A solid state 31P NMR study of the synthesis of phosphorus sulfides from PCI3 and H2S in microporous materials

Article abstract of DOI:10.1139/v98-147

The interaction and reaction of PCl₃ and PCl₃-H₂S mixtures with the microporous materials Silicalite, ALPO-5, NaY, NaX, and NaA have been investigated, with the intention of producing phosphorus sulfide clusters in the pores. A rich chemistry was observed and monitored by solid state ³¹P NMR. The presence of P₄S₃, α-P₄S₅, β-P₄S₆, and P₄S₇ inside the NaY α-cages was demonstrated, as well as a new species that is possibly a third geometric isomer of P₄S₄ with C(3v) symmetry. ¹²⁹Xe NMR showed the exclusion of Xe from the micropores by the phosphorus sulfides. Sulfides lower than P₄S₇ are small enough that they undergo rapid pseudo-isotropic reorientation inside the NaY α-cage.

Full text of DOI:10.1139/v98-147

1660
31
A solid state P NMR study of the synthesis of
phosphorus sulfides from PCl3 and H2S in
microporous materials1
Garry S. H. Lee, Christopher I. Ratcliffe, and John A. Ripmeester
Abstract: The interaction and reaction of PCl3 and PCl3–H2S mixtures with the microporous materials Silicalite,
ALPO-5, NaY, NaX, and NaA have been investigated, with the intention of producing phosphorus sulfide clusters in
the pores. A rich chemistry was observed and monitored by solid state 31P NMR. The presence of P4S3, α-P4S5, β-P4S6,
and P4S7 inside the NaY α-cages was demonstrated, as well as a new species that is possibly a third geometric isomer
of P4S4 with C3v symmetry. 129Xe NMR showed the exclusion of Xe from the micropores by the phosphorus sulfides.
Sulfides lower than P4S7 are small enough that they undergo rapid pseudo-isotropic reorientation inside the NaY α-cage.
Key words: phosphorus sulfides, intrazeolitic cluster synthesis.
Résumé : On a étudié l’interaction et la réaction du PCl3 et de mélanges de PCl3–H2S avec de materiaux microporeux,
comme Silicalite, ALPO-5, NaY, NaX et NaA, dans le but de produire des agrégats de sulfures phosphoreux dans les
pores. On a observé une chimie intéressante qui a été observée par RMN du 31P à l’état solide. Dans les cages-α de
NaY, on a mis en évidence la présence de P4S3, α-P4S5, β-P4S6 et P4S7 ainsi qu’une nouvelle espèce qui pourrait être
un troisième isomère géométrique du P4S4 de symétrie C3v. La RMN du 129Xe montre qu’il y a exclusion du Xe des
micropores par les sulfures phosphoreux. Les sulfures de formules inférieures à P4S7 sont suffisamment petits pour
subir une réorientation pseudoisotropique à l’intérieur de la cage-α de NaY.
Mots clés : sulfures phosphoreux, synthèse d’agrégat intrazéolitique.
[Traduit par la Rédaction] Lee et al. 1667
information on the short-range order of these potentially
useful materials. Currently there is also much interest in
The chemistry of the phosphorus–sulfur system is well
known, having been extensively studied by crystallographic
methods, NMR spectroscopy, and vibrational spectroscopy.
A large number of molecules, all based on expansions of the
P4 tetrahedron of elemental white phosphorus, is now estab-
lished and includes P4Sn (n = 3–10) (1, 2). In mixtures of
phosphorus and sulfur that contain between 0 and 25 at.% P,
glassy materials are formed (3–5). These glasses are particu-
larly attractive to the material sciences due to their optical
and electrical properties, particularly their infrared transmit-
ting properties (6).
The renewed interest in the study of phosphorus sulfides
is based on the recent developments in solid state NMR al-
lowing increased spectroscopic resolution of both ordered
crystalline compounds and disordered systems such as the
glasses (7). This permits the identification of chemical envi-
ronments via chemical shift ranges and provides valuable
nano-clusters with properties much modified from those in
the bulk material, and also for nonlinear optical and semi-
conductor properties. To this end zeolites and other
microporous solids are being used as minireactors and tem-
plates to grow size-restricted clusters and to produce periodic
arrays of clusters or nano-wires (8, 9).
With this in mind and also with a longer term interest in
studying clusters and unusual entrapped species, we at-
tempted to synthesize phosphorus–sulfur compounds within
the cavities of zeolites and other microporous materials, and
used 31P NMR to characterize the species involved. 31P NMR
is perhaps the most useful tool for studying phosphorus
species within zeolites, since many other techniques are
hampered to some extent by the nature of the system, e.g.,
Raman spectroscopy suffers from problems with fluorescence,
IR has problems with strong overlapping vibrational modes
of the lattice, X-ray diffraction has problems with disorder,
EXAFS is not so useful where numerous sites are involved.
Here we describe the results of exploratory experiments on
PCl3 in microporous materials and its reactions with H2S,
focussing particularly on the behaviour with NaY zeolite.
Received February 19, 1998.
G.S.H. Lee,2 C.I. Ratcliffe,3 and J.A. Ripmeester. Steacie
Institute for Molecular Sciences, National Research Council
of Canada, Ottawa, ON K1A 0R6, Canada.
1Published as NRCC No. 40997.
2Present address: Department of Chemistry, University of
Technology, Sydney, Australia.
3Author to whom correspondence may be addressed.
Telephone: (613) 991-1240. Fax: (613) 998-7833.
E-mail: cir@ned1.sims.nrc.ca
Silicalite (UOP), zeolites NaY (LZY-52, Linde), NaX
(13X, Linde), NaA (Linde 4A), and aluminophosphate
ALPO-5 (a gift of S. Zones, Chevron) were dehydrated
Can. J. Chem. 76: 1660–1667 (1998)
© 1998 NRC Canada

Lee et al.
1661
Fig. 1. 31P NMR of PCl3 in Silicalite. (a) Static spectrum
showing chemical shift anisotropy powder pattern for PCl3. The
signal in the vicinity of 3 ppm is probably POCl3 and other
P/O/Cl species. (b) MAS spectrum of a sample with twice the
loading of PCl3, νr = 2.0 kHz.
Fig. 2. 31P NMR MAS spectrum of PCl3 with NaY, νr =
4.5 kHz. Residual molecular PCl3 shows at 220 ppm. Stars
indicate spinning side bands of the line tentatively assigned to
framework-attached SiOPCl2.
8), which show a rich chemistry for phosphorus compounds
inside zeolites. Most reactions seem to produce several prod-
ucts and it is a major problem to identify all of the species
present. It is convenient to describe results first for the sim-
pler systems with only PCl3 and then with the addition of
H2S.
under vacuum, at progressively higher temperatures up to
450°C. All the preparative steps were carried out on a vac-
uum line and great care was taken to exclude all water. Mea-
sured quantities of degassed PCl3 (liquid) were then
condensed into these materials from a side arm and sealed in
5 mm o.d. tubes, allowing 4PCl3 per α-cage for X, Y, and A.
The reactions with H2S were carried out “in situ” after first
condensing (at liquid nitrogen temperature) a known quan-
tity of the gas onto samples of the microporous materials
with PCl3 already adsorbed (reactant ratio range used:
PCl3:H2S = 4:3–4:10). These samples, in 5 mm o.d. tubes,
were then sealed and annealed at various temperatures from
76 to 164°C. One sample was sealed after admitting a small
amount of xenon gas. Solid state NMR spectra were
recorded on a Bruker MSL 200 spectrometer operating at
80.962 MHz for 31P and 55.321 MHz for 129Xe, using a
Chemagnetics magic angle spinning (MAS) pencil probe,
both static and with spinning rates of 1.9–4.5 kHz. The glass
sample tubes were inserted into the zirconia pencil rotors
and held in place by Teflon spacer rings. Spectra were
acquired using single pulses (90° pulse length 2.8–3.0 µs)
with quadrature phase cycling. Positions of isotropic shifts
were determined, when necessary, from MAS spectra
obtained at different spinning rates. 31P spectra were refer-
enced to 85% H3PO4 and typically required 80 scans with
recycle times of up to 100 s. The 129Xe spectra were refer-
enced to Xe gas at zero pressure.
(a) PCl3 with microporous materials
PCl3 is a liquid (mp –93.6°C) with a 31P NMR chemical
shift of 219 ppm (10). When introduced into Silicalite the
spectra, Fig. 1, clearly show that the PCl3 remains intact.
However, the line indicates only a single type of PCl3, with
a small chemical shift anisotropy: δiso = 220 ± 0.4, δ11,22
=
227.8 ± 1.0, δ33 = 204.3 ± 1.0, axially symmetric within er-
ror, and there is no trace of any sharp liquid line. This aniso-
tropy can only mean that the molecule is adsorbed in the
Silicalite channels or on the surface. At the loading levels
used there is not enough surface area to accommodate all the
PCl3 and consequently one must conclude that the signal
most likely arises from molecules in the channels. Similar
behaviour was observed with ALPO-5, which has one-
dimensional channels.
With activated NaY and NaX the spectra initially show
mainly the PCl3 resonance, but with time this is very much
reduced, Fig. 2, and is replaced by rather broad MAS lines
with several spinning side bands that indicate considerable
anisotropy. The isotropic shifts, which are at about 180–
183 ppm for NaY and 153–156 ppm for NaX, can possibly
be assigned to Si-O-PCl2 and (Si-O-)2PCl species (10, 11);
the former were recently observed at 184 ppm in studies of
PCl3 reacting with a silica surface (10), and the latter sug-
gested assignment of a line at 156 ppm observed by
Bogatyrev et al. (11), though, ideally, further substantiation
of these assignments is desirable. The presence of these
chemically attached species would imply the production of
NaCl (and perhaps a little HCl by reaction with any residual
OH groups on the framework). Unfortunately the spectra are
not good enough to determine the chemical shift anisotropies.
With the exception of a short note on 129Xe, all the fol-
lowing discussion focusses on 31P NMR results (Figs. 1–6,
© 1998 NRC Canada

1662
Can. J. Chem. Vol. 76, 1998
Fig. 3. 31P NMR spectra of NaY with PCl3–H2S in a 4:3 ratio
after heating to 76°C. (a) Static; (b) MAS, νr = 3.05 kHz;
(c) 10× expansion of (b).
Fig. 5. 31P NMR spectra of NaY with PCl3–H2S in a 4:5 ratio
after heating to 100°C. (a) Static; (b) MAS, νr = 1.95 kHz;
(c) MAS, νr = 3.02 kHz, 18 days after (a, b). In (c) the spectrum
of P4S7 predominates and P4S3 lines are much reduced. Isotropic
peaks of identified species are indicated by ᭺, P4S3; ᭛, α-P4S5;
ٗ, β-P4S6. The two P4S7 lines are indicated by × = P3,P4; ᭹ =
P2,P1. Double symbols indicate the isotropic shift.
Fig. 4. 31P NMR spectra of NaY with PCl3–H2S in a 4:3 ratio
after heating to 100°C. (a) MAS, νr = 2.0 kHz; (b) MAS, νr =
2.9 kHz, 5 days after (a). Isotropic peaks of identified species
are indicated by ᭺, P4S3; ᭛, α-P4S5; ٗ, β-P4S6.
common impurity in PCl3 (less than 1%), which in the liquid
state has a shift of 3 ppm (10). However, the intensity of the
line in the NaY spectrum cannot be explained by POCl3
alone, and a more likely alternative with a shift in the same
range would be some other P/O/Cl or phosphorus oxy com-
pound produced by reaction of the PCl3 with any residual -
OH groups or defect sites in the framework.
(b) PCl3 and H2S within microporous materials
The main phosphorus species observed in the cases of
Silicalite (with a 4PCl3:3H2S ratio, annealed at 76°C) and
ALPO-5 (with a 4PCl3:5H2S ratio, annealed at 80°C) was
absorbed but unreacted PCl3. In the case of Silicalite when
annealed at 170°C, the main signal was still PCl3 but several
other small signals appeared, at 3.6 and 85 ppm (broad), and
48 and 55 ppm (sharp), which we have not been able to as-
sign with any certainty.
In the reactions combining PCl3 and H2S (in ratios of 4:3,
4:5, and 4:10) with NaY we observe that in samples heated
only to 76°C (Fig. 3), PCl3 remains prominent in the spectra,
and there is a trace of the Si-O-PCl2 species and a strong
unidentified broad signal or signals underlying the whole
These materials also show a line in the 1.8–4.6 ppm range,
which is weak in Silicalite (Fig. 1) and stronger in NaY
(Fig. 2). One possible source of tis line could be POCl3, a
© 1998 NRC Canada

Lee et al.
1663
Fig. 7. Molecular structures of P4S3, α-P4S5, β-P4S6, and P4S7
Fig. 6. 31P NMR spectra of NaY with PCl3–H2S in a 4:10 ratio
after heating to 164°C. (a) Static; (b) MAS, νr = 2.95 kHz.
Isotropic peaks of identified species are indicated by ᭺, P4S3;
᭛, α-P4S5; ٗ, β-P4S6; ᭿, new species.
(small atoms represent P).
window into the NaA large cage would prevent access even
to PCl3. The spectrum showed principally P4S7, again with
many spinning side bands, with some P4S3, represented by
two single lines, Fig. 8c. This was one of the cleanest spec-
tra with only a trace of any other species (a broad line with
several spinning side bands).
We now proceed to a detailed discussion of the evidence
for the phosphorus sulfide assignments, and whether these
molecules are located inside or outside the zeolites:
spectrum. Heating to 100°C or higher, however, initiates a
complex series of reactions, and the spectra show virtually
no PCl3 and numerous new species (Figs. 4–6), which turn
out to be phosphorus sulfides, as anticipated, Fig. 7. Assign-
ments were made by comparison with known chemical shift
ranges from numerous previous studies of phosphorus sul-
fides in solution, the melt, and the solid state, summarized in
Table 1. In the early stages a major identifiable component
is P4S3, with smaller quantities of α-P4S5 and β-P4S6. Again
there are other unidentified species, which in some of the
preparations have considerable intensity and large chemical
shift anisotropies. The 4PCl3:5H2S preparation was moni-
tored for some time after heating, and the lines of P4S3 were
observed to diminish while those of P4S7 grew, Fig. 5c. The
structures of the phosphorus sulfides identified are shown in
Fig. 7.
A preparation of 4PCl3:3H2S with NaX zeolite showed
none of the sharp phosphorus sulfide lines observed for the
NaY samples. The MAS spectra showed a very broad, in-
tense spectrum centred around 108 ppm with rather broad
spinning side bands superimposed, and several sharp lines
between 30 and –12 ppm, with the strongest at 2 ppm. The
broader features could conceivably be phosphorus sulfides
and the sharp features P/O/Cl species. The different behav-
iour of NaX compared to NaY is noteworthy, but any discus-
sion of this is necessarily speculative due to the lack of any
firm assignments in the NaX spectra. The higher cation con-
tent of NaX, however, means that there is also a higher Al
content in the framework, and as a consequence of this there
will likely be more defects where different reactions with
PCl3–H2S and their products could occur.
P4S3
P4S3 is a globular cage molecule (Fig. 7) consisting of a
triangle of three basal Pb atoms each linked via an S atom to
a single apical Pa atom. The two sharp lines, which are seen
in many of the current spectra at 82 ± 2 and –101 ± 2 ppm
in a roughly 1:3 intensity ratio, may be assigned confidently
to Pa and Pb, respectively. The low-frequency resonance is a
unique characteristic of P4S3 that distinguishes it from the
other phosphorus sulfides. There have been numerous previous
studies of P4S3 in solution (12) and the solid state(13–18). In
various solvents the Pa shift is in the range 62–79 ppm and
the Pb shift is from –100 to –135 ppm (12), whereas the
most recent studies of the solid give shifts of 91.0/84.5 and
–87.5 ppm for the room temperature β-phase (13). (Note, in
the solid, Pa appears as two lines due to crystallographically
inequivalent molecules.) Clearly there is considerable varia-
tion depending on the environment, and our results fall well
in the middle of the respective shift ranges.
What is most striking about the P4S3 resonances seen in
the spectra with NaY zeolite is that the lines are isotropic,
since they display no spinning side bands under MAS and
remain quite sharp and featureless when static. On its own
this isotropic behaviour might be taken as evidence for P4S3
being inside the α-cages: Within the Y zeolite α-cage a sin-
gle P4S3 molecule would have considerable rotational free-
dom, and since the cage has elements of Td symmetry the
dynamic averaging of the chemical shift tensor (arising from
pseudo-isotropic reorientation of the molecule) would pro-
duce an isotropic line. This is in marked contrast to the room
temperature crystalline β-phase, which shows rather large
chemical shift anisotropies of several hundred ppm (13).
A reaction with NaA with a 4PCl3:7H2S ratio, annealed at
80°C, was carried out with the notion that the much smaller
© 1998 NRC Canada

1664
Can. J. Chem. Vol. 76, 1998
Table 1. Ranges of reported 31P NMR isotropic shifts for phosphorus sulfides compared with observed lines for the same
in NaY.a
Equivalent P
atoms b
Shift ranges
(ppm)
Shifts in NaY
(ppm) c
Compound
P4S3
Refs.
1
3
4
2
1
1
62–91
(–87)–(–135)
201.9
205–207.5
174.7–185.8
85.7–97
81.9 ± 1
–101 ± 2
12–18
α-P4S4
β-P4S4
26
2, 26
α-P4S5
1
1,1
1
2
2
1
1
1,1
1
2
1
2
2
2
2
1
3
4
231.7–237.4
111.4–139.8
89.6–120.2
129.8–142.3
80.1–87.2
132.3–133.5
114.9–115.8
54.9–65.7
198.9–206.7
178.2–184.8
55.7–61.0
109.5–113
81.9–98
229.6 ± 1
123.8 ± 6
98 ± 3
2, 13, 22
β-P4S5
α-P4S6
2, 27
2
β-P4S6
200 ± 1
178.2 ± 1
59.6 ± 1
110.4 ± 0.5
97.7 ± 0.5
2
P4S7
P4S8
P4S9
P4S10
2, 13–15, 22, 23
28
135.4
13.4
57.9–67.5
57.8–63.6
49.7–60.1
13, 15, 25
13, 15 25
aShift ranges from reported values in solution, melt, and solid phases. α’s and β’s in this instance refer to different molecular structural
isomers (not different crystal phases).
bDistinct P atoms in isolated molecule.
cThis work.
Similar anisotropies were also obtained from studies of P4S3
dissolved in a nematic phase (19). However, there is a sec-
ond plausible explanation: Crystalline P4S3 is known to have
a plastic phase (16, 18, 20, 21), where the molecule sits on
specific lattice sites but reorients rapidly and thus gives rise
to an isotropic NMR spectrum. Although the room tempera-
ture β-phase transforms to the plastic α-phase at 313.7 K, the
α-phase can be supercooled to a lower limit of 259 K (20).
However, it does not always have to be cooled this far down
to transform back to the β-phase; for example in a T1 study
the material supercooled by about 35 degrees (21), though in
a MAS study the supercooled state was only observed for a
few minutes before it reconverted to the β-phase at ambient
temperatures (18). Since the preparation of our zeolite / phos-
phorus sulfides all involved heating to 100°C or more, it is
possible that the isotropic lines are from α-crystalline P4S3
outside the zeolite. Furthermore, resonances for the α-phase
have been reported at 72 and –113 ppm (16) and, more re-
cently, at 80.1 and –104.0 ppm (18). The latter result is very
close to our results for NaY. To distinguish between these
two possibilities, two of the NaY reaction samples
(4PCl3:5H2S and 4PCl3:10H2S) and the NaA reaction sam-
ple were cooled to 77 K for 10 min and then allowed to
warm back to room temperature; the isotropic lines were
unchanged, thus demonstrating unequivocally the presence
of P4S3 inside these zeolites.
α-P4S5 and β-P4S6
The static NMR spectra also show series of weak, sharp
lines with no chemical shift anisotropy, which we assign to
α-P4S5 and β-P4S6 in the zeolite α-cage, see Figs. 4–6 and
Table 1. The identification of three lines for β-P4S6 is rea-
sonably certain; the lines are in the correct shift regions (Ta-
ble 1) and in the correct intensity ratio of 1:2:1 (2). The
assignment of α-P4S5 must rely mainly on the unique high-
frequency shift at 230 ppm (2, 13, 22), as the other expected
resonances are partly obscured by other lines. Plastic (rota-
tor) phases for α-P4S5 and β-P4S6 have not been reported and
indeed, since both have a pendant S atom that makes these
molecules much less globular than P4S3 (see Fig. 7), one
would not expect them to have plastic phases. Consequently,
the argument for these two products being inside the α-cages
and undergoing rapid pseudo-isotropic reorientation is
strong.
P4S7
The major product in the preparation of both NaY with
4PCl3:5H2S, Fig. 8a, and NaA with 4PCl3:7H2S, Fig. 8c,
was P4S7, but this product is in two quite different states.
The experiment with NaA produced crystalline α-P4S7,
which has four resolved peaks at 82, 96, 111.5, and 113 ppm
(13, 23), whereas the spectrum of NaY/P4S7 is definitely not
that of the crystalline α-form and the isotropic shifts at 110.4
© 1998 NRC Canada

Lee et al.
1665
Fig. 8. 31P NMR MAS spectrum of P4S7. (a) NaY/PCl3/H2S
preparation, νr = 3.02 kHz; (b) α-crystalline preparation (no
zeolite), νr = 4.09 kHz; (c) NaA/PCl3/H2S preparation, νr =
4.03 kHz. (a) and (c) both show isotropic P4S3 lines and traces
of broader underlying resonances. (Note that spectrum (a) is the
same as in Fig. 5c). × = P3,P4; ᭹ = P2; ᭜ = P1. In (a) P1, P2
overlap. Double symbols indicate the isotropic shift.
Table 2. 31P NMR isotropic shift and principal components of
the apparent a chemical shift tensors obtained from MAS spectra
for P4S7 in NaY compared with crystalline α-P4S7.
Atom
δiso
δ11
δ22
δ33
b
P3,4
109.4
97.8
113
111.6
96
254
158
271
275
187
170
172
163
109
92
124
131
85
78
84
69
–36
44
–55
–72
20
b
P1,2
c
P4
P3
P2
c
c
36
–6
10
c
P1
82
aThe reader should note that these are not true chemical shift tensors,
since direct and indirect coupling tensors are also involved (see ref. 13).
In particular, the dipolar coupling between the directly bonded P atoms
will be substantial.
bData for P4S7 in NaY (this work).
cData for crystalline α-P4S7 (13). The double entries for P2 and P1
correspond to J-split components.
atom of another molecule, which makes all four P atoms
inequivalent. This well-defined intermolecular interaction
does not occur in the amorphous solid, which shows only
two resonances, at shifts similar to those in the melt and so-
lution. Likewise for a molecule isolated in the α-cage of
NaY, only two resonances would be expected.
Once again we are left with a choice between two equally
plausible assignments for the NaY preparation: the amor-
phous solid outside the zeolite or isolated molecules inside
the cages. Moreover, if the P4S7 molecule is inside the
cages, this time it cannot be undergoing rapid reorientations
because the chemical shift anisotropy is not averaged. The
principal components of the chemical shift anisotropies for
both resonances of NaY–P4S7 were determined by spinning
side-band analysis at various spinning speeds using the
Herzfeld–Berger method (24), which is incorporated in the
Bruker software, see Table 2. From the table we note that
our results, which clearly would not be expected to be the
same as for crystalline α-P4S7 determined previously (13),
nevertheless show anisotropies of similar span and skew.
This rules out any question of dynamic averaging.
There are two main arguments that favour the existence of
single molecules of P4S7 inside the cages of NaY: (a) There
is the circumstantial evidence of the precursors being inside
the cages, and several of the smaller sulfides also being
formed there. (b) The amorphous form of P4S7 has only been
formed previously by quenching the melt rapidly to 77 K
(23), conditions quite remote from our preparation where the
PCl3 was sublimed into the zeolite and H2S gas was added
separately before annealing at 100°C.
and 97.7 ppm are very similar to those observed for the
amorphous form (measured by Bjorholm at 109.5 and
98 ppm (23)). In both spectra the lines show sizeable
anisotropies. The molecular point group symmetry of the
free P4S7 molecule is C2v. There are two distinct pairs of
equivalent P atoms and hence one expects two 31P NMR
resonances for the C2v molecule (aside from any J cou-
plings). In the solid α-phase of P4S7, however, the molecule
is distorted because of a moderately strong and asymmetric
interaction between two P atoms on one molecule and an S
P4S9 and P4S10
Since P4S7 appears to be static it is highly unlikely that
we would see dynamic P4S9 and P4S10, which enables us to
exclude making any assignment of sharp lines in the 50–
60 ppm region to these species. Although their isotropic
shifts fall in this region the static molecules have large
chemical shift anisotropies (13, 15, 25).
© 1998 NRC Canada

1666
Can. J. Chem. Vol. 76, 1998
NaY: The diameter of the Y zeolite α-cage, without taking
into account Na+ cations, is 11.0 Å and the diameter of the
12-ring window is 7.4 Å (30, 31).
NaA: The diameter of the A zeolite α-cage is 11.4 Å, and
the diameter of the 8-ring window is only 4.1 Å (31, 32), ex-
cluding the Na+ ions.
P4S3: The dimensions of P4S3 show a maximum diameter
of 7.5 Å, and the largest diameter perpendicular to the 3-fold
axis is 7.24 Å (calculated from available structural data (20,
33)). Inside the large cages of both NaY and NaA it will
have considerable room. This molecule should therefore be
able to go easily through the 12-ring window of NaY, but it
should not be able to get through the 8-ring window of NaA
zeolite.
A new P4S4 species in NaY ?
The 4PCl3:10H2S NaY preparation heated to 163°C shows
three sharp lines in the static spectrum, at –126.3, 49, and
64 ppm, which do not appear for any of the other samples.
Furthermore, when this material was reexamined after con-
siderable aging, the P4S3 lines and the 49 ppm line had di-
minished considerably, and the lines due to α-P4S5 and β-
P4S6 had disappeared completely, leaving a broad back-
ground centred around zero shift and more prominent lines
at –126.3 and 64 ppm. The latter two lines occur in a
roughly 3:1 ratio and appear to be associated with one spe-
cies. The –126.3 ppm shift is very unusual, and does not ap-
pear to have been observed in the solid state before. The
unique –101 ppm shift seen for P4S3 arises from its ring of
three P atoms, so we suspect that the –126.3 ppm line may
also be due to a P3 ring, perhaps of a P4S3 derivative with a
pendant S on the unique P atom. In this case the shift of
64 ppm for the unique P is in the region expected for S3PϭS
units. Such a P4S4 molecule would be a new geometric iso-
mer with C3V symmetry, and it would still have room to un-
dergo pseudo-isotropic reorientation inside the α-cage. Why
this new isomer would form in preference to the α or β
forms is debatable, but it is likely influenced by the zeolite
framework. A P4S3 molecule protonated on the unique P,
P4S3H+, is another possibility, though this has not been re-
ported before and the S3PH+ unit would not be expected to
have a 64 ppm shift. A chloro derivative seems much less
likely since H2S is in excess, and in fact it can be excluded,
since we have observed the same two resonances in a prepa-
ration involving elemental P with H2S in NaY annealed at
420°C (work to be reported later). One final possibility is
that these resonances are from P4S3 in a second distinct kind
of environment, though it is difficult to envisage where this
might be.
α-P4S5: This is slightly smaller than β-P4S6 with a largest
dimension of 9.3 Å (calculated from the crystal structure
(34)).
β-P4S6: This is basically a P4S7 with one pendant S re-
moved, so we have calculated its size based on the structural
information for P4S7 (see below). Its largest dimension is 9.4
Å, and from the NMR it would appear that this is still small
enough to allow rapid reorientation in the NaY cage.
P4S7: The smallest dimension of P4S7 (calculated from the
crystal structure information (34–36)) is 7.7 Å and the larg-
est is 10.2 Å.
One molecule of P4S7 can therefore fit inside the cage
without too much trouble, but it would be difficult to get
through the window. This has several implications: (a) The
particular interactions that distort the molecules in the α-
phase should not be present. (b) If it is inside the cage the
molecule must have formed in situ from smaller precursors.
(c) Once formed inside, the molecules will remain trapped.
When the presence of Na+ ions is also considered together
with the long dimension of the molecule it is highly likely
that the P4S7 will experience considerable hindrance to rota-
tion, and hence the chemical shift anisotropy will not display
the effects of dynamic averaging.
129Xe NMR results
The characteristic chemical shift of 129Xe in pure NaY
samples arises from exchange of the Xe between the zeolite
cavities and the gas phase. The observed shift is a popula-
tion weighted average of the shifts in the two types of situa-
tion (a high-frequency shift inside, and a low-frequency shift
outside). The shifts for similar amounts of Xe in pure NaY
and in an NaY with 4PCl3:3H2S reaction sample were 87.6
and 16.1 ppm, respectively. The much smaller shift of the
Xe in the reaction sample indicates that the Xe is effectively
excluded from the zeolite, implying that the cages are occu-
pied by the phosphorus compounds. The 16.1 ppm shift,
compared to that of the isolated Xe atom, indicates that the
Xe is under significant pressure, the shift being caused by
collisions with Xe atoms, the HCl reaction product, and any
residual H2S.
P4S9 and P4S10: Isolated P4S10 is a tetrahedral molecule.
The largest dimension (between two pendant S atoms) is
10.24 Å, but if the molecule were rotating pseudo-
isotropically its effective diameter would be 11.71 Å (calcu-
lated from the structural data (34, 35)). P4S9 is structurally
similar to P4S10 but with one pendant S removed, hence its
largest dimensions are virtually identical.
PCl3: PCl3 is a pyramidal molecule with dimensions
(a) along a P—Cl bond of 5.74 Å, (b) along a Cl—-Cl edge
of 6.73 Å, and (c) for the pyramid height of 4.65 Å (calcu-
lated from microwave structural data (37)). It should there-
fore have easy access to the NaY α-cage, and a tetrahedral
cluster of four PCl3 will just fit inside the cage. PCl3 should
not, however, be able to get through the window of the α-
cage of NaA.
Other components of the spectra
Molecular dimensions
We will emphasize again that the picture is incomplete as
there are signals that we have been unable to assign, in par-
ticular, the broad background of signals with large chemical
shift anisotropies. The MAS lines of these signals are them-
selves broad, whereas most crystalline materials would be
expected to show sharp MAS lines. The broadening could be
due to intermediate rate dynamics but a much more likely
An important consideration regarding the encapsulation of
the phosphorus sulfide molecules is whether they can fit in-
side the zeolite cage and, if so, whether they have room to
rotate. All the dimensions in the following discussion take
into account the van der Waals radii of the atoms (O = 1.4, P
= 1.9, S = 1.85, and Cl = 1.8 Å (29)).
© 1998 NRC Canada

Lee et al.
1667
explanation is that it is due to a distribution of species and
(or) a distribution of environments for the same species.
We may be seeing impure and mixed P–S species outside
the zeolite, and any species attached to the framework
(inside or on the external surface) would also give broad
lines with anisotropies.
1. D.E.C. Corbridge. Phosphorus. An outline of its chemistry,
biochemistry and technology. 4th ed. Studies in inorganic
chemistry 10, Elsevier, New York. 1990.
2. T. Bjorholm and H.J. Jakobsen. J. Am. Chem. Soc. 113, 27 (1991).
3. F. Heyder and D. Linke. Z. Chem. 13, 480 (1973).
4. Y. Monteil and H. Vincent. Z. Anorg. Allg. Chem. 428, 259 (1977).
5. R. Blachnik and A. Hoppe. J. Non-Cryst. Solids, 34, 191 (1979).
6. P.C. Taylor. MRS Bull. 12, 36 (1987).
7. M. Tullius, D. Lathrop, and H. Eckert. J. Phys. Chem. 94,
2145 (1990).
8. G.D. Stucky and J.E. MacDougall. Science, 247, 669 (1990).
9. G.A. Ozin, A. Kuperman, and A. Stein. Angew. Chem. Int. Ed.
Engl., 28, 359 (1989).
10. B.A. Morrow, S.J. Lang, and I.D. Gay. Langmuir, 10, 756 (1994).
11. V.M. Bogatyrev, V.V. Brei, and A.A. Chuiko. Theor. Expt.
Chem. (Engl. Transl.), 24, 603 (1988).
12. G. Heckmann and E. Fluck. Z. Naturforsch. 26B, 982 (1971).
13. R.K. Harris, P.J. Wilkes, P.T. Wood, and J.D. Woollins. J.
Chem. Soc. Dalton Trans. 809 (1989).
14. C.Brevard and M.Demarcq. Chem. Phys. Lett. 82, 167 (1981).
15. H. Eckert, C.S. Liang, and G.D. Stucky. J. Phys. Chem. 93,
452 (1989).
16. E.R. Andrew, W.S. Hinshaw, and A.Jasinski. Chem. Phys.
Lett. 24, 399 (1974).
17. M.G. Gibby, A. Pines, W.-K. Rhim, and J.S. Waugh. J. Chem.
Phys. 56, 991 (1972).
18. T. Bjorholm and H.J. Jakobsen. J. Magn. Reson. 84, 204 (1989).
19. N. Zumbulyadis and B.P. Dailey. Chem. Phys. Lett. 26, 273 (1974).
20. T. Chattopadhyay, E. Gmelin, and H.G. Von Schnering. J.
Phys. Chem. Solids, 43, 925 (1982).
21. E.R. Andrew, W.S. Hinshaw, M.G. Hutchins, and A. Jasinski.
Chem. Phys. Lett. 27, 96 (1974).
22. R. Thamm, G. Heckmann, and E. Fluck. Phosphorus Sulfur,
11, 273 (1981).
23. T. Bjorholm. Chem. Phys. Lett. 143, 259 (1988).
24. J. Herzfeld and A.E. Berger. J. Chem. Phys. 73, 6021 (1980).
25. E.R. Andrew, W. Vennart, G. Bonnard, R.M. Croiset, M.C.
Demarcq, and E. Mathieu. Chem. Phys. Lett. 43, 317 (1976).
26. A.M. Griffin, P.C. Minshall, and G.M. Sheldrick. J. Chem.
Soc. Chem. Commun. 809 (1976).
27. A.M. Griffin and G.M. Sheldrick. Acta Crystallogr. Sect. B:
Struct. Crystallogr. Cryst. Chem. B31, 2738 (1975).
28. J.-J. Barieux and M.C. Demarcq. J. Chem. Soc. Chem.
Commun. 176 (1982).
Chemistry
The chemical complexity of these systems must now be
quite apparent, so beyond a few obvious points our com-
ments regarding reactions must necessarily be speculative:
1. A basic assumption is that any -SH will react with P–
Cl bonds, provided the two are able to access each other, so
if PCl3 is in excess all the -SH will be depleted and there
may be P/S/Cl species present, and, vice versa, in reactions
where H2S is in excess, any PCl species will disappear and
there may be P/S/H species produced. The HCl by-product
could itself influence further reactions.
2. We observed P4S3 at the early stages after heating to
100°C or more, and that this is depleted as P4S7 forms, sug-
gesting a building of S onto the P4S3 framework.
3. It was surprising to find P4S3 inside NaA (demon-
strated by the persistence of isotropic lines after cooling to
77 K). Since P4S3 cannot get through the window of NaA it
must be constructed inside the cage, and yet PCl3 is also too
large to get through the window. This suggests some smaller
precursor must first form at the zeolite surface and migrate
+
into the cages, say, perhaps PCl2 .
4. An interesting feature of these experiments is that the
zeolites seem to have a catalytic effect, enhancing the for-
mation of the phosphorus sulfides. The thiosolvolysis reac-
tion of PCl3 in liquid H2S is known to produce some P4S7 as
part of a mixture after 60–70 h (38), while PCl3 and H2S in
the presence of base react to give P4S3, among other prod-
ucts (1, 39). We found, however, that a yellowish white solid
produced after 4 days of reaction between PCl3 and H2S in
the absence of zeolite still showed mainly PCl3 signals in the
31P NMR (over 90% of the “observable” phosphorus), and
no P4S7 was detected in this solid even after 3 months. The
ionic nature of NaY and NaA and their microporosity proba-
bly play important roles in orchestrating the chemistry.
29. F.A. Cotton and G. Wilkinson. Advanced inorganic chemistry.
3rd ed. Interscience, New York. 1972. p. 120.
30. G.R. Eulenberger, D.P. Shoemaker, and J.G. Keil. J. Phys.
Chem. 71, 1812 (1967).
We have clear evidence for the formation of phosphorus
sulfides inside the micropores; however, the chemistry is
rather complicated. Moreover, with the numerous different
products present, the goal of producing a regular array of
identical molecules using the zeolite framework has not
been realized. However, the results point the way: Perhaps
this might be achievable by starting with pure P4S3 instead
of PCl3.
31. W.M. Meier and D.H. Olson. Atlas of zeolite structure types.
3rd ed. Butterworth–Heinemann, London. 1992.
32. J.J. Pluth and J.V.Smith. J. Am. Chem. Soc. 102, 4704 (1980).
33. Y.C. Leung, J. Wazer, S. Van Houten, A. Vos, G.A. Wiegers,
and E.H. Wiebenga. Acta Crystallogr. 10, 574 (1957).
34. A. Vos, R. Olthof, F. van Bolhuis, and R. Botterweg. Acta
Crystallogr. 19, 864 (1965).
35. A. Vos and E.H. Wiebenga. Acta Crystallogr. 8, 217 (1955).
36. D.T. Dixon, F.W.B. Einstein, and B.R. Penfold. Acta
Crystallogr. 18, 221 (1965).
37. P. Kisliuk and C.H. Townes. J. Chem. Phys. 18, 1109 (1950).
38. A.R. Pitochelli and L.F. Audrieth. J. Am. Chem. Soc. 81, 4458 (1959).
39. E. Fluck and B. Neumuller. Z. Anorg. Allg. Chem. 534, 27
(1986).
The authors would like to thank J. Bennett for expert
technical assistance with the NMR instrumentation, and S.
Zones for the kind donation of the ALPO-5.
© 1998 NRC Canada