[Rh2(MEPY)4] and [BiRh(MEPY)4]: Convenient Syntheses and Computational Analysis of Strikingly Dissimilar Siblings

Article abstract of DOI:10.1002/hlca.202100042

[Rh₂(MEPY)₄] is a versatile catalyst for asymmetric synthesis but its preparation requires purification by chromatography on surface-modified silica. A higher yielding procedure based on a more convenient work-up is presented herein. Moreover, a much improved method for the preparation of [BiRh(OTfa)₄] is described, which makes this heterobimetallic complex readily available. Subsequent exchange of the trifluoroacetate ligands opens access to a so far underappreciated family of (chiral) paddlewheel complexes. While [BiRh] complexes comprising four carboxylate ligands are highly adequate for intermolecular asymmetric cyclopropanation reactions, [BiRh(MEPY)₄] as the heterobimetallic cousin of [Rh₂(MEPY)₄] was found to be surprisingly unreactive; DFT calculations uncover the reasons for this inertia.

Full text of DOI:10.1002/hlca.202100042

HELVETICA
chimica acta
Accepted Article
Title: [Rh2(MEPY)4] and [BiRh(MEPY)4]: Convenient Syntheses and
Computational Analysis of Strikingly Dissimilar Siblings
Authors: Alois Fürstner, Lorenz E. Löffler, Michael Buchsteiner, Lee R.
Collins, Fabio P. Caló, and Santanu Singha
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Helv. Chim. Acta 10.1002/hlca.202100042
Link to VoR: https://doi.org/10.1002/hlca.202100042
www.helv.wiley.com

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
[Rh2(MEPY)4] and [BiRh(MEPY)4]: Convenient Syntheses
and Computational Analysis of Strikingly Dissimilar Siblings
Lorenz E. Löffler,[+] Michael Buchsteiner,[+] Lee R. Collins, Fabio P. Caló, Santanu Singha, and Alois Fürstner*
a Max-Planck-Institut für Kohlenforschung, 45470 Mülheim/Ruhr, Germany
email: fuerstner@kofo.mpg.de
Dedicated to Prof. E. Peter Kündig on the occasion of his 75th birthday
[Rh2(MEPY)4] is a versatile catalyst for asymmetric synthesis but its preparation requires purification by chromatography on surface-modified silica. A
higher yielding procedure based on a more convenient work-up is presented herein. Likewise, a much improved method for the preparation of
[BiRh(OTfa)4] is described, which makes this heterobimetallic complex readily available. Subsequent exchange of the trifluoroacetate ligands opens
access to a so far underappreciated family of (chiral) paddlewheel complexes. While [BiRh] complexes comprising four carboxylate ligands are highly
adequate for intermolecular asymmetric cyclopropanation reactions, [BiRh(MEPY)4] as the heterobimetallic cousin of [Rh2(MEPY)4] was found to be
surprisingly unreactive; DFT calculations uncover the reasons for this inertia.
Keywords: Bismuth • Carbenes • Cyclopropanation • Heterobimetallic Complexes • Rhodium
convenient and proficient method. Moreover, it was hoped that
Introduction
replacement of 1 by its heterobimetallic cousin [BiRh(5S-MEPY)4] (13)
A
recent approach to an ensemble of different casbane diterpenes
might lead to even better results: we have previously shown that certain
bismuth-rhodium paddlewheel complexes outperform the parent
dirhodium catalysts in terms of asymmetric induction in intermolecular
cyclopropanation reactions.[13,14] Proof-of-concept notwithstanding, the
preparation of such heterobimetallic complexes also needs to be
improved; the quest for [BiRh(5S-MEPY)4] (13) hence provided an
opportunity to develop a more convenient entry route and, at the same
time, to investigate whether or not formal replacement of one of the
rhodium atoms by the main-group element bismuth entails a better
application profile in this particular case too.
including compounds 4-6 capitalized on the use of the cyclopropyl lactone
3 as a key building block.[1] This compound is readily accessible by
intramolecular cyclopropanation of diazoester
2
using [Rh2(5S-
MEPY)4]2MeCNiPrOH (1a) as the catalyst;[2-12] in our hands,
3 was
obtained in 87% yield and 93% ee.[1]
Results and Discussion
An Improved Method for the Preparation of [Rh2(5S-MEPY)4]2MeCN
(1b). Complex 1a is made from commercial [Rh2(OAc)4] on treatment with
excess methyl 2-pyrrolidone-carboxylate (MEPY-H, 7) in refluxing
chlorobenzene overnight (13 h). Although the ligand exchange is a priori
reversible, the reaction is driven to completion by passing the condensing
solvent through a thimble filled with K2CO3 in a Soxhlet apparatus, which
traps the released and co-evaporating HOAc.[2] To ensure reasonable rates,
7 has to be used in slight excess. The reaction is accompanied by a
characteristic color change, in that the originally green solution turns dark
red when heated overnight. For work up, all volatile materials were
evaported in high vacuum, which causes another striking color change to
give a deep blue/violet solid residue. It is this material that had to be
purified by reversed phase chromatography in order to remove isomeric
Scheme 1. Formation of a key building block for the total synthesis of casbane
derivatives by intramolecular cyclopropanation; 5S-MEPY = methyl 2-pyrrolidinone-
5(S)-carboxylate (methyl 5(S)-pyroglutamate)
Although the reaction worked well even when the catalyst loading was
reduced to 0.6 mol%, a substantial amount of complex 1 was needed to
ensure the material throughput necessary for this extensive total synthesis
campaign. The preparation of 1a is well described in the literature, but its
isolation in analytically pure form hinges upon reversed phase
chromatography using cyanopropyl-modified silica and the reported yield
(58%) is not overly high;[2] therefore we were prompted to develop a more
species[15]
and
afford
bright-red
(cis-2,2)-configured
[Rh2(MEPY)4]2MeCNiPrOH (1a), carrying acetonitrile ligands at the axial
1
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
coordination sites and iPrOH as a co-crystallized solute in the unit cell as
analytically pure form in 81% yield as a single isomer, within the limits of
detection (600 MHz NMR). The identity of the complex follows from its
spectroscopic and analytical data; in contrast to the original procedure,[2,4]
the modified work up furnished 1b free of any solute in the unit cell as
confirmed by X-ray crystallography (Figure 1).
shown by X-ray crystallography.[2,3]
An Improved Procedure for the Preparation of Heterobimetallic [BiRh]-
Paddlewheel Complexes. The chemistry of heterobimetallic [BiRh]
paddlewheel complexes was pioneered by Dikarev and coworkers. The
initial access routes had relied on solid phase syntheses;[17,18] later, these
authors managed to prepare the prototype complex [BiRh(OTfa)4] (9) by
reaction of [Rh2(OTfa)4] (8) with Bi(OTfa)3 and Bi metal in toluene/Ph2O.[19]
One of the recommended procedures for product isolation is sublimation
at 125°C (10 mbar). Although 9, once obtained in pure form, is stable in
3
air for days, the sublimation proved not only rather low yielding ( 40%)
but quite erratic in our hands: in one out of ca. four cases, the mixture
spontaneously turned black, resulting in loss of material. Although the
exact reasons could not be clarified, the decomposition is almost certainly
not caused by inadvertent contact of the crude product with oxygen
and/or moisture, as it occurred equally regularly under rigorously inert
conditions; rather, we suppose that traces of remaining Bi metal, when
superheated in the solid state, might entail deleterious over-reduction of
the product complex which comprises a bismuth center in the unusual
oxidation state +2.
Scheme 2. Practical and high-yielding preparation of 1b avoiding purification by
reversed-phase chromatography
Figure 1. Structure of (cis-2,2)-configured [Rh2(5S-MEPY)4]2MeCN (1b) in the solid
state; H-atoms and disorder of two of the –COOMe groups not shown for clarity; for
the full structure, see the Supporting Information.1
Scheme 3. Practical and high-yielding preparation of the heterobimetallic complex
13 avoiding (erratic) purification by sublimation; TFA = trifluoroacetic acid
We chose to purify an aliquot of the crude blue material by subliming
excess ligand (and possible impurities) off in high vacuum to give an
analytically pure sample of what turned out to be [Rh2(5S-MEPY)4]2(S-
MEPY-H)2 (1c) (Scheme 2). It is the additional ligand used to drive the
reaction which occupies the axial sites; the interaction is obviously so
Gratifyingly, we found that sublimation is not needed altogether; rather,
complex [BiRh(OTfa)4] (9), free of any axial ligands, can be conveniently
isolated by ordinary flash chromatography (Scheme 3). This simple
procedure gave an improved yield of 82% and never met with failure, such
that this precious complex is now deemed a well-accessible starting point
for further investigations.
strong that it persists even upon heating to 130-150°C at 10 mbar.[16]
3
Gratifyingly though, MeCN as solvent is able to displace the axial MEPY-H
ligands to give a red solution, which was adsorbed on ordinary silica.
Rinsing of the resulting deep red solid material with MeCN allowed the
excess 7 to be eluted and hence its re-coordination to be prevented. The
product was then desorbed with MeOH, the filtrate evaporated, and the
resulting solid material dried in vacuo at 100°C to give a turquoise powder,
which was triturated with MeCN in order to provide a stabilizing yet
defined axial ligand. This colorful and convenient procedure is described in
detail in the Experimental Section and photographs are contained in the
Supporting Information; it afforded [Rh2(5S-MEPY)4]2MeCN (1b) in
The trifluoroacetate groups of 9 can be readily displaced by other
carboxylic acids (or other ligands);[20-22] for the low boiling point of the
released TFA, which is again trapped by the Soxhlet method alluded to
above,[2] the reaction proceeds well even in toluene rather than
chlorobenzene. The examples shown in Scheme 4 using HOAc or the well-
proven amino acid derivatives 14[23,24] and 15[ 25-27] illustrate this point. The
resulting complexes 10 or 11 perform particularly well in intermolecular
cyclopropanation reactions of donor-acceptor carbenes, for which
[Rh2(MEPY)4] is not qualified. The level of asymmetric induction clearly
surpasses that reached with the corresponding homobimetallic [Rh2]
complexes carrying the exact same paddlewheel ligands,[13] even though
the rate of reaction is usually slower;[20,28] the comparison shown in Table 1
1
CCDC 2068752 (1b), 2068750 (10a), 2068749 (12), and 2068751 (13) contain the
supplementary crystallographic data for this work. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
2
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
is representative. Catalysts of this type operate by adopting an ,,,-
conformation: the N-phthaloyl-protected amino acid residues point in one
direction and craft a chiral binding site about the Rh-atom as proven by
crystallographic data for an actual reactive carbene intermediate;[ 29-32] this
conformation persists in solution and is likely selectivity-determining.[29,33-
36]
The larger ion radius of Bi(+2) imparts a conical shape onto the ligand
sphere, which in turn makes the chiral binding site about Rh tighter and
more effective.[13] Equally important is the fact that the Bi(+2) center,
though solvent exposed and surrounded only by the “achiral” tBu- or
adamantyl parts of the ligands, does not decompose a diazo derivative
and therefore does not lead to any racemic background reaction.[13] The
structure of 10a in the solid state illustrates these structural attributes
(Figure 2).
Figure 3. Structure of [BiRh(OAc)4] (12) in the solid state; the Rh-atoms account for
the formation of the coordination polymer, whereas the Bi-centers are unligated.1
Table 1. Preliminary Evaluationa
Nr
Substrate
Catalyst
T
Prod
Yield
(%)
nd
ee
(%)
93
(°C)
40b
40b
20b
20b
20b
20
1
2
1b
3
n. r.
17
2
2
13
---
37c
---
rac.
---
---
75
3
16
16
16
16
16
16
16
16
16
16
1b
4
13
[BiRh(hp)4]
[Rh2(S-PTTL)4]
[Rh2(S-PTTL)4]
11a
n. r.
n.r.
17
---
5
---
6
7
90
40
20
17
92
79
8
17
87
86
97
9
10
11
11a
10
10
10
10
17
88
10a
17
52c
85c
87c
97
10b
17
88
98
12
11b
17
a
all reactions were performed in pentane; yields of isolated pure compounds, unless
stated otherwise; b in CH2Cl2; c NMR yield relative to mesitylene as internal standard;
n. r. = no reaction; nd = not determined
Scheme 4. a) Acid, toluene, reflux/Soxhlet (K2CO3), 88% (10a, from 14a); 92% (10b,
from 14b); 95% (11a, from 15a); 99% (11b, from 15b); 94% (12, from HOAc); b) 7,
chlorobenzene, reflux/Soxhlet (K2CO3), 81%
The same ligand exchange process cannot be used to make [BiRh(5S-
MEPY)4], likely because the pK values of the trifluoroacetate ligands in 9
and the amide moiety of MEPYH (7) are not well-matching.2 This
problem was solved by resorting to [BiRh(OAc)4] (12)[21] as the starting
material: even though this complex is a coordination polymer in the solid
state (Figure 3), the more basic acetate ligands exchange with MEPYH
without incident to form [BiRh(5S-MEPY)4]MeCN (13) after
recrystallization from MeCN/iPrOH.
2 The exact mechanism of ligand exchange is unknown; it is reasonable to assume,
however, that one of the initial steps consists in the reversible protonation of the
carbonyl group of the leaving group by the protic incoming ligand. Since the amide
group of MEPY is much less acidic that the carboxylic acids 14 or 15, the use of
complex 12 carrying more basic acetate ligands (rather than trifluoroacetates) was
thought to be advantageous.
Figure 2. Structure of [BiRh(S-PTAD)4] (10a) in the solid state; H-atoms and solute
CHCl3 in the unit cell not shown for clarity; for the full structure, see the Supporting
Information.1
3
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
The structure of 13 in the solid state (Figure 4) is very different from that
of its [Rh2]-congener 1b (Figure 1). For the high oxophilicity of bismuth, all
four O-atoms of the MEPY-ligands bind to this main group metal, thus
forcing the four N-atoms of the amidates to ligate the catalytically active
Rh-center where they form a very tight and arguably rigid C4-symmetric
chiral binding site ([4,0]O-arrangement). Such self-sorting of the donor
atoms is not observed in 1b, which is C2-symmetric; the two Rh centers are
equivalent, featuring mixed coordination spheres comprised of two
mutually cis-configured N- and two O-donor atoms each.
As expected, the computed and the experimental structures of 1b and 13
in the solid state are in good agreement (for details, see the Supporting
Information). The match pertains to the bond distances as well as to the
conformational details, which were well reproduced in either case.
Removal of the axial MeCN ligands as a necessary prelude to any reaction
with a diazo derivative is largely inconsequential in structural terms, as
long as implicit solvent effects of CH2Cl2 were emulated by a conductor-
like polarizable continuum model (CPCM).3 The generated MO diagrams
of “bare” [Rh2(5S-MEPY)4] and [BiRh(5S-MEPY)4], however, reveal the
markedly different electronic nature of these complexes (Figure 5). The
LUMO is strongly metal-centered in either case but noticeably higher in
energy in the heterobimetallic variant (2.55 eV in 13 versus 3.71 eV in 1):
this situation is unfavorable as it renders attack of a diazo derivative onto
the catalyst difficult. In addition to this electronic handicap, the very
narrow pore formed by the C4-symmetric arrangement of the esters about
the axial binding site of 13 likely constitutes an additional steric
impediment for substrate binding.
The change in the orientation of the chiral ligands enforced by
incorporation of Bi(+2) has dramatic consequences: complex 13 was found
incapable of decomposing diazoesters 2 or 16 to the corresponding metal
carbenes; it hence proved inadequate for the formation of 3 by catalytic
intramolecular cyclopropanation, as we had originally hoped. This failure
came as a surprise in view of the good experiences with other chiral [BiRh]
paddlewheel complexes such as 10 and 11 (Table 1).[13] We are aware of
only one complex carrying oxypyridinate (hp) ligands which shows the
same “[4,0]O-pattern”;[37] since no reactivity data had been reported for
[BiRh(hp)4]H2O, we re-prepared this sterically unencumbered complex
and found it to be unreactive too vis-à-vis diazoester 16 (Scheme 4, Table
1). Therefore, electronic rather than geometric factors seem to be largely
accountable for the missing performance, for which we tried to gain better
understanding by computational means.
The HOMO of 13 features significant lobes on the Rh-atom as well as on
the paddlewheel ligands, but not on Bi. Even though the Rh-atom is
surrounded by all four N-atoms as the supposedly stronger donor sites of
an amidate ligand, the HOMO of 13 (4.65 eV) is lower-lying in energetic
terms than that of 1 (4.21 eV). Back-donation from the filled rhodium d-
orbitals to
a bound diazo derivative, however, is necessary for the
extrusion of N2 and the formation of the corresponding carbene complex
to occur. The interactions between the 4d orbitals of [N4]-ligated rhodium
with the 6p orbitals of [O4]-ligated bismuth are obviously weak and not
outweighed by the particular positioning of the heteroatom donors.
Figure 4. Structure of one of the two independent molecules of [BiRh2(5S-
MEPY)4]MeCN (13) in the unit cell, which differ from each other only in minor
conformational detail; H-atoms not shown for clarity.1
Computational Study. A previous investigation from our laboratory into
the electronic character of achiral homo- and heterobimetallic carbene
complexes[28] had been based on all-electron calculations using the BP86
functional[38] and the valence triple- basis set of the Karlsruhe group
together with the scalar relativistic 0th-order regular approximation (ZORA
Hamiltonian), as implemented in the ORCA program suite.[39] The
computational results were carefully calibrated against advanced
spectroscopic data (UV/Vis, resonance-Raman), which proved that the
chosen level of theory was fully adequate;[28] it was therefore also used in
the present context.
Figure 5. Molecular orbital scheme for bare [Rh2(5S-MEPY))4] (1, left) and [BiRh(5S-
MEPY)4] (13, right); the structures were truncated for sake of clarity; energy in eV (for
further details, see the SI).
3 When the calculations are carried out in the gas phase, a conformational change is
observed for [Rh2(MEPY)4] in that one of the esters on each end of the dinuclear
complex turns around to engage the –COOMe atom with the Rh center; this
interaction has a node that raises the energy of the LUMO. This interaction vanishes
when the continuum solvent model for CH2Cl2 is applied.
4
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
coupling constants (J) in Hz. The solvent signals were used as references
and the chemical shifts converted to the TMS scale (CDCl3: δC ≡77.2 ppm;
residual CHCl3 in CDCl3: δH ≡ 7.26 ppm; CD2Cl2: δC ≡ 53.8 ppm; residual
CDHCl2: δH ≡5.32 ppm, [D6]-DMSO: δC ≡39.5 ppm, residual [D5]-DMSO in
[D6]-DMSO: δH ≡ 2.50 ppm). Signal assignments were established using
HSQC, HMBC, COSY, NOESY and other 2D experiments. All commercially
available compounds (abcr, Acros Organics, Alfa Aesar, FluoroChem, TCI
Chemicals, Sigma-Aldrich, Strem Chemicals) were used as received, unless
stated otherwise. For the preparation of previously unknown or not fully
characterized chiral ligands, see the Supporting Information.
The resulting large HOMO/LUMO gap in 13 (E = 2.10 eV versus 0.50 eV in
1) in combination with a narrow binding site likely accounts for the poor
reactivity of [BiRh(5S-MEPY)4]; this notion is in accord with previous
conclusions that carbene formation is the rate-determining step in
reactions catalyzed by achiral [BiRh] complexes.[20] From the conceptual
viewpoint, this result implies that metal-metal bonding is
a prime
reactivity-determinant.[28,40] In order to harness the full potential of
heterobimetallic cooperation in (chiral) carbene chemistry, future catalyst
design must primarily aim at properly adjusting the net effect of the
internal metallo-ligand as the arguably most critical parameter.
Conclusions
[Rh2(5S-MEPY)4]2MeCN (1b). A two-necked 50 mL, round bottom flask
equipped with a magnetic stir bar, an inert gas adapter on the side neck,
and a Soxhlet apparatus topped by a reflux condenser, which also carries
an inert gas adapter, was evacuated, flame-dried, and allowed to cool to
The literature route to [Rh2(MEPY)4] (1), a venerable catalyst in the area of
rhodium carbene chemistry, has been improved in terms of practicality
and yield. The same is true for the previously rather cumbersome
approaches to heterobimetallic [BiRh] paddlewheel complexes, which are
now readily available too. While chiral [BiRh] complexes comprising amino
acid derived carboxylate ligands hold great promise for enantioselective
catalysis,[13,14] the carboxamidate derivative [BiRh(MEPY)4] (13) proved
unreactive. This unexpected observation shows that it does not suffice to
simply extrapolate successful concepts from dirhodium chemistry to the
heterobimetallic estate; rather, the effect of the main-group metallo-
ligand is the dominant reactivity-determining factor. Future work must
hence try to gain deeper understanding for and better control over this
particular parameter. For our long-standing interest in various aspects of
metal carbene chemistry in general,[41-49] we are committed to doing so;
encouraging lead findings along these lines have recently disclosed.[50]
ambient temperature before it was refilled with argon.
A thimble
containing a layer of dry sand and a layer of dry K2CO3 (4.4 g) was
introduced into the Soxhlet apparatus under Ar before the gas adapter on
the side neck was replaced by a glass stopper (instead of the Soxhlet
extractor with the thimble, it is also possible to use a frit with a bridging
side arm; a photograph of this set-up is contained in the Supporting
Information).
The flask was charged with chlorobenzene (22.5 mL), which had been
degassed prior to use by passing a stream of Ar through it for 20 min. Next,
commercial [Rh2(OAc)4] (300.0 mg, 679 µmol) was added, followed by
methyl 2-pyrrolidone-5S-carboxylate (7, 5S-MEPYH, 651.0 mg, 4.7
mmol). The flask was immersed into a pre-heated oil bath (145°C) and the
mixture stirred at this temperature for 13 h, causing a color change from
Experimental Section
4
green to dark red. The mixture was allowed to reach ambient
temperature and the solvent was evaporated in high vacuum (10 mbar)
3
General Information
to give a deep blue/violet solid material. An aliquot of this compound was
purified by subliming the excess ligand and any impurities off (130°C, 10
3
Unless stated otherwise, all reactions were carried out in flame-dried
glassware using anhydrous solvents under argon. The solvents and
reagents were purified by distillation over the indicated drying agents and
were transferred under argon: chlorobenzene (CaH2), toluene (Na/K alloy);
pentane, MeCN, Et3N (absorption solvent purification system based on
molecular sieves). For the synthesis of 1b, MeCN can be used as received.
Flash chromatography: Merck silica gel 40-63 μm with predistilled or HPLC
grade solvents. Preparative HPLC separations were carried out on an
Agilent 1260 Infinity II Preparative LC System. GC analyses: Shimadzu
GCMS-QP2010 Ultra instrument with a FID detector. IR: Alpha Platinum
mbar), leaving
a sample of analytically pure [Rh2(5S-MEPY)4]2(S-
MEPYH)2 (1c) behind; the spectral data are compiled below.
The crude blue compound was dissolved in MeCN (20 mL) under air to give
a dark red solution. Silica gel (8 g) was added with stirring, resulting in
decolorization of the solution. The now dark red solid material was filtered
off and carefully rinsed with MeCN (3 x 50 mL); the combined MeCN
filtrates were discarded. The still red silica was then washed with MeOH (3
x 50 mL) until it was colorless, and the combined blue MeOH filtrates were
evaporated on a rotary evaporator to leave a violet solid material. This
purification step was repeated three times.
ATR instrument (Bruker), wavenumbers (푣̃
) in cm−1. MS: ESI-MS: ESQ 3000
A flame-dried Schlenk flask containing a magnetic stir bar was charged
(Bruker); EI: MAT 8200 (70 eV, Finnigan); accurate mass determinations:
Bruker APEX III FT-MS (7 T magnet), Mat 95 (Finnigan), Thermo Scientific
LTQ-FT, or Thermo Scientific Exactive. Optical rotations ([훼]): A-Krüss
Otronic Model P8000-t polarimeter, wavelength: 589 nm, concentration:
(c/(10 mg/mL)). NMR: Bruker DPX 300, AVIII 400, AVIII 600 or AV600neo
spectrometers (the latter two equipped with cryoprobes) in the solvents
indicated at 25°C; chemical shifts (δ) are given in ppm relative to TMS,

3
with this solid material. The compound was then heated in vacuum (10
4 Reaction time and temperature are important parameters to avoid accumulation of
isomeric [Rh2(5S-MEPY)4] complexes and complexes resulting from incomplete
substitution of the acetate ligands by MEPY, which result in lower yields of product
and/or may compromise the performance of the catalyst, see ref. 15
5
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
mbar) for 14 h to 100°C with gentle stirring until the solid had taken a
turquoise color. The flask was refilled with Ar and allowed to cool to
ambient temperature before dry MeCN (1 mL) was introduced. After
stirring for 5 min, the remaining solvent was evaporated and the product
dried in vacuum to furnish the title complex as a red/violet solid material
was purified by flash chromatography (CH2Cl2/EtOAc, 99:1) to obtain the
title complex as a yellow solid (110 mg, 99%). 1H NMR analysis showed
that one equivalent of EtOAc is bound to the complex. 1H NMR (400 MHz,
CDCl3): δ = 7.90 (d, J = 4.3 Hz, 4H), 7.76 – 7.62 (m, 8H), 4.89 (d, J = 3.9 Hz,
4H), 4.11 (q, J = 6.8 Hz, 2H), 2.04 (d, J = 4.2 Hz, 3H), 1.38 (d, J = 3.9 Hz, 36H),
1.25 (t, J = 5.6 Hz, 3H), 1.15 ppm (d, J = 4.2 Hz, 36H); 13C NMR (101 MHz,
CDCl3): δ = 181.9, 171.6, 168.3, 167.9, 158.3, 132.3, 130.8, 129.5, 123.3, 120.8,
(473 mg, 81%).
= −335.7 (0.014 g/100 mL, CHCl3); 1H NMR (600 MHz,
CDCl3) δ = 4.32  4.27 (m, 2H), 3.95 (d, J = 8.7 Hz, 2H), 3.71 (s, 6H), 3.68 (s,
6H), 2.63  2.59 (m, 4H), 2.31  2.21 (m, 10H), 2.20  211 (m, 4H), 1.97 
1.82 ppm (m, 4H); 13C NMR (151 MHz, CDCl3) δ = 188.6, 188.3, 175.5, 175.2,
115.4, 66.8, 66.7, 52.1, 51.9, 31.8, 31.6, 26.1, 25.4, 3.07 ppm; IR (film)
61.4, 60.6, 35.9, 35.8, 31.3, 28.2, 21.2, 14.3 ppm; IR (ATR): 푣̃ = 1776, 1713,
1591, 1367, 1257, 1185, 1103, 840, 785, 753, 693, 565 cm-1; HRMS (ESI+) for
C72H88BiN4O16Rh [M+Na]+: calcd: 1599.49464, found: 1599.49642.
푣̃ = 2950, 1729, 1608, 1428, 1279, 1193, 1168, 1117, 1043, 987, 686,
[BiRh(S-tertPTAD)4]EtOAc (10b). Prepared analogously from 9 and acid
14b as a yellow solid (205 mg, 92%). 1H NMR analysis showed that one
595 cm−1; HRMS (ESI+): m/z calcd. for C28H38N6O12Rh2 [M+Na−(2×MeCN)]+:
797.00190; found: 797.00231.
1
equivalent of EtOAc is bound to the complex. H NMR (400 MHz, CDCl3): δ
= 7.88 (d, J = 11.6 Hz, 4H), 7.74 – 7.65 (m, 8H), 4.70 (s, 4H), 4.11 (q, J = 7.1 Hz,
2H), 2.04 (s, 3H), 1.97 ppm (d, J = 12.9 Hz, 24H), 1.71 (dd, J = 17.2, 9.3 Hz,
36H), 1.38 (s, 36H), 1.25 ppm (t, J = 7.1 Hz, 3H); 13C NMR (101 MHz, CDCl3):
δ = 181.6, 171.3, 168.5, 168.1, 158.3, 132.3, 130.8, 129.5, 123.3, 120.8, 62.1,
[[Rh2(5S-MEPY)4]2(S-MEPYH)2 (1c). Prepared by sublimation as
1
described above; H NMR (400 MHz, CDCl3) δ = 8.62 (s, 2H), 4.34 (dd, J =
8.3, 6.7 Hz, 2H), 4.23 (dd, J = 8.8, 4.2 Hz, 2H), 3.97 (d, J = 7.0 Hz, 2H), 3.75 (s,
6H), 3.71 (s, 6H), 3.62 (s, 6H), 2.69  2.50 (m, 8H), 2.50  2.31 (m, 4H), 2.30
 2.18 (m, 4H), 2.18  2.08 (m, 4H), 1.93  1.78 ppm (m, 4H): 13C NMR
(100 MHz, CDCl3): δ = 187.7, 187.5, 182.1, 175.7, 175.1, 171.6, 65.6, 65.1, 56.9,
60.5, 39.5, 38.0, 37.1, 35.8, 31.3, 28.9, 21.2, 14.3 ppm; IR (ATR): 푣̃ = 2902,
1775, 1712, 1591, 1365, 1349, 1259, 1239, 1093, 1044, 841, 757, 693, 671, 564
cm-1; HRMS (ESI+) for C96H112BiN4O16Rh [M+Na]+: calcd: 1911.68244, found:
1911.68500.
52.4, 51.9, 51.8, 31.8, 31.5, 30.8, 25.9, 25.7, 24.2 ppm; IR (film) 푣̃ = 2950,
1734, 1666, 1606, 1426, 1356, 1277, 1236, 1193, 1168, 1116, 1042, 988, 926,
794, 685, 594, 509 cm−1; HRMS (ESI): m/z calcd. for C36H50N6O18Rh2 [M+]:
1060.12862; found: 1060.12861.
[BiRh(OAc)4] (12). Prepared analogously from 9 and HOAc (415 mg, 94%).
1H NMR (400 MHz, [D6]-DMSO): δ = 1.88 ppm (s, 12H); 13C NMR (101 MHz,
[D6]-DMSO): δ = 184.9, 22.2 ppm; IR (ATR): 푣̃ = 2939, 1530, 1484, 1378,
[BiRh(OTfa)4] (9). [Rh2(OTfa)4]∙2MeCN (8) (534 mg, 0.722 mmol) was
heated to 80°C under high vacuum (10−3 mbar) for 1 h to remove the axially
bound ligands; during this time, the color of the sample changed from
purple to green. Next, Bi(OTfa)3 (415 mg, 0.757 mmol),[51] freshly ground Bi
metal (817 mg, 3.91 mmol), toluene (40 mL), Ph2O (1.1 mL, 6.93 mmol) and
trifluoroacetic acid (200 µL, 2.61 mmol) were added. The mixture was
stirred at 115°C bath temperature for 16 h and the reaction progress was
monitored by 19F NMR analysis. The remaining Bi metal was allowed to
settle and the supernatant was removed via cannula. The yellow filtrate
was concentrated in vacuo. Most of the remaining Ph2O was sublimed
onto a cold (−30°C) sublimation finger at 50°C under high vacuum (10−3
mbar). The residue was purified by flash chromatography (toluene/MeCN,
gradient 100:0 → 90:10); some black residue sticks on top of the column,
whereas a broad yellow-brown band containing the product was collected
(for photographs, see the Supporting Information). Evaporation of the
product-containing fraction afforded the title complex as a yellow powder
(855 mg, 82%). 19F{1H} NMR (282 MHz, C6D6): δ 74.2 ppm.
1345, 1261, 1042, 804, 688, 623, 545 cm−1; HRMS (ESI+) for C8H12BiO8Rh
[M+Na]+: calcd: 570.92833, found: 570.92784. The spectral data matched
those reported in the literature.[21] Single crystals suitable for X-ray
diffraction were grown by dissolving the sample in the minimum amount
of hot MeOH, filtration of the turbid mixture, and slow cooling of the clear
filtrate to ambient temperature.
[BiRh(5S-MEPY)4] (13). A mixture of [BiRh(OAc)4] (12, 101 mg, 0.184
mmol) and methyl 2-pyrrolidone-5S-carboxylate (7, 208 mg, 1.46 mmol) in
dry chlorobenzene (15 mL) was stirred at reflux temperature for 12 h,
passing the condensing vapor through a Soxhlet apparatus containing a
thimble filled with K2CO3. The mixture was concentrated in vacuo and the
residue was purified by preparative HPLC (150 mm Kromasil 5-C18, 5 µm,
Ø 30 mm, MeCN/H2O = 30/70, v = 35 mL/min, 308 K, t = 4.12 min) to obtain
the title complex as a yellow solid (82 mg, 51%). Single crystals of the
cetonitrile adduct suitable for X-ray diffraction were grown from
a
1
saturated MeCN/2-propanol solution. H NMR (400 MHz, CD2Cl2): δ = 4.19
– 4.13 (m, 4H), 3.73 (s, 12H), 2.44 – 2.16 (m, 10H), 2.01 – 1.89 ppm (m, 6H);
13C NMR (101 MHz, CD2Cl2): δ = 184.5, 176.6, 65.3, 52.6, 32.7, 28.5 ppm; IR
[BiRh(S-tertPTTL)4]EtOAc (11b). A mixture of complex 9 (51 mg, 0.067
mmol) and acid 15b (106 mg, 0.335 mmol) in dry toluene (20 mL) was
stirred at reflux temperature for 3 h, passing the condensing vapor
through a Soxhlet apparatus containing a thimble filled with K2CO3; the
reaction progress was monitored by 19F NMR. Once full conversion was
reached (ca. 3 h), the mixture was concentrated in vacuum and the residue
(ATR): 푣̃ = 2926, 2854, 1735, 1667, 1583, 1535, 1454, 1412, 1357, 1299, 1232,
1195, 1172, 1109, 1023, 986, 922, 794, 769, 686, 538 cm−1; HRMS (ESI+) for
C24H32BiN4O12Rh [M+Na]+: calcd: 903.07678, found: 903.07672.
6
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
Representative Procedure for Cyclopropanation Reactions. Methyl
(1S,2R)-1-(4-methoxyphenyl)-2-phenylcyclopropane-1-carboxylate (17).
A solution of the diazo ester 16 (20.6 mg, 0.1 mmol) in pentane (2 mL) was
added to a solution of complex 11bEtOAc (1.7 mg, 0.001 mmol) and
styrene (57 µL, 0.5 mmol) in pentane (3 mL) at 10°C. The resulting
mixture was stirred for 24 h at 10°C before all volatile materials were
evaporated under reduced pressure. The residue was purified by flash
chromatography (hexane/EtOAc, 20:1) to give the title compound as a
colorless amorphous solid (24.5 mg, 87%, 98% ee). The optical purity was
determined by HPLC (Chiralpak IB,  4.6 mm, 2% 2-propanol in n-
heptane, 1 mL/min, 20 min, UV 225 nm): tR 7.65 min (major) and tR 8.48
min (minor); [훼] = +4.6 (c = 1.6, CHCl3) [ref.: [훼] = +5.1 (c = 1, CHCl3)];[52]
1H NMR (400 MHz, CDCl3): δ 7.09 – 7.04 (m, 3H), 6.96 – 6.91 (m, 2H), 6.80 –
6.74 (m, 2H), 6.69 – 6.63 (m, 2H), 3.72 (s, 3H), 3.66 (s, 3H), 3.07 (dd, J = 9.4,
7.4 Hz, 1H), 2.12 (dd, J = 9.4, 4.8 Hz, 1H), 1.82 ppm (dd, J = 7.2, 4.8 Hz, 1H);
13C NMR (101 MHz, CDCl3): δ 174.8, 158.6, 136.6, 133.0, 128.2, 127.8, 126.9,
126.4, 113.3, 55.2, 52.8, 36.8, 33.3, 20.9 ppm. The recorded data are
consistent with those reported in the literature.[52]
Supplementary Material
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/MS-number.
Acknowledgements
Generous financial support by the Fonds der Chemischen Industrie (Kekulé
stipend to F. P. C.) and the Max-Planck-Gesellschaft is gratefully
acknowledged. We thank Ms. V. A. Tran, Dr. C. Poidevin, Dr. D. Yepes and
Dr. M. van Gastel for advice concerning the computational studies, Ms. C.
Wirtz for NMR assistance, Mr. J. Rust. Dr. R. Goddard and Prof. C. W.
Lehmann for solving the X-ray structures, and the analytical service groups
of our Institute for excellent support.
[15] C. J. Welch, Q. Tu, T. Wang, C. Raab, P. Wang, X. Jia, X. Bu, D. Bykowski, B.
Hohenstaufen, M. P. Doyle, ‘Observations of Rhodium-Containing Reaction
Intermediates using HPLC with ICP-MS and ESI-MS Detection’, Adv. Synth.
Catal. 2006, 348, 821-825.
[16] For a similar complex comprising bridging and axial pyrrolidin-2-one ligands,
see: J. L. Baer, R. S. Lifsey, L. K. Chau, M. O. Ahsan, J. D. Korp, M. Chavan, K.
M. Kadish, ‘Structural, Spectroscopic, and Electrochemical Characterization
Author Contribution Statement
[+] These authors contributed equally
of
Tetrakis-µ-(2-pyrrolidinonato)-dirhodium(II)
and
Tetrakis-µ-(-
valerolactamato)-dirhodium(II)’, J. Chem. Soc. Dalton Trans. 1989, 93-100.
M. B., L. R. C., F. P. C., S. S. performed the experiments, analyzed and
compiled the data, L. E. L. optimized the conditions for the formation of 1
and performed the computational studies, A. F. conceived the project and
wrote the manuscript.

References




7
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA





[46] A. Guthertz, M. Leutzsch, L. M. Wolf, P. Gupta, S. M. Rummelt, R. Goddard, C.
Farès, W. Thiel, A. Fürstner, ‘Half-Sandwich Ruthenium Carbene Complexes
Link trans-Hydrogenation and gem-Hydrogenation of Internal Alkynes’, J. Am.
Chem. Soc. 2018, 140, 31546-3168.

[49] T. Biberger, R. J. Zachmann, A. Fürstner, ‘Grubbs Metathesis Empowered by a
Light-Driven gem-Hydrogenation of Internal Alkynes’, Angew. Chem. Int. Ed.
2020, 59, 18423-18429.
[50] S. Singha, M. Buchsteiner, G. Bistoni, R. Goddard, A. Fürstner, ‘A New Ligand
Design Based on London Dispersion Empowers Chiral Bismuth-Rhodium
Paddlewheel Catalysts’, J. Am. Chem. Soc. 2021 (doi: 10.1021/jacs.1c01972).

‐
‐
‐

8
This article is protected by copyright. All rights reserved.

10.1002/hlca.202100042
Helvetica Chimica Acta
HELVETICA
Entry for the Table of Contents
9
This article is protected by copyright. All rights reserved.