John Popp1, Schirin Hanf1,2, Evamarie Hey-Hawkins1. 1. Faculty of Chemistry and Mineralogy, Institute of Inorganic Chemistry, Leipzig University, Johannisallee 29, 04103 Leipzig, Germany. 2. Department of Chemistry, Cambridge University, Lensfield Road, CB2 1EW Cambridge, U.K.
Abstract
Half-sandwich arene-metal complexes are commonly used for specific applications. Herein, we report facile arene ligand exchange reactions of two ruthenium(II) complexes of tertiary P-stereogenic ferrocenyl phosphines. By mild photochemical activation, the displacement of p-cymene and subsequent tethering by η6-coordination of the terminal phenyl ring of a biphenylyl-substituted ferrocenyl phosphine were enabled. Furthermore, the spontaneous p-cymene displacement in a 2-methoxyphenyl-containing ferrocenyl phosphine and ensuing coordination of the ligand as a P,O chelate were examined. For both reactions, theoretical calculations of the general course of the reaction confirmed the experimental findings. The ease of the controlled arene displacement reported here can offer new pathways for the synthesis and design of novel tailor-made catalysts.
Half-sandwich arene-metal complexes are commonly used for specific applications. Herein, we report facile arene ligand exchange reactions of two ruthenium(II) complexes of tertiary P-stereogenic ferrocenyl phosphines. By mild photochemical activation, the displacement of p-cymene and subsequent tethering by η6-coordination of the terminal phenyl ring of a biphenylyl-substituted ferrocenyl phosphine were enabled. Furthermore, the spontaneous p-cymene displacement in a 2-methoxyphenyl-containing ferrocenyl phosphine and ensuing coordination of the ligand as a P,O chelate were examined. For both reactions, theoretical calculations of the general course of the reaction confirmed the experimental findings. The ease of the controlled arene displacement reported here can offer new pathways for the synthesis and design of novel tailor-made catalysts.
Half-sandwich complexes of group 8 metals
with a characteristic piano-stool architecture are privileged structures
in organometallic and catalysis research,[1−5] and have found wide application in bioinorganic and
medicinal chemistry.[3,6−9] In this context, 1-methyl-4-(propan-2-yl)benzene
(para-cymene, p-cym) plays a predominant
role because it represents an easily accessible formally neutral η6-coordinating arene ligand and its ruthenium(II) complex [{Ru(p-cym)Cl2}2] is commercially available.[10,11] The structural behavior of these complexes is crucial for their
desired application. For example, transfer hydrogenation reactions
proceeding via an outer sphere hydride transfer mechanism require
the retention of the arene ligand on the metal center.[12] Furthermore, the interaction of half-sandwich
arene–metal complexes with biomolecules usually takes place
with the arene ligand remaining at the metal[6,7] (with
only a few exceptions[13]). On the other
hand, examples of hydrogen-borrowing catalysis or C–H arylation
reactions were reported, where loss of the arene ligand is necessary
for the desired reaction to proceed.[14−16] In addition, facile
and controlled arene displacement became a key tool for the synthesis
of new catalysts which is, for example, highlighted by the variety
of transition-metal complexes of tethered arenes.[17]Tethered complexes are an interesting subset of metal–arene
complexes which emerge when the π-coordinating neutral arene
ligand is connected with one of the coordinating donor groups. Here,
phosphines with pendant aromatic groups in close proximity to the
phosphorus atom are predominantly used as polydentate ligands for
the formation of tethered ruthenium complexes.[18−26] Utilizing the chelate effect to stabilize the arene–metal
interaction is one major benefit of tethered complexes compared to
nontethered complexes.[21,27] In the case of chiral ligands,
another advantage is a restrained stereochemical situation which raises
the racemization barrier of the chiral tethered complex.[22,28−30] The arene coordination to form tethered complexes
can be triggered thermally, but usually requires high temperatures.[27] In contrast, tethering triggered by light is
a milder alternative in cases where thermal racemization needs to
be avoided, for example, with P-stereogenic phosphines.[22]For some time, we have been interested
in developing new classes of ferrocenyl phosphines, where our focus
moved from aryl-based ferrocenyl phosphine ligands for rhodium(I)-catalyzed
hydroformylation of olefins[31,32] to immobilized ferrocenylphosphine ligands for ruthenium(II)-catalyzed isomerization of allylic
alcohols.[33] Following our report on the
first example of redox-switchable catalysis involving dendriticphosphine-containing
ligands,[34] our next target was the use
of the corresponding P-stereogenic ferrocenyl phosphine
ligands for asymmetric catalysis.
Results and Discussion
The P-stereogenic
ferrocenyl phosphine boranes 1a,b were successfully synthesized
from an unsymmetrically disubstituted 1,1′-ferrocene[34] and methyl(phenyl)phosphinite boranes (Scheme ) originating from
an (−)-ephedrine-based oxazaphospholidine borane complex.[35,36] This auxiliary-based methodology provides a convenient strategy
for a stepwise preparation of tertiary P-stereogenic
phosphines (98.7 and 96.7% ee for 1a and 1b, respectively). In contrast to similarly prepared monodentate[37] or bidentate[38,39] ligands, 1a,b contain a 4-methoxyphenyl substituent in the 1′-position
of the ferrocenyl group. Replacement of 4-methoxyphenyl in these model
complexes with 4-hydroxyphenyl allows immobilization of these ligands
on a variety of solid supports. Borane-deprotection of 1a,b was achieved by heating in diethylamine, and the free ferrocenylphosphines 2a,b were reacted with [{Ru(p-cym)Cl2}2].
Scheme 1
Synthesis of P-Stereogenic Ferrocenyl Phosphine Boranes 1a,b
In the reaction of biphenylyl-containing ferrocenylphosphine 2a, initially, the monodentate phosphine-ruthenium(II)
complex 3a was formed after 12 h at room temperature,
as indicated by a rather broad signal at 18.7 ppm in the 31P{1H} NMR spectrum (Figure , bottom). However, the reaction was not completed
as revealed by a sharp peak at −22.9 ppm for the free ligand 2a. In addition, a small signal was observed at 41.5 ppm giving
a first indication of the formation of a chelate complex.
Figure 1
31P{1H} NMR spectra of the reaction of 2a with
[{Ru(p-cym)Cl2}2] yielding
monodentate complex 3a and, after light irradiation,
chelate complex 4a.
31P{1H} NMR spectra of the reaction of 2a with
[{Ru(p-cym)Cl2}2] yielding
monodentate complex 3a and, after light irradiation,
chelate complex 4a.As in rare cases, when the 2-biphenylyl substituent
of a phosphine in ruthenium complexes is able to undergo a tethering
process by replacing an η6-coordinating arene ligand,[22] we anticipated a similar process here. Because P-stereogenic phosphines are generally known to be configurationally
stable only under mild conditions and similar ferrocenyl-containing
phosphines were reported to be thermally unstable,[28] we decided to trigger the tethering of 3a by
mild photochemical activation, as described previously.[22,27−29] Indeed, the substitution reaction resulting in formation
of the tethered complex 4a was already completed after
8 h just by irradiation with a conventional desk lamp (singlet at
41.5 ppm in the 31P{1H} NMR spectrum; Figure , top). This reaction
is particularly remarkable because the displacement of the p-cymene fragment as the first step of tethering is generally
known to proceed rather sluggishly or is even reported to be impossible.[22] Usually, arene displacements of this kind are
achieved with labile η6-coordinated ligands, for
example, electron-poor arenes such as alkyl benzoates.[40−42] The solid-state structure of 4a confirmed the η6-coordination of the terminal phenyl ring of the 2-biphenylyl
substituent (Figure ).
Figure 2
Molecular structure of 4a in the solid state. Ellipsoids
at 50% probability; hydrogen atoms are omitted for clarity.
Molecular structure of 4a in the solid state. Ellipsoids
at 50% probability; hydrogen atoms are omitted for clarity.In order to substantiate our experimental observations
and understand why the p-cymene ligand can be easily
displaced, DFT calculations on the complexation reaction with 2a were performed (Figure , see the Supporting Information for further details). The course of reaction for the tethering process
giving 4a was calculated with a focus on key intermediates
only. The first step, the coordination of ferrocenyl phosphine 2a to a [Ru(p-cym)Cl2] fragment
yielding the monodentate complex 3a, is exergonic (−41.4
kJ mol–1). More interestingly, the subsequent displacement
of the p-cymene ligand forming chelate complex 4a further stabilizes the system energetically by 10.3 kJ
mol–1. This second step likely proceeds by stepwise
hapticity decrease of the p-cymene ligand as reported
previously.[16]
Figure 3
Differences in the energy
levels of the reaction of 2a to 3a and 4a calculated by DFT using the ORCA program package: BP86/def2-TZVP
(see the Supporting Information for further
details).
Differences in the energy
levels of the reaction of 2a to 3a and 4a calculated by DFT using the ORCA program package: BP86/def2-TZVP
(see the Supporting Information for further
details).In the reaction of 2-methoxyphenyl-containing ferrocenylphosphine 2b with [{Ru(p-cym)Cl2}2], initially, the expected monodentate complex 3b was formed as indicated by a signal at 17.1 ppm in the 31P{1H} NMR spectrum (Figure , bottom). Additionally, single crystals
of 3b were obtained, and X-ray crystallography confirmed
the proposed structure (Figure ). However, when 3b was dissolved in CDCl3, a new phosphorus-containing compound was formed after several
days at room temperature (Figure , top). A sharp singlet with
a downfield shift to 59.8 ppm in the 31P{1H}
NMR spectrum indicates the formation of a new phosphorus-containing
species.
Figure 5
31P{1H} NMR spectra of the reaction
of 2b with [{Ru(p-cym)Cl2}2], yielding the 1:1 complex 3b and, after
several days at rt, formation of the 2:1 complex 4b by
arene displacement.
Figure 4
Molecular structure of 3b in the solid state. Ellipsoids
at 50% probability, hydrogen atoms are omitted for clarity.
Molecular structure of 3b in the solid state. Ellipsoids
at 50% probability, hydrogen atoms are omitted for clarity.31P{1H} NMR spectra of the reaction
of 2b with [{Ru(p-cym)Cl2}2], yielding the 1:1 complex 3b and, after
several days at rt, formation of the 2:1 complex 4b by
arene displacement.Chromatographic separation of the newly formed
compound followed by successful single-crystal X-ray structure analysis
revealed the formation of the 2:1 complex 4b (Figure ). Accordingly, complex 4b can be obtained in high yield in a 2:1 (ligand/metal) reaction.
In 4b, a RuCl2 fragment is coordinated by
two ligand molecules 2b in a chelating bidentate fashion
via phosphorus and oxygen with complete displacement of the p-cymene ligand. This corresponds well to the 1H NMR spectrum of 4b, showing a downfield shift for
the associated methoxy groups from 3.57 ppm in 3b to
4.80 ppm upon coordination in 4b. The structure of the
complex can be described as a distorted octahedral geometry with two
trans-coordinating chlorido ligands and two coplanar P,O chelates with cis-coordinating oxygen and phosphorus
donors. Out of five possible structural isomers, only the cis–cis–trans
isomer shown in Figure is found because of steric crowding combined with the trans influence
of the involved donor atoms, in line with previously reported examples.[43,44] While the coordination mode of ruthenium observed here in 4b is known, it is usually achieved by reacting a 2-methoxy-substituted
phosphine with ruthenium(III) chloride.[45−47] Only a few examples
are known where an arene displacement leads to this coordination motif.[44,48]
Figure 6
Molecular
structure of 4b in the solid state. Ellipsoids at 50%
probability, hydrogen atoms are omitted for clarity.
Molecular
structure of 4b in the solid state. Ellipsoids at 50%
probability, hydrogen atoms are omitted for clarity.The unexpected ease of arene displacement is remarkable
because it opens new pathways for the synthesis and design of novel
metal complexes. These phosphine-ether ligands and their ruthenium(II)
complexes have received much attention in homogeneous catalysis, mainly
due to their hemilability.[44,49,50] The weakly coordinating oxygen atom readily dissociates generating
a vacant site for substrate binding. Accordingly, the Ru1–O2
and Ru1–O3 bonds in 4b (2.271 and 2.248 Å,
respectively) are much longer than the sum of the covalent radii.[45,51]DFT calculations on the arene displacement of 3b were performed (Figure , see the Supporting Information for further details). Again, the general course of the reaction
with key structures only was calculated. As expected, the reaction
of ferrocenyl phosphine 2b with [{Ru(p-cym)Cl2}2] yielding the monodentate 1:1 complex 3b is exergonic by 93.1 kJ mol–1. The subsequent
formation of the 2:1 complex 4b with loss of the p-cymene ligand is again exergonic by 44.4 kJ mol–1, which qualitatively resembles the experimental observations.
Figure 7
Differences
in the energy levels of the reaction of 2b to 3b and 4b calculated by DFT using the ORCA program package:
BP86/def2-TZVP (see the Supporting Information for further details).
Differences
in the energy levels of the reaction of 2b to 3b and 4b calculated by DFT using the ORCA program package:
BP86/def2-TZVP (see the Supporting Information for further details).
Conclusions
In summary, we have reported two examples
of ruthenium(II) complexes with P-stereogenic ferrocenylphosphines which impressively demonstrate facile arene replacement,
in the case of 4a, or displacement, in the case of 4b. Both compounds were isolated as follow-up products of
the complexation reactions of the P-stereogenic ferrocenylphosphine ligands with [{Ru(p-cym)Cl2}2], which initially gave the expected monodentate complexes 3a,b. For formation of 4a, we applied mild photochemical
activation of 3a to induce tethering, which is unusual
for p-cymene displacement. The ease with which p-cymene loss can be deliberately induced reveals that stable
arene coordination to the metal center may not always be guaranteed
and facile replacement opens new pathways for the synthesis of novel
metal complexes.
Experimental Section
General Procedures and Analytical Methods
In the case
of moisture or air sensitivity, the reactions were conducted under
a nitrogen atmosphere using Schlenk techniques. Diethyl ether, dichloromethane,
toluene, and hexane (isomeric mixture) were obtained from an MBraun
Solvent Purification System SPS-800 and stored over 4 Å molecular
sieves or potassium (in the case of hexane). Tetrahydrofuran (THF)
was dried and distilled from potassium, and diethylamine was dried
and distilled from potassium hydroxide. The starting material 1-bromo-1′-(4-methoxyphenyl)ferrocene[34] was synthesized from 1,1′-dibromoferrocene[52] according to literature procedures. The methyl(phenyl)phosphinite
boranes[35] were synthesized by way of an
(−)-ephedrine-based oxazaphospholidine borane complex. NMR
spectra were recorded with a Bruker AVANCE III HD 400 or Bruker Ascend
400 spectrometer. Numbering schemes for the assignment of NMR signals
are included in the Supporting Information. Mass spectra were obtained with a Bruker ESI-TOF micrOTOF, a Bruker
ESI-qTOF Impact II, and a Bruker Esquire 3000plus spectrometer. Elemental
analyses were determined with a Heraeus Vario EL Analyser. IR spectra
were obtained with a PerkinElmer FT-IR Spectrum 2000 spectrometer.
The samples were measured as KBr pellets. Chiral HPLC was performed
with a Knauer HPLC system with a Smartline PDA 2800 detector (λ
= 233 nm) and a 250 × 4.6 mm Lux 5 μm Amylose-1 column
by Phenomenex. The specific rotations were measured with a Krüss
Optronic P3002RS automatic digital polarimeter using a 1 dm micro
polarimeter tube from Schmidt + Haensch. Melting points were determined
using a Gallenkamp MPD 350 BM 2.5 capillary melting point apparatus
and are reported uncorrected.
Synthesis of Ferrocenyl Phosphine Boranes 1a,b
At −80 °C, BuLi in hexane
(1.10 equiv) was slowly added to a solution of 1-bromo-1′-(4-methoxyphenyl)ferrocene
(1.05 equiv) in THF (0.25 mol L−1). After stirring
for 1 h at −80 °C, the reaction mixture was slowly added
to a solution of the corresponding methyl(phenyl)phosphinite borane
(1.00 equiv) in THF (0.25 mol L−1) at −80
°C. The solution was allowed to warm to room temperature over
a period of 12 h. Water was added, and the aqueous phase was extracted
with diethyl ether. The combined organic layers were washed with saturated
aqueous NaCl and dried over MgSO4. After removal of the
solvent in vacuum, the crude product was purified by column chromatography
on silica with hexane/DCM 4:1 grad. hexane/DCM 1:1.
A solution of 1 in diethylamine (0.025 mol L−1) was heated at 50 °C for 12 h. Diethylamine was removed by
evaporation under reduced pressure, and the crude residue was purified
by column chromatography on degassed, deactivated silica (pretreated
with hexane/triethylamine 95:5) with degassed hexane/DCM 2:1 grad.
hexane/DCM 1:2. In order to determine the enantiomeric excess, 2 was reprotected by reaction with BH3·SMe2 (2.0 mol L−1 in THF) prior analysis by
chiral HPLC.
Synthesis of Ferrocenyl Phosphine Ruthenium(II) Complexes 3b and 4a,b
A solution of di-μ-chlorobis-[(η6-p-cymene)chlororuthenium(II)] (0.50 equiv)
in DCM (0.025 mol L−1) was added to a solution of 2 (1.00 equiv) in DCM (0.025 mol L−1) and
stirred at room temperature until the 31P{1H}
NMR spectrum indicated full conversion of the free phosphine (4a was exposed to the light of a conventional desk lamp while
stirring). The volume of the reaction mixture was reduced under reduced
pressure, and the resulting residue was purified by column chromatography
on degassed silica with degassed pure DCM grad. DCM/ethyl acetate
1:1. The determination of the enantiomeric excess of the metal complexes
was not successful. The results of the screening on a number of phases
in different screening modes were inconclusive; no selectivity between
the isomers was observed (chiral HPLC screening service by Phenomenex).
Complex 4b was obtained as
a byproduct of the reaction to 3b and was separated from 3b by column chromatography. Depending on the reaction time
for 3b, different yields of 4b could be
isolated. In solution, complex 3b slowly forms 4b at room temperature. R: 0.42 (DCM/ethyl acetate 8:1, v/v). mp 210–212 °C. 1H NMR (400.2 MHz, CD2Cl2): δ (ppm)
3.48 (br s, 2H, H11/H14), 3.52–3.53 (m, 2H, H8/H9), 3.55 (br
s, 2H, H7/H10), 3.64–3.65 (m, 2H, H11/H14), 3.66 (br s, 2H,
H12/H13), 3.68–3.69 (m, 2H, H12/H13), 3.75 (s, 6H, H1), 3.82–3.84
(m, 2H, H8/H9), 4.38–4.39 (m, 2H, H7/H10), 4.80 (s, 6H, H26),
6.72 (d, 4H, 3JHH = 8.8 Hz,
H4), 7.10–7.15 (m, 8H, H17, H18), 7.13 (d, 4H, 3JHH = 8.7 Hz, H3), 7.24 (t, 4H, 3JHH = 7.3 Hz, H22, H23), 7.49–7.56
(m, 4H, H21, H24), 7.65–7.69 (m, 2H, H19). 13C{1H} NMR (100.6 MHz, THF-d8): δ (ppm) 55.4 (s, C1), 61.8
(s, C26), 67.0 (s, C7/C10), 68.5 (s, C7/C10), 71.3 (s, C8/C9), 71.9
(s, C8/C9), 72.7 (s, C12/C13), 73.2 (d, 2JCP = 11.0 Hz, C11/C14), 73.9 (s, C12/C13), 78.5 (d, 1JCP = 52.1 Hz, C15), 80.9 (d, 2JCP = 9.2 Hz, C11/C14), 87.1 (s,
C6), 114.2 (s, C24), 114.4 (s, C4), 122.5 (s, C22), 127.4 (d, 3JCP = 6.5 Hz, C18), 127.9 (s,
C3), 129.6 (d, 2JCP = 7.6 Hz,
C17), 129.7 (d, 1JCP = 45.4
Hz, C20), 131.9 (s, C5), 132.5 (s, C21), 133.4 (s, C19), 135.4 (s,
C23), 141.2 (d, 1JCP = 58.4
Hz, C16), 159.3 (s, C2), 164.5 (s, C25). 31P{1H} NMR (162.0 MHz, THF-d8): δ (ppm) 59.8 (s). ESI(+)-MS: m/z (%) 1190.1 (16.7) [M + Li]+, 598.1 (45.3) [2b + Na]+, 506.1 (100) [2b]+. FT-IR (KBr): ν̃ (cm–1) 3444 m, 3080 w, 2993 w, 2936 m, 2834 m, 2234 w, 2035 w, 1956 w,
1610 m, 1576 m, 1525 s, 1471 m, 1455 s, 1440 s, 1386 w, 1303 w, 1273
m, 1247 s, 1226 m, 1163 m, 1134 w, 1094 w, 1072 w, 1030 s, 1005 m,
908 m, 887 w, 829 m, 794 m, 757 m, 729 m, 695 m, 649 w, 629 w, 591
w, 575 w, 553 m, 533 m, 493 m, 450 w, 413 m. Anal. Calcd for C60H54Cl2Fe2O4P2Ru (1184.1): C, 60.83; H, 4.59. Found: C, 60.66; H, 4.26.
X-ray Crystallography
X-ray diffraction studies were
performed with an Oxford Diffraction CCD Xcalibur-S diffractometer
using Mo Kα radiation (λ = 0.71073 Å) and ω-scan
rotation. Data reduction was performed with CrysAlis Pro[53] including the program SCALE3 ABSPACK[54] for empirical absorption correction. Structures
were solved by direct methods and refined by full-matrix least-squares
techniques against F2 by using the SHELX
program package.[55] All nonhydrogen atoms
were refined with anisotropic thermal parameters and all nonboron-bonded
hydrogen atoms were assigned riding isotropic displacement parameters
and constrained to idealized geometries. Structure figures were generated
with Mercury 4.0.0.[56,57]
Computational Details
All calculations were carried
out in the gas phase using the ORCA program package (version 4.0.0.1).[58,59] Geometry optimizations and frequency analysis, to prove the absence
of imaginary frequencies, were obtained employing the BP86[60−62] functional in conjunction with a def2-TZVP[63,64] basis set. The energies from frequency analyses were used for the
investigation of the reaction pathways. For all calculations, atom-pairwise
dispersion corrections with the Becke–Johnson damping scheme
(D3BJ) and density fitting techniques, also called resolution-of-identity
approximation (RI), were utilized.[65,66]
Authors: Matthew P Sullivan; Michél K Nieuwoudt; Graham A Bowmaker; Nelson Y S Lam; Dianna Truong; David C Goldstone; Christian G Hartinger Journal: Chem Commun (Camb) Date: 2018-06-12 Impact factor: 6.222