the SiH beam was about 800 m/s with a speed

4

ratio of about 10 and an angular divergence of about Ϯ2°.

The collisional energy at which this work was carried out is

about 8.0 kcal/mol.

ϩ

3

2

The whole experiment was pulsed, and time zero was

deﬁned as the time when the two beams were crossed. After

ﬂying about 25 cm from the crossed region, the neutral re-

action products were then ionized by a Brink’s type electron

impact ionizer with an electron energy of about 60 eV. The

product ions were mass ﬁltered by a quadrupole mass ﬁlter,

and counted by a Daly ion detector. All time-of-ﬂight ͑TOF͒

spectra were taken at 1 ␮s per channel during the experi-

ment. The TOF spectra shown in this work were all re-

binned for a better S/N ratio without changing the shapes of

the spectra. The total product angular distributions were

measured by rotating the detector. During the experiments

A. The OH¿SiH channel

3

OH products were detected at m/eϭ17, indicating

clearly the existence of the OHϩSiH channel. OH products

are normally difﬁcult to detect using conventional universal

crossed beam apparatus because of the high background at

mass 17 in the detector. This is especially true for reactions

3

described above, the vacuum in the detector ionization re-

Ϫ12

gion was maintained at about 1ϫ10

Torr.

in which fast products are scattered into a large solid angle,

such as O( D)ϩSiH in which the OH products are scat-

1

The time-of-ﬂight spectra and angular distributions of

the neutral products measured in the laboratory ͑LAB͒ frame

were computer-simulated in order to obtain the translational

energy distributions and angular distributions in the center-

of-mass ͑CM͒ frame. In the simulation for a single reaction

channel, normally a few initial CM translational energy dis-

tributions at several CM angles and an initial CM angular

distribution were used as the starting point. If the product

translational energy distribution for the channel is angular

dependent, several CM translational energy distributions are

necessary to describe the angular dependent energy distribu-

tions using linear interpolation. If the translational energy

distribution is not angular dependent, however, a single dis-

tribution is normally used. The method to ﬁnd out how many

CM kinetic energy distributions are required is usually trial

4

tered into the whole solid angle 4␲ ͑see Fig. 2 for the New-

ton diagram for this reaction͒. TOF spectra for the OH prod-

uct were measured at 11 different angles. Figure 3 show six

of these TOF spectra with TOF signals at different angles

normalized. The signals observed are all due to reactive scat-

tering since no other sources can contribute signiﬁcantly to

the mass 17 signal. The total product angular distribution for

the OH product is shown in Fig. 4.

Reasonable ﬁts to the TOF spectra of the OH product at

different lab angles have been obtained using three different

energy distributions at CM angles 0°, 25°, and 180° ͑see

upper panel in Fig. 5͒. Kinetic energy distributions at other

CM angles are obtained by linear interpolation using the

above three distributions. Clearly, the translational energy

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

155.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1834

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Lin, Lee, and Yang

FIG. 4. The total product angular distribution of the OH radical product.

The solid squares are the experimental data, while the solid line is the

simulation.

FIG. 2. Newton diagram for the OHϩSiH₃channel. The circle represents

the OH product scattering Newton circle with 10 kcal/mol kinetic energy

release.

product CM angular distribution P(␪) used in the ﬁtting is

shown in the lower panel of Fig. 5. The CM velocity ﬂux

contour map ͑Fig. 6͒ of the OH product was also obtained

from ﬁtting the angular resolved TOF spectra ͑Fig. 3͒. From

the results ͑Figs. 4, 5, and 6͒ described above, it is clear that

distribution is coupled to the angular distribution, indicating

that reaction products scattering into different CM angles

obviously have different translational energy distributions,

and thus different internal energy distributions. The OH

FIG. 3. The TOF ﬂight spectra of mass 17 ͑OH͒ at six different laboratory

angles. The empty circles are the experimental data, while the solid lines are

the simulated results. These TOF spectra were taken by averaging signals

over 200 000 laser shots.

FIG. 5. The CM kinetic energy distributions of the OH product channels

and the CM product angular distribution used in simulating the experimental

TOF spectra and the laboratory angular distribution.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Dynamical pathways in the OϩSiH reaction

1835

4

FIG. 6. The product ﬂux contour diagram of the OH product channel.

the OH products are scattered predominantly to the forward

direction with respect to the O atom beam. This interesting

observation provides important insight for the OH formation

channel.

FIG. 7. Newton diagram for the binary H and H formation channels. The

circles represents the Newton circles for heavy radicals from H and H2

formation channels with different kinetic energy releases.

2

Since the long-lived complex mechanism normally pro-

duces forward and backward symmetric or isotropic angular

distributions, it is quite obvious that the forward scattered

ments of the OH radical products²²^,²⁹from the O( P, D)

reactions show inverted vibrational distributions, indicating

that the OH radicals are internally hot. This is also consistent

with the conclusion that main forwarded OH products ob-

3

1

3

1

OH products from the O( P, D) reaction with SiH are

4

formed with a totally different mechanism. These forward

OH products are most likely produced from an abstraction

type mechanism, in which the O atom comes in at a relative

larger impact parameter and strips a hydrogen atom from the

silane molecule directly; the intermediate reaction to produce

the forward scattered OH products should be very shorted

lived. In addition to the strongly forward scattered OH prod-

ucts, a signiﬁcant portion of the OH products is backward

scattered ͑see Fig. 5͒. These products are probably due to a

long-lived complex mechanism. If the backward scattered

products ͑90°–180° in Fig. 5͒ are all due to this mechanism

1

served are coming from the O( D) reaction. For the back-

ward scattered OH products (⌰_C_Mϭ180°), averaged trans-

lational energy release is about 10.7 kcal/mol, which is

signiﬁcantly smaller than that for the forward scattered prod-

ucts. This is probably due to the long-lived intermediate for-

mation mechanism in which energy is more randomized in

the internal degrees of freedom of the products. It is possible

that the O( P) reaction contributes more to this mechanism.

The overall O( P) contribution should be minor in the OH

production. In any case, the results presented above clearly

show that an abstraction type reaction mechanism is largely

responsible for the OHϩSiH pathway which is mainly due

to the O( D) reaction, while a more statistical long-lived

intermediate mechanism is also signiﬁcant. From the above

investigations, it is also quite clear that the O( P) reaction is

signiﬁcantly less reactive than the O( D) reaction at 8 kcal/

mol collisional energy. This is, nevertheless, not so surpris-

ing since O( P) is less reactive than O( D) at room tem-

perature by almost two orders of magnitude. Even though the

O( P) reaction cross section increases as temperature rises,

3

͑the same amount of forward scattered products should also

be included͒, the overall contribution to the OH channel

from this mechanism should be about 36%. This is a rather

signiﬁcant contribution, implying that even though the ab-

straction mechanism is the most important for the OH for-

mation, a mechanism involving a long-lived complex which

is likely formed through the O atom insertion could also be

signiﬁcant.

3

1

3

1

3

1

The averaged product translational energy obtained from

this experiment for this channel is about 16 kcal/mol for the

^f^a^s^t^f^o^r^w^a^r^d^s^c^a^t^t^e^r^e^d^p^r^o^d^u^c^t^s⁽^⌰CMϭ0). The available en-

ergy for the OH formation is about 64 kcal/mol for the

3

1

it might still be less important than the O( D) reaction at the

collisional energy of this experiment.

1

3

O( D) reaction, and 18 kcal/mol for the O( P) reaction.

Since it is not likely that all the available energy can be

deposited into the product translational degree of freedom,

especially for such a complicated reaction, the major forward

scattered OH products observed in this study should mostly

B. The H¿H SiOÕH SiOH channel

3

2

The H SiO/H SiOH product was detected at m/e

ϭ47 (H SiO /H SiOH ), which is clearly from the

HϩH SiO/H SiOH channel since no product at higher

3

2

ϩ

3

2

1

come from the O( D) reaction. The high energy limit ͑all

3 2

larger than 40 kcal/mol͒ in the translational energy distribu-

tions also supports this conclusion. Previous LIF measure-

masses from this reaction is present. The Newton diagram

for this channel is shown in Fig. 7. The CM direction for this

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

155.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1836

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Lin, Lee, and Yang

FIG. 8. The TOF ﬂight spectra of mass

ϩ

4

7 (H SiO /H SiOH ) at four different laboratory

3 2

angles. The empty circles are the experimental data,

while the solid lines are the simulated results.

reaction is at about 34° laboratory angle. The experimental

TOF spectra at four different laboratory angles are shown in

Fig. 8. The total product angular distribution of the mass 47

product in the laboratory frame is shown in Fig. 9. The TOF

spectra and the product angular distributions were simulated

using two different CM kinetic energy distributions at 0° and

From the CM angular distribution, it is obvious that the

mass 47 product is essentially backward scattered relative to

the O atom beam. This is quite interesting since the H atom

formation are believed to come from an insertion mechanism

from previous studies, which is unlikely to produce clearly

backward scattered radical products. This indicates that the

backward scattered products are likely from reaction mecha-

nisms other than insertion type. The experimental results

180°, shown in the upper panel of Fig. 10, and the CM prod-

uct angular distribution shown in the lower panel of Fig. 10.

CM kinetic energy distribution at other angles are obtained

using linear interpolation from the distributions at 0° and

1

9

80°. The simulated results are also shown in Fig. 8 and Fig.

in comparison with the experimental results, which are in

rather good agreement. It is interesting to point out that the

product kinetic energy distributions at 0° and 180° degrees in

the CM frame are apparently different from each other. At

CM 180° degree ͑backward direction relative the O atom͒,

the mass 47 product (H SiO/H SiOH) kinetic energy distri-

3

2

bution is shifted toward higher energy in comparison with

the 0° distribution. The product ﬂux contour map for this

reaction channel is constructed from the CM kinetic and an-

gular distributions, and is presented in Fig. 11.

FIG. 9. The total product angular distribution of the H SiO/H SiOH prod-

FIG. 10. The CM kinetic energy distributions of the HϩH SiO/H SiOH

3

2

3

2

uct. The solid squares are the experimental data, while the solid line is the

simulation.

product channels and the CM product angular distribution used in simulat-

ing the experimental TOF spectra and the laboratory angular distribution.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Dynamical pathways in the OϩSiH reaction

1837

4

FIG. 11. The product ﬂux contour diagram for the HϩH SiO/H SiOH prod-

3

2

uct channel.

show that after the O atom attacks the SiH molecule, some

4

of the H SiO/H SiOH radical products bounces back while

3

2

an H atom pops forward along the O atom beam. This back-

ward scattered angular distribution suggests that this reaction

pathway likely occurs at small impact parameters. Since

SiH is a tetrahedral molecule, the backward scattered radi-

4

FIG. 12. The neutral TOF spectra at m/eϭ47, 46, 45, 44 at lab angle ⌰_L

ϭ35° which is close to the LAB angle of the center-of-mass velocity vector.

Open circles: experimental data. Solid lines: ﬁtting curves of various chan-

nels. Please see text for different assignments. All TOF signals have been

normalized to the same number of laser shots. These spectra were typically

taken by averaging signals over 100 000 laser shots.

cals could come from a S 2 type mechanism in which the

N

1

O( D) atom approaches directly at the Si atom along the

departing H atom direction.

The product translational energy distribution peaks at

about 11 kcal/mol with the maximum translational energy

cut off at about 43 kcal/mol. Since the H atom product does

not have any internal energy, the product internal energy for

this channel should all be carried by the H SiO/H SiOH radi-

C. The 2H¿H SiOÕHSiOH channel

2

3

2

cals, which should have about 55 kcal/mol internal energy if

In comparison with the TOF spectrum at mass 47 at 35°

͑the laboratory angle nearly along the CM direction͒, TOF

spectra of ionized products at mass 46, 45, 44 at the same lab

angle are shown together in Fig. 12. The TOF signals at

these different masses have been normalized. From these

TOF spectra it is quite obvious that the shapes of the lower

mass products ͑m/eϭ46, 45, and 44͒ are signiﬁcantly differ-

ent from the TOF spectrum at mass 47, indicating that extra

neutral products at these lower masses are likely present in

addition to the contribution from the dissociative ionization

of mass 47 products. Furthermore, TOF spectra at 44, 45,

and 46 are also substantially different from each other, im-

plying that product channels to produce these lower mass

neutral products are present with contributions from cracking

of higher mass products. Since there are likely multiple con-

tributions to the signals at masses 44. 45, 46, careful analyses

have been carried out to show the least possible number of

channels that are responsible to these signals. The analyses

are done in the following way. For example, since the TOF

spectra at mass 46 are signiﬁcantly different from those at

mass 47, a single component from dissociative ionization of

mass 47 signals would not be sufﬁcient to properly simulate

the mass 46 TOF spectra. Therefore a new component is

1

this reaction is due to the O( D) reaction. Since the maxi-

3

mum available energy of the O( P) reaction with silane for

the H formation channel (H SiOHϩH) is about 49.8 kcal/

2

mol, it is certainly possible that some of these radical prod-

3

ucts observed are due to the O( P) reaction. There is, how-

ever, no documented observation of the H atom formation

3

from the O( P)ϩSiH reaction, while there are experimental

4

evidences that the H formation processes are present in the

1

O( D) reaction. If the observed radical products at mass 47

1

are coming from the O( D) reaction and the available en-

ergy is deposited into the vibrational degrees of freedom of

the H SiO/H SiOH radicals, most of these radical products

3

2

certainly will go through a secondary dissociation to form

H SiO/HSiOHϩH since the energetic H SiO/H SiOH radi-

2

3

2

cal products are clearly above the barriers to the triple prod-

uct channels ͑see Fig. 1͒. Therefore if the products observed

1

are from the O( D) reaction, the majority of the

H SiO/H SiOH products observed at m/eϭ47 are likely

3

2

electronically excited since these products clearly do not go

through a secondary dissociation. Theoretical investigations

of these radicals are desirable in order to better understand

the sources of these processes.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

155.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1838

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Lin, Lee, and Yang

FIG. 13. The TOF ﬂight spectra of mass

ϩ

4

6 (H SiO /HSiOH ) at four different laboratory

2

angles. The empty circles are the experimental data,

while the solid lines are the simulated results. Different

contributions are labeled as in Fig. 12.

added to simulate the mass 46 TOF spectra. For the mass 45

signals, contributions from dissociative ionization of mass 46

and mass 47 are considered at ﬁrst. Similarly, if the TOF

spectra at mass 45 cannot be simulated using the dissociative

ionization of mass 46 and mass 47 products, new contribu-

tions will be added to simulate the TOF spectra. TOF spectra

at mass 44 are also analyzed in a similar way. From these

analyses, a minimum number of reaction channels can be

extracted from the data. These analyses provide at least a

reasonable picture for the complicated dynamical processes

which are hard to study by any other experimental method.

During the analyses, dissociative ionization is assumed to be

independent of internal energy excitation of the radical prod-

ucts. This should be a rather good approximation since high

energy electrons ͑ϳ60 eV͒ are normally used to ionize the

reactive products with much less internal energy.

mass 47 products (H SiO/H SiOH). Therefore, the TOF

3

2

spectra at mass 46 are simulated using two different contri-

butions: the cracking contribution from the mass 47 product,

and a neutral product channel at mass 46 with mass ratio of

4

6:2. A single CM kinetic energy distribution ͑the upper

panel in Fig. 15͒ and a forward–backward symmetric prod-

uct angular distribution ͑the lower panel in Fig. 15͒ are used

to simulate the TOF spectra at different lab angles and the

product laboratory angular distribution. The simulated TOF

spectra and the laboratory angular distribution are shown in

Fig. 13 and Fig. 14 along with the experimental results. The

agreement between the experimental and simulated results is

quite satisfactory. From this simulation, it is quite clear that

the main product observed at mass 46 is the new channel of

mass 46 neutral product, while there is some contribution

from the cracking of the mass 47 product. The neutral mass

Four

typical

TOF

spectra

at

m/e

ϩ

4

6 signals are most likely due to the neutral H SiO/HSiOH

2

ϭ46 (H SiO /HSiOH ) are shown in Fig. 13. The total

2

product angular distribution of mass 46 is also presented in

Fig. 14. Since these TOF spectra are clearly different from

those of mass 47, it is obvious that the observed products at

mass 46 are not purely from the dissociative ionization of the

products. There are two possible sources for these radicals:

the H2ϩH2SiO/HSiOH channel and the 2HϩH2SiO/HSiOH

channel. The kinetic energy limit obtained from the above

simulation is about 38 kcal/mol ͑Fig. 15͒. Since the energetic

FIG. 14. The total product angular distribution of the

ϩ

mass 46 (H SiO /HSiOH ) product. The solid squares

2

are the experimental data, while the solid lines are the

simulations from two different contributions. Different

contributions are labeled as in Fig. 12.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Dynamical pathways in the OϩSiH reaction

1839

4

FIG. 17. The total product angular distribution of the mass

ϩ

4

5 (HSiO /SiOH ) product. The solid squares are the experimental data,

while the solid lines are the simulations from various contributions. Differ-

ent contributions are labeled as in Fig. 12.

1

that the products observed are consistent with the O( D)

3

reaction, not the O( P) reaction in which this triple product

channel is not accessible energetically. The fact that the an-

gular distribution for this channel obtained from the simula-

tion is essentially forward–backward symmetric indicates

that the products are likely produced through a long-lived

intermediate which is possibly formed through the insertion

of the O atom into the Si–H bond.

FIG. 15. The CM translational energy distribution ͑upper panel͒ and CM

angular distribution ͑lower panel͒ which are used to simulate the TOF spec-

tra and the laboratory angular distribution of the extra contribution to the

mass 46 signal using a 46:2 mass ratio. Some of the signals observed at

mass 46 are from dissociative ionization of the mass 47 product.

D. The H ¿H SiOÕHSiOH channel

2

Signals at mass 45 have also been observed. Four typical

ϩ

TOF spectra at m/eϭ45 (HSiO /SiOH ) at different labo-

ratory angles are shown in Fig. 16. The total product angular

distribution of mass 45 is also presented in Fig. 17. The

slower part of the signal at mass 45 can be simulated rather

well using cracking contributions from the H formation

͑mass 47͒ and the 2H formation ͑mass 46͒ channels with

slightly different ratio between the two channels from that at

mass 46. This is likely due to the different cracking patterns

of the original radicals. In addition, there is clearly some

1

limit of the 2HϩH SiO/HSiOH channel from the O( D) re-

2

action is about 40 kcal/mol, it is quite likely that the ob-

served neutral products at mass 46 are due to this channel.

The observed energy cutoff for this channel is much lower

than that of the H formation processes ͑ϳ142 kcal/mol for

H ϩH SiO/HSiOH͒, therefore it is not likely that the ob-

2

served mass 46 signals are due to the H formation pro-

cesses. From the energy diagram in Fig. 1, it is quite obvious

2

FIG. 16. The TOF ﬂight spectra of mass

ϩ

4

5 (HSiO /SiOH ) at six different laboratory angles.

The empty circles are the experimental data, while the

solid lines are the simulated from various contributions.

Different contributions are labeled as in Fig. 12.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

1840

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Lin, Lee, and Yang

implying that this process occurs also likely through a long-

lived complex which is possibly formed by the insertion

mechanism.

E. The 2H ¿SiO channel

2

ϩ

TOF spectra at mass 44 (SiO ) have also been mea-

sured. Figure 19 shows TOF spectra of mass 44 at four dif-

ferent laboratory angles. The measured product angular dis-

tribution of mass 44 is presented in Fig. 20. The main feature

of the TOF spectra of mass 44 can be simulated using the

contributions from dissociative ionization of higher mass

products. An extra fast feature is required to simulate the

TOF spectra of this mass in addition to the dissociative ion-

ization contributions of higher mass products. Since this fea-

ture only exists at mass 44, it is most likely due to the neutral

SiO products. There are two sources of generating SiO prod-

ucts: SiOϩH ϩ2H and SiOϩ2H . It is known that the

FIG. 18. The translational energy distribution ͑with isotropic angular distri-

bution͒ for the H ϩH SiO/HSiOH reaction channel used to simulate the fast

contribution to the TOF spectra and the laboratory angular distribution of

the mass 45 product using a 46:2 mass ratio.

2

SiOϩH ϩ2H channel is not accessible energetically in this

2

reaction. Therefore this new feature at mass 44 is likely from

the SiOϩ2H pathway. A reasonable ﬁt to the TOF spectra

2

and the laboratory angular distribution of mass 44 can be

extra fast feature in the TOF spectra of mass 45 in compari-

son with those of mass 46. This feature is assigned to the

H ϩH SiO/HSiOH process. The kinetic energy distribution

obtained by adding the SiOϩ2H channel using a mass ratio

2

of 44:4. The simulated results ͑see Figs. 19 and 20͒ including

other contributions are in good agreement with the experi-

mental data. Once again, an isotropic angular distribution in

2

for this feature is obtained from simulation, and is shown in

Fig. 18. The energetic limit ͑ϳ120 kcal͒ of the distribution

obtained is consistent with the available energy of the

the CM frame for the SiOϩ2H channel ﬁts best to the ex-

2

1

perimental data, implying this product channel also likely

goes through a long-lived reaction intermediate which is pos-

sibly formed through an insertion mechanism. The CM ki-

netic energy distribution used in the simulation is shown in

Fig. 21. The energy cutoff ͑ϳ125 kcal/mol͒ of this distribu-

tion is quite consistent with that of the SiOϩ2H2 channel,

H ϩH SiO/HSiOH process ͑ϳ142 kcal/mol͒ of the O( D)

2

reaction. This is rather signiﬁcant evidence that these fast

products at mass 45 are from the H SiO/HSiOH products

2

which subsequently undergo dissociative ionization to be-

ϩ

come mass 45 ions (SiHO /SiOH ). This observation is

1

also very similar to that in the O( D)ϩCH reaction, in

4

1

ϳ145 kcal/mol, from the O( D) reaction. This triple product

which nearly all H CO/HCOH products from the H forma-

2

ϩ

40

channel has also been observed previously using laser in-

duced ﬂuorescence technique.²⁹It has to be noted that be-

cause of the complicated nature of this reaction and the rela-

tive contributions of different channels are somewhat

correlated, the analyses provided above should be regarded

as qualitative only.

tion channel were observed at mass 29 (HCO ). The peak

ϳ30 kcal/mol͒ of the kinetic energy distribution is also con-

͑

sistent with a typical reverse barrier height ͑20–30 kcal/mol͒

of a more statistical H elimination process. An isotropic CM

2

angular distribution ﬁts well with the experimental results,

ϩ

FIG. 19. The TOF ﬂight spectra of mass 44 (SiO ) at

four different laboratory angles. The empty circles are

the experimental data, while the solid lines are the

simulated from various contributions. Different contri-

butions are labeled as in Fig. 12.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Dynamical pathways in the OϩSiH reaction

1841

4

study. Three binary product channels, SiH ϩOH,

3

H SiO/HSiOHϩH , H SiO/H SiOHϩH, have been ob-

2

3

2

served. Each of these three channels shows distinctive dy-

namical behaviors. The OH products are largely forward

scattered, implying an abstraction type reaction mechanism

is responsible for this channel. The H SiO/H SiOH products

3

2

are scattered backward relative to the O atom beam, which is

probably due to a SN type reaction mechanism to this chan-

2

nel is possible. An H formation channel has also been also

2

observed. Since the 2HϩH SiO/HSiOH channel shows a

2

forward and backward symmetric distribution, it is likely that

this channel goes through a long-lived (H SiOH)* reaction

3

intermediate formed by O insertion to a S–H bond in SiH₄.

An intriguing 2H ϩSiO channel is also clearly present.

2

ϩ

1

FIG. 20. The total product angular distribution of the mass 44 (SiO ) prod-

uct. The solid squares are the experimental data, while the solid lines are the

simulations from various contributions. Different contributions are labeled

as in Fig. 12.

From the above analyses, the O( D) reaction seems to be

3

more important than the O( P) reaction in most of the chan-

nels.

Even though this study is qualitative in nature, it, never-

theless, provides a good experimental case of a complicated

reaction with multiple pathways and triple product channels.

With improved crossed molecular beam methods, it is now

possible to study the dynamics of these complicated reac-

tions under molecular beam conditions. More experimental

studies on this reaction using either pure O( D) or O( P)

beam are certainly desirable in order to clear some of the

experimental facts. Further theoretical investigations are also

in need in order to better understand this complicated reac-

tion.

F. Relative branching ratios

From the above analyses of the different product chan-

nels, relative branching ratios can also be estimated. Assum-

ing the transmission efﬁciencies for all different ions are the

same, and ionization efﬁciencies for masses 44–48 are twice

that of mass 17 ͑OH͒,⁴¹the OH channel should count about

1

3

48%, while other channels count the rest 52%. Among the

other channels, the 2H ϩSiO channel is estimated to be

2

1

2

1%, the H ϩH SiO/HSiOH channel counts about 15%, the

2 2

HϩH SiO/HSiOH triple product channel is about 21%, and

2

the HϩH SiO/H SiOH channel ͑mass 47͒ is about 5%.

3

2

ACKNOWLEDGMENTS

These branching ratios are only roughly estimated numbers.

It is interesting to point out that triple product processes are

quite important in the title reaction, and the most important

channel is the OH formation channel, similar to the O atom

reactions with hydrocarbon molecules.

This work is supported by the National Science Council,

Academia Sinica of the Republic of China. Financial support

from the China Petroleum Company is also greatly appreci-

ated. We also appreciate helpful discussions with Dr. A. H.

H. Chang and Dr. A. Mebel.

IV. CONCLUSIONS

From the above investigations, a total of six different

1

J. J. Lin, D. W. Hwang, S. Harich, Y. T. Lee, and X. Yang, Rev. Sci.

Instrum. 69, 1642 ͑1998͒.

S. Koda and O. Fujiwara, 21st International Symposium on Combustion

channels have been observed in the OϩSiH reaction in this

4

2

͑

The Combustion Institute, Pittsburgh, 1986͒, p. 1861.

3

J. R. Hartman, J. Famil-Ghiriha, M. A. Ring, and H. E. O’Neal, Combust.

Flame 68, 43 ͑1987͒.

4

5

6

7

8

9

S. Koda and O. Fujiwara, Combust. Flame 73, 187 ͑1988͒.

S. Suga and S. Koda, Jpn. J. Appl. Phys. 27, L1966 ͑1988͒.

J. M. Jasinski and S. M. Gates, Acc. Chem. Res. 24, 9 ͑1991͒.

M. S. Gordon and L. A. Pederson, J. Phys. Chem. 9, 5527 ͑1990͒.

T. Kudo and S. Nagase, J. Phys. Chem. 88, 2833 ͑1984͒.

A. Talibana, H. Fueno, and T. Yamabe, J. Am. Chem. Soc. 108, 4346

͑

1986͒.

1

0

1

B. Ma and H. F. Schaefer III, J. Chem. Phys. 101, 2734 ͑1994͒.

A. Tachibana, M. Koizumi, H. Teramae, and T. Yamabe, J. Am. Chem.

Soc. 109, 1383 ͑1987͒.

S. Saki, M. S. Gordon, and K. Jordan, J. Phys. Chem. 92, 7053 ͑1988͒.

K. Raghavachari, J. Chandrasekhar, and M. J. Frisch, J. Am. Chem. Soc.

1

2

3

104, 3779 ͑1982͒.

14

K. Raghavachari, J. Chandrasekhar, M. J. Gordon, and K. Dykema, J. Am.

Chem. Soc. 106, 5833 ͑1984͒.

S. Su and M. S. Gordon, Chem. Phys. Lett. 204, 306 ͑1993͒.

S. Nagase, T. Kudo, T. Akasaka, and W. Ando, Chem. Phys. Lett. 163, 23

͑1989͒.

1

5

6

FIG. 21. The translational energy distribution ͑with isotropic angular distri-

bution͒ for the 2H ϩSiO reaction channel used to simulate the fast contri-

2

bution to the TOF spectra and the laboratory angular distribution of mass 44

using a 44:4 mass ratio. These extremely fast products are likely due to the

interesting triple product channels.

1

7

T. Akasaka, S. Nagase, A. Yabe, and A. Ando, J. Am. Chem. Soc. 110,

6270 ͑1988͒.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

1842

J. Chem. Phys., Vol. 113, No. 5, 1 August 2000

Lin, Lee, and Yang

1

8

30

R. Atkinson and J. N. Pitts, Jr., Int. J. Chem. Kinet. 10, 1151 ͑1978͒.

T. G. Mkryan, E. N. Sarkisyan, and S. A. Arutyunyan, Arm. Khim. Zh.

S. Koda, S. Suga, S. Tsuchiya, T. Suzuki, C. Yamada, and E. Hirota,

Chem. Phys. Lett. 161, 35 ͑1989͒.

9

3

1

2

3

4, 3 ͑1981͒; S. A. Kharutunyan, T. G. Mkryan, and E. N. Sarkisyan,

A. Tachibana and K. Sakata, Appl. Surf. Sci. 117, 151 ͑1997͒.

S. Satyapal, J. Park, and R. Bersohn, J. Chem. Phys. 91, 6873 ͑1989͒.

C. R. Park and J. R. Wiesenfeld, J. Chem. Phys. 95, 8166 ͑1991͒.

W. Hack and H. Thiesemann, J. Phys. Chem. 99, 17364 ͑1995͒.

C. L. Darling and H. Bernhard Schlegel, J. Phys. Chem. 97, 8207 ͑1993͒.

M. R. Zachariah and W. Tsang, J. Phys. Chem. 99, 5308 ͑1995͒.

A. Tachibana and K. Sakata, Appl. Surf. Sci. 117Õ118, 151 ͑1997͒.

K. Watanabe, E. C. Y. Inn, and M. Zelikoff, J. Chem. Phys. 21, 1026

͑1953͒.

J. J. Lin, D. W. Dennis, Y. T. Lee, and X. Yang, J. Chem. Phys. 109, 1758

͑1998͒.

J. J. Lin, S. Harich, Y. T. Lee, and X. Yang, J. Chem. Phys. 110, 10821

͑1989͒.

Oxid. Commun. 7, 49 ͑1984͒.

2

0

1

2

L. Ding and O. Marshall, J. Chem. Phys. 98, 8545 ͑1993͒.

D. Iida, M. Koshi, and H. Matsui, Isr. J. Chem. 36, 285 ͑1996͒.

C. R. Park, G. D. White, and J. Wiesenfeld, J. Phys. Chem. 92, 152

33

34

35

36

37

38

͑1988͒.

2

3

4

B. S. Agrawalla and D. W. Setser, J. Chem. Phys. 86, 5421 ͑1987͒.

O. Horie, R. Taega, B. Reimann, N. L. Arthur, and P. Potzinger, J. Phys.

Chem. 95, 4393 ͑1991͒.

K. Sugawara, T. Nakanaga, H. Takeo, and C. Matsumura, Chem. Phys.

Lett. 157, 109 ͑1989͒.

2

5

6

3

4

9

0

1

I. R. Slagle, J. R. Bernhardt, and D. Gutman, Chem. Phys. Lett. 149, 180

͑1988͒.

2

7

8

9

M. Koshi, A. Miyoshi, and H. Matsui, J. Phys. Chem. 95, 9869 ͑1991͒.

S. Koda, Prog. Energy Combust. Sci. 18, 513 ͑1992͒.

K. Okuda, K. Yunoki, T. Oguchi, Y. Murakami, A. Tezaki, M. Koshi, and

H. Matsui, J. Phys. Chem. A 101, 2365 ͑1997͒.

The ionization cross-sections for H SiO/H SiOH, H SiO/HSiOH, HSiO/

3

2

SiOH are not known. The 2:1 ionization cross-section ratio is estimated

from the similar carbon species.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

55.33.16.124 On: Mon, 01 Dec 2014 02:40:23

1

Article Doi

Multiple dynamical pathways in the O+SiH4 reaction studied by the crossed molecular beam method

DOI: 10.1063/1.481987

Source and publish data:

Authors:

Article abstract of DOI:10.1063/1.481987

Full text of DOI:10.1063/1.481987

Products guided by the article

R&D Labs maybe for 3352-57-6

Relevant to this article

Hot Product