Regioselective double cycloplatination of 9,10-bis(diphenylphosphino) anthracene

Article abstract of DOI:10.1021/om800356k

Double eyclometalation of 9,10-bis(diphenylphosphino)anthracene (PAnP) by Pt(L)(OTf)₂ (L = diphosphines, OTf = CF₃SO₃) gives rise to the geometrical isomers syn- and anti-[Pt₂(L) ₂(PAnP-H₂)

Full text of DOI:10.1021/om800356k

Organometallics 2009, 28, 1093–1100
1093
Regioselective Double Cycloplatination of
9,10-Bis(diphenylphosphino)anthracene
Jian Hu and John H. K. Yip*
Department of Chemistry, National UniVersity of Singapore, 3 Science DriVe 3, Singapore, 117543
ReceiVed April 21, 2008
Double cyclometalation of 9,10-bis(diphenylphosphino)anthracene (PAnP) by Pt(L)(OTf)2 (L )
diphosphines, OTf ) CF3SO3) gives rise to the geometrical isomers syn- and anti-[Pt2(L)2(PAnP-H2)](OTf)2
(Pt2). The reaction is regioselective, with the syn-isomer being the kinetic product. The cyclomelation is
reversible, and thermodynamic product distribution is obtained after prolonged standing. The ratio of the
isomers is subject to the influence of solvents and ancillary diphosphines. Protonolysis of the Pt-C leads
to a monocyclometalated intermediate. A mechanistic postulate that invokes a preferential electrophilic
attack of the Pt ions at the C-H bonds of the anthracenyl ring is proposed to explain the regioselectivity.
standing the origin of regioselectivity not only is an important
step toward synthetic application3 of the reaction but also
provides insights into important problems such as C-H bond
and metal-ligand activations.
Introduction
Cyclometalation is a facile and common way to convert a
C-H bond to a metal-carbon bond.1 The reaction begins with
coordination of a metal atom to a ligand (i.e., phosphines and
amines) that has pendent alkyl or aryl groups, followed by
cleavage of a C-H bond in a pendent aryl or alkyl group, and
finally formation of metal-carbon bond and metallacycle. The
mechanisms proposed for the C-H bond activation are elec-
trophilic attack, nucleophilic addition, and σ-bond metathesis.1c
Regioselective cyclometalation could happen to ligands that
have more than one C-H bond that is reactive toward the
incoming metal atom.2 Ryabov1c classified regioselective cy-
clometalation into enforced regioselectivity, which is dictated
by electronic or steric factors, and regulated regioselectivity,
which can be altered by changing reaction conditions. Under-
Double cyclometalation is possible for ligands that have two
donor groups and more than one activated C-H bond.1c,2a-q
The reaction can produce geometrical isomers, and their yields
could be different if the attack of the second metal atom is
regioselective. The origin of regioselectivity in double cyclo-
metalation is an intriguing problem. The first question is whether
the first metal atom directs and activates/deactivates the second
attack. It is related to the mechanism of the C-H bond cleavage
and metal-ligand interactions. Another question concerns
electronic, steric, and solvent effects on the regioselectivity,
which is related to the question of whether the reaction is
kinetically or thermodynamically controlled. So far our under-
standing of regioselectivity of double cyclometalation is rather
limited, despite the fact that it was first observed by Trofimenko
more than 30 years ago.2a Subsequent studies on double
cyclometalation of 1,4-disubstituted benzenes showed that the
para-dimetalation is favored over ortho-dimetalation mainly due
to steric factors.2a-q Notably, Steel reported a systematic inves-
tigation of double cyclopalladation of a series of bis(2-pyridyloxy)-
napthalenes,2c,g showing that the reaction is kinetically con-
trolled, but thermodynamic products could be obtained at high
temperature.
Recently we reported the syntheses and electronic spectros-
copy of the doubly cycloplatinated complexes [Pt2(L)2(PAnP-
H2)](OTf)2 (Pt2) (L ) bis(diphenylphosphino)methane (dppm),
1,2-bis(diphenylphosphino)ethane (dppe), 1,3-bis(diphenylphos-
phino)propane (dppp), PAnP ) 9,10-bis(diphenylphosphino)an-
thracene) (Scheme 1).4 Our further study showed that the double
cyclometalation of PAnP by PtII(L)(OTf)2 is very fast, producing
two geometrical isomers, syn- and anti-Pt2. The second cyclo-
platination is found to be regioselective, and the selectivity
depends on solvent and L. Reported in this paper is our effort
* Correspondence author. E-mail: chmyiphk@nus.edu.sg. Fax: 65-
67791691.
(1) (a) Kleiman, J. P.; Dubeck, M. J. Am. Chem. Soc. 1963, 85, 1544.
(b) Chatt, J.; Davidson, J. M. J. Chem. Soc. 1965, 843. (c) Ryabov, A. D.
Chem. ReV. 1990, 90, 403. (d) Canty, A. J.; van Koten, G. Acc. Chem. Res.
1995, 28, 406. (e) Newkome, G. R.; Puckett, W. E.; Gupta, V. K.; Kiefer,
G. E. Chem. ReV. 1986, 86, 451. (f) Dehand, J.; Pfeffer, M. Coord. Chem.
ReV. 1976, 18, 327. (g) Newkome, G. R.; Puckett, W. E.; Gupta, V. K.;
Kiefer, G. E. Chem. ReV. 1986, 86, 451. (h) Omae, I. Organometallic
Intramolecular-Coordination Compounds; Elsevier: Amsterdam, 1986. (i)
Bruce, M. I. Angew. Chem., Int. Ed. Engl. 1977, 16, 73.
(2) (a) Trofimenko, S. J. Am. Chem. Soc. 1971, 93, 1808. (b) Trofimenko,
S. Inorg. Chem. 1973, 12, 1215. (c) Hartshorn, C. M.; Steel, P. J.
Organometallics 1998, 17, 3487. (d) Steenwinkel, P.; James, S. L.; Grove,
D. M.; Kooijman, H.; Spek, A. L.; van Koten, G. Organometallics 1997,
16, 513. (e) van der Boom, M. E.; Liou, S.-Y.; Shimon, L. J. W.; Ben-
David, Y.; Milstein, D. Organometallics 1996, 15, 2562. (f) Slater, J. W.;
Lydon, D. P.; Alcock, N. W.; Rourke, J. P. Organometallics 2001, 20, 4418.
(g) O’Keefe, B. J.; Steel, P. J. Organometallics 2003, 22, 1281. (h) Sumby,
C. J.; Steel, P. J. Organometallics 2003, 22, 2358. (i) Philips, I. G.; Steel,
P. G. J. Organomet. Chem. 1991, 410, 247. (j) Frena´ndez, A.; Frena´ndez,
J. J.; Lo´pez-Torres, M.; Sua´rez, A.; Ortigueira, J. M.; Vila, J. M.; Adams,
H. J. Organomet. Chem. 2000, 612, 85. (k) Crespo, M.; Grande, C.; Klein,
A. J. Chem. Soc., Dalton Trans. 1999, 1629. (l) Zucca, A.; Doppiu, A.;
Cinellu, M. A.; Stoccoro, S.; Minghetti, G.; Manassero, M. Organometallics
2002, 21, 783. (m) Vicente, J.; Abad, J. A.; Rink, B.; Hernandez, F. S.;
Arellano, M. C. Organometallics 1997, 16, 5269. (n) Cardenas, D. J.;
Echavarren, A. M.; Ramirez de Arellano, M. C. Organometallics 1999,
18, 3337. (o) Vicente, J.; Saura-Llamas, I.; Cuadrado, J.; Ramirez de
Arellano, M. C. Organometallics 2003, 22, 5513. (p) Chakladar, S.; Paul,
P.; Mukherjee, A. K.; Dutta, S. K.; Nanda, K. K.; Podder, D.; Nag, K.
J. Chem. Soc., Dalton Trans. 1992, 3119. (q) Chakladar, S.; Paul, P.;
Venkatsubramanian, K.; Nag, K. J. Chem. Soc., Dalton Trans. 1991, 2669.
(r) Holland, A. W.; Bergman, R. G. Organometallics 2002, 21, 2149.
(3) (a) Ritleng, V.; Sirlin, C.; Pfeffer, M. Chem. ReV. 2002, 102, 1731.
(b) Moritani, I.; Fujiwara, Y. Tetrahedron Lett. 1967, 1119. (c) Murai, S.;
Kakiuchi, F.; Sekine, S.; Tanaka, Y.; Kamatani, A.; Sonoda, M.; Chatani,
N. Nature 1993, 366, 529. (d) Hull, K. L.; Sanford, M. S. J. Am. Chem.
Soc. 2007, 129, 11904.
(4) Hu, J.; Lin, R.; Yip, J. H. K.; Wong, K.-Y.; Ma, D.-L.; Vittal, J. J.
Organometallics 2007, 26, 6533.
10.1021/om800356k CCC: $40.75
2009 American Chemical Society
Publication on Web 01/27/2009

1094 Organometallics, Vol. 28, No. 4, 2009
Hu and Yip
Scheme 1
Scheme 2
7
reported methods. Pt(L)(OTf)2 were prepared in situ by reacting
Pt(L)Cl2 with 2 molar equiv of AgOTf in CH3CN/CH2Cl2 based
on modified literature methods.
to understand the regioselectivity of the double cyclometalation.
Two singly metalated complexes, [Pt(dppe)(PAn-H)]ClO4 (4)
(PAn ) 9-(diphenylphosphino)anthracene) and [Pt(dppm)(PAn-
PO-H)]OTf (5) (PAnPO is singly oxidized PAnP) (Scheme 2),
serve as models for the singly cyclometalated intermediate that
precedes the second metalation.
Physical Measurements. NMR experiments were performed on
a Bruker ACF 300, AMX500, or DRX500 spectrometer. All
chemical shifts (δ) are reported in ppm and coupling constants (J)
in Hz. 1H NMR chemical shifts are relative to TMS; the resonance
of residual protons of solvents was used as an internal standard.
31P{1H} NMR chemical shifts were relative to 85% H3PO4 in D2O.
Electrospray ionization mass spectra (ESI-MS) were measured on
a Finnigan MAT 731 LCQ spectrometer. Elemental analyses of
the complexes were carried out at the Elemental Analysis Labora-
tory at the National University of Singapore.
Experimental Section
General Methods. All syntheses were carried out in a N2
atmosphere unless stated otherwise. All the solvents used for
synthesis were purified according to the literature procedures. Triflic
acid was purchased from Sigma-Aldrich. 1a, 1b, 2a, 2b, 3a, 3b,
4,4 9,10-bis(diphenylphosphino)anthracene (PAnP),5 and 9-(diphe-
nylphosphino)anthracene (PAn)6 were prepared according to the
Synthesis of [Pt(dppm)(PAnPO-H)](OTf) (5). A CH2Cl2 solu-
tion (70 mL) of Pt(dppm)(OTf)2 (568 mg, 0.647 mmol) was added
slowly into a CH2Cl2 solution (100 mL) of excess PAnP (1.42 g,
(7) (a) Anderson, G. K.; Davies, J. A.; Schoeck, D. J. Inorg. Chim. Acta
1983, 76, L251. (b) Fallis, S.; Anderson, G. K.; Rath, N. P. Organometallics
1991, 10, 3180. (c) Stang, P. J.; Cao, D. H.; Saito, S.; Arif, A. M. J. Am.
Chem. Soc. 1995, 117, 6273.
(5) Yip, J. H. K.; Prabhavathy, J. Angew. Chem., Int. Ed. 2001, 40, 2159.
(6) Wesemann, J.; Jones, P. G.; Schomburg, D.; Heuer, L.; Schmutzler,
R. Chem. Ber. 1992, 125, 2187.

RegioselectiVe Double Cycloplatination
Organometallics, Vol. 28, No. 4, 2009 1095
Figure 1. 31P{1H} NMR (121.5 M, CD3CN) spectrum of the crude product of reaction 3 in CH2Cl2 (S: syn-isomer; A: anti-isomer). * 195Pt
#
satellites, minor unknown.
2.60 mmol). After stirring for 2 h, all solvent was removed by
rotoevaporation to give a yellow solid. (The following purifying
process was carried out in air.) The solid was extracted with CH3CN
to remove the unreacted PAnP, which was not soluble in CH3CN.
Removal of CH3CN by rotoevaporation gave a yellow solid. Slow
diffusion of Et2O into a CH2Cl2 solution of the solid gave a mixture
of needle- and block-shaped yellow crystals. The needle-shaped
crystals were carefully removed and discarded, and the remaining
block crystals were regrown. This process was repeated several
times to obtain pure compound 5. Single crystals suitable for X-ray
diffraction were grown by slow diffusion of Et2O into a CH3CN
solution of the purified complex. Anal. Calcd (%) for 5 · 2H2O,
C64H53F3O6P4PtS: C, 57.97; H, 4.03. Found (%): C, 57.85; H, 4.05.
1H NMR (500 MHz, CD3CN, δ/ppm): 8.78-8.77 (m, 1H, H5, An),
Figure 2. ORTEP diagram of 2a (50% thermal ellipsoids). H atoms
8.34-8.33 (m, 1H, H4, An), 7.86-7.85 (m, 1H, H8, An), 7.80-7.76
(m, 4H, Ph), 7.69-7.68 (m, 1H, H2, An), 7.67-7.63 (m, 4H, Ph),
7.56-7.54 (m, 8H, Ph), 7.46-7.45 (m, 10H, Ph), 7.36-7.29 (m,
6H, Ph), 7.19-7.17 (m, 2H, H6,7, An), 7.14-7.12 (m, 8H, Ph),
6.81-6.78 (m, 1H, H3, An), 4.85-4.81 (m, 2H, CH2). 31P{1H}
NMR (121.5 MHz, CD3CN, δ/ppm): 44.71 (ddd, P1), 31.23 (d, P4),
-21.99 (dd, P2), -29.81 (dd, P3); 1J(P1-Pt) ) 2807 Hz, 1J(P2-Pt)
) 2418 Hz, 1J(P3-Pt) ) 1424 Hz; 2J(P1-P2) ) 361 Hz, 2J(P1-P3)
) 12 Hz, J(P2-P3) ) 39 Hz, J(P1-P4) ) 4 Hz. ESI-MS(m/z):
1140.4 [M - OTf]+.
Protonolysis of Pt2. In a typical reaction, about 25 mg of the
Pt2 complexes was dissolved in ∼0.5 mL of CD2Cl2 in a NMR
tube to give a yellow solution. A slight excess of HOTf was then
added dropwise until a dark red mixture was obtained, and 1H and
31P{1H} NMR spectra of the mixture were recorded.
and solvent molecules are omitted for clarity.
Results and Discussion
Reactions and Characterizations of Products. In the
following discussion, the cyclometalation reactions of PAnP and
Pt(dppm)(OTf)2, Pt(dppe)(OTf)2, and Pt(dppp)(OTf)2 are re-
ferred to as reactions 1, 2, and 3, respectively. The structures
of 1a, 1b, 3a, 3b, and 4 were reported in our previous paper.4
The 31P{1H} NMR spectra of the products isolated from the
three reactions show two sets of first-order signals with some
of the peaks partly overlapped. A typical spectrum is shown in
Figure 1, which shows 31P{1H} NMR signals of the crude
product of the reaction 3.
The spectrum is dominated by two sets of intense signals
and small peaks of some minor unknown compounds. The
chemical shifts, JP-P and 1JPt-P, are listed in Table ST1. Of the
two sets of intense signals, each of them comprises three
mutually coupled double doublets with an intensity ratio of 1:1:
1. That all the P atoms coordinate to Pt atoms is evident from
X-ray Crystallography. The diffraction experiments were
carried out on a Bruker AXS SMART CCD 3-circle diffractometer
with a sealed tube at 23 °C using graphite-monochromated Mo KR
radiation (λ ) 0.71073 Å). The software used were SMART8a for
collecting frames of data, indexing reflections, and determination
of lattice parameters; SAINT8a for integration of intensity of
reflections and scaling; SADABS8b for empirical absorption cor-
rection; and SHELXTL8c for space group determination, structure
solution, and least-squares refinements on F2. The crystals were
mounted at the end of glass fibers and used for the diffraction
experiments. Anisotropic thermal parameters were refined for rest
of the non-hydrogen atoms. The H atoms were placed in their ideal
positions. A brief summary of crystal data and experimental details
are given in the Supporting Information.
1
the JPt-P satellites of the signals. On the other hand, the two
sets of signals do not couple with each other. This indicates
that the reactions produce two Pt complexes in which the metal
centers are coordinated to three different P atoms. The assign-
ments of the signals have been reported.5 In summary, the
signals that show large 2JP-P coupling constants (343-361 Hz)
are assigned to P1 and P2, while the ones with small 2JP-P (4-39
Hz) are attributed to P3 (see Scheme 1 for the labels of the P

1096 Organometallics, Vol. 28, No. 4, 2009
Hu and Yip
Figure 4. ORTEP diagrams of the cation of 5 (50% thermal
ellipsoids). H atoms, solvent, and the anion are omitted for clarity.
Table 3. Selected Bond Lengths (Å) and Angles (deg) for 5 · 0.5H2O
Figure 3. ORTEP diagram of 2b (50% thermal ellipsoids). H atoms
and solvent molecules are omitted for clarity.
Pt(1)-P(1) 2.255(18)
Pt(1)-P(2) 2.316(2)
Pt(1)-P(3) 2.330(2)
Pt(1)-C(1) 2.052(7)
P(2)-O(1) 1.473(6)
P(1)-Pt(1)-P(2) 174.68(6)
P(1)-Pt(1)-P(3) 103.82(7)
P(2)-Pt(1)-P(3) 71.00(7)
P(1)-Pt(1)-C(1) 83.9(2)
P(2)-Pt(1)-C(1) 101.3(2)
P(3)-Pt(1)-C(1) 172.2(2)
Table 1. Selected Bond Lengths (Å) and Angles (deg) for
2a · 1.25(CH2Cl2)0.5(Et2O)
Pt(1)-P(1) 2.281(3)
Pt(1)-P(2) 2.328(4)
Pt(1)-P(3) 2.313(4)
Pt(1)-C(1) 2.106(1)
P(1)-Pt(1)-P(2) 176.19(11)
P(1)-Pt(1)-P(3) 99.47(12)
P(2)-Pt(1)-P(3) 83.83(13)
P(1)-Pt(1)-C(1) 82.4(3)
P(2)-Pt(1)-C(1) 94.4(3)
Structures. The geometrical isomers can be separated by
fractional crystallization. The X-ray crystal structures of 1a, 3a,
and 3b have been previously reported by us. Figures 2 and 3
show the structures of syn-[Pt2(dppe)2(PAnP-H2)](OTf)2 ·
1.25CH2Cl2 · 0.5Et2O (2a · 1.25CH2Cl2 · 0.5Et2O) and anti-
[Pt2(dppe)2(PAnP-H2)](OTf)2 · 2CH2Cl2 (2b · 2CH2Cl2), which
are representatives for the syn- and anti-isomers. Selected bond
distances and angles are listed in Tables 1 and 2. The structure
of 2a shows two syn-oriented Pt centers attached to the C1 and
C4 positions of the anthracenyl ring, while the C1 and the C5
positions are metalated in the syn-isomers 2b. The syn- and anti-
isomers display approximate C2V and C2h symmetry, respectively.
The Pt-P and Pt-C bond lengths are typical for cyclometalated
Pt-phosphine complexes.4 The anthracenyl rings of the com-
plexes are nearly planar. The metal center is the center of two
five-membered chelated rings that comprise the P atoms of dppe
and PAnP, as well as the substituted C atom. The bite angles
P2-Pt1-P3 for the complexes (2a 83.83(13)°, 2b 82.99(6)°)
are significantly smaller than that of 3a (87.49(5)°) and 3b
(86.88(5)°) but larger than that of 1a (70.74(6)°).
Table 2. Selected Bond Lengths (Å) and Angles (deg) for
2b · 2CH2Cl2
Pt(1)-P(1) 2.2816(15)
Pt(1)-P(2) 2.3175(16)
Pt(1)-P(3) 2.3234(16)
Pt(1)-C(1) 2.073(5)
P(1)-Pt(1)-P(2) 176.90(5)
P(1)-Pt(1)-P(3) 99.67(6)
P(2)-Pt(1)-P(3) 82.99(6)
P(1)-Pt(1)-C(1) 82.54(15)
P(2)-Pt(1)-C(1) 94.82(15)
P(3)-Pt(1)-C(1) 177.76(15)
atoms).9 The two sets of signals are attributed to the geometrical
isomers syn- and anti-[Pt2(L)2(PAnP-H2)](OTf)2. The assignment
is confirmed by the spectra of pure samples of the complexes.
The most important message obtained from the spectra is that
two geometrical isomers are produced in the reactions and the
double cyclometalation is clearly regioselective, as the yields
of the isomers are different in all three reactions.
Reacting Pt(dppe)(ClO4)2 with 1 molar equiv of 9-(diphe-
nylphosphino)anthracene (PAn) gives the mononuclear cyclo-
metalated complex [Pt(dppe)(PAn-H)](ClO4) (4), which has
been fully characterized. The reaction of Pt(dppm)(OTf)2 and
excess PAnP gives the singly cyclometalated [Pt(dppm)(PAnP-
H)](OTf) together with the doubly cyclometalated complexes
(1a and 1b). When exposed to air, the uncoordinated P atom in
[Pt(dppm)(PAnP-H)](OTf) is oxidized and the complex Pt-
(dppm)(PAnPO-H)](OTf) (5) is formed. The 31P{1H} NMR
spectrum of 5 consists of four sets of signals (P1-P4) of equal
intensity. The chemical shifts and signal patterns of the three P
atoms (P1, P2, and P3) coordinating to the Pt ion are similar to
those of 1a, except that P1 is further coupled to the oxidized
P4 with a very small coupling constant of ∼4 Hz. Because of
its long-range coupling to P1, P4 appears as a doublet at δ 31.
Complex 4 is a mononuclear analogue of 2a and 2b. It
consists of a Pt ion attached to the anthracenyl ring at the C1
position.4 The coordination geometry and Pt-C and Pt-P bond
lengths are similar to those of 2a and 2b.
The structure of 5 is shown in Figure 4 (Table 3 for selected
bond distances and angles). One of two P atoms (P1) on PAnP
is coordinated to a Pt ion, which is attached to the C1 position
of anthracene. The coordination geometry of the Pt center and
the Pt-C and Pt-P bond lengths are similar to those of 1a.
The other P atom on PAnP is oxidized to phosphine oxide. The
PdO bond distance (1.473(6) Å) is normal. Hydrogen bonding
between the PdO and the anthracenyl proton at C5 is possible
in view of the C-O and H---O distances of 2.842 and 2.101 Å,
respectively, and the C-H-O angle of 135.72°.
(8) (a) SMART & SAINT Software Reference Manuals, version 4.0;
Siemens Energy & Automation, Inc., Analytical Instrumentation: Madison,
WI, 1996. (b) Sheldrick, G. M. SADABS: A Software for Empirical
Absorption Correction; University of Gottingen: Gottingen, Germany, 1996.
(c) SHELXTL Reference Manual, version 5.03; Siemens Energy & Automa-
tion, Inc., Analytical Instrumentation: Madison, WI, 1996.
(9) (a) Allen, F. H.; Pidcock, A. J. Chem. Soc. A 1968, 2700. (b)
Appleton, T. G.; Clark, H. C.; Manzer, L. E. Coord. Chem. ReV. 1973, 10,
335.
Regioselectivity of the Double Cyclometalations. The
31P{1H} NMR studies clearly show that the double cyclometa-
lation of PAnP by Pt(L)(OTf)2 is regioselective, as the intensities
of the signals of the geometrical isomers are different. The
relative intensities of the signals obtained from direct integration
are essentially the same as those obtained from the inverse-
1
gated H-decoupling spectra and are used to obtain the ratios

RegioselectiVe Double Cycloplatination
Organometallics, Vol. 28, No. 4, 2009 1097
Figure 5. 31P{1H} spectra (CD2Cl2, 121.5 MHz) of products from reaction 1 in 100% CH2Cl2, showing the change of isomer distribution
with reaction time. Peaks labeled as (S) are due to the syn-isomer 3a and (A) the anti-isomer 3b. * 195Pt satellites.
Figure 6. Change of the regioselectivity S of reaction 1 in CH2Cl2/
CH3CN solvent mixtures [CH2Cl2:CH3CN v/v: 25:75 (∆), 50:50
(0), 75:25 (O), 100:0 (b)] with time at room temperature.
Figure 8. Change of the regioselectivity S of the double cyclom-
etalation with time in CH2Cl2 at room temperature. Reaction 1:
solid line. Reaction 2: dashed line. Reaction 3: dotted line.
Table 4. Si and Sf for Reactions 1 and 3
reaction 1
reaction 3
CH3CN % (v/v)
Si
Sf
Si
Sf
0
89
94
96
96
51
56
59
61
78
89
90
85
44
50
56
60
25
50
75
It was discovered that S changes with time as the two
geometrical isomers undergo slow isomerization at room
temperature. Typically an equilibrium is reached after 3-4 h.
Figure 5 shows the 31P{1H} NMR spectral changes for reaction
1 in neat CH2Cl2. All three reactions exhibit an increase in the
minor product, the anti-isomer, at the expense of the major
product, the syn-isomer. It is clear that the cyclometalation is
reVersible (vide infra).
Since the isomerization is catalyzed by the protons released
from the metalation of the C-H bond (vide infra), the S at
different reaction times can be determined by quenching
isomerization by addition of excess triethylamine. Figures 6 and
7 show respectively the variation of S of the reactions 1 and 3
in CH3CN/CH2Cl2 mixtures (v/v ) 0:100, 25:75, 50:50, and
75:25), and Figure 8 compares the S of the three reactions in
neat CH2Cl2.
Figure 7. Change of the regioselectivity S of reaction 3 in CH2Cl2/
CH3CN solvent mixtures [CH2Cl2:CH3CN v/v: 25:75 (∆), 50:50
(0), 75:25 (O), 100:0 (b)] with time at room temperature.
of the isomers. The regioselectivity is defined as S ) [syn]/
([syn] + [anti]) × 100%, where [syn] and [anti] are the
integrated intensities of the 31P{1H} NMR signals of the isomers,
which are proportional to the concentrations of the complexes.

1098 Organometallics, Vol. 28, No. 4, 2009
Hu and Yip
Scheme 3
Table 5. 31P{1H} NMR (in CD2Cl2, 121.5 MHz) Data for I (n ) 3)
Protonolysis. It has been demonstrated that the Pt-C bonds
1J(Pt-P) (Hz)
2J(P-P) (Hz)
of some organoplatinum complexes can be cleavage by pro-
tonation (protonolysis).10 According to the mechanistic proposal
and on the basis of the reversibility of the cyclometalation, it is
expected that the Pt-C bonds of the Pt2 complexes can be
cleaved by acid. Addition of a slight excess of HOTf (∼1.2-
fold) to solutions of 3a and 3b changes the 31P{1H} spectra of
the compounds. Notably, both resulting spectra (SF1) are the
same, consisting of two sets of signals of equal intensity that
do not couple with each other. Each set of signal is composed
of three mutually coupled P atoms in 1:1:1 ratio, and all signals
δ (ppm)
P1
P2
P3
P1A
P2A
P3A
48.17
8.44
-8.05
27.83
-6.54
2.54
2649
2335
1780
2335
2266
3368
P1-P2 ) 340
P2-P3 ) 34
P1-P3 ) 15
P1A-P2A ) 332
a
a
2
2JP2A-3A and JP1A-P3A are unresolved.
Table 4 shows the regioselectivity for reactions 1 and 3 at a
short reaction time (∼5 min) as well as at sufficiently long
reaction time (1-2 weeks). The former is defined as the initial
regioselectivity (Si), and the latter is defined as the final
regioselectivity (Sf).
1
are split by JPt-P coupling. One set of the signals, labeled P1,
1
P2, and P3, has chemical shifts and JPt-P close to those of 3a
and 3b, whereas the other set of signals, labeled P1A, P2A,
and P3A, does not (Table 5). It is believed that the intermediate
I shown in Scheme 3 is formed in the protonolysis. The signals
that have chemical shifts and coupling constants very similar
to those of 3a and 3b are assigned to P1, P2, and P3, which
coordinate to the cyclometalated Pt ion. The other set of signals
is ascribed to P1A, P2A, and P3A bonded to the dangling Pt
ion. It is noted that the 1JPt-P3A (3368 Hz) is exceptionally large,
indicating that P3 is trans to a weak ligand, and accordingly it
is assigned to the P atom opposite the OTf- ion. The 1H NMR
signals of the complex are very broad and cannot be resolved
at either 300 or 193 K. Similar spectra are obtained from the
protonolysis of 1a,b and 2a,b.
All three reactions show a decrease in S in the first 5 h until
an equilibrium is attained after a few days. The Sf is lower than
the corresponding Si. For reaction 1, the concentrations of syn-
and anti-isomers are nearly the same (Sf ) 51%) in CH2Cl2
after 4 days. For reaction 2 in CH2Cl2, 2a remains the major
product even though S decreases from 84% (Si) to 73% (Sf) in
3 days. Reaction 3 in CH2Cl2 reaches thermal equilibrium in
24 h, much faster than the other two reactions, and unlike
reactions 1 and 2, the major product is the anti-isomer (Sf )
44%). But increasing CH3CN to 75% favors the formation of
the syn-isomer as Sf is increased to 60%. In fact, for all three
reactions, Sf increases with the percentage of CH3CN in the
solvent mixtures.
Scheme 3. is a proposed mechanism for the second cyclo-
metalation and isomerization. The second cyclometalation
should start with the coordination of the second Pt ion to the
remaining free PPh2 group of PAnP, forming an intermediate
I. Subsequent attack of the second Pt ion at C4-H4 and C5-H5
leads to the syn- and anti-isomers, respectively. The formation
of a Pt-C bond is accompanied with the release of a proton.
The isomerization begins with the protonolysis of the Pt-C bond
in the doubly cyclometalated complexes. The isomerization is
acid-catalyzed, and it can be quenched by addition of excess
NEt3 as mentioned before.
Interestingly, removal of HOTf by adding excess Et3N to the
protonolyzed mixtures leads to disappearance of the signals of
the intermediate and appearance of the signals of the doubly
cyclometalated complexes with distribution of the isomers being
(10) (a) Romeo, R.; D’Amico, G. Organometallics 2006, 25, 3435. (b)
Belluco, U.; Giustiniani, M.; Graziani, M. J. Am. Chem. Soc. 1967, 89,
6494. (c) De Felice, V.; De Renzi, A.; Panunzi, A.; Tesauro, D. J.
Organomet. Chem. 1995, 488, C13-C14. (d) Hill, G. S.; Rendina, L. M.;
Puddephatt, R. J. Organometallics 1995, 14, 4966. (e) Stahl, S. S.; Labinger,
J. A.; Bercaw, J. E. J. Am. Chem. Soc. 1995, 117, 9371. (f) Ca´mpora, J.;
Lo´pez, J. A.; Palma, P.; Valerga, P.; Spillner, E.; Carmona, E. Angew.
Chem., Int. Ed. 1999, 38, 147. (g) Lersh, M.; Tilset, M. Chem. ReV. 2005,
105, 2471.

RegioselectiVe Double Cycloplatination
Organometallics, Vol. 28, No. 4, 2009 1099
Scheme 4
Scheme 5
close to the Sf. This indicates that acid can speed up the
isomerization and lends further support to the proposal that the
intermediate I is formed in the protonolysis.
distribution, and G4 and G5 are free energy of I4 and I5,
respectively).12
[syn]i [I4] k1k
-2 ) exp
-(G4 - G5)
Mechanistic Postulate. As the cyclometalation is reversible,
the final regioselectivtiy Sf should be the thermodynamic product
distribution.11 The Sf for reaction 1 in CH2Cl2 is 51%, indicating
that the two isomers 1a and 1b are nearly isoenergetic. On the
other hand, the Sf for reaction 3 in CH2Cl2 is 44%, meaning 3b
is slightly more stable than 3a, by ∼0.6 kJ mol-1. These are
very small free energy differences and could be caused by small
differences in solvation energy or steric congestion between the
isomers. A general observation is that the syn-isomer is more
favored as the concentration of CH3CN in the solvent mixture
increases. It can be explained by the fact that the syn-isomer
has a dipole moment larger than that of the anti-isomer, which
has a center of inversion, and therefore can be better solvated
in more polar solvent.
The Si approximates the kinetically controlled product
distribution. It ranges from 78% to 96%, indicating that the syn-
isomers are the major kinetic products, and they are formed
faster than the corresponding anti-isomers. It further implies
that the attack of the second Pt ion to the C4-H4 bond is faster
than to the C5-H5 bond. A reaction scheme is proposed to
explain the regioselectivity (Scheme 4). Many studies on the
mechanism of C-H bond activation by Pt(II) ion support the
formation of an intermediate preceding the formation of
the Pt-C bond and deprotonation.1d,10g Accordingly, the attacks
of the Pt ion at the C4-H4 and C5-H5 bonds of the ring should
lead to intermediates I4 and I5, respectively.
≈
)
(1)
(
)
[anti]i [I5] k-1k2
RT
That the syn-isomer is the major kinetic product implies I4 is
more stable than I5. Assuming the steps I f I4 and I f I5
involve similar transition states, the stability of I4 and I5 should
correlate with the exothermicity of the two steps. In other words,
the C4-H4 bond in I should be more reactive than the C5-H5
bond. If the C-H activation goes through an electrophilic
pathway, then one would expect the C4-H4 bond to be more
electron-rich than the C5-H5 bond. On the other hand, if the
reaction goes through a nucleophilic pathway, the C4-H4 bond
should be more electron-deficient than the C5-H5 bond. It does
not necessarily mean that the two positions must be greatly
different in terms of electron density. In fact, the Si of 96%
observed for reaction 1 in 25:75 CH2Cl2/CH3CN can be
accounted by an energy difference of only ∼8 kJ mol-1 between
the two intermediates.
The electronic effect of the first Pt ion clearly directs the
second attack. Unfortunately, the 1H NMR spectra of the
intermediate I formed by protonolysis cannot be resolved.
Nonetheless, the mononuclear complexes 4 and 5 could provide
some clues about the different reactivity of the two C-H bonds
in I. For both complexes, H4 is more upfield than H5: the two
protons are different by 0.24 ppm (4) and 0.44 ppm (5).
Similarly, the HMQC spectra of the complexes show that C4
is more upfield than C5 by ∼3 ppm in both complexes (SF2
and 3). These suggest that the C4-H4 bond is more electron-
rich than the C5-H5 bond. Accordingly, an electrophilic
pathway is more consistent with the preferential attack of the
second Pt ion at the C4-H4 bond. I4 and I5 could be
As both C-H bond activation and deprotonation are revers-
ible, the regioselectivity can be affected by the rates of
deprotonation as well as the proportions of I4 and I5. However,
it is reasonable to assume that k1′ and k2′ should be similar.
Accordingly, the kinetic product distribution should be mainly
determined by the ratio of I4 and I5, which can be estimated by
eq 1 ([syn]i/[anti]i is the ratio of the isomers in kinetic
(12) (a) Curtin, D. Y. Rec. Chem. Prog. 1954, 15, 111. (b) Hammett,
L. P. Physical Organic Chemistry; McGraw-Hill: New York, 1970; Chapter
5. (c) Seeman, J. I. J. Chem. Educ. 1986, 63, 42.
(11) Berson, J. A. Angew. Chem., Int. Ed. 2006, 45, 4727.

1100 Organometallics, Vol. 28, No. 4, 2009
Hu and Yip
σ-complexes (Scheme 5, left), which have the Pt ion bonded to
either the carbon atom (Wheland intermediate)1c,13 or the C-H
bond, forming an agnostic complex (Scheme 5, right).14 For
both cases, the formation of I4 is favored over that of I5 by the
π-back-donation from the first Pt ion.
picture of the seemingly simple reaction. The regioselectivity
can be influenced by the solvent and the ancillary diphosphine
ligands. The syn-isomers are the kinetic products, but the
reaction is thermodynamically controlled, as the isomers can
undergo acid-catalyzed isomerization. The proposed mechanistic
model invokes a preferential electrophilic attack of the second
Pt ion at the C4-H4 and C5-H5 bonds of the anthracenyl ring,
producing two intermediates, I4 and I5. Subsequently deproto-
nations lead to the syn- and anti-isomers, respectively.
Conclusion
This study demonstrated the double cyclometalation of PAnP
is regioselective and revealed a rather complicated mechanistic
(13) (a) Sweet, J. R.; Graham, W. A. G. Organometallics 1983, 2, 135.
(b) Albrecht, M.; Dani, P.; Lutz, M.; Spek, A. L.; van Koten, G. J. Am.
Chem. Soc. 2000, 122, 11822. (c) Davies, D. L.; Donald, S. M. A.;
Macgregor, S. A. J. Am. Chem. Soc. 2005, 127, 13754. (d) Shilov, A. E.;
Shul’pin, G. B. Chem. ReV. 1997, 97, 2879. (e) Singh, A.; Sharp, P. R.
J. Am. Chem. Soc. 2006, 128, 5998.
(14) (a) Vigalok, A.; Uzan, O.; Shimon, L. J. W.; Ben-David, Y.; Martin,
J. M. L.; Milstein, D. J. Am. Chem. Soc. 1998, 120, 12539. (b) Johansson,
L.; Tilset, M.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2000, 122,
10846. (c) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc.
1996, 118, 5961. (d) Heiberg, H.; Johansson, L.; Gropen, O.; Ryan, O. B.;
Swang, O.; Tilset, M. J. Am. Chem. Soc. 2000, 122, 10831. (e) Kua, J.;
Xu, X.; Periana, R. A.; Goddard, W. A., III. Organometallics 2003, 21,
551. (f) Thomas, J. C.; Peters, J. C. J. Am. Chem. Soc. 2001, 123, 5100.
(g) Tellers, D. M.; Yung, C. M.; Arndtsen, B. A.; Adamson, D. R.; Bergman,
R. G. J. Am. Chem. Soc. 2002, 124, 1400.
Acknowledgment. The authors would like to thank
National University of Singapore and Ministry of Education
Singapore (R-143-000-331-112) for financial support and
Miss Tan Geok Kheng for determining the X-ray crystal
structures.
Supporting Information Available: Crystallographic data, CIF
files of 2a, 2b, and 5, HMQC spectra of 4 and 5, and 31P{1H} NMR
spectrum and data for I. This material is available free of charge
via the Internet at http://pubs.acs.org.
OM800356K