Synthesis and Catalytic Use of Gold(I) Complexes Containing a Hemilabile Phosphanylferrocene Nitrile Donor

Article abstract of DOI:10.1002/chem.201502968

Removal of the chloride ligand from [AuCl(1-κP)] (2) containing a P-monodentate 1′-(diphenylphosphanyl)-1-cyanoferrocene ligand (1), by using silver(I) salts affords cationic complexes of the type [Au(1)]X, which exist either as cyclic dimers [Au(1)]₂X₂ (3a, X=SbF₆; 3 c, X=NTf₂) or linear coordination polymers [Au(1)]_nX_n (3 a′, X=SbF₆; 3 b′, X=ClO₄), depending on anion X and the isolation procedure. As demonstrated for 3 a′, the polymers can be readily cleaved by the addition of donors, such as Cl^-, tetrahydrothiophene (tht) or 1, giving rise to the parent compound 2, [Au(tht)(1-κP)][SbF₆] (5 a) or [Au(1-κP)₂][SbF₆] (4 a), respectively, of which the last two compounds can also be prepared by stepwise replacement of tht in [Au(1-κP)₂][SbF₆]. The particular combination of a firmly coordinated (phosphane) and a dissociable (nitrile) donor moieties renders complexes 3/3′ attractive for catalysis because they can serve as shelf-stable precursors of coordinatively unsaturated Au^I fragments, analogous to those that result from the widely used [Au(PR₃)(RCN)]X catalysts. The catalytic properties of the Au-1 complexes were evaluated in model annulation reactions, such as the synthesis of 2,3-dimethylfuran from (Z)-3-methylpent-2-en-4-yn-1-ol and oxidative cyclisation of alkynes with nitriles to produce 2,5-disubstituted 1,3-oxazoles. Of the compounds tested (2, 3 a′, 3 b′, 3 a, 4 a and 5 a), the best results were consistently achieved with dimer 3 c, which has good solubility in organic solvents and only one firmly bound donor at the gold atom. This compound was advantageously used in the key steps of annuloline and rosefuran syntheses.

Full text of DOI:10.1002/chem.201502968

DOI: 10.1002/chem.201502968
Full Paper
&
Homogeneous Catalysis |Hot Paper|
Synthesis and Catalytic Use of Gold(I) Complexes Containing
a Hemilabile Phosphanylferrocene Nitrile Donor
[a]
ˇ
ˇ
ˇ
ˇ
ˇ
Karel Skoch, Ivana Císarovµ, and Petr Stepnicka*
Abstract: Removal of the chloride ligand from [AuCl(1-kP)]
(2) containing a P-monodentate 1’-(diphenylphosphanyl)-1-
cyanoferrocene ligand (1), by using silver(I) salts affords cat-
ionic complexes of the type [Au(1)]X, which exist either as
cyclic dimers [Au(1)]2X2 (3a, X=SbF6; 3c, X=NTf2) or linear
coordination polymers [Au(1)]nXn (3a’, X=SbF6; 3b’, X=
ClO4), depending on anion X and the isolation procedure. As
demonstrated for 3a’, the polymers can be readily cleaved
by the addition of donors, such as ClÀ, tetrahydrothiophene
(tht) or 1, giving rise to the parent compound 2, [Au(tht)(1-
kP)][SbF6] (5a) or [Au(1-kP)2][SbF6] (4a), respectively, of
which the last two compounds can also be prepared by
stepwise replacement of tht in [Au(1-kP)2][SbF6]. The particu-
lar combination of a firmly coordinated (phosphane) and
a dissociable (nitrile) donor moieties renders complexes 3/3’
attractive for catalysis because they can serve as shelf-stable
precursors of coordinatively unsaturated AuI fragments,
analogous to those that result from the widely used
[Au(PR3)(RCN)]X catalysts. The catalytic properties of the Au-
1 complexes were evaluated in model annulation reactions,
such as the synthesis of 2,3-dimethylfuran from (Z)-3-methyl-
pent-2-en-4-yn-1-ol and oxidative cyclisation of alkynes with
nitriles to produce 2,5-disubstituted 1,3-oxazoles. Of the
compounds tested (2, 3a’, 3b’, 3a, 4a and 5a), the best re-
sults were consistently achieved with dimer 3c, which has
good solubility in organic solvents and only one firmly
bound donor at the gold atom. This compound was advan-
tageously used in the key steps of annuloline and rosefuran
syntheses.
Introduction
Interest in the coordination chemistry[1] of gold has been re-
cently revived, primarily because of rapid developments in the
field of homogeneous gold-catalysed reactions.[2] Compounds
that are typically employed as catalysts (or catalyst precursors)
in gold catalysis are simple AuI/III salts,[2,3] gold–carbene com-
plexes[2,4] and, mainly, stable AuI phosphane complexes of the
type [AuCl(PR3)], which are typically activated in situ by the re-
moval of the metal-bound halide with AgI salts.[2] However, the
latter approach can result in the formation of Au–Ag bimetallic
systems, the reactivity of which may differ from that of the cor-
responding Au-only catalyst. This so-called silver effect in gold
catalysis has stirred up a vigorous debate[5] and has also
prompted a search for defined, silver-free AuI catalysts.[6] In ad-
À
Scheme 1.
dition to the very popular use of the solubilizing NTf2 coun-
terion,[7] the most successful of these newly introduced com-
pounds appear to be cationic complexes of the type
[Au(PR3)(RCN)]+ with easily dissociated nitrile ligands (e.g.,
compounds A and B in Scheme 1),[8,9] which presumably serve
as precursors for the catalytically active (R3P)Au+ species.
Recently, we synthesised 1’-(diphenylphosphanyl)-1-cyano-
ferrocene (1 in Scheme 1),[10] which can be regarded as
a donor-asymmetric[11] analogue of the ubiquitous 1,1’-bis(di-
phenylphosphanyl)ferrocene (dppf).[12] In view of the unexpect-
edly versatile coordination behaviour of 1 towards CuI,[10] we
decided to study the interactions of this ligand with AuI, the
softest Group 11 metal ion.[13] Herein, we describe the synthesis
of structurally unique AuI-1 complexes (Scheme 1, bottom) and
report on their catalytic applications in selected AuI-mediated
organic reactions.
ˇ
ˇ
ˇ
ˇ
ˇ
[a] K. Skoch, Dr. I. Císarovµ, Prof. Dr. P. Stepnicka
Department of Inorganic Chemistry
Faculty of Science, Charles University in Prague
Hlavova 2030, 12840 Prague (Czech Republic)
E-mail: stepnic@natur.cuni.cz
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201502968.
Chem. Eur. J. 2015, 21, 15998 – 16004
15998
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
is only 1.0(2)8) and assume a synclinal eclipsed conformation
with a torsion angle of 71.3(2)8 for C1-Cg1-Cg2-C6 (t; see
Ref. [12a]; note that Cg1 and Cg2 are the centroids of the
cyclopentadienyl rings C(1–5) and C(6–10), respectively). No in-
teractions between the gold atom and the nitrile groups[16] or
aurophilic contacts[17] were detected in the solid-state structure
of 2.
Results and Discussion
Synthesis of the AuI complexes with ligand 1
The syntheses and mutual interconversions of AuI complexes
with phosphanylnitrile
1 as a ligand are illustrated in
Scheme 2. Ligand 1 reacts cleanly and rapidly with [AuCl(tht)]
Complex 2 reacts smoothly with silver(I) salts to give the
corresponding cationic complexes with the general formula
[Au(1)]nXn (Scheme 2).[18] Depending on anion X and the isola-
tion procedure (additives) used, these compounds are isolated
either as symmetric dimers, in which the phosphanylnitrile
connects two gold centres as a P,N-bridge (3), or as coordin-
ation polymers, in which the ligands play a similar role albeit
in an linearly propagating chain (3’). For example, the reaction
of 2 with Ag[SbF6] gives polymeric [Au(1)]n[SbF6]n (3a’). Analo-
gous perchlorate salt 3b’, obtained in a similar manner, is
rather unstable and cannot be crystallised because it readily
decomposes. However, the insolubility of 3b’ in common solv-
ents attests to a similar polymeric structure. In contrast, the re-
action between 2 and AgNTf2 reproducibly affords the reason-
ably soluble dimer [Au(1)]2(NTf2)2 (3c).
The preferred formation of only one type of product under
analogous conditions appears to be controlled by an interplay
of the relative solubility of the hypothetical Au(1)X fragments
(as such or solvated) and their overall crystallisation properties.
The distinct influence of the synthesis conditions on the aggre-
Scheme 2. Synthesis and mutual conversions of AuI complexes with phos-
phanylnitrile 1 (tht=tetrahydrothiophene).
n+
(tht=tetrahydrothiophene) to afford the expected phosphane
gation state of the [Au(1)]n species can be further highlight-
complex [AuCl(1-kP)] (2).[14] In the H and 13C NMR spectra of 2,
ed by the serendipitous isolation of complex 3a,[19] an isomer
to 3a’, in which the structure of the dimeric [Au2(1)2]2+ motif is
associated with two [SbF6]À anions.[20]
1
there are characteristic signals assigned to the phosphanylfer-
rocene ligand, whereas the 31P NMR spectrum displays a res-
onance at dP = +28.1 ppm. The crystal structure of 2 (Figure 1)
reveals the typical linear coordination around the AuI centre.[15]
The ferrocene cyclopentadienyls in P-coordinated 1 are negli-
gibly tilted (the dihedral angle of the cyclopentadienyl planes
Importantly, the reaction that leads to compounds 3 can be
easily reversed by the addition of [Bu4N]Cl as a chloride source
(as demonstrated for 3a’, see Scheme 2). The cleavage of the
multi-gold assemblies can be also achieved by the addition of
other donors, such as 1, tht or even a donor solvent (e.g.,
MeCN). Thus, polymer 3a’ readily dissolves upon the addition
of 1 to afford the monogold(I) species [Au(1-kP)2][SbF6] (4a), in
which the two phosphanylnitrile ligands coordinate as equiva-
lent P-monodentate donors.[21] The same product can be pre-
pared directly by the treatment of [Au(tht)2][SbF6] with two
equivalents of 1. A similar reaction at the Au/1 molar ratio of
1:1 provides a product with an intermediate level of substitu-
tion, [Au(1-kP)(tht)]SbF6 (5a), which can be converted to 4a by
the addition of another equivalent of 1. Complex 5a also re-
sults from cleavage of polymer 3a’ with tht and can be trans-
formed back to the parent complex 2 upon treatment with
[Bu4N]Cl (Scheme 2).
The crystal structures of 3a, 3a’·Me2CO,[22] 3c and 4a were
determined by using X-ray diffraction analysis and are present-
ed in Figure 2 and in the Supporting Information.[23,49] Selected
geometric parameters are given in Table 1. The AuÀP bond
lengths in these compounds do not differ greatly from those
of parent complex 2. A slight yet statistically significant elonga-
tion of the AuÀP bonds in 4a (compared with complexes 3/3’)
can be attributed to steric repulsion of the proximal phos-
phane moieties. The variation in the lengths of the CꢀN bonds
Figure 1. View of the molecular structure of chlorogold(I) complex 2. Select-
ed bond lengths [Š] and angles [8]: AuÀP 2.2287(6), AuÀCl 2.2891(7), C11ÀN
1.144(4), FeÀCg1 1.639(1), FeÀCg2 1.637(1), P-Au-Cl 176.25(2), C1-C11-N
178.8(3).
Chem. Eur. J. 2015, 21, 15998 – 16004
www.chemeurj.org
15999
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
AuÀP bonds in 4a compare well with the data determined for
complexes [Au(PR3)2]X, in which R/X=Me/PF6,[27] Ph/NTf2,[28] Ph/
NO3[29] and FcCH2PPh2/ClO4 (Fc=ferrocenyl).[30]
The conformations of the ferrocene units in complexes 3/3’
are all nearly synclinal eclipsed (ideal value: 728). One of the
two structurally independent molecules of ligand 1 in the
structure of 4a has a similarly compact conformation (Fe2),
whereas the other ligand molecule adopts an anticlinal
eclipsed conformation, which renders the donor substituents
at the ferrocene unit more distant (Fe1).
Catalytic evaluation of the Au-1 complexes
The mutual interconversions of the AuI complexes with ligand
2 described above clearly demonstrate the hemilabile nature[11]
of the cationic Au-1 species, which results from different
strengths of the Au–donor bonds. Apparently, the phosphane
donor moiety acts as a firmly bound pivot in these com-
pounds, whereas the CNÀAu bond can be readily cleaved by
neutral and anionic donors. Such a facile splitting of the
parent structure to provide coordinatively unsaturated frag-
ments, and their possible reassembly to allow for self-stabilisa-
tion of these intermediates, renders these compounds attrac-
tive for use in catalysis.
Figure 2. View of the cations in the crystal structures of 3a, 3a’·Me2CO and
4a. For the conventional displacement ellipsoid plots and structural drawing
of 3c, see the Supporting Information.
Catalytic properties of the Au-1 complexes were evaluated
with some known ring-forming reactions;[31] first, in the cyclisa-
tion of (Z)-3-methylpent-2-en-4-yn-1-ol ((Z)-6) to 2,3-dimethyl-
furan (7 in Scheme 3). In general, this reaction and similar
Table 1. Comparison of the pertinent geometric parameters of 3a,
3a’·Me2CO, 3c and 4a.[a]
3a
3a’·Me2CO
3c
4a[b]
bond lengths [Š]
AuÀP
2.225(2)
2.2246(9)
2.028(3)
1.143(5)
1.651(2)
1.644(2)
2.232(1)
2.035(3)
1.142(5)
1.645(2)
1.650(2)
2.3104(8)/2.3140(8)
–
1.148(4)/1.136(6)
1.649(1)/1.645(2)
1.648(2)/1.643(1)
AuÀN
NꢀC
2.035(4)
1.139(8)
1.645(3)
1.642(3)
FeÀCg1
FeÀCg2
angles [8]
PÀAuÀN
AuÀNꢀC
tilt
Scheme 3. Gold-catalysed cycloisomerisation of (Z)-6 to 2,3-dimethyl-furan
(7).
175.1(1)
168.2(5)
2.9(4)
À60.6(5)
179.4(1)
173.9(3)
3.3(2)
À66.8(3)
173.4(1)
168.0(4)
3.4(3)
À78.2(3)
175.43(2)[c]
–
3.6(2)/4.3(2)
À142.0(2)/75.5(2)
transformations represent an attractive route to furan deriva-
tives, and although a vast number of transition-metal com-
pounds have been tested in this area,[32] applications of AuI
catalysts to this particular cyclisation of 2-en-4-yn-1-ols still
remain quite rare.[33,34]
t
[a] Cg1 and Cg2 are the centroids of cyclopentadienyl rings C1–5 and
C6–10, respectively; tilt is the dihedral angle of the cyclopentadienyl
planes; t is the torsion angle of C1-Cg1-Cg2-C6. [b] Data for ligand
1 (Fe1)/ligand 2 (Fe2). [c] P1-Au-P2 angle.
The results obtained with the AuI-1 complexes (Table 2) indi-
cate a superior performance of the 3-type compounds, which
achieve full conversions of (Z)-6 to 7 at catalyst loadings as
low as 0.01%. Even at this scale, the reactions quickly reach
completion (being typically complete within less than 5 min)
and are strongly exothermic, which becomes evident when the
experiments are performed without any solvent and on
a larger scale. The best results were obtained with complex 3c,
the solubility of which ensures rapid and complete dissolution
of the catalyst in the reaction mixture. However, compounds
with two strongly coordinated ligands (phosphane and chlor-
ide in 2 and 4a), and 5a, proved to be less efficient, which
became particularly evident at low metal loadings.[35]
is only marginal (both in the series and with respect to uncoor-
dinated 1[10]), which indicates that the bonding and back-
bonding components of the AuÀNC dative bond counteract
each other, and thus result in marginal changes in the bond
order. This corresponds to a decrease in the ratio between the
p-acceptor and s-donor abilities of the nitrile donors with re-
spect to, for example, the isonitrile and CO ligands.[24] Overall,
the Au–donor separations and the interligand angles in com-
pounds 3 and 3’ are similar to those reported previously for
[Ph3PAu(NCMe)][SbF6][25] and similar complexes with 2-phos-
phanylbiphenyl ligands (type B in Scheme 1),[25,26] while the
Chem. Eur. J. 2015, 21, 15998 – 16004
www.chemeurj.org
16000
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ments were carried out with the reaction between acetonitrile
and phenylethyne to provide 2-methyl-5-phenyl-1,3-oxazole
(12a in Scheme 5; the crystal structure of 12a is presented in
the Supporting Information).
Table 2. Summary of the catalysis results achieved with AuI-1 complexes
in the cyclisation of (Z)-3-methylpent-2-en-4-yn-1-ol.[a]
Catalyst
Au loading [%]
Yield [%][b]
[AuCl(tht)]
2
3a’
0.1
0.1
0.1
0.01
0.1
0.01
0.1
0.01
0.1
0.1
0.01
0
82
65
quant.
45
quant.
22
98
quant.
0
quant.
17
0
3a’
3b’
3b’
3c
3c
4a
5a
5a
none
Scheme 5. The model Au-catalysed oxidative cyclisation of terminal alkynes
with nitriles to give 1,3-oxazoles and structures of the N-oxides employed in
this reaction.
[a] Conditions: reaction in CHCl3 at RT for 30 min. The yields are the aver-
age of two independent runs and are given relative to the major (Z)-
isomer of the starting enynol ((Z)/(E) ꢁ90:10). [b] Yield determined by
using NMR spectroscopy.
The results (Table 3) clearly differentiated the catalysts.
Whereas the coordinatively saturated complexes 2, 4a and 5a,
as well the precursor [AuCl(tht)], provided 12a with only poor
These promising results led us to demonstrate the useful-
ness of the Au-1 catalysts under practically relevant conditions.
When the cyclisation reaction was carried out with 3c
(0.01 mol%) in the absence of any solvent at a 50 mmol scale
(under ambient conditions for 30 min[36]), it afforded furan 7 in
an isolated yield of 92% after simple distillation.[37] The turn-
over frequency (TOF) for catalyst 3c used in this reaction was
Table 3. Summary of catalysis results obtained with various AuI catalysts
in the model reaction to give oxazole 12a.[a]
Catalyst
Yield with
Yield with
N-oxide 13 [%]
N-oxide 14 [%]
[AuCl(tht)]
2
7
ꢁ1.5
50
33
78
n.a.
n.a.
83
33
88
3a’
3b’
3c
4a
5a
as high as 2”105 hÀ1 [38]
Unfortunately, a further reduction of
.
the catalyst loading to 0.001 mol% markedly decreased the
conversion (only ꢁ35% 7 was formed at 808C over 72 h).
A similar annulation of (Z)-3,7-dimethyl-2,6-octadien-4-yn-1-
ol (8) to give 3-methyl-2-(3-methylbut-2-en-1-yl)furan or rose-
furan (9 in Scheme 4),[39] which is a constituent of natural es-
12
44
n.a.
n.a.
[a] Conditions: phenyl acetylene (0.250 mmol) and N-oxide (0.325 mmol,
1.3 equiv) were reacted in the presence of the Au catalyst (5 mol%) in
acetonitrile (2.5 mL) at 608C for 24 h. The isolated yields are given as the
average of two independent runs; n.a.=not available.
yields, compounds 3 performed much better. Similar to the
previous tests, the best results were obtained with dimer 3c,
which afforded 12a in an isolated yield of 78%. A further in-
crease in the yield, though not for all of the complexes (see
also the results for practically insoluble 3b’ in Table 3), could
be achieved by replacing N-oxide 13 with its more bulky coun-
terpart 14.[42]
The reactions performed next with different substrates
(Scheme 6, data in Table 4) demonstrated that the cyclisation
of ring-substituted phenylacetylenes with 3c and 14 in aceto-
nitrile gives the respective 2-methyloxazoles in very good iso-
lated yields. A similar result was attained with propionitrile, but
Scheme 4. Gold-catalysed cycloisomerisation of 8 to rosefuran (9).
sential oils,[40] required more forcing conditions, most likely be-
cause of the lower reactivity of the internal triple bond present
in the substrate. For example, no cyclisation product was de-
tected when neat enynol 8 was treated with 3c (0.1 mol%) at
room temperature for 20 h, whereas heating the reaction mix-
ture to 808C for 40 h resulted in only a 4% conversion. On in-
creasing the catalyst amount to 0.5 mol%, however, the reac-
tion proceeded with complete conversion within 2 h at 608C
and gave pure rosefuran in 91% yield after column chroma-
tography.
In a continuation of our catalytic tests, we turned to the syn-
thesis of 1,3-oxazoles[41] by an Au-mediated oxidative cyclisa-
tion of alkynes with nitriles in the presence of N-heterocyclic
N-oxides,[42] which offers an attractive alternative to conven-
tional synthetic approaches.[43] The initial screening experi-
Scheme 6. Au-catalysed synthesis of 2,5-disubstituted 1,3-oxazoles.
Chem. Eur. J. 2015, 21, 15998 – 16004
www.chemeurj.org
16001
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
608C for 24 h) and provided analytically pure 15 in 63% yield
after column chromatography, during which the majority of
the unreacted nitrile could also be recovered (2.1 equiv of 18
was isolated). On the whole, this four-step synthesis represents
a high-yield and convergent approach towards annuloline
(ꢁ34% with respect to 16) that obviates the use of advanced
and/or expensive starting materials and reagents and tedious
experimentation, and even avoids unwanted isomerisation at
the double bond in the styryl moiety. As such, it represents
a practical alternative to the methods reported earlier.[45b,c,46]
In addition to the cyclisation route presented above, another
approach to 15 has been investigated based on the Heck
coupling of the respective 2-vinyl-1,3-oxazole and 16.[47d] In
a pilot experiment, oxazole 12 f was employed as a model sub-
strate and was treated with 16 in the presence of a Pd catalyst
under conventional Heck conditions. However, the reaction did
not proceed to any appreciable extent, leading instead to
a complete decomposition of the oxazole, whereas 16 re-
mained unchanged.[47]
Table 4. The substrate scope tests for the reaction to give oxazoles 12.[a]
Nitrile
Alkyne
Product
Yield [%]
MeCN (10a)
10a
10a
10a
10a
EtCN (10 f)
CH2=CHCN (10g)
PhCN (10h)[b]
C6H5CꢀCH (11 a)
4-MeC6H4CꢀCH (11 b)
4-MeOC6H4CꢀCH (11 c)
4-CF3C6H4CꢀCH (11 d)
4-BrC6H4CꢀCH (11 e)
11 a
12a
12b
12c
12d
12e
12 f
12g
12h
88
92
92
72
82
85
46
73
11 a
11 a
[a] Conditions: alkyne, catalyst 3c (5 mol%) and 14 (1.3 equiv) were react-
ed in neat nitrile at 608C for 24 h unless specified otherwise. The isolated
yields are given as the average of two independents experiments. Note:
The first entry is repeated from Table 3 for a comparison. [b] Reaction
with the nitrile (6 equiv) in chlorobenzene (2 mL).
the reaction with the generally more reactive acrylonitrile fur-
nished 12g in only 46% yield.
Encouraged by the successful screening experiments, we set
out to employ this [2+2+1] annulation in the preparation of
Conclusion
a
naturally occurring oxazole alkaloid annuloline (15;
Scheme 7).[44,45] The nitrile required for this cyclisation, (2E)-3-
(3,4-dimethoxyphenyl)-2-propenenitrile (18), was obtained in
Abstraction of the chloride ligand from [AuCl(1-kP)] (2) gives
rise to structurally remarkable cationic complexes [Au(1)]nXn,
the degree of association (n) of which in the solid state (dimer
vs. polymer) can be controlled by the counterion X, presume-
ably through modulation of the solubility and crystallisation
properties of plausible “Au(1)X” intermediates. In these com-
pounds, the structurally flexible 1’-(diphenylphosphanyl)-1-
cyanoferrocene (1) behaves as a bridging hemilabile donor,
which makes use of both of its donor moieties.[48] However,
the relatively weaker coordination of the nitrile groups allows
for an easy disaggregation of these multinuclear compounds
upon the addition of donors and thus makes them an attract-
ive and practical source of coordinatively unsaturated AuI spe-
cies that are potentially capable of self-stabilisation through
equilibria between the mononuclear (solvated) fragments and
their aggregated form. The fact that the [Au(1)]nXn complexes
can indeed serve as a reservoir of catalytically active, low-nu-
clear gold species was established for selected organic
transformations. The catalytic experiments revealed a consist-
ently superior performance of the most soluble derivative, 3c,
Scheme 7. Synthesis of annuloline 15 by Pd-catalysed cross-coupling and
Au-mediated cyclisation.
À
in which the dimeric motif {Au2(1)2}2+ pairs with the NTf2
two steps through a Pd-catalysed Heck coupling of 4-bromo-
veratrole (16) with acrylonitrile and subsequent dehydration of
formed amide (E)-17. The dehydration was associated with
a partial isomerisation at the double bond, which led to an ap-
proximately 90:10 mixture of the (E) and (Z) isomers; however,
the latter isomer could be easily removed by a single recrystal-
lisation from ethyl acetate/heptane. It is notable that the Heck
coupling of 16 with acrylonitrile, which could be suggested as
a direct route to nitrile 18, proved to be less practical due to
anions. This compound, in particular, emerges as an attractive
(shelf-stable and well-defined) catalyst for gold-mediated or-
ganic reactions, which still rely predominantly on ill-defined
species generated in situ from chlorogold(I) complexes with
various supporting ligands and silver(I) salts or on the relatively
unstable cationic B-type complexes.
its lower selectivity (a ꢁ2:1 mixture of (E)- and (Z)-18 was ob- Acknowledgements
tained under otherwise similar conditions).
The subsequent cyclisation of 18 with 11 c and N-oxide 14
was performed in chlorobenzene with three molar equivalents
of the nitrile with respect to 11 c and 5 mol% of 3c as the
catalyst. Gratifyingly, the reaction proceeded smoothly (at
The results presented in this paper were obtained with finan-
cial support from the Czech Science Foundation (project no.
13-08890S) and the Grant Agency of Charles University in
Prague (project no. 108213).
Chem. Eur. J. 2015, 21, 15998 – 16004
www.chemeurj.org
16002
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[16] The shortest intermolecular Au···N distance in the structure of 1 of
3.323(2) Š is found for molecules related by elemental translation along
the crystallographic a axis.
Keywords: gold · homogeneous catalysis · metallocenes ·
phosphane ligands · structure elucidation
[17] Selected reviews: a) H. Schmidbaur, A. Schier, Chem. Soc. Rev. 2012, 41,
370–412; b) H. Schmidbaur, A. Schier, Chem. Soc. Rev. 2008, 37, 1931–
1951; c) H. Schmidbaur, Gold Bull. 2000, 33, 3–10; d) P. Pyykkç, Angew.
Chem. Int. Ed. 2004, 43, 4412–4456; Angew. Chem. 2004, 116, 4512–
4557.
[18] Compounds 3 are best isolated by diluting the crude reaction mixture
(acetonitrile/dichloromethane) with ethyl acetate and hexane and sub-
sequent evaporation under vacuum. In this case, the products are iso-
lated as microcrystalline solids. Other isolation procedures (e.g., direct
precipitation) typically afford glassy solids that adhere to the walls of
the reaction vessel.
[19] Crystals of dimer 3a formed during the attempted preparation of an
asymmetric diphosphane complex (an analogue of 4a) by the reaction
of 3a with tricyclohexylphosphane and crystallisation from chloroform/
diethyl ether.
[20] If the role of the counterion and solid-state effects were negligible, for-
mation of the dimers would be favoured for entropy reasons because
the bonds in both compound types ([Au2(2)2]2+ and [Aun(2)n]n+) are
identical and, thus, no significant contribution from the reaction en-
thalpy to the overall energy balance would be expected.
[21] The conversion of 2 into 4a can also be effected through a one-pot re-
action by successive addition of 1 and then Ag[SbF6] to a solution of 2.
In this case, however, the product is contaminated by an Ag-1 complex
even when the reaction stoichiometry is strictly maintained.
[22] Crystals of 3a’·Me2CO suitable for X-ray diffraction analysis were grown
by recrystallization from acetone/hexane. However, this solvate partly
loses the solvent of crystallization upon drying and prolonged storage.
[23] To complete the series of structurally characterised compounds, we
also attempted to determine the crystal structure of 5a. Unfortunately,
the crystal structure data revealed extensive disorder of the tht ligand
that could not be reliably described.
[1] a) R. J. Puddephatt, in Comprehensive Coordination Chemistry, Vol. 2
(Eds.: G. Wilkinson, R. D. Gillard, J. A. McCleverty), Pergamon, Oxford,
1987, Chapter 55, pp. 861–923; b) M. C. Gimeno, A. Laguna, in Compre-
hensive Coordination Chemistry II, Vol. 6 (Ed.: J. A. McCleverty, T. J.
Meyer), Elsevier, Amsterdam, 2004, Chapter 6.7, pp. 911 –1145.
[2] For recent reviews, see: a) A. S. K. Hashmi, Angew. Chem. Int. Ed. 2005,
44, 6990–6993; Angew. Chem. 2005, 117, 7150–7154; b) A. S. K. Hashmi,
G. J. Hutchings, Angew. Chem. Int. Ed. 2006, 45, 7896–7936; Angew.
Chem. 2006, 118, 8064–8105; c) A. S. K. Hashmi, Chem. Rev. 2007, 107,
3180–3211; d) Z. Li, C. Brouwer, C. He, Chem. Rev. 2008, 108, 3239–
3265; e) A. Arcadi, Chem. Rev. 2008, 108, 3266–3325; f) E. JimØnez-
NfflÇez, A. M. Echavarren, Chem. Rev. 2008, 108, 3326–3350; g) D. J.
Gorin, B. D. Sherry, F. D. Toste, Chem. Rev. 2008, 108, 3351–3378; h) N. T.
Patil, Y. Yamamoto, Chem. Rev. 2008, 108, 3395–3442; i) M. Rudolph,
A. S. K. Hashmi, Chem. Soc. Rev. 2012, 41, 2448–2462; j) Y.-M. Wang,
A. D. Lackner, F. D. Toste, Acc. Chem. Res. 2014, 47, 889–901; k) R. Dorel,
A. M. Echevarren, Chem. Rev. 2015, 115, 9028–9072.
[3] For a few representative examples, see: a) A. S. K. Hashmi, T. M. Frost,
J. W. Bats, J. Am. Chem. Soc. 2000, 122, 11553–11554; b) B. Alcaide, P. Al-
mendros, Beil. J. Org. Chem. 2011, 7, 622–630; c) M. S. Reddy, Y. K.
Kumar, N. Thirupathi, Org. Lett. 2012, 14, 824–827; d) S. Basceken, M.
Balci, J. Org. Chem. 2015, 80, 3806–3814.
[4] For examples of reactions leading to gold-carbene complexes, see:
a) A. S. K. Hashmi, C. Lothschütz, C. Bçhling, T. Hengst, C. Hubbert, F. Ro-
minger, Adv. Synth. Catal. 2010, 352, 3001–3012; b) A. S. K. Hashmi, D.
Riedel, M. Rudolph, F. Rominger, T. Oeser, Chem. Eur. J. 2012, 18, 3827–
3830; c) D. Riedel, T. Wurm, K. Graf, M. Rudolph, F. Rominger, A. S. K.
Hashmi, Adv. Synth. Catal. 2015, 357, 1515–1523.
[5] a) D. Weber, M. R. GagnØ, Org. Lett. 2009, 11, 4962–4965; b) D. Wang, R.
Cai, S. Sharma, J. Jirak, S. K. Thummanapelli, N. G. Akhmedov, H. Zhang,
X. Liu, J. L. Petersen, X. Shi, J. Am. Chem. Soc. 2012, 134, 9012–9019;
c) A. Homs, I. Escofet, A. M. Echavarren, Org. Lett. 2013, 15, 5782–5785;
d) Y. Zhu, C. S. Day, L. Zhang, K. J. Hauser, A. C. Jones, Chem. Eur. J.
2013, 19, 12264–12271.
[6] H. Schmidbaur, A. Schier, Z. Naturforsch. b 2011, 66, 329–350.
[7] a) N. MØzailles, L. Ricard, F. Gagosz, Org. Lett. 2005, 7, 4133–4136; for
an application of this approach, see: b) A. S. K. Hashmi, A. Loos, A. Litt-
mann, I. Braun, J. Knight, S. Doherty, F. Rominger, Adv. Synth. Catal.
2009, 351, 576–582.
[24] M. L. Kuznetsov, Russ. Chem. Rev. 2002, 71, 265–282.
[25] E. Herrero-Gómez, C. Nieto-Oberhuber, S. López, J. Benet-Buchholz,
A. M. Echavarren, Angew. Chem. Int. Ed. 2006, 45, 5455–5459; Angew.
Chem. 2006, 118, 5581–5585.
[26] V. M. Lau, C. F. Gorin, M. W. Kanan, Chem. Sci. 2014, 5, 4975–4979.
[27] U. E. I. Horvath, H. G. Raubenheimer, Acta Crystallogr. Sect. E 2007, 63,
m567.
[28] G. Seidel, C. W. Lehmann, A. Fürstner, Angew. Chem. Int. Ed. 2010, 49,
8466–8470; Angew. Chem. 2010, 122, 8644–8648.
[29] R. J. Staples, C. King, M. N. I. Khan, R. E. P. Winpenny, J. P. Fackler, Jr.,
Acta Crystallogr. Sect. C 1993, 49, 472–475.
[8] C. Nieto-Oberhuber, S. López, A. M. Echavarren, J. Am. Chem. Soc. 2005,
127, 6178–6179.
[9] Examples of the numerous applications of the A- and B-type complexes
can be found in the review articles cited in Ref. [2].
[30] E. M. Barranco, O. Crespo, M. C. Gimeno, A. Laguna, P. G. Jones, B.
Ahrens, Inorg. Chem. 2000, 39, 680–687.
[31] With another series of experiments, we attempted to perform a gold-
catalysed addition of diphenylmethanol across acetonitrile with the Au-
1 complexes. However, under the conditions described in the literature
(5 mol% of 3c, reaction in MeCN at 758C for 22 h), this reaction prod-
uced none of the desired N-(diphenylmethyl)acetamide but only bis-
(diphenylmethyl)ether. For a reference, see: N. Ibrahim, A. S. K. Hashmi,
F. Rominger, Adv. Synth. Catal. 2011, 353, 461–468.
ˇ
ˇ
ˇ
ˇ
ˇ
[10] K. Skoch, I. Císarovµ, P. Stepnicka, Inorg. Chem. 2014, 53, 568–577.
[11] Although both donor moieties present in 1 are classified as soft accord-
ing to the HSAB concept, their different natures can be expected to
result in hemilabile coordination, which is typical for hybrid donors,
see: a) A. Bader, E. Lindner, Coord. Chem. Rev. 1991, 108, 27–110; b) C. S.
Slone, D. A. Weinberger, C. A. Mirkin, Progr. Inorg. Chem. 1999, 48, 233–
350; c) P. Braunstein, F. Naud, Angew. Chem. Int. Ed. 2001, 40, 680–699;
Angew. Chem. 2001, 113, 702–722.
[32] A. V. Gulevich, A. S. Dudnik, N. Chernyak, V. Gevorgyan, Chem. Rev.
2013, 113, 3084–3213.
[33] Reviews: a) P. Belmont, E. Parker, Eur. J. Org. Chem. 2009, 6075–6089;
b) M. Hoffmann, S. Miaskiewicz, J.-M. Weibel, P. Pale, A. Blanc, Beilstein J.
Org. Chem. 2013, 9, 1774–1780. Relevant examples:c) Y. Liu, F. Song, Z.
Song, M. Liu, B. Yan, Org. Lett. 2005, 7, 5409–5412; d) X. Du, F. Song, Y.
Lu, H. Chen, Y. Liu, Tetrahedron 2009, 65, 1839–1845; e) X. Zhang, Z. Lu,
C. Fu, S. Ma, J. Org. Chem. 2010, 75, 2589–2598.
[34] To the best of our knowledge, no Au-catalysed route to 7 has been re-
ported in the literature. For the first report, dealing with the use of Au
catalyst in this type of cyclization, see: A. S. K. Hashmi, L. Schwarz, J.-H.
Choi, T. M. Frost, Angew. Chem. Int. Ed. 2000, 39, 2285–2288; Angew.
Chem. 2000, 112, 2382–2385.
[12] a) K.-S. Gan, T. S. A. Hor, in Ferrocenes: Homogeneous Catalysis, Organic
Synthesis Materials Science (Eds.: A. Togni, T. Hayashi), Wiley-VCH, Wein-
heim, 1995, Chapter 1, pp. 3–104; b) S. W. Chien, T. S. A. Hor, in Ferro-
ˇ
ˇ
ˇ
cenes: Ligands, Materials and Biomolecules (Ed.: P. Stepnicka), Wiley,
Hoboken, 2008, Chapter 2, pp. 33–116; c) G. Bandoli, A. Dolmella,
Coord. Chem. Rev. 2000, 209, 161–196; d) D. J. Young, S. W. Chien,
T. S. A. Hor, Dalton Trans. 2012, 41, 12655–12665.
[13] R. G. Pearson, J. Am. Chem. Soc. 1963, 85, 3533–3539.
[14] Similar [AuCl(L)] complex, where L is (diphenylphosphanyl)acetonitrile,
has been reported in: P. Braunstein, D. Matt, F. Mathey, D. Thavard, J.
Chem. Res. Synop. 1978, 232–233.
[15] The geometry around the gold atom in
1
is similar to that in
[35] It should be noted that our catalysts proved to be inactive in the hydra-
tive cyclisation of 1,3-diphenylbutadiyne to 2,5-diphenylfurane: S.
Kramer, J. L. H. Madsen, M. Rottländer, T. Skrydstrup, Org. Lett. 2010, 12,
2758–2761.
[AuCl(FcPPh2)] (Fc=ferrocenyl), see: K. Rçßler, T. Rüffer, B. Walfort, R.
Packheiser, R. Holze, M. Zharnikov, H. Lang, J. Organomet. Chem. 2007,
692, 1530–1545.
Chem. Eur. J. 2015, 21, 15998 – 16004
www.chemeurj.org
16003
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[36] The reaction flask was cooled in a water bath to avoid overheating.
[37] The distillation removes unreacted (E)-6 that was present as a minor
isomer in the starting alcohol, as well as the catalyst.
[38] For examples of reactions with TOFs in the range of 106 hÀ1, see:
a) M. C. Blanco Jaimes, C. R. N. Bçhling, J. M. Serrano-Becerra, A. S. K.
Hashmi, Angew. Chem. Int. Ed. 2013, 52, 7963–7966; Angew. Chem.
2013, 125, 8121–8124; b) M. C. Blanco Jaimes, F. Rominger, M. M. Per-
eira, R. M. B. Carrilho, S. A. C. Carabineiro, A. S. K. Hashmi, Chem.
Commun. 2014, 50, 4937–4940.
[39] For an analogous, Pd-mediated preparation of 9, see: a) B. Gabriele, G.
Salerno, Chem. Commun. 1997, 1083–1084; b) B. Gabriele, G. Salerno, E.
Lauria, J. Org. Chem. 1999, 64, 7687–7692. This method required
1 mol% of a PdI2/KI catalyst and heating to 1008C for 24 h (GC yield:
85%).
[44] Annuloline was the first isolated and characterised natural compound
that contained an oxazole motif in its structure.
[45] a) B. Axelrod, J. Belzile, J. Org. Chem. 1958, 23, 919–920; b) R. S. Karimo-
to, B. Axelrod, J. Wolinsky, E. D. Schall, Tetrahedron Lett. 1962, 3, 83–85;
c) R. S. Karimoto, B. Axelrod, J. Wolinsky, E. D. Schall, Phytochemistry
1964, 3, 349–355.
[46] a) A. Brossi, E. Wenis, J. Heterocycl. Chem. 1965, 2, 310–312; b) M. P.
Doyle, W. E. Buhro, J. G. Davidson, R. C. Elliott, J. W. Hoekstra, M. Oppen-
huizen, J. Org. Chem. 1980, 45, 3657–3664; c) P. Molina, P. M. Fresneda,
P. Almendros, Heterocycles 1993, 36, 2255–2258; d) F. Besseli›vre, S.
Piguel, F. Mahuteau-Betzer, D. S. Grierson, Org. Lett. 2008, 10, 4029–
4032; e) C. Wan, L. Gao, Q. Wang, J. Zhang, Z. Wang, Org. Lett. 2010, 12,
3902–3905; f) C. W. Cheung, S. L. Buchwald, J. Org. Chem. 2012, 77,
7526–7537; g) D. Kumar, N. M. Kumar, M. P. Tantak, Chem. Biol. Interface
2012, 2, 331–338; h) W.-C. Lee, T.-H. Wang, T.-G. Ong, Chem. Commun.
2014, 50, 3671–3673.
[47] Bromoveratrole 16 (1.05 equiv) and 12 f were reacted in the presence
of as much as 5 mol% Pd(OAc)2 and 20 mol% P(C6H4Me-2)3 in N,N-di-
methylformamide at 1208C for 24 h. An analysis of the crude reaction
mixture by using NMR spectroscopy revealed only the presence of un-
reacted 16 and no signals due to protons at the C=C double bond of
either 12 f or any other 1,2-disubstituted alkene.
[48] For examples of AuI complexes with P,N-donors and their catalytic use,
see: C. Khin, A. S. K. Hashmi, F. Rominger, Eur. J. Inorg. Chem. 2010,
1063–1069.
[49] CCDC 1412074 (2), 1412075 (3a), 1412076 (3a’·Me2CO), 1412077 (3c),
1412078 (4a), 1412079 (12a), 1412080 (17), 1412081 (18), contain the
supplementary crystallographic data. These data can be obtained free
of charge by The Cambridge Crystallographic Data Centre.
[40] a) B. S. Rao, K. S. Subramaniam, Proc. Indian Acad. Sci. 1936, 3, 31–37;
b) G. S. R. S. Rao, B. Ravindranath, V. P. S. Kumar, Phytochemistry 1984,
23, 399–401; c) G. Ohloff, E. Demole, J. Chromatogr. 1987, 406, 181–
183; d) E. Kovµts, J. Chromatogr. 1987, 406, 185–222; e) L. N. Misra, A.
Husain, Planta Med. 1987, 53, 379–380; f) A. O. Tucker, M. J. Maciarello,
J. Essent. Oil Res. 1995, 7, 653–655.
[41] The 1,3-oxazole motif represents an essential component of various
natural and biologically active compounds. For selected reviews, see:
a) X.-M. Peng, G.-X. Cai, C.-H. Zhou, Curr. Top. Med. Chem. 2013, 13,
1963–2010; b) V. S. C. Yeh, Tetrahedron 2004, 60, 11995–12042; c) D.
Davyt, G. Serra, Mar. Drugs 2010, 8, 2755–2780; d) Z. Jin, Nat. Prod. Rep.
2013, 30, 869–915.
[42] W. He, C. Li, L. Zhang, J. Am. Chem. Soc. 2011, 133, 8482–8485.
[43] a) D. C. Palmer, S. Venkatraman, in Chemistry of Hererocyclic Compounds
Vol. 60 (Ed.: D. C. Palmer), Wiley, Hoboken, 2004, Part A, Chapter 1, p.1;
b) G. V. Boyd, in Science of Synthesis (Houben-Weyl), Vol. 11 (Ed.: E.
Schaumann), Thieme, Stuttgart, 2002, Chapter 11.12, p. 383; c) I. J.
Turchi, M. J. S. Dewar, Chem. Rev. 1975, 75, 389–437.
Received: July 29, 2015
Published online on September 23, 2015
Chem. Eur. J. 2015, 21, 15998 – 16004
www.chemeurj.org
16004
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim