The Forgotten Nitroaromatic Phosphines as Weakly Donating P-ligands: An N-Aryl-benzimidazolyl Series in RhCl(CO) Complexes

Article abstract of DOI:10.1002/asia.201701078

The coordination chemistry of a priori weakly σ-donating nitroaromatic phosphines is addressed through a series of nitro-substituted (N-phenyl-benzimidazol-1-yl)diphenylphosphines in Rh^I complexes. From a set of seven such phosphines L=L_xyz ^(′) (x, y, z=0 or 1=number of NO₂ substituents at the 5, 6 and N-Ph para positions, respectively), including the non-nitrated parent L₀₀₀ and its dicationic N-methyl counterpart L₀₀₀′, three LRhCl(COD) and seven L₂RhCl(CO) complexes have been obtained in 72–95 % yield. Despite of a cis orientation of the L and CO ligands, the C=O IR stretching frequency ν_CO varies in the expected sense, from 1967±1 cm^?1 for L_xy0 to 1978±1 cm^?1 for L_xy1, and 2005 cm^?1 for L₀₀₀′. The ¹⁰³Rh NMR chemical shift δ_Rh varies from ?288 ppm for L₀₀₀ to ?316±1 ppm for L_10z or L_01z, and ?436 ppm for L₀₀₀′. The ν_CO and δ_Rh probes thus reveal moderate but systematic variations, and act as “orthogonal” spectroscopic indicators of the presence of nitro groups on the N-Ph group and the benzimidazole core, respectively. For the dicationic ligand L₀₀₀′, a tight electrostatic sandwiching of the Rh-Cl bond by the benzimidazole moities is evidenced by X-ray crystallography (RhCl^δ????CN₂ ⁺ ≈3.01 ?). Along with the LRhCl(CO) complexes, dinuclear side-products (μ-CO)(RhClL)₂ were also obtained in low spectroscopic yield: for the dinitro ligand L=L₀₁₁, a unique 1:6.7 clathrate structure, with dichloromethane as solvate, is also revealed by X-ray crystallography.

Full text of DOI:10.1002/asia.201701078

CHEMISTRY
AN ASIAN JOURNAL
www.chemasianj.org
Accepted Article
Title: The forgotten nitro-aromatic phosphines as weakly donating P-
ligands: a N-aryl-benzimidazolyl series in RhCl(CO) complexes
Authors: Chonhwei Zhu, Emmanuel Gras, Carine Duhayon, Francis
Lacassin, Xiuling Cui, and Remi Chauvin
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: Chem. Asian J. 10.1002/asia.201701078
Link to VoR: http://dx.doi.org/10.1002/asia.201701078
A Journal of
A sister journal of Angewandte Chemie
and Chemistry – A European Journal

10.1002/asia.201701078
Chemistry - An Asian Journal
The forgotten nitro-aromatic phosphines as weakly donating P- ligands:
a N-aryl-benzimidazolyl series in RhCl(CO) complexes†
Chongwei Zhu,[a,b] Emmanuel Gras,[a,b]* Carine Duhayon,[a,b] Francis Lacassin,[a,b]
Xiuling Cui,[c] Remi Chauvin[a,b,c]
*
[a] CNRS, LCC (Laboratoire de Chimie de Coordination), 205 route de Narbonne, BP44099,
31077 Toulouse Cedex 4 (France). Fax: (+33)5 61 55 30 03. E-mail: emmanuel.gras@lcc-
toulouse.fr, chauvin@lcc-toulouse.fr
[b] Université de Toulouse, UPS, ICT-FR 2599, 31062 Toulouse Cedex 9 (France)
[c]
Engineering Research Center of Molecular Medicine, Ministry of Education, Key
Laboratory of Xiamen Marine and Gene Drugs, Institutes of Molecular Medicine and School
of Biomedical Sciences, Huaqiao University, Xiamen, 361021 (China).
Keywords: benzimidazolophosphines, nitroarylphosphines, coordination bond, rhodium
carbonyl complexes, nitro-aromatic chemistry
† This paper is dedicated to Prof. Marian Mikolajczyk on the occasion of his 80th birthday.
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
Abstract. The coordination chemistry of the a priori weakly s-donating nitro-aromatic
phosphines is addressed through a series of nitro-substituted (N-phenyl-benzimidazol-1-
yl)diphenylphosphines in Rh(I) complexes. From a set of seven such phosphines L = Lxyz(') (x,
y, z = 0 or 1 = number of NO2 substituent at the 5, 6 and N-Ph para positions, respectively),
including the non-nitrated parent L000 and its dicationic N-methyl counterpart L000', three
LRhCl(COD) and seven L2RhCl(CO) complexes have been obtained in 72-95 % yield.
Despite of a cis orientation of the L and CO ligands, the C=O IR stretching frequency nCO
varies in the expected sense, from 1967 ± 1 cm–1 for Lxy0 to 1978 ± 1 cm–1 for Lxy1, and 2005
cm–1 for L000'. The 103Rh NMR chemical shift dRh varies from –288 ppm for L000 to –316 ± 1
ppm for L10z or L01z, and –436 ppm for L000'. The nCO and dRh probes thus reveal moderate but
systematic variations, and act as "orthogonal" spectroscopic indicators of the presence of nitro
groups on the N-Ph group and the benzimidazole core, respectively. For the dicationic ligand
L000', a tight electrostatic sandwiching of the Rh–Cl bond by the benzimidazole moities is
+
evidenced by X-ray crystallography (RhCld–…CN2 ≈ 3.01 Å). Along with the LRhCl(CO)
complexes, dinuclear side-products (µ-CO)(RhClL)2 were also obtained in low spectroscopic
yield: for the dinitro ligand L = L011, a unique 1:6.7 clathrate structure, with dichloromethane
as solvate, is also revealed by X-ray crystallography.
2
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
1. Introduction
In the continuing debate on the fundamental nature of the coordination bond,[1] the ligands of
some reference Lewis acid (LA) are commonly classified between the electron-rich and
electron-poor categories: for transition metal LAs, this pre-bonding characteristic makes it
possible to predict strong or weak ligand-to-metal charge transfer. Beside the long studied
class of electron-rich ligands, ranging from alkylphosphines to diaminocarbenes and onium
ylides,[2] the class of electron-poor phosphane counterparts has attracted less systematic
attention but remains challenging for appraising coordination limits while decreasing the
strength of s-donation.[3] The need for electron-poor ligands is, however, more practically
motivated by catalysis issues for rate-determining steps requiring a relatively low electron
density at a metal center balanced by sufficient stability of the complex. Even weak, s-
donation can indeed be compensated by strong π-back bonding, as it happens with phosphites
identified as ligands of choice for catalytic transformations such as hydroformylation.[4] As
esters of phosphorous acid, phosphites are however sensitive to P–O bond protolysis, and two
main kinds of more robust electron-deficient triarylphosphine ligands can be distinguished:
the electron-poor neutral phosphines, mainly represented by fluoroarylphosphine,[ 5 ] N-
heteroarylphosphines (e.g. imidazolophosphines),[6] and the highly electron-poor a-cationic
phosphines, such as imidazolio-, diaminocyclopropenio-, or pyridinio-phosphines.[6a, , ] These
7 8
carbenio-phosphines, optimally described as carbene-phosphenium adducts,[ 9 ] have been
shown to exhibit P-donating ability in ternary C→P→LA coordination motifs,[7a] even in the
case of the extremely electron-poor a-diationic dicyclopropeniophosphines.[ 10 ] While
cyclopropenio-phosphines give particularly stable Pd complexes acting as efficient catalysts
7d,
]
11
(e.g. for enynes cyclo-isomerization),[
imidazolio-phosphines and their complexes are
more reactive, undergoing P–C bond cleavage in the presence of weak nucleophiles (such as
9,12]
Cl– ), thus restricting the scope of possible applications.[
3
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
With the view to enhancing the electron-deficiency of the parent imidazolophosphines
+
without the detrimental effect of the a-positive charge, N-methylation of the formers (CH3
N-coordination) could be replaced by CH-substitution with neutral electron-withdrawing
groups (Figure 1). Examination of this issue drew the authors' attention to the fact that an a
priori simple way to make a triarylphosphine ligand more electron-deficient while remaining
neutral seems to have been forgotten: alternative to fluorination, nitration should indeed be
particularly efficient. Indeed, considering Hammett constants sp and sm as crude means of
comparison of the electron-withdrawing effect of a substituent through an aromatic system,[13]
one of the largest values is attained for NO2 (sm = 0.71, sp = 0.78), standing close to that of
+
the cationic substituent NMe3 (sm = 0.88, sp = 0.82), and much higher than those of other
neutral substituents such as F (sm = 0.34, sp = 0.06), CF3 (sm = 0.43, sp = 0.54) or CN (sm =
0.56, sp = 0.66).[14]
Prior to any coordination chemistry considerations, the oxymoronic red-ox nature of
nitro-substituted arylphosphines can be considered as a basic challenge,[15] at least from the
thermodynamic standpoint (the nitro-nitroso isomerization 4-NO2-C6H4-PPh2→ 4-NO-C6H4-
P(O)Ph2 is calculated to be exothermic by ∆ZPE = –23.2 kcal/mol; see SI).[16] Direct nitration
of aryl-phosphines are prevented by the P-oxidizing potency of the nitronium ion,[17] and most
of related synthetic efforts have focused on oxidized P=O or P=S precursors.[18] Coordination
complexes of nitroarylphosphines are actually limited to the formal cases of such P→O or
P→S derivatives: to the best of the authors' knowledge, indeed, no experimental example of
corresponding P→metal complexes has ever been reported,[19] thus unveiling a quite general
challenge. While the basic triphenylphosphine series will deserve a natural attention in
particular for the reference (4-nitrophenyl)diphenylphosphine (see conclusion),[15b, 20 ] the
forgotten nitration strategy for lowering further the electron deficiency of aromatic phosphine
4
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
ligands is hereafter addressed within the above-summarized concern in the benzimidazolyl
phosphine series (Figure 1a).
LA
LA
PPh2
LA
PPh2
Me
N
N
N
N
NO2
- H
N
Me
PPh2
INO2
O2N
I0
N
I+
ArNO2
Ar
Ar
neutral nitro-benzimidazolophosphines:
cationic benzimidazoliophosphine:
neutral benzimidazolophosphine:
targets...
reference, with stable P–C bond
know, with labile C–P bond
Hal
H
X
X
X
N
X
Y
N
N
Cl–PPh2
N
N
Z
PPh2
H
N
N
X, Y, Z = H, NO2
Hal = I, F
N
Y
Y
Y
H
Z
3a: X = Y = H
Z
2a: X = Y = H
3b: X = NO2, Y = H
3c: X = H, Y = NO2
3d: X = Y = NO2
3e: X = Y = H
2b: X = NO2, Y = H
2c: X = H, Y = NO2
2d: X = Y = NO2
2e: X = Y = H
Z = H
Z = H
1a: X = Y = H
1b = 1c: X = NO2, Y = H
1d: X = Y = NO2
3f: X = NO2, Y = H
3g: X = H, Y = NO2
3h: X = Y = NO2
2f: X = NO2, Y = H
2g: X = H, Y = NO2
2h: X = Y = NO2
Z = NO2
Z = NO2
Figure 1. a) top: alternative manners to tune the P-s-donating ability of
benzimidazolophosphines towards a Lewis acid (LA): C-nitration vs N-alkylation (expected
order of inductive effects: I0 < INO2 < I+); b) bottom: selected targets and synthetic approach.
While modern routes to arylphosphines consist in Cu-, Ni- or Pd-catalyzed Ar–P
coupling reactions from Ar-X + E-PPh2 (X = I, Br, OTf; E = H, Ph, SnBu3, Cl), nitro-ArX
substrates are rarely exemplified,[21] and have even been claimed to be incompatible with
generally efficient conditions.[22] Yet, the most employed approach to tertiary phosphines
remains based on nucleophilic substitution processes at P centers with lithium or Grignard
reactants, as also illustrated in the arene/heteroarene series,[23] in particular for imidazole and
6e,12b]
benzimidazole representatives.[
For the present objective, (N-aryl-1H-benzoimidazol-2-
yl)diphenylphosphines have thus been targeted from the more or less nitrated NH,N2CH
5
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
benzimidazole precursors 1a-d by substitution of ClPPh2 with lithium salts of N-aryl-N2CH
derivatives 2a-h (Figure 1b). The propensity of nitro-arene/heteroarene substrates to undergo
addition or redox processes in the presence of lithium reagents has indeed been reported for
alkyl lithiums only.[24] Moreoever, since N-methyl-5-nitro-1H-benzoimidazol-2-yl lithium has
been shown to undergo selective electrophilic chlorination with NCS,[25] reaction of the 1N-
phenyl congener 2b with the analogous Ph2P–Cl electrophile appears as a natural possibility
to prepare the set of eight more or less nitrated N-aryl-1H-benzoimidazolophosphines 3a-h
(Figure 1). Their coordinating properties were then envisaged to be compared by the nCO
stretching frequencies of carbonyl complexes thereof (LA = Ni(CO)3 in Figure 1a).[26 ] Instead
of Ni(CO)3 serving to define the Tolman's electronic parameters, LA = RhCl(CO) is here
preferred for the sake of comparison with former results on related complexes of chelating
bisimidazolophosphines,[ 27 ] and more generally because of the added value of the
complementary 103Rh NMR probe for rhodium-phosphine complexes.[28]
2. Results and Discussion
2.1. Synthesis of nitro-benzimidazolophosphines
The four NH,N2CH benzimidazole precursors 1a-d were provided either as commercially
available for 1a, or prepared following known procedures; the nitro derivative 1b or 1c (in
tautomeric equilibrium) was thus elaborated from 4-nitro-1,2-benzenediamine by
condensation of triethylorthoester in the presence of a catalytic amount of TiCl4 and the
dinitro derivative 1d was obtained by nitration of 1b-c with fuming HNO3 in concentrated
H2SO4 (see experimental section 4.1 and cited references).
N-phenylation of 1a and 1b-c was implemented by copper-catalyzed C–N coupling
from iodobenzene under conditions reported by Buchwald et al.,[29] yielding 2a in 90 %
yield,[30] and the isomers 2b and 2c in 64 % and 23 % yield, respectively, after separation by
6
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
column chromatography.[ 31 ] The method however failed to produce 2d from 1d, in
correlation with the very low N-nucleophilicity of 1d resulting from the cooperative effect of
the two nitro groups.
N-p-nitrophenylation of 1a-c was carried out by SNAr of 4-fluoro-nitrobenzene,
affording the known nitro derivative 2e in 95 % yield,[30] and the isomers 2f and 2g in 53 %
and 34 % yields, respectively.[32] Using the same procedure from 1d, the trinitro derivative 2h
was isolated in 86 % yield.
CuI (10 %)
1,10-phenanthroline
(20 %)
PhI (0.84 eq.)
O2N
N
N
N
N
+
Cs2CO3 (2.1 eq.)
o
O2N
TiCl4 (10 %)
CH(EtO)3
(1.2 eq.)
DMF, 100 C, 48 h
O2N
NH2
NH2
O2N
N
2b, 64%
2c, 23%
EtOH, r.t., 2h
N
H
p-NO2-C6H5F
(2 eq.)
NaOH (1 eq.)
1b or 1c, 80%
O2N
N
N
N
+
DMF, 70 oC, 5 h
N
O2N
p-NO2-C6H5F
(2 eq.)
X
X
X
N
N
2f, 53%
2g, 34%
NO2
NO2
NaOH (1 eq.)
DMF, 70 oC, 5 h
N
N
H
X
2e: X = H, 95%
2h: X = NO2, 86%
1a: X = H
1d: X = NO2
NO2
Scheme 1. Syntheses of nitro-functionalized benzimidazoles (for the 4,5-X2 isomer of 1d,
1d', see SI).
Phosphinylation of the lithium salts of 2a-c-2e-h by nucleophilic substitution of
ClPPh2 required optimization of the initial deprotonation step. Inspired by the above-
mentioned report on the selective deprotonation and reactivity of the N-methyl congener of
2b,[25] the use of lithium diisopropylamide (LDA) as a base in THF was envisaged, with a
strict control of the temperature at –78°C.[33] Applied to 2b, these conditions indeed allowed
isolation of 3b in 83 % yield, in the absence of any oxyazobenzene side-product.[ 33] The same
procedure gave access in 56 % yield to the reference non-nitrated benzimidazolophosphine
7
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
3a, previously reported to be available in 20 % yield only under metal free conditions (in the
presence of NEt3 as a base).[34] In a whole, application of the LDA procedure to 2a-c-2e-g
afforded 3a-c-3e-g in 45-83 % yield, while only trace amounts of 3h were obtained from the
trinitro substrate 2h. The phosphinylation selectivity was first evidenced by NMR
spectroscopy by the disappearance of the amidine C1H singlet signal at 8.10-8.90 ppm and the
appearance of a 31P{1H} singlet signal in the range -21.3 / -24.5 ppm.
X
Y
X
N
N
1) LDA (1.2 eq.), -78 oC ~ -40 oC, 1 h
N
N
H
PPh2
2) Cl–PPh2 (1.2 eq.), -78 oC ~ r.t.
THF, overnight
Y
Z
Z
2a: X = Y = H
3a: X = Y = H, 56 %
2b: X = NO2, Y = H
2c: X = H, Y = NO2
Z = H
3b: X = NO2, Y = H, 83 %
3c: X = H, Y = NO2, 70 %
2e: X = Y = H
3e: X = Y = H, 62 %
2f: X = NO2, Y = H
2g: X = H, Y = NO2
2h: X = Y = NO2
3f: X = NO2, Y = H, 77 %
3g: X = H, Y = NO2, 45 %
3h: X = Y = NO2, trace
Z = NO2
Scheme 2. Selective C-phosphinylation of nitro-functionalized benzimidazoles.
Single crystals of 3b, 3c, 3e and 3f deposited from solvent mixtures of pentane and
dichloromethane (DCM) were found suitable for X-ray diffraction (XRD) analysis (see SI).[35]
Notably, the structure of the 5-nitrobenzimidazolophosphine 3b was resolved in the
uncommon P61 space group of the hexagonal system. In all the structures, the C–P distances
are virtually identical (1.83 ± 0.01 Å) and the torsion angles between the N–Ar and
benzimidazole mean planes have a median value of 68.5 ± 1.3°. The amidine sp2-C1 atom is
found slightly pyramidalized, with a tilting angle of the C–P bond axis vs the imidazole ring
mean plane of ca 4.7 ± 0.3° for 3c and 3e-f, but reaching 11.3° for 3b.
8
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
2.2. Coordination complexes of nitro-benzimidazolophosphines
With the view to comparing the coordinating properties of the benzimidazolophosphines
through IR nCO values of RhCl(CO) complexes of type 4 (Scheme 3), 3a-c and 3e-h were first
sequentially reacted with [Rh(COD)Cl]2 (for cleavage of the µ-Cl bridges; COD = 1,5-
cyclooctadiene) before displacement of the COD ligand by a second equivalent of the
phosphine and a carbonyl ligand.
Reaction of 3b with half an equivalent of [Rh(COD)Cl]2 in DCM for 2 h at room
temperature led to the COD complex 4b in 95 % yield. P-Rh coordination was established by
the change of the 31P{1H} NMR signal, shifting from -22.88 to +19.00 ppm, and splitting
1
from a singlet to a doublet with a coupling constant of 150.7 Hz characteristic of a JPRh
coupling. Moving the nitro group from the benzimidazole core in 3b to the Ar N-substituent
in 3e did not alter the reactivity with the Rh(I) dimer, giving 4e in 93 % yield. Likewise,
reaction of the dinitro benzimidazolophosphine 3f with [Rh(COD)Cl]2 gave 4f in a similar 94
% yield. Further displacement of the COD ligand of 4b, 4e and 4f was achieved by adding
one additional equivalent of the respective phosphines 3b, 3e and 3f, followed by CO
bubbling for 15 min, thus affording the diphosphine RhCl(CO) complexes 5b, 5e and 5f in 74
± 2 % yield (Scheme 3). In each case, another phosphine-rhodium complex numbered as 6b,
6e or 6f was concomitantly formed as an initially non-identified side-product, the structural
assignment of which could be achieved by crystallographic analysis only (see section 2.4,
Figure 3).
9
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
• sequential:
CO
Rh
X
X
X
X
N
N
N
N
N
N
[Rh(cod)Cl]2
(0.5 eq.)
3b, or
3f or 3e
P
Ph2
PPh2
P
Ph2
P
Ph2
Rh
Cl
N
N
Cl
DCM, r.t.
2 h
CO bubbling
DCM, r.t.
15 min.
Z
Z
Z
Z
3b: X = NO2, Z = H
3e: X = H, Z = NO2
3f: X = Z = NO2
4b: X = NO2, Z = H, 95 %
4e: X = H, Z = NO2, 93 %
4f: X = Z = NO2, 94 %
5b: X = NO2, Z = H, 72 %
5e: X = H, Z =NO2, 74 %
5f: X = Z = NO2, 76 %
• one-pot:
CO
N
N
N
N
N
1) [Rh(cod)Cl]2 (0.25 eq.)
DCM, r.t., 2 h
P
Ph2
Rh
Cl
PPh2
P
Ph2
N
Y
Y
Y
2) CO bubbling
DCM, r.t., 15 min.
Z
Z
Z
3a: Y = Z = H
5a: Y = Z = H, 79 %
3c: Y = NO2, Z = H
3g: Y = Z = NO2
5c: Y = NO2, Z = H, 81 %
5g: Y = Z = NO2, 73 %
Scheme 3. Sequential and one-pot two-step syntheses of RhCl(CO) complexes of nitro-
substituted benzimidazolophosphines. From 6x, an initially non-identified side products
respectively denoted as 6x (x = b-c e-g) were also obtained (see Figure 3, section 2.4).
A one-pot procedure was also attempted, providing the target complexes with slightly
higher overall yields, as illustrated from 3a, 3c and 3g (Scheme 3). Two equivalents of 3c
were thus treated with half an equivalent of [Rh(COD)Cl]2 in DCM for 2 h at r.t., and
subsequently with bubbling carbon monoxide for 15 min., affording 5c in 81 % yield. The
same treatment from 3a and 3g providing directly 5a and 5g in 79 % and 73 % yield,
respectively. In all cases involving a nitro ligand (not in the case of 5a), the side-product of
initially unknown structure 6x (x = b-c, e-g) was also observed (see section 2.4).
For the sake of reference to the original concern, the RhCl(CO) complex 5a' of the
cationic benzimidazoliophosphine 3a' was also targeted. Treatment of 3a with one equivalent
of methyl triflate (MeOTf) in DCM at -78oC actually led to the N-methylated target product
3a' in mixture with the P-methylated isomer 3a" in a 5:1 ratio, as evidenced by the
31
disappearance of the P NMR signal of 3a (-24.49 ppm) and appearance of two signals at –
10
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
15.48 and +14.36 ppm consistent with the structures of 3a' and 3a", respectively. This lack of
N/P selectivity in the reaction of the hard methylating agent MeOTf with imidazolophosphine
is unprecedented, but correlated with the absence of a second P center in the monophosphine
3a as compared to the cases of the N-aryl benzimidazolo- and imidazolo-diphosphines
BIMINAP (2-(diphenylphosphanyl)-1-[2-(diphenylphosphanyl)naphthalen-1-yl]-1H-1,3-
benzodiazole) and BIPHIMIP (2-(diphenylphosphanyl)-1-[2-(diphenylphosphanyl)phenyl-
36
]
1H-imidazole).[6e,12b,
Under the same conditions, treatment of the 5-
nitrobenzimidazolophosphine 3b with MeOTf gave a mixture of the N- and P-methylated
isomers 3b' and 3b" in a 3:2 ratio according (see SI): this further lowering of N/P selectivity
can be interpreted by the electron-withdrawing effect of the nitro group decreasing the
electron density of the N atom more than that of the more remote P atom.
As the two cations 3a' and 3a" could not be efficiently separated, the 5:1 mixture was
used in coordination experiments. Treatment of this mixture with 0.25 mol. equivalent of
[Rh(COD)Cl]2 with respect to 3a', before bubbling carbon monoxide for 15 min. allowed the
complex 5a' to be isolated in a high 85 % yield after removal of 3a" by washing the DCM-
containing residue with cold pentane.
1) [Rh(COD)Cl]2
(0.25 eq.)
DCM, r.t., 2 h
2 TfO
TfO
TfO
Me
Me
Me
MeOTf
(1 eq.)
CO
Me
PPh2
N
N
N
N
N
N
+
PPh2
PPh2
Rh
Cl
P
P
Ph2
Ph2
N
N
DCM,
-78 oC
to r.t.
2 h
N
2) CO (bubbling)
15 min
N
5a': 85%
3a
3a'
3a"
3a' :3a" = 5:1
Scheme 4. N- vs P-methylation of the benzimidazolophosphine 3a, and P-coordination of the
benzimidazoliophosphine product 3a' at a RhCl(CO) center.
1
The P2RhCl(CO) complexes of type 5 have been characterized by H, 13C, 31P and
103Rh NMR and IR spectroscopy (see section 2.3). Remarkably, no MS method, among ESI,
11
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
APCI and MALDI, was found suitable to produce molecular peaks of the complexes,
preventing structural confirmation by HRMS: the lability of primary cationic species
generated in the MS chamber is actually correlated with the weak donating ability of the
monodentate ligands. The structures of 5a-c, 5f-g and 5a' were more precisely determined by
XRD analysis of single crystals grown slowly from DCM solutions (Table 1, Figure 2).
Notably, the structure of the dinitro ligand complex 5f was resolved with the chiral space
group P212121. In the crystal state, all the complexes were found to exhibit a square planar-
like geometry, with a trans arrangement of the two phosphine ligands, and an occupation
disorder of the facing Cl and CO ligands, except for 5f and 5a'. While the Rh-Cl and Rh–C
distances vary in the ranges 2.38 ± 0.03 Å and 1.81 ± 0.02 Å, respectively, the C–P distances
remain close to 2.31 ± 0.01 Å. The C-O distance, varying in the range 1.14 ± 0.03 Å (Table
1), appears not to be a significant indicator of the global electron-donating ability of the
phosphine ligand.[37a] This can be ascribed to the cis-orientation of the P–Rh–CO bonds, and
to the long recognized considerably weaker geometrical cis-influence of a P ligand on the
lengthening of a geminal Rh–L bond (with an even weaker correlated influence on the C=O
bond length for L = CO) as compared to the corresponding trans-influences.[37b] For the a
priori least donating dicationic ligand 3a', the intermediate C-O distance value of 1.14 Å is,
however, also correlated with the unique conformational features of the complex 5a'. In the
neutral complexes 5a-c, 5f-g, indeed, the ImPh2P–(Rh)–PPh2Im unit adopts a staggered
conformation, with perfectly anti-periplanar imidazolyl groups in 5a-c and 5g (Im–P–P–Im
dihedral angles = 180°), and synclinal-gauche imidazolyl groups wrapping the Rh–Cl bond in
5f (Im–P…P–Im dihedral angle = 45°, N2C…Cl ≈ 3.28 Å). In contrast, the dicationic complex
5a' exhibits a perfectly eclipsed conformation of Rh–Cl bond with both the syn-periplanar
imidazolyl groups (Im–P…P–Im dihedral angle = 2.8°); the shortness of the N2C+…Cld–
distances (3.01 ± 0.01 Å) is attributed to the electrostatic effect.
12
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
Table 1. Crystallographic data for the complexes 5a-c, 5f-g and 5a'.
5a
5b
5c
5f
5g
5a’
6f
CCDC no
1496162
1496160
1496161
1496163
1496164
1496165
1496166
Empirical
formula
C51H38Cl
N4OP2Rh
C55H44Cl9
N6O5P2Rh
C51H36Cl
C53H38Cl5
N8O9P2Rh
C53H38Cl5
N8O9P2Rh
C58.5H51Cl8F6 C57.7H47.4Cl15.4
N6O5P2Rh
N4O7P2RhS2
N8O9P2Rh2
Mr
923.19
monoclinic
P21/n
1352.92
triclinic
1013.19
triclinic
P-1
1273.05
1273.05
1548.67
triclinic
P-1
1810.59
triclinic
Cryst. system
Space group
T [K]
orthorhombic orthorhombic
P-1
P212121
100
Pbca
120
P-1
100
100
100
100
100
a [Å]
10.704(3)
11.496(4)
17.080(4)
90
9.5819(3)
10.7141(4)
14.7256(6)
98.0370(16)
8.8037(4)
10.9560(6)
12.0990(6)
87.651(2)
79.310(2)
69.853(2)
1076.24(9)
1.563
14.0668(9)
16.7870(13)
22.524(2)
90
15.9457(5)
11.2917(3)
14.7172(11)
15.0289(11)
10.5619(3)
17.8643(6)
19.6796(7)
97.7610(16)
90.9492(15)
104.4739(14)
3557.48(13)
1.690
b [Å]
c [Å]
29.2082(11) 18.4532(14)
90
90
91.977(3)
109.881(2)
117.436(2)
3316.0(4)
1.551
a [°]
99.466(9) 102.8561(16)
90
b [°]
90
2073.1(5)
1.479
2
102.9773(16)
1407.26(6)
1.596
90
90
g [°]
V [Å3]
5318.8(5)
1.590
5259.08(19)
1.608
Dc
Z
1
1
4
4
2
2
µ(MoKα) [mm-1]
0.599
31654
0.842
0.593
0.698
0.706
0.761
1.146
Refl. measured
86814
52472
79662
13763/0.058
11405
704
76783
7694/0.098
4876
106296
180292
Refl. unique/Rint 4800/0.143 12318/0.026
4759/0.038
4427
21412/0.030 24511/0.035
2792
274
10661
403
16438
811
18966
928
Refl. [I > 3s(I)]
Nb parameters
313
352
R
0.0560
0.0609
1.121
0.0389
0.0348
1.049
0.0266
0.0233
1.073
0.0309
0.0312
1.093
0.0621
0.0751
0.999
0.0560
0.0540
1.030
0.0348
0.0366
1.077
Rw
S = GooF
1.20/-1.99
0.98/-0.67
0.41/-0.72
0.57/-0.41
0.95/-1.03
2.40/-0.96
2.10/-1.27
∆rmax/∆rmin
[e.Å-3]
13
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
5a
5c
5b
2 CF3SO3
5g
5a'
5f
Figure 2. XRD molecular views of 5a-c, 5f-g and 5a' (see section 4.4). Thermal ellipsoids at
the 50 % probability level. For clarity, hydrogen atoms are omitted. The ellipsoid diagrams of
the triflate anions of 5a' are also omitted. The N-Ph and P-Ph ipso atoms reveal a π-stacking
interaction: C8-C14 ≈ 3.26 Å (5a), 3.20 Å (5b, 5c), 3.18 Å (5f), 3.14 Å (5a'); C8-C20 ≈ 2.24
Å (5g); C34-C40 ≈ 3.14 Å (5a'). In the dication 5a', the sandwiching of the Cl atom by the
benzimidazolium moieties is indicated by the short distances: Cl1-C1 ≈ Cl1-C27 = 3.01 ±
0.01 Å, Cl1-N1 ≈ Cl1-N3 = 3.20 Å, Cl1-N2 ≈ Cl1-N4 = 3.41 ± 0.00 Å Cl1-C26 ≈ Cl1-C52 =
3.55 ± 0.05 Å.
14
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
2.3. Tentative analysis of coordination properties of nitro-benzimidazolophosphines
1
Disregarding the classical broadness of the H signals of the cycloctadiene ligand in
complexes 4, all the other signals of NMR spectra of complexes 4 and 5 recorded at room
temperature are particularly sharp (see SI, section 2, pp. 32-51), ruling out any slow
equilibrium between a main trans form (found in the crystal state) and significant proportion
of the putative cis isomer. The latter is indeed expected to be thermodynamically highly
disfavored by the bulkiness of the P environment. Notably, the 31P coordination shift ∆d31P(x)
= d31P(5x) – d31P(3x) (x = a-g, a') is found larger for the cationic ligand (∆d31P(a') = 53.6 ppm)
than for the neutral ligands, for which ∆d31P is found to depend on the presence of NO2 groups
on the benzimidazole core only (Table 2): ∆d31 ≈ 45.5 ± 0.1 ppm for a non-nitrated
P
benzimidazole core (x = a, e), ∆d31 ≈ 46.6 ± 0.3 ppm otherwise (x = b, c, f, g). At last, the
P
1JRhP coupling constant does not vary significantly remaining in the range 129.6 ± 1.5 Hz for
all the neutral complexes, and just reaching 132.8 Hz for the dicationic complex 5a'.
In a more classical approach, the IR stretching frequency of the CO co-ligand (nCO)
31
and 103Rh, P NMR chemical shifts (d103Rh, d31P(5x)) of the (3x)2RhCl(CO) complexes 5x in
CD2Cl2 solutions were selected for appraising the coordinating properties of 3x, x = a-g, a'
(Table 2). The "pre-coordinating" properties of the latter was also tentatively envisaged to be
revealed by its 31P NMR chemical shift in the free state, d31P(3x).
15
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
Table 2. Selected IR, NMR spectroscopic or crystallographic data for the ligands 3a-c, 3f-g,
3a' or complexes thereof.a,b)
charge
Z
xa)
1JRhP
(Hz)
dC-O (5x)h)
nCO d103Rh (5x)f) d31 (5x) ∆d31 (x)g)
d31 (3x)
d31
P
P
P
P
(5x)c)
d)(ppm)e)
(ppm)
(ppm)
(ppm)
(3xO)
(Å)
(cm–1)
(ppm)
0
0
H
H
a
1966 - 8586 -288 20.9
1968 -8613 -316 23.7
45.4
46.7
129.6
129.6
-24.5
-23.0
17.0
17.6
1.110(14)
1.138(4)
b
0
H
c
1969 -8612 -315 23.9
46.4
131.2
-22.5
17.4
1.160(5)
0
0
NO2
NO2
e
f
1979
1978
n.d.
n.d.
22.3
25.0
45.7
46.9
128.3
129.6
-23.4
-21.9
18.1
17.9
-
1.157(3)
0
NO2
H
g
1980 - 8612 -315 25.0
46.3
53.6
129.6
132.8
-21.3
-15.5
17.9
-
1.161(16)
1.141(3)
+1
a' 2005 - 8733 -436 38.1
a) Substitution pattern (x = a-g, a') of N-phenyl-benzimidazole core as defined in Figure 1 (X,
Y, Z = H, NO2) and Scheme 4 (for a'). b) NMR spectra of solutions in CD2Cl2 (see SI); c)
from IR spectra in the solid state recorded by the ATR (attenuated total reflection) technique;
d) chemical shift with respect Rh(acac)3 with the frequency ratio ꢀ' = 3.186447 % of (see
section 4.3); e) chemical shift for the formal frequency ratio ꢀ = 3.160000 % (see section 4.3);
1
f) the JRh-P coupling constants remain similar, varying from 128.3 Hz for 5e to 132.8 Hz for
5a'; g) coordination shift: ∆d31P(x) = d31P(5x) – d31P(3x); h) from X-ray diffraction data (see
Table 2 and SI).
As previously observed in related RhX(COD)(phosphine) complexes,[ 38 ] no
correlation is observed between d103Rh(5x) and either d31P(5x) or d31P(3x). P-coordination to
non-metal hard Lewis acids such an oxygen atom is the singlet spin state (promoted oxygen,
O*) was also envisaged in the corresponding phosphine oxides 3xO, x = a-g, generated by
treatment of 3x with H2O2 (3aO and 3bO were also isolated and characterized: see SI); the
31P NMR shifts of 3xO remain however almost constant at 17.5 ± 05 ppm.
16
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
The nCO and d103 parameters thus appear to play complementary roles for appraising
Rh
the effects of nitro substituents on either the benzimidazole core or pending N-phenyl group,
respectively.
Whatever the nitro-decoration of the benzimidazole core, indeed, a N-phenylated
ligand gives nCO = 1968 ± 1 cm–1 (5a-c) and a N-p-nitrophenylated ligand gives nCO = 1979 ±
1 cm–1 (5e-g). Using the 103Rh NMR probe (see section 4.3), whatever the N-aryl substituent,
the presence of a nitro substituent on the benzimidazole core (at either position 5 or 6) gives
d
103 = –315 ppm, vs d103 = –288 ppm for the non-nitrated reference 5a (the value for 5e
Rh
Rh
could however not be determined due to solubility issues). The nCO and d103 values for the
Rh
dicationic complex 5a' are out of range, but relatively consistent with the expected effects of
"electronegative" substituents on the global donating character of the P atom towards the
RhCl(CO) center.
The sensitivity of the nCO value to p-nitro-substitution at the N-phenyl group is
remarkable (+10 cm–1 for a remote NO2 group, nine-bond away from the CO ligands), in
particular with regard to the vanishing effect of nitro substitution at the benzimidazole core in
spite of π-conjugation with the coordinating P center. Whatever the electronic process at stake
(partial π-conjugation and/or s-inductive attraction through the N–Ar bond), the observation
is in line with former observations of a marked effect of N-nitroaryl and N-fluoroaryl
substitution on intrinsic π-accepting ability or catalytic properties of imidazol-2-ylidene
ligands.[39]
2.4. Dinuclear nitro-benzimidazolophosphine side products: CH2Cl2-clathrate complexes
Each of the nitro-containing complexes 5b-c, 5e-g (Scheme 3) was actually obtained in
mixture with a side product containing a Rh–P bond, as evidenced by 31P NMR spectroscopy
of the crude material, and denoted as 6b-c, 6e-g (see experimental section 4.2). The exact
17
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
structure of theses complexes could however not be assessed by spectroscopic methods. In the
case of the dinitro ligand 3f, however, crystals of 6f deposited from a DCM solution of the
residue obtained from the supernatant of crystals of 5f after reaction of 3f with 4f (Scheme 3).
X-ray diffraction analysis showed a dinuclear structure, where the two Rh centers are in
distorted bipyramidal environments, coordinated by a µ-CO and two P,N-bridging ligands 3f
(Figure 3). The corresponding structures are assigned to the side-products 6b, 6c, 6e and 6g
obtained from 4b, 4c, 4e and 4g, respectively.
A search in the CCDC database for similar complexes featuring such a [Rh2(µ-
CO)Cl2] core gave 18 related examples, five of them featuring P,N coordination to the Rh
centers, with Rh-Rh bonding distances ranging from 2.57 Å to 3.01 Å.[40] The Rh–Rh distance
of 2.63 Å measured in 6f is thus quite short, smaller than the average value of 2.68 Å over the
series. The Rh–Cl bonds of 6f (bent from the Rh–Rh axis by ca 163°) are significantly shorter
than all the corresponding bonds within the CCDC sampling (2.36 Å vs 2.42 to 2.45 Å). The
C–O distance of 1.17 Å measured in 6f is identical to the average value over the CCDC
series. The Rh–P bonds of 6f are found quite short (2.21-2.22 Å), and much shorter than in
the parent complex 5f (2.31 Å). The Rh–N bonds are however found consistently shorter
(2.08-2.10 Å).
The crystals of 6f are actually clathrates, with 6.7 DCM molecules per molecule of 6f.
The high solvent content was actually first suspected after observing disintegration of isolated
crystals in polyisobutene blend oils used as cryo-protectants for XRD analysis (PARATONE
or PARABAR 10312; the use of a perfluoropolyether oil, PFO-XR 75, was finally required).
The solvate molecules are spread over 12 positions, eight of them being partially occupied
and forming a chain with short distances between positions strongly occupied by Cl atoms
(Cl…Cl ≈ 3.5 Å), thus suggesting that the chain cohesion is due to halogen bonding,[41a] today
widely invoked in crystal engineering.[41b-c] Chains of DCM molecules were also proposed to
18
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
occur in a related dirhodium complex, [40i] but the extended van des Waals gallery within the
crystal network of 6f (Figure 3) meets more general challenges in the search for metal-organic
inclusion compounds of the clathrate type or porous coordination polymers,[42a] in particular
for DCM guest molecules which are more currently found hosted by organic lattices.[42b-e]
Within the family of the Werner clathrates, based on a discrete packing of MX2L4 complexes
(M = divalent transition metal ion, X = small anionic ligand, L = pyridine ligand),[42f] and
related Hoffmann clathrates based on a grid of CN-bridged coordination polymer (initiated by
the cadmium(II)-CCl4 clathrate),[42g-h] one may cite the Werner DCM-clathrate [Ni(NCS)2(3-
cyanopyridine)4]•2CH2Cl2.[42i] More particular examples are the recently described dinuclear
clathrates [Cd2I2(µ-L)2]•CH2Cl2 (L = 4,4'-bis(2-methylimidazol-1-ylmethyl)biphenyl),[42j] and
[Ag(NO3)2Pt(µ- benzothiadiazol-4-ylethynyl)(4-diphenylaminopyridine)]•CH2Cl2,.[42k]
X
Y
6b: X = NO2, Y = Z = H
6c: X = Z = H, Y = NO2
6e: X = Y = H, Z = NO2
6f: X = Z = NO2, Y = H
6g: X = H, Y = Z = NO2
Cl
Z
Rh
PPh2
N
N
C
O
N
N
Ph2P
Rh
Cl
Z
Y
X
6f
Figure 3. Lewis structure of the dinuclear side-products formed along with the target
complexes 5b-c, 5e-g (Scheme 3; top), crystallographic molecular view of 6f with thermal
ellipsoids at the 50 % probability level (bottom left), and free van der Waals gallery available
for hosting DCM solvent molecules within the network of molecules of 6f (bottom right).
19
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
3. Conclusion
The disclosed strategy, upstream based on the use of LDA as the optimal base for the
deprotonation of more or less nitrated N-aryl benzimidazole precursors dedicated to react with
ClPPh2, allowed a pioneering exploration of the coordination chemistry of nitro-aromatic
phosphines as neutral electron-poor ligands with decreased s-donating ability. The approach
and results open natural prospects in organometallic chemistry such as the synthesis of tri- or
tetra-nitro counterparts of the ligands 3, the design of chelating diphosphine congeners, and
the comparison with possible analogous complexes of nitrated triphenylphosphines such as
(4-NO2-C6H4)3P standing as a paradigm. As a preliminary analysis, reduction of the P atom of
known p-nitrophenylphosphine oxides with the classical HSiCl3 reagent is however expected
to be tricky: in spite of an early report,[15e] recent results indeed indicate that competing
reduction of the nitro group by HSiCl3 could occur, even under milder conditions.[43] In an
alternative strategy, (4-nitrophenyl)diphenylphosphine, partly described 40 years ago,[15b,20]
was targeted from 4-nitro-iodobenzene and ClPPh2 but could not be isolated, and further
efforts in this direction are naturally envisaged. Nevertheless, targeting the particular category
of the benzimidazolyllphosphine complexes 4 and 5 gave the opportunity of evidencing
unique features such as: (i) empirical binary "orthogonal spectroscopic indicators" of the
presence or not of nitro groups on distinct parts of the ligand (the benzimidazole core and the
N-aryl substituent), based on an observed contrast between two ranges of values separated by
ca 5 times the spectroscopic resolutions: 10 cm–1 and 30 ppm, respectively for the nCO and dRh
probes at the RhCl(CO) center; (ii) in the dicationic complex 5a', an electrostatic sandwiching
of a chloro ligand by two imidazolium rings resulting from a packing effect, and possibly
balanced by enhanced polarization of the Rh–Cl bond; (iii) serendipitous finding of an
organic-organometallic clathrate of high solvent content (the DCM solvate complex 6f). A
more general occurrence of such features could be sought for in other systems involving other
20
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
transition metals (Ni, Pd, Pt...) and ligands or co-ligands thereof. Finally, beyond coordination
chemistry, the results are naturally aimed at exploring the use such complexes in
homogeneous catalysis.
4. Experimental section
4.1. General remarks
THF and diethyl ether were dried and distilled over sodium-benzophenone. Pentane and
dichloromethane (DCM) were dried with a PureSolv-MD-5 Innovative Technology system
for the purification of solvents. [Rh(COD)Cl]2 and other reagents were employed as received
from commercial sources, in particular, solutions of n-BuLi were 2.5 M in hexane. All
reactions were carried out under an argon atmosphere by using Schlenk and vacuum line
techniques. Column chromatography was carried out on silica gel (60 Å, C.C 70-200 µm).
Previously reported procedures were used for the preparation of the phosphines and their
precursors, in particular for the previously described compounds 1b and 1c,[44] 1d and 1e,[45]
2a,[46] and 3a.[47]
4.2. Procedures for the preparation of new compounds
1
For the characterization of the reported compounds by H, 31P ,13C NMR and IR
spectroscopy, by mass spectrometry, and by melting point, see SI
Preparation of 2b and 2c.[45] A mixture of 1b (2.69 g, 16.5 mmol), CuI (0.263 g, 1.38 mmol),
1,10-phenanthroline (0.48 g, 2.76 mmol), Cs2CO3 (9.42 g, 28.9 mmol) was evacuated twice
and back-filled with argon in a dried Schlenk tube, then iodobenzene (1.52 ml, 13.8 mmol)
and DMF (20 ml) was added with a syringe. The reaction mixture was stirred at 100oC for 48
h. After the mixture was cooled down to ambient temperature, the solution was filtered on
silica gel, extracted with EtOAc, dried over MgSO4 and concentrated under reduced pressure.
The residue was purified by column chromatography on silica gel eluted with a 3/2
21
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
EtOAc/pentane mixture to deliver 2b (2.108 g, 8.8 mmol, 64%) and 2c (0.75g, 3.1 mmol,
23%) as yellow solids.
Preparation of 2f and 2g.[48] A mixture of 1b (1.63 g, 10 mmol) and NaOH (0.40g, 10 mmol)
in 20 ml DMF was treated with 4-nitrofluorobenzene (2.12 ml, 20 mmol) at r. t. under argon.
The mixture was stirred at 70oC for 5 h before addition of water. The resulting suspension
was extracted with EtOAc (3× 40 mL), and the organic layer were combined, dried over
MgSO4 and concentrated to dryness under reduced pressure. Column chromatography of the
residue on silica gel eluted with a 3/2 EtOAc/pentane mixture afforded the corresponding
products 2f (1.51 g, 5.3 mmol, 53 %) and 2g (0.97 g, 3.4 mmol, 34 %) as yellow solids.
The same procedure from 1e afforded 2e as a yellow solid in 95 % yield. Starting from 1h,
and changing the chromatographic eluting system to a 1/1 EtOAc/pentane mixture, 2h was
obtained as a yellow solid in 86 % yield.
General procedure for preparation of the nitro-aromatic phosphines 3b-c, 3e-h
A solution of 2a-c, 2e-h (1 equiv.) in THF was treated with a solution of freshly prepared
LDA (1.2 equiv.) at -78oC. The mixture was slowly warmed up to -40oC, then stirred for 1 hat
this temperature. After cooling back to -78oC, chlorodiphenylphosphine was added dropwise
and the solution was slowly warmed up, then stirred at r.t. for 4 h. After addition of a
saturated aqueous NH4Cl solution and dilution with DCM, the aqueous layer was separated
and the organic layer was dried over MgSO4 and concentrated to dryness under reduced
pressure. The residue was purified by column chromatography on silica gel eluted with a 3/1
DCM/pentane mixture to give the corresponding products as yellow solids (for more details,
see SI): 3a (78 %), 3b (83 %), 3c (70 %), 3e (62 %), 3f (77 %), 3g (45 %), 3h (< 5 %).
22
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
General procedure for generation of the phosphine oxides 3xO (x = a-c, e-g)
The phosphine oxides were generated according to a previously reported procedure,[49] by
treatment of a solution 3x in THF with H2O2 (3 equiv.) at room temperature under TLC
monitoring. After addition of NaHCO3 and Na2S2O3 solutions, the reaction mixture was
separated, and the aqueous layer was extracted with DCM. The combined organic layer was
dried over Na2SO4 and concentrated under reduced pressure affording the corresponding
phosphine oxide in quantitative 31P NMR spectroscopic yield without further purification. The
products were analyzed by 31P NMR spectroscopy only (see Table 2), except for 3aO and
3bO which were more completely characterized (see SI).
General procedures for preparation of rhodium complexes
Method A. A solution of 3x (x = b, f, e) and [Rh(cod)Cl]2 (0.5 equiv.) in DCM was stirred at
r. t. for 4 h. After concentration under reduced pressure, the residue was washed with pentane
and dried to afford the corresponding rhodium complex as brown solids: 4b (95 %), 4e (93
%), 4f (94 %).
A solution of 4x and the corresponding phosphine 3x (1 equiv.) in DCM was treated with
bubbling carbon monoxide for 15 min. After concentration under reduced pressure, the
residue was washed with diethyl ether (or a DCM/pentane mixture) to give the corresponding
rhodium complex 5b (60 %), 5e (74 %), 5f (76 %). Minor amounts of the dimeric rhodium
complex 6b (18 %), 6e (14 %) or 6f (15 %) were also detected by 31P NMR spectroscopy of
the crude material (approximate spectroscopic yields given). Crystals of 6f deposited from a
DCM solution could be characterized by XRD analysis (see section 4.4 and SI).
Method B. A solution of 3x (x = a, c, g) and [Rh(cod)Cl]2 (0.25 equiv.) in DCM was stirred
at r. t. for 4 h. Carbon monoxide was then bubbled through the solution for 15 min. After
evaporation of the solvent under reduced pressure, the crude mixture was washed with diethyl
ether (or a DCM/pentane mixture) to give the corresponding rhodium complex 5a (79 %), 5c
23
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
(81 %) or 5g (73 %) as a yellow solid. Minor amounts of the dimeric rhodium complex 6c (4
%) or 6g (13 %) were also detected by 31P NMR spectroscopy of the crude material
(approximate spectroscopic yields given).
Rhodium complex 5a’. To a solution of 3a (152 mg, 0.4 mmol) in DCM (4 mL) at –90oC
was added methyl trifluoromethanesulfonate (45 µL, 0.4 mmol). The mixture was slowly
warmed up to -40oC and stirred for 2 h. After evaporation of the solvent under reduced
pressure, the residue was washed with cold pentane, dried under reduced pressure and then
analyzed as a 5:1 N+Me:P+Me mixture of the methylation products of 3a, which could not be
separated. To a solution of this mixture in DCM (10 mL) was added [Rh(cod)Cl]2 (41 mg,
0.083 mmol), and stirring was pursued at r. t. for 4 h. Carbon monoxide was then bubbled
through the solution for 15 min. The solvent was evaporated under reduced pressure, and the
residue was washed with diethyl ether to afford 5a’ (130 mg, 68 %) as a yellow solid.
4.3. 103Rh NMR spectroscopy
103Rh/31P{1H} HMQC spectra were recorded on a Bruker AV400 spectrometer, equipped with
a three-canal probe H{P}{X}- TBI 5 mm using saturated solutions of 5x in CD2Cl2 (in the
cases of 5e and 5f, saturated concentration did not allow achieving sufficient acquisition in
available time). A preliminary search for the 103Rh resonances was performed through
31P{1H} experiments with 103Rh inversion using a sequence adapted from the literature giving
a narrow range for 103Rh chemical shifts.[50] In this range, precise chemical shifts of 103Rh
were measured via an inverse correlation HMQC using both 103Rh-31P coupling constants.
1
The exact spectrometer H frequency of n0 = 400.1318 MHz corresponds to the resonance of.
tetramethylsilane. The 103Rh frequency ratio ꢀ can be assigned to either the formal value ꢀ =
3.16 % (for which rhodium metal is just an approximate reference giving ꢁ103 ≈ 0 ppm),[51]
Rh
or to the observable value ꢀ' = 3.186447 % defined from the real reference Rh(acac)3 (for
24
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
which ꢁ103 = 0 ppm if saturated in CDCl3 solution).[52] Each frequency ratio defines its own
Rh
chemical shift scale in ppm: d = (n – ꢀ n0)/(ꢀ n0) 106 and d' = (n – ꢀ' n0)/(ꢀ' n0) 106, where n
is the physical resonance frequency observed for the nominal field of the spectrometer. By
elimination of n, the scales are related by the equation: d = d' x + (x–1) 106, where x = ꢀ'/ꢀ =
[51a]
1.008369 (giving d(Rh(acac)3) = 8369 ≈ 8358 ppm for d'(Rh(acac)3) = 0).
The ꢁ103
Rh
values in Table 2 are given in both scales.
4.4. Crystallographic studies
Intensity X-ray diffraction data from single crystals were collected at low temperature on an
Apex2 Bruker diffractometer equipped with a 30W air-cooled microfocus source (λ = 0.71073
Å) or on an Oxford-Diffraction Gemini (λ = 0.71073 or 1.54180 Å). The structures were
solved using direct methods or SUPERFLIP, and refined by means of least-squares
procedures on F using the programs of the PC version of CRYSTALS. Atomic scattering
factors were taken from the International Tables for X-Ray Crystallography.[53] All non-
hydrogen atoms were refined anisotropically. Hydrogen atoms were refined using a riding
model. Absorption corrections were introduced using the program MULTISCAN. For the free
phosphines 3b and 3f, it was not possible to resolve diffuse electron-density residuals
(enclosed solvent molecules). Treatment with the SQUEEZE facility from PLATON resulted
in a smooth refinement.[54]
Supporting information. Supporting information for this article is available on the WWW
under http://dx.doi.org/… It includes yields, spectroscopic and other characterizations of all
1
31
13
new compounds. H, P, C,103Rh-31P{1H} HMQC NMR spectra and crystallographic data
are also provided.
25
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
Acknowledgements. C. Z. thanks the China Scholarship Council for financial support. R. C.
thanks the CNRS for half a teaching sabbatical in 2014-2015 and 2015-2016. A determining
support was also indirectly provided by the ANR program (ANR-11-BS07-016-01). The
authors are also indebted to Dr. Valérie Maraval for efficient logistic management.
References and notes
1 C. Lepetit, V. Maraval, Y. Canac, R. Chauvin, Coord. Chem. Rev. 2016, 308 II, 59-75.
2
a) Nelson, D. J.; Nolan, S. P. Chem. Soc. Rev. 2013, 42, 6723-6753. b) Melaimi, M.;
Soleilhavoup, M.; G. Bertrand, G. Angew. Chem. Int. Ed. 2010, 49, 8810-8849. c) Top.
Organomet. Chem. ‘Late transition metal complexes of neutral h1-carbon ligands: R.
Chauvin, Y. Canac, Eds, Springer, 2010, 30, 1-252. d) J. Vignolle, X. Cattoën, D. Bourissou,
Chem. Rev. 2009, 109, 3333-3384. e) S. Díez-González, N. Marion, S. P. Nolan, Chem. Rev.
2009, 109, 3612-3676. f) F. E. Hahn, M. C. Jahnke, Angew. Chem. Int. Ed. 2008, 47, 3122-
3172. g) Y. Canac, M. Soleihavoup, S. Conejero, G. Bertrand, J. Organomet. Chem. 2004,
689, 3857-3865.
3 C. Maaliki, C. Lepetit, Y. Canac, C. Bijani, C. Duhayon, R. Chauvin, Chem. Eur. J. 2012,
18, 7705-7714.
4 See for example: a) R. L. Pruett, J. A. Smith, J. Org. Chem. 1969, 34, 327-330; b) K.
Nozaki, T. Nanno, H. Takaya, J. Organomet. Chem. 1997, 527, 103-108; c) R. A. Baber, M.
L. Clarke, K. M. Heslop, A. C. Marr, A. G. Orpen, P. G. Pringle, A. Ward, D. E. Zambrano-
Williams, Dalton Trans. 2005, 1079-1085; d) F. Doro, J. N. H. Reek, P. W. N. M. van
Leeuwen, Organometallics 2010, 29, 4440-4447; e) Q. Liu, L. Wu, I. Fleischer, D. Selent, R.
Franke, R. Jackstell, M. Beller, Chem. Eur. J. 2014, 20, 6888-6894.
26
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
5 See for example: a) M. F. Ernst, D. M. Roddick, Inorg. Chem. 1989, 28, 1624-1627; b) M.
Brookhart, W. A. Chandler, A. C. Pfister, C. C. Santini, P. S. White, Organometallics 1992,
11, 1263-1274; c) D. C. Smith Jr., E. D. Stevens, S. P. Nolan, Inorg. Chem. 1999, 38, 5277-
5281; d) D. R. Palo, C. Erkey, Organometallics 2000, 19, 81-86; e) F. Berhal, O. Esseiva, C.
H. Martin, H. Tone, J.-P. Genet, T. Ayad, V. Ratovelomanana-Vidal, Org. Lett. 2011, 13,
2806-2809; f) B. G. Anderson, J. L. Spencer, Chem. Eur. J. 2014, 20, 6421-6432.
6 a) I. V. Komarov, M. Y. Kornilov, A. A. Tolmachev, A. A. Yurchenko, E. B. Rusanov and
A. N. Chernega, Tetrahedron 1995, 51, 11271-11280; b) M. Enders, O. Fritz, H. Pritzkow, Z.
Anorg. Allg. Chem. 2004, 630, 1501-1506; c) P. Antognazza, T. Benincori, S. Mazzoli, F.
Sannicolo, T. Pilati, Phosphorus Sulfur Silicon 1999, 144-146, 405-408. d) T. Benincori, E.
Brenna, F. Sannicolo, L. Trimarco, P. Antognazza, E. Cesarotti, F. Demartin, T. Pilati, G.
Zotti, J. Organomet. Chem. 1997, 529, 445-453; e) N. Debono, Y. Canac, C. Duhayon, R.
Chauvin, Eur. J. Inorg. Chem. 2008, 2991-2999; f) I. Abdellah, N. Debono, Y. Canac, L.
Vendier, R. Chauvin, Chem. Asian. J. 2010, 5, 1225-1231.
7 For pioneering works, see: a) U. Zoller, Tetrahedron 1988, 44, 7413-7426; b) N. Kuhn, J.
Fahl, D. Blaeser, R. Boese, Z. Anorg. Allg. Chem. 1999, 625, 729-734; c) N. Kuhn, M.
Goehner, G. Henkel, Z. Anorg. Allg. Chem. 1999, 625, 1415-1416; d) J. Sirieix, M. Obberger,
B. Betzemeier, P. Knochel, Synlett 2000, 11, 1613-1615; e) M. Azouri, J. Andrieu, M.
Picquet, P. Richard, B. Hanquet, I. Tkatchenko, Eur. J. Inorg. Chem. 2007, 4877-4883.
8 For reviews, see: (a) Y. Canac, C. Maaliki, I. Abdellah, R. Chauvin, New. J. Chem. 2012,
36, 17-27; b) S. Gaillard, J. L. Renaud, Dalton Trans. 2013, 7255-7270; c) M. Alcarazo,
Chem. Eur. J. 2014, 20, 7868-7877; d) M. Alcarazo, Acc. Chem. Res. 2016, 49, 1797-1805.
9 I. Abdellah, C. Lepetit, Y. Canac, C. Duhayon, R. Chauvin, Chem. Eur. J. 2010, 16, 13095-
13108.
27
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
10 a) C. Mboyi, C. Maaliki, A. Mankou Makaya, Y. Canac, C. Duhayon, R. Chauvin, Inorg.
Chem. 2016, 55, 11018-11027; b) G. Mehler, P. Linowski, J. Carreras, A. Zanardi, J. W.
Dube, M. Alcarazo, Chem. Eur. J. 2016, 22, 15320-15327.
11 J. Petuskova, H. Bruns and M. Alcarazo, Angew. Chem., Int. Ed. 2011, 50, 3799-3802.
12 a) B. Vabre, Y. Canac, C. Duhayon, R. Chauvin, D. Zargarian, Chem. Commun. 2012, 48,
10446-10448; b) I. Abdellah, Y. Canac, C. D. Mboyi, C. Duhayon, R. Chauvin, J.
Organomet. Chem. 2015, 776, 149-162; c) B. Vabre, Y. Canac, C. Lepetit, C. Duhayon, R.
Chauvin, D. Zargarian, Chem. Eur. J. 2015, 21, 17403-17414.
13 C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165-195.
14 Generalization of the Hammett equation, originally defined for X/H-substitution (–H ® –
X), to aza-analogation (=CH– ® =N–) and protoaza-analogation (=CH– ® =NH+–) gives the
following values for the corresponding constants: sm(N) = 0.45, sp(N) = 0.76, sm(NH+) = 2.09
and sp(NH+) = 2.34. Comparison with the NO2/H-substitution values (sm = 0.71, sp = 0.78)
indicates that neutral azalogation has a significant, albeit weaker, electron-withdrawing effect,
but suggests that N-protonation (representative of N-alkylation) has a twice stronger effect
(sm(NH+) – sm(N) = 1.64, sp(NH+) – sp(N) = 1.58). See: C. D. Jonhson, The Hammet Equation, pp.
99-101, Cambridge University Press, London, 1973.
15 a) A. Stachlewska-Wroblowa, K. Okon, Bull. Acad. Pol. Sci, Ser. Sci. Chim. 1961, 9, 297-
301; b) G. P. Schiemenz, Chem.Ber. 1966, 99, 514-519; c) G. P. Schiemenz, K. Roehlk,
Chem.Ber. 1971, 104, 1219-1233; d) J. I. G. Cadogan, D. J. Sears, D. M. Smith, J. Chem. Soc.
C 1969, 1314-1318; e) G. P. Schiemenz, P. Nielsen, Phosphorus Sulfur 1985, 21, 259-266.
16 DFT calculations at the B3PW91/6-31G(d,p) level, without solvent correction (vacuum),
using the Firefly software (see SI): a) A. A. Granovsky, Firefly version 8,
28
This article is protected by copyright. All rights reserved.

10.1002/asia.201701078
Chemistry - An Asian Journal
http://classic.chem.msu.su:gran:firefly:index.html) which is partially based on the GAMESS
(US) source code; b) M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S.
Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen, S. Su, T. L. Windus, M.
Dupuis, J. A. Montgomery, J. Comput. Chem. 1993, 14, 1347-1363.
17 G. A. Olah, B. G. B. Gupta, S. C. Narang, J. Am. Chem. Soc. 1979, 5317-5322.
18 a) M. Mąkosza, M. Paszewski, D. Sulikowski, Synlett 2008, 2938-2940; b) J. Xu, P. B.
Zhang, Y. Z. Gao, Y. Y. Chen, G. Tang, Y. F. Zhao, J. Org. Chem. 2013, 78, 8176-8183. c)
M. Hayashi, T. Matsuura, I. Tanaka, H. Ohta, Y. Watanabe, Org. Lett. 2013, 15, 628-631; d)
H. Zhang, X. Y. Zhang, D. Q. Dong, Z. L. Wang, RSC. Adv. 2015, 5, 52824-52831; e) D. V.
Aleksanyan, Z. S. Klemenkova, A. A. Vasil’ev, A. Y. Gorenberg, Y. V. Nelyubina, V. A.
Kozlov, Dalton Trans. 2015, 44, 3216-3226.
19 Model nitroaryl-phosphine ligands in Rh(II), Ru(II) or Fe(II) complexes have however been
calculated at the DFT level of theory: (a) P. Hirva, P. Lahuerta, J. Perez-Prieto, Theor. Chem.
Acc. 2005, 113, 63-68; b) P. Hirva, E. Julio, J. Lloret, P. Lahuerta, J. Perez-Prieto, Inorg.
Chem. 2007, 46, 2619-2626; d) E. Pump, L. Cavallo, C. Slugovc, Monatsch. Chem. 2015,
146, 1131-1141; d) B. Mondal, F. Neese, S. Ye, Inorg. Chem. 2016, 77, 5438-5444.
20
G. P. Schiemenz, M. Finzenhagen, Liebigs Ann. Chem. 1976, 2126-2135;
21 a) N. F. Blank, J. R. Moncarz, T. J. Brunker, C. Scriban, B. J. Anderson, O. Amir, D. S.
Glueck, L. N. Zakharov, J. A. Golen, C. D. Incarvito, A. L. Rheingold, J. Am. Chem. Soc.
2007, 129, 6847-6858. b) N. Nowrouzi, S. Keshtgar, E. B. Jahromi, Tetrahedron Lett. 2016,
57, 348-350.
22 a) F. Y. Kwong, C. W. Lai, Y. Tian, K. S. Chan, Tetrahedron Lett. 2000, 41, 10285-10289;
b) S. E. Martin, M. Bonaterra, R. A. Rossi, J. Organomet. Chem. 2002, 664, 223-227.
29
This article is protected by copyright. All rights reserved.