Vacuum ultraviolet photochemistry of CH4 and isotopomers. II. Product channel fields and absorption spectra

Article abstract of DOI:10.1063/1.1288145

Methyl and methylene fragments are achieved using a complete set of quantum yields for different photodissociation channels. The results depicted that the hydrogen (H) atoms are more abundant in photofragments than deuterium (D) atoms. The ultraviolet absorption spectrum of methane is identical at different temperatures. The quantum yields of H atoms are determined by the measurement of the ratio of areas under the laser-induced fluorescence excitation curves.

Full text of DOI:10.1063/1.1288145

Vacuum ultraviolet photochemistry of CH 4 and isotopomers. II. Product channel fields
and absorption spectra
Jen-Han Wang, Kopin Liu, Zhiyuan Min, Hongmei Su, Richard Bersohn, Jack Preses, and John Z. Larese
Citation: The Journal of Chemical Physics 113, 4146 (2000); doi: 10.1063/1.1288145
View online: http://dx.doi.org/10.1063/1.1288145
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/113/10?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Electronic spectroscopy of the deuterated isotopomers of the NO∙methane molecular complex
J. Chem. Phys. 123, 204305 (2005); 10.1063/1.2125748
Determining the vibrational pattern via overtone cold spectra: C–H methyl stretches of propyne
J. Chem. Phys. 122, 224316 (2005); 10.1063/1.1930832
Dynamics of photodissociation of ethylene and its isotopomers at 157 nm: Branching ratios and kinetic-energy
distributions
J. Chem. Phys. 120, 10983 (2004); 10.1063/1.1740711
Photodissociation of propyne and allene at 193 nm with vacuum ultraviolet detection of the products
J. Chem. Phys. 110, 3320 (1999); 10.1063/1.478197
VUV photochemistry of CH 4 and isotopomers. I. Dynamics and dissociation pathway of the H/D-atom
elimination channel
J. Chem. Phys. 109, 7105 (1998); 10.1063/1.477394
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

JOURNAL OF CHEMICAL PHYSICS
VOLUME 113, NUMBER 10
8 SEPTEMBER 2000
Vacuum ultraviolet photochemistry of CH4 and isotopomers. II. Product
channel fields and absorption spectra
Jen-Han Wang and Kopin Liu
Institute of Atomic and Molecular Science (IAMS), Academia Sinica, Taipei, Taiwan 10764
Zhiyuan Min, Hongmei Su, and Richard Bersohn
Department of Chemistry, Columbia University, New York, New York 10027
Jack Preses and John Z. Larese
Chemistry Department 555A, Brookhaven National Laboratory, Upton, New York 11973
͑Received 12 May 2000; accepted 13 June 2000͒
In part I of this work the relative velocities and anisotropies of the atomic H and D fragments from
methane photolysis at 10.2 eV were measured. In this paper the relative abundance of the methyl
and methylene fragments are reported. A complete set of quantum yields for the different
photodissociation channels of each isotopomer is obtained by combining the two sets of data.
Previously it was found that H atoms are almost four times more likely than D atoms to be ejected;
now it is found that hydrogen molecule photofragments are much richer in H atoms than in D.
Overall, the heavier D atoms are more likely than the H atoms to remain attached to the carbon
atom. An implication for astrophysics is discussed. The VUV absorption spectra of CH4 and CH3D
are almost identical both at room temperature and 75 K. There is, as expected, no variation in the
absorption spectrum with temperature. Evidence is given that all or almost all of the methylene is
3
produced in the a 1A1 and not in the ground B1 state. © 2000 American Institute of Physics.
͓S0021-9606͑00͒00934-X͔
I. INTRODUCTION
rapid. Photodissociation of methane is therefore mainly an
indirect process in that dissociation does not take place from
S1 but from the triplet T1 state or from the S0 ground state.
The triplet state can correlate only to CH3ϩH, whereas the
ground state can correlate to CH3ϩH as well as to
CH2(a 1A1)ϩH2.
Because there is a large excess energy in the nine normal
modes of the molecule and an extensive reorganization of all
the normal modes on the excited surface occurs, the spec-
trum consists of thousands of weak vibronic transitions.
Mebel et al. showed that the two broad humps and the small
bumps are accidental bunches of transitions. They have no
simple physical meaning.
The previous paper ͑part I͒ reported the hydrogen atom
yields, velocity distributions, and anisotropy parameters in
the photodissociation of CH4, CH3D, CH2D2, and CHD3.6
The present paper describes the measurement of the relative
yields of methylene and methyl radicals derived from the
methane isotopomers. With the help of some assumptions, a
complete table of quantum yields is obtained. In addition, the
absorption spectrum of CH4 and CH3D have been measured
both at room temperature and 75 K. Finally, the atomic H/D
ratio at 118.2 nm is reported.
Methane is the most abundant hydrocarbon on earth and
is present in small amounts in the atmospheres of all the
planets. For example, in the atmosphere of Jupiter its mole
fraction is about 2ϫ10Ϫ3.1 Other hydrocarbons are even less
abundant. The most remarkable exception is the atmosphere
of Titan, the largest moon of Saturn which has 2%–3%
methane and heavier hydrocarbons as well. The observations
of Titan made by the Voyager space craft are tantalizing but
incomplete.2 A new space craft, Cassini, is on its way and is
expected to reach Saturn in 2004. In order to construct a
photochemical model for this atmosphere, it is essential to
know the yields of the different photodissociation channels
at 121.6 nm ͑10.2 eV͒. The yields are also of intrinsic inter-
est for the theory of photodissociation of small molecules.
The choice of photon wavelength is determined by two
closely linked considerations. Nearly 92% of the solar VUV
radiation absorbed by methane is due to the 2p-1s emission
of H atoms, the Lyman ␣ line.3 This same transition is used
as a laboratory detector for the hydrogen atoms produced by
absorption.
The electronic absorption of methane begins only around
140 nm ͑ϳ8.8 eV͒.4 The VUV spectrum consists of two not
well-resolved humps which were at one time assigned to two
different electronic transitions. Recent ab initio calculations
by Mebel et al. have finally revealed the nature of the
spectrum.5 The absorption from 9 to 12 eV is due to a tran-
II. EXPERIMENT
A. Total quantum yield for H atoms from CH4
„Columbia…
1
sition to a single, triply degenerate excited state, T2 . The
The total quantum yield of H atoms is determined by
measuring the ratio of the areas under the laser-induced fluo-
rescence excitation curves of CH4 and another molecule with
known quantum yield. H2O is chosen because it has a unit
quantum yield. The experiment is a one-color experiment in
excited state correlates to CH2(b 1B1)ϩH2 which is a very
minor channel with yield р2%. However, on the excited
potential surface are minima with C3V or C2V symmetry.
These minima are not stable because surface crossing is
0021-9606/2000/113(10)/4146/7/$17.00
4146
© 2000 American Institute of Physics
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

4147
J. Chem. Phys., Vol. 113, No. 10, 8 September 2000
VUV photochemistry of CH4
D. Atomic HÕD ratios from methane isotopomers at
118.4 and 121.6 nm „Columbia…
that the Ly␣ is used both to dissociate and to probe the H
atom photoproduct. The Ly␣ was generated by mixing two
photons at 212.6 nm resonant with a two photon transition
and a third photon around 845 nm in a Kr cell containing Ar.
The optimum amount of Ar depends on the laser intensity so
it has to be optimized with each run. The signals from meth-
ane and water vapor were measured six times alternately
with water at 80 mTorr and methane at 50 mTorr. The sig-
nals were normalized to equal pressures and corrected for
absorption of the exciting light and the emitted light within
the 8 cm cubic cell.
All experiments were carried out inside an 8 cm cubic
stainless steel cell. Ly␣ light was generated by four-wave
mixing of a fixed 212.6 nm wavelength with a tunable wave-
length near 845 nm. The pressure of methane was 60 mTorr.
The H/D ratios from the isotopomers CHnD4Ϫn were mea-
sured at two VUV wavelengths under two very different con-
ditions. When Ly␣ was both the dissociating and probing
light, it was naturally made as intense as possible. The coun-
terpropagating 118.4 nm light was made by focusing the
third harmonic, 355 nm of a YAG laser in a mixture of 5
Torr of xenon and 70 Torr of argon. The observed signal was
the sum of a signal proportional to the product of the 118.4
nm intensity times the Ly␣ intensity plus a signal quadratic
in the Ly␣ intensity. Under these conditions the Ly␣ inten-
sity was reduced so that the quadratic effect was 20%–30%
of the main signal. The probe laser was scanned through the
absorption of both H and D and then the 118.4 nm laser was
blocked and the spectrum run again. The second spectrum
was subtracted from the first and the cell was filled with a
fresh gas sample. The uncertainty quoted in the final result is
the rms deviation of the six repetitions of this process. An
additional change in conditions was necessary because of the
high velocity of the hydrogen atoms produced and the often
large variation in the long 500 ns time delay between the
YAG based dissociating laser and the excimer pumped dye
based Ly␣. To prevent selective loss of the faster H atoms,
800 Torr of helium was added. The helium does not quench
the H atom fluorescence and the high pressure keeps the
atoms from migrating out of the probe beam.
B. Photoionization yield measurements „IAMS…
The experiment was carried out in a crossed-beam appa-
ratus as detailed in part I. The VUV laser at 121.6 nm, which
served as both the photolysis and the fragment ionization
photon sources, was generated by a frequency tripling pro-
cess in a Kr gas cell. To avoid other multiphoton complica-
tions, the VUV laser beam was unfocused. It had the dimen-
sions ϳ3 mmϫ5 mm through the interaction region and the
more intense UV beam ͑364.8 nm, Ͻ4 mJ/pulse͒ diverged.
The ion TOF mass spectrometer was operated in the usual
mass mode. The slit aperture, which was used in part I for
velocity measurement, was removed to ensure the detection
of all ions. Research grade CH4(99.99%) and isotopomers
͑98% from Cambridge Isotope Lab.͒ were used as received.
The supersonically cooled beam was composed of 20%CH4
or its isotopomer in He with a total backing pressure of 2
atm. To ensure that the beam is free of clusters, the laser was
timed to fire at the leading part of the pulsed beam. The
VUV photoionization mass spectra were acquired with the
molecular beam on and off, as described in Ref. 7. Over the
mass range of this work, the spectra are essentially back-
ground free. The integration of each peak, to account for the
small difference in width, gives the relative ion yields. The
normalization of the photoionization yield spectra for differ-
ent isotopomers was carried out by several back-to-back
measurements.
III. RESULTS AND DISCUSSION
A. Electronic state of methylene photofragment
There are two distinct observations. One is that when
irradiating with 10.2 eV photons, ion signals appear due to
CHϩ2 and its isotopomers. The second is that when a gas
containing methane and a hundredfold excess of N2 is irra-
diated at 118.2 nm ͑10.49 eV͒, a large IR absorption appears
due to C2H4. If the N2 is replaced by H2, there is no de-
tectible C2H4.
Methylene could, in principle, be formed in either the
ground X 3B1 state or the first excited a 1A1 state, 3156
Ϯ5 cmϪ1 higher. At the outset, we wish to present evidence
that some of the CH2 is produced in the a 1A1 state and some
additional evidence that almost all of it is. The ionization
potential of CH2(X) is 10.3962Ϯ0.0036 eV and that of
CH2(a) is 10.005 eV.9 The Lyman alpha photon whose en-
ergy is 10.199 eV can ionize CH2(a) but not CH2(X). The
ion signals seen of CHϩ2 and of its isotopomers are therefore
due to methylene in the a state.
C. Absorption spectra „Brookhaven…
The absorption spectra of CH4(99%) and CH3D(98%)
͑from Cambridge Isotope Lab.͒ were measured using the
U11 beam line of the National Synchrotron Light Source at
Brookhaven National Laboratory. A 10 cm long cell was
used with LiF windows and a 0.1 nm resolution. The corre-
sponding optimal room temperature pressures for methane
were ϳ0.2 Torr. A program available at the beam line was
used to derive VUV absorption spectra8 from recordings of
VUV transmission through the cell with and without meth-
ane. The program made appropriate normalizations for varia-
tions in source intensity and absorption spectra were derived
from the measurements. Low temperature was maintained by
cooling the cell to 75 K with a closed cycle He cryostat.
Because of temperature variations between the cold cell and
the warm capacitance manometer used to measure pressure,
molar extinction coefficients could not be determined quan-
titatively at low temperature.
The photochemical result implies that all or almost all
the CH2 is formed in the a state. CH2(a) is readily quenched
by N2 to CH2(X). The process
CH a 1A ͒ϩN → H CNN →CH X 3B ͒ϩN
͑
͓
͔
͑
2
2
1
2
2
1
2
is analogous to the very efficient isoelectronic reaction:
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

4148
J. Chem. Phys., Vol. 113, No. 10, 8 September 2000
Wang et al.
TABLE I. Photoion signal strengths ͑arbitrary un͒ of various mass peaks at
10.2 eV photon energy.
M/e
CH4
CH3D
CH2D2
CHD3
14
15
16
17
18
1.52
0.62
0.666
0.952
0.608
0.173
0.957
0.590
0.354
0.685
0.948
0.156
0.192
1
3
O D͒ϩN → NNO →O P͒ϩN .
͑
͓
͔
͑
2
2
When H2 replaces N2, it reacts easily with the nascent ener-
gized singlet methylenes through a methane transition state
just as O(1D) reacts rapidly with H2 through an intermediate
H2O transition state.
FIG. 2. The relative velocity distribution for H atoms photodissociated at
10.2 eV from CH4. The two peaked distribution is decomposed into a fast,
medium, and slow distribution ͑see text and part I͒.
B. Photofragment yields
on their translational and internal energies. Direct measure-
ment of the nascent products when laser-based methods are
available is highly desirable.
For methane, four important dissociation channels have
been found:
The observed ion signals for each isotopomer are listed
in Table I. These are not the relative yields because we do
not know the ratio, designated x, of the ionization cross sec-
tion for CH2 to that for CH3 ,␴(10.2)CH /␴(10.2)CH . We
2
3
deduce the value of x from the previous detailed results on
hydrogen atom velocity distributions and yields from the
photodissociation of methane and its isotopomers at 10.2
eV.6 It turns out that the two ionization cross sections are
almost equal, so that Fig. 1 gives at a glance a picture of the
relative methyl and methylene yields.
There is extensive literature dealing with the classical
photochemistry of methane.10–12 The gas is irradiated with a
lamp for a certain period of time, the products are analyzed
and the elementary processes are deduced. If all the products
were stable molecules, the backward step from the composi-
tion of the irradiated gas to the elementary processes would
be simple. Unfortunately, besides stable molecules, the prod-
ucts include reactive intermediates such as atoms, radicals,
and carbenes which react with each other as well as the par-
ent molecules. Moreover, their reaction rates often depend
CH4ϩh␯→CH3ϩH
͑1a͒
͑1b͒
͑1c͒
͑1d͒
→CH a 1A ͒ϩH
͑
2
1
2
→CH a 1A ͒ϩHϩH
͑
2
1
→CHϩH2ϩH.
In part I it was found that the translational energy distri-
butions of H and D atoms from all the isotopomers are bi-
modal having a low energy and a high energy peak ͑see Fig.
21͒. At first, hydrogen atoms are divided into two classes:
those with so much translational energy that a three-body
dissociation is energetically impossible and those moving
slowly enough that three-body dissociations are possible.
The threshold for channels c and d are 9.23 and 9.05 eV,
respectively. We assume therefore that all atoms with ener-
gies Ͻ1.15 eV result from a three-body dissociation. As
shown in Fig. 2, this assumption means that the area under
the first peak and a lower energy side of the area under the
second peak are assigned to three-body dissociations.
The product hydrogen atoms can be subdivided accord-
ing to whether they could or could not be a product of a
three-body dissociation. Another subdivision is based on the
speed dependent anisotropy parameter, ␤. As explained in
part I, the speed distributions can be fitted by a sum of three
distributions, labeled fast, medium, and slow. The fast distri-
bution has a ␤ϭϪ0.45, the medium distribution has a ␤
which is close to 2 at high energy but diminishes to zero at
low energy. The third slow component has a ␤ close to zero.
Table III in part I lists the fraction of the H and D atoms
assigned to the three categories. Figure 2 shows the decom-
position for methane.
The a channel has, in effect, been divided into two chan-
nels, ‘‘fast’’ and the higher energy part of the ‘‘medium.’’
The fastest hydrogen atoms which have a constant negative
␤ are believed to dissociate at the earliest time. The medium
FIG. 1. A bar graph of the data of Table I showing relative ion signals of the
photofragments of methane and its isotopomers irradiated at 10.2 eV. All
ion signals are on the same arbitrary scale.
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

4149
J. Chem. Phys., Vol. 113, No. 10, 8 September 2000
VUV photochemistry of CH4
TABLE II. Relative H and D ion signals, partitioned between two-and
three-body fragmentation. These numbers were obtained from the areas of
Figs. 7 and 9 of Ref. 6 above and below the threshold for three-body dis-
sociation. Signals are normalized to unity for CH4.
TABLE III. Photodissociation channels of methane isotopomers. The sum
of the yields for any molecule is, by definition, one but the closeness of the
sum to one is used in Tables IV and VI as a check on overall consistency.
CH3D
Yield
CHD3
Yield
CH2D2
Yield
Channel
CH4
CH3D
CH2D2
CHD3
CH2DϩH
CH3ϩD
CHDϩH2
CH2ϩHD
CHD2ϩD
CD3ϩH
CHDϩD2
CD2ϩHD
CHD2ϩH
CH2DϩD
CD2ϩH2
CHDϩHD
CH2ϩD2
CD2ϩHϩH
CHDϩHϩD
CH2ϩDϩD
CDϩHDϩH
CHϩD2ϩH
CHϩHDϩD
CDϩH2ϩD
a1
a2
b1
b2
a1
a2
b1
b2
a1
a2
b1
b2
b3
c1
c2
c3
d1
d2
d3
d4
Ј
Ј
Ј
Ј
Љ
Љ
Љ
Љ
Љ
Љ
Љ
Љ
Љ
Љ
Љ
Љ
Two fragment H atom
Three fragment H atom
Two fragment D atom
Three fragment D atom
0.620
0.380
0.608
0.363
0.032
0.355
0.165
0.071
0.058
0.192
0.061
0.156
0.094
CHDϩHϩH
CH2ϩHϩD
CHDϩDϩD
CD2ϩHϩD
c1
c2
c1
c2
Ј
Ј
CHϩHDϩH
CDϩH2ϩH
CHϩH2ϩD
CDϩHDϩD
CHϩD2ϩD
CDϩD2ϩH
d1
d2
d3
d1
d2
d3
Ј
Ј
Ј
hydrogen atoms have a broader speed distribution, and a ␤
which decreases with decreasing energy. Hydrogen atoms
with lower speeds are formed simultaneously with methyls
with more internal energy. The dissociation process probably
takes progressively longer. When the atom has a kinetic en-
ergy Ͻ1.15 eV(15ϫ105 cm/s), the methyl radical partner
has enough energy to dissociate into CHϩH2. We assume
that this channel switches on immediately above threshold
because the reaction CHϩH2→CH3 is facile, i.e., there is an
angle of approach for which there is no barrier to reaction
and therefore there is no barrier for the reverse reaction.12
Calculation of absolute quantum yields proceeds as fol-
lows. Let the quantum yields of these four channels be de-
noted by a, b, c, and d, respectively. By definition,
dϭ1Ϫ0.62YϪ1.520y/x.
͑8͒
To evaluate the unknown x, the ratio of the photoioniza-
tion cross section of CH2 to that of CH3, one additional
assumption is necessary. This is that the d channel H atoms
are the low energy tail of the ‘‘medium’’ component of the H
atom translational energy distribution. If so, the ‘‘slow’’
component H atoms must be assigned to the c channel. Re-
ferring to Table III of part I, the slow component is a fraction
0.232 of the total H atom yield. Therefore, 2cϭ0.232y is an
additional equation which allows x to be obtained. If c
ϭ0.116y, then
aϩbϩcϩdϭ1.
͑2͒
The total quantum yield for H atom production has been
found by Brownsword et al. to be 0.47Ϯ0.11.13 We have
independently measured the same yield as described in the
experiment section. The ratio of the signal from H2O to that
from CH4 was 1.77Ϯ0.05 and the ratio of the absorption
coefficient of H2O to that of CH4 is (1.6Ϯ0.1)/(2.0Ϯ0.1)
ϭ0.80Ϯ0.08. The total quantum yield assuming unit quan-
tum yield for water is 0.45Ϯ0.10, insignificantly different
from the prior result. We can then write:
xϭ1.520y/ 1Ϫ0.768y͒.
͑9͒
͑
If yϭ0.47Ϯ0.11, then xϭ1.118Ϯ0.420
Substituting these values of x and y in Eqs. ͑5͒–͑8͒, one
obtains the absolute yields:
aϭ0.291Ϯ0.068,
bϭ0.585ϯ0.098,
aϩ2cϩdϭy, where y is 0.47Ϯ0.11.
From Table I we can write the ratio of the CHϩ2 signal to that
of the CHϩ3 :
͑3͒
͑10͒
cϭ0.066Ϯ0.012,
dϭ0.068Ϯ0.013.
The uncertainties included in the above quantum yields are
those due to the Ϯ0.11 uncertainty in y, the total quantum
yields for producing H atoms. The measurement of y is quite
independent of the measurements reported here and in part I.
Moreover, x could be independently measured. Thus it is
hoped that future investigators could use Eqs. ͑5͒–͑8͒ with
improved values of x and y.
The various channels for the dissociation of the isoto-
pomers, CHnD4Ϫn are listed in Table III. The absolute yields
listed in Table IV were calculated using all of the following
data
x bϩc͒/aϭ1.52/0.62ϭ2.452.
͑4͒
͑
In other words, the ratio of the ion signals is equal to the
ratio of the quantum yield for producing CH2 to that for CH3
times the ratio of the photoionization cross sections, x.
Table II contains a set of ion signals from H and D
atoms partitioned between those from two or three-body
fragmentation. The normalization is such that the total H
atom signal from CH4 is taken as unity. However, the total
quantum yield for H atoms from CH4 is y. Therefore, the
quantum yield for H atoms produced by the two-body chan-
nel is
͑1͒ The ‘‘slow’’ component fractions of the total H and D
atom velocity distributions using the data in Table III of
part I.
aϭ0.62y.
͑5͒
Equations ͑2͒–͑4͒ can be solved for the individual quan-
tum yields b, c, and d in terms of x and y;
͑2͒ The ratios of two-fragment to three-fragment yields
͑Table II͒ deduced from the areas of Figs. 7 and 9 of Part
bϭ0.760y/xϪy/2ϩ1/2,
͑6͒
I above and below ϭ15ϫ105 cm/s.
v
cϭ0.760y/xϩy/2Ϫ1/2,
͑7͒
͑3͒ The H/D ratios from Table IV of part I.
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

4150
J. Chem. Phys., Vol. 113, No. 10, 8 September 2000
Wang et al.
TABLE V. Photoionization yields of the observed molecular fragments at
10.2 eV photon energy.
͑4͒ The ion signal ratios of the various fragment masses de-
rived from each molecule of Table I using xϭ1.118; and
͑5͒ The ratio of a unique methyl ion signal from an isoto-
pomer to that of methane.
Fragment
CH4
CH3D
CH2D2
0.073
CHD3
0.619
0.291
0.280
0.015
0.387
0.286
CH2
CH3
Table IV lists the yields of the different photodissocia-
tion channels. Table V lists the yields of specific isotopomers
of the methylene and methyl fragments. The overall picture
is obtained from Table VI which lists just the four yields
from each chemically distinct channel. The sum of the four
yields for methane has been normalized to one. The sum of
the yields of the other isotopomers has not been normalized.
The fact that the sum of their yields is close to one supports
the assumptions used in partitioning the hydrogen atom ve-
locity distributions. No correction has been applied for the
different coefficients for absorption of the 10.2 eV light.
Hydrogen molecules were not directly probed in these
experiments but the relative yields of H2, HD, and D2 can be
determined approximately. The following equations give the
yields of the three hydrogen molecule isotopomers using
Tables II and IV.
CHD
CH2D
CD2
CHD2
CD3
0.402
0.033
0.218
0.167
0.288
0.399
0.073
0.090
H2 /HD ratio is ϳ1.23. Similarly from CHD3, HD and D2
should be formed in equal numbers but the HD/D2 ratio is
ϳ1.32. Finally, CH2D2 should form H2, HD, and D2 in a
1:4:1 ratio but the observed ratio is 1:1.85:0.33.
In part I an important photochemical discovery was that
the atomic H/D ratio from a molecule CHnD4Ϫn was three to
four times the statistically expected ratio n/(4Ϫn) . We
͓
͔
now find that the abundance of H atoms in hydrogen mol-
ecules is also highly nonstatistical. A broad qualitative ex-
planation is that the intramolecular forces acting on H and D
atoms are the same but the H atoms have, classically, twice
the acceleration. This means that an H atom under the action
of a force will pass the point of no return sooner than a D
atom subjected to the same force. However, there are two
kinds of points of no return. If the force is predominantly
radial, the atom which passes a certain C–H separation be-
comes a free atom. If the force is predominantly angular, the
point of no return is an atom–atom distance less than which
a hydrogen molecule has been formed. Obviously forces are
not purely radial or angular. The mixed directions produce
hydrogen atoms with reduced ␤. It would also follow that
there is not a unique C–H or H–H distance beyond which
there is product formation. This is a zero order explanation
which needs further theoretical elaboration. It must be em-
phasized that once the H and D atoms really leave the mol-
ecule, they are detected with equal probability.
The overall results in Table VI can be compared with
previous photochemical studies done on CH4 and CD4.8–10
Substantial isotope effects on the yields of different channels
were found. Fewer atoms are released by CD4 than by CH4.
Less CD is formed from CD4 than CH from CH4. These
conclusions are now seen to be part of a continuous trend
with the number of deuterium atoms in the molecule. The
hydrogen atom yield steadily decreases with increasing deu-
For CH3D,
␾
␾
2ϭb1ϩd2ϩd3ϭ0.334ϩϽ0.063,
H
HDϭb2ϩd1ϭ0.271ϩϽ0.054.
For CHD3,
␾
␾
ϭb2 d1 ϭ0.379ϩϽ0.025,
Ј Ј
HD
ϭb1 ϩd2 ϩd3 ϭ0.287ϩϽ0.010.
Ј
Ј
Ј
D
2
For CH2D2,
␾
ϭ b1 ϩc1 ͒Ϫc1 ϩd4
͑
Љ Љ Љ Љ
H
2
ϭ0.218ϪϽ0.020ϩϽ0.027,
ϭ b2 ϩc2 ͒Ϫc2 ϩd1 ϩd3
Љ
␾
␾
͑
Љ
Љ
Љ
Љ
HD
ϭ0.402ϪϽ0.014ϩϽ0.064,
ϭ b3 ϩc3 ͒Ϫc3 ϩd2
͑
Љ
Љ
Љ
Љ
D
2
ϭ0.073ϪϽ0.014ϩϽ0.037.
The inequalities appear in the above equations because
the small d and c coefficients have not been individually
Љ
determined. From CH3D on statistical grounds H2 and HD
should be formed in equal numbers. The experimental
TABLE IV. Quantum yields of isotopically distinct product channels.
TABLE VI. Fractional yields of four chemically distinct product channels.
X denotes H or D and X2 denotes H2, HD, or D2.
CH4
CH3D
CH2D2
0.166(a1 )
CHD3
0.073(a1 )
Channel
CH4
CH3D
CH2D2
CHD3
0.286(a1)
0.015(a2)
0.334(b1)
0.271(b2)
0.054(c1)
0.009(c2)
0.054(d1ϩd2)
0.008(d3)
0.291(a)
0.585(b)
0.055(c)
0.070(d)
Љ
Ј
0.034(a2 )
0.090(a2 )
͑a͒ CX3ϩX
͑b͒ CX2ϩX2
͑c͒ CX2ϩXϩX
͑d͒ CXϩX2ϩX
Total yield
X yield
0.291
0.584
0.055
0.070
1.000
0.470
0.654
0.719
0.301
0.605
0.063
0.063
1.032
0.490
0.668
0.734
0.200
0.665
0.027
0.065
0.957
0.319
0.730
0.437
0.163
0.667
0.019
0.035
0.884
0.236
0.702
0.336
Љ
Ј
0.218(b1 ϩc1 )
0.287(b1 )
Ј
0.379(b2 )
Ј
0.000(c1 )
Ј
0.019(c2 )
Ј
Љ
Љ
Љ
Љ
Љ
0.402(b2 ϩc2 )
Љ
0.073(b3 ϩc3 )
Љ
0.037(d1 ϩd2 )
Љ
0.027(d3 ϩd4 )
0.25(d1 ϩd2 )
Ј Ј
Љ
Љ
0.040(2c1 ϩc2 )
0.010(d3 )
X2 yield
X/X2
Љ
Љ
Ј
0.014(2c3 ϩc2 )
Љ
Љ
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

4151
J. Chem. Phys., Vol. 113, No. 10, 8 September 2000
VUV photochemistry of CH4
FIG. 4. Room temperature absorption spectra of CH4 and CH3D at 0.192
Torr.
shape at 75 K as at room temperature. This is not a surprise
because at room temperature only 0.7% of methane mol-
ecules are in an excited vibrational state. Only these mol-
ecules would be expected to have different absorption coef-
ficients at the two temperatures. In any case, even the small
protuberances on top of the larger two peaks are reproduced
at low temperature and are therefore not due to rotations.
Some time ago, somewhat coarse-grained spectra of CH4
and CD4 were reported which showed that at 10.2eV CD4
absorbed about 20% more strongly than CH4.16 We have
measured the spectra of CH3D and found, as shown in Fig. 4,
that the spectrum is almost identical in shape and strength to
that of CH4. This means that isotope dependent absorption
does not play a significant role in the photochemistry of
naturally occurring methane.17 As with CH4, the room tem-
perature and 75 K absorption spectrums of CH3D have ex-
actly the same shape.
FIG. 3. Absorption spectra of CH4 at 300 K and at 75 K. The cell was 10 cm
in length and the pressure was 0.192 Torr. Absorption in Figs. 3 and 4 is
ln(I/I0).
teration and, to compensate, the hydrogen molecule yield
steadily increases. In general, the D atom is more likely than
the H atom to remain bonded to the carbon atom. The pro-
cess of dissociation is very likely slower when the masses in
the system are doubled. The lighter atoms are preferentially
ejected either as atoms or as diatomic molecules.
A corollary to the statement that D atoms preferentially
remain attached to the carbon atom is that subsequent reac-
tion products such as ethane and ethylene will be richer in D
atoms than the original methane. This means that the hydro-
carbon molecules found in the atmosphere of Titan, a moon
of Saturn will be enriched in D compared to methane if, in
fact, they are derived from methane through an initial
photodissociation.14
D. Atomic HÕD ratios at 118.4 and 121.6 nm
The results of the LIF measurements of atomic H/D
yields from photodissociation at 121.6 and 118.4 nm are
shown in Table VII. The ratios obtained at 121.6 nm indicate
C. Absorption spectra of CH4 and CH3D at room T
and at 75 K
The calculation of Mebel et al. has shown that the shape
of the VUV spectrum of methane has no obvious physical
interpretation. It is nevertheless interesting to see if there is
any modification of the spectrum caused by lowering the
temperature or by isotopic substitution. The absorption of
cold methane is important for obvious astrophysical reasons
as well as because nozzle-expanded supercold molecular
beams are used in research. The question has been raised as
to whether the absorption of methane is the same at low
temperatures as at room temperature.15 Because of pressure
instabilities at low temperatures in our apparatus, we were
not able to measure an absolute absorption of cold methane
but, as shown in Fig. 3, the spectrum has exactly the same
FIG. 5. LIF excitation spectra at 121.6 nm of H and D atoms from photo-
dissociation of methane isotopomers at 121.6 nm.
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11

4152
J. Chem. Phys., Vol. 113, No. 10, 8 September 2000
Wang et al.
TABLE VII. H/D from methane isotopomers at 118.4 and 121.6 nm. The
third and fifth columns are obtained by dividing the second and fourth
columns by the statistical H/D ratio.
118.4 nm
121.6 nm
4.4Ϯ0.4
1.58Ϯ0.05
0.53Ϯ0.03
CH3D
CH2D2
CHD3
0.94Ϯ0.05
1.68Ϯ0.08
3.03Ϯ0.15
1.47Ϯ0.13
1.58Ϯ0.05
1.59Ϯ0.09
2.83Ϯ0.15
1.68Ϯ0.08
1.01Ϯ0.05
cording to the theory of Mebel et al. the upper electronic
state of methane correlates at long distances to CH2(1B1)
ϩH2. As the emission yield increases, the sum of the inter-
system and internal conversion rates must decrease. The up-
per state is becoming slightly less strongly coupled to lower
states as the energy increases. The whole question certainly
deserves further investigation.
FIG. 6. LIF excitation spectra at 121.6 nm of H and D atoms from photo-
dissociation of methane isotopomers at 118.4 nm in one atm of helium. The
weaker spectra from 121.6 nm alone have been subtracted.
ACKNOWLEDGMENTS
The research of K.L. was supported by the National Sci-
ence Council of Taiwan ͑NSC 87-2119-M-001-009-Y͒ and
the Chinese Petroleum Corp. The research of R. B. was sup-
ported by NASA. Research at Brookhaven National labora-
tory was performed under Contract No. DE-AC02-
98CH10886 with the U.S. Department of Energy and
supported by its Division of Chemical Sciences ͑J.M.P.͒ and
its Division of Material Sciences ͑J.Z.L.͒, Office of Basic
Energy Sciences.
that there is per atom a roughly constant preference for H
atom escape by a factor of 1.55Ϯ0.10. This is in qualitative
but not quantitative accord with the molecular beam RAMP
results of Ref. 6, which were used to calculate Table IV. The
latter indicated a preference for H atom escape in the range
of three to 4.7. We cannot at present identify the reason for
the difference in the two sets of data acquired with different
methods. If the Ly␣H/D ratios of Table VII are used to cal-
culate the yields of different channels instead of those in Ref.
1 S. K. Atreya, M. H. Wong, T. C. Owen, P. R. Mahaffy, H. B. Niemann, J.
dePater, P. Drossart, and Th. Encrenaz, Planet. Space Sci. 47, 1243
͑1999͒.
1, only the yields, a2, a2 , and a2 are appreciably affected.
Ј
Љ
The yield a2 becomes 0.047. The qualitative result is in line
with previous findings on a number of hydrocarbons.18 When
dissociation is from an excited state, dynamics favors the
release of the swifter H atom; from an equilibrated ground
state statistics favors the release of the more weakly bound H
atom because of its higher zero point energy.
Figures 5 and 6 are the LIF spectra of the H and D atoms
dissociated from the partially deuterated methanes. The spec-
tra of Fig. 5 were obtained from low pressure methanes. In
Fig. 6 one atmosphere of He is added to thermalize the na-
scent hydrogen atoms. The atoms can no longer escape from
the laser beam during the necessary delay time; the spectra
are sharpened and thereby the S/N is greatly improved.
The results at 118.4 nm are surprising. In CH3D there is
no preference for H atom release but that preference grows
strongly with increasing numbers of deuterium atoms. The
absorption spectrum is nearly flat between 10.20 and 10.47
eV ͑118.4 nm͒ which does not suggest any drastic change in
dynamics. There is one hint in that the weak CH2 b→a
emission is increasing rapidly as the energy increases.4 Ac-
2 D. F. Strobel, M. E. Summers, and X. Zhu, Icarus 100, 512 ͑1992͒.
3 S. K. Atreya, S. G. Edgington, Th. Encrenaz, and A. H. Feuchtgruber,
Proceedings of the Conference ‘‘The Universe as seen by ISO’’ ͑European
Space Agency SP-427, March 1999͒, p. 149.
4 L. C. Lee and C. C. Chiang, J. Chem. Phys. 78, 688 ͑1983͒.
5 A. M. Mebel, S.-H. Lin, and C.-H. Chang, J. Chem. Phys. 106, 2612
͑1997͒.
6 J.-H. Wang and K. Liu, J. Chem. Phys. 109, 7105 ͑1998͒.
7 L. H. Lai, Y.-T. Hsu, and K. Liu, Chem. Phys. Lett. 307, 385 ͑1999͒.
8 J. C. Sutherland, P. C. Keck, K. P. Griffin, and P. Z. Tacacs, Nucl. In-
strum. Methods 195, 375 ͑1982͒.
9 B. Ruscic, M. Litorja, R. L. Asher, J. Phys. Chem. A 103, 8625 ͑1999͒.
10 A. H. Laufer and J. R. McNesby, J. Chem. Phys. 49, 2272 ͑1968͒.
11 P. J. Ausloos and S. G. Lias, Annu. Rev. Phys. Chem. 22, 85 ͑1971͒; R.
Gorden, Jr. and P. Ausloos, J. Chem. Phys. 46, 4823 ͑1967͒.
12 R. G. Macdonald and K. Liu, J. Chem. Phys. 93, 2443 ͑1990͒.
13 R. A. Brownsword, M. Hillenkamp, T. Laurent, R. K. Vaisa, H-R. Volpp,
and J. Wolfrum, Chem. Phys. Lett. 266, 1258 ͑1997͒.
14 J. R. McNesby, J. Atmos. Sci. 26, 594 ͑1969͒.
15 N. S. Smith and F. Raulin, J. Geophys. Res. 104, 1873 ͑1999͒.
16 A. H. Laufer and J. R. McNesby, Can. J. Chem. 43, 3487 ͑1965͒.
17 B-M. Cheng, E. P. Chew, C-P. Liu, M. Bahou, Y-P. Lee, Y. L. Yung, and
M. F. Gerstell, Geophys. Res. Lett. 26, 3657 ͑1999͒.
18 R. Bersohn and B. Katz, in Isotope Effects in Gas-Phase Chemistry, edited
by J. A. Kaye, ACS Symposium Series, 502 ͑1992͒, p. 66.
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:
128.193.164.203 On: Mon, 22 Dec 2014 04:01:11