Reaction of mercury(0) with the I2 adduct of tetraphenyldithioimidodiphosphinic acid (SPPh2)2NH (HL) - Crystal structures of [Hg(HL)I2] and HgL2

Article abstract of DOI:10.1002/ejic.200400279

The complex [Hg(HL)I₂] (1) has been synthesised by reacting liquid Hg(0) in Et₂O under mild reaction conditions with the I ₂ adduct of HL, HL·I₂, while HgL₂ (2) has been obtained from the reaction of compound 1 with HL in CH₃CN. A single-crystal X-ray investigation of 1 shows four independent molecules in the asymmetric unit, each of which contains an Hg^II ion coordinated to two iodine atoms and two sulfur atoms of one bidentate neutral ligand in a distorted tetrahedral coordination geometry. Compound 2 consists of two anionic ligands coordinated to an Hg^II ion, which again displays a distorted tetrahedral coordination sphere. The reaction of 2 with HI (55 wt.-% in water) affords [Hg(HL)₂](I)₂ (3). Compounds 1, 2, and 3 have been characterised by FT-IR and ³¹P NMR spectroscopy. Density functional calculations suggest that compound 3 should feature a distorted tetrahedral coordination around the metal centre, with unequal Hg-S bond lengths. ( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004).

Full text of DOI:10.1002/ejic.200400279

FULL PAPER
Reaction of Mercury(0) with the I2 Adduct of
Tetraphenyldithioimidodiphosphinic Acid (SPPh2)2NH (HL) ؊
Crystal Structures of [Hg(HL)I2] and HgL2
M. Carla Aragoni,[a] Massimiliano Arca,[a] M. Bonaria Carrea,[a] Francesco Demartin,[b]
Francesco A. Devillanova,[a] Alessandra Garau,[a] Francesco Isaia,*[a] Vito Lippolis,[a] and
G. Verani[a]
Keywords: Bridging ligands / Coordination chemistry / Density functional calculations / Iodine / Mercury / S ligands
The complex [Hg(HL)I2] (1) has been synthesised by reacting
liquid Hg(0) in Et2O under mild reaction conditions with the
I2 adduct of HL, HL·I2, while HgL2 (2) has been obtained from
the reaction of compound 1 with HL in CH3CN. A single-
plays a distorted tetrahedral coordination sphere. The reac-
tion of 2 with HI (55 wt.-% in water) affords [Hg(HL)2](I)2 (3).
Compounds 1, 2, and 3 have been characterised by FT-IR
and 31P NMR spectroscopy. Density functional calculations
crystal X-ray investigation of 1 shows four independent mo- suggest that compound 3 should feature a distorted tetrahe-
lecules in the asymmetric unit, each of which contains an dral coordination around the metal centre, with unequal
HgII ion coordinated to two iodine atoms and two sulfur
atoms of one bidentate neutral ligand in a distorted tetrahe-
dral coordination geometry. Compound 2 consists of two
anionic ligands coordinated to an HgII ion, which again dis-
Hg−S bond lengths.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
aimed at improving our understanding of the nature of do-
nor-I2 bonding, we have recently decided to investigate the
metal activation reaction by using tetraphenyldithioimidod-
iphosphinic acid HN(Ph2PS)2 (HL), which contains both
group 15 and 16 elements, as the donor/complexing agent.
Our choice was driven both by the donor ability of phos-
phane-sulfide compounds[4] towards I2 and by the well-
known intrinsic ability of L to give chelation and to favour
the preferred geometry required by the metal ion.[5,6] Gen-
erally, the reaction of HL with metal ions leads to the for-
mation of the anion L, which has the negative charge de-
localised over the SPNPS moiety, and is capable of coordin-
ating a metal centre through the two sulfur atoms, to form
six-membered MSSP2N metallacycles.[5,6]
The oxidation/complexation reaction of metal powders
by diiodine adducts of phosphane and polyfunctional
thione donors has been investigated thoroughly in recent
years. Numerous papers on this subject have shown this
synthetic route to be a useful tool for obtaining coordi-
nation compounds with unusual geometries, stoichio-
metries, and oxidation numbers at the metal centre. In fact,
owing to the majority use of R3E·X2 adducts (R ϭ alkyl,
aryl; E ϭ P, As, Sb; X2 ϭ I2, Br2, IBr), novel metal tertiary
phosphane, arsane, and stibine complexes have been pre-
pared by McAuliffe and co-workers.[1] Recently, another
class of powerful oxidation reagents based on the I2 and
IBr adducts of N,NЈ-dimethylperhydrodiazepine-2,3-di-
thione has been employed successfully to oxidise gold metal
in a one-step reaction under mild conditions. This reaction
has found potential industrial applications in the fields of
electronics and the recovery of precious metals from waste
materials.[2]
Thanks to our extensive experience in the area of I2 ad-
ducts of sulfur- and selenium-containing donors,[3] mainly
In this context, the I2 adduct of HL, HL·I2, has shown an
interesting potential in the oxidation/complexation of zero-
valent metals in a low-polar solvent, as demonstrated by
the results described above. Antimony(0) powder, for ex-
ample, has been easily oxidised to SbIII in a single step to
give the unusual, X-ray characterised, dinuclear complex
[a]
`
Dipartimento di Chimica Inorganica ed Analitica, Universita
di Cagliari,
S. S. 554 bivio per Sestu, 09042 Cagliari, Italy
E-mail: isaia@unica.it
[b]
Dipartimento di Chimica Strutturale
e
Stereochimica
`
Inorganica, Universita di Milano,
Via G. Venezian 21, 20133 Milano, Italy
4660
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejic.200400279
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668

Reaction of Mercury(0) with the I2 Adduct of Tetraphenyldithioimidodiphosphinic Acid
FULL PAPER
[LSb(µ-S)(µ-I)2SbL].[7] Similarly, the adduct HL·I2 reacts ally characterised complexes bearing HL have been re-
with palladium and cobalt powders to produce the com- ported in the literature, and due to the current interest in
plexes PdL2 and CoL2.[9]
mercury complexes with dichalcogenoimidodiphosphinate
[8]
Since there is a current interest in the chemistry of mer- derivatives,[14aϪ14c] we expanded our study on 1 and 2 by
cury, which stems from its widespread industrial use as well characterising the complexes by X-ray single-crystal dif-
as its inherent toxicity and hazardous effects on human fractomety and 31P NMR and FT-IR spectroscopy; DFT
health,[10,11] and considering the promising results obtained calculations on the simplified model complexes
with this synthetic route for other zero-valent metals, we [Hg{N(H2PS)2}2] and [Hg{HN(H2PS)2}2I2], and on the ion
report here on the reaction of HL·I2 with liquid mercury [Hg{HN(H2PS)2}2]2ϩ were carried out as well.
with a view to opening new perspectives for the recovery of
The X-ray structures of 1 and 2 are reported in Figure 1
and 2, respectively; selected bond lengths and angles are
given in Table 1. The asymmetric unit of compound 1 con-
tains four independent [Hg(HL)I2] molecules with the metal
centres labelled Hg(1), Hg(2), Hg(3), and Hg(4), respec-
tively. Each molecule consists of an HgII ion coordinated to
two iodide ions and one HL ligand, with the metal ion
showing a very distorted tetrahedral coordination geometry
mercury(0) from industrial waste materials.[12]
Results and Discussion
Synthesis and X-ray Crystal-Structure Determination of
[Hg(HL)I2] and HgL2
[XϪHgϪX angles (X
ϭ
I, S) in the range
The reaction between equimolar quantities of the adduct
HL·I2 (generated in situ) and Hg ([HL·I2] ϭ 2.22 ϫ 10Ϫ3
, 25 °C) was carried out under nitrogen and anhydrous
conditions in diethyl ether solution. The dark-red colour of
the reaction mixture, due to the HL·I2 charge-transfer ad-
duct, disappeared almost completely within six hours with
formation of an insoluble whitish precipitate. At this stage
of the reaction, the concentration of HL·I2 in the mixture
is negligible, as determined from the very low intensity of
the blue-shifted band of iodine[13] at 460 nm; moreover, no
phosphorus-containing species was detectable by 31P NMR
spectroscopy. Elemental analysis of the separated solid and
31P NMR spectroscopy confirmed the formation of
[Hg(HL)I2] (1) with the ligand in its neutral form. In this
reaction, the zero-valent mercury is oxidised to HgII ac-
cording to Equation (1).
91.14(8)Ϫ135.66(4)°]. The HgSPNPS metallacycle displays
a pseudo-boat conformation similar to that found in other
metal complexes containing the anionic ligand L,[5] with a
(1)
Figure 1. ORTEP plot of one of the crystallographically indepen-
dent molecules of [Hg(HL)I2] (1); thermal ellipsoids are shown at
50% probability; hydrogen atoms, except for the one bonded to the
nitrogen, have been omitted for clarity
Complexes containing metal ions coordinated to undis-
sociated HL are very scarce in the literature since anything
less than strictly anhydrous reaction conditions or the use
of polar solvents always provokes imido-proton dissociation
and the consequent formation of the deprotonated ligand
L.[5a,6e] In this respect, in the case of the reactions with co-
balt(0) and palladium(0) we were able to isolate the species
[Co(HL)I2] and [Pd(HL)I2], but not to obtain suitable crys-
tals since crystallization attempts always led to the forma-
[8]
tion of complexes CoL2 and PdL2.[9]
The reaction in Equation (1) was also carried out in the
presence of an excess of the adduct (1:2 Hg:HL·I2 molar
ratio) in order to synthesise complexes with a different stoi-
chiometry. In addition to the formation of complex 1, a
small amount of white crystals of stoichiometry HgL2 (2)
was obtained after concentration of the reaction solution.
Although complex 2 has been reported previously,[6e] its
crystal structure has not been solved, and the complex has
Figure 2. ORTEP plot of [Hg(L)2] (2); thermal ellipsoids are de-
picted at 50% probability; hydrogen atoms have been omitted for
only partly been investigated. Moreover, since no structur- clarity
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4661

F. Isaia et al.
FULL PAPER
˚
Table 1. Selected interatomic distances (A) and angles (°) in the
structures of 1 and 2
[Hg(HL)I2] (1)
Hg(1)ϪI(1)
Hg(1)ϪI(2)
Hg(1)ϪS(1)
Hg(1)ϪS(2)
S(1)ϪP(1)
S(2)ϪP(2)
P(1)ϪN(1)
P(2)ϪN(1)
Hg(3)ϪI(5)
Hg(3)ϪI(6)
Hg(3)ϪS(5)
Hg(3)ϪS(6)
S(5)ϪP(5)
S(6)ϪP(6)
P(5)ϪN(3)
P(6)ϪN(3)
2.662(1)
2.691(1)
2.692(3)
2.709(3)
1.960(4)
1.970(4)
1.674(9)
1.686(9)
2.681(1)
2.678(1)
2.665(4)
2.679(4)
1.964(5)
1.980(5)
1.650(9)
Hg(2)ϪI(3)
Hg(2)ϪI(4)
Hg(2)ϪS(3)
Hg(2)ϪS(4)
S(3)ϪP(3)
S(4)ϪP(4)
P(3)ϪN(2)
P(4)ϪN(2)
Hg(4)ϪI(7)
Hg(4)ϪI(8)
Hg(4)ϪS(7)
Hg(4)ϪS(8)
S(7)ϪP(7)
S(8)ϪP(8)
P(7)ϪN(4)
2.683(1)
2.676(1)
2.660(4)
2.681(4)
1.966(4)
1.983(5)
1.645(9)
1.687(9)
2.694(1)
2.656(1)
2.715(3)
2.681(3)
1.962(4)
1.969(4)
1.659(9)
1.691(10) P(8)ϪN(4)
1.672(9)
I(1)ϪHg(1)ϪS(1) 118.79(8)
S(5)ϪHg(3)ϪI(6) 118.46(9)
I(1)ϪHg(1)ϪI(2)
S(1)ϪHg(1)ϪI(2)
I(1)ϪHg(1)ϪS(2)
S(1)ϪHg(1)ϪS(2)
I(2)ϪHg(1)ϪS(2) 116.54(8)
S(3)ϪHg(2)ϪI(4) 117.32(9)
S(3)ϪHg(2)ϪS(4)
I(4)ϪHg(2)ϪS(4)
S(3)ϪHg(2)ϪI(3)
I(4)ϪHg(2)ϪI(3)
134.00(3)
91.45(8)
95.07(8)
96.27(9)
S(5)ϪHg(3)ϪS(6)
I(6)ϪHg(3)ϪS(6)
S(5)ϪHg(3)ϪI(5)
I(6)ϪHg(3)ϪI(5) 128.49(5)
S(6)ϪHg(3)ϪI(5) 120.29(9)
I(8)ϪHg(4)ϪS(8) 117.43(8)
I(8)ϪHg(4)ϪI(7) 135.66(4)
S(8)ϪHg(4)ϪI(7)
I(8)ϪHg(4)ϪS(7)
S(8)ϪHg(4)ϪS(7)
96.96(10)
95.24(8)
94.55(7)
97.31(11)
95.65(9)
95.45(8)
91.14(8)
96.17(8)
95.87(9)
129.42(5)
S(4)ϪHg(2)ϪI(3) 118.43(9)
HgL2 (2)
2.5167(8) S(1)ϪP(1)
I(7)ϪHg(4)ϪS(7) 114.92(8)
Figure 3. View along the axis connecting Hg(2)ϪN(2) (a) and
Hg(4)ϪN(4) (b) in complex 1
HgϪS(1)
2.0221(9)
2.0221(9)
2.0157(9)
2.0135(9)
1.590(2)
1.585(2)
111.44(3)
104.42(3)
111.54(2)
HgϪS(2)
HgϪS(3)
HgϪS(4)
2.5436(7) S(2)ϪP(2)
2.5430(7) S(3)ϪP(3)
2.5569(8) S(4)ϪP(4)
Compound 2 consists of two anionic L ligands coordi-
nated to an HgII ion, which displays tetrahedral coordi-
nation geometry. Distortion from the idealized tetrahedral
geometry is less pronounced than in complex 1, the
SϪHgϪS angles ranging from 104.42(3)° to 112.13(2)°.
Each of the two pseudo-boat metallacycles arising from co-
ordination of the anionic ligand to the metal ion displays
P(1)ϪN(1)
P(2)ϪN(1)
S(1)ϪHgϪS(2)
S(1)ϪHgϪS(3)
S(1)ϪHgϪS(4)
1.587(2)
1.583(2)
112.13(2)
108.93(3)
108.30(3)
P(3)ϪN(2)
P(4)ϪN(2)
S(2)ϪHgϪS(3)
S(2)ϪHgϪS(4)
S(3)ϪHgϪS(4)
P and an S atom of the ring acting as the stern and bow of an almost exact C2 symmetry, with the twofold axis joining
the boat. Two out of the four independent complex mol- the metal and the N atoms. However, the two rings have
ecules, Hg(2) and Hg(3), conform to an almost exact C2 opposite configurations within the same molecule. As a re-
symmetry, the twofold axis passing through the Hg and N sult of deprotonation and coordination, the PϪS distances
˚
atoms and the puckering of the HgSPNPS metallacycle hav- increase from 1.937(1)Ϫ1.950(1) A in the free ligand to
˚
ing the same absolute configuration. As regards the other 2.0135(9)Ϫ2.0221(9) A in 2 and, accordingly, the PϪN
two [those containing Hg(1) and Hg(4)] the conformation bond lengths decrease from 1.672(2)Ϫ1.683(2) to
˚
of the ring is much more distorted and the absolute con- 1.583(2)Ϫ1.590(2) A.
figuration is the opposite of that found in the former two
Complexes 1 and 2 are easily distinguishable by IR spec-
complex molecules (see Figure 3a and 3b). Such distortions, troscopy since, upon coordination, HL and L have a differ-
together with some variability in the length of chemically ent π-electron delocalisation over the chelating ring
equivalent interactions, are due to packing effects, which HgSPNPS. The νasym(PNP) stretching band, which is the
play an important role in determining the actual symmetry most sensitive to metal coordination,[6e] occurs in HL as a
of the crystal. The PϪS distances are, on average, only strong absorption at 922 cmϪ1; this increases by about 300
slightly, although significantly, longer than those found in cmϪ1 when the imido proton is removed. As regards the
the
uncoordinated
ligand
HL[15]
[1.969(5)
vs. ν(PϪS) stretching frequencies, a 30Ϫ60 cmϪ1 and 50Ϫ80
1.937(1)Ϫ1.959(1) A]. Conversely, no significant variation cmϪ1 decrease is expected upon metal-ion chelation of HL
˚
is detected for the PϪN bonds [average 1.670(9) vs. and L, respectively. The IR spectra of 1 and 2 support these
˚
1.672(2)Ϫ1.683(2) A].
findings: the ν(PϪS) stretching frequencies measured in
4662
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668

Reaction of Mercury(0) with the I2 Adduct of Tetraphenyldithioimidodiphosphinic Acid
FULL PAPER
complex 1 (606 and 592 cmϪ1) and in complex 2 (576 and CH2Cl2, 25 °C) gives a 31P NMR spectrum with two reso-
561 cmϪ1) agree with a lower PϪS bond order than the free nances, one due to the free ligand and one due to the com-
ligand HL (645 and 622 cmϪ1 for the antisymmetric and plex. Variable-temperature NMR experiments were carried
symmetric stretching modes, respectively);[6d,6e] the out to as low as Ϫ80 °C, the lowest temperature for which
νasym(PNP) stretching frequency (922 cmϪ1 in HL)[6d,6e] ap- 31P NMR spectroscopic data were recordable, with no sig-
pears at 916 cmϪ1 in 1 and 1239 and 1219 cmϪ1 in 2; these nificant change in the spectra, suggesting the presence of
values are within the expected frequency range for single- different conformers.
bonded PϪN and double-bonded PϭN stretching regions,
respectively.
Reaction Chemistry of [Hg(HL)I2] and HgL2
The 31P CP MAS NMR spectrum of complex 1 shows
the presence of a single broad resonance at δ ϭ 59.5 ppm
with a spectral line-width of 300 Hz, suggesting that the
eight phosphorus atoms in the asymmetric unit have a very
similar environment. Conversely, the 31P CP MAS NMR
spectrum of 2 shows the presence of three resonances at δ ϭ
40.0 (∆ν1/2 ϭ 155 Hz), 37.8 (∆ν1/2 ϭ 163 Hz), and 33.9 ppm
(∆ν1/2 ϭ 148 Hz), with a relative intensity ratio of approxi-
mately 1:1:2. Evidently, small bond-length differences in the
anionic ligands L imply a different charge distribution over
the SPNPS fragment and lead to a significant effect on the
phosphorus chemical shifts. This effect is not observed for
complex 1, where the neutral ligand HL is involved in the
metal coordination. The 31P NMR spectrum (in CH2Cl2)
The attempts to synthesise new complexes by reacting
[Hg(HL)I2] (1) with increasing amounts of HL proved un-
successful ([1] ϭ 2.5 ϫ 10Ϫ3 ; HL to 1 molar ratio 3:1;
CH2Cl2, 25 °C). However, the same reaction carried out in
a more polar solvent like CH3CN (HL to 1 molar ratio 1:1;
25 °C) leads to the quantitative formation of complex 2,
thus confirming that in polar solvents HL is prone to lose
the imido proton when complexed to a metal.[6e,8,9]
The reaction of complex 1 with hydroiodic acid (55 wt.-
% in water) was also evaluated since it may prove useful in
freeing the ligand from the complex. Examination of the
equimolar reaction between HI and compound 1 (CH2Cl2,
25 °C) by 31P NMR spectroscopy revealed only one signal
at δ ϭ 57.3 ppm, which is only slightly high-field shifted
with respect to the resonance of complex 1; conversely,
when a twentyfold excess of HI is employed the resonance
is found at δ ϭ 56.6 ppm, indicating the presence of free
HL in the solution [Equation (2)]. Raman spectra carried
out on the organic phase confirmed the presence of free
HL, whereas in the water phase a peak at 121 cmϪ1 is in-
of 1 exhibits a single resonance at δ ϭ 57.8 ppm (∆ν1/2
ϭ
8.1 Hz), which confirms the presence of the ligand in the
complex in its neutral form when in solution too, since de-
protonation/complexation of the ligand results in a high-
field shift of the 31P NMR signal of about 20Ϫ30 ppm. The
corresponding chemical shift for HL in the same solvent is
δ ϭ 56.6 ppm (∆ν1/2 ϭ 3.6 Hz), and the small change
caused by coordination is in keeping with the structural
similarities between the free and bound ligand in the solid
state. The absence of any Hg/P coupling suggests that, at
room temperature, the ligand HL is labile on the NMR ti-
mescale. In fact, as increasing amounts of HL are added to
the solution of the complex up to a 3:1 molar ratio, only
dicative of the formation of the anion [HgI4]2Ϫ [16]
.
(2)
Equation (1) and (2) could obviously be very important
one phosphorus resonance is observed, although the chemi- as far as a recovery process is concerned, since they allow
cal shift decreases towards the value of δ ϭ 56.6 ppm as a not only an easy oxidation of mercury(0) to mercury(), but
consequence of the weighted average of the individual shifts also the possibility to free the metal from the complex 1
of the contributing species 1 and free HL. Low-temperature and to recover the ligand HL.
measurements were carried out to as low as Ϫ70 °C, the
Reaction of 2 with hydroiodic acid (55 wt.-% in water) in
lowest temperature for which 31P NMR spectroscopic data a 1:2 molar ratio in CH2Cl2 results in a considerable change
were recordable, with no Hg/P coupling or significant shifts in the 31P NMR spectrum of the complex (Figure 4a and
in the position of the resonance being observed. The broad- 4b). Indeed, the signal of 2 at δ ϭ 39.5 ppm disappears
ening of the signal with decreasing temperature (25 °C, completely and a new signal is observed at low field (δ ϭ
∆ν1/2 ϭ 8.1 Hz; Ϫ70 °C, ∆ν1/2 ϭ 10.4 Hz) is indicative of 57.6 ppm). The shift of the signal and its value, which is
the expected slowing of the chemical-exchange process. Ob- similar to that of complex 1 (δ ϭ 57.8 ppm, CH2Cl2),
viously, we were not able to achieve the temperature at strongly indicate the successful protonation of the anionic
which the complex is ‘‘frozen out’’ on the NMR timescale. ligand in the complex. Slow concentration of the dichloro-
In the case of complex 2, the 31P NMR spectrum methane solution afforded a white powder, whose micro-
(CH2Cl2) shows only one resonance at δ ϭ 39.7 ppm analysis suggests a stoichiometry [Hg(HL)2](I)2 (3). The 31
P
(∆ν1/2 ϭ 6 Hz) flanked by mercury satellites with a 2J- CP MAS NMR spectrum of 3 shows a broad band at δ ϭ
199
31
Hg- P)
coupling of 101 Hz. Thus, in solution at room 59.1 ppm (∆ν1/2 ϭ 494 Hz), confirming the presence of the
temperature, unlike the situation found in the solid state, ligands in their neutral form in the complex. IR bands at
the four phosphorus atoms in the complex are equivalent, 565 cmϪ1 [ν(PϪS)], 922 cmϪ1 [ν(PϪN)], and 2631 cmϪ1
(
,
indicating that the three resonances found in the solid state which originates from the first overtone of the δ(NϪH)
are due to solid-state effects. Addition of an equimolar bending vibration[6e] at 1325 cmϪ1, further support coordi-
amount of HL to a solution of 2 ([2] ϭ 1.1 ϫ 10Ϫ2 ; nation of HL to the metal in 3. In addition, the molar con-
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4663

F. Isaia et al.
FULL PAPER
atoms, the pVDZ basis set of Schafer et al.[20,21] was used.
The basis set for the metal atom is more difficult to choose
since, due to the high atomic number of mercury, relativistic
effects have to be taken into account. Thus, for 4 and 5, we
used the SBKJC with effective core potentials (ECP),[21,22]
since the LanL2DZ basis set previously used successfully
for Pd and In in 7 and 8 is not available for Hg. Geometry
optimizations on 4 and 5, performed starting from the
structural data of 2 and 1, respectively, show a very good
agreement with the experimental data. Thus, in both com-
plexes, the central mercury atom is coordinated by the li-
gand donor atoms in a distorted tetrahedral arrangement
˚
with average HgϪS bond lengths of 2.611 A (experimental
[23]
˚
2.543 A).
Analogously, in the case of 5, the average
˚
HgϪS and HgϪI bond lengths (2.889 and 2.743 A, respec-
10Ϫ3 ); (b) solution of 2 after addition of HI (55 wt.-% in H2O; tively) are very close to the experimental values determined
Figure 4. 31P NMR spectra in CH2Cl2 (25 °C) of: (a) 2 (3.58 ϫ
2/HI 1:2 molar ratio); (c) the solution from (b) after addition of
DMAN (2/HI/DMAN 1:2:1 molar ratio); (d) the solution from (c)
after addition of more DMAN (2/HI/DMAN 1:2:2 molar ratio)
˚
for 1 (2.700 and 2.688 A, respectively). In agreement with
the experimental structures of 1 and 2, a comparison of the
optimized structural features of the ligand moiety in 4 and
5 shows that, upon protonation, we obtain a lengthening
of the PϪN bond lengths (1.606 and 1.709 A in 4 and 5,
respectively) and a shortening of the PϪS bond lengths
ductivity of 3 (660 µS·cm2·molϪ1, CH2Cl2, 25 °C) is con-
siderably higher than that of 2 (93 µS·cm2·molϪ1, CH2Cl2,
25 °C). The protonation reaction is reversibl. A very strong
base such as 1,8-bis(dimethylamino)naphthalene (DMAN)
was added to a solution of 3 in CH2Cl2 in order to obtain
1:1 and 1:2 molar ratios of complex 3/DMAN. In the for-
mer case the neutral complex 2 is partly regenerated (Fig-
ure 4c) and in the latter fully regenerated (see d in Figure 4),
as shown by the 31P NMR peak at δ ϭ 39.5 ppm. The
reactions involving complexes 2 and 3 in CH2Cl2 are sum-
marised in Scheme 1.
˚
˚
(2.031 and 1.959 A in 4 and 5, respectively). In order to
better understand the interaction between the metal ion and
the ligands, a qualitative interpretation can be obtained
from the fragment molecular orbital approach (FMO)
within the formalism of the extended Hückel theory
(EHT).[24] The interaction diagram between the 6s and 6p
atomic orbitals of the Hg2ϩ ion and those of the ligands is
shown in Figure 5, calculated at the DFT-optimised geo-
metries. For both 4 and 5, the KohnϪSham HOMOs are
nonbonding orbitals built from the p atomic orbitals of the
donor atoms with a small contribution from the filled 5dxy
and 5dxz orbitals of the central metal atom. The LUMOs
are nonbonding in nature and they are made up of a combi-
nation of the 6s atomic orbital of the Hg ion σ-interacting
with the p-orbitals of the donor atoms pointing towards the
central metal ion. Confident in the agreement between the
calculated geometries of 4 and 5 and their parent com-
plexes, we extended our DFT calculations at the same level
of theory to 6 so as to acquire information on the geometri-
cal features of the parent cationic complex 3. In this case
Scheme 1
Quantum-Chemical Calculations
In the light of the novelty of complex 1, and in order to the Hg ion is also coordinated by the two ligand units in a
provide further insight into the electronic properties of the distorted tetrahedral fashion. In agreement with what em-
title compounds, we performed hybrid-DFT calculations on erges from the comparison of the calculated structural fea-
complexes 1 and 2. We also determined the geometrical fea- tures of 4 and 5, upon imido protonation, model 6 shows
˚
tures of complex 3. Theoretical calculations were performed an increase in the HgϪS (average value 2.628 A) and the
˚
on the model complexes [Hg{N(H2PS)2}2] (4) and PϪN (average value 1.698 A) bond lengths compared to
[Hg{HN(H2PS)2(I)2}] (5), in which the phenyl groups have those calculated for 4, while the PϪS bond lengths are
˚
been replaced by hydrogen atoms, and subsequently on the slightly shortened (average value 1.993 A). Since 4 and 6
model complex [Hg{HN(H2PS)2}2]2ϩ (6). As previously re- are isoelectronic, their frontier orbitals are very similar in
ported by us for similar model complexes, hybrid-DFT cal- shape and composition. In particular, the KohnϪSham
culations have given reliable results in understanding both HOMOs of 4 and 6 (see a in Figure 6) are slightly anti-
the structural and electronic features of the model com- bonding orbitals built from the p atomic orbitals of the do-
plexes [Pd{N(H2PS)2}2][8] (7) and [In{N(H2PS)2}I2][9b] (8). nor atoms with a small contribution from the filled 5dxy
Encouraged by our recent results,[17] the mPW1PW[18] func- and 5dxz orbitals of the central metal atom. On the other
tional was preferred to the B3LYP functional,[19] which was hand, in the case of 5 (c in Figure 5), the HOMO and the
previously employed for 7 and 8. As regards the ligand HOMOϪ1 are almost entirely centered on the iodides, and
4664
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668

Reaction of Mercury(0) with the I2 Adduct of Tetraphenyldithioimidodiphosphinic Acid
FULL PAPER
Figure 5. EHT-interaction diagram showing the interaction between the empty atomic orbitals of the Hg2ϩ central ion and the filled
ones of the ligands HL in the model complexes 4 and 5 [HL ϭ H2P(S)ϪNHϪP(S)H2]; the dashed lines connect FMOs which contribute
at least 15% to the formation of molecular orbitals; the hybrid-DFT frontier orbitals are also depicted (contour value 0.05)
they are therefore nonbonding in nature. The LUMOs of obtained from the reaction of HgX2 (X ϭ Br, I) with HL.
the three complexes are all antibonding in nature, and are The reaction of 2 with concentrated HI leads to the species
due to the combination of the 6s atomic orbitals of the Hg [Hg(HL)2](I)2 (3; see Scheme 2). X-ray diffraction analyses
ion σ-interacting with the p-orbitals of the donor atoms and DFT calculations on simplified model complexes have
pointing towards the central metal ion (see b and d in Fig- been performed to elucidate the structural and electronic
ure 5 for 4 and 5, respectively).
properties of 3. Complex 1 readily reacts with an excess of
HI to yield the anion [HgI4]2Ϫ with the release of HL,
which can be re-utilised, hopefully opening new perspec-
tives for the recovery of Hg(0) from mercury-containing
waste materials.
Scheme 2
Experimental Section
Figure 6. Sketches of KohnϪSham HOMOs and LUMOs calcu-
lated for 4 (a and b, respectively), and 5 (c and d, respectively)
Materials and Instrumentation: All reagents were used as purchased
from Aldrich. Diethyl ether was distilled from over LiAlH4 shortly
before use. CH2Cl2 was distilled from P2O5 and stored under N2.
All reactions were carried out under an atmosphere of dry nitrogen.
31P{1H} NMR spectra were recorded on a Varian Unity Inova
400 MHz operating at 161.9 MHz. The 31P chemical shifts are ref-
erenced to an external standard of 85% H3PO4 (δ ϭ 0.0 ppm). 31P
NMR CP MAS NMR spectra were calibrated indirectly from the
H3PO4 peak (δ ϭ 0.0 ppm). Variable temperature 31P NMR spec-
Conclusion
The oxidizing ability of the adduct HL·I2 in Et2O
towards liquid Hg(0) leads to the synthesis of the air stable
complex [Hg(HL)I2] (1) in a single oxidation step. To the
best of our knowledge, this complex represents the first ex-
ample of a structurally characterised complex of HL. The
reaction of 1 with an excess of HL in MeCN affords the
troscopic data on compound 1 were collected from a 1.1 ϫ 10Ϫ2

solution in CH2Cl2 [temp. (°C), δ, ∆ν1/2: 25, 57.7, 8.1; 0, 57.9, 8.6;
Ϫ20, 58.1, 10.0; Ϫ40, 58.3, 10.2; Ϫ60, 58.3, 10.5, Ϫ70, 58.4, 10.4].
31P NMR data for compound 2 were collected from a 1.2 ϫ 10Ϫ2
neutral air stable complex [Hg(L)2] (2), which can also be  solution in CH2Cl2 [temp. (°C), δ, ∆ν1/2: 25, 39.5, 6.1; Ϫ20, 39.5,
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668
www.eurjic.org  2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 4665

F. Isaia et al.
FULL PAPER
9.0; Ϫ40, 39.5, 9.5; Ϫ70, 39.1, 12.9, Ϫ80, 39.2, 14.0]. IR spectra
were measured as KBr (4000Ϫ400 cmϪ1) or polyethylene pellets
(400Ϫ50 cmϪ1) on a Bruker IFS 55 FT-IR spectrometer. FT Ra-
man spectra were recorded on a Bruker FRS 100 Fourier-transform
Raman spectrometer, operating with a diode-pumped Nd:YAG ex-
citing laser at 1064 nm. Conductometric measurements were car-
ried out at 25 °C.
using the SHELXS 97 program,[27] followed by difference Fourier
synthesis, and refined by full-matrix least-squares procedures on F2
using the SHELXL-97 program.[28] All non-H atoms were refined
anisotropically and H atoms were introduced at calculated posi-
tions and thereafter incorporated into a riding model with
Uiso(H) ϭ 1.2Ueq(C). For compound 1 the search for an alternative
symmetry other than the triclinic one reported in Table 2 was done
using Le Page’s MISSYM (PLATON) routine.[29] The check re-
vealed that the triclinic unit cell can be transformed metrically into
an almost-monoclinic C-centred unit cell with a ϭ 24.485(3), b ϭ
Synthesis: HL was prepared according to ref. 25. The adduct HL·I2
was prepared in situ in distilled Et2O just before use from equimo-
lar amounts of HL and I2.
˚
24.506(3), c ϭ 19.230(3) A, α ϭ 90.42(1), β ϭ 103.25(1), γ ϭ
89.61(1)°. However, with this choice, the agreement between equiv-
alent reflections was poor (Rint ϭ 0.434) thus ruling out the pos-
sibility of the symmetry being monoclinic.
CCDC-228871 (for 1) and -228872 (for 2) contain the supplemen-
tary crystallographic data for this paper. These data can be ob-
tained free of charge at www.ccdc.cam.ac.uk/conts/retrieving.html
[or from the Cambridge Crystallographic Data Centre, 12 Union
Road, Cambridge CB2 1EZ, UK; Fax: ϩ44-1223-336033; E-mail:
deposit@ccdc.cam.ac.uk].
[Hg(HL)I2] (1): A mixture of HL (0.100 g, 0.222 mmol) and I2
(0.0565 g, 0.223 mmol) in Et2O (100 mL) was stirred at room tem-
perature under N2 until complete dissolution of the reagents.
Liquid mercury (0.0236 g, 0.222 mmol) was then added while stir-
ring. This reaction was continued at room temperature under N2
for about one day. The resulting air-stable whitish solid was isolated
by suction filtration, washed with diethyl ether, and dried in vacuo,
to give 0.172 g of 1 (0.211 mmol; yield 95%). Stable colourless crys-
tals were obtained upon dissolution of 1 in anhydrous CH2Cl2 and
subsequent slow evaporation of the solvent. C24H21HgI2NP2S2
(903.87): calcd. C 31.89, H 2.34, N 1.55, S 7.09; found C 32.0, H
2.5, N 1.6, S 7.4. IR (KBr): ν˜ ϭ 3054w cmϪ1, 2966w, 2924w, 2855w,
2606brw, 1435vs, 1384m, 1303m, 1184m, 1107vs, 916s, 804m, 782s,
733vs, 720vs, 685vs, 622w, 606ms, 592s, 506m, 487m, 464m, 441w.
31P NMR (CH2Cl2): δ ϭ 57.8 ppm. M.p. 228Ϫ230 °C (dec.).
Table 2. Crystal data and details of refinements for [Hg(HL)I2]
and HgL2
Compound
[Hg(HL)I2] (1)
HgL2 (2)
Formula
C24H21HgI2NP2S2 C48H40HgN2P4S4
Mol. mass (amu)
Crystal system
Space group
903.87
1097.53
triclinic
P1 (no. 2)
[Hg(L)2] (2). Method a: A mixture of HL (0.100 g, 0.222 mmol)
and [Hg(HL)I2] (0.200 g, 0.222 mmol) in CH3CN (100 mL) was
stirred at room temperature for 4 h. The solid white product was
collected by filtration and then dissolved in hot CH2Cl2. The re-
sulting solution was allowed to evaporate in air to yield a white
solid. Yield: 0.230 g (0.210 mmol, 95%). Method b:[6e] A solution
of HL (0.125 g, 2.774. mmol) in CHCl3 (10 mL) was added to a
solution of HgBr2 (0.050 g, 1.387 mmol) in Et2O (10 mL). The re-
action mixture was stirred for 1 h at room temperature and then
refluxed for 20 min. The solid product was separated by filtration,
washed twice with a 1:1 acetone/diethyl ether mixture, and dried
under vacuum to give 0.136 g of 2 (89% yield). C48H40HgN2P4S4
(1097.5): calcd. C 52.32, H 3.67, N 2.55, S 11.68; found C 53.0, H
3.9, N 2.6, S 12.0. IR (KBr): ν˜ ϭ 3054m cmϪ1, 2918w, 2846w,
1475w, 1436s, 1384w, 1239s, 1219s, 1174s, 1103s, 1025w, 997w,
786w, 742m, 714s, 698vs, 618w, 576s, 561vs, 514s, 495m. 31P NMR
(CH2Cl2): δ ϭ 39.7 ppm. M.p. Ͼ 255 °C
triclinic
¯
¯
P1 (no. 2)
˚
a (A)
17.262(2)
17.380(2)
19.230(3)
99.59(1)
99.05(1)
90.05(1)
5616(1)
8
13.626(1)
13.794(1)
14.336(1)
82.21(1)
66.10(1)
69.44(1)
2306.6(4)
2
1.580
3.692
0.735Ϫ1.000
8871
0.0243, 0.0613
˚
b (A)
˚
c (A)
α (°)
β (°)
γ (°)
3
˚
U (A )
Z
ρcalcd (Mg·mϪ3
)
2.138
7.957
0.602Ϫ1.000
µ(Mo-Kα) (mmϪ1
Transmission factors
)
Unique refl. with I Ͼ 2σ(I) 9240
Final R and wR2 indices 0.0496, 0.1225
Computations: Quantum-chemical DFT calculations were carried
out with the Gaussian 98 (rev. A11)[30] suite of programs on the
model complexes [Hg{N(H2PS)2}2] (4), [Hg{N(H2PS)2}I2] (5), and
[Hg{HN(H2PS)2}2]2ϩ (6), with Adamo and Barone’s mPW1PW
functional[18] and pVDZ Schafer, Horn and Ahlrichs pVDZ basis
sets for H, N, P, and S.[20,21] For Hg and I, the SBKJC basis set
with effective core potentials was used.[22] In the case of 4, the cal-
culations were also performed with the Stuttgart RLC ECP basis
set for Hg.[23] Integration was performed numerically using a total
of 7500 points for each atom (Finegrid option). The geometry opti-
mizations of 4, 6 and 5 were performed starting from the exper-
imental geometries determined for 1 and 2, avoiding introducing
any symmetry constraint. DFT calculations were performed on an
SGI Origin 3800 equipped with 128 GByte of RAM. The results
of the calculations were analysed with the Molekel 4.3[32a] and
Molden 3.9[32b] programs. The quantum mechanically optimised
structure was verified by normal-mode harmonic frequency analy-
[Hg(HL)2](I)2 (3): A dichloromethane solution (10 mL) of 2
(0.020 g, 0.018 mmol) and a hydroiodic acid solution (55 wt.-% in
water; 0.0092 g, 0.04 mmol) were vigorously stirred for 1 h at 25
°C. The colourless two-phase mixture was then passed through a
phase-separation filter (Whatman, Phase Separator 1PS) to remove
the water. After concentration of the solution to about 5 mL, n-
hexane (5 mL) was added dropwise and a white solid precipitated.
This was washed with n-hexane and dried to give 0.018 g of 3 (74%
yield). C48H42HgI2N2P4S4 (1353.4): calcd. C 42.60, H 3.13, N 2.07,
S 9.48; found C 43.0, H 3.2; N 2.2, S 10.0. IR (KBr): ν˜ ϭ 2631w
cmϪ1, 1436vs, 1325m, 1218m, 1106s, 922s, 785s, 741s, 716s, 688vs,
649m, 561s, 513m. 31P NMR (CH2Cl2): δ ϭ 57.6 ppm. M.p.
203Ϫ204 °C
X-ray Crystallography: Crystal data and refinement details for
structure determinations of both compounds are reported in
Table 2. Intensity data were acquired at room temperature on a sis. Harmonic frequencies were obtained using the second deriva-
SMART CCD diffractometer (ω scans). All data sets were cor-
tives of the DFT energy, computed by numerical differentiation
of the DFT energy gradients. The qualitative EHT analyses were
rected for Lorentz-polarisation effects and for absorption using the
SADABS routine.[26] The structures were solved by direct methods obtained with the CACAO (Computer Aided Composition of
4666
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668

Reaction of Mercury(0) with the I2 Adduct of Tetraphenyldithioimidodiphosphinic Acid
FULL PAPER
[12]
Atomic Orbitals) program package[32c] on the hybrid-DFT-op-
timised geometry. The interaction diagram was generated using the
fragment molecular orbital (FMO) approach.[32d] Parameters used
in EHT calculations were taken from the standard database of
CACAO.
The largest source of recycled mercury-containing products is
the spent catalysts used in the production of chlorine and caus-
tic soda, followed by electrical apparatus, dental amalgams,
batteries, and other instruments such as thermometers. See: L.
G. Twidwell, R. J. Thompson, JOM-J. Miner. Met. Mater. Soc.
2001, 53, 15Ϫ17.
K. F. Purcell, J. C. Kotz, Inorganic Chemistry, W. B. Saunders
Company, 1977, chapter 5, p. 209Ϫ212.
V. Garcia-Montalvo, J. Novosad, P. Kilian, A. M. Z. Sla-
win, P. Garcia y Garcia, M. Lopez-Cardoso, G. Espinosa-
[13]
Acknowledgments
[14] [14a]
This research was financially supported by the ‘‘Regione Autonoma
della Sardegna’’. We are grateful to CINECA (Consorzio Interun-
iversitario per il Calcolo Automatico dellЈItalia Nord Orientale) for
providing computational facilities.
Perez, R. Cea-Olivares, J. Chem. Soc., Dalton Trans. 1997,
[14b]
1025Ϫ1029.
O. Bumbu, A. Silvestru, C. Silvestru, J. E.
Drake, M. B. Hursthouse, M. E. Light, J. Organomet. Chem.
[14c]
2003, 687, 118Ϫ124.
D. J. Crouch, P. M. Hatton, M. Hel-
liwell, P. O’Brien, J. Raftery, Dalton Trans. 2003, 2761Ϫ2766.
S. Husebye, K. Maartmann-Moe, Acta Chem. Scand., Ser. A
1983, 37, 439Ϫ441.
[1] [1a]
S. M. Godfrey, C. A. McAuliffe, R. G. Pritchard, J. M.
[15]
[16]
[1b]
Sheffield, Inorg. Chim. Acta 1999, 292, 213Ϫ219.
S. M.
Godfrey, N. Ho, C. A. McAuliffe, R. G. Pritchard, Angew.
Chem. Int. Ed. Engl. 1996, 33, 2344Ϫ2345. [1c] C. A. McAuliffe,
S. M. Godfrey, A. G. Mackie, R. G. Pritchard, Angew. Chem.
Int. Ed. Engl. 1992, 31, 919Ϫ921, and references cited therein.
M. L. Delwaulle, Bull. Soc. Chim. Fr. 1955, 247, 1294Ϫ1299.
[17] [17a]
M. C. Aragoni, M. Arca, C. Denotti, F. A. Devillanova,
E. Grigiotti, F. Isaia, F. Laschi, V. Lippolis, L. Pala, A. M. Z.
Slawin, P. Zanello, J. D. Woollins, Eur. J. Inorg. Chem. 2003,
[2] [2a]
`
L. Cau, P. Deplano, L. Marchio, M. L. Mercuri, L. Pilia,
[17b]
1291Ϫ1295.
M. C. Aragoni, M. Arca, F. Demartin, F. A.
[2b]
A. Serpe, E. F. Trogu, Dalton Trans. 2003, 1969Ϫ1974.
Bigoli, P. Deplano, M. L. Mercuri, M. A. Pellinghelli, G.
Pintus, A. Serpe, E. F. Trogu, J. Am. Chem. Soc. 2001, 123,
1788Ϫ1789.
Pellinghelli, G. Pintus, M. A. Serpe, E. F. Trogu, Chem. Com-
mun. 1998, 2351Ϫ2352.
lanova, V. Lippolis, M. L. Mercuri, M. A. Pellinghelli, E. F.
F.
Devillanova, F. Isaia, F. Lelj, V. Lippolis, A. Mancini, L. Pala,
Eur. J. Inorg. Chem. 2004, 3099Ϫ3109.
[18]
[2c]
C. Adamo, V. Barone, J. Chem. Phys. 1998, 108, 664Ϫ675.
F. Bigoli, P. Deplano, M. L Mercuri, M. A.
[19] [19a]
[19b]
A. D. Becke, J. Chem. Phys. 1993, 98, 1372Ϫ1377.
A.
[19c]
[2d]
D. Becke, J. Chem. Phys. 1993, 98, 5648Ϫ5652.
Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785Ϫ789.
C. Lee, W.
F. Bigoli, P. Deplano, F. A. Devil-
[20]
A. Schafer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97,
2571Ϫ2577.
Trogu, Inorg. Chim. Acta 1998, 267, 115Ϫ121.
[3] [3a]
M. C. Aragoni, M. Arca, F. Demartin, F. A. Devillanova,
[21] [21a]
A. Garau, F. Isaia, V. Lippolis, G. Verani, Trends Inorg. Chem.
W. J. Stevens, M. Krauss, H. Basch, P. G. Jasien, Can. J.
[3b]
[21b]
1999, 6, 1Ϫ18.
M. C. Aragoni, M. Arca, F. Demartin, F.
Chem. 1992, 70, 612Ϫ630.
Chem. Phys. 1993, 5555Ϫ5565.
T.R. Cundari, W. J. Stevens, J.
A. Devillanova, A. Garau, F. Isaia, V. Lippolis, G. Verani, Co-
[3c]
[22]
ord. Chem. Rev. 1999, 184, 271Ϫ290.
M. C. Aragoni, M.
Basis sets were obtained from the Extensible Computational
Chemistry Environment Basis Set Database, Version 6/19/03,
as developed and distributed by the Molecular Science Com-
puting Facility, Environmental and Molecular Sciences Labor-
atory which is part of the Pacific Northwest Laboratory, P.O.
Box 999, Richland, Washington 99352, USA, and funded by
the U.S. Department of Energy. The Pacific Northwest Labora-
tory is a multi-program laboratory operated by Battelle Mem-
orial Institute for the U.S. Department of Energy under con-
tract DE-AC06Ϫ76RLO 1830.
Arca, A. J. Blake, F. A. Devillanova, W.-W. du Mont, A. Ga-
rau, F. Isaia, V. Lippolis, G. Verani, C. Wilson, Angew. Chem.
Int. Ed. 2001, 40, 4229Ϫ4232.
[4]
[4a] M. Arca, F. A. Devillanova, A. Garau, F. Isaia, V. Lippolis,
G. Verani, F. Demartin, Z. Anorg. Allg. Chem. 1998, 624,
745Ϫ749. [4b] M. Arca, F. Demartin, F. A. Devillanova, A. Ga-
rau, F. Isaia, V. Lippolis, G. Verani, J. Chem. Soc., Dalton
Trans. 1999, 3069Ϫ3073.
[5] [5a]
I. Haiduc, Comprehensive Coordination Chemistry II, vol 1
[23]
(Ed.: A. B. P. Lever), Elsevier Pergamon, Amsterdam, Oxford,
The same calculation on 4 has also been performed with the
larger Stuttgart RLC ECP basis set for the mercury atom. The
calculated molecular orbital composition as well as the geo-
metrical features are very close to those calculated with the
SBKJC ECP basis set. Calculated HgϪS distances: 2.583,
2004, p. 323Ϫ347. [5b]C. Silvestru, J. E. Drake, Coord. Chem.
[5c]
Rev. 2001, 223, 117Ϫ216.
Chem. Rev. 1998, 176, 451Ϫ481, and references cited therein.
T. Q. Ly, J. D. Woollins, Coord.
[6] [6a]
C. G. Pernin, J. A. Ibers, Inorg. Chem. 2000, 39, 1216Ϫ121.
[6b] C. G. Pernin, J. A. Ibers, Inorg. Chem. 2000, 39, 1222Ϫ1226.
˚
2.564, 2.637, and 2.666 A.
[6c] O. Siiman, C. P. Huber, M. L. Post, Inorg. Chim. Acta 1977,
25, L11ϪL14.
[24]
[25]
[26]
[27]
[28]
[29]
[30]
R. Hoffmann, W. N. Lipscomb, J. Chem. Phys. 1962, 36,
2179Ϫ2189.
A. Schmidpeter, R. H. Bohm, H. Groenger, Angew. Chem.
1964, 76, 860Ϫ861.
SADABS Area-Detector Absorption Correction Program,
Bruker AXS, Inc. Madison, WI, USA (2000).
SHELXS-97; G. M. Sheldrick, Acta Crystallogr., Sect. A 1990,
46, 467Ϫ473.
G. M. Sheldrick, SHELXL-97, Program for Crystal Structure
Refinement, Universität Göttingen, Germany, 1997.
A. L. Spek, Acta Crystallogr., Sect. A 1990, 46, supplement
C34, MS-02.01.05.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgom-
ery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C.
Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B.
[6d]
O. Siiman, J. Vetuskey, Inorg. Chem. 1980,
19, 1672Ϫ1680. [6e] G. P. McQuillan, I. A. Oxton, Inorg. Chim.
Acta 1978, 29, 69Ϫ75.
M. Arca, F. A. Devillanova, A. Garau, F. Isaia, V. Lippolis, G.
Verani, G. L. Abbati, A. Cornia, Z. Anorg. Allg. Chem. 1999,
625, 517Ϫ520.
[7]
[8]
G. L. Abbati, M. C. Aragoni, M. Arca, A. C. Fabretti, F. A.
Devillanova, A. Garau, F. Isaia, V. Lippolis, G. Verani, J.
Chem. Soc., Dalton Trans. 2001, 1105Ϫ1110.
[9] [9a]
M. C. Aragoni, M. Arca, A. Garau, F. Isaia, V. Lippolis,
G. L. Abbati, A. C. Fabretti, Z. Anorg. Allg. Chem. 2000, 626,
[9b]
1454Ϫ1459.
G. L. Abbati, M. C. Aragoni, M. Arca, F. A.
Devillanova, A. C. Fabretti, A. Garau, F. Isaia, V. Lippolis, G.
Verani, Dalton Trans. 2003, 1515Ϫ1519.
[10]
[11]
W. Kaim, B. Schwedersky, Bioinorganic Chemistry: Inorganic
Elements in the Chemistry of Life, an introduction and Guide,
Wiley-Interscience, New York 1991, chapter 17, p. 338Ϫ343.
H. G. Seiler, H. Sigel, A. Sigel, Handbook on Toxicity of Inor-
ganic Compounds, Marcel Dekker, Inc., New York and Basel,
1988, chapter 35, p. 419Ϫ436.
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4667

F. Isaia et al.
FULL PAPER
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L.
Andres, M. Head-Gordon, E. S. Replogle, J. A. Pople, GAUS-
SIAN 98 (Revision A 11), Gaussian, Inc., Pittsburgh, PA, 1998.
Weber, Swiss Center for Scientific Computing, Manno (Switz-
[32b]
erland), 2000.
G. Schaftenaar, J. H. Noordik, MOLDEN
3.6: a pre- and post-processing program for molecular and elec-
tronic structures, J. Comp.-Aided Mol. Design 2000, 14,
123Ϫ124. [32c] C. Mealli, D. M. Proserpio, J. Chem. Educ. 1990,
[32d]
67, 399Ϫ402.
1397Ϫ1412.
R. Hoffmann, J. Chem. Phys. 1963, 39,
[31] [31a] W. Kuechle, M. Dolg, H. Stoll, H. Preuss, Mol. Phys. 1991,
[31b]
74, 1245Ϫ1263.
A. E. Reed, L. A. Curtiss, F. Weinholt,
Received April 5, 2004
Early View Article
Published Online October 18, 2004
Chem. Rev. 1988, 88, 899Ϫ926.
[32] [32a]
MOLEKEL 4.3, P. Flükiger, H.P. Lüthi, S. Portmann, J.
4668
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 4660Ϫ4668