Keto-acid chloride and diketone intermediates in the acylation of ferrocene by arylene bis-acid chlorides. Crystal and molecular structure of 4-ferrocenylcarbonyl-4′-chlorocarbonylbiphenyl: Intermolecular association by means of strong C-HO hydrogen bonds and ππ stacking interactions

Article abstract of DOI:10.1016/S0022-328X(99)00252-1

Acylation of ferrocene with arylene bis-acid chlorides yields deep-red polymeric products. However, under controlled conditions two types of intermediate can be isolated and characterised: the keto-acid chlorides FcCO(arylene)COCl 1 [Fc=(C₅H₅)Fe(C₅H₄); a arylene=4,4′-C₆H₄C₆H₄, b arylene=1,3-C₆H₄, c arylene=1,4-C₆H₄], followed in the case of 1a and 1b by the diketones FcCO(arylene)COFc 2a and 2b, although no conversion of 1c to 2c occurs. Similarly, use of 1,3,5-benzenetricarbonyl trichloride yielded 1,3,5-C₆H₃(COFc)₃. In the crystal structure of 1a, there are four independent molecules: pairs of molecules are linked into dimers by strong C-HO hydrogen bonds, and these dimers are linked into chains by ππ stacking interactions.

Full text of DOI:10.1016/S0022-328X(99)00252-1

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 585 (1999) 331–340
Keto-acid chloride and diketone intermediates in the acylation of
ferrocene by arylene bis-acid chlorides. Crystal and molecular
structure of 4-ferrocenylcarbonyl-4%-chlorocarbonylbiphenyl:
intermolecular association by means of strong C–H···O hydrogen
bonds and p···p stacking interactions
S. Zaka Ahmed, George Ferguson 1, Christopher Glidewell *
School of Chemistry, Uni6ersity of St Andrews, St Andrews, Fife KY16 9ST, UK
Received 23 February 1999; received in revised form 20 April 1999
Abstract
Acylation of ferrocene with arylene bis-acid chlorides yields deep-red polymeric products. However, under controlled conditions
two types of intermediate can be isolated and characterised: the keto-acid chlorides FcCO(arylene)COCl 1 [Fc=(C5H5)Fe(C5H4);
a arylene=4,4%-C6H4C6H4, b arylene=1,3-C6H4, c arylene=1,4-C6H4], followed in the case of 1a and 1b by the diketones
FcCO(arylene)COFc 2a and 2b, although no conversion of 1c to 2c occurs. Similarly, use of 1,3,5-benzenetricarbonyl trichloride
yielded 1,3,5-C6H3(COFc)3. In the crystal structure of 1a, there are four independent molecules: pairs of molecules are linked into
dimers by strong C–H···O hydrogen bonds, and these dimers are linked into chains by p···p stacking interactions. © 1999 Elsevier
Science S.A. All rights reserved.
Keywords: Ferrocene derivatives; Acylation by arylene bis-acid chlorides; Intermolecular H-bonding; p–p Stacking
1. Introduction
this contrasts with most other types of ferrocene-con-
taining polymers, whose syntheses from ferrocene gen-
erally require several steps [2–5].
In Friedel–Crafts acylation of ferrocene, the deacti-
vating effect of a newly introduced acyl substituent
leads to (a) a large difference in rates between the first
and second acylation steps, so that monoacyl products
are readily isolable and (b) exclusively hetero-annular
diacylation [1]. When bis-acid chlorides are employed in
the acylating agent, the formation of linear polyketone
polymers is thus to be expected, while tris-acid chlo-
rides should lead to dendrimeric polymers. One advan-
tage of polymer synthesis using acylation methods is
that ferrocene-containing polyketone polymers can be
produced from ferrocene itself in a single-step process:
In seeking methods for the controlled growth of such
polymers, we have now investigated the early steps in
the acylation of ferrocene with some simple arylene
bis-acid chlorides, and here we report the isolation and
characterisation of some of the early intermediates 1
and 2, including the single-crystal X-ray structure of 1a.
The structure of 1a exhibits supramolecular association
employing a combination of strong C–H···O hydrogen
bonds and aromatic p···p stacking interactions.
* Corresponding author. Fax: +44-1334-46308.
E-mail address: cg@st-andrews.ac.uk (C. Glidewell)
1 On leave from: Department of Chemistry and Biochemistry,
University of Guelph, Guelph, Ontario N1G 2W1, Canada.
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(99)00252-1

332
S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
10H, 2×C5H5), 4.62 (m, 4H) and 4.97 (m, 4H) (2×
C5H4), 7.76 and 8.04 (8H, AB, 2×C6H4): l(C) 70.3 (d,
C5H5), 71.5 (d), 72.7 (d) and 78.2 (s) (C5H4), 127.1 (d),
128.8 (d), 139.1 (s) and 143.2 (s) (C6H4), 198.5 (s, CO).
IR (Nujol) w(CO) 1645 cm−1. Procedure C gave, using
similar elution conditions, 4%-chlorocarbonyl-4-biphenyl
ferrocenyl ketone (1a) in 30% yield. Anal. Calc. for
C24H17ClFeO2: C, 67.2; H, 4.0. Found: C, 67.3; H,
3.8%. NMR l(H) 4.26 (s, 5H, C5H5), 4.67 (m, 2H) and
4.98 (m, 2H) (C5H4), 7.75 and 8.05 (4H, AB), 7.82 and
8.28 (4H, AB) (2×C6H4): l(C) 70.4 (d, C5H5), 71.6 (d),
72.9 (d) and 78.1 (s) (C5H4), 127.3 (d), 128.9 (d), 139.8
(s) and 142.0 (s) (C6H4COFc), 127.7 (d), 132.1 (d),
132.4 (s) and 147.0 (s) (C6H4COCl), 168.0 (s, COCl),
198.1 (s, COFc). IR (Nujol) w(CO) 1755 cm−1
.
2.1.3. Reactions with isophthaloyl dichloride
Procedure A produced a deep-red powder, isolated
1
by filtration, whose H-NMR spectrum showed reso-
nances characteristic of C5H5, C5H4 and 1,3-
biphenylene groups in the approximate ratio 1:20:20. In
procedure B, chromatographic elution with CH2Cl2
provided an 83% yield of 1,3-bis(ferrocenecar-
bonyl)benzene (2b) in 83% yield. Anal. Calc. for
C28H22Fe2O2: C, 67.0; H, 4.4. Found: C, 66.8; H, 4.5%.
NMR l(H) 4.26 (s, 10H, 2×C5H5), 4.65 (m, 4H) and
4.95 (m, 4H) (2×C5H4), 7.58 (t, J=12 Hz, 1H), 8.10
(d, J=12 Hz, 2H) and 8.58 (s, 1H) (C6H4): l(C) 70.3
(d, C5H5), 71.5 (d), 72.8 (d) and 77.7 (s) (C5H4), 128.0
(d), 128.4 (d), 130.9 (d) and 139.5 (s) (C6H4), 198.5 (s,
CO). IR (Nujol) w(CO) 1640 cm−1. Using similar elu-
tion conditions, procedure C gave a very small quantity
of (3-chlorocarbonylphenyl) ferrocenyl ketone (1b)
(yield B10%). NMR l(H) 4.30 (s, 5H, C5H5), 4.70 (m,
2H) and 4.90 (m, 2H) (C5H4), 7.66 (m, 1H), 8.18 (d,
1H0, 8.32 (d, 1H0 and 8.84 (s, 1H) (C6H4): l(C) 70.8
(d, C5H5), 71.5 (d), 73.2 (d) and 77.4 (s) (C5H4), 129.4
(d), 130.9 (d), 133.2 (s), 133.9 (d), 134.7 (d) and 140.4
(s) (C6H4), 168.0 (s, COCl), 198.1 (s, COFc). IR (Nujol)
2. Experimental
2.1. Synthesis
2.1.1. Acylation reactions using arylene bis(carbonyl
chlorides): general procedures
Stoichiometric quantities of ferrocene dissolved in
dry CH2Cl2 were added to a well-stirred solution con-
taining the acid chloride and anhydrous AlCl3. This
mixture was then stirred either at 0°C or at reflux
temperature for the required time. Work-up consisted
of quenching the reaction with ice/water, filtration or
extraction of the organic material with CH2Cl2, wash-
ing of the organic material until neutral. For soluble
products, the extract was then dried, concentrated and
chromatographed on silica. Three variants were em-
ployed: in procedure A, one equivalent of ferrocene per
mole of acid chloride was employed, and the reaction
mixture was refluxed for 8 h, producing copious quanti-
ties of a deep-red solid, isolated by filtration; in proce-
dure B, two equivalents of ferrocene were employed,
and reactions were refluxed for 8–10 h; in procedure C,
one equivalent of ferrocene was used, the reaction was
run at 0°C and its progress was monitored by TLC
until the abundance of the first intermediate was judged
to be at a maximum, when the reaction was quenched.
w(CO) 1770 cm−1
.
2.1.4. Reactions with terephthaloyl dichloride
Procedure B with 24 h reflux, provided (4-chlorocar-
bonylphenyl) ferrocenyl ketone (1c) in 20% yield. NMR
l(H) 4.25 (s, 5H, C5H5), 4.68 (m, 2H) and 4.90 (m, 2H)
(C5H4), 7.98 and 8.24 (4H, AB, C6H4): l(C) 70.4 (d,
C5H5), 71.4 (d), 73.4 (d) and 77.2 (s) (C5H4), 128.2 (d),
131.3 (d), 135.1 (s) and 145.4 (s) (C6H4), 167.9 (s,
2.1.2. Reactions with 4,4%-bis(chlorocarbonyl)biphenyl
Procedure A produced a deep-red powder, isolated
by filtration, whose H-NMR spectrum showed reso-
COCl), 198.1 (s, COFc). IR (Nujol) w(CO) 1765 cm−1
.
1
nances characteristic of C5H5, C5H4 and 4,4%-biphenylyl
groups in the approximate ratio 1:3:3. In procedure B
chromatography provided, upon elution with 10%
petrol in CH2Cl2, 4,4%-bis(ferrocenecarbonyl)biphenyl
2a in 80% yield. Anal. Calc. for C34H26Fe2O2: C, 70.5;
H, 4.5. Found: C, 70.5; H, 4.4%. NMR l(H) 4.27 (s,
2.1.5. Acylation of ferrocene using
1,3,5-benzenetricarbonyl trichloride
A mixture of 1,3,5-benzenetricarbonyl trichloride
(0.66 g, 2.5 mmol), anhydrous AlCl3 (2.0 g, 15 mmol)
and ferrocene (1.67 g, 9 mmol) in dry CH2Cl2 (100 cm3)
was refluxed for 16 h. The mixture was worked up as

S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
333
Table 1
Table 1 (Continued)
Experimental details for 1a
Computer programs
Crystal data
Chemical formula
Chemical formula weight 428.68
Data collection
Cell refinement
CAD4-PC Software [6]
SET4 and CELDIM in CAD4-PC Software
[6]
C
24H17ClFeO2
Crystal system
Space group
Monoclinic
Cc
Data reduction
DATRD2 in NRCVAX-96 [7]
SOLVER in NRCVAX-96
NRCVAX-96 and SHELXL-97 [9]
PLATON [8]
NRCVAX-96, SHELXL-97 and
WordPerfect macro PREP-8 [10]
Structure solution
Structure refinement
Graphics
Preparation of material
for publication
,
a (A)
14.8547(9)
18.3258(17)
24.757(2)
90
97.043(6)
90
7769.4(11)
16
1.466
,
b (A)
,
c (A)
h (°)
i (°)
k (°)
3
,
V (A )
usual and chromatographed on silica, when CH2Cl2
eluted unreacted ferrocene (1.1 g, 5.9 mmol) and 1,3,5-
tris(ferrocenecarbonyl)benzene (0.5 g, 23% based on
ferrocene consumed). NMR l(H) 4.30 (s, 15H, 3×
C5H5), 4.70 (m, 6H) and 4.95 (m, 6H) (3×C5H4), 8.78
(s, 3H, C6H3): l(C) 71.2 (d, C5H5), 72.2 (d), 73.8 (d)
and 778.7 (s) (C5H4), 131.7 (d) and 140.4 (s) (C6H3),
Z
Dcalc (Mg m−3
Radiation type
)
Mo–Ka
0.71073
,
Wavelength (A)
No. of reflections for cell 25
parameters
q Range for cell
8.90–14.29
measurement (°)
v (mm−1
)
0.931
294(1)
Plate
0.38×0.35×0.04
Orange
198.3 (s, CO). IR (Nujol) w(CO) 1645 cm−1
.
Temperature (K)
Crystal form
Crystal size (mm)
Crystal colour
2.2. X-ray crystallography
Crystals of 1a suitable for single-crystal X-ray dif-
fraction were grown form CH2Cl2 solution. Details of
the crystal data, diffraction data collection, and struc-
ture solution and refinement [6–10] are given in Table
1: the 2% intensity decay during data collection was
allowed for by linear scaling.
Data collection
Diffractometer
Data collection method
Absorption correction
No. of measured
reflections
Enraf–Nonius CAD
q/2q Scans
Gaussian
4
12 534
No. of independent
reflections
No. of observed
reflections
Criterion for observed
reflections
Rint
12 232
4011
The systematic absences permitted two possible space
groups, Cc and C2/c. The unit cell dimensions, together
with an estimated density of ca. 1.4–1.5 g cm−3 typical
of the densities found in other acyl ferrocenes, indicated
Z=16, i.e. Z%=4 [11] for Cc or Z%=2 for C2/c.
However, attempts to solve the structure, either by
direct methods or by Patterson methods, were uni-
formly unsuccessful in both of the possible space
groups. Careful inspection of the intensity statistics had
already shown that the population of the parity groups
eee and eeo was far higher than that of the groups ooe
and ooo, while the other four groups were systemati-
cally absent. This intensity distribution suggested an
alternative approach to the structure solution.
Use of only the eee and eeo data, and conversion to
a quarter sized unit cell (a“a/2, b“b/2, c“c) with an
appropriate origin shift allowed a straightforward Pat-
terson solution in P21/c with Z%=1. Development of
this solution revealed all the carbon atoms of the
ferrocenyl fragment with full occupancy, and two
COC6H4C6H4COCl fragments, one bonded to each cy-
clopentadienyl ring and each having site occupation
factor 0.5. Refinement of this disordered P21/c solution
using only the eee and eeo groups of the original data,
suitably transformed, converged at R=0.05.
[I\2|(I)]
0.015
25.12
qmax (°)
Index ranges
−17BhB17, 0BkB21, −34BlB27
3
No. of standard
reflections
Frequency of standard
reflections (min)
Intensity decay (%)
120
2.0
Refinement
Refinement on
R[F2\2|(F2)]
wR(F2)
F2
0.0464
0.1349
0.835
S
No. of reflections used in 12 232
refinement
No. of parameters used
No. of restraints
H-atom treatment
Weighting scheme
863
1780
Riding
w=1/[|2(F2o)+(0.0582P)2], where
P=(F2o+2F2c)/3
0.003
(D/|) max.
−3
,
Dzmax (e A
)
0.280
−0.262
−3
,
Dzmin (e A
)
Source of atomic
scattering factors
International Tables for
Crystallography (vol. C)
Conversion of the P21/c solution back to the original
cell and testing the various possibilities for the location

334
S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
Fig. 1. Views of the four independent molecules in the crystal structure of 1a, showing the atom-labelling scheme and the rotational disorder of
the chlorocarbonyl groups in molecules 1, 3 and 4. Displacement ellipsoids are drawn at the 30% probability level.

S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
335
Table 2 (Continued)
Table 2
Fractional atomic coordinates and equivalent isotropic displacement
a
2
Atom
x
y
z
Ueq
,
parameters (A ) for 1a
a
C33
C34
C35
C36
C37
C38
C39
0.1838(7)
0.2479(6)
0.2381(6)
0.0200(8)
0.2348(5)
0.2922(6)
0.3236(5)
0.4058(6)
0.3479(6)
0.3467(6)
0.4039(7)
0.4405(6)
0.3556(4)
0.3984(5)
0.4491(4)
0.4569(4)
0.4140(4)
0.3634(4)
0.2986(6)
0.2609(5)
0.6165(4) −0.0836(2) 0.056(3)
0.5863(4) −0.0914(2) 0.069(3)
0.5334(4) −0.0610(3) 0.064(3)
0.5109(4) −0.0230(3) 0.057(3)
0.5411(4) −0.0152(3) 0.062(3)
0.5940(4) −0.0456(3) 0.066(3)
0.6726(4) −0.1160(2) 0.077(3)
0.6944(10) −0.1207(6) 0.082(4)
0.7136(5) −0.1530(3) 0.093(3)
0.7120(13) −0.1438(8) 0.093(3)
0.6848(11) −0.1144(7) 0.082(4)
0.2243(3) 0.054(3)
0.2242(3) 0.053(3)
0.1787(4) 0.047(3)
0.1821(3) 0.073(4)
0.2063(4) 0.082(4)
0.2529(4) 0.087(4)
0.2575(3) 0.077(4)
0.2137(4) 0.068(4)
0.0743(2) 0.056(3)
0.0468(3) 0.061(3)
0.0160(3) 0.061(3)
0.0127(2) 0.052(3)
0.0403(3) 0.051(3)
0.0711(3) 0.054(3)
0.1047(3) 0.053(3)
0.0906(3) 0.082(3)
Atom
x
y
z
Ueq
Fe1
C11
C12
C13
C14
C15
C16
C17
0.39476(17) 0.07787(14) 0.28922(8) 0.0511(8)
0.3482(7) 0.0268(6)
0.3441(3) 0.053(3)
0.3122(4) 0.069(4)
0.2679(3) 0.069(4)
0.2724(3) 0.062(4)
0.3195(4) 0.060(4)
0.3130(3) 0.071(4)
0.2806(4) 0.065(3)
0.2370(3) 0.067(4)
0.2424(4) 0.076(4)
0.2894(4) 0.074(4)
0.4221(2) 0.052(3)
0.4478(3) 0.064(3)
0.4791(3) 0.061(3)
0.4848(2) 0.052(3)
0.4591(3) 0.055(3)
0.4277(3) 0.061(3)
0.3914(3) 0.055(3)
0.4053(3) 0.073(3)
0.5801(2) 0.060(3)
0.5866(2) 0.061(3)
0.5556(3) 0.063(3)
0.5180(3) 0.049(3)
0.5115(2) 0.060(3)
0.5425(3) 0.064(3)
0.6132(3) 0.085(4)
0.6406(10) 0.128(4)
0.6126(6) 0.118(3)1
0.6494(5) 0.128(4)
−0.0173(7)
0.0273(8)
0.0922(8)
0.0877(8)
0.1502(6)
0.0874(4)
0.1168(5)
0.2089(6)
0.2716(4)
0.2423(5)
0.1170(7)
0.0486(7)
0.2884(5)
0.2021(5)
0.1760(5)
0.2363(6)
0.3226(5)
0.3487(4)
0.3178(6)
0.3925(9)
0.2316(5)
0.274(3)
0.3738(7) −0.0263(5)
0.3227(8) −0.0134(6)
0.2656(7)
0.2814(7)
0.4995(7)
0.5274(7)
0.4731(7)
0.4116(7)
0.4279(7)
0.3624(7)
0.4263(5)
0.3991(5)
0.3079(6)
0.2439(4)
0.2711(6)
0.3931(8)
0.4625(6)
0.2260(5)
0.3126(5)
0.3374(5)
0.2757(6)
0.1891(5)
0.1642(4)
0.1988(7)
0.256(2)
0.0477(6)
0.0725(5)
0.1408(7)
0.0895(6)
0.1001(6)
0.1579(7)
0.1831(6)
0.1016(4)
0.1457(5)
0.1965(4)
0.2031(4)
0.1590(4)
0.1082(4)
0.0428(6)
0.0091(5)
0.3630(4)
0.3331(4)
0.2807(4)
0.2584(3)
0.2883(4)
0.3406(4)
0.4188(4)
0.4486(18)
0.4411(8)
0.4601(9)
C310
C311
C312
C313
C314
C315
C316
C317
O31
C321
C322
C323
C324
C325
C326
C327
O32A
Cl3A
O32B
Cl3B
Fe4
C18
C19
C110
C111
C112
C113
C114
C115
C116
C117
O11
C121
C122
C123
C124
C125
C126
C127
O12A
Cl1A
O12B
Fe2
0.4349(8)
0.11508(18) 0.82839(13) 0.20849(9) 0.0500(7)
C41
C42
C43
C44
C45
C46
C47
C48
0.0158(7)
−0.0214(7)
0.0182(7)
0.0799(7)
0.0784(7)
0.1806(8)
0.1349(7)
0.1701(7)
0.2374(7)
0.2440(6)
0.0298(6)
0.0626(6)
0.0893(6)
0.0832(6)
0.0504(6)
0.0237(6)
−0.0033(7)
−0.0497(5)
0.1661(6)
0.1478(7)
0.1212(7)
0.1129(6)
0.1312(6)
0.1578(6)
0.1992(5)
0.2312(11)
0.1950(5)
0.2182(13)
0.2101(11)
0.8932(6)
0.8413(5)
0.8530(6)
0.9122(6)
0.9371(5)
0.7760(7)
0.7246(6)
0.7310(6)
0.7863(6)
0.8142(6)
0.9407(4)
0.9160(3)
0.9658(5)
1.0402(5)
1.0649(3)
1.0151(5)
0.8839(6)
0.8318(5)
0.1745(3) 0.046(3)
0.2034(4) 0.060(3)
0.2503(3) 0.053(3)
0.2503(3) 0.056(3)
0.2035(4) 0.050(3)
0.1588(3) 0.071(4)
0.1845(4) 0.076(4)
0.2325(4) 0.074(4)
0.2365(3) 0.076(4)
0.1910(4) 0.060(3)
0.0925(3) 0.052(3)
0.0521(3) 0.074(4)
0.0200(3) 0.074(4)
0.0283(3) 0.059(3)
0.0687(3) 0.061(3)
0.1008(3) 0.056(3)
0.1236(3) 0.051(3)
0.1060(3) 0.066(3)
0.0844(7)
0.1187(11)
0.39785(16) 0.58005(14) 0.29418(8) 0.0483(7)
C21
C22
C23
C24
C25
C26
C27
C28
0.4970(7)
0.5317(6)
0.4838(7)
0.4194(7)
0.4276(7)
0.3436(9)
0.3765(8)
0.3299(9)
0.2681(8)
0.2766(8)
0.4794(6)
0.4470(6)
0.4211(6)
0.4275(5)
0.4598(6)
0.4858(6)
0.5160(8)
0.5590(6)
0.3415(5)
0.3638(6)
0.3915(6)
0.3970(5)
0.3747(6)
0.3469(6)
0.3129(4)
0.2918(7)
0.3083(4)
0.6415(6)
0.5957(6)
0.6113(6)
0.6669(6)
0.6855(5)
0.5220(7)
0.4751(6)
0.4923(7)
0.5498(7)
0.5682(6)
0.6946(4)
0.6714(3)
0.7223(5)
0.7964(4)
0.8196(3)
0.7687(5)
0.6404(6)
0.5868(5)
0.9483(4)
0.8768(4)
0.8286(3)
0.8520(4)
0.9235(4)
0.9716(3)
0.9992(4)
0.9830(6)
1.0914(3)
0.3265(3) 0.051(3)
0.2931(4) 0.055(3)
0.2484(3) 0.060(3)
0.2541(3) 0.049(3)
0.3024(3) 0.049(3)
0.3418(3) 0.090(4)
0.3084(4) 0.082(4)
0.2637(3) 0.091(4)
0.2694(4) 0.089(4)
0.3177(5) 0.092(4)
0.4096(3) 0.055(2)
0.4505(3) 0.064(3)
0.4821(2) 0.062(3)
0.4729(3) 0.051(2)
0.4320(3) 0.061(3)
0.4004(2) 0.056(3)
0.3778(3) 0.059(3)
0.3964(3) 0.082(3)
0.5717(2) 0.066(3)
0.5845(2) 0.077(3)
0.5518(3) 0.072(3)
0.5062(3) 0.051(2)
0.4934(2) 0.066(3)
0.5262(2) 0.067(3)
0.6074(3) 0.079(3)
0.6477(3) 0.080(3)
0.5909(2) 0.146(2)
C49
C410
C411
C412
C413
C414
C415
C416
C417
O41
C421
C422
C423
C424
C425
C426
C427
O42A
Cl4A
O42B
Cl4B
C29
C210
C211
C212
C213
C214
C215
C216
C217
O21
C221
C222
C223
C224
C225
C226
C227
O22
1.1973(4) −0.0690(2) 0.073(3)
1.1262(4) −0.0842(2) 0.096(4)
1.0744(3) −0.0533(3) 0.095(4)
1.0937(4) −0.0073(3) 0.061(3)
1.1647(5)
0.0079(2) 0.072(3)
1.2166(4) −0.0230(3) 0.079(4)
1.2520(5) −0.1006(3) 0.090(4)
1.2365(19) −0.1364(5) 0.136(5)
1.3447(3) −0.0889(3) 0.066(2)
1.3148(8) −0.0911(9) 0.066(2)
1.2196(11) −0.1556(4) 0.136(5)
Refined site-occupation factors
O12A and
Cl1A
O12B and
Cl1B
Cl2
Fe3
C31
C32
0.489(11)
0.11857(16) 0.33022(14) 0.20654(9) 0.0507(8)
0.1680(6)
0.1344(6)
0.2855(6)
0.2306(5)
0.1507(3) 0.048(3)
0.1788(3) 0.053(3)
0.511(13)

336
S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
Table 2 (Continued)
a
Atom
x
y
z
Ueq
O32A and
Cl3A
O32B and
Cl3B
O42A and
Cl4A
O42B and
Cl4B
0.697(11)
0.303(11)
0.563(11)
0.437(11)
For 3a and 3b the NMR integration indicated mean
(
values, n, of ca. 3 and 24, respectively: it seems plausible
that chain length is limited by the solubility of the
growing polymer in the reaction medium, here of rather
low polarity.
a Ueq=(1/3)SiSjUija*a*aiaj.
i
j
The principal purpose of the present study, however,
has not been the characterisation of these polymeric
products, but rather the characterisation of the early
intermediates in polymer formation: by appropriate
choice of reaction conditions it is possible to isolate and
identity two early intermediates. When reactions with the
same two acylating agents were carried out using a
two-fold excess of ferrocene, so that no hetero-annular
diacylation was required before the complete consump-
tion of the acylating agent, the diketones 2a and 2b could
be isolated in good yields. These products were identified
by elemental analysis and by means of their particularly
simple and characteristic 1H- and 13C-NMR spectra. The
spectra for each of 2a and 2b contained the characteristic
resonances of a mono-substituted ferrocene, the charac-
teristic resonances of the symmetrically disubstituted
arylene units 4,4%-C6H4C6H4 in 2a and 1,3-C6H4 in 2b,
together with a single ketonic carbonyl resonance in the
13C spectrum. Both crystallised with a very tiny crystal
habit and, despite repeated efforts, no crystals suitable
for single-crystal X-ray diffraction could be obtained.
When the reaction which gave compounds 2a and 2b
were monitored by TLC, it was apparent that, in each
case, formation of the diketone 2 was preceded by
formation of another compound. TLC showed initial
formation of an intermediate less polar than 2, whose
abundance grew to a maximum and then fell away as the
abundance of 2 increased: the procedures used for the
optimum production of compounds 2a and 2b involved
reactions run until this less polar intermediate has wholly
disappeared.
of the COC6H4C6H4COCl groups confirmed the space
group as Cc. Refinement, with the aryl and cyclopenta-
dienyl rings modelled as planar regular pentagons and
hexagons, respectively, was possible only in Cc. No
solution or refinement was feasible in C2/c. At this point
it was clear that the four independent iron atoms in Cc
generated an almost exact P21/c pattern of coordinates
in the quartered cell. Because of the correlation intro-
duced into the refinement by this pseudo-symmetry, the
disorder of the CoCl groups in three of the four indepen-
dent molecules, and the large number of parameters
required for full anisotropic refinement, the subsequent
refinement in Cc was carried out with the application of
heavy restraints on the values of the displacement
parameters. Refinement of the Flack parameter [12] gave
a value of 0.56(5), implying racemic twinning [9].
Refined atom coordinates are in Table 2, and selected
geometric parameters are in Table 3. Fig. 1 shows the
four independent molecules in the structure, Fig. 2 shows
the hydrogen-bonded dimer formed between molecules
are types 1 and 3, Fig. 3 shows one pair of molecules
linked by aromatic p···p stacking, Fig. 4 shows one of the
(
chains running parallel to the [110] directions, and Fig.
(
5 shows a stereoview of the stacked [110] and [110] chains.
3. Results and discussion
3.1. Acylation reactions
Friedel–Crafts acylation of ferrocene using equimolar
quantities of ferrocene and either 4,4%-bis(chlorocar-
bonyl)biphenyl or isophthaloyl dichloride lead, using
extended reaction times in refluxing dichloromethane, to
the precipitation of deep-red sparingly soluble materials,
In order to isolate and identify these early intermedi-
ate, reactions of 1:1 stoichiometry designed to optimise
their formation were conducted at 0°C with frequent
monitoring by TLC: when the abundance of the early
intermediate was judged to be at a maximum, each
reaction was quenched and worked up in the usual
manner. In this way, compounds 1a and 1b were isolated,
albeit in very low yield in the case of 1b. The constitution
of 1a was established by elemental analysis and NMR,
and was confirmed by single-crystal X-ray structure
analysis (see Section 3.2). The 13C-NMR spectra of 1a
1
whose H-NMR spectra were consistent with their for-
mulation as ferrocene-terminated polyketone polymers
of type 3.

S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
337
and 1b both contained two resonances in the carbonyl
region, one at ca. l 198 characteristic of a ketone and
a second at ca. l 168; this is consistent with the
presence of either a carboxylic acid or an acid chlo-
ride, and initially it was thought that the presence of a
carboxyl group arising from hydrolysis during the
work-up procedure was the more plausible. However
the X-ray analysis of 1a showed plainly that no hy-
drolysis had occurred. The near identity of the 13C
chemical shifts in 1a and 1b, particularly for the car-
bonyl groups, suggested that 1b was also an acid chlo-
ride, rather than a carboxylic acid. This was confirmed
by the lack of the characteristic carboxylic acid ab-
sorption in the IR spectra of 1a and 1b: in addition
the w(CO) values in 1a and 1b were within the range
normally observed for aroyl chlorides, but significantly
higher than those normally observed for aromatic car-
boxylic acids.
isomeric terephthaloyl dichloride was employed no
diketone product was formed, even in prolonged reac-
tions. The reaction conditions which led, in 8–10 h, to
good yields of 2a and 2b gave only 1c, even after 24 h.
This product was again identified as an acid chloride
by virtue of its carbonyl shift at l 167.9, virtually
identical with those in 1a and 1b, by the lack of
carboxylic IR absorption, and by the high value of
w(CO). Plainly the subsequent reaction of 1c as an
acylating agent is extremely slow, while that of 1b is
the fastest of the three acid chlorides of type 1, as
judged by the very low maximal abundance of 1b and
the high isolated yield of 2b.
With 1,3,5-benzenetricarbonyl trichloride, although
the reaction with ferrocene is also slow, a 24 h reflux
in CH2Cl2 permits the isolation of modest quantities
of the triketone 4c, again readily identified by its ex-
1
tremely simple H- and 13C-NMR spectra. These com-
By contrast with the acylation behaviour of isoph-
thaloyl dichloride to form 1b and then 2b, when the
prise the characteristic resonances for a single type of
Table 3
,
Selected torsional angles (°), hydrogen bonds (A, °) and p···p interactions for 1a
(a) Torsional angles
C122–C121–C127–O12A
C122–C121–C127–Cl1A
C122–C121–C127–O12B
C122–C121–C127–Cl1B
14.0(8)
−165.9(8)
−163.8(11)
16.2(11)
C113–C114–C124–C123
C213–C214–C224–C223
C313–C314–C324–C323
C413–C414–C424–C423
−25.6(7)
20.6(7)
−21.2(7)
19.1(8)
C222–C221–C227–O22
C222–C221–C227–Cl2
−12.2(5)
167.9(4)
C112–C111–C117–O11
C212–C211–C217–O21
C312–C311–C317–O31
C412–C411–C417–O41
37.6(14)
−35.8(14)
39.3(14)
C322–C321–C327–O32A
C322–C321–C327–Cl3A
C322–C321–C327–O32B
C322–C321–C327–Cl3B
−163.9(8)
16.2(8)
19.9(10)
−160.1(10)
−41.9(12)
C12–C11–C117–O11
C22–C21–C217–O21
C32–C31–C317–O31
C42–C41–C417–O41
2.7(1)
−10.5(17)
−2.4(15)
0.8(14)
C422–C421–C427–O42A
C422–C421–C427–Cl4A
C422–C421–C427–O42B
C422–C421–C427–Cl4B
−16.9(6)
−163.2(5)
−179.1(9)
0.9(9)
(b) Hydrogen bonds
D–H
H···A
D···A
D–H···A
C123–H123···O31 a
C223–H223···O41 b
C323–H323···O11 c
C423–H423···O21 d
0.93
0.93
0.93
0.93
2.37
2.49
2.39
2.45
3.265(13)
3.391(11)
3.294(12)
3.368(11)
161
164
164
168
(c) p···p Interactions
Ring 1
,
,
Ring 2
Dihedral angle (°)
Perp. distance (A)
Offset (A)
e
C12n
C22n
C32n
C42n
1.1
1.8
3.52
3.72
1.32
1.26
f
a Symmetry codes: 0.5+x, 0.5−y, 0.5+z.
b 0.5+x, 1.5−y, 0.5+z.
c −0.5+x, 0.5−y, −0.5+z.
d −0.5+x, 1.5−y, −0.5+z.
e x, 1−y, 0.5+z.
f x, 2−y, 0.5+z.

338
S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
(
Fig. 4. View of one of the [110] chains built from molecules of types
1 and 3: atoms are depicted as in Fig. 3.
Fig. 2. The cyclic dimer formed by molecules of types 1 and 3,
showing the R22(18) ring formed by the two C–H···O hydrogen bonds.
mono-substituted ferrocene together with a single aro-
matic resonance in the 1H spectrum and, in the 13C
spectrum, two aromatic resonances, one C–H and one
quaternary, and a ketonic carbonyl resonance.
Aside from the characteristic resonances of a mono-
1
substituted ferrocene, the H spectrum of 4c contains
only a single sharp absorption, while the 13C spectrum
contains just three other resonances, two arising from
the aryl ring and one from the ketonic carbonyl group.
Monitoring of this acylation by TLC showed that, as
expected, there were two intermediates formed before
4c and these are plausibly identified as the mono- and
bi-ferrocenyl species 4a and 4b, respectively, although
no attempt was made to isolate these.
3.2. Crystal and molecular structure of 1a
Fig. 3. View of one pair of p-stacked molecules, formed by molecules
of types 1 and 3, showing the ring offset approximately equal to the
radius of an aryl ring. For the sake of clarity, only the major
conformation of the chlorocarbonyl groups is shown, and H atoms
are omitted.
The asymmetric unit of compound 1a consists of four
independent molecules (Fig. 1), and it was clear from
an early stage in the refinement that this compound is

S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
339
indeed an acyl chloride rather than a carboxylic acid,
as demonstrated by the bond distances and the be-
haviour of the displacement parameters. In three of
these molecules, those containing Fe1, Fe3 and Fe4,
the terminal chlorocarbonyl groups is disordered with
two orientations of this fragment differing by a 180°
rotation about the exocyclic C(aryl)–C(O)Cl bond
(Fig. 1). For the three molecules disordered in this
manner, the two orientations are in the ratios
0.51(1):0.49(1), 0.70(1):0.30(1) and 0.56(1):0.44(1), re-
spectively. Further evidence that there are no COO
groups in 1a is provided by the absence of the char-
acteristic association into R22(8)dimers [13].
As is often observed in systems where there is an
excess of hard [14] hydrogen-bond acceptors, in this
case the ketonic oxygen atoms, there are C–H···O
hydrogen bonds with aryl carbon atoms acting as the
donors. The pairwise development of such interac-
tions serves to link pairs of molecules into cyclic
dimers (Table 3). The C–H group C123–H in
molecule 1 at (x, y, z) acts as a donor towards the
ketonic oxygen O31 in molecule 3 at (0.5+x, 0.5−y,
0.5+z), while C323–H in molecule 3 at (0.5+x,
0.5−y, 0.5+z) in turn acts as donor to O11 in
molecule 1 at (x, y, z), so generating a dimeric motif
(Fig. 2) characterised by an R22(18)ring [13]. Similarly
molecule 2 at (x, y, z) and molecule 4 at (0.5+x,
1.5−y, 0.5+z) are linked by a pair of C–H···O hy-
drogen bonds into a second R22(18)motif. These four
independent C–H···O hydrogen bonds (Table 3) all
while ketonic carbonyl groups are second only to sul-
foxide and sulfone SꢀO groups as neutral hydrogen-
bond acceptors. The two independent R22(18)rings are
both approximately centrosymmetric around the
points (0.5, 0.125, 0.5) and (0.5, 0.625, 0.5), respec-
tively: in the P21/c solution, these rings would be
precisely centrosymmetric.
These dimers are themselves linked into chains by
means of p···p stacking interactions [16] in which aryl
rings of molecules in neighbouring dimer units are
essentially parallel with perpendicular separations of
,
3.5–3.7 A and with their ring centroids offset by ap-
proximately the radius of an aryl ring (Fig. 3 and
Table 3), the ideal for such interactions [16]. The ring
C12n (n=1–6) in molecule 1 within the dimer cen-
tred near (0.5, 0.125, 0.5) interacts in this manner
with ring C32n in molecule 3 within the dimer
around (0, 0.625, 0.5), while molecule 3 within the
(0.5, 0.125, 0.5) dimer interacts similarly with
molecule 1 in the dimer around (1, −0.375, 0.5).
(
These interactions generate a chain parallel to [110]
and passing through the point (0.5, 0, 0.5) (Fig. 4).
The dimers formed by molecules of types 2 and 4
form a similar chain, parallel to the first but passing
through the point (0.5, 0.5, 0.5). Each of these chains
is in fact generated by the action of the C-centring
operation in the space group Cc.
This array of parallel chains utilises only half the
contents of the unit cell, and lies largely in the do-
main 0.2BzB0.75. A similar array lying in the do-
main 0.7BzB1.25 is generated from the first by the
action of the glide planes and runs parallel to the
[110] direction. The entire structure is built up from a
stack of these arrays, having alternating directions in
a criss-cross mesh, where the angle between the direc-
tions of the crossing chains is ca. 78° (Fig. 5).
,
have d(H···O) (un-normalised)B2.50 A, d(C···O)B
,
3.40 A, and angle(C–H···O)\160°. Hence all should
be regarded as strong examples of their type [15]. The
strength of these hydrogen bonds depends upon the
fact that aromatic carbon atoms are amongst the best
donors in C–H···A (A, acceptor) hydrogen bonds,
(
Fig. 5. Stereoview of the stacked [110] and [110] chains.

340
S.Z. Ahmed et al. / Journal of Organometallic Chemistry 585 (1999) 331–340
5. Supplementary material
4. Conclusions
Additional data, comprising full details of the
molecular geometries and the displacement parame-
ters, have been deposited at the Cambridge Crystallo-
graphic Data Centre, CSD-116264 for compound 1a.
The acylation of ferrocene using arylene bis-acid
chlorides proceeds via the keto-acid chlorides 1 and
the diketones 2. However, at no stage was any evi-
dence found for hetero-annular bis-diacylation of a
single ferrocene nucleus to provide products of type
Fe[C5H4CO(arylene)COCl]2. Hence acylation of an
unsubstituted ferrocene nucleus by the species 1a and
1b, to give 2a and 2b, respectively, is significantly
faster than the further acylation of the intermediates
of type 1 by the initial bis-acid chlorides: on the
other hand, the acylation reactions of 1c appear to be
exceedingly slow, so that no diketone 2c was ob-
tained. Hence hetero-annular bis-diacylation occurs
for 2a and 2b, but not for 1a or 1b.
Acknowledgements
We are indebted to Dr J.F. Gallagher and Dublin
City University for funds to purchase the X-ray tube
used for the data collection. S.Z.A. thanks the Com-
mittee of Vice-Chancellors and Principals (UK), and
the University of St Andrews for financial support.
The isolation of the primary reaction products as
the acid chlorides 1 rather than as the corresponding
carboxylic acids implies very slow hydrolysis in the
aqueous work-up and subsequent chromatography. In
general, acyl halides undergo rapid hydrolysis only if
they are appreciably soluble in water. Aromatic acyl
halides are poorly soluble in water and hence un-
dergo much slower hydrolysis [17]: thus even benzoyl
chloride is scarcely soluble in water, so allowing ben-
zoylations of, for example, alcohols and amines to be
conducted efficiently in alkaline aqueous solution. The
acid chlorides 1 are expected to be even less soluble
in water than benzoyl chloride, so leading to ex-
tremely slow hydrolysis.
References
[1] D.E. Bublitz, K.L. Rinehart Jr., Org. React. 17 (1969) 1.
[2] P. Gomez-Elipe, P.M. MacDonald, I. Manners, Angew. Chem.
Int. Ed. Engl. 36 (1997) 762.
[3] R. Rulkens, D.P. Gates, J.K. Pudelski, D. Balaishis, D.F. McIn-
tosh, A.J. Lough, I. Manners, J. Am. Chem. Soc. 119 (1997)
10976.
[4] G.E. Southard, M.D. Curtis, Organometallics 16 (1997) 5618.
[5] P. Zanello, M. Fontani, S.Z. Ahmed, C. Glidewell, Polyhedron
17 (1998) 4155.
[6] CAD4-PC Software, Version 1.1, Enraf–Nonius, Delft, Holland,
1992.
[7] E.J. Gabe, Y. Le Page, J.-P. Charland, F.L. Lee, P.S. White, J.
Appl. Cryst. 22 (1989) 384.
[8] A.L. Spek, PLATON Molecular Geometry and Graphics Pro-
gram, Version November 1998, University of Utrecht, Holland,
1998.
[9] G.M. Sheldrick, SHELXL-97, Program for the refinement of
crystal structures, University of Go¨ttingen, Germany, 1997.
[10] G. Ferguson, PREP8—a WordPerfect-5.1 macro to merge and
polish CIF format files from NRCVAX and SHELXL-97 programs,
University of Guelph, Canada, 1998.
[11] C.P. Brock, J.D. Dunitz, Chem. Mater. 6 (1994) 1118.
[12] H.D. Flack, Acta Crystallogr. A39 (1983) 876.
[13] J. Bernstein, R.E. Davis, L. Shimoni, N.-L. Chang, Angew.
Chem. Int. Ed. Engl. 34 (1995) 1555.
The crystal structure of compound 1a was origi-
nally undertaken to provide proof of constitution: de-
spite
the
aqueous
work-up
followed
by
chromatography on silica, compounds 1a–1c were all
deduced from spectral and analytical data to be acid
chlorides rather than carboxylic acids. The single-crys-
tal X-ray diffraction study certainly provided unam-
biguous proof of the acid chloride constitution and,
in addition, revealed extensive intramolecular associa-
tion involving both strong C–H···O hydrogen bonds
and aromatic p···p interactions. Thus a study origi-
nally intended simply as a structure proof turned out
to reveal highly interesting supramolecular interac-
tions.
[14] D. Braga, F. Grepioni, K. Biradha, V.R. Pedireddi, G.R. Desir-
aju, J. Am. Chem. Soc. 117 (1995) 3156.
[15] (a) G.R. Desiraju, Acc. Chem. Res. 24 (1991) 290. (b) G.R.
Desiraju, Acc. Chem. Res. 29 (1996) 441.
[16] C.A. Hunter, Chem. Soc. Rev. 23 (1994) 101.
[17] M.A. Ogliaruso, J.F. Wolfe, in: C.J. Moody (Ed.), Comprehen-
sive Organic Functional Group Transformations, vol. 5, Perga-
mon, Oxford, 1995, p. 26.
.