Transition Metal Complexes of Diazenes XLI: Cobalt Catalyzed Addition of Internal Alkynes to 1,2-diaryldiazenes: Formation of 2,3-Dihydrocinnolines, mono- and distilbenylazobenzenes

Article abstract of DOI:10.1016/S0022-328X(98)00404-5

The CoH3(PPh3)3 catalyzed addition reaction betwen azobenzenes and internal alkynes produces 2-stilbenylazobenzenes and isomeric 2,3-dihydrocinnolines which are intermediates in the rhodium catalyzed formation of N-anilinoindoles from the same substrates.

Full text of DOI:10.1016/S0022-328X(98)00404-5

Journal of Organometallic Chemistry 558 (1998) 91–101
Transition metal complexes of diazenes XLI1: cobalt catalyzed addition
of internal alkynes to 1,2-diaryldiazenes: formation of
2,3-dihydrocinnolines, mono- and distilbenylazobenzenes
U. Du¨rr, F.W. Heinemann, H. Kisch *
Institute for Inorganic Chemistry of the Uni6ersity Erlangen-Nu¨rnberg, Egerlandstr. 1, D-91058 Erlangen, Germany
Received 29 September 1997; received in revised form 27 November 1997
Abstract
The CoH3(PPh3)3 catalyzed addition reaction between azobenzenes and internal alkynes produces 2-stilbenylazobenzenes and
isomeric 2,3-dihydrocinnolines which are intermediates in the rhodium catalyzed formation of N-anilinoindoles from the same
substrates. The structures of six adducts were resolved by single crystal X-ray structural analyses. In both reactions the insertion
of the alkyne into a M–H bond and ortho-metalation of the 1,2-diazene are key steps of the catalytic cycle. When only one phenyl
ring of the diazene is substituted like in 3,5-dichloro- and 3,5-difluoroazobenzene, selectively this ring is ortho-metalated in the
case of cobalt while only the unsubstituted or both rings are attacked in rhodium catalysis. This difference enables a two-step
regioselective cobalt catalyzed synthesis of the corresponding N-anilinoindole via isolation of 2-stilbenyl-3,5-difluoroazobenzene
and subsequent acid catalyzed rearrangement to the indole derivative. Results obtained with 4-methoxytolan reveal that insertion
into the Co–H bond produces regioisomers the ratio of which is determined by the sign of the triple bond polarization. It is
postulated that insertion initially affords a cis-alkenyl ligand which undergoes an efficient cis–trans isomerization except when
decafluorotolan is the alkyne; in this case also the cis-alkenyl product was isolated. Catalytic amounts of HOAc lead to a tenfold
increase of reaction rate. No reaction is observed when sterically demanding substitutents like mesityl and adamantyl are
introduced to the triple bond. © 1998 Elsevier Science S.A. All rights reserved.
Keywords: Cobalt catalysis; Hydride complexes; 1,2-Diaryldiazenes; Alkyne insertion; Stilbenylazobenzene; 2,3-Dihydrocinnoline;
Ortho-metalation
1. Introduction
CoH3(PPh3)3, new 2:1 adducts were isolated in the form
of distilbenylazobenzenes 4 which may undergo ring
closure to the isomeric 2,3-dihydrocinnolines 2 (Scheme
1) [2]. The former were obtained exclusively since no
isomerization occurred when the azobenzene derivative
was substituted in the meta-positions by CH3 or Cl. It
is noted that in all cases the stilbenyl group had a
trans-configuration except when decafluortolan was
used, in this case also the cis-stilbenyl isomer of the
corresponding 2,3-dihydrocinnoline could be isolated
([2]b). Upon conducting the reaction in Et2O or THF
solution at room temperature (r.t.), it was possible to
isolate 1:1 adducts which have a structure of the 2-stil-
benylazobenzene type 3 [3]. The formation of these
During our studies on the activation of 1,2-diazenes
we found that the reaction between aromatic azo com-
pounds and internal alkynes catalyzed by Wilkinson’s
catalyst RhCl(PPh3)3 in boiling 1-PrOH/HOAc solu-
tions afforded 1:1 adducts which have the structures of
hitherto unknown N-anilinoindole derivatives [1].
When this reaction was performed in a melt of the
substrates and the rhodium complex was replaced by
* Corresponding author. Fax: +49 9131 857363; e-mail:
kisch@anorganik.chemie.uni-erlangen.de
1 For XL, see: H. Kisch, Z. Naturforsch. 52B (1997) 994.
0022-328X/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.
PII S0022-328X(98)00404-5

92
U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
products was rationalized by a catalytic cycle [3] out of
which some key intermediates are illustrated in Scheme
2.
or 80 (1) min, the absorbance recovers to its original
value.
It is assumed that after loss of dihydrogen the reac-
tion of A with tolan affords the cis-stilbenyl complex B.
Subsequent coordination and ortho-metalation of
azobenzene leads to the intermediate trans-C, wherein
the stilbenyl ligand and the ortho-metalated ring are in
trans-position. Isomerization to cis-C has to proceed
before reductive elimination to the 1:1 adducts 1 and 3
can occur. Alternatively, a second insertion of the
alkyne, ortho-metalation of the stilbenylphenyl ring and
reductive elimination from intermediary D may pro-
duce the 2:1 adducts 2 and 4. However, no information
is available on the factors which control the regioselec-
tivity of alkyne insertion when an unsymmetrical tolan
derivative is used.
In rhodium catalysis, 3,3%,5,5%-tetramethylazobenzene
does not react to the corresponding N-anilinoindole
derivative due to steric hindrance in the ortho-metala-
tion step, while 3,5-difluoroazobenzene afforded a mix-
ture of two isomeric indoles and 3,5-dichloroazo-
benzene exclusively yielded that indole derivative where
the unsubstituted phenyl ring has been ortho-metalated.
Opposite to that, in the cobalt catalyzed reaction
3,3%,5,5%-tetramethylazobenzene and tolan afforded the
expected 2-stilbenylazobenzene of type 3 [3]. Since the
latter could be converted to the N-anilinoindole
through a HOAc catalyzed rearrangement [3], this
route seemed to open a path to hitherto not accessible
indole derivatives. In the following we therefore report
on experiments with 3,5-substituted azobenzenes and
on the influence of varying steric and electronic proper-
ties of the alkyne on the regioselectivity of the reaction.
2.1. Reactions with tolan and 4-methoxytolan
As also observed in the melt reaction [2], the reaction
between azobenzene and tolan in solution afforded the
2:1 adduct 2a for which a single crystal X-ray analysis
could be performed [4]. In agreement with previous
results, according to which in the absence of meta-sub-
stituents in the azobenzene component the ring closure
of the intermediate distilbenylazobenzene to the 2,3-di-
hydrocinnoline derivative occurs, 2a has a trans-stil-
benyl-2,3-dihydrocinnoline structure (Table 1, Fig. 1).
The presence of an ortho-quinoid system is evidenced
by the two exocyclic double bonds C11–N1 (134.6 pm)
and C16–C3 (137.5 pm), and by the alternating bond
lengths between C11–C12 (145.1), C12–C13 (135.2),
C13–C14 (142.4), C14–C15 (134.2) and C15–C16
(144.9 pm). These data excellently agree with those of
2. Results and discussion
The addition of solid CoH3(PPh3)3 to a solution of
an 1,2-diazene and an alkyne derivative in the molar
ratio of 1:2 in diethyl ether or THF at r.t. [3] afforded
the already known 2b [2], the new 2:1 adducts 4c and
2a, and the new 1:1 adducts 3b and 3d which could only
be purified by preparative HPLC (Scheme 1). In gen-
eral, the 2,3-dihydrocinnolines 2 are dark red powders
while the isomeric distilbenylazobenzenes 4 form or-
ange powders; the 2,3-dihydrocinnolines 1 and isomeric
monostilbenylazobenzenes 3 were obtained as red oils
or orange powders, respectively. The 2,3-dihydrocinno-
line structure is suggested by the photochromic behav-
ior [2,3]. Irradiating a diluted red solution of 1 or 2 for
about 1 min leads to a color change to yellow and a
decreased absorbance at about 500 nm due to a ring
opening to the stilbenylazobenzenes of type 3 or 4,
respectively. Upon standing in the dark for about 10 (2)
Scheme 1. Products isolated from the reaction of 1,2-diaryldiazenes
with mono- and disubstituted alkyne derivatives.

U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
93
was formed in only 50% of the amount of 3d. Of the
two expected transition states E (Scheme 3) seems to be
favored over F as also observed in the rhodium cata-
lyzed indole synthesis [1].
The N1–N2 distances amount to 124.7(2) (3b) and
124.4(4) pm (3d) and exhibit only small deviations from
the values obtained for unsubstituted azobenzene (124.3
pm) [6]; the two phenyl rings of the azobenzene frag-
ment C11–C16 and C21–C26 are slightly twisted by 12
(3b) and 4° (3d), respectively. The C3–C4 distances
(133.3(3) (3b), 132.9(6) (3d) pm) are similar to the
double bond length in unsubstituted trans-stilbene
(131.8 pm) [7] and the interplanar angle between the
double bond plane C16–C3–C4 and the azobenzene
phenyl ring C11–C16 amount to 63 (3b) and 71° (3d)
indicating that the ortho-fluoro substituent in 3b (van
der Waals radius=135 pm) [8] has less steric influence
than the ortho-methyl group (200 pm) in 3d. The two
phenyl rings of the stilbenyl fragment C31–C36 and
C41–C46 form with the double bond plane C16–C3–
C4 interplanar angles of 32 and 43° (3b), 35 and 42°
(3d) and are twisted by 75 and 77°, respectively. As in
the case of 2a the rather long C16–C3 distances
(150.3(3) (3b) and 150.8(5) (3d) pm) and the relatively
large interplanar angles mentioned above suggest a
rather weak interaction between the azobenzene and
stilbene fragments.
Scheme 2.
the previously characterized 2b [5]. The two exocyclic
double bonds are twisted by 18° as indicated by inter-
planar angles of 6 and 12°, formed by the plane of
C11–C16 with the planes of C11–C16–C3 and C16–
C11–N1, respectively. The N–N distance of 135.1 pm
is much shorter than expected for a simple single bond
(147 pm) suggesting an extension of the conjugated
system across N2. This is supported by the planarity of
N2 as indicated by the angles N1–N2–C4 (120.2°),
C4–N2–C21 (123.8°) and N1–N2–C21 (115.7°). The
plane generated by N2, N1, C21 and C4 forms with the
ortho-quinoid system and with the phenyl ring C21–
C26 interplanar angles of 38 and 30°, respectively.
Extended conjugation is further evidenced by the N2–
C21 bond length of 142.8 pm which is about 4 pm
shorter than expected. The short C3–C31 distance of
147.5 pm suggests that also the phenyl ring C31–C36
participates in electron delocalization; the dihedral an-
gle with the quinoid fragment is 49°. sp3-Hybridization
of C4 follows from the bond angles N2–C4–C3
(104.5), C3–C4–C41 (113.7) and C41–C4–N2 (116.6°)
and the bond lengths of N2–C4 (148.1), C4–C3 (152.3)
and C4–C41 (151.3 pm). The bond length C5–C12 of
152.5 pm indicates that the trans-stilbenyl group ex-
hibits only minor interactions with the ortho-quinoid
system. Correspondingly, the latter and the plane gen-
erated by C61, C6, C5, C51 form an interplanar angle
of 112°.
No product formation could be observed in the
reaction between azobenzene and 4-nitrotolan, while
4-methyltolan afforded the expected adducts but sepa-
ration of the regioisomers was not achievable by
HPLC. The same experimental problem arose when in
the latter reaction azobenzene was replaced by 3,5-
dichloroazobenzene.
2.2. Reactions with disubstituted tolan deri6ati6es
When 3,5-dichloroazobenzene and the disubstituted
4,4%-dimethoxytolan were employed, the 2:1 adduct 4c
was obtained (Fig. 3). Due to the presence of two
sterically demanding stilbenyl substituents the almost
coplanar arrangement of the two azobenzene phenyl
rings C11–C16 and C21–C26, as observed in the
monostilbenyl compounds 3b and 3d, is prevented and
the interplanar angle increased to 76°. 4,4%-Dimethyl-
tolan in the same reaction afforded very labile 1:1 and
2:1 adducts which could not be isolated as analytically
pure materials.
The reactions between tolan or the monosubstituted
4-methoxytolan and 3,5-difluoroazobenzene or 3,3%,5,5%-
tetramethylazobenzene, afforded the 1:1 adducts 3b and
3d, respectively, the structures of which were resolved
by single crystal X-ray analyses (Table 1, Fig. 2). Both
compounds are 2-(trans-stilbenyl)azobenzenes and in 3d
the C2-carbon atom of the alkyne became attached to
the dimethylphenyl ring; the C1-bound regioisomer 3d%
Introduction of sterically demanding substituents at
the triple bond resulted in a complete inhibition as
indicated by the absence of any addition products in
the reactions attempted between azobenzene or 4,4%-
dimethylazobenzene with bis(adamantyl)- or bis(me-
sityl)acetylene. When one aromatic group of the alkyne
is replaced by a methyl group as in the reaction of
1-phenylpropyne with azobenzene, no corresponding

94
U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
Table 1
Selected bond distances (pm)
2a [4]
3b
4c
3d
cis-3f
4f
N(1)–N(2)
N(1)–C(11)
N(2)–C(21)
C(3)–C(4)
C(16)–C(3)
C(5)–C(6)
C(12)–C(5)
135.1(7)
124.7(2)
142.3(2)
142.9(3)
133.3(3)
150.3(3)
120.9(5)
145.6(5)
144.8(5)
133.3(6)
148.8(5)
132.5(5)
149.2(5)
124.4(4)
142.3(5)
143.3(5)
132.9(6)
150.8(5)
124.5(3)
143.1(3)
142.7(3)
133.4(4)
148.8(4)
132.7(5)
149.2(5)
125.1(4)
143.6(5)
141.8(5)
132.3(6)
149.6(5)
134.6(8)
142.8(8)
152.3(9)
137.5(9)
126.9(11)
152.5(10)
adducts but a mixture of oligomerization products were
obtained; according to MS analysis they contained one
molecule of the diazene and up to seven alkyne
molecules.
vis, HPLC and MS data with the cis-isomer. According
to HPLC the ratio of cis-3f:trans-3f was 7:5.
The molecular structure of cis-3f reveals no special
features except the presence of a cis-decafluorostilbenyl
group and a less twisted arrangement of the two stil-
benyl phenyl rings (a twist angle of 174° as compared
to an average of 76° observed for 3b and 3d). Surpris-
ingly, 4f exhibits trans-configuration at both de-
cafluorostilbenyl substituents.
When 2,2%,4,4’,6-pentamethylazobenzene and de-
cafluorotolan were employed, no conversion of the
substrates could be observed. Apparently, the de-
cafluorostilbenyl ligand in B (Scheme 2) decreases the
electron density at the cobalt center to such an extent
that in the following step nucleophilic attack at the
2,4-dimethylphenyl ring of the azobenzene ligand can-
not occur anymore.
The unique formation of the cis-stilbenyl adduct
mentioned above indicates that the electron-withdraw-
ing properties of the pentafluorophenyl groups either
prevent the complete isomerization of an initially
formed cis-alkenyl ligand or promote a cis-insertion of
the alkyne. Although there is no general rule for the
stereochemistry of alkyne insertion into M–H or M–C
bonds, in most cases a cis-insertion is the primary step
which often is followed by cis/trans-isomerization. For
the latter various mechanisms have been proposed,
including protonation at the alkenyl ligand [10] or
metal center [11], and pathways through radical [12]
and coordinatively unsaturated intermediates [13].
Bergman et al. postulated that in the reaction between
Ni(acac)(PPh3)Ph with PhCCR (acac, acetylacetonato,
R=Ph or CH3) the cis/trans isomerization of the |-
alkenyl ligand occurs via rearrangement to a carbe-
nium–carben intermediate [14] as depicted for the
present case in Scheme 5. Although isomerization may
occur also in the first intermediate B, stabilization of
the negatively polarized cobalt center should be more
efficient in C and we propose that it proceeds through
the transition state G. Destabilization of the positive
charge by the pentafluorophenyl group should decrease
the stability of G and therefore disfavor isomerization
which has to compete with reductive elimination to the
1:1 adducts 1 and 3.
To investigate the influence of electron withdrawing
substituents on the course of the reaction, decafluoro-
tolan was employed. This was the only alkyne which in
the melt reaction with 4,4%-dimethylazobenzene at 90°C
afforded the corresponding 2,3-dihydrocinnoline with a
cis-stilbenyl substituent; no reaction occurred when all
meta-positions are substituted like in 3,3%,5,5%-te-
tramethylazobenzene [9]. When the reaction was per-
formed in THF solution, different to the preceding
reactions with other alkynes, an almost complete con-
version of the alkyne to the 2:1 adduct 2e took place
when 4,4%-dichloroazobenzene was employed. FD-MS
analysis indicated the absence of any 1:1 adducts. How-
ever, these were obtained when 3,5-dichloroazobenzene
was used (Scheme 4). The structures of cis-3f and 4f
were resolved by X-ray analysis (Fig. 4, Table 1), the
structure of trans-3f follows from comparison of UV–
Fig. 1. Molecular structure of 2a; hydrogen atoms are omitted for
clarity.

U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
95
Fig. 2. Molecular structure of 3b and 3d; hydrogen atoms are omitted for clarity.
2.4. Influence of acids
2.3. Rearrangement of 3b to the corresponding
N-anilinoindole
Since catalytic amounts of HOAc increased yields
and turnover numbers of the rhodium catalyzed indole
synthesis [1], the effect of various acids was also investi-
gated on the cobalt catalyzed reaction between azoben-
zene and tolan. In the absence of an acid stirring of the
solution was necessary for 24 h to obtain 80% conver-
In the RhCl(PPh3)3 catalyzed reaction between 3,5-
difluoroazobenzene and tolan, a mixture of two iso-
meric N-anilinoindoles was obtained [15] in the ratio of
5b:5b%=5:16, with 5b% being the indole in which the
unsubstituted phenyl ring has been ortho-metalated.
Correspondingly, two 2-stilbenylazobenzene derivatives
of type 3 should be involved as intermediates. Since in
all cobalt catalyzed reactions of unsymmetrically substi-
tuted azobenzenes only those products were obtained
which arrived from ortho-metalation of the substituted
phenyl ring, exclusively the expected 3b and no other
regioisomer could be detected (Scheme 6). Refluxing of
3b in 1-BuOH/HOAc at 115°C, which are the standard
conditions for the rhodium catalyzed indole formation
[1], afforded the isomeric N-anilinoindole 5b in 95%
yield. Although this preparation of 5b involves two
steps it offers a convenient access to a single isomer.
Fig. 3. Molecular structure of 4c; hydrogen atoms are omitted for
Scheme 3.
clarity.

96
U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
Column chromatography: Al2O3 (ICN, neutral, activ-
ity I); l=40 cm, ¥=2.0 cm; petroleum
ether:tetrahydrofuran ratio 20:1 (v/v) as eluting agent.
HPLC: Knauer HPLC pump 64 with analytical and
preparative pump head, Knauer UV–vis filter photo-
meter at u=220 nm as detector.
Analytical measurements: precolumn (30 mm×8
mm) attached to main column (250 mm×8 mm) and
both filled with Spherisorb ODS2, 5 mm (RP C18),
eluting with CH3CN:H2O ratio 5:1 (v/v) at a flow rate
of 5.0 ml min−1
.
Sample preparation: 100 ml of the reaction solution
was withdrawn, evaporated to dryness, and redissolved
in 0.5 ml of CH3CN; 20 ml of this solution was then
injected.Preparative isolations: 1 ml of the acetonitrile
solution was injected; identical elution agent and filling
material were used while the size of precolumn and
main column were 30 mm×32 mm and 250 mm×32
mm, respectively; the flow rate was 35 ml min−1
.
1H- and 13C-NMR spectra were measured in chloro-
form-d1 solutions at 400 or 270 and 100 or 67 MHz,
respectively, and IR spectra in KBr unless otherwise
noted. The following instruments were used: NMR,
Jeol FT-JNM-LA 400 and Jeol FT-JNM-EX 270; IR,
Perkin–Elmer 983 and FT IR 1600; UV–vis, Shimadzu
UV-3101 PC; MS, Jeol MStation 700 and Varian MAT
212.
Scheme 4. Products isolated from the reaction of 3,5-dichloroazoben-
zene with decafluorotolan.
CoH3(PPh3)3 was prepared according to literature
[16]. Azobenzene (Merck) and tolan (Fluka) were com-
mercially available. 3,3%,5,5%-Tetramethylazobenzene
and 4,4%-dichloroazobenzene were synthesized from the
corresponding aniline derivatives [17], 3,5-dichloroa-
zobenzene and 3,5-difluoroazobenzene from the corre-
sponding 3,5-dihalogenoanilines and nitrosobenzene
[18]. Decafluorotolan was synthesized according to lit-
erature [19] from tetrabromoethylene which was ob-
tained via dehydrohalogenation of pentabromoethylene
[20].
sion. Contrary to that, the presence of acetic acid
induced a complete conversion of the substrates to the
2:1 adduct 2a as the unique product within 2 h. No
intermediate 1:1 adduct 1a or the corresponding N-
anilinoindole could be detected by HPLC and MS
analyses. Stronger organic acids like trifluoro acetic
acid or para-toluene sulfonic acid did not improve the
reaction. Best results were obtained when the molar
ratio of azobenzene: tolan: catalyst: HOAc was
1.0:2.0:0.35:0.15. When the amount of HOAc was in-
creased up to equimolar with azobenzene, the reaction
was completely inhibited.
The positive effect of HOAc is further demonstrated
by the reaction of 4-methyltolan with azobenzene which
afforded the expected 2,3-dihydrocinnolines in form of
the 1:1 and 2:1 adducts 1g and 2g, respectively. Upon
addition of HOAc the yields of 1g and 2g increased
from 11 and 10% to 12 and 38%, respectively.
3.1. Synthesis of the 2,3-dihydrocinnolines 1 and 2, the
2-trans-stilbenylazobenzenes 3 and the 2,6-di(trans-
stilbenyl)azobenzenes 4
3.1.1. General procedure
All reactions have been performed in diethyl ether or
THF solution. Unless otherwise noted, the general pro-
cedure for the synthesis of the mixtures of 1 and 2 or 3
and 4 is described in the following.
The structures of the new 1:1 adducts 3b and 3d and
the new indole 5b follow from comparison of extensive
NMR data (1H, 13C, 1H–1H-COSY, 1H–13C-COSY)
with those of unsubstituted azobenzene and trans-stil-
bene and from calculation of the 13C-NMR data with
the increment system for substituted benzenes and
olefinic C-atoms described in the literature (for atom
numbering see Scheme 7) [21].
3. Experimental section
All reactions were conducted under a nitrogen atmo-
sphere, solvents were dried and purified by standard
procedures and nitrogen-saturated after distillation.
Acetonitrile (Ferak) used for the HPLC measurements
was sonicated and stored under nitrogen.

U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
97
Fig. 4. Molecular structure of cis-3f and 4f; hydrogen atoms are omitted for clarity.
3.1.3. 2-(trans-Stilbenyl)-3,5-difluoroazobenzene (3b)
3.1.2. 2,3,4-Triphenyl-8-(trans-stilbenyl)-
and 2,3,4-triphenyl-5,7-difluoro-8-(trans-stilbenyl)-2,3-
dihydrocinnoline (2b)
2,3-dihydrocinnoline (2a)
Azobenzene (728 mg, 4.0 mmol) and tolan (1426 mg,
8.0 mmol) were dissolved under stirring in 20 ml of
diethyl ether. Thereafter about 700 mg (0.82 mmol) of
CoH3(PPh3)3 were added in one portion whereby the
color of the solution changed within 5 min from orange
over black–green and dark-brown to red. After stirring
for another 24 h at r.t., 5 g of Al2O3 were added; after
removing the solvent, the dark red residue was chro-
matographed at Al2O3 with light petroleum ether:THF
solution 20:1 (v/v). The first orange–red fraction
yielded 425 mg (1.18 mmol, 30%) of an orange oil of 1a
and some by-products, the following red fraction af-
forded 130 mg of a red powder of 2a [3] (0.24 mmol,
6%), turn-over number (TON): (mmol 1+2)/mmol cat-
alyst: 1.7. The orange–red oil was purified by prep.
HPLC yielding 100 mg (0.28 mmol) of 1a [3]. Single
crystals of 2a were obtained from CH2Cl2:MeOH ratio
1:2 (v/v) at r.t.
3,5-Difluoroazobenzene (872 mg, 4.0 mmol), tolan
(1426 mg, 8.0 mmol) and CoH3(PPh3)3 (745 mg, 0.88
mmol) in 3 ml of diethyl ether were used. The color of
solution changed within 2 min from orange over black
to dark red. Yields after column chromatography were:
985 mg of an orange oil of 3b (2.48 mmol, 62%) and
by-products, 223 mg (0.39 mmol, 10%) of a red powder
of 2b (TON 3.3). Yield after purification by prep.
HPLC: 575 mg of an orange powder of 3b. 3b: MS: FD
m/z 396 [M]+. Calc. for C26H18F2N2 (396.4): C, 78.77;
H, 4.58; N, 7.07%. Found: C, 79.19; H, 4.72; N, 6.78%.
1H-NMR (400 MHz, CDCl3) l (ppm): 7.59–6.90 [m,
18H, 4-H, 6-H, 8-H, 9-H 10-H, 11-H, 12-H,
PhCꢀCHPh (11H)]. 13C-NMR (100.4 MHz, CDCl3) l
(ppm): 99.4 (dd, 2J(F,C)=23 Hz, 4J(F,C)=4 Hz, C-6),
2
105.9 (t, J(F,C)=26 Hz, C-4), 123.2 (C-8 and C-12),
Scheme 5.
Scheme 6.

98
U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
3.1.5. 2-(trans-4-Methoxystilbenyl)-3,3%,5,5’-
tetramethylazobenzene (3d/3d%) and
2,6-di(trans-4-methoxystilbenyl)-3,3%,5,5%-
tetramethylazobenzene (4d/4d%)
A sample of 715 mg (3.0 mmol) of 3,3%,5,5%-te-
tramethylazobenzene, 1250 mg (6.0 mmol) of 4-
methoxytolan and 585 mg (0.69 mmol) of CoH3(PPh3)3
in 7 ml of THF were used. The color of the solution
changed after addition of the catalyst from orange to
green–black. Yields after column chromatography
were: 451 mg of an orange oil of 3d/3d% (1.01 mmol,
34%) and by-products, 150 mg (0.24 mmol, 9%) of an
orange powder of 4d/4d% (TON 1.8). Yield after purifi-
cation by prep. HPLC: first fraction: 160 mg of 3d/3d%;
second fraction: 215 mg of an orange oil of 3d/3d%. 3d:
Scheme 7.
4
1
MS: FD m/z 446 [M]+. H-NMR (270 MHz, CDCl3) l
(ppm): 7.40 (s, 1H, 10-H), 7.37–7.15 (m, 8H, 8-H,
12-H, 14-H, 16-H to 20-H), 7.12 (s, 1H, 6-H), 6.95 (s,
1H, 4-H), 6.91 (d, 2H, 17%-H and 19%-H), 6.62 (d, 2H,
16%-H and 20%-H), 3.65 (s, 3H, 23%-O–CH3), 2.41 (s, 3H,
21%-CH3), 2.27 (s, 6H, 22-CH3, 22%-CH3), 2.12 (s, 3H,
21-CH3). 13C-NMR (67.7 MHz, CDCl3) l (ppm): 20.1
(C-21), 21.8 (C-22 and C-22%), 22.0 (C-21%), 55.7 (C-23%),
114.2 (C-17% and C-19%), 114.3 (C-6), 121.5 (C-8 and
C-12), 127.1 (C-16 and C-20), 127.4 (C-18), 128.9 (C-17
and C-19), 129.8 (C-14), 130.5 (C-16% and C-20%), 131.2
(C-15%), 132.8 (C-10), 134.4 (C-4), 137.0 (C-3), 137.8
(C-5), 138.4 (C-15), 138.6 (C-2), 139.0 (C-9 and C-11),
144.3 (C-13), 151.7 (C-1), 153.9 (C-7), 159.1 (C-18%).
The corresponding 13C signals of 3d% are the same as for
3d within 0.2 ppm. UV–vis (CH2Cl2) umax [nm] (m [M−1
cm−1]): 459 (700), 309 (18000). HPLC (CH3CN:H2O
ratio 5:1 (v/v), 5ml min−1) retention time [min] (inte-
grated area): after column chromatography: 14.7 (2.2×
106) 3d%, 15.4 (4.5×106) 3d (ratio 3d:3d% was 2:1). After
prep. HPLC: first fraction: 14.7 (2.2×106) 3d%, 15.4
(1.4×106) 3d (ratio 3d/3d%=2/3); second fraction: 14.7
(7.4×105) 3d%, 15.4 (4.4×106) 3d (ratio 3d/3d%=6/1).
Single crystals of 3d were obtained from CH2Cl2:MeOH
ratio 1:2 (v/v) at −20°C. 4d/4d%: MS: FD m/z 655
[M]+. Calc. for C46H42N2O2 (656.3): C, 84.11; H, 6.75;
N, 4.26%. Found: C, 84.06; H, 6.64; N, 4.26%. UV–vis
(CH2Cl2) umax [nm] (m [M−1 cm−1]): 461 (2300), 302
(59000).
2
124.3 (dd, J(F,C)=19 Hz, J(F,C)=4 Hz, C-2), 126.5
(C-16 and C-20), 127.2 (C-18%), 127.5 (C-18), 128.1
(C-17% and C-19%), 128.4 (C-16% and C-20%), 128.5 (C-17
and C-19), 128.9 (C-9 and C-11), 131.6 (C-12), 132.1
(C-15%), 132.6 (C-14), 137.2 (C-15), 142.6 (C-13), 152.4
(C-7), 152.5 (C-1), 160.6 (dd, 1J(C,F)=245 Hz,
3J(C,F)=12 Hz, C-5), 162.6 (dd, 1J(C,F)=246 Hz,
3J(C,F)=12 Hz, C-3). UV–vis (CH2Cl2) umax [nm] (m
[M−1 cm−1]): 302 (24300). HPLC (CH3CN:H2O ratio
5:1 (v/v), 5 ml min−1) retention time [min] (integrated
area):
7.3
(5.2×106)
3b,
10.1
(8.5×104)
[H(PhCꢀCPh)2H]. Single crystals of 3b were obtained
from CH2Cl2:MeOH ratio 1:2 (v/v) at −20°C. 2b: MS:
FD m/z 574 [M]+. For further characterization see ref.
[2].
3.1.4. 2-(trans-4,4%-Dimethoxystilbenyl)-3,5-
dichloroazobenzene (3c) and 2,6-di(trans-4,4%-
dimethoxystilbenyl)-3,5-dichloroazobenze ne (4c)
A sample of 753 mg (3.0 mmol) of 3,5-dichloroa-
zobenzene, 1430 mg (6.0 mmol) of 4,4%-dimethoxytolan
and 735 mg (0.86 mmol) of CoH3(PPh3)3 in 13 ml of
THF were used. The color of the solution changed after
addition of the catalyst only from orange to green–
black and not to red since distilbenylazobenzenes 4
absorb at about 500 nm much weaker than the dihy-
drocinnolines 2. Yields after column chromatography
were: 553 mg of an orange oil of 3c (1.13 mmol, 38%)
and by-products, 382 mg (0.53 mmol, 17%) of an
orange powder of 4c (TON 1.9). Purification of 3c by
prep. HPLC was not performed, since decomposition of
3c occurred. 4c: MS: FD m/z 726 [M]+. Calc. for
C44H38Cl2N2O4 (728.2): C, 72.63; H, 4.99; N, 3.85%.
Found: C, 72.77; H, 5.22; N, 3.84%. UV–vis (CH2Cl2)
umax [nm] (m [M−1 cm−1]): 510 (1000), 302 (59000).
Single crystals of 4c were obtained from CH2Cl2:MeOH
ratio 1:2 (v/v) at r.t.
3.1.6. 2-(4-Chlorophenyl)-3,4-bis(pentafluorophenyl)-8-
(decafluorostilbenyl)-2,3-dihydrocinnoline (2e)
A total of 502 mg (2.0 mmol) of 4,4%-dichloroazoben-
zene, 1432 mg (4.0 mmol) of decafluorotolan and 580
mg (0.68 mmol) of CoH3(PPh3)3 in 13 ml of THF were
used. The color of the solution changed after addition
of the catalyst within 5 min from orange over black to
dark red. The formation of an 1:1 adduct could not be
detected, neither by thin layer chromatography nor by

U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
99
over black to dark red. Upon adding HOAc as co-cata-
lyst, the conversion of reactants increased to about 90%
and the reaction time decreased to about 2 h while the
yield reached values normally obtained only in reaction
with unsubstituted tolan. Yields after column chro-
matography were: 136 mg of a red oil of 1g/1g% (0.36
mmol, 12%) and by-products, 647 mg (1.14 mmol, 38%)
of a dark red powder of 2g/2g% (TON 3.6). Purification
of 1g/1g% by prep. HPLC was not performed. 2g/2g%:
MS: FD m/z 566 [M]+. Calc. for C42H34N2 (566.8): C,
89.01; H, 6.05; N, 4.94%. Found: C, 88.25; H, 6.34; N,
4.22%. UV–vis (CH2Cl2) lmax [nm] (e [M−1 cm−1]): 512
(10000), 288 (43000).
FD-MS. Yield after column chromatography (light
petroleum ether:diethyl ether ratio 10:1 (v/v)) was: 1242
mg (1.28 mmol, 64%) of a dark red powder of 2e (TON
1.9). 2e: MS: FD m/z 966 [M]+. Calc. for
C40H8Cl2F20N2 (966.0): C, 49.69; H, 0.83; N, 2.90%.
Found: C, 50.28; H, 0.76; N, 2.38%. UV–vis (CH2Cl2)
umax [nm] (m [M−1 cm−1]): 523 (8700), 240 (38000).
3.1.7. 2-(cis-Decafluorostilbenyl)-3,5-
dichloroazobenzene (cis-3f), 2-(trans-decafluorostilbenyl)
-3,5-dichloroazobenzene (trans-3f) and 2,6-di(trans-
decafluorostilbenyl)-3,5-dichloroazobenzene (4f)
A total of 502 mg (2.0 mmol) of 3,5-dichloroazoben-
zene, 1432 mg (4.0 mmol) of decafluorotolan and 955
mg (1.13 mmol) of CoH3(PPh3)3 in 12 ml of THF were
used. The color of the solution changed after addition
of the catalyst immediately from orange to black. Yield
after column chromatography [light petroleum
ether:THF ratio 20:1 (v/v) up to methylene chlo-
ride:diethyl ether ratio 1:1 (v/v)] was: 1267 mg of a
mixture of three different products. HPLC
(CH3CN:H2O ratio 5:1 (v/v), 5 ml min−1) retention
time [min] (integrated area): 22.6 (1.3×106) cis-3f, 26.2
(1.0×106) trans-3f, 41.4 (1.2×107) 4f. After purifica-
tion with prep. HPLC: first fraction: 50 mg (0.08 mmol,
4%); second fraction: 72 mg (0.12 mmol, 6%); third
fraction: 778 mg (0.81 mmol, 40%) 4f, all three frac-
tions were obtained as orange powders (TON 0.9).
HPLC (CH3CN:H2O ratio 5:1 (v/v), 5 ml min−1) reten-
tion time [min] (integrated area): first fraction: 22.6
(1.0×106) cis-3f, 26.2 (3.9×104) trans-3f (ratio cis-3f/
trans-3f=25/1); second fraction: 22.6 (1.3×105) cis-3f,
26.2 (9.0×105) trans-3f (ratio cis-3f/trans-3f=1/7);
third fraction: 41.4 (7.2×106) 4f. cis-3f: MS: FD m/z
608 [M]+; UV–vis (CH2Cl2) umax [nm] (m [M−1 cm−1]):
457 (900), 328 (26000); trans-3f: MS: FD m/z 608
[M]+. UV/Vis (CH2Cl2) umax [nm] (m [M−1 cm−1]): 459
(600), 327 (28000). 4f: MS: FD m/z 966 [M]+. Calc. for
C40H8Cl2F20N2 (966.0): C, 49.69; H, 0.83; N, 2.90%.
Found: C, 49.82; H, 0.61; N, 2.86%. UV–vis (CH2Cl2)
umax [nm] (m [M−1 cm−1]): 476 (1000), 272 (48000).
Single crystals of cis-3f and 4f were obtained from
CH2Cl2:MeOH ratio 1:2 (v/v) at −20°C and r.t.,
respectively.
3.1.9. 2-(trans-4-Methylstilbenyl)-3,5-dichloroazoben-
zene (4h/4h%) and 2,6-di(trans-4-methylstilbenyl)-3,5-
dichloroazobenzene (3h/3h%)
A total of 502 mg (2.0 mmol) of 3,5-dichloroazoben-
zene, 768 mg (4.0 mmol) of 4-methyltolan and 175 mg
(0.21 mmol) of CoH3(PPh3)3 in 20 ml of THF were
used. The color of the solution changed after addition
of the catalyst from orange to green–black. Yields after
column chromatography were: 261 mg of an orange oil
of 3h/3h% (0.59 mmol, 30%) and by-products, 230 mg
(0.36 mmol, 18%) of an orange powder of 4h/4h% (TON
4.5). Purification of 3h/3h% by prep. HPLC was not
attempted. 4h/4h%: MS: FD m/z 635 [M]+. Calc. for
C42H34Cl2N2 (634.2): C, 79.36; H, 5.07; N, 4.41%.
Found: C, 79.23; H, 5.06; N, 4.05%. UV–vis (CH2Cl2)
umax [nm] (m [M−1 cm−1]): 501 (1000), 292 (51000).
3.2. Isomerization of 2-(trans-stilbenyl)-3,5-difluoroazo-
benzene (3b) to N-anilino-2,3-diphenyl-4,6-difluoro-
indole (5b) catalyzed by HOAc
A 23.6 mg (5.96 mmol) sample of the orange powder
of 3b, dissolved in 5 ml of 1-BuOH and 3.40 ml of
HOAc ([3b]=[HOAc]=1.19×10−2 mol l−1), were
heated to 115°C. The color of the solution changed
within 30 min from orange to pale yellow and a blue
fluorescence (uexc=366 nm) was observed. The reaction
was stopped after 2.5 h and a pale yellow residue (5b)
was obtained after solvent removal. HPLC analysis: 5b
(4.3 min, initial value: area=7.8×104, final value:
area=9.3×105), 3b (7.3 min, initial value: area=
6.2×105, final value: area=5.6×104). 5b: MS: EI m/z
(%): 396 (95) [M]+, 319 (50), [M−Ph]+, 358 (10)
3.1.8. 2,3/4-Diphenyl-4/3-p-tolyl-2,3-dihydrocinnoline
(1g/1g%) and 2,3/4-diphenyl-4/3-p-tolyl-8-(trans-
stilbenyl)-2,3-dihydrocinnoline (2g/2g%)
1
[M−2F]+, 304 (100) [M−NH–Ph]+. H-NMR (400
A total of 546 mg (3.0 mmol) of azobenzene, 1152
mg (6.0 mmol) of 4-methyltolan and 355 mg (0.42
mmol) of CoH3(PPh3)3 in 4 ml of diethyl ether were
used. Before adding the catalyst, 11.5 ml (0.20 mmol) of
HOAc were added. The color of the solution changed
after addition of the catalyst within 5 min from orange
MHz, CDCl3) l (ppm) (for atom numbering see ref.
[2]): 7.35–7.14 [m, 12H, 11-H, 13-H, Ph–H (10 H)],
6.89 (t, 1H, 3J(H,H)=7 Hz, 12-H), 6.85 (dd, 1H,
4
3J(F,H)=9 Hz, J(H,H)=2 Hz, 7-H), 6.66 (dt, 1H,
4
3J(F,H)=9 Hz, J(H,H)=2 Hz, 5-H), 6.53 (s, 1 H,
3
4
8-H), 6.48 (dd, 2 H, J(H,H)=8 Hz, J(H,H)=1 Hz,

100
U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
Table 2
Crystal data and summary of data collection and structure refinement of 3b, 3d, 4c, cis-3f and 4f
Compound
3b
3d·CH2Cl2
4c
cis-3f
4f
Molecular formula
Molecular weight
Crystal system
Space group
a (pm)
b (pm)
c (pm)
h (°)
i (°)
C
26H18F2N2
C32H32Cl2N2O
530.19
Monoclinic
P21/n
907.0(2)
2727.0(3)
1170.0(1)
90.0
103.9(1)
90.0
2.808(7)
4
C44H38Cl2N2O4
728.22
Triclinic
P1
992.6(4)
1325.3(4)
1462.0(8)
92.48(4)
100.59(3)
103.95(2)
1.827(1)
2
C26H8Cl2F10N2
609.24
Monoclinic
P21/n
1571.8(3)
1074.0(2)
1586.2(3)
90.0
118.50(1)
90.0
C40H8Cl2F20N2
967.38
Triclinic
P1
1010.8(2)
1096.3(2)
1628.6(3)
90.02(1)
97.67(1)
97.19(1)
1.774(1)
2
396.42
Triclinic
P1
789.7(10)
1124.2(10)
1236.9(10)
69.96(1)
82.03(1)
79.26(1)
1.010(2)
2
k (°)
V (nm3)
2.353(1)
4
Z
v (mm−1)
0.090
1.303
0.259
1.257
0.225
1.323
0.376
1.720
0.326
1.811
Dcalc. (g cm−3
)
Crystal size (mm)
Color
2q range
Scan rate (deg min−1)
No. measured reflections
Independent reflections
Observed reflections [Fo=4|(F)]
Extinction coefficient k
No. parameters refined
R1 [Fo=4|(F)]
0.70×0.30×0.15
Orange
4.3–54.0
3–30
5301
4332
1616
0.070(4)
344
0.0478
0.1177
0.60×0.40×0.40
Orange
4.6–52.0
3–30
6422
5489
1899
0.008(1)
455
0.0655
0.1767
0.60×0.40×0.40
Orange
4–50
3–30
9837
6420
4448
0.003(3)
614
0.084
0.80×0.60×0.40
Orange
4.7–54.1
6–60
6331
5150
3297
0.0061(9)
394
0.0503
0.1313
0.40×0.40×0.15
Orange
4.3–54.0
3–30
9048
7714
3964
0.0061(7)
610
0.0605
0.1416
wR2 (all data)
0.237
10-H, 14-H). 13C-NMR (100.4 MHz, CDCl3)
l
The solvent was removed and the residue chro-
matographed at Al2O3 [petroleum ether:THF ratio
10:1 (v/v)] yielding a dark red oil. MS: FD m/z: 537
[M]+.
2
4
(ppm): 92.8 (dd, J(F,C)=26 Hz, J(F,C)=4 Hz, C-
7), 97.0 (t, 2J(F,C)=24 Hz, C-5), 110.5 (dd,
2J(F,C)=19 Hz, 4J(F,C)=2 Hz, C-3a), 112.7 (C-10
and C-14), 113.1 (C-3), 121.4 (C-12), 126.5 (C-18%),
127.8 (C-16% and C-20%), 128.2 (C-16 and C-20), 128.3
(C-18), 129.5 (C-11 and C-13), 129.6 (C-15), 130.78
(C-17% and C-19%), 130.81 (C-17 and C-19), 133.9 (C-
15%), 138.3 (C-2), 137.8 (t, 2J(F,C)=12 Hz, C-7a),
146.7 (C-9), 156.5 (dd, 1J(F,C)=252 Hz, 3J(F,C)=
When 5.72 ml (0.10 mmol) of HOAc and 540 mg
(0.64 mmol) of CoH3(PPh3)3 were added ([HOAc]=
3.3×10−2 mol l−1), a TLC showed after 2 h a yel-
low (azobenzene), an orange (1a) and a red fraction
(2a); the reaction was qualitatively faster as compared
to adding 11.45 ml (0.20 mmol) of HOAc and 360 mg
(0.42 mmol) of CoH3(PPh3)3 ([HOAc]=6.7×10−2
mol l−1), a simultaneously performed experiment.
When 57.2 ml (1.0 mmol) of HOAc and 260 mg
(0.31 mmol) of CoH3(PPh3)3 were added ([HOAc]=
0.33 mol l−1), the catalyst was deactivated and no
conversion of the substrates could be observed.
Upon adding 11.48 ml (0.15 mmol) of trifluoro
acetic acid and 300 mg (0.35 mmol) of CoH3(PPh3)3
([CF3COOH]=5.0×10−2 mol l−1), a mixture of
azobenzene, 1a and 2a was observed as indicated by
TLC after 2 h. Additionally, a green residue of deac-
tivated catalyst precipitated.
1
3
14 Hz, C-4), 159.7 (dd, J(F,C)=240 Hz, J(F,C)=
12 Hz, C-6). 19F-NMR (376.0 MHz, CDCl3, CFCl3
as reference) l (ppm): −111.3 (C-4-F), −113.7 (C-6-
F).
3.3. Influence of 6arious acids
The standard amounts of the substrates were 182
mg (1.0 mmol) of azobenzene and 356 mg (2.0
mmol) of tolan dissolved in 3 ml of diethyl ether.
Then 8.59 ml (0.15 mmol) of HOAc and 300 mg (0.35
mmol) of CoH3(PPh3)3 were added ([HOAc]=5.0×
10−2 mol l−1). The color of the solution changed
within 2 min from orange over black to dark red.
After stirring for 2 h, a thin layer chromatogram
(TLC) [Al2O3, petroleum ether:THF ratio 20:1 (v/v)]
indicated complete conversion of the substrates to 2a.
A reaction with 28.5 mg (0.15 mmol) of p-toluene
sulfonic acid as additive and 525 mg (0.62 mmol) of
CoH3(PPh3)3 ([p-CH3C6H4SO3H]=5.0×10−2 mol l−
1) afforded a mixture of azobenzene, 1a and 2a as
indicated by TLC after 2 h, no precipitation of a
green powder of deactivated catalyst was observed.

U. Du¨rr et al. / Journal of Organometallic Chemistry 558 (1998) 91–101
101
References
3.4. Photochromic beha6ior
[1] (a) P. Reisser, Y. Wakatsuki, H. Kisch, Monatsh. Chemie 126
(1995) 1. (b) U.R. Aulwurm, J.U. Melchinger, H. Kisch,
Organometallics 14 (1995) 3385.
[2] (a) G. Halbritter, F. Knoch, A. Wolski, H. Kisch, Angew. Chem.
Int. Ed. Engl. 106 (1994) 1676; Angew. Chem. Int. Ed. Engl. 33
(1994) 1603. (b) G. Halbritter, F. Knoch, H. Kisch, J. Organomet.
Chem. 492 (1995) 87.
[3] U. Du¨rr, F.W. Heinemann, H. Kisch, J. Organomet. Chem. 541
(1997) 307.
[4] U. Du¨rr, F. Knoch, H. Kisch, Z. Kristallogr. 212 (1997) 173. Atom
numbering has been changed for the sake of comparison with the
mono- and distilbenylazobenzenes.
[5] H. Kisch, F. Knoch, D. Fenske, G. Halbritter, Z. Kristallogr. 210
(1995) 222.
[6] C.J. Brown, Acta Crystallogr. 21 (1966) 146.
[7] C.J. Finder, M.G. Newton, N.L. Allinger, Acta Crystallogr. B 30
(1974) 411.
Irradiating a diluted solution of the 2,3-dihydrocin-
noine 2a ([2a]B10−3 mol l−1) with a laboratory
halogen lamp (12 V/20 W) for 1 min (u\375 nm)
leads to a color change from red to yellow and to a
decreasing absorbance at about 500 nm due to the
ring opening to the corresponding 2,6-distilbenyla-
zobenzene 4a. Upon standing in the dark for 10 min,
the absorbance recovered to its original value. The
corresponding ring closure reaction of 3a produced by
irradiation of 1a needed a reaction time of 80 min.
3.5. Crystal structure determinations
Suitable single crystals of 3b, 3d·CH2Cl2, 4c, cis-3f
and 4f were sealed under N2 in glass capillaries. In-
tensity data were collected at 200 K on a Siemens P4
diffractometer using Mo–Kh radiation (u=71.073
pm, graphite monochromator). Data were corrected
for Lorentz and polarization effects while an absorp-
tion correction has not been applied. All structures
were solved by direct methods [22] and refined using
full-matrix least-squares procedures on F2 values
(SHELXL-93 [23]). All non-H atoms were refined
with anisotropic displacement parameters. The posi-
tions of all H atoms could be localized in a difference
fourier synthesis. With the exception of the hydrogen
atoms of the CH2Cl2 solvate molecule in 3d·CH2Cl2
where both the positional and a thermal parameter
were kept fixed during the refinement, all other H
atoms were refined isotropically. For all compounds
an empirical extinction correction has been applied
using the function (see Table 2). Selected crystallo-
graphic data and further details of the data collection
and refinement of 3b, 3d·CH2Cl2, 4c, cis-3f and 4f are
summarized in Table 2 [24].
[8] A. Streitwieser, C.H. Heathcock, in: Organische Chemie, 2nd
reprint of 1st ed., Verlag Chemie, Weinheim, 1990, p. 118.
[9] G. Halbritter, Personal Communication.
[10] H. Boon, M. Rausch, J. Chem. Soc. Chem. Comm. (1977) 397.
[11] H. Werner, J. Wolf, U. Schubert, K. Ackermann, J. Organomet.
Chem. 317 (1986) 327.
[12] H. Clark, C. Wong, J. Am. Chem. Soc. 99 (1977) 7073.
[13] D. Hart, J. Schwartz, J. Organomet. Chem. 87 (1975) 11.
[14] (a) J.M. Huggins, R.G. Bergman, J. Am. Chem. Soc. 103 (1981)
3002. (b) J.M. Huggins, R.G. Bergman, J. Am. Chem. Soc. 101
(1979) 4410.
[15] U.R. Aulwurm, Ph.D. Thesis, University Erlangen-Nu¨rnberg,
Erlangen, Germany, 1996.
[16] A. Sacco, M. Rossi, Inorg. Synth. XII (1970) 12.
[17] K.H. Schu¨ndehu¨tte, in: E. Muller (Ed.), Houben-Weyl: Methoden
der Organischen Chemie, vol. 10/3, Thieme Verlag, Stuttgart,
1965, p. 373.
[18] K.H. Schu¨ndehu¨tte, in: E. Muller (Ed.), Houben-Weyl: Methoden
der Organischen Chemie, vol. 10/3, Thieme Verlag, Stuttgart,
1965, p. 332.
[19] R.G. Gastinger, E.F. Tokas, M.D. Rausch, J. Org. Chem. 43
(1978) 159.
[20] W. Taylor, A.M. Ward, J. Chem. Soc. 34 (1934) 2006.
[21] M. Hesse, H. Meier, B. Zeeh, Spektroskopische Methoden in der
organischen Chemie, 3rd revised ed, Thieme Verlag, Stuttgart,
1987, p. 152.
[22] G.M. Sheldrick, Acta Crystallogr. A 46 (1990) 467.
[23] G.M. Sheldrick, SHELXL-93. Program for crystal structure
refinement, University of Go¨ttingen, Germany, 1993.
[24] Further details of the single crystal X-ray analyses can be obtained
from the Fachinformationszentrum Karlsruhe, D-76344 Eggen-
stein-Leopoldshafen by citing the depository numbers CSD-
402565 (2a), CSD-407241 (3b), CSD-407242 (3d·CH2Cl2),
CSD-407243 (cis-3f), CSD-407244 (4c) and CSD 407245 (4f), the
names of the authors and the references.
Acknowledgements
This work was supported by Deutsche Forschungs-
gemeinschaft and Fonds der Chemischen Industrie.
We thank Prof. Dr D. Sellmann for assistance in the
single crystal X-ray analyses.
.