Formation of η2 C-H agostic rhodium arene complexes and their relevance to electrophilic bond activation

Article abstract of DOI:10.1021/ja982534u

Reaction of the ligand 1,3-bis((di-tert-butylphosphino)methyl)benzene (1a) with the [RhCO]⁺ fragment in THF resulted in clean formation of the crystallographically characterized his-chelated complex 2a which contains an η² agostic Rh

Full text of DOI:10.1021/ja982534u

J. Am. Chem. Soc. 1998, 120, 12539-12544
12539
Formation of η2 C-H Agostic Rhodium Arene Complexes and Their
Relevance to Electrophilic Bond Activation
Arkadi Vigalok,† Olivier Uzan,† Linda J. W. Shimon,‡ Yehoshoa Ben-David,†
Jan M. L. Martin,*,† and David Milstein*,†
Contribution from the Departments of Organic Chemistry and Chemical SerVices, The Weizmann Institute
of Science, RehoVot 76100, Israel
ReceiVed July 17, 1998
Abstract: Reaction of the ligand 1,3-bis((di-tert-butylphosphino)methyl)benzene (1a) with the [RhCO]+
fragment in THF resulted in clean formation of the crystallographically characterized bis-chelated complex 2a
which contains an η2 agostic Rh C-H bond. Both the NMR data and the X-ray crystal structure show strong
interaction between the metal center and the agostic C-H bond, which results in high acidity of the agostic
proton. Reaction of 2a with a weak organic base (NEt3 or collidine) affords the known cyclometalated complex
3. Reaction of the new ligand 1,3-bis((di-tert-butylphosphino)methyl)-4,5,6-trimethoxybenzene (1b) with the
[RhCO]+ fragment in THF gives the analogous to 2a agostic complex 2b. Analysis of the NMR data and the
reactivity of both 2a and 2b showed that there is very little, if any, contribution of a metal arenium structure.
This interpretation is supported by B3LYP/LANL2DZ density functional calculations on model compounds.
Thus, deprotonation of η2 aromatic C-H agostic complexes can be proposed as an alternative route to
electrophilic metalation of aromatic compounds.
Introduction
Scheme 1
Aromatic C-H activation by metal complexes is an area of
much current interest.1 One of the generally accepted mecha-
nisms for metal-assisted inter- or intramolecular carbon-
hydrogen bond cleavage involves a concerted metal insertion
into this bond (oxidative addition) preceded by the formation
of an η2 metal arene complex.2 Another general mechanism is
believed to proceed via an electrophilic attack at the ipso-carbon
to form a metal arenium (Wheland) complex followed by proton
elimination1d,3,4 (Scheme 1). Surprisingly, despite this general
belief, no such complex has ever been isolated in a reaction of
a metal complex with an aromatic compound. Moreover, the
kinetics of electrophilic metalation of arenes by transition metal
complexes is not always reminiscent of aromatic electrophilic
substitution. For example, the F values for this reaction are
often substantially lower than those observed in organic SEAr
reactions and the regioselectivity with substituted aromatics does
not always correspond to the expected pattern.1a
such complexes are exceedingly rare.6 As no chemical proper-
ties have been reported, it is unclear whether this type of
interaction results in increased acidity of the agostically bound
proton, as has been observed in aliphatic agostic complexes.7
If so, the increased acidity might lead to facile deprotonation
of the arene to give the metalated product without actual metal
insertion into the C-H bond, i.e., a process similar to electro-
philic metalation.8 Here we demonstrate that, using an ap-
It had been suggested that in intramolecular processes the
aromatic C-H bond activation might occur via an η2 C-H
agostic intermediate or transition state,5 although examples of
† Department of Organic Chemistry.
‡ Department of Chemical Services.
(1) For recent reviews see: (a) Shilov, A. E.; Shul’pin, G. B. Chem.
ReV. 1997, 97, 2879. (b) Crabtree, R. H. Chem. ReV. 1985, 85, 245. (c)
Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H. Acc. Chem.
Res. 1995, 28, 154. (d) Taylor, R. Electrophilic Aromatic Substitution; John
Wiley & Sons: New York, 1990; Chapter 5, p 173. (e) Sen, A. Acc. Chem.
Res. 1988, 21, 421.
(2) (a) Jones, W. D.; Feher, F. J. J. Am. Chem. Soc. 1982, 104, 4240.
(b) Jones, W. D.; Feher, F. J. Acc. Chem. Res. 1989, 22, 91.
(3) For detailed studies on late transition metal complexes, see: (a)
Aoyama, Y.; Yoshida, T.; Sakurai, K.; Ogoshi, H. Organometallics 1986,
5, 168. (b) Shul’pin, G. B.; Nizova, G. V.; Nikitaev, A. T. J. Organomet.
Chem. 1984, 276, 115. (c) Sweet, J. R., Graham, W. A. G. J. Am. Chem.
Soc. 1983, 105, 305.
(5) Lavin, M.; Holt, E. M.; Crabtree, R. H. Organometallics 1989, 8,
99.
(6) (a) Dani, P.; Karlen, T.; Gossage, R. A.; Smeets, W. J. J.; Spek, A.
L.; van Koten, G. J. Am. Chem. Soc. 1997, 119, 11317. (b) Albeniz, A. C.;
Schulte, G.; Crabtree, R. H. Organometallics 1992, 11, 242.
(7) Agostic metal complexes are normally viewed as intermediates (or
transition states) toward C-H oxidative addition. For reviews, see: (a) Hall,
C.; Perutz, R. N. Chem. ReV. 1996, 96, 3125. (b) Brookhart, M.; Green, M.
L. H.; Wong, L.-L. Prog. Inorg. Chem. 1988, 36, 1. (c) Crabtree, R. H.;
Hamilton, D. G. AdV. Organomet. Chem. 1988, 28, 299.
(8) Agostic interactions that do not lead to an increase in acidity and
subsequent deprotonation of aromatic C-H bonds were reported: Perera,
S. D.; Shaw, B. L. J. Chem. Soc., Dalton Trans. 1995, 3861.
(4) For a review on cyclometalation, see: Ryabov, A. D. Chem. ReV.
1990, 90, 403.
10.1021/ja982534u CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/17/1998

12540 J. Am. Chem. Soc., Vol. 120, No. 48, 1998
Vigalok et al.
Table 1. Selected Bond Lengths (Å) and Angles (deg) for 2a
(a) Bond Lengths
propriate metal precursor and ligand system, it is possible to
isolate the agostic tntermediate for the aromatic C-H cyclo-
metalation process,7 which can be easily deprotonated. Sig-
nificantly, very little, if any, contribution from a metal arenium
structure is observed. This provides an attractive alternative to
the electrophilic aromatic substitution mechanism, which does
not require a metal-stabilized arenium intermediate. Evidence
for aromatic metalation by scandium complexes via a σ-bond
metathesis mechanism without participation of the aromatic
π-system was presented.9
Rh(1)-C(1)
Rh(1)-C(11)
Rh(1)-H(11)
Rh(1)-P(2)
Rh(1)-P(3)
C(1)-O(1)
1.820 (5)
2.273 (5)
1.950
2.330 (2)
2.325 (2)
1.146 (6)
C(11)-C(12)
C(12)-C(13)
C(13)-C(14)
C(14)-C(15)
C(15)-C(16)
C(11)-C(16)
1.408 (8)
1.398 (7)
1.384 (8)
1.385 (8)
1.403 (7)
1.386 (9)
(b) Angles
P(2)-Rh(1)-P(3)
P(2)-Rh(1)-C(1)
P(2)-Rh(1)-C(11)
167.32 (5) Rh(1)-H(11)-C(11) 97.57
96.2 (2) Rh(1)-C(11)-H(11) 58.28
83.9 (2) C(12)-C(11)-C(16) 123.9 (5)
H(11)-Rh(1)-C(11) 24.15
C(12)-C(11)-Rh(1) 109.1 (3)
Results and Discussion
While studying the mechanisms of C-H and C-C bond
activation in PCP-type ligands with various metals, we examined
the reactivity of unsaturated, relatively electron-poor rhodium-
(I) precursors. Surprisingly, although various platinum group
metals readily cyclometalate ligand 1a, as first shown by Shaw
and co-workers,10 no C-H bond activation was observed when
1a was treated with an equimolar amount of [Rh(C2H4)(CO)-
(solv)n]+ (n ) 1, 2)11 in THF. Rather, the reaction smoothly
produced complex 2a which contains an η2 C-H bond (eq 1).
Complex 2a is an air-stable compound and has been fully
spectroscopically characterized including a single-crystal X-ray
analysis.
Figure 1. (a) ORTEP view of a molecule of 2a with the thermal
ellipsoids at 50% probability. The hydrogen atoms (except H(11)) are
omitted for clarity. (b) Molecular view of a molecule of 2a showing
the mutual orientation of the aromatic ring and the metal center.
represents a large upfield shift relative to the ipso-carbon of
cyclometalated PCP Rh complexes. Noteworthy, there is no
observable coupling between the Rh atom and the ipso-C, which
can be as large as 30 Hz in similar Rh(I) PCP systems,12
indicating that the interaction between the two atoms is weak.
1
We also note the relatively low JHC coupling constant of 123
Hz, as compared to 163 and 165 Hz for the other two aromatic
C-H bonds.13
Orange plates of 2a suitable for X-ray analysis were obtained
by slow crystallization of a CH2Cl2 solution of 2a at room
temperature. Selected bond distances and bond angles are given
in Table 1. The structure of 2a clearly demonstrates (Figure
1) a strong metal interaction with the ipso-C-H bond, while
the aromaticity of the ring remains practically intact. In good
agreement with the 13C NMR data, the distance between the
ipso-carbon and the Rh atom is extremely long (2.273(5) Å)
relative to regular rhodium-carbon σ-bonds and it is even longer
than that in closely related Rh olefin complexes.14 The
rhodium-ipso-carbon bond length in neutral aromatic PCP
complexes is normally substantially shorter (cf. 1.999(7) Å in
ClRh(H)[C6H3-1,3-(CH2P(t-Bu)2)2]15 and 2.02(2) Å in ClRh-
(CH3)[C6H(CH3)2(CH2P(t-Bu)2)2],12a respectively). The carbon-
carbon distances inside the ring are equal within the experi-
mental error and are of values expected for aromatic compounds.
The 31P NMR spectrum of complex 2a exhibits a doublet at
34.90 ppm (JRhP ) 100.4 Hz). An important feature of the 1H
NMR spectrum of 2a is the high-field shift of the proton attached
to the ipso-carbon giving rise to a doublet at 4.13 ppm.
However, the large value of the 2JRhH coupling constant of 18.1
Hz suggests some sort of “agostic” interaction between the metal
and this proton (vide infra). In the 13C NMR spectrum of 2a
the ipso-carbon atom exhibits a signal at 110.95 ppm, which
(9) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan,
M. C.; Santasiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am. Chem. Soc.
1987, 109, 203.
(12) (a) Rybtchinski, B.; Vigalok, A.; Ben-David, Y.; Milstein, D. J.
Am. Chem. Soc. 1996, 118, 12406. (b) Vigalok, A.; Ben-David, Y.; Milstein,
D. Organometallics 1996, 15, 1838.
(10) Moulton, C. J.; Shaw, B. L. J. Chem. Soc., Dalton. Trans. 1976,
1020.
(13) Lowering of 1JCH is often used as an indication for the presence of
agostic interactions. See refs 7b,c for more details.
(14) (a) Vigalok, A.; Shimon, L. J. W.; Milstein, D. J. Am. Chem. Soc.
1998, 120, 477. (b) Vigalok, A.; Milstein, D. J. Am. Chem. Soc. 1997, 119,
7873.
(11) The complex [Rh(C2H4)(CO)(solv)n]+ (n ) 1, 2) was prepared in
situ according to a general procedure reported in: (a) Schrock, R. R.;
Osborn, J. A. J. Am. Chem. Soc. 1971, 93, 3089. The source was the known
[Rh(C2H4)(CO)Cl]2: (b) Powell, J.; Shaw, B. L. J. Chem. Soc. A 1968,
211.
(15) Nemeh, S.; Jensen, C.; Binamira-Soriaga, E.; Kaska, W. C.
Organometallics 1983, 2, 1442.

η2 C-H Agostic Rhodium Arene Complexes
J. Am. Chem. Soc., Vol. 120, No. 48, 1998 12541
Table 2. Comparative Analysis of Selected NMR Data (ppm) for
Complexes 2a and 2b and for Parent Ligands 1a and 1b,
Respectively (CDCl3, 23 °C)
Scheme 2
ligand vs complex
NMR
signal
δ(1a)
δ(2a)
∆δ
δ(1b)
δ(2b)
∆δ
para-C
128.10 138.34
10.24 149.42 159.98
1.02 146.09 148.90
12.93 129.23 144.57
10.56
2.81
15.34
meta-C 126.60 127.62
ortho-C 141.24 154.17
ipso-C
ipso-H
130.63 110.95 -19.68 127.15 104.77 -22.38
7.25 4.13 -3.12 7.11 4.13 -2.98
temperature, rapid formation of the cyclometalated hydrido
chloride Rh(III) complex 4 took place in a quantitative yield
within a few seconds (eq 3). By contrast, reaction with [Rh-
The H(11) atom (which was located independently) strongly
interacts with the Rh(1) (1.950 Å), which was also evident from
the 1H NMR spectrum of 2a. It is noteworthy that this hydrogen
is remoVed from the aromatic plane by ca. 17°, whereas the
C-H bond length of 0.93 Å is in the range observed in X-ray
diffraction measurements for unactivated hydrocarbons.16 The
calculated rbp value of 0.73 is indicative of a very strong η2
C-H agostic interaction in 2a.16
Although the X-ray structure of 2a does not show noticeable
distortion from aromaticity in the ring, the removal of the ipso-
C-H bond from the plane suggests that the contribution of the
arenium (Wheland type) form cannot be excluded. The agostic
proton in 2a is somewhat acidic, and it undergoes slow exchange
with excess of D2O (∼30% conversion after 18 h at room
temperature in THF). Moreover, the reaction with relatively
weak bases such as NEt3 or collidine results in deprotonation
of 2a and formation of the known Rh(I) carbonyl complex Rh-
(CO)[C6H3-1,3-(CH2P(t-Bu)2)2]10 (3). The reaction is reversed
upon addition of HOTf (eq 2). While deprotonation of aliphatic
(3)
(C2H4)(CO)(solv)n]+ (n ) 1, 2) in THF smoothly produced the
agostic complex 2b (eq 1). Significantly, only very small
changes in the NMR spectra of 2b were noted in comparison
to those of 2a. Similar to the case of 2a, the agostic proton
bound to the ipso-carbon in 2b shows a doublet at 4.13 ppm
(JRhH ) 17.6 Hz) in the 1H NMR spectrum and the ipso-carbon
lacks the 1JRhC coupling and appears at 104.77 ppm as a triplet
(JPC ) 4.4 Hz) in the 13C NMR spectrum. Table 2 compares
selected 13C chemical shifts of complexes 2a and 2b to those
of the parent ligands 1a and 1b, respectively. Interestingly, the
∆δ between the complex and ligand signals is nearly identical
for the two complexes, indicating that contribution of the
arenium structure in 2a and 2b is similar and not large as
compared with the neutral C-H aromatic agostic structure.
Moreover, the H/D exchange of the agostic protons in both 2a
and 2b with D2O proceeded with comparable rates, as was
detected by 31P NMR spectroscopy (Figure 2), showing that
the substituents in the ring do not play a significant role in the
acidity of the agostic proton.
(2)
agostically bound protons is a known process,17 it has not yet
been demonstrated with aromatic hydrocarbons. In the latter,
the HX elimination during metalation is solely attributed to the
electrophilic pathway.1,3,18 If 2a is however viewed as an
arenium species, then its thermodynamic stability is somewhat
unexpected, as most of the protonated Wheland intermediates
can be observed only at low temperatures and/or in the presence
of excess of superacids.19
Obviously, the arenium form must be influenced by electron-
donating or -withdrawing substituents on the ring. To explore
a substituent effect, the new ligand 1b, bearing three methoxy
groups in the aromatic ring, was synthesized (Scheme 2). When
1b was reacted with 0.5 equiv of [Rh(COE)2Cl]2 at room
Although the neutral [RhCl] fragment easily oxidatively adds
the C-H bond in PCP-type ligands to give the corresponding
Rh(III) hydrido chloride complexes, the cationic [RhCO]+
fragment is not sufficiently electron rich to insert into the C-H
bond. Complex 2a is invariably stable in the -60 to +90 °C
temperature range. This stability is a result of the strong C-H
agostic interaction with the cationic metal center.20 This agostic
(16) The real values of C-H bond distances are somewhat larger than
those obtained from the X-ray studies. See, for example: Crabtree, R. H.;
Holt, E. M.; Lavin, M.; Morehouse, S. M. Inorg. Chem. 1985, 24, 1986.
(17) (a) Speckman, D. M.; Knobler, C. B.; Hawthorne, M. F. Organo-
metallics 1985, 4, 1692. (b) Kanamori, K.; Broderick, W. E.; Jordan, R.
F.; Willett, R.; Legg, J. I. J. Am. Chem. Soc. 1986, 108, 7122.
(18) A reviewer suggested that deprotonation of 2a can take place from
an unobserved Rh(III) carbonyl hydride, which can be in equilibrium with
the agostic 2a. While an equilibrium involving fluxional rhodium hydride
complexes of bis(di-tert-butylphosphino)pentane was reported in the fol-
lowing references, no evidence for such an equilibrium was obtained in
our rigid system: (a) Crocker, C.; Errington, R. J.; Markham, R.; Moulton,
C. J.; Shaw, B. L. J. Chem. Soc., Dalton Trans. 1982, 387. (b) Crocker, C.;
Errington, R. J.; Markham, R.; Moulton, C. J.; Odell, K. J.; Shaw, B. L. J.
Am. Chem. Soc. 1980, 102, 4373. In addition, such an equilibrium cannot
explain the deuteration of 2a with D2O.
(19) (a) Koptyug, V. A. Contemporary Problems in Carbonium Ion
Chemistry III; Topics in Current Chemistry 122; Springer-Verlag: Berlin,
1984. See also: (b) Brouwer, D. M.; Mackor, E. L.; MacLean, C. In
Carbonium Ions; Olah, G. A., Schleyer, P. v. R., Eds.; Wiley: New York,
1970; Vol. 2, p 837. (c) Farcasiu, D. Acc. Chem. Res. 1982, 15, 46. (d)
Farcasiu, D.; Marino, G.; Miller, G.; Kastrup, R. V. J. Am. Chem. Soc.
1989, 111, 7210. (e) Xu, T.; Barich, D. H.; Torres, P. D.; Haw, J. F. J. Am.
Chem. Soc. 1997, 119, 406 and references therein.
(20) For recent examples of stabilizing agostic interactions in cationic
late transition metal complexes with bulky phosphine ligands, see: (a)
Huang, D.; Huffman, J. C.; Bolinger, J. C.; Eisenstein, O.; Caulton, K. G.
J. Am. Chem. Soc. 1997, 119, 7398. (b) Cooper, A. C.; Streib, W. E.;
Eisenstein, O.; Caulton, K. G. J. Am. Chem. Soc. 1997, 119, 9069. (c)
Ujaque, G.; Cooper, A. C.; Maseras, F.; Eisenstein, O.; Caulton, K. G. J.
Am. Chem. Soc. 1998, 120, 361.

12542 J. Am. Chem. Soc., Vol. 120, No. 48, 1998
Vigalok et al.
Figure 2. 31P{1H} NMR spectrum (CDCl3) of the reaction of complexes 2a (5 mg, 0.007 mmol) and 2b (5 mg, 0.007 mmol) with D2O (8 µL,
0.400 mmol) in THF (1 mL) after 18 h at room temperature. Larger excesses of D2O resulted in partial deprotonation of the complexes.
Scheme 3
Chart 1
A contour plot of the computed electron density of 5b in the
Cipso-Hipso-Rh plane is given in Figure 3. The picture of the
electron density between the ipso-C-H bond and the Rh atom
is strongly suggestive of an agostic interaction. Between the
Rh and Cipso atom, a Lo¨wdin bond order26 of 0.43 is found;
interaction results in increased acidity of the aromatic proton
which can now be easily removed even by very weak bases.
However, formal positive charge transfer from the metal to the
aromatic ring to form an arenium-type species takes place, if at
all, to only a minor extent. There are many instances of
aromatic metalation where the oxidative addition and electro-
philic mechanisms cannot be easily distinguished.21 Thus, it is
possible that metalation of aromatic hydrocarbons can occur
via the formation and deprotonation of an η2 C-H agostic
intermediate, the reactivity of which mimics the electrophilic
activation mechanism (Scheme 3).
that between Rh and Hipso is only 0.13. As expected, the Cipso
-
Hipso bond is somewhat weakened by this interaction; the
Lo¨wdin bond order is 0.73 (compared to 0.87 for the other ring
C-H bonds), while the computed Cipso-Hipso bond length is
0.026 Å longer than the other aromatic C-H bonds.
The calculations offer several gauges for arenium (Wheland)
character. Relative to those of 5a, the ring bond lengths in 5b
are affected only quite weakly: Cipso-Cortho +0.011 Å, Cortho
-
To further verify our interpretation, we have carried out
B3LYP/LANL2DZ22-24 density functional calculations on
model compounds for 1a and 2a using Gaussian 94.25 The
model compounds, denoted 5a and 5b (Chart 1), respectively,
solely differ by the substitution of smaller methyl groups for
the tert-butyl groups. Our computed structure for 5b agrees
quite closely with the corresponding parts of the X-ray structure
of 2a, except for somewhat more realistic C-H bond distances
and (because of the absence of the triflate counterion) Cs
symmetry.
Cmeta -0.007 Å, Cmeta-Cpara -0.006 Å. For comparison, the
corresponding values for the arenium compound of 5a, denoted
5c here, at the same level of theory are +0.088, -0.019, and
+0.009 Å. Noticeable alternation is seen in the Lo¨wdin bond
orders in the benzene ring of 5b: ipso-ortho 1.33, ortho-meta
1.46, and meta-para 1.49, compared to 1.44, 1.43, and 1.50,
(25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Revision D.4;
Gaussian, Inc.: Pittsburgh, PA, 1995.
(21) See, for example: (a) Brainard, R. L.; Nutt, W. R.; Lee, T. R.;
Whitesides, G. M. Organometallics 1988, 7, 2379. (b) Cordone, R.; Harman,
W. D.; Taube, H. J. Am. Chem. Soc. 1989, 111, 2896.
(22) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(23) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(24) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270, 284, 299.
(26) Natiello, M. A.; Medrano, J. A. Chem. Phys. Lett. 1984, 105, 180.

η2 C-H Agostic Rhodium Arene Complexes
J. Am. Chem. Soc., Vol. 120, No. 48, 1998 12543
stored under high-purity nitrogen on molecular sieves (3 Å). [Rh-
(C2H4)(CO)Cl]2,11b 1a,10 and 4,6-dichloromethyl-1,2,3-trimethoxyben-
zene27 were prepared according to literature procedures.
1H, 31P, and 13C NMR spectra were recorded at 400, 162, and 100
MHz, respectively, using a Bruker AMX400 spectrometer. 1H and 13
C
chemical shifts are reported in ppm downfield from TMS and referenced
to the residual solvent h1 (7.24 ppm chloroform-d, 7.15 ppm benzene-
d6) and all-d solvent peaks (77.00 ppm chloroform, 128.00 ppm
benzene), respectively. 31P chemical shifts are in ppm downfield from
H3PO4 and referenced to an external 85% phosphoric acid sample. All
measurements were performed at 21 °C unless otherwise specified.
Density functional calculations were carried out with the Gaussian
9426 package running on a DEC Alpha 500 workstation, using the
B3LYP22,23 exchange-correlation functional and the LANL2DZ24 basis
set/relativistic effective core potential combination. An initial structure
for a model compound of 2a was generated by symmetrizing the X-ray
structure of 2a to Cs, substituting methyl groups in standard geometries
for the tert-butyl groups, and manually adjusting the overly short X-ray
C-H distances. The geometry of this compound was then fully
optimized, as were (for comparison) structures of [RhCO]+ and of a
model compound for 1a generated by deleting the [RhCO]+ moiety
from the model compound for 2a. A cross section of the electron
density in the Cipso-Hipso-Rh plane was generated using Spartan,28
while charge distributions were obtained using the natural population
analysis (NPA)29 method and bond orders were obtained by the
Lo¨wdin26 method.
Figure 3. Plot of the B3LYP/LANL2DZ electron density in the Cipso
Hipso-Rh plane.
-
respectively, in 5a. The alternation is however much smaller
than in 5c (1.10, 1.54, and 1.44, respectively).
Preparation of a THF Solution of [Rh(C2H4)(CO)(solv)n]+ (n )
1, 2). To 0.5 mL of a THF solution of [Rh(ethylene)(CO)Cl]2 (10 mg,
0.026 mmol) was added 14 mg of AgOTf (0.054 mmol) in 1 mL of
THF. The mixture was allowed to remain in the dark for 30 min, after
which the precipitate of AgCl (7 mg, 0.049 mmol) was filtered off.
The resulting pale yellow solution was used for further reactions with
1a and 1b.
The dihedral angle Cmeta-Cortho-Cipso-Hipso in 5b differs
from the ideal planar value (180°) by only 16.3°, compared to
61.6 and 48.8° for the two ipso hydrogens in 5c. Furthermore,
taking the difference between the NPA populations of [RhCO+]
and 5a on one hand and 5b on the other hand, we see that most
of the positive charge remains concentrated on the Rh and
particularly its -P(alkyl)2 ligands, while the sum of all the
charge differences on the non-ipso ring carbons is only 0.19,
or 0.26 when the charges borne by their hydrogens are included
in the total. In short, while we do have an arenium contribution,
it is a very weak one, and an interpretation of 5b, viz. 2a, in
terms of an agostic interaction rather than a Wheland complex
certainly appears to be valid.
In conclusion, we have prepared and fully characterized novel
chelated Rh(I) complexes with an η2 aromatic C-H agostic
interaction. These complexes demonstrate relatively strong
C-H aciditysso far unobserved for aromatic compoundssand
they can be deprotonated by weak organic bases to give the
corresponding cyclometalated products. The observed reactivity
of these agostic complexes is similar to that attributed to metal
arenium complexesscrucial intermediates in electrophilic aro-
matic metalation. However, the NMR data analysis and
deuterium-labeling experiments have shown that there is very
little, if any, contribution of a metal arenium structure in these
molecules. This interpretation is supported by the computed
electron density, bond orders, charge distributions, and bond
distances obtained using B3LYP/LANL2DZ density functional
calculations. Thus, strong agostic interaction between a metal
center and an aromatic compound followed by deprotonation
can be considered as an alternative to the generally expected
electrophilic aromatic metalation involving arenium intermedi-
ates.
Ligand 1a. The full NMR data (including the 13C NMR) have not
been previously reported.10
31P{1H}NMR (δ, ppm) (CDCl3): 34.22 (s). 1H NMR (δ, ppm): 7.25
(br s, 1H, ipso-ArH), 7.12-7.13 (overlapped signals of p- and o-ArH,
3H), 2.81 (d, JHH ) 3.1 Hz, 4H, CH2P), 1.10 (d, JPH ) 10.8 Hz, 36H,
t-Bu). 13C{1H} NMR (δ, ppm): 141.24 (br d, JPC ) 11.5 Hz, o-Ar),
130.63 (t, JPC ) 7.5 Hz, ipso-C), 128.10 (s, p-Ar), 126.60 (dd, JPC
)
9.3 Hz, JPC ) 1.9 Hz, m-Ar), 31.69 (d, JPC ) 21.9 Hz, C(CH3)3), 29.75
(d, JPC ) 12.9 Hz, C(CH3)3), 28.41 (d, JPC ) 23.1 Hz, CH2P). The
assignments were confirmed by 13C DEPT and 13C-1H 2D correlation
experiments.
Complex 2a. (a) To a THF solution of [Rh(C2H4)(CO)(solv)n]+ (n
) 1, 2) was added 1 mL of a THF solution of 1a (22 mg, 0.056 mmol).
The resulting solution was allowed to stand for 1 h, concentrated in a
vacuum (to ca. 50% of the volume), and then poured into 5 mL of
pentane. The yellow precipitate was filtered off with a cotton pad,
washed with pentane, and redissolved in CH2Cl2. Removal of the
solvent in a vacuum gave pure 2a (27 mg, 77%).
(b) To a CH2Cl2 solution of (CO)Rh[C6H3-1,3-(CH2P(t-Bu)2)2]10 (3)
was added 1.2 equiv of HOTf. The 31P NMR spectrum showed
quantitative formation of 2a.
31P{1H}NMR (δ, ppm) (CDCl3): 34.90 (d, JRhP ) 100.4 Hz). 1H
NMR (δ, ppm): 7.63 (t, JHH ) 8.2 Hz, 1H, p-ArH), 7.24 (d, JHH ) 8.2
Hz, 2H, m-ArH), 4.13 (d, JRhH ) 18.1 Hz, 1H, ipso-ArH), 3.60 (AB
quart, JHH ) 16.3 Hz, 4H, CH2P), 1.45 (vt, JPH ) 7.6 Hz, 18H, t-Bu),
1.23 (vt, JPH ) 7.3 Hz, 18H, t-Bu). 13C{1H} NMR (δ, ppm): 188.60
(dt, JRhC ) 90.5 Hz, JPC ) 11.8 Hz, RhCO), 154.17 (td, JPC ) 4.2 Hz,
JRhC ) 1.7 Hz, o-Ar), 138.34 (br d, p-Ar), 127.62 (t, JPC ) 5.6 Hz,
m-Ar), 110.95 (t, JPC ) 4.4 Hz, ipso-C), 37.37 (t, JPC ) 8.5 Hz,
C(CH3)3), 35.11 (td, JPC ) 9.1 Hz, JRhC ) 1.3 Hz, C(CH3)3), 29.95 (t,
JPC ) 10.7 Hz, CH2P), 29.76 (t, JPC ) 2.4 Hz, C(CH3)3), 29.00 (t, JPC
Experimental Section
(27) Monte, A. P.; Waldman, S. R.; Marona-Lewicka, D.; Wainscott,
D. B.; Nelson, D. L.; Sanders-Bush, E.; Nichols, D. E. J. Med. Chem. 1997,
40, 2997.
General Procedures. All operations with air- and moisture-sensitive
compounds were performed in a nitrogen-filled glovebox (Vacuum
Atmospheres with a MO-40 purifier). All solvents were reagent grade
or better. Pentane, benzene, and THF were distilled over sodium/
benzophenone ketyl. All solvents were degassed and stored under high-
purity nitrogen after distillation. All deuterated solvents (Aldrich) were
(28) Spartan User’s Guide, Version 4.1; Wavefunction, Inc.: Irvine, CA,
1995.
(29) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88,
899.

12544 J. Am. Chem. Soc., Vol. 120, No. 48, 1998
Vigalok et al.
) 2.8 Hz, C(CH3)3). IR (film): 1981 cm-1 (s). Anal. Found (calc):
C, 46.28 (46.57); H, 6.30 (6.01).
m-OCH3), 36.27 (t, JPC ) 7.6 Hz, C(CH3)3), 34.58 (td, JPC ) 8.9 Hz,
JRhC ) 1.4 Hz, C(CH3)3), 29.78 (t, JPC ) 2.7 Hz, C(CH3)3), 29.38 (t,
JPC ) 2.5 Hz, C(CH3)3), 28.70 (td, JPC ) 11.6 Hz, JRhC ) 2.6 Hz,
CH2P). Anal. Found (calc): C, 52.40 (52.05); H, 7.99 (8.09).
Ligand 1b. A mixture of di-tert-butylphosphine (10.440 g, 71.5
mmol) and 4,6-dichloromethyl-1,2,3-trimethoxybenzene27 (7.55 g, 28.6
mmol) in acetone (75 mL) was refluxed with stirring under argon for
24 h, resulting in white crystals of the diphosphonium salt. The salt
was isolated and washed with cold acetone and with pentane to remove
unreacted starting material. The salt was then dissolved in 90 mL of
distilled degassed water, and to the clear solution was added 30 g of
NaOAc in 90 mL of distilled degassed water. The organic product
was extracted twice with 200 mL of CH2Cl2, and the organic layer
was separated from the mixture and dried over Na2SO4. The solvent
was evaporated under vacuum to give the product which was
contaminated with phosphine oxides. Purification by column chro-
matography (silica) afforded pure 1b in 55% yield (7.6 g).
X-ray Analysis of the Structure of 2a. Complex 2a was crystal-
lized from CH2Cl2 at room temperature to give orange crystals. Crystal
data: C26H44F3O4P2RhS, orange prism, 0.3 × 0.3 × 0.3 mm3,
monoclinic, P21/c, a ) 8.510(2) Å, b ) 23.825(5) Å, c ) 14.801(3)
Å, â ) 93.97(3)° from 25 reflections, T ) 110 K, V ) 2993.7(11) Å3,
Z ) 4, fw ) 674.52, Dc ) 1.497 Mg/m3, µ ) 0.795 mm-1
. Data
collection and treatment: Rigaku AFC5R four-circle diffractometer,
Mo KR, graphite monochromator (λ ) 0.710 73 Å), 5107 reflections
collected, 1.62° e θ e 27.49°, 0 e h e 11, -2 e k e 16, -17 e l e
17, ω scan method, scan width 1.4°, scan speed 2°/min, typical half-
height peak width 0.25°, 3 standards collected 27 times each, with a
6% change in intensity, 4764 independent reflections (Rint ) 0.0325).
Solution and refinement: The structure was solved by the Patterson
method (SHELXS-96). Full-matrix least-squares refinement was based
on F2 (SHELXL-93). Idealized hydrogens were placed and refined in
a riding mode, with the exception of H(11) on C(11), which was located
independently and refined freely; 345 parameters with no restraints;
final R1 ) 0.0492 (based on F2) for data with I > 2σ(I) and R1 )
0.0743 for all data based on all 4759 reflections; goodness-of-fit on F2
31P{1H}NMR (δ, ppm) (CDCl3) 31.41 (s). 1H NMR (δ, ppm): 7.11
(br t, ipso-ArH), 3.85 (s, 6H, m-OCH3), 3.82 (s, 3H, p-OCH3), 2.74
(d, JPH ) 3.3 Hz, 4H, CH2P), 1.12 (d, JPH ) 10.7 Hz, 36H, t-Bu).
13C{1H} NMR (δ, ppm): 149.42 (t, JPC ) 2.2 Hz, p-Ar), 146.09 (s,
m-Ar), 129.23 (d, JPC ) 11.5 Hz, o-Ar), 127.15 (t, JPC ) 10.32 Hz,
ipso-C), 60.75 (s, m-OCH3), 60.51 (s, p-OCH3), 31.79 (d, JPC ) 23.7
Hz, C(CH3)3), 29.71 (d, JPC ) 13.2 Hz, C(CH3)3), 21.98 (d, JPC
24.3 Hz, CH2P).
)
) 1.015; largest electron density ) 0.878 e/Å-3
.
Complex 2b. Complex 2b was prepared and purified analogously
to 2a.
31P{1H}NMR (δ, ppm) (CDCl3): 39.76 (d, JRhC ) 101.5 Hz). 1H
NMR (δ, ppm): 4.13 (d, JRhH ) 17.6 Hz, 1H, ipso-ArH), 3.91 (s, 3H,
p-OCH3), 3.86 (s, 6H, m-OCH3), 3.56 (AB quart, JHH ) 16.3 Hz, 4H,
CH2P), 1.44 (vt, JPH ) 7.6 Hz, 18H, t-Bu), 1.21 (vt, JPH ) 7.3 Hz,
18H, t-Bu). 13C{1H} NMR (δ, ppm): 189.10 (dt, JRhC ) 89.1 Hz, JPC
) 11.6 Hz, RhCO), 159.98 (br m, p-Ar), 148.90 (td, JPC ) 5.1 Hz,
JRhC ) 0.9 Hz, m-Ar), 144.57 (td, JPC ) 4.3 Hz, JRhC ) 1.7 Hz, o-Ar),
104.77 (t, JPC ) 4.4 Hz, ipso-C), 62.40 (s, m-OCH3), 60.69 (s, p-OCH3),
36.95 (td, JPC ) 8.1 Hz, JRhC ) 0.6 Hz, C(CH3)3), 35.14 (td, JPC ) 9.1
Hz, JRhC ) 1.3 Hz, C(CH3)3), 29.28 (t, JPC ) 2.5 Hz, C(CH3)3), 28.99
(t, JPC ) 2.8 Hz, C(CH3)3), 23.48 (t, JPC ) 10.4 Hz, CH2P). IR (film):
1992 cm-1 (s).
Complex 4. To a suspension of [Rh(COE)2Cl]2 (15 mg, 0.02 mmol)
in 1 mL of THF was added 1 mL of a THF solution of 1b (20 mg,
0.041 mmol). The resulting yellow solution showed quantitative
formation of 4 on the basis of 31P NMR spectroscopy. The solvent
was evaporated, and the resulting yellow solid was dried in a vacuum.
Yield: 25 mg (98%).
Acknowledgment. This work was supported by the Israel
Science Foundation, Jerusalem, Israel, and by the MINERVA
Foundation, Munich, Germany. The authors thank B. Rybtch-
inski for valuable discussions. D.M. is the holder of the Israel
Matz Professorial Chair of Organic Chemistry. J.M.L.M. is a
Yigal Allon Fellow, an Honorary Research Associate (“Onder-
zoeksleider in Eremandaat”) of the National Science Foundation
of Belgium (NFWO/FNRS), and the Incumbent of the Helen
and Milton A. Kimmelman Career Development Chair. O.U.
participates in the Franco-Israeli Scientific Cooperation Program
sponsored by the French Embassy in Israel.
Supporting Information Available: Text describing the
crystal structure determination of 2a, an ORTEP diagram of
2a, tables of crystal data and structure refinement details, atomic
coordinates, bond lengths and angles, anisotropic displacement
parameters, and hydrogen atom coordinates for complex 2a, and
a table giving the B3LYP/LANL2DZ computed structures of
5a-c in Cartesian coordinates (Å) (11 pages, print/PDF). See
any current masthead page for ordering information and Web
access instructions.
31P{1H}NMR (δ, ppm) (CDCl3): 74.15 (d, JRhC ) 114.3 Hz). 1H
NMR (δ, ppm): 3.85 (s, 3H, p-OCH3), 3.83 (s, 6H, m-OCH3), 3.14
(AB quart, JHH ) 17.9 Hz, 4H, CH2P), 1.36 (m, 36H, 2 overlapped
t-Bu), -27.96 (dt, JRhH ) 52.0 Hz, JPH ) 11.9 Hz, 1H, RhH). 13C-
{1H} NMR (δ, ppm): 154.34 (br d, JRhC ) 33.2 Hz, ipso-C), 147.20
(td, JPC ) 8.2 Hz, JRhC ) 2.0 Hz, m-Ar), 141.62 (s, p-Ar), 139.26 (td,
JPC ) 10.4 3 Hz, JRhC ) 1.3 Hz, o-Ar), 60.51 (s, p-OCH3), 60.03 (s,
JA982534U