Experimental and DFT study of cyclodehydration and acetylation of ferrocenyl diols

Article abstract of DOI:10.1016/j.molstruc.2012.03.071

Racemic ferrocenyl diols, i.e. ferrocenyl(2-hydroxymethylphenyl)methanol (2), ferrocenyl-2-(2-hydroxymethylphenyl)ethanol (7), and ferrocenyl(2-(2- hydroxyethyl)phenyl)methanol (9) have been prepared by reduction of corresponding ketoesters using NaBH₄ in a mixture EtOH and Et ₂O. In the course of these reactions new cyclic ethers 1-ferrocenyl-2-oxaindane (3), 3-ferrocenylisochromane (8), and 1-ferrocenylisochromane (10) have been isolated as side-products. Intramolecular cyclizations of ferrocenyl diols occur in both acidic and neutral medium. Density functional theory (BP86) calculations were used to explain the mechanism of these cyclodehydrations. Acid catalyzed reaction follows the classical S_N1 mechanism, whereas the cyclodehydration in neutral medium is described as an S_N2 reaction. X-ray diffraction analysis of new cyclic ether products has been performed. Monoacetates 11, 13 and 15 have been obtained in the reaction of ferrocenyl diols 2, 7, and 9, respectively, and acetic anhydride. Stereoselective acylation of racemic diols by vinyl acetate have been catalyzed by various lipases, and the best stereoselectivity has been observed for the diol 2 in the presence of Penicillium camembertii lipase.

Full text of DOI:10.1016/j.molstruc.2012.03.071

Journal of Molecular Structure 1019 (2012) 7–15
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Structure
journal homepage: www.elsevier.com/locate/molstruc
Experimental and DFT study of cyclodehydration and acetylation of ferrocenyl diols
a
b
c
a,
d,
⇑
⇑
´
´
ˇ
Jasmina Lapic , Aleksandar Višnjevac , Mario Cetina , Senka Djakovic , Valerije Vrcek
,
a
´
Vladimir Rapic
a Department of Chemistry and Biochemistry, Faculty of Food Technology and Biotechnology, University of Zagreb, Pierottijeva 6, HR-10000 Zagreb, Croatia
b
-
ˇ
Ruder Boškovi´c Institute, Physical Chemistry Division, Bijenicka c. 54, HR-10000 Zagreb, Croatia
c Department of Applied Chemistry, Faculty of Textile Technology, University of Zagreb, Prilaz baruna Filipovic´a 28a, HR-10000 Zagreb, Croatia
d
ˇ
Faculty of Pharmacy and Biochemistry, University of Zagreb, A. Kovaci´ca 1, HR-10000 Zagreb, Croatia
a r t i c l e i n f o
a b s t r a c t
Article history:
Racemic ferrocenyl diols, i.e. ferrocenyl(2-hydroxymethylphenyl)methanol (2), ferrocenyl-2-(2-hydrox-
ymethylphenyl)ethanol (7), and ferrocenyl(2-(2-hydroxyethyl)phenyl)methanol (9) have been prepared
by reduction of corresponding ketoesters using NaBH4 in a mixture EtOH and Et2O. In the course of these
reactions new cyclic ethers 1-ferrocenyl-2-oxaindane (3), 3-ferrocenylisochromane (8), and 1-ferroceny-
lisochromane (10) have been isolated as side-products. Intramolecular cyclizations of ferrocenyl diols
occur in both acidic and neutral medium. Density functional theory (BP86) calculations were used to
explain the mechanism of these cyclodehydrations. Acid catalyzed reaction follows the classical SN1
mechanism, whereas the cyclodehydration in neutral medium is described as an SN2 reaction. X-ray dif-
fraction analysis of new cyclic ether products has been performed. Monoacetates 11, 13 and 15 have been
obtained in the reaction of ferrocenyl diols 2, 7, and 9, respectively, and acetic anhydride. Stereoselective
acylation of racemic diols by vinyl acetate have been catalyzed by various lipases, and the best stereose-
lectivity has been observed for the diol 2 in the presence of Penicillium camembertii lipase.
Ó 2012 Elsevier B.V. All rights reserved.
Received 15 February 2012
Received in revised form 29 March 2012
Accepted 29 March 2012
Available online 5 April 2012
Keywords:
Ferrocenyl diols
Biocatalysis
Acetylation
Cyclodehydration
BP86 functional
Reaction mechanism
1. Introduction
the b-position of the side chain have received much less attention
due to the paucity of suitable homochiral starting material. (R)-
Enantiomerically pure or enriched diols are important interme-
diates and synthetic building blocks for numerous natural products,
pharmaceuticals and fine chemicals. Various racemic alkane-1,2-
diols bearing a tertiary benzylic alcohol stereogenic center were
regioselectively acetylated with vinyl acetate in the presence of li-
pases and esterases in organic solvents. The primary hydroxyl func-
tion reacts faster in the lipase-catalyzed acylation of 1,2-diols in
organic media, while in the alcoholysis of the corresponding dies-
ters a primary function is hydrolyzed in a highly regioselective
manner [1]. Lipase-catalyzed acetylation seems to be an ideal
choice for obtaining homochiral products from racemic 2-acylated
1,2-diol derivatives having bulkiness in b-position to the acetoxy
moiety [2]. Optically active ferrocenes have been widely used in
asymmetric catalysis as chiral ligands [3]. Several ferrocenyl al-
kenes were converted using Sharpless’ standard conditions to ferr-
ocenyl diols in good yields and in enantiomeric excesses up to >99%
[4]. (R)-2-Ferrocenyletane-1,2-diol have been prepared in both
enantiomeric forms by asymmetric dihydroxylation of vinyl
ferrocene [4]. Ferrocenyl derivatives which have a stereocentre in
(ꢀ)-1-Acetoxy-3-ferrocenylpropane-2-ol and (S)-(+)-1,2-diacet-
oxy-3-ferrocenylpropane were prepared in enantioselective
acylation of ( )-3-ferrocenylpropane-1,2-diol mediated by lipase
from Pseudomonas cepacia in high chemical yield and enantiomeric
excess [3,5]. The enantioselective synthesis of b-ferrocenyl-b-ami-
no alcohols are included the catalytic asymmetric dihydroxylation
of 1-ferrocenyl alkenes and the regio- and stereoselective azide
substitution of the hydroxyl group adjacent to the ferrocene moi-
ety. These compounds are a new class of central chiral ferrocene
derivatives suitable for the elaboration of auxiliaries and ligands
for asymmetric synthesis [6].
In this paper we present a preparation and characterization of
racemic ferrocenyl diols 2, 7 and 9 as substrates and corresponding
acetates 11–16 as standards for study of stereoselectivity in biocat-
alyzed acylation (Chart 1). Each of these diols contains one second-
ary and one primary carbinol group separated by benzene ring. We
specifically pay attention to the mechanism underlying the forma-
tion of diols which is accompanied by the follow-up reaction, i.e.
subsequent cyclodehydration of these diols. The latter process re-
sults in formation of cyclic ethers 3, 8 and 10, not reported earlier
(except the ether 10). Spectral analyses (X-ray, NMR, and IR) and
quantum-chemical calculation (DFT method) were used to investi-
gate these structures in more details.
⇑
Corresponding authors. Tel.: +385
1 4605086; fax: +385 1 4836 083 (S.
´
ˇ
Djakovic), tel.: +385 1 6394441; fax: +385 1 4856201 (V. Vrcek).
ˇ
E-mail addresses: sdjakov@pbf.hr (S. Djakovic´), valerije@pharma.hr (V. Vrcek).
0022-2860/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molstruc.2012.03.071

´
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
8
Chart 1. Ferrocenyl diols prepared in this work.
2. Results and discussion
2.1. Synthesis and spectral characterization
The compounds 2, 7 and 9 were prepared as substrates and cor-
responding monoacetates as standards in our on-going research on
lipase-catalyzed resolution of centrochiral racemic ferrocene aryl-
diols [7].
The starting compounds, o-ferrocenoylbenzoic acid, o-(ferro-
cenoylmethyl)benzoic acid and o-(ferrocenoylphenyl)acetic acid,
are described in the literature as products of Friedel–Crafts reac-
tion of ferrocene with phthalic or homophtalic anhydride [8]. In
our experiments, we prepared these aromatic acids in 53% or 34%
yield, and converted them to the corresponding ketoesters 1, 4
.
and 5. By the action of BF3 Et2O in methanol on the o-ferrocenoyl-
benzoic acid, the methyl o-ferrocenoylbenzoate (1) was prepared
in 86% yield, while methyl o-(ferrocenoylmethyl)benzoate (4)
and methyl o-(ferrocenoylphenyl)acetate (5) were obtained in
reaction with CH2N2 in methanol. Reduction of ferrocene ketoest-
ers 1, 4 and 5 by the molar excess of NaBH4 in a mixture EtOH
and Et2O gave about 80% of the corresponding diols 2, 7 and 9
accompanied with 17% of 1-ferrocenyl-2-oxaindane (3), 20% of 3-
ferrocenylisochromane (8) and 16% of 1-ferrocenylisochromane
(10) (Scheme 1 and 2).
Scheme 2. Preparation of ferrocenyl diols 2 and 9 and their corresponding cyclic
ethers 8 and 10, respectively.
In order to prepare mono- and diacetates, the reaction of diols
(2, 7 and 9) with acetic anhydride has been performed at room
temperature. Only the corresponding monoacetates 11, 13 and
15 have been obtained, while diacetates 12, 14 and 16 were not
isolated. This suggests that the acetoxyl group in the
a-position
to the ferrocenyl moiety is unstable in the acid conditions and
readily hydrolyzed to the hydroxyl group. Under the same condi-
tions, along with monoacetates (Scheme 3), the starting diols and
cyclic ethers (3, 8 and 10) were isolated.
Scheme 3. Preparation of mono- and diacetates from ferrocenyl diols.
signals at 3443 and 3334 cmꢀ1. These red-shifted absorptions indi-
cate that each of the OH groups participate in hydrogen bonds. The
assignment of the IR bands of 7 to the corresponding intramolecu-
lar interactions is possible if calculated geometry parameters are
taken into account. Thus, in the most stable conformer of 7, the
0
One- (1H, APT) and two-dimensional NMR measurements
(COSY, NOESY), FTIR spectra, and density functional theory
(BP86) calculations were combined in order to explore structural
properties of diols 2, 7, and 9.
It is known that the absorption band of alcohol is assigned as a
very sharp signal at 3600 cmꢀ1 which is a result of unhindered
stretching of hydroxyl group, whereas the absorption at 3550–
3200 cmꢀ1 corresponds to species with OH group involved in
hydrogen bonding. Accordingly, the absorption at 3377 and
3435 cmꢀ1 observed in 2 and 9, respectively, (CH2Cl2 solution, ca
10ꢀ2 M) can be attributed to associated hydroxyl group. In con-
trast, the IR spectra of a ca 10ꢀ2 M solution of FcCH(OH)C6H5 [9]
in dichloromethane showed the narrow strong absorptions about
3600 cmꢀ1 which was indication of free OH group. FTIR spectrum
of 7 in dichloromethane solution (c ꢁ 10ꢀ2 M) shows two distinct
OH group is orientated toward the cyclopentadienyl ring, while
a
the OH group is suitably positioned for interaction with iron. As
a
shown in our earlier report [10] electrostatic interactions between
0
OH group and
p
-electrons, and OH group and metal center give
a
a
rise to lower (3334 cmꢀ1 in our case) and higher (3443 cmꢀ1) fre-
quency absorption peak, respectively.
In 1H NMR spectra of 2 in CDCl3 solution, the signal for hydroxyl
proton was found at 2.39 ppm for FcCH(OH )C6H5 [9] and at
a
0
2.35 ppm for C6H5CH2OH , while the signals of the diol 2 have been
a
0
shifted to lower field (H 4.20 and H 5.15 ppm), which is, again, an
a
a
Scheme 1. Preparation of ferrocenyl diol 2 and the corresponding cyclic ether 3.

´
9
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
Table 1
Lipase-catalyzed asymmetric transformations of these racemic
diols to the chiral monoacetates were performed by the action of
vinyl acetate as acetyl donor in the presence of various lipases at
27 °C [7]. The preliminary results indicated that diol 2 showed
the best stereoselectivity by vinyl acetate in the presence of Peni-
cillium camembertii lipase as catalyst.
1H NMR data of 2, 7 and 9 and standard (reference) compounds (c = 10ꢀ2 m).
Compound
dOH/ppm (in CDCl3)
dOH/ppm (in DMSO)
a
a0
a
a0
C6H5CH2OH
FcCH(OH)C6H5
2
7
9
–
2.35
–
5.15
4.59
3.92
–
–
5.31
4.88
4.15
5.2
–
5.16
5.09
4.21
2.39a
4.42
4.49
4.19
2.2. Molecular structures of 3 and 8
a
See Ref. [9].
In a wish to investigate molecular structures of cyclic ethers, we
prepared single crystals of cyclic ethers 3 and 8 and their struc-
tures were determined by X-ray crystal structure analysis
(Fig. 1). Structure of cyclic ether 10 has been published previously
[11].
indication of hydrogen bond interactions (Table 1). In the strongly
coordinating solvent [D6]-DMSO, the signals for OH and OH at
5.30 and 5.16 ppm are found, respectively. A significant shift to
0
a
a
The conformation of the cyclopentadienyl (Cp) rings is eclipsed
in both structures. The values of the corresponding C–Cg–Cg–C
pseudo-torsion angles, defined by joining two eclipsing Cp carbon
atoms through the ring centroids (Cg) range from ꢀ0.3(3) to 0.2(3)°
in 3 and 7.6(4) to 8.4(5)° in 8. Tilt angle between Cp rings in 3 is
slightly larger [2.5(2)°] than in 8 [0.9(4)°]. These two structures dif-
fer in the orientation of benzene ring with respect to the cyclopen-
tadienyl (Cp) rings. While C14AC19 ring in 3 forms an angle of
82.46(17)° with respect to the C1AC5 ring, being virtually perpen-
dicular to it, C14AC15/C17AC20 ring in 8 is almost parallel to
C1AC5 ring [17.0(3)°]. Intriguingly, benzene ring in previously re-
ported 1-ferrocenylisochromane [11], which also contains six-
membered heterocyclic ring as 8 has similar orientation as its
counterpart in the structure of 3 [corresponding dihedral angle
equals 88.07(9)°]. The heterocyclic five-membered ring in 3 adopts
an envelope conformation, while the heterocyclic six-membered
ring in 8 reveals a half-chair conformation.
the lower fields for OH can be an indication of its intermolecular
hydrogen bond interactions or interaction with the solvent DMSO.
a
0
The chemical shift of OH is almost the same as in CDCl3 solution,
a
0
suggesting the involvement of OH group in the formation of
a
intramolecular hydrogen bond only. The NOESY spectra obtained
0
0
from compound 2 reveal OH interactions with both CH and CH .
a
a
a
In contrast, the hydroxyl group OH interacts with CH only. These
a
a
0
interactions suggests a proximity between OH and CH in diol 2.
a
a
An additional support for this comes from optimized geometry of
2 (see Supplementary material) in which the calculated distance
0
between the two hydrogens OH and CH is 2.8 Å, whereas the
a
a
0
corresponding distance between OH and CH is much longer
a
a
(4.7 Å).
In 1H NMR spectra of 7 in CDCl3 solution, the signal for hydroxyl
0
proton located at 4.49 ppm (OH ) and 4.59 ppm (OH ) while in [D6]-
a
a
DMSO solvent are found at 4.8 and 5.09 ppm. These shifts to a lower
field (in comparison with the reference compounds) and chemical
shift values in the [D6]-DMSO solution for both hydroxyl groups
could be indication for associated hydroxyl functions. The NOESY
spectra of diol 7 showed more complex interactions in comparison
with the spectra of diol 2. It is in agreement with calculated structure
of 7 (see Supplementary material) in which all respective distances
Both, 3 and 8, contain two CAHꢂ ꢂ ꢂ
p interactions while structure
of 3 contains one additional CAHꢂ ꢂ ꢂO weak hydrogen-bond. The
latter CAHꢂ ꢂ ꢂO hydrogen-bond in
3
[C8-Hꢂ ꢂ ꢂO12, 3.466(4) Å;
162(3)°] self-assembles the molecules into chains which are fur-
ther extended into two-dimensional network (see Supplementary
data Fig. S1) by two CAHꢂ ꢂ ꢂ
p interactions. In one of them, hydro-
0
0
0
favor mutual interactions between CH , CH , OH , and OH .
a
a
a
a
gen atom of Cp ring C1AC5 points to the benzene ring of neighbor-
ing molecule [Hꢂ ꢂ ꢂCg(C14AC19), 2.78(4) Å; C4AHꢂ ꢂ ꢂCg, 161(3)°],
while in the other one benzene ring hydrogen atom is directed to-
wards Cp ring C1AC5 [Hꢂ ꢂ ꢂCg(C1AC5), 2.96 Å; C17AHꢂ ꢂ ꢂCg, 147°].
IR spectrum of 9 in dichloromethane solution (c ꢁ 10ꢀ2 M) con-
tains the signal at 3435 cmꢀ1 which indicates again that hydroxyl
groups participate in hydrogen bond interactions. As well, in the 1H
NMR spectrum (c ꢁ 10ꢀ2 M) of 9 in CDCl3 solution the OH signals
In 8, two-dimensional network is formed by CAHꢂ ꢂ ꢂ
p interac-
have been shifted to lower field (4.19 (a
) and 3.92 (a0) ppm) in
tions in which benzene ring carbon atoms act as hydrogen-bond
donors (see Supplementary data Fig. S2). In one case, hydrogen
atom points to the Cp C1AC5 ring [Hꢂ ꢂ ꢂCg(C1AC5), 2.95 Å;
C17AH17ꢂ ꢂ ꢂCg, 170°], while in the second one it points to the
comparison with the chemical shifts for reference compounds (Ta-
ble 1). In the 1H NMR spectra in [D6]-DMSO, the signal for OH is at
a
0
4.15 ppm and for OH is at 4.21 ppm.
a
Fig. 1. ORTEP drawings of the structure 3 (left) and 8 (right) with atom numbering. Displacement ellipsoids for nonhydrogen atoms are scaled at the 30% probability level.

´
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
10
neighboring
benzene
ring
[Hꢂ ꢂ ꢂCg(C14AC20),
2.86 Å;
of adsorption of diol onto Brønsted acid centers, protonation by
C20AH20ꢂ ꢂ ꢂCg, 164°]. Thus, supramolecular structure of 8 is exclu-
one hydroxyl group, removal of water with the formation of carbo-
cation, and cyclization [14]. In this work, the similar mechanism
for cyclodehydration of ferocenyl-derived diols 2 and 7 is proposed
and investigated in details by means of DFT calculations.
sively determined by CAHꢂ ꢂ ꢂ
p
interactions, while in 3, this type of
interaction influences significantly the formation of a two-dimen-
sional supramolecular network.
Acid catalyzed intramolecular cyclization of 2 is initiated
(Scheme 4) by protonation of
a-hydroxyl group resulting in pro-
2.3. Cyclodehydration of ferrocenyl diols
tonated intermediate 2a [15]. The protonated hydroxyl group (i.e.
water molecule) departs from the ferrocenyl moiety and forms a
carbocationic intermediate, which is presented here as an ion-neu-
tral complex 2b. The carbocation 2b is located at a free energy min-
imum deeper than the initial structure 2a, indicating the stability
of this ferrocenyl-stabilized carbocation. The alternative proton-
ation of a0-hydroxyl group results in benzyl-stabilized carbocation
2b0 which is 126 kJ/mol less stable than ferrocenyl-stabilized car-
bocation 2b (Table 2).
Cyclic ethers 3, 8, and 10 were formed as side-products in the
course of our work-up procedure where the traces of acid are inev-
itable, but also in a dry organic solvent (e.g. benzene or CH2Cl2)
where proton donors are not expected to exist. In order to explain
the mechanism by which cyclic ethers are formed in two different
media, quantum-chemical calculations were performed. The cyclic
ether 3 results from intramolecular five-membered ring closure,
whereas ethers 8 and 10 are products of six-membered ring forma-
tion processes.
It is well-established that the reaction of 1,4- and 1,5-diols, un-
der various catalytic conditions, leads to the corresponding five-
and six-membered cyclic ethers, respectively [12,13]. Herewith
we present that cyclodehydration of ferrocenyl diol 2 occurs in
both acidic and neutral medium. The similar mechanism is ex-
pected to be operative in cases of cyclodehydration of 7 and 9 in
which cyclic ethers 8 and 10, respectively, are formed. In organic
solvent (benzene in our case) the formation of cyclic ether is a slow
process, due to relatively large barrier for dehydration (see below),
whereas in the presence of acid traces, the protonation could lead
to enhanced dehydration. It has been shown that different hetero-
geneous surfaces that include acidic sites (silica gel in our case) are
suitable and sufficient to promote cyclodehydration of 1,4-diols
[14].
Table 2
Gibbs free energies G298 (in Hartree; at 298.15 K)a and solvation energies (in kJ/mol)b
calculated for species involved in cyclodehydrations of diol 2 and protonated diol 2a.
Structure
G298
D
Gsolv
DG298 + DGsolv
2
ꢀ2111.153870
ꢀ2111.154146
ꢀ2111.149859
ꢀ2111.081933
ꢀ2111.164271
ꢀ2111.524990
ꢀ2111.503146
ꢀ2111.533887
ꢀ2111.492395
ꢀ2111.534900
ꢀ2111.535231
ꢀ47.8
0.0
+1.2
+10.4
20
ꢀ45.9
200
TSSN2
3b
2a
2a0
2b
2b0
TSSN1
3a
ꢀ47.9
ꢀ50.8
+185.9
ꢀ19.1
+28.9
+62.1
0.0
+126.0
+18.1
+11.1
ꢀ39.6
ꢀ288.9
ꢀ313.0
ꢀ294.4
ꢀ277.3
ꢀ273.6
ꢀ279.8
a
Calculated at the BP86/AE1(+)//BP86/AE1 level.
Calculated with PCM/UFF/BP86/AE1 method in the model solvent of
Different reaction mechanisms for cyclodehydration of 2 under
non-acid and (slightly) acid condition are considered. According to
earlier experimental results, cyclic ether is formed by the sequence
b
e = 8.93
(dichloromethane).
Scheme 4. Proposed reaction mechanisms of cyclodehydration of diol 2 in acid and neutral medium calculated at the BP86/AE1 level of theory.

´
11
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
Scheme 5. The most stable conformers of ferrocenyl diol 2 and its protonated form 2a calculated at the BP86/AE1 level of theory.
Only two conformers 2a and 2a0 of protonated diol were located
(Scheme 5), with the former being more stable for 33.2 kJ/mol. It is
found that benzyl moiety must be placed in exo position in order to
successfully locate the stable minimum structure. During confor-
mation search several starting geometries of protonated diols con-
verged to the more stable carbocation, as water is spontaneously
eliminated during geometry optimizations. The particular confor-
mations of 2a and 2a0 are well suited for stabilization by OHꢂ ꢂ ꢂFe
intramolecular interactions. This interaction between metal center
with typical hydrogen-bond donors is usually weak, but can some-
times be observed [16] and has recently been counted among the
‘‘diverse world of unconventional hydrogen bonds’’ [17]. The struc-
ture 2a is also characterized by intramolecular hydrogen bonding
between the two oxygen atoms which results in additional stabil-
ization of this conformer.
ion 2b. It is characterized by one imaginary frequency which corre-
sponds to the intramolecular ring closure. The relatively low
barrier calculated for this process explains the ease of formation
of side-product 3 during our work-up procedure.
The protonated cyclic ether 3a is easily converted to neutral
form 3b which is ca. 20 kJ/mol more stable than the starting neu-
tral diol 2. This suggests that acid-catalyzed cyclodehydration of
2 is both kinetically and thermodynamically feasible process.
Contrary to the case of 2a, cyclodehydration of 2 in dry benzene
(acid-free solvent) should be much slower process. The same cyclic
ether product 3 has been isolated, but the reaction 2 ? 3b in or-
ganic solvent, without acid traces, is kinetically less favorable.
The reaction starts from neutral ferrocene-containing diol 2 which
exists in several conformations spanning a range of ca. 10 kJ/mol
(Table 2). The most stable conformer 2 contains the benzyl moiety
The most stable form of water-complexed carbocation 2b is lo-
cated by placing water molecule in a variety of locations to sample
the different arrays of interaction networks available between car-
bocation and water. Initial configurations were created using a lo-
cally modified version of the stochastic search method [18]. Several
water configurations for water-complexed cation were located, but
only the most stable structure was considered in this study.
After the carbocation 2b is formed, with formally empty p orbi-
in egzo position, while a-hydroxyl group interacts with iron center
(Scheme 5). This type of interaction has been investigated in details
earlier (see above). In contrast, the least stable conformer 200, in
which OHꢂ ꢂ ꢂFe interaction is absent, is characterized by OHꢂ ꢂ ꢂ
p
intramolecular interactions. It has been shown that the two types
of intramolecular interactions, OHꢂ ꢂ ꢂFe and OHꢂ ꢂ ꢂCp, are in compe-
tition with each other in ferrocene-containing alcohols [10]. Both
experimental and theoretical studies suggest that interaction of
the hydroxyl group with the d-orbitals of iron has more stabilizing
tal at the a-carbon center, the nucleophilic substitution takes place
in which a0-hydroxyl group attacks at carbocationic center. This
intramolecular SN1 reaction results in a protonated cyclic ether
3a. The transition state structure TSSN1 (Fig. 2) for this process
has been located and is only 18.1 kJ/mol less stable than carbocat-
effect than the interaction with ferrocenyl p-electrons [10,19].
Diol 2 undergoes SN2 like cyclodehydration mechanism. The
transition state structure TSSN2 is located and is calculated
185.9 kJ/mol less stable than the diol 2. Therefore, the energy bar-
Fig. 2. BP86/AE1 optimized geometries of transition states for SN1 and SN2 type of cyclodehydration reactions in protonated ferrocenyl diol 2a and neutral ferrocenyl diol 2,
respectively. BP86/AE1(+) values are in italics. Selected bond distances are in Å.

´
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
12
rier for cyclodehydration of neutral diol is one order of magnitude
higher that the corresponding barrier for protonated form. The
structure TSSN2 is characterized by one imaginary frequency which
correspond to five-membered ring closure process simultaneous
6500 Hz in time domain f1. Data were collected into a
2048 ꢃ 256 acquisition matrix and processed using a 2 K ꢃ 1 K
transformed matrix with zero filling in f1 domain for both experi-
ments. Sine multiplication was performed before Fourier transfor-
mations. In HMBC spectra the delay for long-range couplings was
set to 60 ms. Digital resolutions were 3.18 and 81.38 Hz per point
in f2 and f1, respectively in HMQC spectra and 7.33 and 122.82 Hz
per point in f2 and f1, respectively in HMBC spectra.
with intramolecular hydrogen transfer from a0- to
a-hydroxyl
group (see Fig. 2). The latter process presents intramolecular acid
catalysis which promotes the elimination of
(i.e. water molecule).
a-hydroxyl group
The final product is the cyclic ether 3b depicted here as a water-
complexed (monosolvated) minimum. While kinetically unviable,
this product is thermodynamically favored (Table 2).
4.2. Compounds preparation, FTIR and NMR spectroscopic analyses
4.2.1. Methyl o-ferrocenoylbenzoate (1)
A solution of 1 g (3.3 mmol) of o-ferrocenoylbenzoic acid in
13 mL of methanol containing 2.2 mL of BF3 ꢃ Et2O was refluxed
for 2 h, cooled, and poured into saturated aqueous NaHCO3 solu-
tion. The organic layer was washed with water, dried over MgSO4,
evaporated and the crude product purified by preparative thin
layer chromatography on silica gel with CH2Cl2/EtOAc (9:1) gave
890 mg (86%) of dark red oil.
3. Conclusions
Reduction of ferrocene ketoesters 1, 4 and 5 by the molar excess
of NaBH4 in a mixture EtOH and Et2O gave about 80% of the corre-
sponding diols 2, 7 and 9. Under the same conditions, the intramo-
lecular cyclization of diols with formation of cyclic ethers 3, 8 and
10 occurred. By virtue of density functional theory (BP86) calcula-
tions the mechanism of these cyclodehydrations has been ex-
plained in details. Acid catalyzed reaction follows the classical
SN1 mechanism, whereas the cyclodehydration in neutral medium
is described as an SN2 reaction. X-ray diffraction analysis of new
cyclic ether products 3 and 8 has been performed and compared
to an earlier reported structure of cyclic ether 10. By the action
of acetic anhydride on diols 2, 7 and 9 in hexane solution, the cor-
responding monoacetates 11, 13 and 15 were obtained, while
diacetates 12, 14 and 16 were not isolated. Stereoselective acyla-
tion of racemic diols by vinyl acetate have been catalyzed by vari-
ous lipases, and the best stereoselectivity has been observed for the
diol 2 in the presence of P. camembertii lipase, which could be ex-
plained by the proximity of the primary carbinol group and stere-
ogenic center.
IR (CH2Cl2):
m = 3103 (w, CAH, Fc), 3065 (w, CAH, Ph), 1731 (s,
COester), 1661 (s, COketone) cmꢀ1
.
1H NMR (CDCl3): d = 7.94 (1H, s,
0
0
0
0
H6 ), 7.65 (3H, m, H3 ,4 ,5 ), 4.60 (2H, m, H2,5), 4.52 (2H, m, H3,4),
4.21 (5H, s, CH-Fc, Cpunstubst), 3.63 (3H, s, COOCH3) ppm. 13C
NMR, APT (CDCl3): d = 200.85 (COketone), 169.94 (COester), 142.17
0
0
0
0
0
(C1 ), 132.05 (C3 ), 129.92 (C4 ), 129.60 (C5 ), 129.28 (C2 ), 127.68
0
(C6 ), 79.90 (C1, Fc), 72.45 (C2,5), 70.65 (C3,4), 69.87 (CpunstubstAFc),
52.20 (CH3) ppm.
4.2.2. Methyl o-(ferrocenoylmethyl)benzoate (4) and methyl o-
(ferrocenoylphenyl)acetate (5)
The mixture of o-(ferrocenoylmethyl)benzoic acid and o-(ferr-
ocenoylphenyl)acetic acide 80 mg (0.23 mmol) was dissolved in
methanol (14 mL), and the excess of ethereal solution of diazo-
methane was added. After standing in refrigerator for 24 h, solvent
was evaporated and the residue purifed by TLC with CH2Cl2/EtOAc
(9:1) as eluent gave 14 mg of orange crystals 4, 41 mg of red crys-
tals 5 and 23 mg of orange-red crystals lactone 6.
4. Experimental
4.1. General methods
4: IR (CH2Cl2):
m = 3105 (w, CAH, Fc), 3052 (w, CAH, Ph), 1721
(s, COester), 1681 (s, COketone) cmꢀ1
.
1H NMR (CDCl3): d = 8.01 (1H,
0 0
0
0
Melting points were determined with a Buechi apparatus. The
IR spectra were recorded for KBr pellets or CH2Cl2 solutions with
a Bomem MB100 Mid FT IR spectrophotometer. The 1H and 13C
NMR spectra of CDCl3 solutions were recorded on a Varian EM
360 or Varian Gemini 300 spectrometer with tetramethylsilane
as internal standard. Products were purified by preparative thin
d, H3 ), 7.48 (1H, t, H4 ), 7.35 (1H, t, H5 ), 7.27 (1H, d, H6 ), 4.86
(2H, s, H2,5), 4.55 (2H, m, H3,4), 4.50 (2H, s, CH2) 4.26 (5H, s, CH-
Fc, Cpunsubst) 3.84 (3H, s, COOCH3) ppm. 13C NMR, APT (CDCl3):
0
0
d = 200.75 (COketone), 157.59 (COester), 136.08 (C1 ), 131.60 (C3 ),
0
0
0
0
131.49 (C4 ), 130.20 (C5 ), 129.82 (C2 ), 126.64 (C6 ), 79.90 (C1),
71.54 (C2,5), 70.65 (CpunsubstAFc), 69.00 (C3,4), 51.44 (CH3), 44.57
(CH2) ppm.
layer chromatography on silica gel (Merck, Kieselgel 60 HF254
)
and by recrystallization from benzene. All the compounds were
prepared under argon atmosphere in dry solvents.
5: IR (CH2Cl2):
m
= 3105 (w, CAH, Fc), 3065 (w, CAH, Ph), 1743
(s, COester), 1647 (s, COketone) cmꢀ1
.
1H NMR (CDCl3): d = 7.78 (1H,
0
0
0
0
1D and 2D (1H, APT, gCOSY, gHSQC and gHMBC) NMR spectra
compounds 2, 7 and 9 were recorded at ambient temperature on
the Avance DRX500 spectrometer working at 500.13 MHz for 1H
and equipped with 5 mm diameter inverse detection probe with
z-gradient. Sample concentration in DMSO-d6 was 20 mg mlꢀ1
with TMS as internal standard. 1D 1H and APT NMR spectra were
acquired with 10,000 Hz and 18,000 Hz spectral windows using
64 K data points. Digital resolution was 0.15 Hz and 0.27 Hz per
point. 2D gCOSY spectra were acquired with sweep width of
6700 Hz in both dimensions into 2 K data points with 256 incre-
ments. Spectra were zero-filled in f1 dimension to 1 K and pro-
cessed using unshifted sine bell window function. Digital
resolution was 3.25 Hz per point in both dimensions. Number of
scans was 2. The gHSQC and gHMBC spectra were recorded with
a relaxation delay of 1.5 s and 8 scans per increment. For HMQC
experiment the spectral width was 20,800 Hz in acquisition do-
main f2 and 6500 Hz in time domain f1. For HMBC experiment
the spectral width was 31,000 Hz in acquisition domain f2 and
d, H3 , J = 7.5 Hz), 7.45 (1H, t, H4 , J = 7.7 Hz), 7.39 (2H, d, H5 ,6
,
J = 7.5 Hz), 4.81 (2H, s, H2,5), 4.59 (2H, s, H3,4), 4.28 (5H, s, CH-Fc,
Cpunsubst) 3.85 (2H, s, CH2) 3.67 (3H, s, COOCH3) ppm. 13C NMR,
0
APT (CDCl3): d = 201.31 (COketone), 171.84 (COester), 139.56 (C1 ),
0
0
0
0
0
132.61 (C2 ), 131.67 (C4 ), 130.38 (C5 ), 128.82 (C3 ), 126.49 (C6 ),
79.23 (C1), 72.54 (C2,5), 71.65 (C3,4), 70.05 (CpunsubstAFc), 51.98
(CH3), 38.50 (CH2) ppm.
6: IR (CH2Cl2):
m = 3105 (w, CAH, Fc), 3059 (w, CAH, Ph), 1741
0
(s, CO) cmꢀ1
.
1H NMR (CDCl3): d = 8.29 (1H, d, H3 , J = 7.8 Hz),
0
0
7.69 (1H, t, H4 , J = 7.5 Hz), 7.45 (1H, t, H5 , J = 7.5 Hz), 7.42 (1H, d,
0
H6 , J = 7.5 Hz), 6.55 (1H, s, CH), 4.82 (2H, s, H2,5), 4.44 (2H, m,
H3,4), 4.21 (5H, s, CH-Fc, Cpunsubst) ppm.
4.2.3. General procedure for preparation of diols 2, 7 and 9
To a solution of 0.9 mmol ketoesters (1, 4 and 5) in 5 mL mix-
ture of EtOH: Et2O (1: 1), NaBH4 (253 mg, 6.7 mmol) was added
in portions. The mixture was refluxed for 2 h and worked up in
the usual manner. The preparated diols were a yellow oils which

´
13
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
0
crystallized after standing in refrigerator. A small amount of cyclic
ethers (3, 8 and 10) were obtained together with the diols.
2: yellow crystals, (238 mg, 82%), Mp. 88–89 °C. Anal. Calcd for
CAOAC) cmꢀ1
.
1H NMR (CDCl3): d = 8.14 (1H, s, H3 ), 7.57 (1H, s,
0
0
0
H6 ), 7.42 (2H, m, H4 ,5 ), 5.38 (1H, s, CH), 4.38 (2H, m, H2,5), 4.13
(5H, s, CH-Fc, Cpunsubst) 4.03 (2H, m, H3,4), 3.32 (2H, m, CH2O),
0
C
18H18O2Fe (322.19): C, 67.10; H, 5.63. Found: C, 67.35; H, 5.78. IR
(CH2Cl2): = 3377 (b, OH), 3097 (w, CAH, Fc), 3054 (w, CAH, Ph)
cmꢀ1 1H NMR (CDCl3): d = 7.37 (1H, s, H6 ), 7.32 (1H, s, H4 ), 7.31
(1H, s, H ), 7.29 (1H, s, H ), 5.76 (1H, s, CH ), 5.15 (1H, m, OH ),
2.78 (2H, m, CH2) ppm. 13C NMR, APT (DMSO): d = 135.26 (C2 ),
0
0
0
0
0
m
133.98 (C1 ), 129.24 (C6 ), 126.78 (C3 ), 126.31 (C5 ), 124.62 (C4 ),
88.77 (C1), 72.91 (CH), 69.00 (CpunsubstAFc), 68.22 (C5), 68.18 (C2),
67.86 (C3), 67.76 (CH2O), 66.55 (C4), 33.69 (CH2) ppm.
0
0
.
30
50
0
a
a
0
0
4.80 (1H, d, CH2, , J = 12.04 Hz), 4.63 (1H, d, CH2, , J = 12.04 Hz),
10: yellow crystals, (45 mg, 16%), Mp. 92–93 °C. Anal. Calcd for
C19H17OFe (317.19): C, 71.95; H, 5.40. Found: C, 72.21; H, 5.58. IR
a
a
4.42 (1H, s, OH ), 4.31 (2H, s, H2,5), 4.19 (5H, s, CH-Fc, Cpunsubst),
a
4.13 (1H, s, H3), 4.06 (1H, s, H4). 1H NMR (DMSO): d = 7.36 (1H, s,
(CH2Cl2):
m = 3079 (w, CAH, Fc), 3017 (w, CAH, Ph), 1283 (s,
0
0
0
0
0
H6 ), 7.32 (1H, s, H4 ), 7.31 (1H, s, H3 ), 7.30 (1H, s, H5 ), 5.71 (1H,
CAOAC) cmꢀ1
.
1H NMR (CDCl3): d = 7.25 (1H, m, H6 ), 7.14 (2H,
0
0
0
s, CH , J = 5.1 Hz), 5.31 (1H, d, OH , J = 5.1 Hz), 5.16 (1H, t, OH ,
m, H4 ,5 ), 7.10 (1H, s, H3 ), 5.61 (1H, s, CH), 4.30 (1H, s, H2), 4.22
(5H, s, CH-Fc, Cpunsubst) 4.16 (1H, s, H5), 4.11 (1H, s, H3), 4.06 (1H,
m, CH2) 4.03 (1H, s, H4), 3.82 (1H, m, CH2), 2.83 (2H, m, CH2)
0
a
a
a
,
J = 5.0 Hz), 4.33 (1H, pt, H2,
0
J = 5.1 Hz), 4.69 (2H, d, CH2,
a
J = 6.3 Hz), 4.26 (1H, pt, H5), 4.20 (5H, s, CH-Fc, Cpunsubst), 4.15
0
0
(1H, pt, H3), 4.07 (1H, pt, H4) ppm. 13C NMR, APT (DMSO):
ppm. 13C NMR, APT (CDCl3): d = 137.18 (C2 ), 133.13 (C1 ), 128.62
0
0
0
0
0
0
0
0
d = 141.69 (C1 ), 138.95 (C2 ), 127.72 (C5 ), 127.37 (C4 ), 127.17
(C6 ), 126.35 (C3 ), 126.06 (C5 ), 125.37 (C4 ), 90.05 (C1), 74.57
(CH), 68.292 (C5), 68.79 (CpunsubstAFc), 68.28(C2), 66.83 (C3),
66.77 (C4), 62.07 (CH2), 28.49 (CH2) ppm.
0
0
(C6 ), 126.95 (C3 ), 89.58 (C1), 68.63 (CpunsubstAFc), 68.67 (C ),
a
2
5
3
4
0
68.07 (C ), 67.58 (C ), 66.33 (C ), 66.15 (C ), 60.6 (C ) ppm.
a
7: yellow crystals (242 mg, 80%), Mp. 94–95 °C. Anal. Calcd for
C
19H20O2Fe (336.22): C, 67.88; H, 5.99. Found: C, 67.63; H, 5.71.
4.2.4. General procedure for preparation of acetates 11–16
A solution of 0.192 mmol of diols (2, 7 and 9) in 4 mL of hexane
containing a 0.480 mmol of pyridine and 0.192 mmol acetic anhy-
dride was stirred at room temperature overnight, and thereafter
dissolved in diethyl ether. The organic layer was washed with
aqueous hydrochloric acid (5%), a saturated aqueous solution of
NaHCO3, and with water. The organic extract was dried, evapo-
rated and the residue purifed by TLC on silica gel with CH2Cl2 as
eluent.
IR (CH2Cl2):
m = 3443, 3334 (m, OH) 3097 (w, CAH, Fc), 3055 (w,
0
CAH, Ph) cmꢀ1
.
1H NMR (CDCl3): d = 7.37 (1H, s, H6 ), 7.32 (1H, d,
0
0
0
H3 , J = 1.8 Hz), 7.27 (1H, d, H5 , J = 1.8 Hz), 7.20 (1H, d, H4 ,
0
J = 1.8 Hz), 4.76 (1H, d, CH, J = 11.79 Hz), 4.59 (1H, t, OH , J = 3.58
a
HZ), 4.49 (1H, s, OH ), 4.27 (2H, s, CH2,b), 4.24 (2H, s, H5,2), 4.22
a
(5H, s, CH-Fc, Cpunsubst), 4.18 (2H, s, H3,4), 3.04 (2H, d, CH2, ,
0
a
0
J = 7.94 Hz) ppm. 1H NMR (DMSO): d = 7.35 (1H, s, H6 ), 7.20 (3H,
0
0
0
s, H3 ,4 ,5 ), 5.09 (1H, t, OH , J = 5.33 Hz), 4.88 (1H, d, OH ,
0
a
a
0
J = 5.98 Hz), 4.59 (1H, d, CH , J = 4.19 Hz), 4.51 (2H, m, CH2, ),
11: orange oil, (38.46 mg, 55%). IR (CH2Cl2):
m
= 3375 (b, OH)
a
a
4.24 (1H, s, H2), 4.18 (5H, s, CH-Fc, Cpunsubst), 4.18 (2H, s, H5,3),
3097 (w, CAH, Fc), 3059 (w, CAH, Ph), 1731 (s, C@O) cmꢀ1
.
1H
0
0
4.08 (2H, s, H4), 3.07 (1H, dd, CH2,b, J = 13.46 Hz), 2.87 (1H, dd,
NMR (CDCl3): d = 7.50 (1H, d, H6 , J = 7.8 Hz), 7.35 (1H, s, H3 ),
0
0
0
CH2,b, J = 13.74 Hz) ppm. 13C NMR, APT (DMSO): d = 140.61 (C2 ),
7.30 (1H, s, H5 ), 7.25 (1H, s, H4 ), 5.69 (1H, d, CH , J = 3.3 Hz),
a
0
0
0
0
0
137.83 (C1 ), 130.62 (C6 ), 127.88 (C3 ), 126.96 (C4 ), 126.07 (C5 ),
5.21 (1H, s, CH2,b), 4.26 (5H, s, CH-Fc, Cpunsubst), 4.22 (1H, s, H2),
4.20 (1H, s, H5), 4.17 (1H, s, H3), 4.15 (1H, s, H4), 2.69 (1H, d,
94.05 (C1), 69.37 (CH ), 68.26 (CpunsubstAFc), 67.26 (C2), 66.86
a
(C5), 65.14 (C3), 63.08 (C4), 61.35 (CH2,b), 40.56 (CH2, ) ppm.
OH , J = 3.l3 Hz), 2.05 (3H, s, CH3) ppm. 13C NMR, APT (CDCl3):
0
a
a
0
0
0
9: yellow crystals, (257 mg, 85%), Mp. 95–96 °C. Anal. Calcd for
d = 170.70 (C@O), 141.59 (C2 ), 132.73 (C1 ), 129.27 (C6 ), 128.42
0
0
0
C
19H20O2Fe (336.22): C, 67.88; H, 5.99. Found: C, 68.01; H, 6.12. IR
(CH2Cl2): = 3435(b, OH) 3103 (w, CAH, Fc), 3065 (w, CAH, Ph)
cmꢀ1
(C3 ), 127.52 (C4 ), 126.63 (C5 ), 94.06 (C1), 68.38 (CpunsubstAFc),
m
68.11 (CH ), 67.94 (C2), 67.74 (C2), 67.38 (C3), 66.53 (C4), 63.59
a
0
.
1H NMR (CDCl3): d = 7.30 (1H, d, C6 , J = 7.5 Hz), 7.25 (1H,
(CH2,b), 20.94 (CH3) ppm.
0
0
0
d, C4 , J = 6 Hz), 7.20 (1H, t, C3 , J = 6 Hz), 7.15 (1H, d, C5 ,
13: orange oil, (46.5 mg, 64%). IR (CH2Cl2):
m
= 3443 (b, OH)
J = 7.5 Hz), 5.75 (1H, s, CH ), 4.30 (2H, s, H2,5), 4.26 (5H, s, CH-Fc,
3098 (w, CAH, Fc), 3056 (w, CAH, Ph), 1721 (s, C@O) cmꢀ1
.
1H
a
0
0
Cpunsubst), 4.19 (1H, s, OH ), 4.12 (1H, s, H3), 3.97 (1H, s, H4), 3.92
NMR (CDCl3): d = 7.34 (1H, d, H6 , J = 7.8 Hz), 7.27 (1H, s, H3 ),
a
0
0
(1H, s, OH ), 3.85 (2H, m, CH2, ), 2.96 (1H, m, CH2,b), 2.85 (1H,
7.25 (1H, s, H5 ), 7.20 (1H, s, H5 ), 5.07 (1H, s, CH2, ), 4.57 (1H, t,
0
0
a
a
a
0
m, CH2,b) ppm. 1H NMR (DMSO): d = 7.18 (1H, d, C6 , J = 7.8 Hz),
CH , J = 6.3 Hz), 4.25 (2H, s, H2,5), 4.19 (5H, s, CH-Fc, Cpunsubst),
a
0
0
7.16 (1H, d, C3 , J = 7.8 Hz), 7.13 (1H, d, C4 , J = 7.8 Hz), 7.12 (1H, s,
4.17 (1H, d, OH , J = 12 Hz), 4.04 (2H, s, H3,4), 3.01 (2H, s, CH2,b),
a
0
C5 ), 5.59 (1H, s, CH ), 4.23 (5H, s, CH-Fc, Cpunsubst), 4.21 (1H, pt,
2.08 (3H, s, CH3) ppm. 13C NMR, APT (CDCl3): d = 170.70 (C@O),
a
0
0
0
0
0
OH ), 4.19 (2H, pt, H2,5), 4.15 (1H, s, OH ), 4.03 (2H, pt, H3,4),
137.33 (C2 ), 134.19 (C1 ), 130.69 (C6 ), 129.64 (C3 ), 128.28 (C4 ),
0
a
a
0
3.96 (1H, m, CH2, ), 3.79 (1H, m, CH2, ), 2.96 (2H, m, CH2,b
)
126.57 (C5 ), 93.06 (C1), 70.66 (CH ), 68.17 (CpunsubstAFc), 67.89
0
0
a
a
a
0
0
ppm. 13C NMR, APT (DMSO): d = 137.65 (C1 ), 133.25 (C2 ), 128.85
(C2), 67.79 (C5), 67.13 (C3), 64.86 (C4), 64.27 (CH2,b), 41.07 (CH2, ),
0
a
0
0
0
0
(C5 ), 126.59 (C4 ), 126.32 (C6 ), 125.70 (C3 ), 90.48 (C1), 73.87
20.94 (CH3) ppm.
(CH ), 68.86 (CpunsubstAFc), 68.66 (C2), 68.08 (C5), 66.83 (C3),
15: orange oil (42.12 mg, 58%). IR (CH2Cl2):
m
= 3430 (b, OH)
.
3105 (w, CAH, Fc), 3063 (w, CAH, Ph), 1744 (s, C@O) cmꢀ1 1H
0 0 0 0
a
66.54 (C4), 61.77 (CH2, ), 28.15 (CH2,b) ppm.
0
a
3: yellow crystals, (44 mg, 17%), Mp. 99–101 °C. Anal. Calcd for
18H16OFe (304.17): C, 71.08; H, 5.30. Found: C, 71.25; H, 5.42. IR
NMR (CDCl3): d = 7.44 (1H, s, H6 ), 7.21 (1H, s, H3 ,5 ,4 ), 5.69 (1H,
C
m, CH ), 5.21 (1H, s, CH2,b), 4.25 (2H, s, H2,5), 4.19 (5H, s, CH-Fc,
a
Cpunsubst), 4.04 (2H, s, H3,4), 3.03 (2H, m, CH2,b ), 2.59 (1H, s, OH ),
0
(CH2Cl2):
m
= 3099 (w, CAH, Fc), 3028 (w, CAH, Ph), 1306 (s,
a
0
CAOAC) cmꢀ1
.
1H NMR (CDCl3): d = 7.27 (1H, m, H6 ), 7.21 (1H,
2.08 (3H, s, CH3) ppm. 13C NMR, APT (CDCl3): d = 170.70 (C@O),
0
0
0
0
0
0
0
0
m, H4 ), 7.17 (1H, s, H3 ), 7.16 (1H, s, H5 ), 6.01 (1H, s, CH), 5.08
(1H, m, CH2), 4.21 (1H, s, H2), 4.16 (1H, s, H5), 4.12 (5H, s, CH-Fc,
Cpunsubst) 4.06 (1H, s, H3) 3.99 (1H, s, H4) ppm. 13C NMR, APT
137.33 (C2 ), 134.19 (C1 ), 130.69 (C6 ), 129.64 (C3 ), 128.28 (C4 ),
0
126.57 (C5 ), 93.06 (C1), 70.66 (CH ), 68.17 (CpunsubstAFc), 67.89
a
(C2), 67.79 (C5), 67.13 (C3), 64.86 (C4), 64.27 (CH2,b), 41.07 (CH2, ),
0
a
0
0
0
0
(CDCl3): d = 141.14 (C1 ), 139.35 (C2 ), 127.43 (C4 ), 126.97 (C5 ),
20.94 (CH3) ppm.
0
0
121.95 (C3 ), 120.83 (C6 ), 89.15 (C1), 81.49 (CH), 72.37 (CH2O),
68.54 (CpunsubstAFc), 68.42 (C5), 67.46 (C2), 66.26 (C3), 66.09 (C4)
ppm.
4.3. X-ray crystallography
8: yellow crystals, (57 mg, 20%), Mp. 112–113 °C. Anal. Calcd for
19H17OFe (317.19): C, 71.95; H, 5.40. Found: C, 72.15; H, 5.65. IR
Crystals suitable for X-ray single crystal structure study were
grown by slow evaporation from CH2Cl2 solution. The intensities
were collected at the room temperature, on an Enraf Nonius
C
(CH2Cl2): = 3081 (w, CAH, Fc), 3020 (w, CAH, Ph), 1278 (s,
m

´
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
14
CAD4 diffractometer using graphite-monochromated CuK
tion (k = 1.54179 Å). The XCAD4 software was applied for data
reduction [20]. Absorption correction based on the set of eight suit-
a
radia-
Single point energy calculations were performed at the BP86/
AE1 level employing the polarizable continuum model (PCM) of
Tomasi and co-workers [33] (using the parameters of dichloro-
methane and UFF radii from the united atom topological model).
able
w-scan reflexions was performed [21]. Both structures were
solved using direct methods with SIR2004 [22] and refined using
the full matrix least-squares refinement based on F2, with SHELX97
[23]. All non-hydrogen atoms were refined anisotropically. The
hydrogen atoms attached to carbon atoms of cyclopentadienyl
rings, as well as carbon atoms C11 and C13 in 8 have been found
in Fourier map and refined freely. All other hydrogen atoms in 8
and all hydrogen atoms in 3 were treated using appropriate riding
models, with SHELXL97 defaults. PLATON [24] was used for struc-
ture analysis and molecular visualization. CCDC 228979 and CCDC
228980 contain supplementary crystallographic data for 3 and 8.
These data can be obtained free of charge via http://www.ccdc.ca-
m.ac.uk/conts/retrieving.html, or from the Cambridge Crystallo-
graphic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK;
fax: +44 1223 336 033; or e-mail: deposit@ccdc.cam.ac.uk.
Acknowledgments
This research was supported by the Ministry of Science, Educa-
tion and Sports of the Republic of Croatia (Grant Nos. 119-
1193079-3069).
Appendix A. Supplementary material
BP86/AE1 optimized coordinates of selected structures and
additional crystallographic data are available. Supplementary data
associated with this article can be found, in the online version, at
http://dx.doi.org/10.1016/j.molstruc.2012.03.071.
Crystal data for 3: C19H18FeO, Mr = 318.18, orthorhombic, Pbca
(No. 61); a = 17.588(2), b = 7.4000(10), c = 22.725(3) Å; V =
References
2957.7(7) Å3; Z = 8; dx = 1.429 g cmꢀ3
;
l(CuK
a ;
) = 8.121 mmꢀ1
[1] (a) A.J.M. Janssen, A.J.H. Klunder, B. Zwanenburg, Tetrahedron 47 (1991) 7409;
(b) A. Bosetti, D. Bianchi, P. Cesti, P. Golini, S. Spezia, J. Chem. Soc. Perkin Trans.
1 (1992) 239;
S = 1.032; R = 0.0765 and wR = 0.2183 for 190 parameters and
2219 independent reflections with I P 2r(I) measured in the range
3.89° – h – 76.52°.
(c) K. Lemke, F. Theil, A. Kunath, H. Schick, Tetrahedron: Asymmetry 7 (1996)
971;
(d) F. Theil, J. Weidner, J. Ballschuh, A. Kunath, H. Schick, J. Org. Chem. 59
(1994) 388;
Crystal data for 8: C18H16FeO, Mr = 304.16, monoclinic, P21/c (No.
14); a = 5.8583(2), b = 28.1470(10), c = 8.2480(3) Å, b = 93.120(3)°;
V = 1358.03(8) Å3; Z = 4; dx = 1.488 g cmꢀ3
;
l(CuK
a) = 8.817
(e) A. Liljeblad, E. Vänttinen, L. Kanerva, Chirality 11 (1999) 432;
(f) R.P. Hof, R.M. Kellogg, Tetrahedron: Asymmetry 5 (1994) 565.
[2] G. Egri, E. Baitz-Gács, L. Poppe, Tetrahedron: Asymmetry 7 (1996) 1437.
[3] (a) A. Togni, T. Hayashi, Ferrocenes, VCH Verlagsgesellschaft mbH, Weinheim,
Germany, 1995;
mmꢀ1; S = 1.219; R = 0.0410 and wR = 0.1005 for 229 parameters
and 2341 independent reflections with I P 2r(I) measured in the
range 3.14° – h – 76.34°.
(b) A. Patti, G. Nicolosi, Tetrahedron: Asymmetry 11 (2000) 815–822;
(c) C.J. Richards, A. Locke, Tetrahedron: Asymmetry 9 (1998) 2377.
[4] (a) W.G. Jary, J. Baumgartner, Tetrahedron: Asymmetry 9 (1998) 2081;
(b) H. Becker, M.A. Soler, K.B. Sharpless, Tetrahedron 51 (1995) 1345.
[5] A. Patti, G. Nicolosi, Tetrahedron: Asymmetry 10 (1999) 2651.
[6] M. Catasús, A. Bueno, A. Moyano, M.A. Maestro, J. Mahía, J. Organomet. Chem.
642 (2002) 212.
[7] S. Djakovic´, PhD thesis, Faculty of Science, University of Zagreb, Croatia, 2000.
[8] K.L. Rinehart Jr., R.J. Curby Jr., P.E. Sokol, J. Am. Chem. Soc. 79 (1957) 3420.
[9] (a) S.I. Goldberg, D.W. Mayo, J.A. Alford, J. Org. Chem. 28 (1963) 1708;
(b) C. Glidewell, R.B. Klar, P. Lightfoot, C.M. Zakaria, G. Ferguson, Acta Cryst.
B52 (1996) 110.
4.4. Computational details
Geometries were fully optimized at the RI-BP86 level as imple-
mented in the G03 program [25], employing the exchange and cor-
relation functionals of Becke [26] and Perdew [27], respectively,
together with a fine integration grid (75 radial shells with 302
angular points per shell), the augmented Wachters’ basis [28] on
Fe (8s7p4d), 6-31G⁄ basis [29] on all other elements, denoted as
AE1. This and comparable DFT levels have proven quite successful
for transition-metal compounds and are well suited for the
description of structures, energies, vibrational frequencies, and
other properties [30]. Harmonic frequencies were computed from
analytical second derivatives.
Test calculations for the ferrocene–water complex indicated
that interaction energies suffer from substantial basis-set superpo-
sition error (BSSE) at that level. This error is significantly reduced
when the 6-31+G(d,p) basis, denoted AE1(+), is employed (e.g.
for ferrocene + water the raw interaction energy with this aug-
mented basis is ꢀ1.9 kcal/mol, which is reduced to ꢀ1.2 kcal/mol
after correction for BSSE with the Counterpoise method [31]).
These interaction energies, as well as the relative energies of se-
lected structures involved in cyclodehydration reactions, are virtu-
ally identical when the molecules are fully optimized with this
augmented AE1(+) basis, or when just single-point energies are
calculated for the BP86/AE1 geometries. We thus adopted the lat-
ter protocol and report AE1(+) single point energies throughout
(i.e. at the BP86/AE1(+)//BP86/AE1 level in Table 2).
ˇ
[10] V. Vrcek, M. Bühl, Organometallics 25 (2006) 358.
ˇ
[11] M. Cetina, S. Djakovic´, V. Rapic´, A. Golobic, Acta Cryst. C59 (2003) 328.
[12] G. Bouchoux, N. Choret, R. Flammang, J. Phys. Chem A 101 (1997) 4271.
[13] M. Fleisher, V. Stonkus, L. Leite, E. Lukevics, Int. J. Quant. Chem. 88 (2002) 670.
[14] (a) Y.V. Subba Rao, S.J. Kulkarni, M. Subrahmanyam, A.V. Rama Rao, J. Org.
Chem. 59 (1994) 3998;
(b) I. Bucsi, A. Molnar, M. Bartok, G.A. Olah, Tetrahedron 51 (1995) 3319.
[15] Protonation of a0-hydroxyl group results in protonated diol intermediate
which is less stable than the protonated diol 2a.
[16] E.S. Shubina, N.V. Belkova, L.M. Epstein, J. Organomet. Chem. 537 (1997) 17.
[17] N.V. Belkova, E.S. Shubina, L.M. Epstein, Acc. Chem. Res. 38 (2005) 624.
[18] (a) M.J. Saunders, Comput. Chem. 25 (2004) 621;
ˇ
(b) V. Vrcek, O. Kronja, M.J. Saunders, Chem. Theory Comput. 3 (2007) 1223.
[19] (a) A.W. Baker, D.E. Bublitz, Spectrochim. Acta 22 (1966) 1787;
(b) D.S. Trifan, R. Bacskai, J. Am. Chem. Soc. 82 (1960) 5010.
[20] K. Harms, S. Wocadlo, XCAD-4: Program for Processing CAD4 Diffractometer
Data, University of Marburg, Germany, 1995.
[21] A.C.T. North, D.C. Phillips, F.S. Mathews, Acta Crystallogr. A24 (1968) 351.
[22] M.C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G.L. Cascarano, L. De Caro, C.
Giacovazzo, G. Polidori, R. Spagna, J. Appl. Cryst. 38 (2005) 381.
[23] G.M. Sheldrick, SHELX97: Program for the Refinement of Crystal Structures,
Universität Göttingen, Germany, 1997.
[24] A.L. Spek, J. Appl. Crystallogr. 36 (2003) 7.
[25] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A.
Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox,
H.P. Hratchian, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E.
Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y.
Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S.
Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K.
Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J.
Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L.
Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M.
IRC calculations (intrinsic reaction coordinate as implemented
in Gaussian 03) were performed at the corresponding level of the-
ory to identify the minima connected through the transition state.
The initial geometries used were that of the corresponding transi-
tion states structures, and the paths were followed in both direc-
tions from that point. This method verified that
transition structure indeed connected the presumed energy mini-
mum structures [32].
a given

´
15
J. Lapic et al. / Journal of Molecular Structure 1019 (2012) 7–15
ˇ
´
Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A.
Pople, Gaussian 03, revision D.01, Gaussian, Inc., Wallingford, CT, 2004.
[26] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
(b) J. Lapic´, M. Cetina, D. Šakic´, S. Djakovic´, V. Vrcek, V. Rapic, Arkivoc IX (2010)
257;
ˇ
(c) A. Šaric´, V. Vrcek, M. Bühl, Organometallics 27 (2008) 394.
[27] (a) J.P. Perdew, Phys. Rev. B 33 (1986) 8822;
[31] S.F. Boys, F. Bernardi, Mol. Phys. 10 (1970) 553.
[32] C. Gonzales, H.B. Schlegel, J. Phys. Chem. 94 (1990) 5523.
[33] As implemented in G03: (a) V. Barone, M. Cossi, J. Tomasi, J. Comput. Chem. 19
(1998) 404;
(b) J.P. Perdew, Phys. Rev. B 34 (1986) 7406.
[28] (a) A.J.H. Wachters, J. Chem. Phys. 52 (1970) 1033;
(b) P.J. Hay, J. Chem. Phys. 66 (1977) 4377.
[29] (a) W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys. 56 (1972) 2257;
(b) P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213.
[30] (a) W. Koch, M.C. Holthausen, A Chemist’s Guide to Density Functional Theory,
Wiley-VCH, Weinheim, 2000;
(b) M. Cossi, G. Scalmani, N. Rega, V. Barone, J. Chem. Phys. 117 (2002)
43;
(c) M. Cossi, O. Crescenzi, J. Chem. Phys. 19 (2003) 8863.