Gold(III) complexes catalyze deoximations/transoximations at neutral pH

Article abstract of DOI:10.1002/anie.201007269

Golden solution: A neutral solution of AuBr₃, containing [AuBr₂(OH)₂]^- in equilibrium with [AuBr ₃(OH)]^- and [AuBr₄]^-, promotes the chemoselective hydrolysis of robust oximes into carbonyl compounds without racemization (see scheme). The food additive diacetyl acts as a NH ₂OH-trapping agent, thus avoiding the formation of gold nanoparticles and allows the reaction to run catalytically. Copyright

Full text of DOI:10.1002/anie.201007269

DOI: 10.1002/anie.201007269
Gold-Catalyzed Reactions
Gold(III) Complexes Catalyze Deoximations/Transoximations at
Neutral pH**
Carles Isart, David Bastida, Jordi Burꢀs,* and Jaume Vilarrasa*
Dedicated to Professor Josꢀ Barluenga on the occasion of his 70th birthday
Table 1: Screening of MXn and additives.[a]
The reaction of gold trihalides with reducing agents such as
thiols, sodium borohydride, and hydroxylamine results in gold
nanoparticles (AuNPs).[1] In particular, NH2OH is used to
facilitate the growth of smaller particles or the formation of
thin films of gold.[2] Our interest in developing a mild catalytic
Entry
MXn, (mol%)
Additive[b]
Conv. [%]
method for the conversion of nitro, oxime, and nitrone groups
into carbonyl groups led us to examine the interaction of
gold(III)[3] with simple oximes as a potential approach for
deoximation.[4] The hydrolysis of the coordination complexes
of oximes and glyoximes with gold(III)[5] has not been
1
2
3
4
5
6
7
8
Sc(OTf)3 (20)
LaCl3 (20)
FeCl2 (20)
FeBr3 (20)
RuCl3 (20)
RhCl3 (20)
PdCl2 (20)
PtCl2 (20)
PtCl4 (20)
CuI (20)
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
none
0
0
0
ꢀ3
ꢀ7
ꢀ6
ꢀ7
0
=
investigated. If a partial hydrolysis of oximes into RR’C O
and NH2OH took place, NH2OH would be oxidized in situ by
gold(III), and the equilibrium would be shifted to the right.
Our first challenge was to find a water soluble and stable
gold complex that could catalyze the hydrolysis of the robust
oxime group[6] at neutral pH, and when possible at room
temperature. A polyfunctional molecule containing groups
that are prone to hydrolysis should survive under these
reaction conditions. Lewis acids that are soluble and stable in
aqueous media, that is Sc(OTf)3, LaCl3·7H2O, FeXn, RuCl3,
RhCl3, PtCl4, CuX2, AuX3, InBr3, and related compounds
were screened as catalysts for the hydrolysis of 4-phenyl-2-
butanone oxime (Table 1). Only AuBr3 (99.9%) and AuCl3
(99.99%) promoted the hydrolysis of 4-phenyl-2-butanone
oxime when the pH was adjusted to pH 7 with a standard
solution of NaOH (1.000m, 99.99%;[7] entries 19–27). Except
for AuX3, none of these salts was a suitable initiator or
catalyst at pH 4–8;[8] in fact, in most cases when the solutions
were neutralized, the corresponding hydroxides or oxide
hydrates precipitated out of solution, as expected. Despite
their lack of solubility in water, we also examined a
platinum(II) salt (entry 8), copper(I) salts (entries 10–14),[9]
and a gold(I) salt (entry 17), because of their success as
catalysts in other contexts, but no effect was observed. Among
all these common transition-metal cations, only the gold(III)
9
0
10
11
12
13
14
15
16
17
19
20
21
22
23
24
25
26
27
28
29
30
ꢀ4
Cu2Cl2 (10)
0
[c]
Cu2(OTf)2 (10)
0
0
10
ꢀ4
0
CuBr2 (20)
Cu(OTf)2 (20)
Ag2SO4 (20)
InBr3 (20)
AuCl (20)
AuCl3 (20)
AuBr3 (20)
AuBr3 (50)
AuBr3 (5)
AuBr3 (5)
AuBr3 (5)
AuBr3 (5)
AuBr3 (5)
AuBr3 (5)
–
none
none
none
none
ꢀ2
27
37[d]
100[d]
60
61
70
88
100
100[d]
0
acetone
CH3COCH2COOEt
CH3COCH2COCH3
formaldehyde hydrate
CH3COCOOCH3
CH3COCOCH3
CH3COCOCH3
CH3COCOCH3
CH3COCOCH3
AlBr3 (5)
FeBr3 (5)
0[e]
5[f]
[a] An aqueous solution of NaOH (1.000m) solution was added to the
solution of MXn in H2O/EtOH (1:4) until a pH of 7 was achieved, and
then the oxime was added. [b] Used 100 mol% except in the case of
acetone, which was used in excess (entry 22). [c] Copper(I) trifluoro-
methanesulfonate·toluene, 99.99%. [d] The same result was achieved
when using THF, CH3CN, 1,4-dioxane, 2-propanol, and MeOH instead of
EtOH (always with 20% H2O, v/v). [e] Under these reaction conditions,
Ba2+, Ca2+, Cr3+, Mg2+, Mn2+, Mn3+, Ni2+, Sb3+, Sn2+, Sn4+, Ti4+, and Zn2+
also gave 0% yield. [f] RuCl3, RhCl3, PdCl2, CuI, Cu(OTf)2, and Ag2SO4
gave <2% yield. Tf=trifluoromethanesulfonyl. THF=tetrahydrofuran.
[*] C. Isart, D. Bastida, Dr. J. Burꢀs, Prof. Dr. J. Vilarrasa
Dept Quꢁmica Orgꢂnica, Universitat de Barcelona
Av. Diagonal 647, 08028 Barcelona, Catalonia (Spain)
E-mail: jvilarrasa@ub.edu
[**] The MICINN (Spanish Government; CTQ 2009-13590) and Gen-
eralitat of Catalunya (2009 SGR 825) are acknowledged for financial
support. The University of Barcelona (UB) is acknowledged for a
studentship to C.I. (2006–2010) and the Fundaciꢃ Cellex de
Barcelona is acknowledged for a postdoctoral fellowship to J.B.
(Sept. 2009-June 2010). We are grateful to Dr. I. Fernꢄndez and L.
Ortiz (UB MS Service) for their help with the ESIMS spectra.
species worked at neutral pH. However, 50 mol% of gold(III)
was required to complete the hydrolysis of 4-phenyl-2-
butanone oxime (entry 21), because the active AuIII species
was reduced to Au0/I NPs by NH2OH, as expected.
Thus, the next challenge was to develop a catalytic version
of this hydrolysis. We explored the possibility of trapping
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201007269.
Angew. Chem. Int. Ed. 2011, 50, 3275 –3279
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3275

Communications
NH2OH with carbonyl compounds that are capable of
forming oximes that are more stable than the starting oxime
(Table 1, entries 22–27). The best results for this transoxima-
tion process were obtained with methyl pyruvate
(CH3COCOOCH3,
entry 26)
and
with
diacetyl
(CH3COCOCH3, entry 27).[10] At the same time in silico
screening, at the MP2 level of theory,[11] indicated that the
=
diacetyl monoxime (CH3COC( NOH)CH3) is especially
stable compared with many other oximes. The AuIII species
is crucial because without AuBr3 the exchange between 4-
phenyl-2-butanone oxime and diacetyl did not occur
(entry 28), even at 1008C (microwave reactor).
To rule out the participation of the commonly reported
impurities contained in 99.9% AuBr3 (at the ppm level;
Aldrich trace analysis), the effect of 5 mol% or more of the
following salts at pH 7 was studied in the presence of diacetyl:
AlBr3, BaCl2, CaCl2, CrCl3, MgBr2, MnCl2, MnF3, NaBr
(200 mol%), NiBr2, SbCl3, SnCl2/SnCl4 (1:1), TiCl4/Ti(OiPr)4
(1:1), and ZnBr2. All were inactive (entry 29).
The AuIII species involved as either the initiator(s) or
catalyst(s) were examined by electrospray ionization mass
spectroscopy (ESIMS, negative mode). Aqueous solutions of
AuBr3 were adjusted to different pH values by the addition of
an NaOH solution and investigated by ESIMS methods to
afford the spectra shown in Figure 1; these spectra were
registered with a potential of 20 V because no signals were
observed below this value, whereas above 20 V fragmenta-
tions of the main species occurred. At pH 5.4 (Figure 1a) the
major anionic species [AuBr3(OH)]ꢁ appeared as a quartet as
a result of the almost 1:1 79Br/81Br isotopic distribution, but an
equilibrium was noted between this complex and [AuBr4]ꢁ
(quintet) and [AuBr2(OH)2]ꢁ (triplet), as a result of a quick
exchange of bromide and hydroxide ions (Scheme 1). At
pH 7.0, [AuBr2(OH)2]ꢁ predominated (Figure 1b). Finally, at
pH 7.9, the less intensely colored species [AuBr2(OH)2]ꢁ and
[AuBr(OH)3]ꢁ (doublet) were predominant (Figure 1c). All
this data agrees with the spectrophotometric data previously
reported for these species.[12]
Figure 1. ESIMS (negative mode, 20 V) results for 0.1m AuBr3 99.9%
+ NaOH. a) Added ca. 100 mol% of NaOH, pH 5.40. The major
species is [AuBr3(OH)]ꢁ (m/z 450.73/452.73/454.73/456.73) in equilib-
rium with AuBr4 (quintet, centered at 516.64) and [AuBr2(OH)2]ꢁ
ꢁ
(triplet, centered at 390.81). b) Added ca. 170 mol% of NaOH,
pH 7.00. c) With ca. 240 mol% of NaOH, pH 7.89.
the Beckmann rearrangement. In general, 5 mol% of AuBr3
was enough for the full conversion of each ketoxime and
=
diacetyl into the ketone and CH3COC( NOH)CH3. In some
cases only 2 mol% of AuBr3 was required either at room
temperature or at 608C. Under these reaction conditions: 1)
an ester group (entry 7) was not hydrolyzed, 2) mono- and
dioximes of diketones (entries 8 and 9) did not cyclize and the
diketones generated did not undergo self-aldol reactions, 3)
conjugate additions to a,b unsaturated ketoximes (entries 11
and 12) or to the resulting ketones were not observed, and 4)
the CHNO2 stereocenter shown in entry 13 did not epimerize.
Hydroxy protecting groups such as tBuMe2Si (TBS) and 1,2-
diol protecting groups such as isopropylidene were not
affected,[14] which was expected because the medium was
not acidic. The enantiopure compounds of entries 16 and 17[15]
did not racemize and no epimerization was noted in the case
of d-fructose oxime (entry 18). Free hydroxy groups did not
interfere with the hydrolysis (entries 17 and 18).
Scheme 1. Gold(III) species in equilibria.
In practice, AuIII solutions at pH 5.4 were slightly more
catalytically active (50% conversion after 6 h, under the
reaction conditions stated in Table 1, entry 27, with 4-phenyl-
2-butanone oxime) than those at pH 7.0 (50% conversion
after 8 h), whereas those at pH 7.9 were less catalytically
active (50% conversion after ca. 12 h) and those at pH ꢂ 8.5,
where [Au(OH)4]ꢁ predominated (Scheme 1), were com-
pletely inactive. The addition of NaBr (from 10 equiv to
100 equiv) stopped the progress of these reactions.[13]
We then examined the scope of the reaction at pH 7. Our
protocol could be applied to many ketoximes (Table 2),
without any formation of the corresponding amides through
3276
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 3275 –3279

Table 2: Gold(III)-Catalyzed transoximation of ketoximes.
Table 3: Gold(III)-catalyzed conversion of aldoximes into aldehydes.
Entry
1
Oxime
Aldehyde yield [%]
Entry Oxime
Ketone Entry Oxime
yield
[%]
Ketone
yield
[%]
100[a]
92[a,c]
1
2
98
10
11
100[d]
91[a]
2
100[b]
100
3
4
100[b]
100[b]
3
100[c]
96
12
13
92[b]
60[b]
4
5
93[a,c]
5
6
91[b,d]
100
100
14
15
100[a]
97[a,c]
90[a]
=
(CH2)14C NOH
6
7
7
100[a,e]
100[a]
94
16
17
92[a]
91
[a] Reaction carried out at RT. [b] Reaction carried out at 608C. [c] With
2 mol% of AuBr3 at pH 7. [d] 10 mol% of AuBr3 at pH 7 and 120 mol%
of diacetyl. [e] In 1:4 D2O/THF the anomeric mixture[16] of b- and a-d-
arabinopyranose (major) was isolated.
8
9
100[d]
18
100[c,e]
[a] Reaction carried out at 608C. [b] At 1008C for 6 h. [c] With 2 mol% of
AuBr3 at pH 7. [d] 200 mol% of diacetyl was added. [e] In D2O/THF
(1:4)the expected mixture[16] of b-d-fructopyranose (major), b-d-fructo-
furanose, and a-d-fructofuranose was obtained.
The application of our protocol to aliphatic and aromatic
aldoximes (Table 3, entries 1–5) gave aldehydes without
dehydration to nitriles. THF was a preferable solvent to
EtOH or MeOH, because contamination of the products with
hemiacetals and acetals is avoided. The oxime of Garnerꢀs
aldehyde (entry 6) and d-arabinose oxime (entry 7) were
hydrolyzed with full retention of the configuration of their
a stereogenic centers.
Finally, to gain insight into the role of the AuIII complexes,
several experiments were followed by ESIMS methods.[11]
One representative example is shown in Figure 2. In this
experiment we chose a solution of AuBr3 at pH 6. A saturated
solution of 4-phenyl-2-butanone oxime in water was added to
the stirred AuIII solution and after analysis the spectrum in
Figure 2a was obtained. The intensities of the triplet, quartet,
and quintet had decreased and new peaks had appeared
(Figure 2a, see arrows) that corresponded to new complexes:
[AuBr(OH)2(oximate)]ꢁ (1, doublet; see Scheme 2 for struc-
tures), [AuBr2(OH)(oximate)]ꢁ (3, triplet), and [AuBr3-
(oximate)]ꢁ (5, quartet). These new complexes are most
likely explained by the replacement of one bromide ion by the
Figure 2. ESIMS spectra of the reaction of gold(III) complexes with 4-
phenyl-2-butanone oxime in H2O. a) Approximately at pH 6.0, after
adding the oxime. b) After 2 h, when the wine-red color of the solution
had faded; regions where the main hydrated intermediates could
appear are expanded. See the main text and Scheme 2 for discussion.
oxime with concomitant loss of a proton. The many spectra
obtained at different times over the course of the reaction
Angew. Chem. Int. Ed. 2011, 50, 3275 –3279
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3277

Communications
seem to indicate that the Brꢁ/OHꢁ and HBr/H2O exchanges
occur quickly at standard concentrations to give equilibrium
mixtures including oximate-containing complexes
diacetyl the complexes of type A (Scheme 3, bottom) may
isomerize to the complexes of type B (by hydrolysis and/or by
an O to N exchange). It is likely that these B species are
involved in the known redox process that affords AuNPs and
“HNO” (2HNO ! N2O + H2O).[18]
(Scheme 2). After 2 hours (Figure 2b) the signals of
In summary, neutral AuBr3 solutions catalyze the deox-
imation of 25 oximes, many of which are really robust,[6,19] by
transoximation with diacetyl at room temperature (in most
cases). No other metallic salts catalyzed this reaction between
pH 4 and 8. The reaction conditions are extremely mild:
a stereocenters do not epimerize, other functional groups and
standard protecting groups are inert, ketoximes do not
undergo Beckmann rearrangements, and aldoximes are not
converted into nitriles. It is a paradox that the secret of
success is to avoid the formation of AuNPs, which are so
important in other contexts, because their formation results in
the disappearance of the AuIII species. Thus, a new episode in
the gold catalysis story is uncovered here. Studies of the
performance of the [AuBrx(OH)y]ꢁ species in other reactions
at neutral pH, where the central atom is still surprisingly
active as a Lewis acid, are in progress.
Scheme 2. Gold(III) species involved in Figure 2.[17]
[AuBrx(OH)y]ꢁ had almost disappeared and the main peaks
now corresponded to the oximato-containing complexes (1, 3,
5). Their hydrated forms (2, 4; Scheme 2), which are the
obvious transient intermediates in any oximation or hydro- Experimental Section
Representative deoximation procedure: Diacetyl (27 mL, 26 mg,
lytic deoximation procedure, were obscured by noise. Never-
theless, by enhancing the sensitivity by 100-fold, even these
relatively unstable intermediates could be observed (Fig-
ure 2b, left expansions). In contrast, we could not detect the
hydrated form [AuBr3(O-NH-CRR’OH]ꢁ (signals too low).
0.3 mmol) and the AuIII stock solution[7] (175 mL, 0.015 mmol) were
added to a solution of the ketoxime (0.3 mmol) in ethanol/water (4:1
v/v, 2 mL). The mixture was stirred at ca. 208C for 15 h in a closed vial.
An N2 atmosphere was unnecessary. Afterwards, the solution was
diluted with dichloromethane (10 mL), and water was then added and
the layers were separated. The aqueous layer was re-extracted with
dichloromethane. The organic extracts were dried over Na2SO4. The
solvent was removed under reduced pressure and the crude reaction
mixture was analyzed by 1H NMR spectroscopy. Several reactions
were performed later at 1.0 or 2.0 mmol scale. When purification was
required it was carried out by flash column chromatography on silica
gel with hexanes/EtOAc (1:1).
With the acetone oxime we observed the O-oximato
ꢁ
species[11] [AuBr2(OH)(O-N CMe2)] (m/z 443.86/445.86/
=
447.86) and its hydrated form (m/z 461.82/463.82/465.82), in
spite of its relatively short half life. Experiments with the 15N-
labeled acetone oxime[11] validated the structures of the
intermediates. In the presence of diacetyl, complexes of
diacetyl monoxime with AuIII were also detected (m/z 471.86/
473.85/475.85).
In Scheme 3, we suggest a plausible general mechanism at
a pH value between 5.4 and 7.9 for the transoximation
between acetone oxime and diacetyl. In the absence of
Received: November 18, 2010
Published online: March 1, 2011
Keywords: gold · homogeneous catalysis · ketones ·
.
reaction mechanisms
[1] For reviews on AuNPs, see: a) D. A. Giljohann, D. S. Seferos,
W. L. Daniel, M. D. Massich, P. C. Patel, C. A. Mirkin, Angew.
Chem. 2010, 122, 3352 – 3366; Angew. Chem. Int. Ed. 2010, 49,
3280 – 3294; b) R. Sardar, A. M. Funston, P. Mulvaney, R. W.
Murray, Langmuir 2009, 25, 13840 – 13851; c) A. Corma, H.
Garcꢁa, Chem. Soc. Rev. 2008, 37, 2096 – 2126.
[2] K. R. Brown, L. A. Lyon, A. P. Fox, B. D. Reiss, M. J. Natan,
Chem. Mater. 2000, 12, 314 – 323, and references therein.
[3] For very recent examples of synthetic applications of gold(III),
see: a) R. Nakajima, T. Ogino, S. Yokoshima, T. Fukuyama, J.
Am. Chem. Soc. 2010, 132, 1236 – 1237; b) N. D. Shapiro, F. D.
Toste, J. Am. Chem. Soc. 2008, 130, 9244 – 9245; c) A. B. Cuenca,
G. Mancha, G. Asensio, M. Medio-Simꢂn, Chem. Eur. J. 2008, 14,
1518 – 1523. For reviews on catalysis that include reactions of
AuX3, see: d) A. S. K. Hashmi, Angew. Chem. 2010, 122, 5360 –
5369; Angew. Chem. Int. Ed. 2010, 49, 5232 – 5241; e) Y.
Yamamoto, H. D. Gridnev, N. T. Patil, T. Jin, Chem. Commun.
2009, 5075 – 5087, and reviews cited therein.
Scheme 3. Plausible mechanism for the hydrolysis of acetone oxime in
the presence of gold(III) and diacetyl at neutral pH.
3278
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 3275 –3279

[4] Most deoximation procedures involve strong oxidants, reducing
agents, acidic reaction conditions, or stoichiometric amounts of
other reagents. For a recent review, see: A. Corsaro, M. A.
Chiacchio, V. Pistarꢃ, Curr. Org. Chem. 2009, 13, 482 – 501.
[5] P. Bourosh, O. Bologa, Y. Simonov, N. Gerbeleu, J. Lipkowski,
M. Gdaniec, Inorg. Chim. Acta 2007, 360, 3250 – 3254, and
references therein.
[6] The stability of oximes to hydrolysis is paradigmatic; see: J.
Kalia, R. T. Raines, Angew. Chem. 2008, 120, 7633 – 7636;
Angew. Chem. Int. Ed. 2008, 47, 7523 – 7526, and references
therein.
[10] We prefer the food additive diacetyl to methyl pyruvate because:
1) it is cheaper, 2) it has a lower boiling point and therefore any
excess of additive may be readily removed, 3) its monoxime is
more soluble in water thus making the workup easier, and 4) it
has two equivalent CO groups per molecule. In aqueous media it
is only partially hydrated; for example, see: K. Miyata, K.
Nakashima, M. Koyanagi, Bull. Chem. Soc. Jpn. 1989, 62, 367 –
371.
[11] See the Supporting Information.
[12] A. Usher, D. C. McPhail, J. Brugger, Geochim. Cosmochim. Acta
2009, 73, 3359 – 3380.
[7] An aqueous solution of NaOH (1.000m) was added to a solution
of AuBr3 in water (0.100m) until pH 7.0 was obtained (moni-
tored with a pH meter). The resulting wine-red solution that was
0.086m AuIII (stock solution) was stable for months when stored
in a closed vial. AuCl3 was almost as active as AuBr3 but
solutions of AuCl3 at pH 7.0 were unstable (a precipitate was
formed).
[8] Under the reaction conditions stated in Table 1 (entries 1–20)
but without neutralization, EtOH/H2O (4:1) solutions of CuBr2
(pH 2.17), Cu(OTf)2 (pH 1.90), and RuCl3 (pH 1.37) turned out
to be more efficient than those of AuBr3 (pH 1.45, 30% yield of
ketone), and PtCl4 solutions gave an outcome similar to AuBr3,
whereas all the remaining salts were very inefficient (< 15% of
ketone). However, the addition of only 100 mol% of NaOH
(pH value ꢂ 4) deactivated CuII, RuIII, and PtIV salts completely.
In other words, only AuIII complexes are still active at neutral pH
values.
[13] This is understandable if the substitution of the oxime for a
bromide ion, in the gold complex, is one of the key reaction
steps, which seems likely.
[14] TBSO- and isopropylidene-containing fragments, which were
synthesized in our lab during the total syntheses of macrolides,
were used as models.
[15] As checked by measuring the optical rotation (entry 16, Table 2;
entry 6, Table 3) and HPLC analysis using a chiral stationary
phase of the TBS derivative of the ketone of entry 17 in Table 2.
[16] J. B. Lambert, G. Lu, S. R. Singer, V. M. Kolb, J. Am. Chem. Soc.
2004, 126, 9611 – 9625, and references therein.
[17] Elucidating the possible reactivity differences between E and
Z oximes and between cis- and trans-aurates is outside the scope
of this work.
[18] V. Soni, R. N. Mehrotra, Transition Met. Chem. 2003, 28, 893 –
898.
[19] Provided that they are less robust than the very stable diacetyl
monoxime. The mono- and dioxime of 2,4-pentanedione can be
hydrolyzed, with 5 mol% of AuBr3 at pH 7, overnight, by adding
110 mol% and 220 mol% of diacetyl, respectively, whilst the
oxime of the methyl pyruvate requires heating and a much larger
excess of diacetyl.
[9] For a recent review on copper(I)-catalyzed azide–acetylene
cycloadditions (click reactions), see: a) J. E. Hein, V. V. Fokin,
Chem. Soc. Rev. 2010, 39, 1302 – 1315; for the Cu2(OTf)2-
catalyzed formation of tetrazoles from azides and cyanides,
see: b) L. Bosch, J. Vilarrasa, Angew. Chem. 2007, 119, 4000 –
4004; Angew. Chem. Int. Ed. 2007, 46, 3926 – 3930.
Angew. Chem. Int. Ed. 2011, 50, 3275 –3279
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3279