Gold(I) and mercury(II) xenon complexes

Article abstract of DOI:10.1002/anie.200351208

Noble-gas ligands: For the first time it has been possible to prepare a gold(I) xenon complex ([(F₃As)AuXe]⁺, see picture) and a mercury(II) xenon complex. The syntheses of these compounds start from precursors with weakly coordinated metal ions and take place in a superacid medium.

Full text of DOI:10.1002/anie.200351208

Communications
Coordination of Noble-Gas Atoms
Gold(i) and Mercury(ii) Xenon Complexes**
In-Chul Hwang, Stefan Seidel, and Konrad Seppelt*
During detailed investigation of the reaction which produced
[1]
the first gold xenon compound [AuXe4]2+[Sb2F11]À
a
2
number of further gold(ii) xenon compounds were discov-
ered.[2] In only one instance could a gold(iii) xenon compound
be detected.[2] However, gold occurs most frequently in its
complexes as gold(i). At a first glance a gold(i) xenon bond
may not appear particularly favored as gold(i) has a fully
occupied 5d shell, but this consideration totally overlooks the
relativistic effect, which, in the periodic system, reaches a first
maximum for gold.[3] The [AuXe]+ ion has already been
detected by mass spectrometry.[4] According to experi-
mentation and calculation this cation has a bond energy of
Figure 1. Crystal structure of 1 (ORTEP representation, thermal ellip-
soids set at 50% probability).For numerical values see Table 1.
Table 1: Selected structural data for [(F3As)AuXe]+ [Sb2F11]À (1) and
[HgXe]2+ [SbF6]À [Sb2F11]À (2).
Separations
r [pm] and
angles [8]
Crystal structure
analysis
1
MP2 calculation
[(F3As)AuXe]+[a]
30 Æ 3 kcalmolÀ1 [4]
; a gold–xenon bond length of 276.1 pm was
predicted. A linear [AuXe2]+ ion could also exist.[5]
rAu-Xe
rAu-As
rAs-F
260.72(6)
231.52(8)
165.9(5), 166.4(4), 169.7
166.7(5)
260.5
230.6
The fundamental difficulty in the preparation of a gold(i)
xenon compound is that gold(i) always exists in the form of
complexes whose ligands, predictably, bind more effectively
than xenon. Therefore a displacement reaction appears to be
impossible. On the other hand, a highly promising educt
would be a compound of an (almost) naked Au+ ion
rAu···F(Sb)
aAs-Au-Xe
aF-As-F
284.8–325.8
173.26(2)
97.1(3), 98.5(3),
99.3(3)
–
180
101.3
and
a
weakly coordinating anion. The compound
2
[HgXe]2+[b]
261.3
[HgXe]2+·6HF[c]
[(F3As)Au]+ [SbF6]À, which according to structural analysis
contains a gold(i) center that is only complexed on one side,[6]
appears to be a suitable model for the desired educt.
Complexation with xenon could be achieved by exchange of
the only weakly basic [SbF6]À ion for a still weaker basic
anion. Thus the reaction of [(F3As)Au]+ [SbF6]À with xenon in
SbF5-rich HF/SbF5 solution leads to the xenon complex
[(F3As)AuXe]+ [Sb2F11]À (1).
rHg-Xe
rHg···F
276.9(4)
259.4(2), 249.9(2),
246,1(2)
267.2
256.8
245.4(2), 237.3(2),
227.9(2)
75.24(6), 80.49(6),
83.73(7)
236.2
86.7
aXe-Hg···F
131.06(6), 137.57(6),
140.18(6)
134.4
Complex 1 crystallizes as colorless needles which are
stable at room temperature. Crystal-structure analysis shows
that the [(F3As)AuXe]+ and [Sb2F11]À ions in 1 interact only
weakly with each other (Figure 1 and Table 1). The smallest
Au···F separation is 284.8 pm. The As-Au-Xe unit is almost
linear (173.26(2)¤); at 260.72(6) pm the gold(i)–xenon bond in
1 is as short as the gold(iii)–xenon bond in [AuXe2F]2+
[SbF6]À [Sb2F11]À and significantly shorter than gold(ii)–
xenon bonds. With the help of a basis set for gold specially
optimized for the MP2 approximation, calculation of the
structure of the [(F3As)AuXe]+ ion is achieved with astonish-
ing accuracy (Table 1). The calculated bond energy of
32.7 kcalmolÀ1 between [(F3As)Au]+ and Xe agrees well
[a] Total energy, with zero-point energy correction: À455.812215 a.u.
[b] Total energy, with zero-point energy correction: À167.697137 a.u.
[c] Total energy, uncorrected: À769.5592037 a.u.
with the bond energy in the [AuXe]+ ion.[4] Ab initio
calculation also allows the prediction of the frequency of
the gold–xenon vibration n(AuXe) (147.5 cmÀ1). The Raman
À1
~
spectrum shows a band at n = 138.3 cm which doubtlessly
overlaps with the d(SbFSb) band of the anion (Figure 2 and
Table 2).[7] If 1 is labeled with 136Xe (c.f. natural xenon isotope
~
mixture: Mr = 131.3) this band appears displaced at n =
À1
À1
~
136.7 cm . The band at n = 190.7 cm shows the only other
isotope effect (displacement to 189.5 cmÀ1, at the limit of
measurement accuracy). This band is assigned to the gold–
arsenic vibration which, because of the linearity of the As-Au-
Xe framework, is strongly coupled with the gold–xenon
vibration. It is noticeable that the intensity of the n(AuXe)
band is significantly lower than that of the n(AuXe4, A1g) band
in the [AuXe4]2+ ion. The latter band is extraordinarily strong
[*] Prof.Dr.K.Seppelt, Dr.I-.C.Hwang, Dr.S.Seidel
Institut für Chemie
Anorganische und Analytische Chemie
Freie Universität Berlin
Fabeckstrasse 34–36, 14195 Berlin (Germany)
Fax: (+49)30-838-53310
E-mail: seppelt@chemie.fu-berlin.de
and dominates the Raman spectrum of [AuXe4]2+ [Sb2F11]À .
[**] This work was supported by the Deutsche Forschungsgemeinschaft
and the Fonds der Chemischen Industrie.
2
Here too calculation gives an answer: the calculated Raman
4392
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200351208
Angew. Chem. Int. Ed. 2003, 42, 4392 –4395

Angewandte
Chemie
Figure 2. Raman spectrum of 1.For numerical values see the Experi-
mental Section, assignment see Table 2.
Figure 3. 129Xe NMR signal of a solution of 1 in HF/SbF5 at À308C in
Table 2: Experimental and calculated (MP2) vibrational frequen-
cies [cmÀ1] of the [(F3As)AuXe]+ ion in 1.
the presence of xenon.In addition to the sharp signal of elemental
xenon (right) the highly broadened signal for the [(F3As)AuXe]+ ion
arising from quadrupole interactions can be seen.
Raman[a]
MP2[a]
Assignment[b]
764.4(25)
752.6(10)
354.9(4)
747.6(25)
764.2(8)
ns AsF3
ne AsF3
ds AsF3
de AsF3
n AuAs
n AuXe
de AuAsF3
de XeAuAs
by HF/SbF5.[11] In agreement with this observation the first
ionization energy of AsF3 (13.0 eV) is higher by 0.72 eV than
that of PF3 (12.28 eV).[12] These particularities of AsF3 are
usually explained by inadequate shielding by the fully
occupied 3d shell (transition-metal contraction).
349.3 (1.5)
261.9(0.9)
207.9 (0.6)
147.5 (2.3)
138.0 (0.4)
40.2 (0.1)
260.0(sh)[c]
190.7(2)
138.3(50[d])
117.2(3[d])
The Hg2+ ion is isoelectronic with the Au+ ion. However,
compounds that contain uncomplexed Au+ ions have not
been reported, whereas compounds such as HgF2 which
contain uncomplexed Hg2+ ions are known. If HgF2 reacts
with an excess of SbF5 in the presence of xenon above room
temperature highly refractive crystals are formed which
decompose at room temperature in moist air with the release
of gas. In dry air the compound is stable at room temperature,
and under xenon pressure up to 608C. This compound was
identified as [HgXe]2+ [SbF6]À [Sb2F11]À (2, see Figure 4) by
[a] Values in parenthesis are relative intensities.[b] Strongest compo-
nents.[c] sh =shoulder.[d] Superimposed with d(SbFSb) of [Sb2F11]À
ions.
intensity of the n(AuXe4, A1g) band is greater than that of the
n(AuXe) band in the [(F3As)AuXe]+ ion by a factor of 25.
In the 129Xe NMR spectrum (129Xe: I = 1= , natural
2
abundance 26.4%), because of the xenon excess necessary
in the solution, a signal for the [(Fe3As)AuXe]+ ion and one
for atomic xenon are expected. However, at room temper-
ature only one signal at d = À5180.7 ppm is observed; clearly
rapid exchange is taking place. On cooling the sample to
À30¤C this signal shifts and splits (Figure 3). The complicated
appearance of the signal can be explained by 197Au–129Xe
(
197Au: I = 3= , 100%) and 75As–129Xe coupling (75As: I = 3= ,
2 2
100%). The large quadrupole moment of one or both I = 3=
2
nuclei leads to extensive broadening of the signals; the 129Xe
NMR signal of the [XeCl]+ ion (35Cl/37Cl, I = 3= ) has a similar
2
shape.[8] In addition in the spectrum at À308C the anticipated
sharp signal for free xenon is also observed.
AsF3 is clearly such a poor ligand that xenon can compete
with it, even when AsF3 is present in excess. In contrast the
[(F3P)2Au]+ ion is formed immediately and irreversibly with
the homologue of AsF3, PF3, under identical conditions.[6]
Apart from the [(F3As)Au]+ and [(F3As)AuXe]+ ions we are
unaware of any arsenic-bonded AsF3 complexes, whereas a
large number of PF3 complexes are known.[9]) In HF/AsF5 the
AsF3 unit binds as a ligand to trivalent lanthanide cations, but
in these species coordination takes place through a fluorine
atom.[10] In contrast to PF3, however, AsF3 is not protonated
Figure 4. Crystal structure 2 (ORTEP representation, thermal ellipsoids
set at 50% probability).The immediate environment of an [HgXe] 2+
2+
ion is shown.The coordination sphere of the Hg ion has a highly
distorted, capped octahedral configuration.
Angew. Chem. Int. Ed. 2003, 42, 4392 –4395
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4393

Communications
Crystal structure analysis: a suitable crystal was mounted onto a
specially constructed apparatus[13] with cooling in an inert atmosphere
on a Bruker SMART CCD1000 TM diffractometer and analyzed.[14]
After semi-empirical absorption correction by equalization of like-
symmetry reflections (SADABS), structure solution and refinement
was carried out with the SHELX programs.[15,16]
crystal-structure analysis. At 276.93(4) pm the mercury–
xenon bond is about as long as the gold(ii)–xenon bond in
the [AuXe4]2+ ion. The coordination sphere of the mercury
center is completed by six fluorine atoms at distances between
227.9(2) and 259.4(2) pm. Compound 2 has a highly distorted
capped structure in which the xenon atom assumes the
capping position. For calculation on 2 with ab initio methods
we replaced the fluoroantimonate anions with HF molecules
and assumed a regular, capped octahedral structure. The
calculation gave a mercury–xenon separation of 267.2 pm and
mercury–fluorine separations of 236.2 and 256.8 pm (see
Table 1). In view of the simplifications the agreement with
experiment is satisfactory in our view. The calculated
mercury–xenon bond energy for [HgXe]2+·6HF is
24.5 kcalmolÀ1.
Ab initio calculations: Gaussian98 program.[17] 2nd-order Møl-
ler–Plesset approximation (MP2) for electron correlation as imple-
mented in Gaussian. Basis set: 6-311G(d,p) for F and
H as
implemented in Gaussian. Au and Hg, relativistically corrected
basis set for MP2, optimized by P. Schwerdtfeger and R. Wesendrup,
in each case 9s 9p 6d 4f with pseudopotentials for 60 core electrons.
Xe 6s, 6p, 3d, 1f with relativistically corrected pseudopotentials for
46 core electrons. As: 3s, 3p, 1d with pseudopotentials for 28 core
electrons.[18]
2: HgF2 (130 mg, 0.54 mmol) and SbF5 (4.06 g, 18.73 mmol) were
placed into a PFA reaction tube under exclusion of moisture. HF
(280 mg, 14.0 mmol) was condensed onto this mixture with a stainless
steel apparatus at À1968C. On warming to room temperature a
colorless solid formed which remained partly in solution. The mixture
was then evaporated to dryness in vacuum at room temperature.
10 mg of the colorless residue was placed into a thick-walled glass
ampoule of 3-mm inner diameter and treated with SbF5 (660 mg,
3.0 mmol). Liquid xenon (220 mL) was condensed into the mixture
and the tube was sealed by melting. On warming to room temperature
the SbF5 and xenon mixed. The reaction mixture was homogenized in
an ultrasound bath and then heated at 808C for 6 h. Slow cooling to
room temperature (0.058CminÀ1) afforded colorless platelet crystals
of 2. The pure, colorless product was obtained by decantation of the
xenon/SbF5 excess at low temperature (À1008C); m.p. ꢀ 308C
(decomp.). The reaction is quantitative. Raman (1064 nm, 258C):
Neither 129Xe nor 199Hg NMR spectra could be obtained
for 2 because the compound decomposed in HF solution. The
Raman spectrum shows
a relatively intense band at
139.8 cmÀ1 (138.7 cmÀ1 for [Hg136Xe]2+) which can be assigned
to the mercury–xenon vibration and again is superimposed
with the d(SbFSb) band of the [Sb2F11]À ion. Only the anion
combination [SbF6]À/[Sb2F11]À was detected following reac-
tions under a number of conditions. The Raman spectrum of a
sample of crystals, several of which were studied crystallo-
graphically, is identical to the spectrum of the bulk product.
Consequently the reaction yields quantitatively a uniform
product.
~
n(Irel) = 714(75), 703(20), 687(10), 673(18), 652(100), 640(35), 594(3),
Finally, reference is made to a peculiarity of this mercury–
xenon compound. All noble-gas compounds, including the
novel gold–xenon compounds, require elemental fluorine in
at least one step of their synthesis (e.g. for the synthesis
AuF3). Compound 2 is the first noble-gas compound whose
synthesis occurs without the use of elemental fluorine: SbF5
and HgF2 are obtainable by metathesis reactions. Our results
show that a number of xenon–metal complexes could be
prepared if it is possible to obtain educts with sufficiently
weakly coordinated metals ions.
572(5), 553(15), 525(10), 385(3), 341(5), 329(9), 306(8), 289(35),
264(20), 227(30), 209(7), 194(3), 181(2), 140(40), 136(30),
119(10) cmÀ1. Crystal structure analysis was carried out as described
above.[19]
Further details on the crystal structure investigations may be
obtained from the Fachinformationszentrum Karlsruhe, 76344,
Eggenstein-Leopoldshafen, Germany (fax: (+ 49)7247-808-666; e-
mail: crysdata@fiz-karlsruhe.de) on quoting the depository numbers
CSD-412994 (1) and CSD-412993 (2)
Received: February 18, 2003 [Z51208]
Keywords: gold · mercury · noble gases · superacid systems
.
Experimental Section
1: Under the exclusion of moisture, AuF3 (210 mg, 0.8 mmol) and
SbF5 (2.29 g, 10.6 mmol) were placed into a polyperfluorovinyl ether/
tetrafluoroethylene copolymer (PFA) reaction tube. Anhydrous HF
(800 mg, 40 mmol) and AsF3 (350mg, 2.6 mmol) were condensed into
the tube with a stainless steel vacuum apparatus at À1968C. Warming
to room temperature led to gas evolution (AsF5) and the formation of
a small amount of black precipitate, probably gold. When the gas
evolution ceased xenon (800 mg, 6 mmol) was condensed into the
mixture at À1968C. The reaction tube was sealed and the reaction
mixture cooled from room temperature to À508C. Compound 1
crystallized out as colorless needles which immediately turned black
on exposure to air. Yield: 750 mg, m.p. 628C (decomp.). Removal of
the adhering HF/SbF5 in vacuum leads to decomposition of the
compound.
[1] S. Seidel, K. Seppelt, Science 2000, 290, 117 – 118.
[2] T. Drews, S. Seidel, K. Seppelt, Angew. Chem. 2002, 114, 470 –
473; Angew. Chem. Int. Ed. 2002, 41, 454 – 456.
[3] P. Pyykkö, Chem. Rev. 1988, 88, 563 – 594.
[4] D. Schröder, H. Schwarz, J. Hrusak, P. Pyykkö, Inorg. Chem.
1998, 37, 624 – 632.
[5] P. Pyykkö, J. Am. Chem. Soc. 1995, 117, 2067 – 2070.
[6] R. Küster, K. Seppelt, Z. Anorg. Allg. Chem. 2000, 626, 236 –
240.
[7] H. Willner, M. Bodenbinder, R. Bröchler, G. Hwang, S. J. Rettig,
J. Trotter, B. von Ahsen, U. Westphal, V. Jonas, W. Thiel, F.
Aubke, J. Am. Chem. Soc. 2001, 123, 588 – 602.
[8] S. Seidel, K. Seppelt, Angew. Chem. 2001, 113, 4318 – 4320;
Angew. Chem. Int. Ed. 2001, 40, 4225 – 4227.
[9] T. Kruck, Angew. Chem. 1967, 79, 27 – 43; Angew. Chem. Int. Ed.
Engl. 1967, 6, 53 – 67.
[10] M. Tramsek, P. Benkic, A. Turicnik, G. Tavcar, B. Zemva, J.
Fluorine Chem. 2002, 114, 143 – 148.
129Xe NMR (HF/SbF5, 110.45 MHz, XeOF4, 258C): d =
À5180.7 ppm (s), 129Xe NMR (HF/SbF5, 110.45 MHz, XeOF4,
~
À308C): 5149.9–5150.3 ppm (m); Raman (1064 nm, 258C): n(Irel) =
764.4(25), 752.6(10), 676.5(sh), 666.3(60), 657.7(sh), 646.0(100),
591.0(10), 354.9(4), 316.9(6), 293.3(15), 282.9(sh), 260.0(sh),
[11] R. Küster, T. Drews, K. Seppelt, Inorg. Chem. 2000, 39, 2784 –
2786.
226.5(15), 190.7(2), 151.8(3), 138.3(50), 117.2(2) cmÀ1
.
4394
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 4392 –4395

Angewandte
Chemie
[12] W. A. Potts, H. J. Lempka, D. G. Streets, W. C. Price, Philos.
Trans. R. Soc. London Ser. A 1970, 268, 59 – 76.
[13] T. Kottge, D. Stalke, J. Appl. Crystallogr. 1993, 26, 615 – 619.
[14] 1: orthorhombic, Pna21, a = 1588.7(2), b = 813.3(1), c =
1094.5(2) pm, V= 1414 ” 106 pm3, T= À1008C, Z = 4, MoKa
radiation, graphite monochromator, scan with 0.38 in w, 10 s
irradiation time per measurement, 1800 measurements, 17436
measured reflections, 4316 independent reflections, 173 param-
eters, R(F ꢁ 4s(F)) = 0.0275, wR2 = 0.065.
[15] G. Sheldrick, Program for Crystal Structure Solution, Universi-
tät Göttingen, 1986.
[16] G. Sheldrick, Program for Crystal Structure Refinement, Uni-
versität Göttingen, 1993.
[17] Gaussian98 (RevisionA.7), M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G.
Zakrzewski, J. A. Montgomery, 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. 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, Gaussian, Inc., Pittsburgh, PA, 1998.
[18] Institut für Theoretische Chemie der Universität Stuttgart,
http://www.theochem.uni-stuttgart.de, pseudopotentials, elec-
tron core potentials.
[19] 2: orthorhombic, P21/c, a = 1372.2(3), b = 1048.1(1), c =
1052.5(1) pm, b = 94.56(1)8, V= 1508.8 ” 106 pm3, T= À1008C,
Z = 4, MoKa radiation, graphite monochromator, scan width 0.38
in w, 10 s irradiation time per measurement, 1800 measurements,
24460 measured reflections, 4583 independent reflections, 200
parameters, R (F ꢁ 4s(F)) = 0.0205, wR2 = 0.0427.
Angew. Chem. Int. Ed. 2003, 42, 4392 –4395
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4395