A dynamic NMR study of sulphur inversion in transition metal complexes of 2,6-dithiaspiro[3.3]heptane (2,6-DTSH), 2,6-DTSH-2-oxide and 2,6-DTSH-2,2-dioxide

Article abstract of DOI:10.1016/S0277-5387(98)00445-8

The following complexes were synthesised [W(CO)₅L¹], [{W(CO)₅}₂L¹], [W(CO)₅L²], [W(CO)₅L³], [{PdCl₂(PPh₃)}₂L¹], [PdCl₂(PPh₃)L²] and [PdCl₂(PPh₃)L³] where L¹=2,6-dithiaspiro[3.3]heptane (2,6-DTSH), L²=2,6-DTSH-2-oxide and L³=2,6-DTSH-2,2′-dioxide. In solution these complexes exhibit pyramidal inversion of the metal-coordinated sulphur atom(s), rates and activation energies of which were evaluated by total NMR bandshape analysis. ΔG^? values were in the range 45-50 kJ mol^-1. These magnitudes are discussed in terms of ring geometry distortions and sulphur lone pair interactions.

Full text of DOI:10.1016/S0277-5387(98)00445-8

Polyhedron 18 (1999) 1345–1353
A dynamic NMR study of sulphur inversion in transition metal complexes
of 2,6-dithiaspiro[3.3]heptane (2,6-DTSH), 2,6-DTSH-2-oxide and
2,6-DTSH-2,2-dioxide
ˇ
*
Edward W. Abel, Keith G. Orrell , Mark C. Poole, Vladimir Sik
Department of Chemistry, The University of Exeter, Exeter EX4 4QD, UK
Received 9 November 1998; accepted 11 December 1998
Abstract
The following complexes were synthesised [W(CO)5L1], [hW(CO)5j2L1], [W(CO)5L2], [W(CO)5L3], [hPdCl2(PPh3)j2L1],
[PdCl2(PPh3)L2] and [PdCl2(PPh3)L3] where L152,6-dithiaspiro[3.3]heptane (2,6-DTSH), L252,6-DTSH-2-oxide and L352,6-DTSH-
2,29-dioxide. In solution these complexes exhibit pyramidal inversion of the metal-coordinated sulphur atom(s), rates and activation
energies of which were evaluated by total NMR bandshape analysis. DG‡ values were in the range 45–50 kJ mol21. These magnitudes
are discussed in terms of ring geometry distortions and sulphur lone pair interactions.
1999 Elsevier Science Ltd. All rights reserved.
Keywords: Tungsten; Palladium; 2,6-Dithiaspiro[3.3]heptane; Sulphur inversion; Dynamic NMR
1. Introduction
by dynamic NMR (DNMR) techniques. We have therefore
undertaken an NMR study of S inversion in a variety of
transition metal complexes of 2,6-DTSH (L1) and its
2-oxide (L2) and 2,2-dioxide (L3) derivatives (Fig. 1). The
ligand 2,6-DTSH formed both mono- and di-nuclear
complexes with the metal moieties W(CO)5 and
PdCl2(PPh3), namely [W(CO)5L1], [hW(CO)5j2L1],
[PdCl2(PPh3)L1] and [hPdCl2(PPh3)j2L1], whereas its
mono-oxide and di-oxide derivatives (L2 and L3) formed
only mononuclear complexes. Analogous complexes of
Cr(CO)5 and Mo(CO)5 were also prepared but were not
fully characterised on account of their unstable nature.
2,6-Dithiaspiroheptanes have been used [1] as ligands
linking Ru(II) and Ru(III) centres to investigate the
phenomenon of electron tunnelling, particularly in bio-
logical systems [2]. The use of substitutionally inert Ru(II)
and Ru(III) as donor and acceptor centres for this process
followed the original work of Taube [3]. These spiro
compounds were chosen on account of their rigidity so that
electron transport would have to occur via the bridge and
not directly from one Ru atom to another as might occur in
less rigid complexes [4]. An intervalence absorption band
at 910 nm provided evidence of electron tunnelling over a
˚
distance of 11.3 A for the two-ring spiro bridge compound
2,6-dithiaspiro[3.3]heptane (2,6-DTSH) [1].
2. Experimental
Metal-coordinated sulphur atoms undergo pyramidal
inversion [5] and it was thought possible that, as a
consequence of electron–electron interaction between the
sulphur lone pairs, the kinetics of inversion of one metal-
bound S atom might influence the kinetics of a second
such atom when the spacer ligand is 2,6-DTSH. We have
shown previously [5] that pyramidal inversion of S atoms
is greatly facilitated by coordination to transition metals as
a result of appreciable (p–d) p stabilization of the
inversion transition state, and can be monitored accurately
2.1. Synthesis of ligands
Extensive improvements were made to the earlier litera-
ture syntheses [6,7] of 2,6-dithiaspiro[3.3]heptane (2,6-
DTSH), 2,6-DTSH-2-oxide and 2,6-DTSH-2,2-dioxide,
details of which are given below.
2.2. 2,6-DTSH (L1)
Pentaerythrityl tetrabromide (20.06 g, 51.7 mmol) was
added to an aqueous solution (100 cm3) of sodium
*
Corresponding author.
0277-5387/99/$ – see front matter
PII: S0277-5387(98)00445-8
1999 Elsevier Science Ltd. All rights reserved.

1346
E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
Fig. 1. The 2,6-dithiaspiro[3.3]heptane ligands, L1, L2 and L3, with their atom numbering and symmetry point groups.
sulphide nonahydrate (50.45 g, 210.2 mmol). The reaction
mixture was stirred vigorously at just below reflux tem-
perature for 1.5 h precisely in the presence of ‘Adogen’
phase transfer catalyst (ca. 1 g). On cooling, the aqueous
phase was decanted off and the product mixture taken up
in CH2Cl2. After washing with water (4350 cm3), vac-
uum distillation gave the title compound (4.81 g, 36.4
mmol, 70% yield, compared to 51% by reaction in aqueous
ethanol [6] or 68% by a less convenient method [7]) as a
pungent, colourless, viscous liquid which crystallised on
standing (mp5318C). NMR: (1H) d53.28 (s, 8H, –CH2 –),
(13C) d539.8 (C1,3,5,7), 52.2 (C4).
powder (mp581.7–81.98C. Lit. 818C [6]). NMR: (1H)
d53.96 (AA9BB9,2H); 3.24 (s, 2H); 3.20 (s, 2H); 3.15
(AA9BB9, 2H). (13C) d538.4 (C7); 38.6 (C5); 39.5 (C4);
64.7 (C1,3). IR: nSO51053 cm21 (CH2Cl2 solution).
Elution with ethyl acetate–methanol [10:1] and solvent
removal under reduced pressure gave 2,6-dithias-
piro[3.3]heptane-2,6-dioxide (0.88 g, 5.9 mmol, 10%) in a
similar physical form to the title compound (mp5141.3–
141.88C. Lit. 1378C [6]). NMR: (1H) d53.28 (m, 4H);
3.71 (m, 4H); (13C) d526.8, 62.8, 64.2. IR: nSO51048
cm21 (CH2Cl2 solution).
Both compounds were obtained as colourless, needle-
like crystals on recrystallisation from hot benzene.
2.3. 2,6-DTSH-2-oxide (L2)
A 30% hydrogen peroxide solution (6.73 g, 59.4 mmol)
was added, dropwise to a vigorously stirred solution of
2,6-dithiaspiro[3.3]heptane (7.92 g, 58.9 mmol) in
ethanoic acid (400 cm3). After stirring for 30 h the solvent
was removed under reduced pressure and the residue
washed free of unoxidised disulphide with hot petroleum
spirit (5350 cm3). The two oxidised species formed (Rf5
0.01 and 0.4 with ethyl acetate–methanol [10:1]) were
separated by chromatography. Elution with ethyl acetate–
methanol [50:1] and solvent removal under reduced pres-
sure gave the title compound (4.00 g, 27.0 mmol, 46% –
without recovery of unoxidised disulphide) as a fine white
2.4. 2,6-DTSH-2,2-dioxide (L3)
To 2,6-dithiaspiro[3.3]heptane (7.96 g, 60.0 mmol) in
acetone (150 cm3) was added potassium permanganate
(21.5 g, 136.1 mmol) in small portions over 8 h. After 50 h
stirring, the reaction mixture was filtered and the solvent
removed under reduced pressure. The residue was partially
dissolved in dichloromethane and the insoluble disulphone
filtered off. On again removing solvent under reduced
pressure the residue was washed free of unoxidised
disulphide with hot petroleum spirit (6350 cm3). Re-
crystallisation from boiling ethanol gave the title com-

E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
1347
pound (4.22 g, 25.7 mmol, 43% – with recovery of
2.5.3. PdCl2(PPh3 ) complexes of DTSH, DTSH-2-oxide
and DTSH-2,29-dioxide
unoxidised disulphide) as colourless, needle-like crystals
(mp5116.2–116.78C. Lit. 116.58C [6]). NMR: (1H) d5
3.45 (s, 4H, –CH2SCH2 –); 4.22 (s, 4H, –CH2SO2CH2 –);
(13C) d533.8 (C4); 38.0 (C5,7); 76.0 (C1,3) [9]. IR:
nSO asym.51325 cm21 (vs), nSO sym.51070 cm21 (s).
These complexes were prepared from the appropriate
molar quantities of [PdCl2(PPh3)]2 and the ligands. Re-
actions proceeded rapidly at ambient temperature in chlori-
nated solvents and could be monitored visually. Crystalline
products (orange, needle-like crystals) were obtained by
slow evaporation of homogeneous dichloromethane–hex-
ane solutions.
2
2
2.5. Synthesis of complexes
2.5.1. Tungsten pentacarbonyl complexes of 2,6-DTSH,
2,6-DTSH-2-oxide and 2,6-DTSH-2,2-dioxide
2.6. Physical methods
[W(CO)5(2,6-DTSH-2-oxide)] and [W(CO)5(2,6-DTSH-
2,2-dioxide)] were both prepared by stirring a 5–10%
molar excess of [W(CO)5(THF)] with the ligand in THF
solvent at 2308C for 10 h. On solvent removal under
reduced pressure, the residue was redissolved in dichloro-
methane, filtered and then recrystallised several times from
dichloromethane–hexane [1:1] at 2208C to give the
desired complexes as glassy, bright yellow crystals.
Although [hW(CO)5j2(2,6-DTSH)] was prepared in
good yield by interaction of a several-fold excess of
[W(CO)5(THF)] with 2,6-DTSH, an analytically and spec-
troscopically pure sample of the complex was easier to
obtain as a by-product in the synthesis of [W(CO)5(2,6-
DTSH)] as follows:
To 2,6-DTSH (2.47 g, 18.68 mmol) in THF at 2788C
was added, dropwise, a THF solution (100 cm3) of
[W(CO)5(THF)] hgenerated from 2.98 g, 7.47 mmol of
[W(CO)6]j at the same temperature. After stirring for a
further 8 h, with an accompanying deep orange to pale
yellow colour change, the solvent was removed under
reduced pressure to give an oily yellow residue. This was
washed with warm hexane (5380 cm3), the collected
washings being filtered and left at 2208C for 24 h.
Recrystallisation of the residue remaining from dichloro-
methane–hexane [1:3] at 2208C gave yellow crystals of
[hW(CO)5j2(2,6-DTSH)] (0.229 g, 0.294 mmol, 3.5%).
The granular yellow solid, obtained from cooling the warm
hexane washings, was washed with cold hexane (2208C,
5350 cm3) to remove free ligand and then recrystallised
again from hexane (80 cm3) to give yellow microcrystals
of [W(CO)5(2,6-DTSH)] (2.05 g, 4.49 mmol, 53%).
Elemental analyses were carried out by Butterworth
Laboratories Ltd., Teddington, Middlesex, London. Un-
corrected melting temperatures were recorded on a digital
Gallenkamp apparatus. IR spectra were recorded in either
dichloromethane or n-hexane solutions on a Perkin-Elmer
881 spectrophotometer. 1H NMR spectra were recorded at
250 MHz on a Bruker AM250 FT spectrometer. A
standard B-VT1000 variable temperature unit was used to
control the probe temperature, which was calibrated using
a Comark digital thermometer. Spectra were recorded as
solutions in either CDCl3 or CD2Cl2 solvents. NMR
bandshape analyses were performed using the authors’
version of the DNMR3 program [8]. Spectra were recorded
over as wide a temperature range as possible and fittings
made for at least six temperatures. Systematic errors in the
calculation of activation energy data were minimised by
allowing as far as possible for the temperature dependen-
cies of the chemical shifts and natural line widths, the
21
*
*
latter expressed as (pT 2 ) values, T 2 values being based
on measurements at the lowest temperatures achieved. The
treatment of errors followed that of Binsch and Kessler [9].
3. Results
Analytical, IR and melting point data for the complexes
I–VII are given in Table 1. Infrared spectroscopy was able
to differentiate clearly between the sulphoxide and sul-
phone functionalities. For sulphoxides a single S=O
stretching mode (1058–1060 cm21) was observed whilst,
for sulphone functionalities, symmetric and antisymmetric
SO2 stretching modes were observed in the ranges 1073–
1095 cm21 and 1330–1334 cm21 in accordance with
literature data [10].
2.5.2. Cr(CO)5 complexes of 2,6-DTSH, 2,6-DTSH-2-
oxide and 2,6-DTSH-2,2-dioxide.
Attempts to prepare [Cr(CO)5(2,6-DTSH-2-oxide)] at
ambient
temperature
resulted
in
the
species
3.1. NMR studies
[hCr(CO)5jn(2,6-DTSH)](n51 and 2) as evidenced by
their 1H NMR signals. Attempts to prepare
[hCr(CO)5jhW(CO)5j(2,6-DTSH)] by the reaction of
[W(CO)5(2,6-DTSH)] with [Cr(CO)5(THF)] at 2308C
resulted in the observation of the desired complex (ca. 60%
yield by H NMR spectroscopy) plus the four complexes
[hM(CO)5jn(2,6-DTSH)] (M5Cr, W; n51, 2). Separation
was not attempted.
The W(CO)5 complexes I–IV were studied in detail by
variable temperature H NMR and their chemical shift and
1
scalar coupling constant data collected in Table 2. Assign-
ments of signals refer to the hydrogen labelling in Fig. 2.
In all cases the spectra showed gross changes on cooling
the solutions as a result of the slowing down of the
pyramidal inversion of W-coordinated sulphur atoms.
1

1348
E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
Table 1
Synthetic, melting point, IR spectroscopic and analytical data for complexes I–VII
Complex
No.
Yield
%
mp
8C
M–CO stretch
cm21
S–O stretch
cm21
Elemental analysisc
C(%)
H(%)
[W(CO)5(2,6-DTSH)]
I
53
79.7
2071(w), 1941(vs)a,
–
26.25 (26.50)
1.60 (1.75)
2077(w), 1947(vs), 1937(m)b
[hW(CO)5j2(2,6-DTSH)]
[W(CO)5(2,6-DTSH-2-oxide)]
II
III
62
47
132.6
118.0
2074(w) 1930(vs)a
–
23.15 (23.10)
25.40 (25.45)
0.95 (1.00)
1.75 (1.70)
2074(w), 1930(vs)a
1060(vs)a
2071(w), 1946(vs), 1939(vs)b
[W(CO)5(2,6-DTSH-2,2-dioxide)]
IV
68
112.8
2078(w), 1940(vs)a
1334(vs)a, 1073(s)a
24.55 (24.60)
1.65 (1.65)
2076(w), 1948(vs), 1942(vs)b
[hPdCl2(PPh3)2j(2,6-DTSH)]
[PdCl2(PPh3)(2,6-DTSH-2-oxide)]
[PdCl2(PPh3)(2,6-DTSH-2,2-oxide)]
V
VI
VII
79
44
69
194.9
184.2
183.4
–
–
–
–
45.95 (47.00)
3.90 (3.95)
d
d
1058(vs)a
1330(vs), 1095(s)a
44.80 (45.75)
3.75 (3.85)
a In CH2Cl2 solution.
b In very dilute hexane solution.
c Calculated values in parentheses.
d Not obtained.
Inversions were rapid on the 1H timescale at room
temperature but became slow by ca. 260 to 2808C.
In the case of [W(CO)5(2,6-DTSH)](I), the room tem-
perature spectrum consists of two singlets due to the
methylene protons of each four-membered ring. On cool-
ing, the higher frequency signal, due to the methylenes of
the metal-bound sulphur ring, split into an apparent AB
quartet whereas the lower frequency signal, due to the
methylenes of the other ring, split into two singlets (Fig.
3). A symmetry plane bisects these latter methylenes and
so their geminal hydrogens are chemically equivalent
although, in principle, they are magnetically non-equiva-
lent. However, no four-bond H–H couplings were detected
between the methylenes of this ring and so this magnetic
non-equivalence was not manifest and two singlet signals
persist at low temperature. The splitting (decoalescence) of
the C and D signals on cooling was followed by bandshape
analysis and reliable rate data obtained (Table 3).
terns (labelled AB and CD in Fig. 3). The low temperature
assignments given in Table 2 are somewhat tentative as
they depend on the long range shielding/deshielding effect
of the W(CO)5 moiety on the methylene hydrogens of the
other ring.
The spectral changes of this complex may be rational-
ised on the basis of the different symmetries associated
with rapid and slow pyramidal inversion of both S atoms.
The complex, like the free ligand, possesses D2d symmetry
when S inversions are rapid at ambient temperatures and
only a single type of methylene group exists. In the
absence of S inversion the symmetry of the complex
reduces to C2 and two pairs of methylene groups become
chemically distinguishable with geminal hydrogen distinc-
tion within each group. Hence, two AB quartet patterns (in
the absence of cross-ring coupling) are detected. It is
apparent from the spectrum at 2708C (Fig. 3) that two of
these hydrogens do, however, exhibit weak long-range
coupling. This could be either adjacent-ring or cross-ring
four-bond coupling. Results for complex III (see below)
suggest that cross-ring coupling may be more likely.
The 1H spectrum of [hW(CO)5j2(2,6-DTSH-2-ox-
ide)](II) consists of a singlet at room temperature which
splits into two slightly overlapping AB quartet-like pat-
Table 2
1H NMR parameters for the W(CO)5 complexes I–IV
Complex
Temperature
8C
Solvent
Chemical shifts
Spin–spin coupling
constants
dA
dB
dC
dD
2JAB
2JCD
I
30
260
30
280
30
270
30
270
CDCl3
CDCl3
3.75
3.85
3.35
3.63
3.44
3.98
|3.7b
4.46
3.27
211.0
210.5
|0
II
CD2Cl2
CD2Cl2
CD2Cl2
CD2Cl2
CD2Cl2
CD2Cl2
3.83
3.755
3.71a
|3.7b
3.92
3.61
3.764
|3.7b
4.26
29.8
III
IV
3.76
4.32
|3.7b
|3.90d
?
?
c
c
|3.90d
?
?
e
e
a Also dE / G 53.26, dF / H 54.01.
b
Also dE 53.35, dF 54.20, dG 53.25, dH 53.93.
c
2JAB, 2JCD could not be measured; 2JEF 511.5 Hz, 2JGH 511.7 Hz, 4JFH 56.6 Hz.
d uda 2dbu very small giving an AA9BB9 spectrum near the A4 limit.
e
2JAB, 2JCD could not be measured.

E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
1349
Fig. 2. The metal complexes of the ligands L1, L2 and L3, with the hydrogen labelling and symmetry point groups.
The spectral changes of complex II can be described in
terms of a spin problem, using the Haigh notation [11], of
type
On cooling the complex to ca. 2708C the arresting of the
S inversion should, in principle, lead to distinction between
all eight hydrogen environments. In reality, the four
hydrogens of the –S=O ring are clearly distinguished (Fig.
5) but the chemical shifts of the hydrogens A→D of the
metal-bound ring are closely similar at 2708C and only a
complex multiplet is detected. The assignments of the
signals E→H are based on the assumption that the –S=O
moiety will deshield those hydrogens which are cis to it,
i.e. hydrogens F and H. It is noteworthy that at 2708C,
these signals show a surprisingly strong four-bond mutual
scalar coupling, an analogous coupling not being observed
between the lower frequency pair of signals E and G, This
must be associated with a considerable distortion of the
ring geometry by the –S=O group.
k
[ABCD]2á[BADC]2
where the labels refer to the methylene hydrogens. If all
scalar couplings other than geminal couplings are ignored
then this problem reduces to
k
k
ABáBA and CDáDC.
This description was used as the basis of the bandshape
analysis and excellent fits between experimental and
computer simulated spectra were obtained (Fig. 4).
1
The H spectrum of [W(CO)5(2,6-DTSH-2-oxide)](III)
The spectral changes of this complex are described by
the dynamic spin problem
is more complex than the others because of its lower
molecular symmetry. At ambient temperatures where the S
inversion is rapid there are four chemically distinct CH2
groups. Those associated with the W←S bound ring
(labelled A–D, Fig. 2) are essentially singlets due to the
absence of the appreciable four-bond cross-ring couplings
whereas those of the sulphoxide ring (labelled E–H) are
detected as a four-spin AA9XX system, [12], this pattern
ABCDáBADC plus EFGHáGHEF
However, since the changes associated with the hydrogen
set A → D are not clearly observed, this part of the
spectrum was not fitted. Thus, total bandshape analysis
was carried out on the system EFGHáGHEF, where, in
2
2
addition to including the geminal couplings JEF and JGH
,
2
2
being dominated by the geminal couplings JEF and JGH
.

1350
E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
Fig. 3. The ‘static’ low temperature 250 MHz 1H spectra of (a) complex I (2608C), (b) complex II (2708C) and (c) complex IV (2708C). Solvent in all
cases was CD2Cl2.
4
the cross-ring coupling JEF (56.6 Hz) was also incorpo-
rated. The simulated spectra are shown in Fig. 5, from
which a set of rate constants was obtained, Table 3.
The complex [W(CO)5(2,6-DTSH-2,2-dioxide)](IV)
possesses the same symmetry as complex I and therefore
comparable spectra are expected. At room temperature two
singlets due to the methylenes of the two rings are
detected. On cooling the high frequency signal splits into
an equal intensity pair of singlets, whereas little change
occurs in the other signal. The low temperature limiting
spectrum (2708C) is shown in Fig. 3. The Cs symmetry of
this structure in the absence of S inversion leads to the
expectation of four distinct chemical shifts. A closer
inspection of the lowest frequency signal at d|3.90 reveals
additional weak signals characteristic of an AA9BB9
system [12] with a very small internal chemical shift,
Table 3
Best-fit rate constants used in the computer simulation of the 1H spectra of complexes I–IV and VI
T/8C
kI /s21
kII /s21
kIII /s21
kIV /s21
kVI /s21
260
250
245
240
235
230
225
220
215
210
25
3.3
10.3
23.0
42.5
80.0
144
250
440
–
–
–
–
–
–
–
9.2
23.5
46.0
79.0
136
221
375
–
–
–
–
–
–
–
–
12.2
–
56.0
–
157
250
430
680
1100
2450
30.1
53.5
95.0
170
300
480
780
2000
–
13.2
30.0
68.0
139
310
–
930
–
–
1670
–
0
10
–
–
–
–
–

E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
1351
have been induced by the sulphone function of the other
ring, causing the geminal environments to be magnetically
very similar. The other signals in the low temperature
spectra due to C,C9 and D,D9 are strictly part of an
AA9XX9 system, but line splittings are not fully resolved
4
because of the weakness of the cross-ring couplings JEG
,
,
4JEH 4JFG and JFH. This is in interesting contrast to
complex III where the S=O ring distortion strongly
favoured one four-bond coupling, presumed to be JFH
4
.
The spectral changes with temperature of the C and D
signals, treated as singlets, were computed in the usual
way and rate data calculated, Table 3.
Analogous NMR studies were performed on the
PdIICl2(PPh3) complexes of the ligands L1, L2 and L3.
These complexes in solution behaved in a similar manner
to the corresponding W(CO)5 complexes with pyramidal
inversion of the metal-bound S atom being fast on the
NMR time-scale at room temperature and slow at tempera-
tures below ca. 2508C. The 1H parameters of these
complexes are collected in Table 4. Low temperature
solution spectra were obtained for complexes V and VI,
and a full bandshape analysis performed on the spectra of
VI. In general, the spectra were of lower quality than those
of the W(CO)5 complexes, perhaps due to the more limited
solubility of these PdII species. The spectra of the dinu-
clear complex V are similar to those of II with four
chemical shifts resolved at low temperatures, but the scalar
1
Fig. 4.Variable temperature 250 MHz H spectra of complex II in CD2Cl2
with computer simulated spectra using ‘best-fit’ rate constants shown
alongside.
unA 2nBu, causing the spin system to be near its A4 limit.
This is unexpected by comparison with the spectrum of
complex I where unA 2nBu is appreciable. In complex IV
distortion of the geometry of the W←S bound ring must
Fig. 5. Variable temperature 250 MHz 1H spectra of complex III in CD2Cl2 with theoretically simulated spectra based on ‘best-fit’ rate constants shown
alongside. The weak sets of lines indicated by asterisks are due to a trace of free ligand. The signal labelling refers to Fig. 2.

1352
E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
Table 4
1H NMR parameters for the PdCl2(PPh3) complexes V–VII
Complex
Temperature
Solvent
Chemical shifts
Spin–spin coupling constants J/Hz
8C
dA
dB
dC
dD
JAB
JCD
V
30
250
30
250
30
CDCl3
CDCl3
CDCl3
CDCl3
CDCl3
3.92a
b
b
4.01
3.34
4.52
3.87
?
?
VI
VII
3.64c
3.87
3.68c
4.39
e
e
|3.79d
|3.79d
|3.53
|3.53
?
?
a Broad band (line width at half height |20 Hz).
b Not measurable due to poor spectral resolution.
c
Also dE / G 53.27, dF / H 54.17.
d
Assignments uncertain. Also dE 53.51, dF 54.75, dG 53.21, dH 53.89.
e
JAB, JCD could not be measured; 2JEF 511.8 Hz, 2JGH 511.3 Hz, 4JFH 56.4 Hz.
couplings were less easily identified. Complex VI gave the
expected spectral changes on cooling but complex VII gave
a low temperature spectrum which could not be analysed.
W(CO)5 complexes of four-membered S-heterocyclic
rings, the DG‡ value for the PdII complex, VI, of 50.4 kJ
mol21 is significantly lower than the DG‡ values obtained
previously [14] for the trans- PdII dichloride complexes of
the
four-membered
rings
and
4. Discussion
were 61.8 and 60.9 kJ mol21 respectively.
The activation parameters for pyramidal sulphur inver-
sion in both the W(CO)5 and PdCl2(PPh3) complexes of
the spiro ligands L1, L2 and L3 are given in Table 5. With
the exception of complex III, the data fall within narrow
limits. For example, DG‡ data, which are least prone to
This implies considerable distortion of the four-membered
rings in the present spiro compounds compared to in-
dividual thiabutane rings. This considerable lowering of
the activation energy for pyramidal inversion suggests an
enlargement of the C–S–C angle containing the metal-
bound sulphur atom so that there is easier access to the
planar sp2 transition state of the inversion process [5]. An
X-ray crystal structure would be needed to confirm this.
The change of heterocyclic sulphur to sulphoxide and to
sulphone certainly leads to further distortions of the four-
membered rings. This is particularly apparent in the
complexes III and VI where there is a single very strong
four-bond interaction between the methylene pairs rather
than two equal-magnitude interactions expected for a more
symmetrical ring geometry.
systematic error [13], are in the range 47.5–50.4 kJ mol21
.
Errors quoted for DG‡ data are as defined by Binsch and
Kessler [9]. Values for the mono- and di-nuclear W(CO)5
complexes of L9 are almost identical, implying that there is
no interaction between the sulphur lone pairs which can be
measured in terms of their pyramidal inversion characteris-
tics. Thus, electron–electron interactions of the type
observed in RuII /RuIII complexes of this same ligand as an
intervalence absorption band in the near IR region do not
appear to be as pronounced in these analogous dinuclear
W0 and PdII complexes. The activation energy values are
also independent of these transition metals. Whilst there do
not appear to be any previous energy data of S inversion in
The DNMR results of [W(CO)5L2] (III) are somewhat
anomalous. Despite good agreement between experimental
and theoretical bandshapes being achieved (Fig. 5), the
Table 5
Activation parameters for complexes I–IV and VI
Complex
I
Ea / kJ mol21
log10(A/s21
)
DH‡ /kJ mol21
DS‡ /J K21 mol21
DG‡a /kJ mol21
No. data points
7
58.0
60.7
53.6
60.4
76.9
60.9
52.5
60.7
57.4
14.6
60.2
13.7
60.1
18.4
60.2
13.4
60.2
14.0
56.0
60.7
51.5
60.4
74.9
60.9
50.5
60.7
55.3
28.4
63.1
11.4
61.7
99.9
63.6
5.0
63.1
16.3
65.0
47.5
60.2
48.1
60.1
45.1
60.2
49.0
60.2
50.4
II
III
IV
VI
7
6
6
8
61.3
60.3
61.3
60.2
a At 298.15 K.

E.W. Abel et al. / Polyhedron 18 (1999) 1345–1353
1353
DH‡ and DS‡ values are exceptionally high and not easily
References
explained. The very high positive DS‡ value accounts for
the rather lower DG‡ value for this complex compared to
the others. The small amount of impurity in this sample
(see Fig. 5) might have led to some systematic error in the
NMR fittings. We see no reason why the solution be-
haviour of this complex should differ significantly from the
others and therefore we do not attach as much reliability to
these data, particularly to the DS‡ value.
This work shows how the methylene hydrogens of the
four-membered rings of these dithiaspiroheptane complex-
es are sensitive NMR probes to the inversion dynamics of
the metal-coordinated S atoms with the orthogonality of
the four-membered rings leading to strikingly different
ways in which the hydrogens of the two rings respond to
the S inversion process.
[1] C.A. Stein, N.A. Lewis, G. Seitz, J. Am. Chem. Soc. 104 (1982)
2596.
[2] B. Chance, D.C. Devault, H. Frauenfelder, R.A. Marcus, J.R.
Shrieffer, N. Sutin (Eds.), Tunnelling in biological systems, New
York: Academic Press, 1979, p. 506.
[3] H. Taube, Pure Appl. Chem. 44 (1975) 25.
[4] C.A. Stein, H. Taube, J. Am. Chem. Soc. 100 (1978) 1635.
[5] E.W. Abel, S.K. Bhargava, K.G. Orrell, Prog. Inorg. Chem. 1 (1982)
32.
[6] H.J. Backer, K.J. Kerning, Recl. Trav. Chim. Pay-Bas 52 (1934)
449.
[7] H. Fujihara, K. Imaoka, N. Furukawa, S. Oae, Heterocycles 16
(1981) 1701.
[8] D. Kleier, G. Binsch, Program DNMR3, Quantum Chemistry
Program Exchange, Indiana University, USA, 1970.
[9] G. Binsch, H. Kessler, Angew. Chem. Int. Ed. Engl. 19 (1980) 411.
[10] D. Socrates, Infrared characteristic group frequencies, Wiley-Inter-
science, 1980, p. 112.
[11] C.W. Haigh, J. Chem. Soc. A (1970) 1682.
[12] R.J. Abraham The analysis of high resolution NMR spectra,
Amsterdam: Elsevier, 1971.
Acknowledgements
¨
[13] J. Sandstrom, Dynamic NMR spectroscopy, London: Academic
Press, 1982.
We are grateful to the EPSRC for a research studentship
for one of us (M.C.P.).
[14] E.W. Abel, M. Booth, K.G. Orrell, J. Chem. Soc., Dalton Trans.
(1979) 1994.