Bisphosphine phenol and phenolate complexes of Mn(i): Manganese(i) catalyzed Tishchenko reaction

Article abstract of DOI:10.1039/c8dt02933d

We synthesized new organomanganese complexes using the phenolic "pincer" type ligand H-POP. The coordination chemistry of H-POP with Mn(i) was explored, revealing a wide range of binding motifs. Finally, we found that complex 1 catalyzes the formation of benzyl benzoate from benzaldehyde in a Tishchenko reaction.

Full text of DOI:10.1039/c8dt02933d

Dalton
Transactions
View Article Online
View Journal
COMMUNICATION
Bisphosphine phenol and phenolate complexes
of Mn(I): manganese(I) catalyzed Tishchenko
Cite this: DOI: 10.1039/c8dt02933d
reaction†
Received 18th July 2018,
Accepted 14th August 2018
a
Karthika J. Kadassery, a Samantha N. MacMillanb and David C. Lacy
*
DOI: 10.1039/c8dt02933d
rsc.li/dalton
We synthesized new organomanganese complexes using the phe- Yamane’s bis(diphenylphosphine)phenol ligand.9,13 The syn-
nolic “pincer” type ligand H-POP. The coordination chemistry of thesis is a two step process, the first of which involves dialkyla-
H-POP with Mn(I) was explored, revealing a wide range of binding tion of p-cresol with formaldehyde and amination with
motifs. Finally, we found that complex 1 catalyzes the formation of aqueous dimethylamine forming 2,6-bis[(dimethylamino)
benzyl benzoate from benzaldehyde in a Tishchenko reaction.
methyl]-4-methylphenol (H-NON) in 57% yield (Scheme 1).
The diamine is heated in neat diphenylphosphine wherein a
quinone methide intermediate,14 generated by liberation of di-
methylamine, is attacked by the phosphine. Excess Ph2PH is
important in the final step to ensure that the reaction con-
tinues in solvent, which is easily removed afterwards by
washing with copious amounts of petroleum ether and recov-
ered for future reactions. The oxygen sensitive H-POP product
is a sticky, colorless transparent gel.
With ligand in hand, we attempted various methods of
installing Mn(I) ions to H-POP. In a previous report, we showed
that treatment of Mn2(CO)10 with 6 eq. Me3NO in rigorously
dry toluene afforded the tetranuclear complex [{Mn(CO)3}2
(µ-ONMe3)]{(µ-CO3)[Mn(NMe3)(CO)3]}2 (Mn4).15 Complex Mn4
contains two basic carbonato ligands and two trimethylamine
ligands that participate in deprotonating water and form the
tetrameric complex [Mn(CO)3(µ3-OH)]4. In an analogous reac-
tion, replacing water with H-POP and treating it with in situ
prepared Mn4 afforded POPMn(CO)3 (1) in 39% crystalline
yield. Crystallization of 1 at −35 °C provided suitable crystals
for XRD and revealed meridional Mn(CO)3 coordination
(Fig. 1). Recrystallization of 1 at room temperature however
afforded pure dimeric complex (12) with each of the two Mn(I)
centers binding to one phosphine arm in both ligands (Fig. 2).
The dimeric species 12 is sparingly soluble in organic solvents
and appears to be a sink as heating or irradiating 1 causes its
irreversible conversion to the dimer.
Ruthenium pincer compounds are excellent catalysts in a
variety of applications but suffer from the fact that ruthenium
is an expensive metal.1 In principle, analogous chemistry
should be accessible with low-spin Mn(I) d6 ions because of its
similarity in ionic radii to d6 Ru(II) ions.2 Indeed, recent
advances with manganese catalysts with pincer ligands attests
to this possibility of replacing Ru with non-precious Mn and
warrants further exploration of the coordination chemistry of
Mn(I) on new platforms.3–7 To this end, we have explored the
coordination chemistry and catalytic activity of the first Mn(I)
complexes with bisphosphine phenol and phenolate ligands
and report our findings herein. While POP “pincer” ligands
have been extensively studied,8 phenolic based POP “pincer”
phosphine ligands have not been explored with the exception
of d8 ions such as Rh(I), Ni(II), and Pd(II).9–11 In part, the lack
of phenolic ligands in organometallic catalysis is due to the
fact that electron rich metal ions often oxidatively add across
the phenolic C–O bond, forming a traditional PCP pincer
complex.12 This C–O bond cleavage reaction unfortunately
hinders coordination studies of organometallic phenolic com-
plexes. Conveniently, Mn(I) ions are not prone to undergo oxi-
dative addition and therefore enabled our investigation of the
coordination chemistry of phenolic platforms and possibly dis-
cover new applications in catalysis.
Our entry into the phenolic Mn(I)POP complexes utilized a
known synthetic route to prepare the cresol variant of
aDepartment of Chemistry, University at Buffalo, State University of New York,
Buffalo, New York 14260, USA. E-mail: dclacy@buffalo.edu
bDepartment of Chemistry and Chemical Biology, Cornell University, Ithaca,
New York 14853, USA
†Electronic supplementary information (ESI) available. CCDC 1847482–1847485.
For ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c8dt02933d
Scheme 1 Synthesis of H-POP.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Communication
Dalton Transactions
Fig. 3 Molecular structure of 2; except for the H-atom attached to O5,
H-atoms are removed for clarity (ellipsoids at 50% probability). Color
scheme same as Fig. 1. Selected bond distances and angles: Mn1–Br1 =
2.519(1); Mn1–P = 2.355(1); Mn1–Cave = 1.85 ( 0.04); O5–Br1 = 4.536(2);
O5–Br2 = 3.243(2); Mn2–Br = 2.515(1).
Fig. 1 Molecular structure of 1; H-atoms and DCM solvent molecules
have been removed for clarity (ellipsoids at 50% probability). Color
scheme: Mn = magenta; O = red; P = orange; Br = tan; C = grey.
Selected bond distances (Å) and angles (°): Mn–O4 = 2.051(2); Mn–P1 =
2.3117(6); Mn–P2 = 2.3153(6); Mn–Cave = 1.83 ( 0.04); C22–O4–Mn =
103.4(1); P1–Mn–P2 = 152.38(2).
Scheme 2 Synthesis of H-POP Mn(I) Precursors (yields calculated for
isolated crystalline material).
Fig. 2 Molecular structure of 12; H-atoms and THF solvent molecules
have been removed for clarity (ellipsoids at 50% probability). Color
scheme same as Fig. 1. Selected bond distances (Å) and angles (°): Mn–
O4 = 2.071(2); Mn–P1 = 2.2689(6); Mn–P2 = 2.3125(6); Mn–Cave = 1.83
THF/hexane layering (Scheme 2) in ≥80% crystalline yield, but
the reaction is essentially quantitative. The molecular structure
of 2 shows that the oxygen atom of H-POP does not directly
interact with either of the two Mn ions and that each of the
Mn centers is coordinated to one of the phosphine arms of the
(
0.04); C22–O4–Mn = 124.1(1); P1–Mn–P2 = 176.48(3).
The synthesis of 1 through Mn4 is not high yielding and ligand. The OH moiety is H-bonded to one of the Br ligands as
therefore we explored alternative synthetic routes to 1. indicated by a short O⋯Br distance of 3.245(2) Å. The remain-
Treatment of H-POP with Mn(CO)5Br in a 1 : 1 stoichiometry in der of the coordination sphere for each of the two Mn ions is
THF or DCM for 3 h results in an incomplete coordination occupied by four carbonyl ligands. Assuming Cs symmetry, the
1
reaction between the ligand and metal. Monitoring the reac- structure is consistent with the H NMR and FTIR (solid state
tion with 31P NMR revealed three major species (Fig. S13†), ATR and solution state transmission) spectrum (Fig. S10–12†),
one of which is free unbound ligand (31P NMR −16 ppm). The which has four CO stretches and indicates that the two Mn
other two species are partially bound ligand (free phosphine centers are equivalent by at least a plane of symmetry in
arm, 31P NMR −18 ppm) and {H-POP}{Mn(CO)4Br}2 (2) (31P solution.
NMR 40 ppm), the latter of which was characterized by X-ray
Complex 2 reacts slowly with excess ligand at room temp-
crystallography (Fig. 3). Complex 2 was independently syn- erature forming a new species (31P NMR 55 ppm) designated
thesized by mixing H-POP with Mn(CO)5Br in a 1 : 2 stoichio- 1·HBr. When the mixture is exposed to static vacuum, several
metry in THF at room temperature and recrystallizing from unidentified species form that disappear upon heating to
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Dalton Transactions
Communication
50 °C and results in the complete conversion to the same forms a species with a resonance at 53.92 ppm (31P NMR).
1·HBr species at 55 ppm (Fig. S13†). 1·HBr is conveniently gen- These shifts of about 2 ppm in the 31P NMR is expected for
erated by mixing H-POP with Mn(CO)5Br in a 1 : 1 stoichio- Mn(I) carbonyl-bisphosphine complexes with different co-
metry, degassing with freeze pump thaw cycles, and heating at ordinated halides;17 this serves as our primary evidence that
50 °C for two days (Scheme 3). While no crystal structure of the halide is bonded to the Mn(I) ions in [1HX]2 (X = Cl, Br, I).
1·HBr was obtained, the CHN analysis supports the assigned The new halido species have FTIR spectra essentially identical
empirical formula of 1·HBr·THF and is an essentially quanti- to 1·HBr with the exception of a shift in the νOH from
tative process.
3300 cm−1 for X = Br to 3255 cm−1 for X = Cl and 3360 cm−1
The 31P NMR spectrum of 1·HBr actually contains two for X = I indicating that the O–H groups of the H-POP ligand
peaks (Fig. S14†), one at 55.10 ppm and a major peak at are H-bonded to the halide ion (Fig. S20 and S21†).
55.35 ppm that we tentatively assign to a dimer and a
Desiring to generate a 16-electron complex for catalysis, we
monomer, respectively. The major monomeric component treated 1, 2, and 1·HBr with various reagents to abstract
(Scheme 3), designated [1H]Br, can be isolated from the bromide or CO ligand. For instance, treatment of 2 with one
mixture by washing solid 1·HBr with toluene/Et2O and collect- equiv. of KN(SiMe3)2 in THF at room temperature or −78 °C
ing the remaining solids (the washings however always contain produced complicated mixtures that convert to 1 or its dimer,
both species). The 1H NMR spectrum of [1H]Br in DCM dis- 12, in less than a day. Interestingly, 1 is also produced when 2
plays peaks similar in position to 1 and 2 but significantly is treated with two equivalents of Me3NO in THF at room
more broad. The FTIR spectrum of [1H]Br contains three νCO temperature or −78 °C along with a white ppt characterized as
in a pattern that is expected for a monomeric [H-POPMn(CO)3] Me3N·HBr. Attempts to remove the bromide ligands of 1·HBr
Br formulation with meridional CO coordination (Fig. S17†). were performed in DCM with AgOTf, but this resulted in the
Treatment of this tentatively assigned monomer [1H]Br with formation of the triflate salt of the mono-protonated dimer
Et3N furnishes 1 and [Et3NH]Br in quantitative yield. However, (12H-OTf) and other intractable products. The 31P NMR spec-
the mixture of the monomeric and dimeric species of 1·HBr trum of 12H-OTf contains two doublets at 73.22 and
produces an equivalent mixture of 1 and 12 when reacted with 63.52 ppm (JPP = 43.56 Hz) (see ESI†). The same phosphorous
Et3N. The process is apparently reversible as a reaction signal was observed upon attempts to remove Br from 1·HBr
between 1 in THF and HBr(aq) forms 1·HBr (vide infra); a with Li[B(ArF5)4] and also treatment of H-POP with [Mn(CO)5]
similar reversible ligand protonation/deprotonation was OTf. Efforts to generate a 16-electron dicarbonyl complex from
recently observed by Kirchner and coworkers.16
1 were made by treating 1 with heat, photons and/or Me3NO.
A possible structure for the second component of 1·HBr is a As mentioned earlier, heating or irradiating 1 led to dimeriza-
dimeric molecule, [1HBr]2, similar to 12, with inner sphere tion to 12. Treatment of 1 with Me3NO led to no reaction at
bromido ligands and non-coordinating phenol O-atoms room temperature and heating the mixture led to decompo-
(Scheme 3). The –OH group on phenol might be hydrogen sition of the metal complex (the solution turned colorless
bonded to the coordinated Br ligand, similar to the interaction from yellow and showed no CO peaks in FTIR).
we observed in 2. Furthermore, the structure proposed for
Despite our current lack of success on producing 16 elec-
[1HBr]2 can easily be reconciled with the synthetic protocol for tron Mn(I) species with the H-POP platform, we also explored
1·HBr from 2. Yet another string of evidence supporting our the efficacy of 1 as a hydrogenation catalyst under various con-
assignment is that treatment of 1 in THF with HCl(aq) forms a ditions. During these investigations, we serendipitously discov-
new single species with a resonance at 57.85 ppm in the 31P ered that 1 catalyzes the Tishchenko reaction with benz-
NMR spectrum. Similarly, treatment of 1 in THF with HI(aq) aldehyde in toluene (Scheme 4).18 For instance, a 4 mol%
loading of 1 in toluene with 0.4 M benzaldehyde and heating
at 120 °C for 24 hours furnished benzyl benzoate with 26%
conversion (TON ≈ 7, 24 h). Importantly, 1 does not react with
benzyl alcohol in toluene, which serves as good evidence that
the formation of benzyl benzoate proceeds through
a
Tishchenko reaction rather than one that involves dehydro-
genation of alcohols or intermediate hemiacetals.19–21
We also tested other H-POP complexes (Table S1†) and
interestingly found that 12 converts exactly twice as much reac-
Scheme 3 Preparation and proposed speciation of 1·HBr in THF.
Scheme 4 Mn(I) catalyzed Tishchenko reaction.
This journal is © The Royal Society of Chemistry 2018
Dalton Trans.

View Article Online
Communication
Dalton Transactions
References
1 C. Gunanathan and D. Milstein, Chem. Rev., 2014, 114,
12024–12087.
2 T. Zell and R. Langer, ChemCatChem, 2018, 10, 1930–1940.
3 F. Kallmeier and R. Kempe, Angew. Chem., Int. Ed., 2018,
57, 46–60.
Scheme 5 Possible Mn(I) intermediates in the Tishchenko reaction.
4 B. Maji and M. K. Barman, Synthesis, 2017, 49, 3377–3393.
5 M. Garbe, K. Junge and M. Beller, Eur. J. Org. Chem., 2017,
4344–4362.
6 A. Mukherjee, A. Nerush, G. Leitus, L. J. W. Shimon,
Y. Ben-David, N. A. E. Jalapa and D. Milstein, J. Am. Chem.
Soc., 2016, 138, 4298–4301.
7 S. Elangovan, C. Topf, S. Fischer, H. Jiao, A. Spannenberg,
W. Baumann, R. Ludwig, K. Junge and M. Beller, J. Am.
Chem. Soc., 2016, 138, 8809–8814.
8 G. M. Adams and A. S. Weller, Coord. Chem. Rev., 2018,
355, 150–172.
tant than 1. Considering the formation of 12 from 1 via heat or
photolysis, we propose that a phosphine arm in 1 or 12 dis-
sociates opening up a vacant site for the Mn(I)-phenoxide
center to interact with benzaldehyde. Typical catalysts for the
Tishchenko reaction are in fact metal alkoxides and therefore
might provide some rationale as to why 1, and not 2, catalyzes
the reaction. Specifically, the deprotonated phenoxide moiety
in 1 serves to facilitate proton transfer or even a metal–ligand
cooperative (MLC) step. The latter MLC reaction could result in
a Mn(I)-hydrido intermediate with a dearomatized POP ligand
in a quinone methide structure (Scheme 5).
9 C. Zhu, N. Yukimura and M. Yamane, Organometallics,
2010, 29(9), 2098–2103.
In summary, we have explored the synthesis, coordination
chemistry, and catalysis of a new pincer platform with Mn(I). A
general finding is that the phenolic oxygen atom appears to
weakly bind to Mn(I). Our conclusion on this matter is inferred
by the small PhC–O–Mn angles in 1, long or non-bonding
Mn–O interactions, and our observation that the δOH and νOH
do not shift much from the various complexes. This weak
O–Mn interaction might also be responsible for the observed
vast landscape of accessible species with the H-POP ligand,
many of which were only observed as intermediates in
31P NMR experiments. In addition, the studies herein are
restricted to the cresol variant with only phenyl substituents
on the phosphine arms. Undoubtedly, the tunability of
phenols and phosphine ligands will furnish a wealth of new
chemistry, which we are actively pursuing.
10 M. E. van der Boom, S.-Y. Liou, Y. Ben-David,
L. J. W. Shimon and D. Milstein, J. Am. Chem. Soc., 1998,
120, 6531–6541.
11 A. Vigalok, B. Rybtchinski, Y. Gozin, T. S. Koblenz, Y. Ben-
David, H. Rosenberg and D. Milstein, J. Am. Chem. Soc.,
2003, 125, 15692–15693.
12 M. E. van der Boom, T. Zubkov, A. D. Shukla,
B. Rybtchinski, L. J. W. Shimon, H. Rozenberg, Y. Ben-
David and D. Milstein, Angew. Chem., Int. Ed., 2004, 43,
5961–5963.
13 K. Kellner, S. Rothe, E. M. Steyer and A. Tzschach,
Phosphorus Sulfur Relat. Elem., 1980, 8(3), 269–273.
14 Đ. Škalamera, C. Bohne, S. Landgraf and N. Basarić, J. Org.
Chem., 2015, 80(21), 10817–10828.
15 K. J. Kadassery, S. K. Dey, A. E. Friedman and D. C. Lacy,
Inorg. Chem., 2017, 56, 8748–8751.
16 D. Himmelbauer, B. Stöger, L. F. Veiros and K. Kirchner,
Conflicts of interest
Organometallics,
2018,
DOI:
10.1021/acs.
organomet.8b00193.
17 S. J. A. Pope and G. Reid, J. Chem. Soc., Dalton Trans., 1999,
There are no conflicts to declare.
1615–1621.
18 S. A. Morris and D. G. Gusev, Angew. Chem., Int. Ed., 2017,
56, 6228–6231.
Acknowledgements
Financial support was provided by the University at Buffalo 19 N. Menashe and Y. Shvo, Organometallics, 1991, 10, 3885–
start-up funds and an ACS Petroleum Research Fund
3891.
(ACS-PRF-57861-DN13). This work was completed using the 20 J. Zhang, G. Leitus, Y. Ben-David and D. Milstein, J. Am.
resources of the Chemistry Instrument Center (CIC), University
Chem. Soc., 2005, 127, 10840–10841.
at Buffalo, SUNY, Buffalo, NY; che-ic@buffalo.edu for sample 21 D. H. Nguyen, X. Trivelli, F. Capet, J.-F. Paul, F. Dumeignil
inquiries.
and R. M. Gauvin, ACS Catal., 2017, 7, 2022–2032.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2018