Cyclization of Terphenyl-Bisfluorenols: A Mechanistic Study of the Regioselectvity

Article abstract of DOI:10.1002/chem.201901457

The mechanism of the bicyclization reaction of a series of terphenyl-bisfluorenols into dispiro[fluorene-9,6′-indeno[1,2-b]fluorene-12′,9′′-fluorene] and dispiro[fluorene-9,6′-indeno[2,1-a]fluorene-12′,9′′-fluorene] is reported. Through a combined experim

Full text of DOI:10.1002/chem.201901457

A Journal of
Accepted Article
Title: CYCLIZATION OF TERPHENYL-BISFLUORENOL: A
MECHANISTIC STUDY OF THE REGIOSELECTVITY
Authors: Cyril Poriel, Frederic Barriere, Joelle Rault-Berthelot, and
Damien Thirion
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201901457
Link to VoR: http://dx.doi.org/10.1002/chem.201901457
Supported by

10.1002/chem.201901457
Chemistry - A European Journal
CYCLIZATION OF TERPHENYL-BISFLUORENOL: A MECHANISTIC
STUDY OF THE REGIOSELECTVITY
Cyril Poriel,* Frédéric Barrière,* Joëlle Rault-Berthelot, Damien Thirion
Univ Rennes, CNRS, ISCR - UMR 6226, F-35000 Rennes, France
Introduction
For the last thirty years of research in material science, spirobifluorene-based compounds have
been extensively studied for applications in organic electronics and especially in organic light-
emitting diodes (OLED).[1-3] The incorporation of spiro linkages in small molecular units is
nowadays a powerful molecular technique to design efficient and stable organic semi-
conductors: fluorophores for OLEDs,[3-8] high-triplet host materials for phosphorescent
OLEDs,[2, 9-12] thermally activated delayed fluorescence hosts,[13-15] and electron-donor[16-18] or
non-fullerene acceptors[19-22] for solar cells. The spirobifluorene fragment combines the
physical advantages of an orthogonal spiro configuration (high thermal and morphological
stability) and the appealing properties of the fluorene unit (e.g. high quantum yield, facile and
versatile chemical functionalization).[23] However, organic electronics is not the only
application which has benefited from spirobifluorene-based compounds. One can cite for
example heterogeneous and homogeneous catalysis (e.g. epoxidation, cyclopropanation…),[24-
29] coordination polymers,[30] or surfaces modification.[31]
Scheme 1: Examples of intramolecular cyclization of fluoren-9-ol derivatives.
From a synthetic point of view, generating spiro-configured compounds[32] and especially
differently substituted spirobifluorenes is usually performed through intramolecular
electrophilic cyclization of a fluoren-9-ol in the presence of a Lewis or a Brønsted acid (Scheme
1).[3, 33-36] If the pendant phenyl ring of the biphenyl core is substituted at its para position the
intramolecular cyclization step leads to a single product, i.e. 2-substituted-spirobifluorene being
hence regioselective (Scheme 1-top). However, if the pendant phenyl unit is substituted at its
meta position, the electrophilic attack may occur on two different positions leading to the
formation of two regioisomers (1- and 3-substituted SBF) being hence non-regioselective. This
feature is due to a rotation around the biphenyl bond during the cyclization process (Scheme 1-
bottom) and has been previously observed notably by Lee and coworkers.[37] Although this
synthetic approach is very interesting to synthesize differently substituted spirobifluorenes, it
remains to date almost absent from the literature. This absence is due to the difficulty to direct
the regioselectivity of the reaction towards one single isomer. For the last ten years, our group
has worked on different ways to favour such intramolecular cyclization reactions used to
generate molecules with a spiro carbon linkage. These works have led to new generations of
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
fused spirobifluorenes namely dispiro[fluorene-9,6'-indeno[1,2-b]fluorene-12',9''-fluorene] or
DSF-IF type 2 isomers and dispiro[fluorene-9,6'-indeno[2,1-a]fluorene-12',9''-fluorene] or
DSF-IF type 3 isomers (Scheme 2), which have shown promising performance when
incorporated as host for phosphors in phosphorescent OLEDs or as fluorophore in OLEDs[4, 36,
38] and even recently as hole transporting materials in perovskite solar cells.[18] These molecules
possess either the dihydroindeno[1,2-b]fluorenyl core (type 2 isomers) or the
dihydroindeno[2,1-a]fluorenyl core (type 3 isomers), widely introduced in many organic semi-
conductors.[36]
The synthetic approach towards regioisomers 2 and 3 involves, in the last step, the
intramolecular bicyclization of a bisfluorenol substituted terphenyl 1, Scheme 2, and leads to
the formation of the two previously mentioned positional isomers. As these positional isomers
can be separated by column chromatography, this versatile reaction has been used to produce a
palette of fluorophores.[39, 40] In this work, as a development to our preliminary note,[41] we wish
to report on the mechanism of this bicyclization reaction and particularly on the origin of the
surprising regioselectivity observed. Through a combined experimental and theoretical study,
we report herein on the rational ways of favouring the formation of one of the two possible
positional isomers. As dihydroindenofluorene positional isomers are nowadays widely used
building units in organic electronics,[1, 42-44] (and their antiaromatic counterparts as well[45]) this
work allows to better understand one of their synthetic access described in the last years.
Results and Discussion
Scheme 2: Intramolecular electrophilic cyclization of difluorenols 1a-g: Synthesis of DSF-IF
2a-g and 3a-g.
2
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
The synthesis of 2a-g and 3a-g molecules involves the treatment of the difluorenols 1a-g
.
precursors [39] by a Lewis acid (BF3 OEt2), Scheme 2, leading to a quantitative intramolecular
1
bicyclization. The crude mixture was further analysed by H NMR spectroscopy in order to
determine the ratio of regioisomers 2 and 3 formed. In the present work, have also been
investigated the most used cyclization conditions originally used in literature to generate spiro
carbon linkages (i.e. AcOH/HCl).[46]
1
Table 1. Ratio of 2/3 isomers (%) determined by H NMR spectroscopy (all these reactions
were run 1h in the presence of BF3 etherate except in the case of AcOH-HClaq)
AcOH-
DCM
rt
DCM
reflux
CH3CN
rt
CH3CN
reflux
DMSO
150°C
CF3CH2OH
reflux
Entry
R
HClaq
reflux
96/4
80/20
43/57
45/55
25/75
24/76
0/100
H
Et
t-Bu
99/<1
93/7
98/2
91/9
96/4
91/9
94/6
78/22
65/35
35/65
-
91/9[a]
87/13
69/31
37/63
-
2a/3a
2b/3b
2c/3c
2d/3d
2e/3e
2f/3f
79/21
33/67
30/70
32/68
27/73
0/100
70/30
34/66
27/73
26/74
16/84
0/100
74/26
60/40
44/56
38/62
0/100
77/23
56/44
39/61
36/64
0/100
4-FPh
4-t-BuPh
4-nonylPh
3,5-di-t-BuPh
27/73
0/100
27/73
0/100
2g/3g
[a] starting materials diols was not soluble in the solvent.
1. Case of the least encumbering and non-substituted R = H, reaction of 1a leading to 2a/3a.
We have first investigated the cyclization reaction of the non-substituted difluorenol 1a
leading to the formation of the two DSF-IFs 2a/3a (Table 1). The first conditions investigated
are the most common ones used in the literature to generate spiro carbon linkages, i.e. acetic
acid/HCl-reflux.[3] In these conditions, the cyclization of 1a provides a mixture of the two
regioisomers 2a/3a, the 2a isomer being however strongly favoured (2a/3a: 96/4). The
modification of the reaction conditions strongly alters the ratio of isomers formed. The
cyclization of 1a performed in an apolar[47] solvent such as dichloromethane at room
temperature is almost fully regioselective providing almost exclusively the isomer 2a with only
formation of less than 1% of the other isomer 3a.* However, switching to more polar and donor
solvent,[47] such as acetonitrile, the amount of 3a over 2a at room temperature consistently
increases remaining nevertheless very low (4%). Interestingly, the temperature of the reaction
allows to increase the amount of 3a and seems to have a significant effect on the ratio of isomers
formed. In acetonitrile, the 2a/3a distribution is 96/4 at room temperature and is 91/9 at reflux.
Thus, switching to a higher temperature and/or to a more polar and donor solvent (CH3CN,
DMSO and CF3CH2OH, Table 1) the amount of 3a over 2a consistently increases, reaching a
maximum of 9% at the reflux of CH3CN or CF3CH2OH. Despite moderate in the present case,
these effects will be stronger in the examples discussed next where the fluorene moieties bear
different substituents.
2. Case of aliphatic substitution R = Et, reaction of 1b leading to 2b/3b.
The incorporation of an ethyl group on the 2,7-positions of the fluorene moieties (ethyl-
difluorenol 1b, Scheme 2) leads to a significant alteration of the ratio of isomers formed (2b/3b)
compared to 2a/3a. Indeed, in dichloromethane and acetonitrile at room temperature, the ratio
of isomer 3b respectively reaches 7% and 21%, a significant increase compared to the ratio of
* It is important to mention that for all the solvents investigated in this study, no cyclization reaction was detected prior to the
addition of acid.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
its unsubstituted analogue 3a in the same conditions (less than 1% and 4% respectively, Table
1). More impressively, the amount of 3b over 2b, in polar and donor solvents (reflux), CH3CN
(2b/3b: 70/30), DMSO (2b/3b: 78/22) and CF3CH2OH (2b/3b: 87/13), is significantly
enhanced compared to the amount of 3a over 2a, reaching a maximum of 30% at reflux of
acetonitrile, Table 1. At this stage, one can hence conclude that increasing the steric hindrance
on the fluorene units of the starting difluorenol (through the incorporation of ethyl groups) has
a significant effect on the regioselectivity of the reaction. This feature is nevertheless surprising
at first sight as it clearly favours the more encumbered 3-type isomer possessing a cofacial
arrangement of the fluorene moieties that is the less energetically stable isomer (vide infra). As
stated above in the case of 2a/3a, the temperature is also of great importance on the 2b/3b
isomers distribution. Note that the cyclization cannot be run at room temperature in DMSO and
CF3CH2OH because of the poor solubility of the reactants.
Thus, the increase of (i) the temperature of the reaction, (ii) the steric hindrance (substitution
of fluorene units with ethyl groups) and (iii) the polarity and donor ability of the solvent all
increase the amount of type-3 isomers, albeit the most sterically hindered and less energetically
isomers (Table 2).
3. Case of a more encumbering aliphatic substituent R = t-Bu, reaction of 1c leading to 2c/3c.
Although the solvents and temperature effects seem to follow a similar trend for the cyclization
of 1a and 1b, the preferentially formed isomer is always the type-2 isomer (2a or 2b). Since the
steric hindrance has a crucial impact on the regioselectivity of the reaction, bulky t-butyl groups
were introduced on the 2,7-positions of the fluorene units to confirm this hypothesis. Thus, at
room temperature, the 2c/3c ratio is 74/26 in dichloromethane and is almost fully reversed in
acetonitrile, 33/67. Compared to the series a and b presented above, we thus noted that the ratio
of 3-type isomers is impressively enhanced in all the solvents: from less than 1% for 3a to 7%
for 3b and up to 26% for 3c in dichloromethane (at room temperature) and from 4% for 3a to
21% for 3b and up to 67% for 3c in acetonitrile (at room temperature). The distribution of 2c/3c
ratio was hence strongly modified compared series a and b (vide supra), highlighting the
importance of the steric hindrance in this cyclization reaction. Thus, the ratio of 3-type isomers
seems directly related to the bulkiness of the substituents borne by the starting difluorenols (t-
butyl for 1c > ethyl for 1b > hydrogen for 1a). Interestingly, when the bulkiness of the R group
is sufficiently large such as in 2c/3c, the observed tuning range through solvent effects is also
the largest. This is a key point. One can hence conclude from this first series of cyclization
reactions of alkyl-substituted difluorenols 1a-c that (i) low polarity and weakly donor solvent,
(ii) low temperature and (iii) small R groups on the fluorene moieties all favour the synthesis
of 2-type positional isomers. The opposite conditions lead gradually and preferably to the 3-
type isomers although they are more encumbered and less energetically stable than the 2-type
isomers (Table 2). The three effects involved in the ratio of isomers formed need to be addressed
herein: the effect of substitution at the fluorene moieties, the solvent effect and the temperature
effect. To assist this discussion, molecular modelling of the type-2 and type-3 isomers, the pro-
2 or pro-3 conformers and their corresponding transition state towards the Wheland
intermediates have been performed.
4. Molecular modelling of pro-2a/3a and pro-2c/3c transition states
For such reaction considered as an addition/elimination process, it is widely admitted that the
electrophile interacts with the -system passing through a sigma bonded intermediate (called
sigma complex, arenium ion or Wheland intermediate) and leading to the final product.[48] As
4
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
stated by Sainsbury in the book 'Aromatic Chemistry', "the true nature of the transition state is
not easily predicted, but it is reasonable to assume that the transition state resembles the sigma
complex".[48] However, in the present case, the Wheland intermediate cannot be involved in the
regioselectivity of the reaction as the double cyclization has already been done and the syn or
anti configuration is already fixed. In order to explain the formation of the two regiosomers,
we reasonably assume that the bicyclization reaction unfolds sequentially (Scheme 3). The first
cyclization leads to a carbocation with a spirobifluorene core, substituted at its C2 carbon by a
pendant phenyl unit, itself ortho-substituted by a fluorene group linked through its C9 carbon.
At this stage, we consider the formation of interconvertible carbocationic pro-2 or pro-3
conformers which afford during the second cyclization steps the corresponding Wheland-type
intermediate[15] that finally leads to the regioisomers 2/3 (Scheme 3) after deprotonation.
Scheme 3: Cyclization of difluorenols 1a-g and the corresponding intermediates postulated in
the mechanism of formation of regioisomers 2a-g/3a-g.
It is hence relevant to examine theoretically the structure and energy of these intermediates and
transition states in the pro-2 and pro-3 conformation (Scheme 3) as a function of the nature of
the substitution at the fluorene units. However, it is first important to state that irrespective of
the substitution of the fluorene units, type-2 isomers are always thermodynamically more stable
than their type-3 analogues (Table 2). Indeed, in type-3 isomers, the steric hindrance induced
by the two cofacial fluorene units leads to less stable molecules compared to the type-2 isomers
with the two fluorene units on two opposite sides of the dihydroindenofluorene core. Thus, the
difference in stability is 0.36 eV (8.30 kcal/mol) and 0.34 eV (7.84 kcal/mol) for 2a/3a (R=H)
and 2b/3b (R=Et) respectively. Branching a bulkier group on the fluorene (2c/3c R=t-Bu)
increases the relative stability of the type-2 isomer to 0.51 eV (11.76 kcal/mol). As we have
shown experimentally that the more the difluorenol is encumbered the more the reaction
becomes selective towards type-3 isomers, the difference observed in the relative stability of
the two isomers indicates that this is certainly not a key parameter in the selectivity observed in
the bicyclization reaction. Rather the encumbrance at the level of the transition state before
cyclization to the Wheland intermediate is likely to be more relevant, as discussed next.
5
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
Table 2: Calculated relative stability of 2a-g vs. 3a-g.
Entry
2a/3a 2b/3b
2c/3c
2d/3d 2e/3e 2f/3f* 2g/3g
0.36
0.34
0.51
0.34
0.38
0.37
(8.53)
0.48
(11.07)
E (eV)
(kcal/mol)
(8.30) (7.84) (11.76) (7.84) (8.76)
*Calculations have been performed with an ethyl chain instead of a nonyl chain.
For R = H the energy of the pro-2a or pro-3a carbocation conformers only differs by 0.04 eV
(0.92 kcal/mol) and that of the corresponding Wheland intermediates lies higher in energy (by
0.17 (3.92 kcal/mol) and 0.55 eV (12.68 kcal/mol) respectively).28 A similar trend is found for
the R=Et and R=t-Bu derivatives. Examination of the transition states between the carbocations
and the Wheland intermediates for R = H, Et or t-Bu shows that they lie at about 0.55 eV (12.68
kcal/mol) higher than the corresponding pro-2 carbocation and ca. 0.80 eV (18.45 kcal/mol)
higher than the corresponding pro-3 carbocation. In other words, although the 2/3 ratio changes
as the substituent encumbrance increases the calculated energy of the transition states remains
very similar from 2a to 2c and decreases only slightly from 3a to 3c. Consequently, there is at
first sight only a subtle energy trend that could explain the differences in the increasing ratio
toward the type 3 isomer as the bulkiness of the R group increases. It must be concluded that,
although all intermediates are energetically accessible, increased bulkiness of the R group
gradually prevents the facile structural access to the pro-2 transition state. This can be illustrated
in Figure 1 where the two calculated pro-2a and pro-3a transition states (R=H) are pictured.
Although free of any steric hindrance, the optimized structure of the pro-2a transition state
(Figure 1, left)* shows that any substitution at the 2,7-positions on each fluorene group (see H2
and H7 coloured in blue in Figure 1) will inevitably lead to an encumbrance between the
substituents. We believe that this structural feature is at the origin of the selectivity observed
herein. As the size of R increases, the actual accessibility to the pro-2 transition state should
then be more difficult compared to that of the pro-3 transition state. This becomes more evident
when comparing the optimized geometry of the pro-2c and pro-3c transition states with the
bulkier t-Bu substituent (Figure 2). We reckon that the key explanation to the selectivity of this
sequential and irreversible reaction as a function of the size of R, lies in the respective structure
and encumbrance of the two pro-2 and pro-3 transition states.
pro-2a and pro-3a transition states
Figure 1: Views of transition states of pro-2a (left) and pro-3a (right). Substitution at the
position of hydrogen atoms coloured in blue will inevitably induce steric hindrance in the pro-
2 transition states (left) but not in the pro-3 transition states (right).
* All the log files with the (attempted) optimized geometry are available as supporting information for viewing
and manipulating the structures presented in Figures 1-4 (see experimental part)
6
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
pro-2c and pro-3c transition states
Figure 2: Views of transition states of pro-2c (left) and pro-3c (right) with hydrogen atoms
omitted for clarity.
5. Case of encumbering para-substituted phenyl substituents R=4-fluoro-phenyl; 4-t-butyl-
phenyl or 4-nonyl-phenyl, reaction of 1d-f leading to 2d-f/3d-f.
In order to confirm that the regioselectivity of this reaction finds its origin in the encumbrance
of the fluorene units of the two pro-2 and pro-3 transition states, three other systems have been
investigated. In this second step, para-substituted phenyl rings were introduced on the fluorene
units to extend the study of the influence of such bulky substituents on the cyclization reaction.
In the light of the structure of the pro-2a and pro-2c transition states presented above, the goal
of increasing the length of the substituents borne by the fluorene was to disfavour access to the
pro-2 transition states. We investigated difluorenol derivatives substituted with 4-fluoro-phenyl
(1d); 4-t-butyl-phenyl (1e) and 4-nonyl-phenyl (1f). The three different substituents borne by
the aryl ring (-F, -t-Bu, and -C9H19) possess different size and these substituents should increase
the amount of type 3 isomers formed at the expense of type 2 isomers.
Compared to the difluorenols 1a-c bearing relatively small substituents directly connected to
the fluorene units, the cyclization of aryl-substituted difluorenols 1d-f led in all case to an
impressive increase of the amount of the isomer 3, Table 1. However, when branching the
fluorene units with a para-substituted phenyl moiety (2d-f/3d-f) the relative stability of the
type-2 isomers is similar to that of the unsubstituted (2a/3a) or the ethyl substituted fluorene
(2b/3b) isomers (0.34 to 0.38 eV range, 7.84 to 8.76 kcal/mol), Table 2. This trend shows that
the effect of the encumbrance of a para-substituted phenyl group is low and does not
significantly affect the magnitude of the relative stability of the type-2 isomers compared to
unsubstituted or alkyl-substituted fluorene groups. Again, this parameter cannot be taken into
account to explain the regioselectivity.
The cyclization of the 4-fluorophenyl-difluorenol 1d, bearing the smallest substituent in para
position of the phenyl ring, i.e. a fluorine atom, leads to a 2d/3d distribution of 60/40 at room
temperature in dichloromethane. By increasing the temperature, the 2d/3d distribution is 56/44,
highlighting as noted above that the amount of 3d is slightly increased by thermal input. Thus,
in these conditions, the ratio of the type 3-isomer is significantly increased compared to alkyl
substituted fluorenols (from 23% for 3c to 44% for 3d) but the general trend is the same. In
polar solvents, the amount of 3d is strongly increased compared to non-polar solvents, reaching
63% in CF3CH2OH, 65% in DMSO and 73% in acetonitrile. The isomer 3d is now the major
isomer formed in all the polar solvents tested. This is a striking difference compared to the case
of isomers 2a-c/3a-c discussed above.
7
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
By increasing the size of the substituent in para position of the phenyl ring with a t-butyl group
(1e) or a nonyl chain (1f), the regioselectivity toward the type-3 isomer was further increased.
The amount of type-3 isomer goes from 40% for 3d, to 56% for 3e and to 62% for 3f in
dichloromethane at room temperature. Increasing the temperature leads again to a slight
enhancement of the type-3 isomers. When the reaction is carried out in polar solvents, the
amount of 3e and 3f is again increased clearly confirming the trend exposed above. Thus, at the
reflux of acetonitrile, the ratio of 3e reaches 74 % and that of 3f reaches 84%. The latter case is
the highest regioselectivity toward type-3 isomer found for this series of compounds.
Regarding the effect of the temperature, of the solvent and of the size of the substituents borne
by the fluorene units, the same trend is hence observed for all these molecules (bearing alkyl or
aryl groups). Nevertheless, in the aryl series, it is important to mention that in polar solvents
i.e. DMSO, MeCN and CF3CH2OH, the major isomer obtained is now always the type-3, which
was not the cases with small alkyl substituents. Thus, examination of the results compiled in
Table 1 clearly shows that as R becomes more encumbering (from R = H to R = 4-nonylphenyl),
the distribution of isomers gradually shifts towards the formation of type-3 at the expense of
type-2 molecules. As the bulkiness of the R group changes from small to large, the isomer
distribution becomes even almost totally reversed (2a/3a = 99/1 and 2f/3f = 16/84).
6. Molecular modelling of pro-2d/3d and pro-2e/3e transition states
Consideration of the optimized geometry of the transition states between the carbocation and
the Wheland intermediate, the aryl-substituted compounds (2d-f/3d-f) display the same
expected features as the previously discussed alkyl or unsubstituted congeners (2a-c/3a-c).
Indeed, as can be seen in Figure 3 (and with the manipulation of the optimized geometry files
provided as supporting information), the introduction of a para-substituted phenyl substituent
at the fluorene moieties, (here with the pro-2d/3d R=PhF (top) and pro-2e/3e R=Ph-t-Bu
(bottom) taken as examples) leads to a potential steric hindrance at the level of the transition
state between two phenyl groups from opposite fluorene moieties (Figure 3, left).
The increased length and encumbrance brought about by the para-substituted phenyl at the 2.7
position of the fluorene moieties does not affect the accessibility to the pro-3 transitions states
structure (Figure 3, right) as it can be visualized by manipulating the structures provided in the
supporting information: Lengthening the longitudinal substitution at the fluorene 2,7 position
do not provoke an encumbrance between the substituents of opposite fluorene groups. On the
contrary this steric hindrance between the substituents of opposite fluorene groups in the pro-2
isomers (Figure 3, left) can be anticipated and is better appreciated by by manipulating the
structures provided in the supporting information. Regarding the relative energy between the
carbocation and the transition states for the alkyl- or non-substituted (R=H) congeners vs. the
aryl substituted ones, one notes that this energy gap increases more significantly in the case of
the pro-2 isomers (for example by ca. 0.15 eV or 3.46 kcal/mol for pro-2d vs. pro-2a) than for
the pro-3 isomers where the values are closer (for example by ca. 0.02 eV or 0.46 kcal/mol for
pro-3d vs. pro-3a). Again, this energy trend is subtle but consistent with the increased ratio of
3-type isomers with increased bulkiness of the fluorene substituents. Thus, the size of the
substituents grafted on the fluorene units leads to pro-2 and pro-3 transitions states possessing
drastically different steric hindrance, which undoubtedly drives the regioselectivity of the
isomers formed. However, one cannot exclude at this stage, an additional effect at the transition
state between the two pro-2 and pro-3 carbocation rotamers, although molecular modeling
failed to provide support for this latter hypothesis as no relevant transition state were found
between the rotamers.
8
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
pro-2d and pro-3d transition states
pro-2e and pro-3e transition states
Figure 3: Views of transition states of pro-2d (top left) and pro-3d (top right), pro-2e (bottom
left) and pro-3e (bottom right) with hydrogen atoms omitted for clarity, fluorine atoms appear
in green.
7. Case of the very bulky meta-substituted R=3,5-di-t-butylphenyl substituent, reaction of 1g
leading exclusively to 3g.
As the results presented above clearly evidence that the steric hindrance induced by the
substituents allows to increase the selectivity towards the type-3 isomer, a very bulky
substituent has been finally investigated. However, to regioselectively obtain a type-3 isomer,
the phenyl substituent should be functionalized in meta positions and not in para in order to
completely prevent the formation of type-2 isomers. Thus, the bulky 3,5-di-t-butylphenyl
fluorenol 1g was investigated, bearing two bulky t-Bu groups in meta positions of the phenyl
ring and a regioselective reaction was observed. Indeed, in all the investigated solvents, the
cyclization of 1g exclusively leads to the formation of the isomer 3g with no trace of the other
9
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
isomer 2g. In this last example, the relative stability of the type-2 isomer (0.48 eV or 11.07
kcal/mol, Table 2) is nevertheless similar to that of the t-Bu substitution case (2c/3c) (0.51 eV,
11.76 kcal/mol) confirming the conclusions drawn above. Unfortunately, molecular modelling
of the pro-3g transition state failed to converge to a stable structure which we assign in part to
the number of rotating t-Bu groups in the molecule. However, a structure of a plausible pro-3g
transition state and consistent with transition states obtained for other type 3 isomers is
nevertheless presented in Figure 4 (middle) together with the optimized pro-3g carbocation
(left) and the optimized pro-3g Wheland intermediate (right). The reaction is quantitatively
regioselective and Figure 4 represents the credible sequence toward the least energetically
stable isomer 3g. It is important to note that calculations attempts of the pro-2g transition state
yielded irrelevant structures with respect to cyclization toward the Wheland intermediate.
pro-3g carbocation
Estimated pro-3g transition state pro-3g Wheland intermediate
Figure 4: Views of optimized geometry of the pro-3g carbocation (left) and Wheland
intermediate (right). The estimated structure of pro-3g transition state (middle) is given as
illustration only as the calculation failed to converge to a stable structure. Hydrogen atoms are
omitted for clarity.
This example clearly highlights the direct influence of the steric hindrance induced by the
substitution of the phenyl rings on the regioselectivity of the reaction.
Conclusion
In summary, the cyclization of alkyl- and aryl-substituted difluorenols 1a-g appears to be
connected to the same parameters. Indeed, low polarity and weakly donor solvents, low
temperature and small substituents on the fluorene moieties all favour the synthesis of 2-type
positional isomers. However, the most important feature, which drives the regioselectivity is
directly linked to the size and the position of the substituents borne by the fluorene units. Indeed,
a small substituent induces the formation of type-2 positional isomers whereas a bulky one leads
to the other family of regioisomers, i.e. type-3 positional isomers. In the light of the steric
hindrance of the transition states, the origin of the regioselectivity has been pinpointed and is
10
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
not related to the relative stability of the regioisomers themselves (Table 2) but to the relative
structure and accessibility of the transition states.
As the three parameters involved, i.e. solvent, temperature and steric hindrance can be
combined in many different ways, we are convinced that this cyclization reaction still deserves
to be investigated to fully explore its potential. Our recent works on the same cyclization
reaction involving a different central terphenyl backbone possessing a meta linkage[4, 49] instead
of a para linkage show the importance of the fluorene rotation on the selectivity of the reaction,
Figure 5. Indeed, preliminary results presented in Figure 5 show that the syn [2,1-b] isomer is
almost always favoured (with either R=H or R=t-Bu) over the anti [1,2-a] isomer with a degree
of modulation of the regioselectivity far less important than for the [2,1-a]/[1,2-b] couple
reported here. These results show the peculiarity of the present systems and the importance of
the central terphenyl backbone on the selectivity of the cyclization reaction.
Figure 5. Top. Synthesis of a different generation of dispirofluorene-Indenofluorene postional
isomers constructed on the meta terphenyl backbone. Bottom. ratio of isomers formed (%,
determined by 1H NMR spectroscopy).
11
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
Acknowledgments.
DT thanks the Région Bretagne for a studentship. We thank Dr Maxime Romain (University of
Rennes 1) for his help in some cyclization experiments and the first results on the cyclization
of meta difluorenol compounds. This work was granted access to the computing ressources of
CINES (Montpellier, allocation 2018-A0020805032 awarded by GENCI).
Experimental section
Synthesis. The synthesis/characterization of the difluorenols 1a-g and the characterizations of
corresponding DispiroFluorene-IndenoFluorenes 2a-g and 3a-g have been previously
reported.[39]
General procedure for the cyclization reactions: Difluorenol 1a-g (10 mg) was dissolved in the
solvent (10 mL) and stirred for 15 min at room temperature or at the reflux of the solvent. Boron
trifluoride etherate (48% BF3) (5 µL) was added and the resulting mixture was stirred for one
hour (until total conversion of the starting difluorenol) at the desired temperature. The crude
mixture was evaporated to dryness and the ratio of isomers determined by 1H NMR.
In the case of cyclization reactions in AcOH, a slightly different procedure was performed.
AcOH (10 mL) was heated to reflux before to add difluorenol 1a-g (10 mg) and 20 µL of
concentrated hydrochloric acid (35%) was added and the resulting mixture was stirred for one
hour (until total conversion of the starting difluorenol) at the desired temperature. The crude
mixture was evaporated to dryness and the ratio of isomers determined by 1H NMR.
Molecular modelling.
Files with the optimized geometry are available as supporting information for viewing and
manipulating the structures: S1pro2aTS.log and S2pro3aTS.log (Figure 1); S3pro2cTS.log and
S4pro3cTS.log (Figure 2); S5pro2dTS.log, S6pro3dTS.log, S7pro2eTS.log and S8pro3eTS.log
(Figure 4); S9pro3gTS.log, S10pro3gCat.log, and S11pro3gW.log (Figure 5).
Transition states geometry optimizations were performed using Density Functional Theory [50,
51]
using the default convergence criteria implemented in the Gaussian software,[52] version
[53-55]
2003/D02 or more recent releases. The hybrid Becke-3 parameter exchange functional
and the Lee-Yang-Parr non-local correlation functional[56] (B3LYP) were used together with
the 6-31G* basis set.[57] Either a guessed structure of the transition state was used as input with
the opt=TS keyword or two structures of the previously optimized carbocations and Wheland
intermediates[41] were used as input with the opt=QST2 keyword. Full convergence was
obtained for all transition state models considered with the exception of pro-3g. Frequency
calculations for pro-2a/3a, pro-2c/3c and pro-3d showed only one significant negative
frequency around -300 cm-1. No further frequency calculations were carried out. Despite the
failed convergence of the transition state of pro-3g assigned to the rotation of the 8 t-Bu groups
(or 24 Me groups), the structure obtained by the optimization sequence has been used for an
illustration purpose. Attempts to optimize the transitions states between the rotamers of the
carbocations or of the transition state of pro-2g led to irrelevant structures. The .log files of the
selected structures presented in the paper are provided as supporting information so that they
can be readily opened and manipulated with the GaussView (version 5.0.9) software that was
used to generate the Figures.
12
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
References
[1]L. Sicard, H.-C. Li, Q. Wang, X.-Y. Liu, O. Jeannin, J. Rault-Berthelot, L.-S. Liao, Z.-Q.
Jiang, C. Poriel, Angew. Chem. Int. Ed. 2019, 58, 3848–3853.
[2]C. Poriel, J. Rault-Berthelot, J. Mater. Chem. C 2017, 5, 3869-3897
[3]T. P. I. Saragi, T. Spehr, A. Siebert, T. Fuhrmann-Lieker, J. Salbeck, Chem. Rev. 2007,
107, 1011-1065.
[4]M. Romain, D. Tondelier, J.-C. Vanel, B. Geffroy, O. Jeannin, J. Rault-Berthelot, R.
Métivier, C. Poriel, Angew. Chem. Int. Ed. 2013, 52, 14147-14151.
[5]L.-H. Xie, J. Liang, J. Song, C.-R. Yin, W. Huang, Curr. Org. Chem. 2010, 14, 2169-2195.
[6]S. Lee, B. Kim, H. Jung, H. Shin, H. Lee, J. Lee, J. Park, Dyes Pigm. 2017, 136, 255-261.
[7]H. Lee, H. Jung, S. Kang, J. H. Heo, S. H. Im, J. Park, J. Org. Chem 2018, 83, 2640-2646.
[8]X. Yang, X. Xu, G. Zhou, J. Mater. Chem. C 2015, 3, 913-944.
[9]C. Quinton, S. Thiery, O. Jeannin, D. Tondelier, B. Geffroy, E. Jacques, J. Rault-Berthelot,
C. Poriel, ACS Appl. Mater. Interfaces 2017, 9, 6194–6206.
[10]L.-S. Cui, Y.-M. Xie, Y.-K. Wang, C. Zhong, Y.-L. Deng, X.-Y. Liu, Z.-Q. Jiang, L.-S.
Liao, Adv. Mater. 2015, 27, 4213-4217.
[11]Y. Liu, L.-S. Cui, X.-B. Shi, Q. Li, Z.-Q. Jiang, L.-S. Liao, J. Mater. Chem. C 2014, 2,
8736-8744.
[12]M.-M. Xue, Y.-M. Xie, L.-S. Cui, X.-Y. Liu, X.-D. Yuan, Y.-X. Li, Z.-Q. Jiang, L.-S.
Liao, Chem. Eur. J. 2016, 22, 916-924.
[13]Y.-K. Wang, Q. Sun, S.-F. Wu, Y. Yuan, Q. Li, Z.-Q. Jiang, M.-K. Fung, L.-S. Liao, Adv.
Funct. Mat. 2016, 26, 7929-7936.
[14]Y.-K. Wang, S.-H. Li, S.-F. Wu, C.-C. Huang, S. Kumar, Z.-Q. Jiang, M.-K. Fung, L.-S.
Liao, Adv. Funct. Mater. 2018, 28, 1706228 (1-10).
[15]Z. Yang, Z. Mao, Z. Xie, Y. Zhang, S. Liu, J. Zhao, J. Xu, Z. Chi, M. P. Aldred, Chem.
Soc. Rev. 2017, 46, 915-1016.
[16]I. Bulut, P. Chavez, S. Fall, S. Mery, B. Heinrich, J. Rault-Berthelot, C. Poriel, P.
Leveque, N. Leclerc, RSC Adv. 2016, 6, 25952-25959.
[17]X.-D. Zhu, X.-J. Ma, Y.-K. Wang, Y. Li, C.-H. Gao, Z.-K. Wang, Z.-Q. Jiang, L.-S. Liao,
Adv. Funct. Mater. 2019, 29, 1807094.
[18]W. Yu, J. Zhang, X. Wang, X. Liu, D. Tu, J. Zhang, X. Guo, C. Li, Solar RRL 2018, 2,
1800048.
[19]J. Yi, Y. Wang, Q. Luo, Y. Lin, H. Tan, H. Wang, C.-Q. Ma, Chem. Commun. 2016, 52,
1649-1652.
[20]K. C. Song, R. Singh, J. Lee, D. H. Sin, H. Lee, K. Cho, J. Mater. Chem. C 2016, 4,
10610-10615
[21]X.-F. Wu, W.-F. Fu, Z. Xu, M. Shi, F. Liu, H.-Z. Chen, J.-H. Wan, T. P. Russell, Adv.
Funct. Mat. 2015, 25, 5954-5966.
[22]S. Li, W. Liu, M. Shi, J. Mai, T.-K. Lau, J. Wan, X. Lu, C.-Z. Li, H. Chen, Energy
Environ. Sci 2016, 9, 604-610.
[23]L. Sicard, C. Quinton, J.-D. Peltier, D. Tondelier, B. Geffroy, U. Biapo, R. Métivier, O.
Jeannin, J. Rault-Berthelot, C. Poriel, Chem. Eur. J. 2017, 23, 7719-7723.
[24]K. Ding, Z. Han, Z. Wang, Chem. Asian. J. 2009, 4, 32-41.
[25]X. Cheng, Q. Zhang, J.-H. Xie, L.-X. Wang, Q.-L. Zhou, Angew. Chem. Int. Ed. 2005, 44,
1118-1121.
[26]X. Cheng, J.-H. Xie, S. Li, Q.-L. Zhou, Adv. Synth.Catal. 2006, 348, 1271-1276.
[27]C. Poriel, Y. Ferrand, P. Le Maux, J. Rault-Berthelot, G. Simonneaux, Tetrahedron Lett.
2003, 44, 1759-1761.
13
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
[28]C. Poriel, Y. Ferrand, P. Le Maux, J. Rault-Berthelot, G. Simonneaux, Inorg. Chem.
2004, 43, 5086-5095.
[29]Y. Ferrand, C. Poriel, P. Le Maux, J. Rault-Berthelot, G. Simonneaux, Tetrahedron
Asymmetry 2005, 16, 1463-1472.
[30]F. Moreau, N. Audebrand, C. Poriel, V. Moizan-Baslé, J. Ouvry, J. Mater. Chem. 2011,
21, 18715-18722.
[31]J. Jalkh, S. Thiery, J.-F. Bergamini, P. Hapiot, C. Poriel, Y. R. Leroux, J. Phys. Chem. C
2017, 121, 14228-14237.
[32]C. Dalinot, V. Jeux, L. Sanguinet, T. Cauchy, M. Allain, Y. Morille, V. Bonnin, P.
Leriche, ACS Omega 2019, 4, 4571-4583.
[33]Y. Wu, J. Zhang, Z. Bo, Org. Lett. 2007, 9, 4435-4438.
[34]Y. Wu, J. Zhang, Z. Fei, Z. Bo, J. Am. Chem. Soc. 2008, 130, 7192-7193.
[35]C. Xu, A. Wakamiya, S. Yamaguchi, J. Am. Chem. Soc. 2005, 127, 1638-1639.
[36]C. Poriel, J. Rault-Berthelot, Acc. Chem. Res. 2018, 51, 1818-1830.
[37]H. Lee, J. Oh, H. Y. Chu, J.-I. Lee, S. H. Kim, Y. S. Yang, G. H. Kim, L.-M. Do, T.
Zyung, J. Lee, Y. Park, Tetrahedron 2003, 59, 2773-2779.
[38]C. Poriel, J. Rault-Berthelot, D. Thirion, J. Org. Chem. 2013, 73, 886-898.
[39]D. Thirion, C. Poriel, R. Métivier, J. Rault-Berthelot, F. Barrière, O. Jeannin, Chem. Eur.
J. 2011, 17, 10272-10287.
[40]D. Thirion, C. Poriel, F. Barrière, R. Métivier, O. Jeannin, J. Rault-Berthelot, Org. Lett.
2009, 11, 4794-4797.
[41]C. Poriel, F. Barrière, D. Thirion, J. Rault-Berthelot, Chem. Eur. J. 2009, 15, 13304-
13307.
[42]M. Romain, C. Quinton, D. Tondelier, B. Geffroy, O. Jeannin, J. Rault-Berthelot, C.
Poriel, J. Mater. Chem. C 2016, 4, 1692-1703.
[43]S. Bebiche, I. Bouhadda, T. Mohammed-Brahim, N. Coulon, J. F. Bergamini, C. Poriel,
E. Jacques, Solid-State Electron. 2017, 130, 49-56.
[44]M. Romain, M. Chevrier, S. Bebiche, T. Mohammed-Brahim, J. Rault-Berthelot, E.
Jacques, C. Poriel, J. Mater. Chem. C 2015, 3, 5742-5753.
[45]C. K. Frederickson, B. D. Rose, M. M. Haley, Acc. Chem. Res. 2017, 50, 977-987.
[46]R. G. Clarkson, M. Gomberg, J. Am. Chem. Soc. 1930, 52, 2881-2891.
[47]W. Linert, Y. Fukuda, A. Camard, Coord. Chem. Rev. 2001, 218, 113-152.
[48]M. Sainsbury, Aromatic Chemistry, Oxford University press Inc., New York, 1994.
[49]C. Poriel, R. Métivier, J. Rault-Berthelot, D. Thirion, F. Barrière, O. Jeannin, Chem.
Commun. 2011, 47, 11703-11705.
[50]P. Hohenberg, W. Kohn, Phys. Rev. 1964, 136, B864-B871.
[51]R. G. Parr, W. Yang, Density-Functional Theory of Atoms and Molecules, Oxford
University Press: Oxford, 1989.
[52]M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman,
R. Scalmani, G. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X.
Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. J. Montgomery, J. E. Peralta, F. Ogliaro, M.
Bearpark, J. J. Heyd, E. Brother, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J.
Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M.
Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C.
Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. C. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski,
G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B.
14
This article is protected by copyright. All rights reserved.

10.1002/chem.201901457
Chemistry - A European Journal
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09, Revision B.01, Gaussian,
Inc., Wallingford CT 2010
[53]A. D. Becke, Phys. Rev. 1988, 38, 3098-3100.
[54]A. D. Becke, J. Chem. Phys. 1993, 98, 1372-1377.
[55]A. D. Becke, J. Chem. Phys. 1993, 98, 5648-5652.
[56]C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785-789.
[57]P. C. Hariharan, J. A. Pople, Chem. Phys. Lett. 1972, 16, 217-219.
15
This article is protected by copyright. All rights reserved.