Molecular dissociation and two low-temperature high-pressure phases of H2S

Article abstract of DOI:10.1103/PhysRevB.69.214102

We performed powder x-ray experiments on hydrogen sulfide (H₂S) under high pressure at room and low temperatures. Phase V appearing above 27 GPa at room temperature was confirmed by x ray to be associated with a molecular dissociation. The molecular dissociation was first observed at 150 K at a higher pressure around 43 GPa. Two new low-temperature high-pressure phases III′ and IV′ were discovered. Spiral chains of sulfur atoms are found in both phases as well as in phase IV.

Full text of DOI:10.1103/PhysRevB.69.214102

PHYSICAL REVIEW B 69, 214102 (2004)
Molecular dissociation and two low-temperature high-pressure phases of H2S
Hiroshi Fujihisa,1 Hiroshi Yamawaki,1 Mami Sakashita,1 Atsuko Nakayama,2 Takahiro Yamada,1 and Katsutoshi Aoki3
1Institute for Materials and Chemical Process, National Institute of Advanced Industrial Science and Technology (AIST),
1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan
2Research Center for Advanced Carbon Materials, National Institute of Advanced Industrial Science and Technology (AIST),
1-1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan
3JAERI/SPring-8, 1-1-1, Kouto, Mikazuki-cho, Sayo-gun, Hyogo 679-5148, Japan
(Received 27 January 2004; published 7 June 2004)
We performed powder x-ray experiments on hydrogen sulfide ͑H2S͒ under high pressure at room and low
temperatures. Phase V appearing above 27 GPa at room temperature was confirmed by x ray to be associated
with a molecular dissociation. The molecular dissociation was first observed at 150 K at a higher pressure
around 43 GPa. Two new low-temperature high-pressure phases III and IV were discovered. Spiral chains of
Ј
Ј
sulfur atoms are found in both phases as well as in phase IV.
DOI: 10.1103/PhysRevB.69.214102
PACS number(s): 61.50.Ks, 61.10.Nz, 62.50.ϩp, 64.70.Kb
I. INTRODUCTION
we performed powder x-ray diffraction experiments up to
48 GPa at room temperature, and found the existence of el-
emental sulfur above 27 GPa. Our second purpose was to
explore a low-temperature high-pressure region from which
no diffraction experiments have been reported. In this study,
we measured x-ray patterns as low as 30 K and up to
43 GPa, and discovered two new phases.
Hydrogen sulfide ͑H2S͒ shows successive phase transi-
tions when cooled. As illustrated at the bottom of Fig. 1,
¯
¯
these are phase I (cubic Fm3m), phase II (cubic Pa3), and
then phase III (orthorhombic Pbcm). According to a neutron
diffraction experiment,1 the molecular rotations of phases I
and II freeze in phase III. On compression at room tempera-
ture, the other successive phase transitions have been
observed: phase I,2 phase I (cubic P213), phase IV2–4
II. EXPERIMENTAL
Ј
(tetragonal I41/acd)5 and phase V.6 The molecular rotations
Gaseous H2S with a stated purity of 99.999% was solidi-
fied in an agate mortar cooled at 77 K by liquid nitrogen.
Then the solid H2S was ground into a fine powder. The pow-
der was mounted on a diamond anvil high-pressure cell
in phases I and I are thought to stop in phase IV.2
Ј
Molecular dissociation is one of the interesting phenom-
ena observed in molecular solids under pressure. There have
been two opinions of the dissociation pressure of ͑H2S͒ de-
rived from optical experiments: 46 GPa7,8 and 27 GPa.9 A
precise interpretation of the Raman and IR spectra of D2S10
revealed that the spectra taken above 27 GPa in the unload-
ing process were identical to those of pure sulfur. The exis-
tence of elemental sulfur above 27 GPa has not been con-
firmed by previous x-ray experiments because all diffraction
peaks above this point became too weak to identify its struc-
ture.
The hydrogen-bond ordering of H2S is also an interesting
subject. In a Raman experiment covering pressures up to
about 20 GPa,11 phase boundaries were determined from the
spectral changes in the S-H stretching region of
2400–2600 cm−1. Obtaining a result indistinguishable from
that found from Raman spectra, a neutron diffraction experi-
ment on D S has determined the I -II boundary to be around
Ј
2
5.3 GPa and 235 K.12 The atomic positions of rotationally
disordered deuterium of phase
I
have also been
Ј
established.12 No other diffraction study below 200 K and
above 6 GPa has been reported. In order to clarify whether
the hydrogen bond in phase IV is ordered or not, molecular-
dynamics (MD) simulations have been actively pursued.13–15
If a disordered hydrogen bond remains in phase IV, a new
phase with a different bonding state due to ordering should
be found at lower temperatures.
FIG. 1. (Color online) The pressure-temperature phase diagram
of H2S. The solid curves are phase boundaries determined by a
Raman experiment.11 The dotted lines between I and II are deter-
Ј
mined by a neutron experiment.12 The triangles (run 1), the inverse
triangles (run 2) the circles (run 3), the diamonds (run 4), and the
squares (run 5) show the measured points from five independent
samples. Three dashed lines are determined in this report.
Our first purpose was to investigate the molecular disso-
ciation of H2S from a structural point of view. To this end,
0163-1829/2004/69(21)/214102(5)/$22.50
69 214102-1
©2004 The American Physical Society

HIROSHI FUJIHISA et al.
PHYSICAL REVIEW B 69, 214102 (2004)
(DAC) cooled down to 77 K, pressurized to a few GPa, and
warmed up to room temperature while holding the pressure
constant. These operations were carried out in a gaseous ni-
trogen atmosphere to avoid an H2O frost.
Five independent experimental runs were carried out at
room and low temperatures. In runs 1 and 2, for experiments
at room temperature, a sample chamber of 100 ␮m in diam-
eter and 40 ␮m in thickness was made by drilling a hole into
a rhenium foil. In runs 3 and 4, for experiments up to 20 GPa
at low temperature, a sample chamber of 130 ␮m in diam-
eter and 80 ␮m in thickness was prepared. A sample cham-
ber of 70 ␮m in diameter and 40 ␮m in thickness was set up
for run 5 which was designed for an experiment with pres-
sures up to 50 GPa at low temperature. In runs 3, 4, and 5,
respectively, a rhenium, an SUS301 (an austenite stainless
steel), and a PK (a spring steel, Fe 98%) gasket were used.
The DAC pressure was controlled by introducing helium gas
through a cryostat chamber into a membrane chamber at-
tached to the DAC itself. The temperatures were monitored
by two Si diode sensors and two thermocouples placed on
the DAC holders.
Angle-dispersive powder x-ray diffraction experiments
were performed on the BL-18C beam line at the Photon Fac-
tory of the High Energy Accelerator Research Organization
(KEK). The x-ray beams were monochromatized to a wave-
length of 0.6198 Å ͑20.00 keV͒ and the diffracted x rays
were recorded on an imaging-plate (IP) detector supplied by
Fuji Photo Film Co., Ltd. Typical exposure times were 1 h
for the room temperature experiments and 5 min for the low-
temperature experiments. The samples thus prepared gave
smooth Debye-Scherrer rings and no contamination peaks of
H2O which led to reliable structure analyses.
FIG. 2. (Color online) (a) The powder pattern of H2S obtained
at 42 GPa. (b) The powder pattern of recovered sample after un-
loading from 48 GPa to ambient. The negative slopes of the back-
grounds below 9 degrees are due to air scattering of the direct
beams.
measured the diffraction pattern for the recovered sample
using a conventional 18 kW x-ray generator with a Mo ro-
tating anode ͑0.7107 Å͒, an 80 ␮m pin-hole collimator, and
the IP for 50 h of exposure. The dots and the profile in Fig.
2(b) represent the powder diffraction pattern and the Rietveld
fit produced with the most likely candidate structure of sulfur
phase I. The solid was thus found to contain a form of sulfur
phase I, consisting of S8 rings. The pattern’s background
level, arising from the remaining amorphous component, was
still high.
The x-ray data revealed that the molecular dissociation of
the H2S started at 27 GPa at room temperature, a result in
good agreement with the results of the previous optical
measurement.9,10 After the dissociation, the sample behaved
like elemental sulfur. This may explain why the metallization
pressure of 96 GPa reported for H2S8 is very close to that of
95 GPa for sulfur.18 In further agreement with the results of
optical experiments, the x-ray results indicate that phases V
and VI are not pure phases of H2S.10 Since the results did not
depend on the chemical identity of the gasket, the dissocia-
tion is not likely to be initiated by a chemical reaction with
gasket material.
III. ROOM-TEMPERATURE RESULTS
In agreement with previous reports, the successive phase
transitions at room temperature were observed in runs 1 and
2. Figure 2(a) displays the powder diffraction pattern of
phase V at 42 GPa. The peaks, first appearing at about
27 GPa, grew continuously until a pressure of 48 GPa was
reached, though a characteristic high background level was
evident. The vertical bars in Fig. 2(a) represent the positions
and the intensities measured for the high-pressure sulfur
phase II at 42 GPa by Akahama.16 They corresponded well to
the peaks of phase V, with an exception of one peak at 13.5
degrees ͑d=2.64 Å͒. As suggested by Sakashita,10 this peak
possibly comes from an intermediate component that still has
an S-H bond. Phase I of elemental sulfur ͑␣-S8͒ undergoes
amorphization from 18–27 GPa, and on further compres-
sion, a crystalline phase II appears.17,18 These findings sug-
gest that the origin of the high background observable in Fig.
2(a) can be attributed to the existence of the amorphous
phase of sulfur.
The presence of solid sulfur was confirmed by an unload-
ing experiment. Here, the diffraction patterns showed no
change until 5 GPa. With a further decrease in pressure to
about 1 GPa, a part of the sample bubbled and leaked from
the sample chamber, but most of the sample remained solid
even at ambient conditions. To examine its structure, we
IV. LOW-TEMPERATURE RESULTS
The patterns obtained in runs 3 and 4 at 1 GPa below
100 K were different from those of phase III (Pbcm Ref. 1).
We name this phase III . A Rietveld analysis, shown in
Ј
214102-2

MOLECULAR DISSOCIATION AND TWO LOW…
PHYSICAL REVIEW B 69, 214102 (2004)
a singlet. Hence, in Fig. 1 we located a temperature insensi-
tive horizontal line at 12 GPa, shown with the dashed line.
The inset of Fig. 3(b) displays the structure without hydro-
gen atoms. The first nearest neighbor S-S distance was
3.41 Å at 3.8 GPa. As in phases IV and III , we can see
Ј
spiral chains. A low-temperature high-pressure phase, which
has been observed by Raman measurement at 25 K in a pres-
sure span of 3.3–8.0 GPa,19 is now identified as this phase
IV . The report has already mentioned that the Raman spec-
Ј
tra at 25 K drastically changed from 0 to 3 GPa. This very
likely corresponds to successive phase transitions of
III-III -IV .
Ј
Ј
In run 5, the molecular dissociation was investigated
along an isobar line at 43 GPa warming from 49 to 270 K. It
was first observed at 150 K, a temperature almost one half of
that ͑298 K͒ for the onset of dissociation at 27 GPa. The
boundary showed a negative slope of −0.11 GPa/K, as il-
lustrated by the dashed line at the top right of Fig. 1. We
additionally measured Raman spectra of this dissociated
sample at room temperature in its unloading process from 48
to 4 GPa to examine existence of hydrogen molecules. The
S-S peaks were clearly observed around 500 cm−1 as same as
the D2S study.10 In spite of an hour accumulation with a
1 mW Ar ion laser at the sample position, there were no
trace of the H-H vibration around 4200 cm−1. The separated
hydrogen atoms may creep into the sulfur lattice, the gasket,
or the diamond anvils.
FIG. 3. (Color online) Rietveld fits and their derived structures
of (a) H S phase III at 1.1 GPa and 38 K, and (b) H S phase IV
Ј
Ј
2
2
at 3.8 GPa and 27 K. Broad peaks labeled W represent diffraction
from the plastic film of the cryostat window. The solid circles
beneath the patterns show the differences between phases III
Ј
and IV .
Ј
V. DISCUSSION
Fig. 3(a), revealed that its structure belongs to an orthorhom-
bic space group Pcca made up of eight molecules in the unit
cell. The lattice parameters at 1.1 GPa and 38 K were a
=10.189±0.001 Å, b=4.040±0.001 Å, c=7.809±0.001 Å.
The sulfur atoms were located on two different sites:
Phases III , IV and V have similar intrachain and inter-
Ј Ј
chain bonds and can exist below 100 K, at which point all
molecular rotations must stop. The existence of phases IV
Ј
and IV from 25 K to room temperature means that the same
bonding state has been kept. Therefore, the hydrogen bonds
in these phases would be fully ordered, even at room tem-
perature. The contradiction in the MD,13,15 as to whether the
hydrogen bond in phase IV is ordered or disordered, has now
been clarified experimentally.
4c ͑0,y=0.469±0.004,1/4͒
and
4d ͑1/4,0,z=0.017
±0.002͒. The inset of Fig. 3(a) displays the structure without
hydrogen. The first nearest neighbor S-S distance, depicted
with sticks, was 3.67±0.03 Å. One can see spiral chains
along the a-axis, just as was found in phase IV.5 At 1.1 GPa
and room temperature, the first nearest neighbor S-S distance
of phase I is 3.9 Å; the molecules are hydrogen bonded.
Figure 4 displays the pressure dependence of the inter-
atomic distances of the sulfur atoms for phases I, II , III , IV,
Ј
Ј
and IV . Since the variation of distances with temperature
Ј
was thought to be much smaller than that with pressure, no
data for variation of temperature was included in the plot.
Each first nearest-neighbor distance of phases III , IV , and
Since the value of the S-S distance in phase III of
3.67±0.03 Å is significantly shorter than the 3.9 Å seen in
Ј
phase I, the molecules in phase III are not hydrogen bonded.
Ј
Ј
Ј
IV corresponds to the intrachain distance (S-S distance), and,
hence, can be used for discussion of the nature of the bond
A new intermediate phase was observed in the region of
4–10 GPa and 30–250 K; this region is in contact with the
boundaries of phases III , II and I , and has been considered
between the sulfur atoms. The S-S distance of phase III
Ј
Ј
Ј
(3.67 Å at 1.1 GPa) is almost same as the shortest S-S con-
tact distance of phase III [3.672 Å at 0 GPa (Ref. 1)]. They
are close to twice the sulfur van der Waals radius ͑3.60 Å͒
to be stable for a tetragonal phase IV. We name this phase
IV . By a Rietveld analysis, as shown in Fig. 3(b), its space
Ј
group was determined to be an orthorhombic Ibca including
16 molecules. The lattice parameters at 3.8 GPa and
shown with dW in Fig. 4. On the other hand, the long dis-
v
27 K were a=7.574±0.001 Å, b=7.389±0.001 Å,
c
tances in phase III [4.280 and 4.382 Å (Ref. 1)], which con-
=9.829±0.002 Å. The sulfur atoms were located on two dif-
ferent sites: 8c ͑x=0.010±0.004,0,1/4͒ and 8d ͑1/4,y
=0.206±0.003,0͒. The axial ratio a/b was 1.025 at 3.8 GPa,
decreasing with pressure to almost 1.0 at 12.0 GPa. It be-
came hard to judge whether the strongest peak of the 12 GPa
data at 54 K, 138 K, and room temperature was a doublet or
tain no hydrogen atoms, are compressed efficiently below
4.04 Å in phase III by pressure. In phase IV where the S-S
Ј
Ј
from 3.41 to 3.12 Å, the electron clouds of the sulfur atoms
must somehow overlap each other. In our previous data con-
cerning phase IV at room temperature,5 the S-S distance
shrank down to 2.96 Å at the highest pressure reached,
214102-3

HIROSHI FUJIHISA et al.
PHYSICAL REVIEW B 69, 214102 (2004)
The dissociation process has been precisely investigated
by a recent UV-visible-IR absorption measurement, and was
shown to be initiated photochemically or thermally by laser
irradiation above 24 GPa.21 The slope of −0.11 GPa/K indi-
cates that a temperature elevation of just 30 K above room
temperature decreases the dissociation pressure to 24 GPa.
Therefore we conclude that the dissociation is induced ther-
mally. It should be noted that all transitions of H2S including
the molecular dissociation did not depend on x-ray wave-
length, intensity, nor exposure time. Therefore the dissocia-
tion would not be affected by the x-ray probe itself.
Endo et al. have proposed a monoclinic Pc structure for
phase IV.22 This structure shows complicated distortions in
atomic arrangement and interatomic distances due to its very
low symmetry. The monoclinic cell with the same length of a
and c axes might be converted to an orthogonal system. The
11.4 GPa diffraction pattern they used for structural analysis
involved the strongest peak with the doublet feature. Our
pattern at 11.9 GPa at room temperature also showed the
doublet feature for the strongest peak, and the overall profile
was essentially in agreement with that of the 11.4 GPa data.
By a Rietveld analysis, the pattern was satisfactorily repro-
FIG. 4. (Color online) Interatomic sulfur-sulfur distances of H2S
under pressure. The data drawn with the triangles, the inverse tri-
angles, the circles, the diamonds, and the squares correspond to the
phases I, I , III , IV , and IV, respectively. Our previous data for
Ј
Ј
Ј
phases I, I , and IV at room temperature (Ref. 5) are included in this
Ј
plot. The open circles at 0 GPa show the phase III data at 1.5 K
(Ref. 1). The horizontal arrows represent twice the sulfur van der
Waals radius ͑ dW͒ and twice the sulfur constant energy radius
v
͑2r*͒. The vertical dashed line represents the IV -IV boundary at
Ј
12.0 GPa.
duced with the phase IV model of the orthorhombic Ibca
Ј
structure. The relationship between the Pc cell ͑am,bm,cm͒
and the Ibca cell ͑a0,b0,c0͒ was written as: a0=am+cm, c0
=cm−am, b0=2bm. Thus the previous pattern was taken at
23.3 GPa. the sulfur constant energy radius of 2.92 Å (Ref.
20) shown with 2r* in Fig. 4, which is a criterion for the
existence of overlap of the molecular orbitals or of covalent
bond formation.10,21 The present low-temperature experiment
extended significantly the pressure span for structural mea-
surement as a result of the suppression of the dissociation
reaction, and the S-S distances of phase IV were measured
up to 43 GPa. The S-S distance decreases to 2.74 Å at a
pressure of 43 GPa, below beyond the critical value of
2.92 Å. At this pressure, intrachain covalency should
strengthen. After the dissociation, this S-S interaction may
become fully covalent bonds in sulfur phase II and amor-
phous sulfur.
11.4 GPa in a stable region of phase IV and could be ex-
Ј
plained as consistent with the Ibca structure which is prefer-
able to the Pc structure.
ACKNOWLEDGMENTS
The present work was carried out as a part of the CREST
(Core Research for Evolutional Science and Technology)
project of the Japan Science and Technology Corporation
and with the approval of the Photon Factory Program Advi-
sory Committee (Proposal Nos. 98G272 and 00G202).
1 J. K. Cockcroft and A. N. Fitch, Z. Kristallogr. 193, 1 (1990).
2 H. Shimizu, Y. Nakamichi, and S. Sasaki, J. Chem. Phys. 95,
2036 (1991).
and Technology of High Pressure, edited by M. H. Manghnani,
W. J. Nellis, and M. F. Nicol (Universities Press, Hyderabad,
2000), p. 391.
3 S. Endo, N. Ichimiya, K. Koto, S. Sasaki, and H. Shimizu, Phys.
Rev. B 50, 5865 (1994).
10 M. Sakashita, H. Fujihisa, H. Yamawaki, and K. Aoki, J. Phys.
Chem. A 104, 8838 (2000).
4 H. Shimizu, H. Murashima, and S. Sasaki, J. Chem. Phys. 97,
7137 (1992).
11 H. Shimizu, H. Yamaguchi, S. Sasaki, A. Honda, S. Endo, and M.
Kobayashi, Phys. Rev. B 51, 9391 (1995).
5 H. Fujihisa, H. Yamawaki, M. Sakashita, K. Aoki, S. Sasaki, and
H. Shimizu, Phys. Rev. B 57, 2651 (1998).
12 J. S. Loveday, R. J. Nelmes, S. Klotz, J. M. Besson, and G.
Hamel, Phys. Rev. Lett. 85, 1024 (2000).
6 S. Endo, A. Honda, S. Sasaki, H. Shimizu, O. Shimomura, and T.
Kikegawa, Phys. Rev. B 54, R717 (1996).
13 R. Rousseau, M. Boero, M. Bernasconi, M. Parrinello, and K.
Terakura, Phys. Rev. Lett. 83, 2218 (1999).
7 H. Shimizu, T. Ushida, S. Sasaki, M. Sakashita, H. Yamawaki,
and K. Aoki, Phys. Rev. B 55, 5538 (1997).
14 T. Ikeda, Phys. Rev. B 64, 104103 (2001).
15 R. Rousseau, M. Boero, M. Bernasconi, M. Parrinello, and K.
Terakura, Phys. Rev. Lett. 85, 1254 (2000).
8 M. Sakashita, H. Yamawaki, H. Fujihisa, K. Aoki, S. Sasaki, and
H. Shimizu, Phys. Rev. Lett. 79, 1082 (1997).
9 M. Yamagishi, H. Furuta, S. Endo, and M. Kobayashi, in Science
16 Y. Akahama (private communication).
17 Y. Akahama, M. Kobayashi, and H. Kawamura, Phys. Rev. B 48,
214102-4

MOLECULAR DISSOCIATION AND TWO LOW…
PHYSICAL REVIEW B 69, 214102 (2004)
6862 (1993).
20 M. L. Huggins, J. Am. Chem. Soc. 75, 4126 (1953).
18 H. Luo, S. Desgreniers, Y. K. Vohra, and A. L. Ruoff, Phys. Rev.
Lett. 67, 2998 (1991).
21 T. Kume, Y. Fukaya, S. Sasaki, and H. Shimizu, Rev. Sci. In-
strum. 73-6, 2355 (2002).
19 A. Anderson, S. Demoor, and R. C. Hanson, Chem. Phys. Lett.
140, 471 (1987).
22 S. Endo, A. Honda, K. Koto, O. Shimomura, T. Kikegawa, and N.
Hamaya, Phys. Rev. B 57, 5699 (1998).
214102-5