The synthesis, aromaticity, and NMR properties of [14]annulene fused organometallics. Determination of the effective bond localizing ability ('relative aromaticity') and diamagnetic anisotropy of several organometallic moieties

Article abstract of DOI:10.1021/ja9729138

The aromatic ring current probe, the bridged [14]annulene dimethyldihydropyrene, 1, is used to investigate the bond localization effects of the organometallic species tricarbonylchromiumbenzene, 37, hexamethylbenzenerutheniumbenzene, 38, tricarbonylmanganesecyclopentadienyl, 40, and pentamethylcyclopentadienylrutheniumcyclopentadienyl, 39, when fused to the annulene. Benzo[a]dimethyldihydropyrene 2 was converted to the two isomers of its tricarbonylchromium complex, 9 and 10, in 60% yield using ligand exchange with naphthalene and to its hexamethylbenzene-ruthenium complexes 11 and 12 with [RuCl₂(HMB)]₂ and AgBF₄. The cyclopentadienide fused dihydropyrene 5 was synthesized from dimethyldihydropyrene 1 in eight steps, and then on reaction with pentacarbonyl manganese bromide gave the tricarbonylmanganese complex of 5, as two isomers 28 and 29 in 61% yield, and with (Cp*RuCl₂)(n) gave 70% yield of the dihydropyrene annelated ruthenocene isomers 26 and 27. Bis(tricarbonylchromium) complexes 32-34 were obtained in 50% yield from the dibenzannulene 30. The ¹H NMR spectra for each complex were analyzed in detail with regard to ring current shielding of the internal methyl protons and ring current deshielding of the external protons. The McGlinchey equation was used to remove diamagnetic anisotropy effects, and the resultant chemical shift values were consistent with coupling constant results, which together yielded bond fixation data of the annulene, caused by the organometallic. The relative bond fixing ability of the organometallics was found to be in the order: Cp-Ru-Cp* > benzene-Ru²⁺-(HMB) > Cp-Mn(CO)₃ > benzene-Cr(CO)₃ > benzene. Alternance parameter evidence is presented that the complexes are aromatic, and the authors discuss the relative aromaticity of the complexes to benzene.

Full text of DOI:10.1021/ja9729138

J. Am. Chem. Soc. 1998, 120, 1785-1794
1785
The Synthesis, Aromaticity, and NMR Properties of [14]Annulene
Fused Organometallics. Determination of the Effective Bond
Localizing Ability (“Relative Aromaticity”) and Diamagnetic
Anisotropy of Several Organometallic Moieties
Reginald H. Mitchell,* Yongsheng Chen, Nasr Khalifa,† and Pengzu Zhou
Contribution from the Department of Chemistry, UniVersity of Victoria, P.O. Box 3065, Victoria,
British Columbia, Canada V8W 3V6
ReceiVed August 18, 1997
Abstract: The aromatic ring current probe, the bridged [14]annulene dimethyldihydropyrene, 1, is used to
investigate the bond localization effects of the organometallic species tricarbonylchromiumbenzene, 37,
hexamethylbenzenerutheniumbenzene, 38, tricarbonylmanganesecyclopentadienyl, 40, and pentamethylcyclo-
pentadienylrutheniumcyclopentadienyl, 39, when fused to the annulene. Benzo[a]dimethyldihydropyrene 2
was converted to the two isomers of its tricarbonylchromium complex, 9 and 10, in 60% yield using ligand
exchange with naphthalene and to its hexamethylbenzene-ruthenium complexes 11 and 12 with [RuCl2(HMB)]2
and AgBF4. The cyclopentadienide fused dihydropyrene 5 was synthesized from dimethyldihydropyrene 1 in
eight steps, and then on reaction with pentacarbonyl manganese bromide gave the tricarbonylmanganese complex
of 5, as two isomers 28 and 29 in 61% yield, and with (Cp*RuCl2)n gave 70% yield of the dihydropyrene
annelated ruthenocene isomers 26 and 27. Bis(tricarbonylchromium) complexes 32-34 were obtained in 50%
yield from the dibenzannulene 30. The 1H NMR spectra for each complex were analyzed in detail with regard
to ring current shielding of the internal methyl protons and ring current deshielding of the external protons.
The McGlinchey equation was used to remove diamagnetic anisotropy effects, and the resultant chemical shift
values were consistent with coupling constant results, which together yielded bond fixation data of the annulene,
caused by the organometallic. The relative bond fixing ability of the organometallics was found to be in the
order: Cp-Ru-Cp* > benzene-Ru2+-(HMB) > Cp-Mn(CO)3 > benzene-Cr(CO)3 > benzene. Alternance
parameter evidence is presented that the complexes are aromatic, and the authors discuss the relative aromaticity
of the complexes to benzene.
Introduction
-0.44, respectively.1a This is caused by a strong bond fixation
effect that occurs whenever two aromatic species are fused,2
and affects both fused rings such that adjacent bonds alternate.
This can be seen experimentally in the magnitude of adjacent
3J coupling constants, which become larger for the bond of
greater π-bond order and smaller for the more single-like bond
than those found in the unfused aromatics.2f,g Thus for 2 and
We have established1 that the bridged [14]annulene 1, trans-
10b,10c-dimethyl-10b,10c-dihydropyrene (DMDHP), is an ex-
cellent NMR probe for ring current and bond localization effects.
Fusion of an aromatic species such as benzene or naphthalene
to 1 as in 2 and 3, respectively, strongly reduces the ring current
3
3 the J values for the bonds b and a in the macrocyclic ring
are 8.85 and 6.52, and 9.07 and 6.17, respectively, while for 1
are both 7.54 Hz.1a The degree of both the bond alternation
effect and the reduction of ring current depend on the bond
localization energy of the ring that is fused, i.e., benzene in 2
and naphthalene in 3, and thus comparison of the NMR data
for any fused system, such as 4 with that obtained for 2 and 1,
permits some comment on the relative aromaticity of the fused
ring relative to benzene.1
(2) (a) Mitchell, R. H.; Carruthers, R. J.; Mazuch, L.; Dingle, T. W. J.
Am. Chem. Soc. 1982, 104, 2544-2551. (b) Mitchell, R. H.; Yan, J. S. H.;
Dingle, T. W. J. Am. Chem. Soc. 1982, 104, 2551-2559. (c) Mitchell, R.
H.; Williams, R. V.; Dingle, T. W. J. Am. Chem. Soc. 1982, 104, 2560-
2571. (d) Mitchell, R. H.; Williams, R. V.; Mahadevan, R.; Lai, Y. H.;
Dingle, T. W. J. Am. Chem. Soc. 1982, 104, 2571-2578. (e) Mitchell, R.
H. Isr. J. Chem. 1980, 20, 294-299. (f) Cremer, D.; Gu¨nther, H. Liebigs
Ann. Chem. 1972, 763, 87-108. (g) Gu¨nther, H.; Shyoukh, A.; Cremer,
D.; Frish, Liebigs Ann. Chem. 1978, 150-164. (h) K. H. Balaban, A. T.;
Banciu, M.; Ciorba, V. Annulenes, Benzo-, Hetero-, Homo-DeriVatiVes; CRC
Press: Boca Raton, FL, 1987; Vol. II, Chapter 7.
that is present in 1 and hence reduces the ring current shielding
of the internal methyl protons from δ -4.25 to -1.62 and
† Present address: Chemistry, Camosun College, Victoria, BC.
(1) (a) Mitchell, R. H.; Iyer, V. S.; Khalifa, N.; Mahadevan, R.;
Venugopalan, S.; Weerawarna, S. A.; Zhou, P. J. Am. Chem. Soc. 1995,
117, 1514-1532. (b) Mitchell, R. H.; Iyer, V. S. J. Am. Chem. Soc. 1996,
118, 722-726. (c) Mitchell, R. H.; Iyer, V. S. J. Am. Chem. Soc. 1996,
118, 2903-2906. (d) Mitchell, R. H.; Khalifa, N. A.; Dingle, T. W. J. Am.
Chem. Soc. 1991, 113, 6696-6697.
S0002-7863(97)02913-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/14/1998

1786 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Mitchell et al.
In this paper we fuse several organometallic species to 1.
This is of interest because, while ferrocene is commonly
regarded as aromatic, other species such as tricarbonylchro-
mium(0)-benzene are less obvious, in part (see below) because
of the approximately 2 ppm upfield shift that is observed for
its protons, relative to benzene. This paper thus compares the
relative bond fixing ability, which we equate to aromaticity, of
some benzene and cyclopentadienide metal complexes.
and -1.158 for the minor isomer 10. The hexamethylbenzene
ruthenium(II) complexes, 11 and 12, as 2BF4- salts, were also
Syntheses
7
prepared as very dark red crystals from 2 and [RuCl2(HMB)]2
As representative examples, we considered metal complexes
of the benz[a]annulene, 2, and the cyclopentadienide fused
annulene, 5, to be suitable targets, since [e]-fused annulenes
such as 6 added the additional complication of easy valence
isomerization to the cyclophane-diene isomer 7.2b,3 At the start
of this work, no metal complexes of benzannulenes were known.
Indeed few metal complexes of any larger annulene were
known,4 probably because of the very limited accessibility and
stability of either group of compounds. We first5 attempted
reaction of 2 with both Cr(CO)6 and Cr(CO)3(MeCN)3 under
normal thermal conditions, but obtained only decomposition of
2. We thus considered the more mild THF catalyzed ligand
exchange reaction6 of (η6-naphthalene)Cr(CO)3, 8.
using AgBF4 at room temperature. However, in this case, the
pair of isomers (3:2 ratio) could not be separated. The major
isomer showed its internal methyl protons at δ -0.517 and
-0.864, while those of the minor were at δ -0.558 and -0.891.
Synthesis of cyclopentadienide derivatives proceeded from
5, which we reported in preliminary form in 1991.1d The route
used to 5 is shown in Scheme 1.
Scheme 1a
It was not immediately obvious whether 2 would exchange
with the naphthalene in 8, since the exchange reaction is
normally used for benzene derivatives. Since in naphthalene
two 6π systems are competing, some loss of resonance energy
occurs for the uncomplexed ring in 8. If 8 were to exchange
with 2, the loss of resonance energy would be in a 14π rather
than a 6π system, and thus might be expected to be favorable.
We thus thought that the exchange reaction was worth trying.
In the event, reaction of 2 and 8 in ether containing THF as
catalyst at 50-60 °C gave 60% of the desired product, as a
mixture of the two expected isomers 9 and 10 in a 3:1 ratio.
The major isomer, 9, could be separated by fractional crystal-
lization as deep red crystals, mp 189-190 °C. The overall
structures of 9 and 10 were supported by NMR data (see below),
mass spectra, and elemental analysis. The relative stereochem-
istries of 9 and 10 probably could be made from the NMR
spectra using anisotropy arguments; however, an X-ray structure
determination on 9 unambiguously indicated the stereochemistry
shown, where the Cr(CO)3 fragment is attached on the opposite
side to the closest attached internal methyl group. The two
isomers are then easily distinguished by the chemical shifts of
their internal methyl protons (note: these shifts are determined
in the same solution), δ -0.871 and -0.977 for 9, and δ -0.807
aConditions: (a) (EtO)2P(O)CH2COOEt/NaH; (b) Pd/H2; (c) aq
NaOH; (d) (COCl)2; (e) BF3‚OEt2; (f) LAH then HCl; (g) KH/THF.
Formylation of 1 using Cl2CHOCH3 and TiCl4 gave 75% of
the 2-formyl derivative 13, along with 20% of the 1-formyl
isomer.8 Wittig-Horner reaction of 13 gave almost quantita-
tively the dark red ester 14, mp 146-147 °C, which could be
hydrogenated in 86% yield to the green ester 15. Hydrolysis
then gave the acid 16, mp 165 °C. Cyclization to the ketone
18 failed with both HF and polyphosphoric acid, and reaction
of 16 with thionyl chloride or aluminum chloride caused
extensive decomposition. Reaction of the acid 16 with oxalyl
chloride did, however, yield the acid chloride 17, IR at 1802
cm-1, which was directly treated with BF3-etherate in CH2Cl2
at room temperature to give cyclized ketone 18 (green, mp 210-
211 °C) in overall 63% yield. Reduction of 18 with LiAlH4
was expected to give the alcohol; however, the sulfuric acid
used in workup caused extensive dehydration and decomposi-
tion. Use of 20% hydrochloric acid dramatically improved the
yield and enabled fused cyclopentadiene 19 to be isolated in
81% yield, as dark green crystals, mp 114-116 °C. These were
only stable for about 2 days (<0 °C), so were always
chromatographed quickly before use. The internal methyl
protons of 19 appeared at δ -4.15 and -4.16, almost
unperturbed from those in 1 at δ -4.25. The red anion 5 could
be generated from green 19 in THF under argon using either
(3) Mitchell, R. H.; Iyer, V. S.; Mahadevan, R.; Venugopalan, S.; Zhou,
P. J. Org. Chem. 1996, 61, 5116-5120.
(4) Fischer, E. O.; Ruhle, H.; Vogel, E.; Grimme, W. Angew. Chem.,
Int. Ed. Engl. 1966, 5, 518-519. Stockel, K.; Sondheimer, F.; Clarke, T.
A.; Guss, M.; Mason, R. J. Am. Chem. Soc. 1971, 93, 2571-2572. Hilken,
G.; Kinkel, T.; Schwamborn, M.; Lex, J.; Schmickler, H.; Vogel, E. Angew.
Chem., Int. Ed. Engl. 1982, 21, 784.
(5) For a preliminary report, see: Mitchell, R. H.; Zhou, P.; Venugopalan,
S.; Dingle, T. W. J. Am. Chem. Soc. 1990, 112, 7812-7813.
(6) Kundig, E. P.; Perret, C.; Spichiger, S.; Bernardinelli, G. J. Orga-
nomet. Chem. 1985, 286, 183-200.
(7) Bennett, M. A.; Matheson, T. W. J. Organomet. Chem. 1979, 175,
87-93. Bennett, M. A.; Matheson, T. W.; Robertson, G. B.; Smith, A. K.;
Tucker, P. A. Inorg. Chem. 1980, 19, 1014-1021.
(8) Mitchell, R. H.; Lau, D. Y. K.; Miyazawa, A. Org. Prep. Proced.
Int. 1996, 28, 713-716.

[14]Annulene Fused Organometallics
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1787
potassium hydride or methyllithium. Protonation of the anion
with water returned 19 and its isomer 20 (δ(Me) -4.09) in a
2:3 ratio.
From the anion 5, the desired targets were the “ferrocene like”
molecules 21 and 22. Bis(π-indenyl)iron, 23, has been prepared
treatment with FeCl2 (rather very unstable Fe(II) salts), nor does
in our hands dibenzocyclopentadienide (fluorenide), whereas
benzocyclopentadienide (indenide) does. In the latter case, our
calculations indicate only about 10% of the negative charge is
delocalized around the benzene ring. Despite these discouraging
results, we noted that Gassman20a had been able to prepare (η5-
fluorenyl)(η5-Cp*)ruthenium(II), and that Mulay20b has sug-
gested that ruthenocene has less electron density in the five-
membered rings than does ferrocene, and so we decided that
ruthenocenes might be worth pursuing.
20,21
Reaction of (Cp*RuCl2)n
with anion 5 at room temper-
ature in THF was successful and gave the ruthenocene in 70%
yield as a 3.7:1 mixture of the two isomers 26 and 27. Neither
from the indenyl anion and both ferric chloride9 and ferrous
chloride.10 As well, mixed ferrocenes such as 24 have been
prepared10 by reacting two different cyclopentadienide anions
with ferrous chloride in dry THF. Indeed, we were able to
obtain both 23 and 24 under these conditions, but failed to obtain
either 21 or 22 when anion 5 was used. Use of FeCl2‚2THF,
which dramatically improves the yields of some ferrocenes,11
did not change the result, nor did use of Fe(acac)2,12 nor did
use of Cp* anion or [Cp*Fe(acac)],13 reversal of addition order,
or change in temperature. Both cyclopentadienide14 and in-
denide15 form air-stable thallium(I) complexes; however, only
decomposition was observed from 5. Gill and Mann16 have
used the cation 25 to deliver the iron-Cp+ unit to a number of
chromatography, fractional crystallization, nor sublimation
would separate these isomers, though they were obtained
analytically pure as the mixture, with correct MH+ pattern
around 506 in the CI mass spectrum. The major isomer,
assigned structure 26 by analogy to 9, showed internal methyl
protons at δ -0.548 and -0.691, while those of the minor
isomer were more separated (compare 9 and 10 above) at δ
-0.473 and -0.820. The compounds were air-sensitive.
Following this success, we anticipated that reaction of anion
5 with dipyridine tricarbonylmanganese(I) bromide,22 which is
prepared from pentacarbonylmanganese(I) bromide,23 would
yield the two tricarbonylmanganese compounds 28 and 29. This
electron rich arene substrates under photolytic conditions. We
saw no reason a cyclopentadienide anion should not be used in
this process, and indeed photolysis of 25 and potassium
cyclopentadienide in DMF for 4 h, using an ordinary 150-W
flood lamp, yielded 80% of ferrocene, while potassium indenide
yielded 69% of the mixed ferrocene 24. However, both the
anion 5 and fluorenyl anion failed in this reaction.
Thus, so far we have failed to obtain an iron complexed
derivative of 5. We believe this to be caused in part by too
much delocalization of the negative charge around the annulene
ring (40% by π-SCF calculation17 ). In support of this, we note
that both tetracyano-18 and penta(carbomethoxy)-19 cyclopen-
tadienide (which both have strongly electron-withdrawing
groups on the cyclopentadienide) do not form a ferrocene on
reaction failed. However, reaction of the anion 5 with penta-
carbonylmanganese(I) bromide itself gave the two isomers in
61% yield in a 2.5:1 ratio. The major isomer was assigned
structure 28, and showed internal methyl protons at δ -0.728
and -0.784, while 29 was at δ -0.670 and -0.976. Again
satisfactory analysis and mass spectral data were obtained, but
the two isomers could not be separated. Note that in each case,
the major isomer (9, 26, and 28) was assigned with the closest
attached methyl on the opposite side (i.e., as far away as
possible) to the metal fragment, and that these isomers showed
internal methyl protons with the least different chemical shifts.
(9) Paulson, P. L.; Wilkinson, G. J. Am. Chem. Soc. 1954, 76, 2024-
2026.
(10) King, R. B.; Bisnette, M. B. Inorg. Chem. 1964, 3, 796-801.
(11) LeVanda, C.; Bechgaard, K.; Cowan, D. O.; Mueller-Westerhoff,
U. T.; Eilbracht, P.; Candella, G. A.; Collins, R. L. J. Am. Chem. Soc.
1976, 98, 3181-3187.
(12) Buckingham, D. A.; Gorges, R. C.; Henry, J. T. Aust. J. Chem.
1967, 20, 281-296.
(13) Bunel, E. E.; Valle, L.; Manriquez, J. M. Organometallics 1985, 4,
1680-1682.
(14) Meister, H. Angew. Chem. 1957, 69, 533-534.
(15) Bennett, M. A.; Matheson, T. W.; Robertson, G. B.; Smith, A. K.;
Tucker, P. A. Inorg. Chem. 1980, 19, 1014-1021.
(16) Gill, T. P.; Mann, K. R. Inorg. Chem. 1980, 19, 3007-3010.
(17) PCMODEL v 5.1 (PCWIN) was used from Serena Software, Box
3076, Bloomington, IN 47402-3076.
(18) Webster, O. W. J. Am. Chem. Soc. 1966, 88, 4055-4060.
(19) Bamfield, P.; Cookson, R. C.; Crabtree, A.; Henstock, J.; Judec, J.;
Johnson, A. W.; Whitear, B. Chem. Ind. (London) 1964, 1313-1314.
(20) (a) Gassman, P. G.; Winter, C. H. J. Am. Chem. Soc. 1988, 110,
6130-6135. (b) Mulay, L. N.; Withstandley, V. J. Chem. Phys. 1965, 43,
4522-4523.
(21) Oshima, N.; Suzuki, H.; Moro-Oka, Y. Chem. Lett. 1984, 1161-
1164; Tilley, T. D.; Grubbs, R. H.; Bercaw, J. E. Organometallics 1984, 3,
274-278.
(22) Lynch, T. J.; Helvenston, M. C. Organometallics 1989, 8, 1959-
1963.
(23) Abel, E. W.; Wilkinson, G. J. Chem. Soc. 1959, 1501-1505.

1788 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Mitchell et al.
Bis-Complexes. Our recent1a,24 improved route to the
dibenzannulenes 30 and 31 gave us opportunity to attempt bis-
Table 1. Experimental Chemical Shifts in CDCl3 of the Internal
Methyl Protons of the Compounds Studieda
compd
δ(Me)a
δ(Me)b
∆δ
1
2
9
-4.25
0.000
0.008
0.106
0.351
0.340
0.330
0.143
0.347
0.056
0.306
0
-1.618
-0.871
-0.807
-0.517
-0.558
-0.548
-0.473
-0.728
-0.670
+0.64
-1.626
-0.977
-1.158
-0.864
-0.891
-0.691
-0.820
-0.784
-0.976
10
11
12
26
27
28
29
32
33
34
complexation.25a The cisoid-fused system 30 is much more
stable than the transoid-fused 31, and has two benzenoid rings
that could be complexed. Formally 31 has only one benzenoid
ring, unless biradicaloid2c structures are invoked. Indeed,
reaction of 30 with (tricarbonylchromium)naphthalene at 60 °C
for 30 h gave 50% of the bis-complex as a mixture of three
isomers in a 12:2:1 ratio. These gave clear CI MS peaks for
MH+ at 605. The major isomer would be expected to be 32,
and this could be fractionally crystallized pure, mp 219-220
°C. It and the minor isomer 33 are symmetrical and thus show
only one type of methyl proton at δ +0.64 and +0.96,
respectively, while the nonsymmetrical isomer 34 shows two
+0.58
+0.860
0
+0.422
0.438
a Note (Me)a and (Me)b are not yet assigned. Their chemical shifts
are determined in the same molecule and thus ∆δ is given to three
decimal places and was consistent to (0.002.
peaks at δ +0.86 and +0.42.25b A small amount of mono-
complexed product 35 is also formed. Disappointingly, the less
stable transoid-isomer 31 decomposes under these complexing
conditions, and no complexed product could be isolated.
Attempted formation of the bis[(hexamethylbenzene)ruthe-
nium(II)] complexes of 30 and 31 failed; only the mono-complex
36 could be obtained as a mixture of two isomers.
Figure 1. Diagram to show parameters r and θ used in eq 1.
between the methyl groups are really rather small, especially
in comparison to the difference in shift from 2.
It would seem that the through-space anisotropy effect at this
distance is rather small. Fortunately the elegant work of
McGlinchey26 enables us to prove this point. McGlinchey has
made use of the McConnell equation together with a composite
geometric term for the organometallic moiety to determine ø
values for several organometallic groups. We can thus use his
ø value for the Cr(CO)3 fragment to determine the through-
space (de)shielding of this group on any of the protons in our
compounds containing the -Cr(CO)3 group. Application to
compound 9 is shown in Figure 1.
Diamagnetic Anisotropy Effects
We have shown1a that the change in chemical shifts that occur
between 4 and 2 can be used to estimate the relative aromaticity
of the annelating ring in 4 to benzene. Central to this hypothesis
is the assumption that the chemical shifts of the internal methyl
protons and the most distant external ring protons are too far
away from the annelating ring to be affected anisotropically
through space. This we have demonstrated for relatively planar
systems such as 3, 6, 30, or 31.1a,2a-c It is not quite so obvious
whether this holds true for complexed systems such as 9. Any
change in bond-fixation (delocalization) in the macrocyclic ring
caused by complexation of the annelating ring would, of course,
affect both methyl groups equally. A through-space effect of
the metal complexing group might be expected to affect the
methyl groups differently since, of course, one methyl group is
above the plane of the large ring, while the other is below it. In
fact, despite this, as Table 1 shows, the chemical shift differences
For -Cr(CO)3, the three carbonyl groups are considered a
“super-carbonyl” along the C3 axis. The center is 3.3 Å from
the Cr atom. The distance from this center to the proton under
consideration is r (Å) and θ (deg) is the angle made by r and
the C3 axis as shown in Figure 1. McGlinchey’s data then yields
eq 2.
In general, σ (ppm, + ) shielding),
σ ) -øSI × 1036 (3 cos2 θ - 1)/12πr3
(1)
and using McGlinchey’s value for øSI -2124 × 10-36 m3/
molecule
σ ) 56.341(3 cos2 θ - 1)/r3
(2)
(24) Mitchell, R. H.; Chen, Y.; Iyer, V. S.; Lau, D. Y. K.; Baldridge, K.
K.; Siegel, J. S. J. Am. Chem. Soc. 1996, 118, 2907-2911.
(25) (a) For a preliminary report see: Mitchell, R. H.; Chen, Y.
Tetrahedron. Lett., 1996, 37, 6665-6668. (b) In the preliminary report,25a
the chemical shifts for the methyl protons of the minor isomers 33 and 34
were erroneously reported.
where σ is the shielding (ppm) expected for the proton caused
by the anisotropic effect of the Cr(CO)3.
(26) McGlinchey, M. J.; Burns, R. C.; Hofer, R.; Top, S.; Jaouen, G.
Organometallics 1986, 5, 104-109.

[14]Annulene Fused Organometallics
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1789
Table 2. Calculated Shielding’s (σ, ppm) and Experimental
Chemical Shifts (δ) for the Protons of Compounds 9 and 10
the benzene ring through space to be 0.062 and 0.047 ppm for
the closest and furthest methyl group, respectively (note: they
actually differ by 0.008 ppm). Even if the entire ring current
of the complexed benzene ring in 9 were lost, the through-
space effect on the methyl protons would not be expected to be
larger than in 2. Thus we believe that the change in chemical
shift of the protons on going from 2 to 9, reflects a real change
in ring current in the macrocyclic ring, and we will show below
that this is caused by a change in the delocalization in the
macrocyclic ring.
proton
9:σ
9:δ
10:σ
10:δ
Mep
Med
1
2
3
4
5
6
7
8
9
10
11
12
-0.104
+0.118
-0.011
-0.013
-0.006
+0.005
+0.053
+0.255
+0.456
+0.456
+0.453
+0.443
+0.193
+0.022
-0.871
-0.977
7.179
6.882
7.021
7.276
7.276
7.244
5.939
5.613
5.503
6.567
7.594
7.002
-0.030
+0.090
+0.001
-0.008
-0.009
+0.001
+0.045
+0.233
+0.456
+0.472
+0.473
+0.451
+0.185
+0.337
-0.807
-1.158
7.24
6.925
7.095
7.385
7.287
7.244
5.826
5.645
5.453
6.489
7.563
7.012
The data for the minor isomer 10 are not quite so good. As
can be seen from Table 2, the observed chemical shift difference
between the methyl groups is 0.351 ppm. Using eq 1 above,
the calculated difference is 0.120 ppm. This may be because
of small geometry differences between 9 and 10. In fact, in
the literature,29 the R- and â-Cr(CO)3 complexes of methyl
podocarpate adopt slightly different geometries. According to
our molecular model calculations, the skeleton of this isomer
twists slightly away from the plane of the benzene ring
(especially around atoms 12, 12a, 1, probably because of a
greater nonbonded Mep-CO interaction in 10 than in 9. For
example, the average deviation of the macrocyclic ring carbons
from the plane of the benzene ring in 9 is only 4°, with
maximum value 12°, while in the case of isomer 10, the average
value is 10° and the maximum value is 42°. This extra
deformation of the ring carbons in 10 bends the methyl groups
slightly outward from the center of the macrocyclic ring, and
thus might be expected to perturb the observed chemical shifts
away from their values in 9. From the point of view of
aromaticity calculations that we wish to make below, we thus
believe the more planar isomer 9 to be the better choice. Indeed
for the isomers 26/27 and 28/29, similar chemical shift differ-
ences are seen between the pairs of methyl groups, and
molecular model calculations lead to the same conclusions as
for 9/10 above, and thus only data from the major isomers 9,
11, 26, and 28 will be used in the aromaticity calculations below.
McGlinchey26 reported the value of øSI for the Cr(CO)3
“supercarbonyl” moiety to be -2124 × 10-36 m3/molecule.
Using his technique, this can be determined experimentally from
the difference in chemical shift ∆δ(Me) of the same methyl
group in isomers 9 and 10, and the difference in the factors ∆f,
where f ) (3 cos2 θ - 1)/r3 using
The parameters r and θ were obtained from PCMODEL17
minimized structures, which in our experience27 gives rather
good geometries in both the dihydropyrenes and cyclophanes,
agreeing well with X-ray data where available. Whereas the
values of r and θ for the ring hydrogens are unambiguous, those
for the methyl groups, which can rotate, are not. We tried
averaging the (3 cos2 θ - 1)/r3 term (which we will call the
factor, f), and then determining σ, averaging θ and r first to
determine f and then σ, and finally considering the methyl
protons to be at a point at the center of the circle that they sweep
out on rotation. This latter method gave the most consistent
results, though it should be stressed that the differences were
rather small; for example, for the methyl groups of compound
9, values for σ for the proximal methyl group were found to be
-0.077, -0.090 and, -0.104 ppm, respectively, and for the
distal methyl group +0.119, +0.117, and +0.118 ppm, respec-
tively (the proximal methyl group, Mep, is on the same side as
the Cr(CO)3), and the distal methyl group, Med , is on the
opposite side).
The values of σ calculated using eq 2 for each proton in
compounds 9 and 10 are given in Table 2, together with the
experimental chemical shifts. If compounds 9 and 10 had
exactly the same geometry, then the calculated value of σ (and
the experimental chemical shift) for each ring proton in
compound 9 would be identical with the corresponding proton
in 10. PCMODEL calculations indicate that the geometries are
not exactly the same (see below), and thus slightly different σ
values are calculated and chemical shifts observed.
øSI ) 12π x 10-36∆δ(Me)/∆f
The calculated chemical shifts for the two methyl groups of
9 after removal of the (de)shielding due to the Cr(CO)3
anisotropy are -0.975 (Mep) and -0.859 (Med). This difference
is small (∆δ ) 0.116 ppm), suggesting that the skeleton of 9 is
not much different from that in 2 where ∆δ(Me) ) 0.008 ppm.
The average chemical shift of the two methyl groups of 9 in
the absence of the Cr(CO)3 anisotropy is thus -0.92 ((0.06)
ppm, and this is the value we will use in the aromaticity
calculations below. Clearly the Cr(CO)3 group does not have
a large through space (anisotropic) effect on the methyl protons
or indeed on any of the more distant protons (see Table 2).
Even for the isolated proton adjacent to the complexed ring,
(H-6), σ is only calculated to be +0.25 ppm! Thus the change
in chemical shift on complexation, δ(9) - δ(2) of about 0.7
ppm, is not a through-space anisotropy effect of the Cr(CO)3
group. Nor is it a through-space anisotropy effect of the benzene
ring. The Memory equation28 can be used to calculate the
deshielding of the methyl protons in 2 caused by the field of
Thus for 9, using Med of 9 and Mep of 10, δ(Med-Mep) )
-0.170 ppm and fd - fp ) 0.0026342, giving ø ) -2432 ×
10-36 m3/molecule, in good agreement with McGlinchey’s value.
This method relies on the fact that the geometry of the molecule
does not change between isomers. This is more true for Med
of 9 and Mep of 10 than for Mep of 9 and Med of 10, because
of the slight bending of the skeleton as discussed above. The
greater twist occurs further from the complexed benzene ring,
and hence it is best to use the methyl groups closest to the
complexed ring. In a similar fashion, the corresponding methyl
groups of the pairs 11/12, 26/27, and 28/29 can be used to
determine the øSI values for the -Ru2+(HMB), -RuCp*, and
-Mn(CO)3 fragments, respectively. These are reported in Table
3.
As explained above, under Table 2, to use the chemical shifts
of the internal methyl protons and the most distant protons (Hd,
(29) Mailvaganam, B.; Perrier, R. E.; Sayer, B. G.; McCarry, B. E.; Bell,
R. A.; McGlinchey, M. J. J. Organomet. Chem. 1988, 354, 325-340;
Cambie, R. C.; Clark, G. R.; Gourdie, A. C.; Rutlidge, P. S.; Woodgate, P.
D. J. Organomet. Chem. 1985, 297, 177-184.
(27) Mitchell, R. H. AdV. Theor. Interesting Mol. 1989, 1, 135-199.
(28) Memory, J. D. J. Magn. Reson. 1977, 73, 241-244.

1790 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Mitchell et al.
Table 3. Experimental McGlinchey26 Diamagnetic Anisotropy
Values, øSI (in Units 10-36 m3/Molecule)
Table 5. Coupling Constant Ratios for the Compounds Studied
compd Jb (Hz) Ja (Hz)
Jb/Ja
δ(Meave
)
[Jb/Ja]calc (eq 5)
a
molecular fragment
øSI
9
11
26
28
8.97
6.30
1.42
∼1.38
1.42
-0.92
-0.72
-0.62
-0.76
1.42
1.45
1.46
1.44
b
-Cr(CO)3
-2432
-2368
-478e
-836
∼9.1
∼6.6
c
-Mn(CO)3
8.95
8.72
6.30
6.41
-Ru(Cp*)d
1.36
-Ru2+(HMB)d
a McGlinchey26 suggests (10% error. b McGlinchey26 value )
-2124, and center of anisotropy taken as 3.3 Å from Cr along the C3
axis. c Center of anisotropy taken as 3.2 Å from Mn along the C3 axis.
d Center of anisotropy taken as the metal atom. e A øcgs value for the
parent RuCp2 has been reported,20b,30a which when converted gives a
øSI value of about -325 for Ru(Cp).
Table 6. RA (RBFA) Compared to Benzene of Several Metal
Complexes
complex
compd δ(Meave) δ(Hd) RA (eq 6) RA (eq 7)
benzene-Cr(CO)3, 37
9
-0.92 6.87
-0.72 6.73
-0.62 6.57
-0.76 6.74
1.27
1.34
1.38
1.33
1.25
1.39
1.55
1.38
benzene-Ru2+(HMB), 38 11
Cp-Ru-Cp*, 39
Cp-Mn(CO)3, 40
26
28
Table 4. Anisotropy Free Chemical Shift Values, δcalc, for Use in
Relative Aromaticity Calculations
compd
Mep
Med
Meave
Hd
Hd[eq 4]
the macrocyclic ring, reducing the ring current and hence
reducing the shielding of the internal methyl groups. There is
no substantial through-space anisotropy effect. The relative
bond fixing ability of the various complexes can thus be
compared to that of benzene (from the data of 2), and this we
equate to relative aromaticities.
9
11
26
28
-0.975
-0.667
-0.557
-0.729
-0.859
-0.780
-0.687
-0.782
-0.92 ( 0.06
-0.72 ( 0.06
-0.62 ( 0.07
-0.76 ( 0.03
6.87
6.75
6.57
6.74
6.87
6.79
6.75
6.81
Figure 1) to determine the relative aromaticities of these
fragments, the through-space anisotropies of the metal com-
plexing fragment must first be removed. With the øSI values
in hand, this is now possible using
Relative Bond Fixing Ability (RBFA) [Relative
“Aromaticity” (RA)]
Having now ruled out anisotropy effects, and having previ-
ously established1a that ring current changes are correlated to
bond delocalization changes in systems such as 4, we can now
estimate the RA ()RBFA) of the complexed species studied.
This can be done1a using eqs 6 for Meave and 7 for Hd:
δcalc ) δexp + σ
(3)
where σ is the shielding (ppm) calculated using eq 1. Calculated
anisotropy free values of δ for Mep and Med and Hd are thus
shown in Table 4. Also in Table 4 are shown the average value
for Med and Mep, Meave, together with the value of Hd calculated
using1a
RA ) [4.25 - δ(Meave)]/2.63
(6)
The results are shown in Table 6. As can be seen, both
δ(Meave) ) 17.515 - 2.685δ(Hd)
(4)
RA ) [8.11 - δ(Hd)]/0.99
(7)
This latter value is important in the sense that we have
previously shown1a that the chemical shift of the internal methyl
protons is linearly related to that of the external proton Hd, and
thus if the value of Hd calculated from eq 4 differs by much
from the experimental value of Hd, an indication that a nonring
current effect is operating is obtained. As δ(Me) changes from
-4.25 to +1.00, δ(Hd) changes from 8.11 to 6.15.
equations yield the same order: Cp-Ru-Cp* > benzene-Ru2+
(HMB) > Cp-Mn(CO)3 > benzene-Cr(CO)3 > benzene.
-
Thus these all appear to have a greater bond fixing ability
of the annulene ring than does benzene itself, and similarly
they also resist bond fixation in their own rings (by the
annulene ring) to a greater extent than does benzene (see
below) and this behavior we equate to mean “effectively”
that they have 30-40% greater aromaticity than benzene
itself!
As can be seen, excellent agreement is found for Hd , in
support of the fact that the major change is one of a changed
ring current, which we will now show is caused by a difference
in bond delocalization. For a series of annelated annulenes of
type 4, we have shown that there is a linear relationship between
the chemical shift of the internal methyl protons and the ratio
Are the Complexes 37-40 Really Aromatic?
Ferrocene and related species such as 39 are generally
considered to be aromatic. However, benzene-chromiumtri-
carbonyl, 37, stirs up more controversy. Bursten and Fenske,30b
on the basis of the ease of Friedel-Crafts acylation reactions
claim that 37 is “less aromatic” than both benzene and
cyclobutadiene-irontricarbonyl, and that the latter is “more
aromatic” than ferrocene or CpMn(CO)3, 40. Very recently in
this journal,31 Simion and Sorensen claim that 37 is actually
antiaromatic on the basis of magnetic susceptibility exhaltation
calculations. Fortunately, Gu¨nther has made extensive studies2f,g
on the effects of fusing both 4nπ- and (4n+2)π-systems to
benzene. When an aromatic system (Ar) is fused to benzene,
3
of adjacent J coupling constants, eq 5:1a
δ(Meave) ) 7.99(Jb/Ja) - 12.29
(5)
where Jb ) 3J(H-1,2) and Ja ) 3J(H-2,3) (see Figure 1). If this
relationship holds for the series of compounds in Table 4, we
can be reasonably certain that, as in the case of the benzenoid
annelated annulenes 4, it is the change in bond fixation that is
causing the change in ring current. Thus Table 5 shows the
experimental Jb/Ja values and calculated ones based on eq 5.
Clearly the annulene portion of the complexed macrocycles does
show the expected amount of bond fixation consistent with the
strength of the observed ring current.
(30) (a) Ritchie, G. L. D.; Cooper, M. K.; Calvert, R. L.; Dennis, G. R.;
Phillips, L.; Vrbancich, J. J. Am. Chem. Soc. 1983, 105, 5215-5219. Mulay,
L. N.; Mulay, I. L. Anal. Chem. 1966, 38, 501-513. (b) Bursten, B. E.;
Fenske, R. F. Inorg. Chem. 1979, 18, 1760-1765.
(31) Simion, D. V.; Sorensen, T. S. J. Am. Chem. Soc. 1996, 118, 7345-
7352.
We believe that the above analysis clearly indicates that in a
metal complexed benzannulene such as 9, the major effect that
the metal has on the system is to change the bond fixation in

[14]Annulene Fused Organometallics
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1791
3
as in 41, the J coupling constant Jb > Ja. However, when an
substantially less localized than the uncomplexed ring in 6 or
2. Unfortunately, the X-ray structure of 9 would not refine
sufficiently to give accurate bond lengths. We hope in the future
to overcome this at low temperature. The cyclopentadienyl
coupling constants in the complexes 26 and 28 are also
approximately equal, indicating that these rings are not much
localized at all.
antiaromatic system (AAr) is fused to benzene, as in 42, then
the reverse is true, Jb < Ja.
Clearly, in looking at the sum of all the above results, the
metal complexes 37-40 cause substantial bond fixation and
reduce the ring current when fused to an annulene, more so
than does benzene, and themselVes resist bond fixation, more
so than does benzene. Does this make them more “aromatic”
than benzene? In the absence of some other effect which is
causing these results, we would argue “yes”.
Thus for naphthalene, Jb ) 8.28 Hz and Ja ) 6.85 Hz, while
for biphenylene Jb ) 6.80 Hz and Ja ) 8.24 Hz. He has
generalized this in terms of an alternance parameter, Q, which
is the ratio of the π-SCF bond orders Pb/Pa, which ranges from
0.78 toward 1 (at large n) for 4nπ systems and 1.26 toward 1
(at large n) for (4n+2)π systems. The bond orders, Px are
related linearly to the coupling constant Jx by eq 8:
The relative bond localization energies (RBLE) compared to
benzene of the complexes 37-40 can be estimated from eq 9:1a
Px(SCF) ) 0.1043Jx - 0.120
(8)
RA(Me) ) 0.978(RBLE) + 0.04
(9)
These values are not strongly changed by charge in the fused
ring; thus the benzocycloheptatrienyl anion, an 8π AAr system,
shows a Q value of 0.89, while the aromatic 6π systems,
benzocyclopentadienyl anion and benzotropylium cation, show
Q values of 1.22 and 1.23, respectively.2g Clearly for the
compounds studied in this work (see Table 5), in each case for
the analogous bonds, Jb is substantially greater than Ja, indicating
that unequivocally the fused systems 37-40 are all behaving
in the same way that strongly aromatic systems do.32 Since
we have shown2d that eq 6 holds reasonably well for the
[a]-fused dihydropyrenes, it, together with the data in Table 5,
can be used to calculate Q values for 37 (1.52), 39 (1.46), and
40 (1.47). These can be compared with the value for benzene
itself (1.39) derived from the coupling constants of the uncom-
plexed benzannulene 2, or 1.36 from Gu¨nther’s theoretical bond
orders.2f Since a consequence of fusing two aromatic rings is
that both rings show bond fixation relative to the parent rings,
comparison of two aromatic systems with a common probe as
in 4 has as a consequence the fact that the stronger aromatic
ring will bond fix the probe ring the most. Clearly from either
the coupling constant or the alternance parameter data, the
complexes bond fix the parent annulene more than does benzene,
in support of our contention that they are “stronger aromatics”.
A second consequence is that in a series involving the same
probe molecule, the most aromatic fused ring will itself be the
least bond fixed.1 Thus above, if the metal complexed species
are indeed more aromatic or more bond fixing than benzene,
then the benzene ring in 2 should be more localized than the
and decrease in the order 39 (1.39) > 38 (1.35) > 40 (1.34) >
37 (1.28) > benzene (1.00).
Since the BLE for benzene is 2034 to 2635 kcal/mol, our results
would suggest that the complexes are harder to bond localize
than benzene by some 6-10 kcal/mol. The question then
becomes why?
Siegel36 has suggested that the results for 9 might be explained
if the Cr tripod adopts a specific ground-state conformation,
which favors one of the resonance forms of the dihydropyrene.
This would suggest that in 9, this specific rotomer is favored
by at least 6 kcal/mol. That would seem somewhat high based
on rotational barrier experimental results for nonalkyl-substituted
arenes and fused arenes (except biphenylene).37 The (MM2+π)17
rotational energy barriers in these complexes, are indeed
predicted to be small, but, as Siegel remarks,36 this method gives
rather low values. Moreover, it is not obvious how this
argument would work for the other ligands. It is reasonably
easy to see why the complexed species might resist bond fixation
more than benzene itself, however, since to obtain good orbital
overlap with the metal orbitals, more or less bond equal arenes
are preferred. For the Cr(CO)3 complexes, benzene is com-
plexed in preference to naphthalene, the end rings in phenan-
threne in preference to the center ring.37a Is this aromaticity?
Recently the driving force behind the so-called aromaticity of
benzene has been questioned.38 Is it σ or π driven? Perhaps it
does not matter, because the end result is a “bond-equal”
benzene. Since our complexes 37-40 cause more bond fixation
when fused to dihydropyrene than does benzene, and themselves
prefer to remain more “bond-equal” than does even benzene
under similar circumstances, we would argue this is aromatic,
even though we are not yet sure of the driving force. We hope
that these results will stimulate theoreticians into probing this
matter further.
3
corresponding aromatic ring in 9, 11, 26, or 28. While the J
coupling constants for the benzene ring protons, H-7,8, H-8,9,
and H-9,10 in 2 cannot be separated at 360 MHz, the
corresponding ones in the [e]-fused isomer 6 can and are 8.26
and 6.93 Hz, indicating about the same degree of bond
3
localization as in naphthalene. The benzene ring J coupling
(34) Dewar, M. J. S.; De Llano, C. J. Am. Chem. Soc. 1969, 91, 789-
795.
constants in the complex 9, however, are much more equal and
are 6.9, 6.5, and 6.4 Hz, similar to those (6.55 and 5.5 Hz) in
the complexed ring of tricarbonylchromiumnaphthalene.33 This
indicates to us that the complexed benzene ring in 9 is
(35) Paldus, J.; Li, X. Isr. J. Chem. 1991, 31, 351-362.
(36) Nambu, M.; Hardcastle, K.; Baldridge, K. K.; Siegel, J. S. J. Am.
Chem. Soc. 1992, 114, 369-371.
(37) (a) Muetterties, E. L.; Bleeke, J. R.; Wucherer, E. J.; Albright, T.
A. Chem. ReV. 1982, 82, 499-525. (b) Albright, T. A. Acc. Chem. Res.
1982, 15, 149-155.
(38) Gobbi, A.; Yamaguchi, Y.; Frenking, G.; Schaefer, H. F. Chem.
Phys. Lett. 1995, 244, 27-31; Glendening, E. D.; Faust, R.; Streitweiser,
A.; Vollhardt, K. P. C.; Weinhold, F. J. Am. Chem. Soc. 1993, 115, 10952-
10957. Hiberty, P. C.; Ohanessian, G.; Shaik, S. S.; Flament, J. P. Pure
Appl. Chem. 1993, 65, 35-45. Jug, K.; Koster, A. M. J. Am. Chem. Soc.
1990, 112, 6772-6777. Shaik, S. S.; Hiberty, P. C.; Lefour, J. M.;
Ohanessian, G. J. Am. Chem. Soc. 1987, 109, 363-374.
(32) (a) When an antiaromatic system is less planar and cyclic conjugation
is reduced, the effect on the benzene ring is less, and the coupling constants
are only slightly perturbed from those in benzene,32b unlike the case of
biphenylene. However, the coupling constant ratio, Jb/Ja is never >1.2 when
an antiaromatic system is fused to benzene. (b) Scott, L. T.; Kirms, M. A.;
Gu¨nther, H.; Puttkamer, H. J. Am. Chem. Soc. 1983, 105, 1372-1373.
Kirms, M. A. Ph. D. Thesis, University of Nevada, Reno, 1982; p 107.
(33) Deubzer, B.; Fritz, H. P.; Kreiter, C. G.; Ofele, K. J. Organomet.
Chem. 1967, 7, 289-299.

1792 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Mitchell et al.
space group P1, with a ) 14.516 Å, b ) 9.062 Å, c ) 7.772 Å. An
ORTEP diagram is provided in the Supporting Information to show
the skeleton. We intend to attempt to obtain a low-temperature structure
in the future.
The minor isomer, 10, could not be obtained pure, but its 1H NMR
data was obtained by subtraction: (360 MHz) δ 7.56 (d, J ) 6.69 Hz,
H-11), 7.36 and 7.29 (AB, J ) 9.19 Hz, H-4,5), 7.24 (s, H-6), 7.24 (d,
J ) ∼9 Hz, H-1), 7.10 (d, J ) 6.39 Hz, H-3), 7.01 (d, H-12), 6.93 (dd,
H-2), 6.49 (d, J ) 6.9 Hz, H-10), 5.83 (dd, J ) 6.7, 1.2 Hz, H-7), 5.65
(ddd, H-8), 5.45 (ddd, H-9), -0.807 and -1.158 (s, -CH3).
Bis-Complexes
In our preliminary report25 on the bis-complex 32, we stated
that only 6% of the ring current of the [14]annulene ring
remains. The current work supports that figure, since from
Table 2, for compound 9, σ’s for Mep and Med are almost equal
but opposite in sign. Since in 32 Med for one Cr(CO)3 group
is Mep for the other Cr(CO)3, each Me group suffers a shielding
and a deshielding of equal magnitude, leaving the observed
chemical shift almost identical with a calculated anisotropy free
shift. Our ring current estimate of 6% thus stands. Clearly
fusion of two molecules of 37 to dihydropyrene 1 almost, but
not completely, kills the annulene ring current. Certainly they
do so more than do two benzene rings (18% residual current).25
Ruthenium Complexes 11 and 12. A solution of benzo[a]-
dimethyldihydropyrene 21a (74 mg, 0.26 mmol) in anhydrous propylene
carbonate (5 mL) was added to a mixture of [Ru(HMB)Cl2]27 (88 mg,
0.13 mmol) and AgBF4 (204 mg, 1.05 mmol) in anhydrous propylene
carbonate (5 mL) at 20 °C and was stirred for 12 h under argon. The
reaction mixture was then poured into diethyl ether (250 mL) with
stirring and the product was collected by filtration. The residue was
washed with ether, dissolved in a minimum of acetone, and then filtered.
The filtrate was added dropwise to ether with vigorous stirring to give
dark red crystals of product, 11 and 12, 100 mg (53%) as a (3:2)
mixture, which could not be separated by fractional recrystallization.
1H NMR (360 MHz, CD3CN) [assigned by COSY] δ 8.17 (d, J ) 6.7
Hz, H-11 (both isomers, δ 8.50 and 8.43 in (CD3)2CO) (NOESY to
H-10), 7.71 and 7.70 (d, J ) 9.1 Hz, H-5), 7.63 and 7.57 (d, J ) 9.1
Hz, H-4), 7.49-7.45 (m, H-1, H-10), 7.40-35 (m, H-6, H-10′), 7.32-
7.28 (m, H-9), 7.26 and 7.24 (d, J ) 6.7 Hz, H-12), 7.10 (d, J ) ∼6.6
Hz, H-3), 6.96 (d, J ) 6.1 Hz, H-7), 6.63-6.55 (m, H-2), 6.51 and
6.47 (t, J ) 6.0 Hz, H-8), 2.31 and 2.21 (s, HMB-CH3), -0.632,
-0.966 and -0.667, -0.985 (s, DHP-CH3, 3:2 ratio before purifica-
tion. In CDCl3: δ -0.517, -0.864 and -0.558, -0.891. In (CD3)2-
1
In the H NMR spectrum of the nonsymmetrical isomer 34, a
3
9.5 Hz J coupling for the 6-7 annulene protons can be seen,
larger than any of the Jb values in Table 5, supporting the very
extensive bond fixation in this molecule.
Conclusions
We have described the synthesis of 13 metal complexed
annelated trans-dimethyldihydropyrenes (bridged [14]annulenes)
and have presented their proton magnetic resonance spectra in
detail. Analysis of the latter with respect to the change in ring
current in the macrocyclic ring, and in the change in coupling
constants, reveals that in each metal complex the [14]annulene
is bond localized to a greater extent than in the corresponding
uncomplexed annulene. This effect is not local, nor is a through-
space anisotropy effect, but is a real change in the delocalization
of the whole π-system. The relative bond fixing ability of
several organometallic moieties was thus estimated by their
effects on the annulene ring and this ability was discussed in
terms of the aromaticities of the systems involved. Not all
authors would agree with our conclusions concerning aroma-
ticity, but we hope that this work, the first extensive investigation
of metal complexed annulenes, will provoke theoreticians to
explore this matter further!
CO: δ -0.484, -0.850 and -0.499, -0.911); UV (acetone) λmax (ꢀmax
)
nm 341 (37 500), 400 (33 500), 615sh (9800); FAB(LSI) MS (m-NBA
matrix) 633 (M - BF4), 546 (M - 2BF4), 531 (M - 2BF4, -CH3),
516 (M - 2BF4, -2CH3) [dihydropyrenes normally show strong peaks
corresponding to loss of one and then both internal methyl groups].
Ethyl trans-3-(2′-trans-10b,10c-Dimethyl-10b,10c-dihydropyre-
nyl)propenate (14). Triethyl phosphonoacetate (1.32 mL, 6.64 mmol)
was added to a suspension of NaH (162 mg, 6.76 mmol) in dry THF
(35 mL) at 0 °C under argon. After hydrogen gas evolution had
subsided, 2-formyldimethyldihydropyrene 138 (1.90 g, 6.15 mmol) in
dry THF (80 mL) was added dropwise. The reaction mixture was
allowed to warm to 20 °C, and then was refluxed for 1 h. After cooling,
saturated aq NH4Cl and dichloromethane were then added, and the
organic layer was washed, dried, and evaporated to give a maroon solid.
This was chromatographed over SiGel using dichloromethane-PE (1:
20) as eluant and gave 2.0 g (98%) of 14 as dark red crystals from
ether-pentane, mp 146-147 °C: 1H NMR (250 MHz) δ 8.68 (s,
H-1′,3′), 8.59 and 8.50 (AB, J ) 7.9 Hz, H-4′,5′,9′,10′), 8.46 (d, J )
7.8 Hz, H-6′,8′), 8.30 (d, J ) 15.8 Hz, H-3), 8.03 (t, H-7′), 6.91 (d,
H-2), 4.35 (q, J ) 7.1 Hz, -OCH2-CH3), 1.40 (t, -CH2CH3), -3.85
and -3.87 (s, -CH3); 13C NMR (62.9 MHz) δ 167.3, 146.2, 139.0,
136.4, 128.2, 125.8, 124.8, 124.2, 124.0, 123.5, 117.9, 60.4, 31.1, 30.3,
Experimental Section
The same general procedures were used as are given in ref 1a.
[7,8,9,10,10a,10b-η6]-trans-12b,12c-Dimethyl-12b,12c-dihydrobenzo-
[a]pyrene chromium(0)tricarbonyl, (9, 10). A mixture of the benzo-
[a]dimethyldihydropyrene 21a (40 mg, 0.14 mmol), naphthalenetricar-
bonylchromium39 (60 mg, 0.22 mmol), and THF (70 µL, 0.86 mmol)
in diethyl ether (2 mL) was stirred magnetically in a sealed vial at
50-60 °C for 14 h. The mixture was cooled, preabsorbed on silica
gel, and chromatographed on SiGel using dichloromethane-PE (1:4)
as eluant. After a small amount of starting material, the complexes 9
and 10 were eluted as a red solid, 40 mg (60%) as a 3:1 ratio (by
NMR). Recrystallization from dichloromethane yielded mostly isomer
9, as deep red crystals, mp 189-190 °C: 1H NMR (360 MHz) δ 7.59
(d, J ) 6.53 Hz, 1, H-11), 7.28 (s, 2, H-4,5), 7.24 (s, 1, H-6), 7.18 (d,
J ) 8.97 Hz, 1, H-1), 7.02 (d, J ) 6.30 Hz, 1, H-3), 7.00 (d, 1, H-12),
6.88 (dd, 1, H-2), 6.57 (d, J ) 6.9 Hz, 1, H-10), 5.94 (dd, J ) 6.5, 0.9
Hz, 1, H-7), 5.61 (dt, J ) 6.4, 1.1 Hz, 1, H-8), 5.50 (ddd, 1, H-9),
-0.871 and -0.977 (s, 3 each, -CH3), in benzene-d6 H-4,5 appear as
a doublet, J ) 9.24 Hz; 13C NMR (62.9 MHz) δ 232.6, 140.7, 140.3,
139.1, 130.5, 128.2, 126.6, 126.0, 125.2, 123.3, 121.1, 120.8, 119.8,
101.5, 98.2, 91.8, 91.7, 90.7, 88.0, 37.7, 36.9, 20.0, 18.4; IR 1950,
1860 cm-1; UV λmax (ꢀmax) nm 220 (34 400), 336 (39 600), 435 (20 600);
CI MS m/z 419 (MH+). Anal. Calcd for C25H18CrO3: C, 71.76; H,
4.34. Found: C, 71.60; H, 4.39. An X-ray structure determination
was carried out on this isomer, which clearly indicated the carbon
skeleton, but would not complete using a least-squares refinement to
give the hydrogen atoms (R ) 0.066). The crystal system was triclinic,
15.1, 14.8, 14.4; IR 1698, 1613, 1270, 1174, 1149 cm-1; UV λmax (ꢀmax
)
nm 348 (79 400), 378 (21 400), 402 (32 400), 522 (20 900), 608 (310),
669 (200); EI MS m/z 330 (M+). Anal. Calcd for C23H22O2: C, 83.60;
H, 6.72. Found: C, 83.43; H, 6.78.
Ethyl 3-(2′-trans-10b,10c-Dimethyl-10b,10c-dihydropyrenyl)pro-
panoate (15). A solution of the ester 14 (2.03 g, 6.15 mmol) in ethyl
acetate (100 mL) was stirred under hydrogen in the presence of Pd/C
(10%, 5 mg) at 20 °C for 2.5 h, when 1 equiv of H2 had been consumed.
The mixture was then carefully filtered (argon), washing the catalyst
with dichloromethane, and then the filtrate was evaporated. The green
gummy residue was chromatographed over SiGel using pentane as
eluant, and gave 15 as a dark green gum, 1.75 g (86%), which would
not crystallize: 1H NMR (250 MHz) δ 8.61 and 8.56 (AB, J ) 7.8
Hz, H-4′,5′,9′,10′), 8.55 (d, J ) 7.7 Hz, H-6′,8′), 8.45 (s, H-1′,3′), 8.04
(t, H-7′), 4.13 (q, J ) 7.1 Hz, -OCH2CH3), 3.66 (t, J ) 7.7 Hz, H-2),
2.99 (t, H-3), 1.17 (t, -CH2CH3), -4.159 and -4.164 (s, -CH3); 13
C
NMR (62.9 MHz) δ 172.6, 136.9, 136.0, 135.6, 123.6, 123.4 (×2),
123.1, 122.3, 60.2, 36.6, 32.9, 29.7, 29.5, 14.1(×2), 13.7; IR 1730 cm-1
;
(39) Desobry, V.; Kundig, E. P. HelV. Chim. Acta 1981, 64, 1288-1297.
UV λmax (ꢀmax) nm 338 (95 500), 355 (16 600), 377 (40 800), 455

[14]Annulene Fused Organometallics
J. Am. Chem. Soc., Vol. 120, No. 8, 1998 1793
(6800), 470 (7100), 529 (85), 594 (89), 625 (110), 642 (170); EI MS
m/z 332 (M+). HRMS Calcd for C23H24O2: 332.178. Found: 332.168.
3-(2′-trans-10b,10c-Dimethyl-10b,10c-dihydropyrenyl)propan-
oic Acid (16). The ester 15 (2.00 g, 6.02 mmol) was refluxed in aq 1
M NaOH (75 mL) for 15 h, cooled to 20 °C, and acidified carefully
with concentrated aq HCl; the resulting solid was collected by vacuum
filtration. The solid was chromatographed over SiGel using ether as
eluant to give 1.65 g (90%) of acid, as short dark green needles from
cyclohexane-PE, mp 165-166 °C: 1H NMR (250 MHz) δ 10.98 (bs,
-COOH), 8.59 and 8.53 (AB, J ) 7.7 Hz, H-4′,5′,9′,10′), 8.55 (d, J )
7.7 Hz, H-6′,8′), 8.44 (s, H-1′,3′), 8.05 (t, H-7′), 3.66 (t, J ) 7.9 Hz,
H-3), 3.06 (t, H-2), -4.19 and -4.20 (s, -CH3); 13C NMR (62.9 MHz)
δ 179.1, 137.1, 136.3, 135.2, 123.9, 123.6, 123.4, 122.6 (×2), 36.4,
32.6, 29.9, 29.7, 14.2, 13.9; IR 3200-2400, 1683, 865, 811 cm-1; UV
λmax (ꢀmax) nm 338 (53 700), 351 (13 200), 377 (19 500), 455 (3300),
470 (4700), 529 (120), 629 (180), 642 (300); EI MS m/z 304 (M+).
Anal. Calcd for C21H20O2: C, 82.86; H, 6.62. Found: C, 82.53; H,
6.38.
immediately, then the alkenes 19 and 20 were obtained in a 2:3 ratio.
These could not be separated by chromatography, but the H NMR
spectrum of 20 could be obtained by subtraction: (250 MHz) δ 8.78-
8.40 (m, 8H), 7.42 (d, J ) 5.6 Hz, H-7), 6.94 (dt, J ) 5.6, 2.1 Hz,
H-8), 4.26-4.21 (m, H-9,9′), -4.09 (s, 6H, -CH3).
1
Ferrocene by Photolysis of 25. Potassium hydride (60 mg, 1.5
mmol) was added to cyclopentadiene (100 mg, 1.5 mmol) in dry DMF
(6 mL) under argon, and was stirred for 15 min, and then complex
2516 (450 mg, 1.2 mmol) was added. The reaction mixture was then
irradiated with a 150-W tungsten garden flood lamp for 5 h, and then
was extracted with PE (3 × 20 mL). The organic extracts were
concentrated and gave 170 mg (80%) of orange crystals of ferrocene,
1
identical with a commercial sample (TLC, mp, H NMR).
Indenylferrocene 24. Potassium hydride (45 mg, 1.1 mmol) was
added to indene (130 mg, 1.1 mmol) in dry DMF (6 mL) under argon,
and the mixture was stirred for 30 min, when complex 2516 (370 mg,
1.0 mmol) was added. The reaction mixture was then irradiated with
a 150-W tungsten garden flood lamp for 3 h, and then was extracted
with PE (4 × 30 mL). The organic extracts were concentrated and
gave a red-violet solid which was chromatographed over basic alumina
using PE as eluant. Eluted first was ferrocene (10 mg). Eluted second
was 163 mg (69%) of indenylferrocene 24, as red-violet crystals from
PE, mp 75 °C, with identical properties as those described.40
trans-11b,11c-Dimethyl-7,8,11b,11c-tetrahydro-9-oxo-9H-cyclo-
penta[a]pyrene (18). Excess oxalyl chloride (250 mg, 2 mmol) was
added to the acid 16 (200 mg, 0.66 mmol) in dichloromethane (15
mL) and the mixture was stirred at 20 °C for 6 h. Removal of solvent
and excess oxalyl chloride under vacuum yielded a dark green solid; a
sample gave an IR (KBr disk) which indicated conversion to acid
chloride 17 by the peak at 1802 cm-1 and disappearance of the peak at
1683 cm-1. The residue was dissolved in dry dichloromethane (100
mL) and BF3‚Et2O (0.8 mmol) was added; the mixture was stirred for
9 h. Aqueous potassium carbonate (10%) was then added, and the
organic layer was collected, dried, and evaporated. The dark green
residue was chromatographed on SiGel using dichloromethane as eluant
and gave 120 mg (63%) of ketone 18 as dark green plates from
[6a,7,8,9,9a-µ5]-trans-11b,11c-Dimethyl-11b,11c-dihydrocyclopen-
ta[a]pyrene-pentamethylcyclopentadienylruthenium (26, 27). Me-
thyllithium (0.15 mL, 1.4 M in ether) was added to alkene 19 (50 mg,
0.19 mmol) in dry deaerated THF (2 mL) at -78 °C under argon, and
20,21
then the mixture was stirred at 20 °C for 15 min, when (Cp*RuCl2)n
(58 mg, 0.19 mmol) was added and the reaction stirred for a further
30 min. The reaction was then heated to 40 °C and stirred for a further
15 min. After cooling, the solvent was evaporated and the residue
was dissolved in cyclohexane and chromatographed over basic alumina
using deaerated cyclohexane as eluant to give 65 mg (70%) of rose-
violet solid, a mixture of isomers 26 and 27 in a 3.7:1 ratio, which
could not be separated by recrystallization from ethanol or sublimation
at 235 °C at 0.1 mmHg. The compounds are air sensitive. Because
of the unequal ratio, the NMR peaks of 26 could be assigned: 1H NMR
(360 MHz) δ 6.95 and 6.90 (AB, J ) 9.09 Hz, H-5,4), 6.94 and 6.67
(AB, J ) 6.41 Hz, H-10,11), 6.88 (d, J ) 8.95 Hz, H-1), 6.72 (s, H-6),
6.68 (d, J ) 6.30 Hz, H-3), 6.57 (dd, H-2), 4.99 (d, J ) 2.45 Hz, H-9),
4.70 (d, J ) 2.45 Hz, H-7), 4.39 (t, H-8), 1.68 (s, 15H, Cp-CH3), -0.548
and -0.691 (s, -CH3); 13C NMR (62.9 MHz) δ 139.2, 137.3, 135.8,
135.4, 126.6, 126.0, 123.9, 122.0, 121.4, 121.1, 120.5, 113.7, 85.2
(Cp*), 85.0, 83.9, 74.3, 71.6, 69.3, 39.7, 38.3, 20.1, 18.9, 11.5 (Cp-
CH3). The 1H NMR internal methyl protons of the minor isomer
appeared at δ -0.473 and -0.820.
1
cyclohexane, mp 210-211 °C; H NMR (250 MHz) δ 9.66 and 8.51
(AB, J ) 7.3 Hz, H-10,11), 8.61 and 8.49 (AB, J ) 8.1 Hz, H-5,4),
8.50 (d, J ) 7.7 Hz, H-1), 8.47 (d, J ) 7.7 Hz, H-3), 8.45 (s, H-6),
7.95 (t, H-2), 3.79-3.68 (m, 2, H-7), 3.09-3.01 (m, 2, H-8), -3.72
and -3.73 (s, -CH3); 13C NMR (62.9 MHz) δ 208.9, 153.6, 142.2,
136.9, 136.8, 130.0, 128.1, 126.9, 126.2 (×2), 126.0, 123.7, 123.3,
120.3, 119.9, 37.3, 32.3, 31.0, 27.5, 14.8, 14.4; IR 1670 cm-1; UV
λ
max (ꢀmax) nm 367 (20 000), 471 (730), 558 (75), 594 (75), 611 (110),
619 (95), 679 (550); EI MS m/z 286 (M+). Anal. Calcd for C21H18O:
C, 88.08; H, 6.34. Found: C, 87.71; H, 6.43.
trans-11b,11c-Dimethyl-11b,11c-dihydro-7H-cyclopenta[a]py-
rene (19). The ketone 18 (100 mg, 0.35 mmol) and LiAlH4 (20 mg,
0.5 mmol) were stirred in dry THF (20 mL) for 1.5 h at 20 °C and
then ice-water was added to destroy excess hydride followed by 20%
aq HCl (10 mL) and dichloromethane (30 mL), and the reaction was
then stirred for 1 h. The organic layer was washed, dried, and
evaporated to yield a green solid which was chromatographed quickly
on SiGel using pentane as eluant to yield 77 mg (81%) of alkene 19 as
dark green plates from methanol, mp 114-116 °C: 1H NMR (360
MHz) δ 8.86 and 8.67 (AB, J ) 7.7 Hz, H-10,11), 8.75 (s, H-6), 8.62
and 8.60 (AB, J ) 7.5 Hz, H-5,4), 8.57 (d, J ) 7.6 Hz, H-1,3), 8.07
(t, H-2), 8.00 (ddt, J ) 5.6, 1.9, 0.8 Hz, H-9), 6.99 (dt, J ) 5.7, 2.0
Hz, H-8), 4.07 and 3.98 (AB of t, J ) 24.2 Hz, H-7,7′), -4.15 and
-4.16 (s, -CH3); 13C NMR (90.6 MHz) δ 139.6, 138.3, 136.6, 136.3,
136.0, 135.2, 130.9, 128.6, 123.1, 122.9 (×2), 122.6 (×2), 122.2, 119.7,
118.9, 40.7, 30.4, 30.3, 14.4, 13.8; UV λmax (ꢀmax) nm 356 (37 200),
390 (13 500), 471 (3100), 597 (62), 616 (68), 644 (115), 658 (90); CI
MS m/z 271 (MH+). Anal. Calcd for C21H18: C, 93.28; H, 6.72.
Found: C, 92.77; H, 7.00.
trans-11b,11c-Dimethyl-11b,11c-dihydrocyclopenta[a]pyrene, Ion-
(1-) (5). Potassium hydride (5 mg) was added to a solution of alkene
19 (27 mg, 0.1 mmol) in dry deaerated THF-d8 (0.4 mL) in an argon
filled glovebox. The intense red solution was filtered into an NMR
tube which was previously flushed with argon: 1H NMR (discussed in
detail in ref 1d); 13C NMR (62.9 MHz, THF-d8) δ 137.1, 136.5, 136.0,
131.6, 130.6, 124.6, 124.5, 124.3, 123.8 (×2), 123.4, 123.3, 123.1,
120.6, 119.8, 119.3, 118.1, 31.5, 30.6, 15.5, 14.3; UV (THF-d8) λmax
(ꢀmax) nm 357 (95 500), 391 (55 000), 454 (28 200), 477 (26 900), 593
(2500), 645 (3500), 654 (2900). This solution decomposed if left to
stand at about 20° for 2 h. If wet THF was added to the anion
On the mixture of isomers: UV λmax (ꢀmax) nm 249 (36 300), 358
(55 000), 502 (sh, 600), 538 (sh, 450), 564 (sh, 300); CI MS m/z 507
(MH+, correct Ru isotope pattern). Anal. Calcd for C31H32Ru: C,
73.63; H, 6.38. Found: C, 73.36; H, 6.08.
[6a,7,8,9,9a-µ5]-trans-11b,11c-Dimethyl-11b,11c-dihydrocyclopen-
ta[a]pyrene-tricarbonylmanganese (28, 29). Methyllithium (0.17
mL, 1.4 M in ether) was added to alkene 19 (60 mg, 0.23 mmol) in
dry deaerated THF (2 mL) at -78 °C under argon, and then the mixture
was stirred at 20 °C for 15 min, when manganese pentacarbonyl
bromide23 (62 mg, 0.23 mmol) was added and the reaction stirred for
a further 10 min. The reaction was then heated to 40 °C and stirred
for a further 20 min. After cooling, the solvent was evaporated and
the residue was dissolved in cyclohexane and chromatographed over
basic alumina using deaerated cyclohexane as eluant to give 57 mg
(61%) of complex 28/29 in a 2.5:1 ratio. Recrystallization from
dichloromethane-pentane yielded an analytically pure dark orange-
brown sample, but would not separate the isomers. Because of the
unequal ratio, the NMR peaks of 28 could be assigned: 1H NMR (250
MHz) δ 7.16 and 6.83 (AB, J ) 6.32 Hz, H-10,11), 7.15 (s, H-6),
7.13 and 7.11 (AB, J ) 8.76 Hz, H-5,4), 7.05 (d, J ) 8.72 Hz, H-1),
6.89 (d, J ) 6.31 Hz, H-3), 6.74 (dd, H-2), 5.54 (bs, H-9), 5.09 (bs,
H-7), 5.07 (bs, H-8), -0.728 and -0.784 (s, -CH3); 13C NMR (62.9
MHz) δ 226, 140.7, 139.4, 138.4, 135.8, 127.3, 126.6, 126.0, 124.3,
(40) King, R. B.; Bisnette, M. B. Inorg. Chem. 1964, 3, 796-801.

1794 J. Am. Chem. Soc., Vol. 120, No. 8, 1998
Mitchell et al.
124.1, 120.8, 118.4, 116.5, 85.1, 77.8, 76.2, 74.5, 74.0, 37.9, 37.1, 20.4,
18.9. The 1H NMR internal methyl protons of the minor isomer
appeared at δ -0.670 and -0.976. On the mixture of isomers: IR
2028 and 1935 cm-1; UV λmax (ꢀmax) nm 346 (30 200), 473 (1400),
646 (180); CI MS m/z 409 (MH+). Anal. Calcd for C24H17MnO3: C,
70.58; H, 4.20. Found: C, 70.16; H, 4.24.
d8) δ 234.03, 233.99, 142.1, 141.9, 137.0, 135.0, 129.2, 128.3, 124.5,
123.3, 119.7, 118.4, 103.9, 102.6, 98.7, 98.4, 94.6, 93.7, 92.9, 92.1,
91.5, 91.3, 90.3, 88.7, 41.8, 41.1, 21.7, 20.3; IR 1944 and 1856 cm-1
;
UV (CH2Cl2) λmax (ꢀmax) nm 337 (48 100), 434 (23 800) tail to 630
nm. EI MS m/z 604 (M+); Anal. Calcd for C32H20Cr2O6: C, 63.58;
H, 3.34. Found: C, 63.72; H, 3.69.
Bis(tricarbonylchromium) Complexes of Dibenzannulene 30, (32,
33, 34). A solution of dibenzannulene 301a (200 mg, 0.613 mmol)
and naphthalenetricarbonylchromium (1.0 g, 3.8 mmol) in ether (15
mL) and THF (1 mL) in a heavy screw capped vial was stirred
magnetically at 60 °C for 30 h. After cooling to 20 °C, 15 mg of
isomer 32 directly crystallized from the reaction mixture. The filtrate
was directly preabsorbed onto SiGel and chromatographed using PE-
ethyl acetate (10:1) as eluant. Unchanged 30 and monocomplex 35
were sometimes present in trace amounts and were eluted first; next
eluted was a further 170 mg of the mixture of bis-complexes 32-33-
The third isomer, 33, could not be obtained pure, but was present in
the mother liquors above; its internal methyl protons appeared at δ
0.96 (CDCl3) and 0.99 (THF-d8).
Monoruthenium Complex 36. A solution of dibenzannulene 301a
(25 mg, 0.075 mmol) in anhydrous propylene carbonate (5 mL) was
added to a mixture of [Ru(HMB)Cl2]27 (75 mg, 0.113 mmol) and AgBF4
(110 mg, 0.565 mmol) in anhydrous propylene carbonate (5 mL) and
was stirred for 12 h and then was poured in to ether (200 mL) with
stirring. The crude product was collected by filtration, dissolved in a
minimum of acetone, and filtered again. Ether was slowly diffused
into this acetone solution and 25 mg (45%) of dark red crystals of
complex 36 was formed. This was a 1:1 mixture of the two isomers,
and these could not be separated by fractional crystallization: 1H NMR
(360 MHz, (CD3)2CO, mixture of isomers) δ 8.35 and 8.30 (d, J ) 8.4
Hz, H-12), 8.02 and 7.96 (d, J ) 6.6 Hz, H-11), 7.79 (s, H-7), 7.76-
6.90 (m), 2.14 and 2.05 (s, HMB-CH3), 0.69, 0.29 and 0.58, 0.20 (s,
-CH3); FAB (LSI) MS m/z 683 (small, M - BF4), 596 (M - 2BF4),
581 (M - 2BF4,CH3), 566 (M - 2BF4,2CH3).
1
34. The total yield was 50% and the H NMR spectrum indicated a
total ratio of isomers 12:1:2. The first isomer eluted was 32, which
upon fractional crystallization from cyclohexane yielded pure isomer
32 as red-orange crystals, mp 219-220 °C (dec): 1H NMR (360 MHz)
δ 6.85 (s, H-13,14), 6.55 (s, H-6,7), 6.49 (s, H-5,8), 6.13 (d, J ) ∼ 7
Hz, H-1,12), 5.52 (d, J ) ∼7 Hz, H-4,9), 5.48 (t, J ) ∼ 7 Hz, H-3,
10), 5.39 (t, J ) ∼ 7 Hz, H-2,11), 0.64 (s, -CH3); 13C NMR (90.6
MHz) δ 232.9, 141.3, 134.0, 127.6, 122.9, 118.5, 101.5, 96.9, 91.9,
91.7, 90.7, 89.0, 41.0, 21.5; IR 1950, 1875 cm-1; UV (CH2Cl2) λmax
(ꢀmax) nm 336 (31 100), 438 (20 300) tail to 630 nm. EI HR MS: Calcd
for C32H20Cr2O6: 604.0070. Found: 604.0093. Anal. Calcd: C,
63.58; H, 3.34. Found: C, 62.91; H, 3.08.
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council of Canada and the University
of Victoria for financial support of this work.
The second isomer eluted was mostly 34, which could be obtained
pure by fractional crystallization from cyclohexane-THF as red-orange
1
Supporting Information Available: Tables of r, θ, f, σ (see
eqs 2 and 3) for compounds 9, 10, 11, 12, 26, 27, 28, 29 (6
pages). An ORTEP diagram of 9 to indicate the carbon
skeleton. See any current masthead page for ordering and Web
access instructions.
crystals, mp 214-216 °C (dec); H NMR (360 MHz, THF-d8) δ 7.03
(s, H-13,14), 6.71 (s, H-8), 6.71 and 6.65 (AB, J ) 9.5 Hz, H-7,6),
6.67 (s, H-5), 6.45 (d, J ) 6.9 Hz, H-12), 6.30 (d, J ) 6.9 Hz, H-1),
5.78 (dd, J ) ∼7, 1 Hz, H-9), 5.78 (t, J ) ∼7 Hz, H-3), 5.69 (t, J )
∼7 Hz, H-10), 5.63 (dd, J ) 6.7, 1.1 Hz, H-11), 5.55 (dt, J ) ∼7, 1.1
Hz, H-11), 5.49 (dt, J ) ∼7, ∼1 Hz, H-2), 0.90 and 0.45 (s, -CH3)
[in CDCl3 these were at δ 0.86 and 0.42]; 13C NMR (90.6 MHz, THF-
JA9729138