Potassium-alkane interactions within a rigid hydrophobic pocket

Article abstract of DOI:10.1002/anie.201207962

A NON-issue: Potassium complexes of an extremely rigid and sterically encumbered NON-donor ligand have been prepared, and the solid-state structures (see figure) feature remarkably short potassium-alkane distances. DFT calculations highlight the presence of an electrostatic cation-induced dipole potassium-alkane interaction supported by interactions between the alkane and the surrounding ligand framework. Copyright

Full text of DOI:10.1002/anie.201207962

.
Angewandte
Communications
DOI: 10.1002/anie.201207962
Alkali Metals
Potassium–Alkane Interactions within a Rigid Hydrophobic Pocket**
Nicholas R. Andreychuk and David J. H. Emslie*
Metal–alkane complexes are of importance because of their
involvement in alkane C H activation reactions[1] and hydro-
À
carbon adsorption in alkali-metal-containing zeolites.[2] How-
ever, observable metal–alkane complexes are scarce as
a consequence of the poor donor/acceptor character of
À
alkanes and the low polarity of C H bonds. Examples
detected by NMR spectroscopy include [(C5R5)Re(CO)2-
(alkane)],[3,4] [(C5R5)M(CO)(PF3)(alkane)] (M = Re or
Mn),[4]
[TpRe(CO)2(alkane)],[5]
[(PONOP)Rh(CH4)]+
{PONOP = 2,6-(tBu2PO)2C5H3N},[6] and [(C6Et6)W(CO)2-
(n-pentane)],[7] but none of these complexes have proven
sufficiently robust to allow isolation or crystallization. At the
other end of the spectrum are the crystallographically
characterized metal–alkane complexes[8] which have not
been observed in solution. The only members of this group
are the iron(II) double A-frame porphyrin–heptane complex
reported by Reed and co-workers,[9] the uranium(III)–alkane
complexes reported by Meyer and co-workers,[10] and a rho-
dium(I) norbornane complex reported by Weller and co-
workers,[11] and in all cases the metal–alkane interaction is
considered to possess some degree of covalency, perhaps with
additional stabilization from interactions between the alkane
and the ligand framework. Herein we describe potassium
complexes of a new highly rigid and sterically encumbered
NON-donor ligand, all of which feature remarkably short
intermolecular potassium–alkane distances in the solid state.
Palladium-catalyzed coupling of 4,5-dibromo-2,7-di-tert-
butyl-9,9-dimethylxanthene with 2 equivalents of 2,6-dimesi-
tylaniline afforded 4,5-bis(2,6-dimesitylanilino)-2,7-di-tert-
butyl-9,9-dimethylxanthene [H2(XAT); Scheme 1], which is
an extremely sterically hindered analogue of the known 4,5-
bis(2,6-diisopropylanilino)-2,7-di-tert-butyl-9,9-dimethylxan-
thene[12] and 4,5-bis(2,4,6-trimethylanilino)-2,7-di-tert-butyl-
9,9-dimethylxanthene[13] pro-ligands. Reaction of H2(XAT)
with excess KH in toluene yielded the dipotassium salt, and
filtration and layering with hexanes at À308C deposited
vibrant yellow X-ray quality crystals of [K2(XAT)(n-hexa-
ne)]·toluene (1; Scheme 1 and Figure 1). The potassium atoms
in 1 are bound to bridging amido and ether donors, forming
Scheme 1. Synthesis of H2(XAT) and 1–6. DPEPhos=bis{2-(diphenyl-
phosphino)phenyl}ether], Dmp=2,6-dimesitylphenyl, Mes=mesityl.
Figure 1. Two views of the X-ray crystal structure of [K2(XAT)-
(n-hexane)]·toluene (1). Hydrogen atoms and lattice solvent are
omitted for clarity and ellipsoids are at 50% probability. K-C distances
less than 3.50 ꢀ are highlighted as dotted lines.
a square pyramidal K2N2O core with oxygen in the apical site.
However, an unexpected feature is the close approach of
a molecule of n-hexane to K(1), with a K(1)-C(1S) distance of
3.284(4) ꢀ.
The Fe-C distances in Reedꢁs iron porphyrin heptane
complex are 2.5 and 2.8 ꢀ (the heptane molecule and the Fe
atoms are disordered; calculated Fe-C distances for methane,
ethane, propane, and butane complexes are 2.68–2.70 ꢀ),[9]
the U-C distances in Meyerꢁs uranium–alkane complexes
range from 3.731(8) to 3.864(7) ꢀ (the calculated U-C
distance for the methylcyclohexane complex is 3.974 ꢀ),[10]
and the Rh-C distances in Wellerꢁs rhodium norbornane
complex are 2.480(11) and 2.494(10) ꢀ.[11] To enable a rough
comparison between the M-C (M = metal) distances in the
more ionic uranium complex and complex 1, ionic radii for
U3+ and K+ (1.03 and 1.38 ꢀ for a coordination number of
six)[14] may be subtracted from the crystallographic M-C
distances, yielding values of 2.70–2.83 and 1.90 ꢀ, respec-
tively. The K-C distance in complex 1 is therefore notably
[*] N. R. Andreychuk, Prof. D. J. H. Emslie
Department of Chemistry, McMaster University
1280 Main Street West, Hamilton, ON, L8S 4M1 (Canada)
E-mail: emslied@mcmaster.ca
Homepage: http://www.chemistry.mcmaster.ca/emslie/
emslie.html
[**] D.J.H.E. thanks the NSERC of Canada for a Discovery Grant. N.R.A.
thanks the NSERC of Canada and the province of Ontario for CGS-M
and OGS support. We are also grateful to H. A. Jenkins for help with
X-ray crystallography, and to P. W. Ayers and I. Vargas-Baca for
helpful discussions regarding DFT calculations.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201207962.
1696
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1696 –1699

Angewandte
Chemie
short, and even falls at the lower end of the range of K-C
distances observed for face-on potassium–benzene and potas-
sium–toluene interactions, which are typically 3.2 to
3.5 ꢀ.[15,16] The potassium–alkane interaction in 1 can be
surmised to involve a weak electrostatic (primarily cation-
induced dipole)[17] potassium–alkane interaction stabilized by
interactions between the alkane and the hydrophobic ligand
pocket.
An analogous intermolecular potassium–alkane interac-
tion is not observed at K(2), perhaps as a result of crystal
packing forces as the para-methyl carbon C(48) of a mesityl
group in an adjacent [K2(XAT)(n-hexane)] molecule is
positioned 3.538(3) ꢀ from K(2). However, both potassium
atoms in 1 are forced into close proximity with flanking
mesityl groups and the xanthene backbone, leading to a large
number of K-Carene and K-Cmethyl distances that are below
3.50 ꢀ (Figure 1). In particular, the intramolecular K-Cmethyl
distances K(2)-C(56) and K(1)-C(76) are 3.180(3) and
3.230(3) ꢀ, respectively. For comparison, the intramolecular
K–CHR3 interactions in the sterically encumbered [{KSi-
(SiMe3)3}2],[16]
[KC(SiMe3)3]n,[18]
and
[K2(O{SiMe2C-
(SiHMe2)2}]n[19] range from 3.138(3) to 3.433(3) ꢀ.
To further probe the disposition of the K2(XAT) moiety to
interact with the hydrocarbon solvent, alternative crystalliza-
tion conditions were explored, yielding X-ray quality crystals
of [K2(XAT)(n-pentane)]·toluene (2), [K2(XAT)(3-methyl-
pentane)]·3-methylpentane (3), [K2(XAT)(cyclopentane)]·
cyclopentane (4), [K2(XAT)(toluene)]·0.5toluene (5), and
[K2(XAT){(Me3Si)2O}2] (6) (Scheme 1 and Figure 2). The
central core of structures 2–6 is analogous to that in 1 (each
potassium atom is NON-coordinated and engages in intra-
molecular potassium–carbon interactions with surrounding
mesityl groups), and in every case either one (2–5) or two (6)
intermolecular K–H3CR or K–H2CR2 interactions are
observed. These interactions involve the 1-position of n-
pentane and 3-methylpentane, one of the CH2 groups in
cyclopentane, and a methyl group of toluene and hexame-
thyldisiloxane, leading to K-C distances of 3.358(5) ꢀ in 2,
3.215(5) ꢀ in 3, 3.48(1) and 3.62(3) ꢀ in 4,[20] 3.285(7) and
3.305(9) ꢀ in 5, and 3.282(5) and 3.332(5) ꢀ in 6 (bound
cyclopentane in 4 and toluene in 5 are disordered over 2
positions). In 5, toluene bridges between adjacent molecules
through K–Carene interactions with distances of 3.240(7),
3.425(9), and 3.433(8) ꢀ (Figure 2). The K-C-C angles in
primary alkyl complexes 1, 2, and 3 are 1178, 1548, and 1708,
respectively, the K-Cmethyl-C angles in 5 are 998 and 1088, and
the K-C-Si angles in 6 are 171 and 1768.
Figure 2. X-ray crystal structures of a) [K2(XAT)(n-pentane)]·toluene (2),
b) [K2(XAT)(3-methylpentane)]·3-methylpentane (3), c) [K2(XAT)-
(toluene)]·0.5toluene (5), d) [K2(XAT)(cyclopentane)]· cyclopentane (4),
and e) [K2(XAT){(Me3Si)2O}2] (6). Hydrogen atoms and lattice solvent
are omitted for clarity. Only one of the two orientations of cyclo-
pentane and toluene are shown in the structures of 4 and 5. Ellipsoids
are shown at 50% probility for 2–5 (collected at 100 K) and 30%
probability for 6 (collected at 223 K). K-C distances below 3.50 ꢀ are
highlighted as dotted lines.
nature of the potassium–alkane interaction in 3, which is the
complex with the shortest K-Calkane distance. Geometry
optimization using the BLYP-D3-BJ functional yielded a sub-
stantially shorter K-Calkane interaction (3.218 ꢀ) than was
observed using BLYP (4.176 ꢀ), and is consistent with
significant stabilization of the potassium–alkane interaction
through dispersion interactions. The calculated K-Calkane
distance in 3 shows excellent agreement with the crystallo-
graphic distance [3.215(5) ꢀ], and the three hydrogen atoms
on C(1S) are located 2.82, 3.08, and 3.15 ꢀ from K(1). The
BLYP-D3-BJ functional also adequately reproduced the K-C-
C angle (168.18; see 170.28 in the X-ray crystal structure), the
conformation of 3-methylpentane, and the K-N, K-O, and
intramolecular K-C distances. However, the exact position of
3-methylpentane (especially the two more remote carbon
atoms), within the binding pocket deviates to some extent
from that in the X-ray structure (see Figure S4 in the
Supporting Information).
Compounds 1–6 illustrate the extent to which intermo-
lecular K–H3CR and K–H2CR2 interactions are a common
feature of the solid-state structures of K2(XAT). However,
1
attempts to observe alkane or O(SiMe3)2 binding by H or
13C NMR spectroscopy in 3-methylpentane/[D8]toluene
(À808C),
3-methylpentane
(À1108C),
cyclopentane
(À808C), or O(SiMe3)2 (À608C; 1H NMR spectroscopy
only) were unsuccessful, possibly as a result of rapid exchange
between free and bound solvent.
Complex 3 was investigated through a fragment approach
which considered the interaction between the K2(XAT) and
3-methylpentane fragments in the geometries adopted in the
calculated structure of 3. Within this approach, the energy
DFT calculations (ADF 2012.01, BLYP with and without
Grimmeꢁs DFT-D3-BJ dispersion correction, TZ2P all-elec-
tron, gas phase, VWN, ZORA) were carried out to probe the
decomposition analysis[21] of Ziegler and Rauk[22] (DEint
=
Angew. Chem. Int. Ed. 2013, 52, 1696 –1699
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1697

.
Angewandte
Communications
DEelec + DEorb + DEdisp + DEPauli; BSSE correction included)
yielded a total interaction energy (DEint) of À54.2 kJmolÀ1,
which comprises DEelec = À31.6 kJmolÀ1 (electrostatic inter-
action energy, calculated using frozen charge distributions for
both fragments), DEorb = À16.6 kJmolÀ1 (orbital interaction
energy; this term includes all contributions resulting from
intrafragment polarization), DEdisp = À87.3 kJmolÀ1 (disper-
sion interactions), and DEPauli = 81.2 kJmolÀ1 (Pauli repul-
sion). The total interaction energy, DEint, differs from the true
interaction energy only by the energy needed to bring the
fragments from their optimum geometries to their geometries
in 3. This preparation energy (DEprep) is less than 5% of the
value of DEint.
the hydrophobic binding pocket to stabilize the potassium–
alkane interaction in 3 therefore extends beyond the realm of
dispersion interactions.
In summary, this work presents the first main-group-
metal–alkane interactions to have been observed crystallo-
graphically, and provides a unique opportunity for the
computational study of well-defined potassium–alkane inter-
actions. The combination of cation-induced dipole electro-
static bonding supported by interactions between the alkane
and the surrounding framework is a feature common to both 3
and the potassium–alkane interactions proposed to occur in
certain potassium-containing zeolites.[2] The effectiveness of
the rigid hydrophobic binding pocket in K2(XAT) to promote
and stabilize even very weak potassium–alkane interactions
(as shown crystallographically in the solid state and computa-
tionally in the gas phase) also suggests that in combination
with catalytically relevant metals, ligands featuring a rigid
hydrophobic binding pocket may have untapped potential in
The closed-shell interactions (DEdisp and DEPauli) roughly
cancel out, leaving a net contribution of À6.0 kJmolÀ1. DEint is
therefore approximately equal to DEelec + DEorb. The SCF
deformation density isosurfaces in Figure 3 highlight both the
À
alkane C H activation chemistry.
Experimental Section
Full experimental and characterization details for H2(XAT) and 1–6,
and the details of DFT calculations on structures 3 and 3’ are included
in the Supporting Information. CCDC 903943 (1), 903944 (2), 903945
(5), 903946 (4), 903947 (3) and 903948 (6) contain the supplementary
crystallographic data for this paper. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
Figure 3. SCF deformation density isosurfaces (set to Æ0.0005 in
a and Æ0.0002 in b) from fragment analysis of 3. Blue represents
increased electron density and red represents depleted electron density
relative to the initial K2XAT and 3-methylpentane fragments.
Received: October 2, 2012
Published online: January 4, 2013
Keywords: alkali metals · alkanes · ligand design ·
noncovalent interactions · potassium
degree of polarization within the alkane fragment and the
extent to which the electron density around potassium
remains largely unchanged, indicating that DEorb is primarily
due to a cation-induced dipole electrostatic interaction rather
.
À
than s-donation from alkane C H bonds to potassium.
Negligible Hirshfeld charges on the K2(XAT) and alkane
fragments (<Æ 0.003) support this interpretation.
[1] a) B. A. Arndtsen, R. G. Bergman, T. A. Mobley, T. H. Peterson,
Acc. Chem. Res. 1995, 28, 154 – 162; b) J. A. Labinger, J. E.
Bercaw, Nature 2002, 417, 507 – 514; c) R. H. Crabtree, J. Chem.
Soc. Dalton Trans. 2001, 2437 – 2450.
[2] a) A. M. Ferrari, K. M. Neyman, S. Huber, H. Knozinger, N.
Rosch, Langmuir 1998, 14, 5559 – 5567; b) S. Calero, D. Dub-
beldam, R. Krishna, B. Smit, T. J. H. Vlugt, J. F. M. Denayer,
J. A. Martens, T. L. M. Maesen, J. Am. Chem. Soc. 2004, 126,
11377 – 11386; c) B. Liu, B. Smit, Phys. Chem. Chem. Phys. 2006,
8, 1852 – 1857.
[3] a) S. Geftakis, G. E. Ball, J. Am. Chem. Soc. 1998, 120, 9953 –
9954; b) J. A. Calladine, O. Torres, M. Anstey, G. E. Ball, R. G.
Bergman, J. Curley, S. B. Duckett, M. W. George, A. I. Gilson,
D. J. Lawes, R. N. Perutz, X. Z. Sun, K. P. C. Vollhardt, Chem.
Sci. 2010, 1, 622 – 630.
[4] G. E. Ball, C. M. Brookes, A. J. Cowan, T. A. Darwish, M. W.
George, H. K. Kawanami, P. Portius, J. P. Rourke, Proc. Natl.
Acad. Sci. USA 2007, 104, 6927 – 6932.
Calculations were also performed on a simplified model
for 3 (structure 3’) in which XAT methyl and mesityl groups
have been replaced by hydrogen atoms, but the coordinates of
all other atoms are identical to those in the BLYP-D3-BJ/
TZ2P structure of 3. This model lacks much of the hydro-
phobic pocket surrounding the metal, and fragment analysis
yielded values of À32.3, À15.8, À10.7, À36.5, and
30.6 kJmolÀ1 for DEint, DEelec, DEorb, DEdisp, and DEPauli
,
respectively. The less-negative interaction energy in 3’ high-
lights the importance of the hydrophobic pocket in stabilizing
the potassium–alkane interaction. However, DEdispÀDEPauli is
approximately the same in 3 and 3’, so the more-negative
interaction energy for 3 can be considered to arise from more-
negative DEelec and DEorb contributions. Differences in DEelec
will primarily reflect the number of electron–nucleus inter-
actions between the alkane and the ligand framework, and the
more negative DEorb for 3 versus 3’ implies stabilization of the
polarized alkane through interactions with the surrounding
pocket (Figure 3b indicates polarization of the adjacent
ligand framework by the polarized alkane). The ability of
[5] S. B. Duckett, M. W. George, O. S. Jina, S. L. Matthews, R. N.
Perutz, X. Z. Sun, K. Q. Vuong, Chem. Commun. 2009, 1401 –
1403.
[6] W. H. Bernskoetter, C. K. Schauer, K. I. Goldberg, M. Brook-
hart, Science 2009, 326, 553 – 556.
[7] R. D. Young, D. J. Lawes, A. F. Hill, G. E. Ball, J. Am. Chem.
Soc. 2012, 134, 8294 – 8297.
1698
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1696 –1699

Angewandte
Chemie
[8] The binding of ethane and propane to iron centers in an
extended metal – organic framework was recently characterized
by neutron diffraction: E. D. Bloch, W. L. Queen, R. Krishna,
J. M. Zadrozny, C. M. Brown, J. R. Long, Science 2012, 335,
1606 – 1610.
[9] D. R. Evans, T. Drovetskaya, R. Bau, C. A. Reed, P. D. W. Boyd,
J. Am. Chem. Soc. 1997, 119, 3633 – 3634.
[10] I. Castro-Rodriguez, H. Nakai, P. Gantzel, L. N. Zakharov, A. L.
Rheingold, K. Meyer, J. Am. Chem. Soc. 2003, 125, 15734 –
15735.
[11] S. D. Pike, A. L. Thompson, A. G. Algarra, D. C. Apperley, S. A.
Macgregor, A. S. Weller, Science 2012, 337, 1648 – 1651.
[12] C. A. Cruz, D. J. H. Emslie, L. E. Harrington, J. F. Britten, C. M.
Robertson, Organometallics 2007, 26, 692 – 701.
[13] R. M. Porter, A. A. Danopoulos, Polyhedron 2006, 25, 859 – 863.
[14] R. D. Shannon, Acta Crystallogr. Sect. A 1976, 32, 751 – 767.
[15] a) J. Hao, H. Song, C. Cui, Organometallics 2009, 28, 3100 – 3104;
b) M. Niemeyer, Inorg. Chem. 2006, 45, 9085 – 9095; c) P. B.
Hitchcock, M. F. Lappert, R. Sablong, Dalton Trans. 2006, 4146 –
4154; d) W. J. Evans, D. B. Rego, J. W. Ziller, Inorg. Chem. 2006,
45, 3437 – 3443; e) G. C. Forbes, A. R. Kennedy, R. E. Mulvey,
B. A. Roberts, R. B. Rowlings, Organometallics 2002, 21, 5115 –
5121; f) G. Bai, H. W. Roesky, M. Noltemeyer, H.-G. Schmidt, J.
Chem. Soc. Dalton Trans. 2002, 2437 – 2440; g) C. J. Schaverien,
J. B. Vanmechelen, Organometallics 1991, 10, 1704 – 1709;
h) G. R. Fuentes, P. S. Coan, W. E. Streib, K. G. Caulton,
Polyhedron 1991, 10, 2371 – 2375.
[16] K. W. Klinkhammer, Chem. Eur. J. 1997, 3, 1418 – 1431.
[17] a) E. V. Anslyn, D. A. Dougherty, Modern physical organic
chemistry, University Science Books, Sausalito, CA, 2006;
b) A. W. Ehlers, C. G. de Koster, R. J. Meier, K. Lammertsma,
J. Phys. Chem. A 2001, 105, 8691 – 8695.
[18] C. Eaborn, P. B. Hitchcock, K. Izod, A. J. Jaggar, J. D. Smith,
Organometallics 1994, 13, 753 – 754.
[19] L. J. Bowman, K. Izod, W. Clegg, R. W. Harrington, J. D. Smith,
C. Eaborn, Dalton Trans. 2006, 502 – 508.
[20] The K-C distances in 4 should be viewed with some caution since
bound cyclopentane is disordered over two positions and
À
restraints had to be applied to ensure reasonable C C bond
distances (DFIX was used to set all five C-C distances to 1.41 ꢀ
with an ESD of 0.01 ꢀ).
[21] F. M. Bickelhaupt, E. J. Baerends in Reviews in Computational
Chemistry, Vol. 15 (Eds.: K. B. Lipkowitz, D. B. Boyd), Wiley-
VCH, New York, 2000, pp. 1 – 86.
[22] a) T. Ziegler, A. Rauk, Inorg. Chem. 1979, 18, 1755 – 1759; b) T.
Ziegler, A. Rauk, Inorg. Chem. 1979, 18, 1558 – 1565.
Angew. Chem. Int. Ed. 2013, 52, 1696 –1699
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 1699