Drawbacks Arising from the High Steric Congestion in the Synthesis of New Dendritic Polyalkylaromatic Polyradicals

Article abstract of DOI:10.1021/jo9622385

The synthesis of the highly strained tris(α,α-bis(pentachlorophenyl)-2,4,5,6-tetrachlorotolyl)methane (8), which is the precursor of the first generation of the polyradical series III, has been achieved by exhaustive chlorination of compound 11. Stepwise

Full text of DOI:10.1021/jo9622385

J . Org. Chem. 1997, 62, 9009-9017
9009
Dr a w ba ck s Ar isin g fr om th e High Ster ic Con gestion in th e
Syn th esis of New Den d r itic P olya lk yla r om a tic P olyr a d ica ls
†
†
‡
†,
Daniel Ruiz-Molina, J aume Veciana, Fernando Palacio, and Concepci o´ Rovira *
Institut de Ci e` ncia de Materials de Barcelona (CSIC), Campus U.A.B, 08193 Bellaterra (Spain) and
Instituto de Ciencia de Materiales de Arag o´ n (CSIC-UZ) Universidad de Zaragoza,
E-50009 Zaragoza (Spain)
Received December 3, 1996 (Revised Manuscript Received September 4, 1997X)
The synthesis of the highly strained tris(R,R-bis(pentachlorophenyl)-2,4,5,6-tetrachlorotolyl)methane
(8), which is the precursor of the first generation of the polyradical series III, has been achieved by
exhaustive chlorination of compound 11. Stepwise divergent synthesis, by successive Friedel-
Crafts reactions, of the precursors of dendritic series II and III was not possible since the limit
generation was reached in the earlier synthetic steps due to high steric congestion. Several
polyradical mixtures derived from 8 have also been prepared, being stable under ambient conditions.
The formation, electrochemistry, and magnetic properties of these polyradical species are discussed.
The steric hindrance of highly chlorinated hydrocarbon 8 prevents the formation of tetraradical
1
5.
In tr od u ction
initiator core (Chart 1). All are based on highly chlori-
nated 1,3-connected polyarylmethyl radical centers for
which theory predicts ferromagnetic couplings between
the neighboring spins and, therefore, high-spin ground
states for the resulting dendrimers. Moreover, due to
their hyperbranched topologies, they must have large
energy gaps between the ground states and the first
excited ones. High thermal and chemical stability are
indeed expected for all the members of these series due
to the great steric hindrance produced by the chlorine
atoms in the ortho positions that proved to be very
The development of magnetic materials with meso-
scopic size is an area of increasing interest due to the
new and exotic properties expected for such compounds,
taking into account that nanometer size structures can
manifest quantum-mechanics effects at the macroscopic
scale.1 The potential use of such compounds requires
specific chemical and structural characteristics such as
a high stability, uniformity, and a well defined size.
These compounds, however, conflict with the realization
that “engineering down” approaches are unpractical.
Hence, with the possibility of “engineering up” from an
initiator core, large and complex molecules have become
an increasingly attractive prospect.
effective in the stabilization of a wide number of similar
mono- and biradicals.5 The series of dendrimer polymers
I differs from series II and III in the nature, size, and/or
multiplicity (or branching) of their central core unit, N
and the three series differ in their branching-juncture
multiplicities, N . Thus, the series I has a hyperbranched
topology with N ) 3 and N ) 4, and the series II has
) 6 and N ) 4 while the series III has a lower level
of branching with N ) 3 and N ) 2 and the topology of
a three-coordinated Caley tree.
c
,
2
Dendrimers, also called cascade molecules or cauli-
flower compounds, are step growth macromolecules with
3
b
symmetrical three-dimensional structures with a sharply
defined size. These compounds show unusually interest-
ing properties such as fractality, monodispersivity, me-
soscopic dimensions, and endoreceptor properties, being
suggested as units for the construction of three-dimen-
c
b
N
c
b
c
b
Up to now, we have obtained the first polyradical
generation of the series I which, as predicted, is a robust
quartet that displays a high stability and show typical
starbust dendrimeric characteristics such as a fractal
character and a spheroidal shape.6 Nevertheless, the
second generation of this dendritic series revealed a drop
2
sional devices on a molecular level (nanotechnology). On
the other hand, purely organic magnetic materials are
current topics of great interest, one of the general
successful approaches being those based on π-conjugated
polyradicals with topologically polarized π spins.4 Thus,
combination of the last type of polyradicals with a
dendritic nature seems a proper way to achieve mesos-
copic organic magnetic materials.
,7
(4) (a) Mataga, N. Theor. Chim. Acta 1968, 10, 1972. (b) Ovchinikov,
A. A. Theor. Chem. Acta 1978, 47, 297. (c) Veciana, J .; Rovira, C.;
Crespo, M. I.; Armet, O.; Domingo, V. M.; Palacio, F. J . Am. Chem.
Soc. 1991, 113, 2552. (d) Rajca, A.; Utamapanya, S.; Xu, J . J . Am.
Chem. Soc. 1991, 113, 9235. (e) Rajca, A.; Utamapanya, S.;
Thayumanavan, S. Ibid. 1992, 114, 1884. (f) Rajca, A.; Utamapanya,
S. Ibid. 1991, 115, 2396. (g) Rajca, A.; Rajca, S. Ibid. 1995, 117, 806.
Following this approach, we started the study of three
different series of polyradical dendrimers, differing in the
(
h) Fujita, I.; Teki, Y.; Takui T.; Kinoshita, T.; Itoh, K.; Miko, F.;
†
Insitut de Ciencia dels Materials de Barcelona.
Instituto de Ciencia de Materiales de Arag o´ n.
Abstract published in Advance ACS Abstracts, November 15, 1997.
1) (a) Sessoli, R.; Tsai, H. L.; Schake, A.; Wang, S.; Vincent, J . B.;
Sawaki, Y.; Iwamura, H.; Izuoka, A.; Sugawara, T. Ibid. 1990, 112,
4074. (I) Nakamura, N.; Inoue, K.; Iwamura, H.; Fujioka, T.; Sawaki,
Y. Ibid. 1992, 114, 1484. (j) Matsuda, K.; Nakamura, N.; Takahasi,
K.; Inoue, K.; Koga, N.; Iwamura, H. Ibid. 1995, 117, 5550. (k)
Nakamura, N.; Inoue, K.; Iwamura, H. Angew. Chem., Int. Ed. Engl.
1993, 32, 872. (l) Matsuda, K.; Nakamura, N.; Inoue, K.; Koga, N.;
Iwamura, H. Bull. Chem. Soc. J pn. 1996, 69, 1483.
(5) (a) Ballester, M. Acc. Chem. Res. 1985, 380. (b) Armet, O.;
Veciana, J .; Rovira, C.; Riera, J .; Casta n˜ er, J .; Molins, E.; Rius, J .;
Miravitlles, C.; Olivella, S.; Brichfeus, J . J . Phys. Chem. 1987, 91, 5608.
(c) Veciana, J .; Rovira, C.; Armet, O.; Riera, J .; Casta n˜ er, J .; Vincent,
E.; Radhakrishna, P.; Mol. Cryst. Liq. Cryst. 1988, 156, 301. (d)
Veciana, J .; Rovira, C.; Crespo, M. I.; Armet, O.; Domingo, V. M.;
Palacio, F. J . Am. Chem. Soc. 1991, 113, 2552. (e) Ballester, M.;
Pascual, I.; Carreras, C.; Vidal-Gancedo, J . J . Am. Chem. Soc. 1994,
116, 4205.
‡
X
(
Folting, K.; Gatteschi, D.; Christou, G.; Hendrickson, D. N. J . Am.
Chem. Soc. 1993, 115, 1804. (b) Chien, C. L. Recent Advances in
Magnetism and Magnetic Materials; Huang, H. L., Kuo, P. C., Eds.;
World Scientific: Singapore, 1989. (c) Kahn, O. Molecular Magnetism;
VCH: New York, 1993. (d) Awschalom, D. D.; DiVicenzo, D. P.; Smyth,
J . F. Science 1992, 258, 414.
(
2) (a) Tomalia, D. A.; Naylor, A. M.; Goddard III, W. A. Angew.
Chem., Int. Ed. Engl. 1990, 29, 138. (b) Tomalia, D. A. Adv. Mater.
994, 6, 529. (c) V o¨ gtle, F.; Issberner, J .; Moors, R. Angew. Chem., Int.
Ed. Engl. 1994, 33, 2413. (d) Frechet, J . M. J . Science 1994, 263, 1715.
3) Young, J . K.; Baker, G. R.; Newkome, G. R.; Morris, K. F.;
J ohnson, C. S., J r. Macromolecules 1994, 27, 3464.
1
(
S0022-3263(96)02238-4 CCC: $14.00 © 1997 American Chemical Society

9
010 J . Org. Chem., Vol. 62, No. 26, 1997
Ruiz-Molina et al.
Ch a r t 1
of the effective S value.7 Such a drop has also been
observed by Rajca et al. in analogous dendritic polyradi-
cals with up to 7, 15, and 31 sites for ferromagnetically
coupled electrons that surprisingly show spin averages
of 3, 7/2, and 5/2, respectively. The authors have at-
tributed tentatively this phenomena to the presence of
defects in these π conjugated systems that interrupt the
ferromagnetic coupling to the extent that some parts of
the polyradicals become effectively uncoupled.8
of formation of carbanions with defects arising from
negative charges missing specially in those centers that
are more shielded. Since the central core unit of series
II and III is larger than that of series I, it seemed
promising to move toward these two new series in order
to decrease the hindrance and, as a result, to obtain
higher generations of stable polyradical series.
We now report the results concerning the synthesis of
the polytriphenylmethanes precursors of the first genera-
tion of series II and III as well as the synthesis and the
most relevant properties of the polyradicals that belong
to the first generation of series III. Unexpectedly, the
topologies of these new polyradicals not only do not
decrease but even increase the steric congestions and
affect the formation of these new radicals.
Due to the promising properties of these polyradical
dendrimers our aim was focused to the design of alterna-
tive molecular topologies, those of the series II and III,
that overcome such steric congestions. Polychlorotri-
phenylmethyl polyradicals are usually synthesized fol-
lowing three successive steps: (1) obtaining the poly-
triphenylmethane precursors, (2) generation of the cor-
responding polycarbanions by acid-base reactions, and
Resu lts a n d Discu ssion
(3) oxidation of such polycarbanions to the corresponding
polyradicals. One of the main drawbacks lies in the
difficulty in growing well-characterized polytriphenyl-
methane precursors (step 1). Due to the large number
of bulky chlorine atoms, the limit generation in the
precursors of series of polyradicals can be achieved only
at early generations. The limit generation is a self-
limiting step characterized by the presence of high steric
congestion and over which it is very difficult to continue
growing new generations without structural defects.
Concerning the synthesis of the polyradicals, the limiting
step is the generation of the polycarbanions. The pres-
ence of large steric congestions increase the probability
Syn th esis a n d Ch a r a cter iza tion of th e P r ecu r sor
P olytr ia r ylm eth a n es. The synthetic approach was
based on the divergent dendrimeric growth methodol-
ogy.2 The first step is the functionalization of the
initiator core of the series II, i.e., of tris(2,4,6-trichlo-
rophenyl)methane (1), by the introduction of new branch
2
junction elements (CHCl groups) in each meta position
of the phenyl rings. Thus, compound 2 was obtained in
high yield by a Friedel-Crafts alkylation of 1 with
chloroform. Each introduced CHCl group is a potential
2
alkylating agent, with a branching multiplicity of 2,
which can be used for the construction of the next
generation of this series of compounds (Scheme 1). To
obtain the precursor of the polyradical of the first
generation, reaction of 2 with pentachlorobenzene is
required. To continue toward the precursor of the next
member of series II, compound 3, three consecutive
Friedel-Crafts reactions must be accomplished: first
that of 2 with 1,3,5-trichlorobenzene, second the reaction
,3
9
(6) (a) Veciana, J .; Rovira, C. Magnetic Molecular Materials Gatte-
schi, D.; Kahn, O.; Miller, J . S.; Palacio, F. Eds. NATO ASI Series
Kluwer Academic Dordrecht, 1991, 121. (b) Ventosa, N.; Ruiz, D.;
Rovira, C.; Veciana, J . Mol. Cryst. Liq. Cryst. 1993, 232, 333. (c)
Veciana, J , Rovira, C.; Ventosa, N.; Crespo, M. I.; Palacio, F. J . Am.
Chem. Soc. 1993, 115, 57.
(7) Ventosa, N. Ph.D. Thesis. Insitut Quimic de Sarria, 1996,
Barcelona, Spain.
(
(
8) Rajca, A. Chem. Rev. 1994, 94, 871 and references therein.
9) Ballester, M.; Riera, J .; Casta n˜ er, J .; Rovira, C.; Armet, O.
of the obtained compound 4 with CHCl
3
, and finally the
Synthesis 1986, 64.
resulting compound with pentachlorobenzene.

Synthesis of New Dendritic Polyalkylaromatic Polyradicals
Sch em e 1
J . Org. Chem., Vol. 62, No. 26, 1997 9011
Sch em e 2
Unfortunately, the reaction of 2 with pentachloroben-
zene or 1,3,5-trichlorobenzene via an acid-catalyzed
Friedel-Crafts reaction failed along different experimen-
tal conditions tested. To all appearances, the propeller-
like conformation of triarylmethane units and the high
effective volume of chlorine atoms originate a high
shielding in the earlier dendrimeric generations that
prevents this highly disfavored reaction to proceed in the
desired way.10 We, therefore, tried the reaction of 2 with
benzene provided that benzene has not such bulky
substituents and, consequently, displays a smaller van
der Waals volume as well as an enhanced reactivity with
respect to chlorinated analogues. Unfortunately, all
attempts to isolate the condensation compound 5 were
once more unsuccessful since in all reactions tested we
obtained a complex mixture of products. The results
suggest that the limit generation had been reached in
the first synthetic step, due to the steric hindrance, and
furthermore that the reactive intermediates formed
undergo further reactions such as dealkylations or hy-
dride additions, which are typical of reversible Friedel-
Triphenylmethane 6 was synthesized by a Lewis acid-
catalyzed condensation of 1,2,3,5-tetrachlorobenzene with
chloroform. Once we synthesized the initiator core, we
moved toward the synthesis of the first generation of
series III using a two-stage reaction. First, 6 was
alkylated with chloroform, yielding the triarylmethane
7 that has three branch junction elements (Scheme 2).
However, as in the previous series, the second step, i.e.,
the subsequent Friedel-Crafts reaction of 7 with pen-
tachlorobenzene and 1,3,5-trichlorobenzene, failed. Both
reactions were further attempted by varying the key
reaction parameters such as the ratios between aryl,
catalyst, and alkylating agent and the reaction temper-
ature and time. In all conditions assayed, complex
mixtures of products were obtained in which it was not
possible to identify the desired compounds by spectro-
scopic techniques. The same discouraging results were
obtained in the reaction of 7 with pentafluorobenzene,
despite the smaller volume of the aromatic substrate and
the fact that pentafluorobenzene gives Friedel-Crafts
reactions with chloroform and dichloromethane in high
1
1
12
Crafts reactions.
Therefore, to avoid such big steric
yields.
hindrance problems we turned our attention toward a
new series of highly chlorinated hydrocarbons, the pre-
cursors of series III (Chart 1). The nature of the
structural components in this series was maintained as
in series II, although, the multiplicity of the core and
branches was strongly decreased. As a starting precursor,
acting as the initiator core of the series III, we chose
tris(2,3,4,6-tetrachlorophenyl)methane (6), which only
contains three reacting sites instead of the six of com-
pound 1, the precursor of the initiator core of series II.
Polytriphenylmethane compounds obtained from this
compound by successive Friedel-Crafts reactions were
intended to act as the precursors of dendrimeric polyradi-
cals of series III.
Finally, reaction of 7 with benzene was successful since
it gave the polytriphenylmethane 11 (Scheme 3) in
variable yields depending upon the reaction conditions
(Table 1). In fact, extreme conditions such as a high
temperature (reflux) and large excess of benzene and the
Lewis acid, AlCl , with respect to the alkylating agent
3
7, molar ratios larger than 60:1 and 3:1, were required
to achieve high yields (80%).
The results demonstrate that 7 can act as an alkylating
agent, although it did not react with polyhalogenated
benzenes. This lack of reactivity could be ascribed to the
enhancement of the steric hindrance of the reaction
intermediates and also to the electronic deactivation that
precludes the desired reactions. It should be noticed that
the optimum reaction conditions to obtain 11 were also
used in the previous attempted reactions of 7 with
polyhalogenated benzenes reaching to the same previous
disappointing results. To sum up, although it is well-
(10) The large steric hindrance and the negative inductive effects
produced by the chlorine atoms in both the substrate and the alkylating
agent make it necessary to use extreme reaction conditions, such as
an excess of the Lewis acid, high temperatures, long times, and to carry
out the reaction in absence of solvent.
(
11) Roberts, R. M.; El-Khawaga, A. M.; Sweeney, K. M.; El-Zohry,
M. F. J . Org. Chem. 1987, 52, 1591.
(12) Beckert, W. F.; Lowe, J . U. J . Org. Chem. 1967, 32, 1212.

9
012 J . Org. Chem., Vol. 62, No. 26, 1997
Ruiz-Molina et al.
Sch em e 3
Ta ble 1. Differ en t Rea ction Con d ition s for th e
Due to the presence of bulky chlorine atoms, the aromatic
rings in 6, 8, and 11 as well as the CHCl
groups in 2
and 7 have highly restricted rotations, giving rise to a
a
Syn th esis of 11
2
C6H6/6b
AlCl3/6b
time (h)
yield (%)
large number of atropoisomers with different populations.
For this reason, H NMR spectra of all these compounds
consist of multiplets on account of the non equivalence
of protons belonging to different atropoisomers. All the
reported compounds are moderately soluble in typical
organic solvents such as CH Cl , CHCl , C H , and THF.
2 2 3 6 6
Amazingly, the solubility increases with increasing the
molecular weight, a general trend that was previously
4
00
00
200
200
200
200
200
200
200
1
7
7
7
7
0.3
1
14
7
7
0.5
0.5
0.5
0.5
0.5
0.5
0.5
2.5
3.5
6.5
30
88
89
85
c
27
d
65
62
68
1
2
4
1
1
1
1
1
1
1
7
observed by Neenan et al. in a different series of 1,3,5-
a
15
All the experiments were performed using the same amount
phenylene-based hydrocarbon dendrimers.
Another
b
of the alkylating agent 6 and reflux temperature. Molar ratio.
relevant characteristic of these branched molecules is the
tendency to entrap solvent when crystallized in a wide
c
Starting material recovered. d Complex mixture containing car-
bonilic compounds.
variety of solvents such as C
6 6 6 3
H , n-C H14, CHCl , and
CCl . This makes necessary a thermal treatment of the
4
known that asymmetrical branching, such as that of
series III, should lead to a classical relationship between
density and molecular mass, the results presented here
show that, despite topology changes, the limit generation
was reached at the first steps.
compounds under vacuum in order to perform their
complete characterization. Inclusion compounds of triph-
enylmethane with benzene in a 1:1 ratio were reported
16
earlier. More recently we reported the affinity of the
perchlorotriphenylmethyl radical and some of its radical
derivatives to form inclusion compounds with cyclic small
As the last attempt to synthesize the polytriaryl-
methane 8, the precursor of the second generation of
series III, we performed an exhaustive perchlorination
17
molecules. X-ray structures of the chlatrates showed
that guest molecules were placed into channels in the
1
3
of 11 using the BMC reagent.
This reagent is a
17
structure with a host/guest ratio of 1:1. Obtaining a
powerful chlorinating agent that was extensively used
in the perchlorination of a great number of triaryl-
methanes.14 However, this was the first time in which
such a chlorinating agent has been used with a molecule
such as 11 with a large structural complexity. Indeed,
reaction of 11 with the BMC reagent eventually yielded
the highly chlorinated compound 8. IR analysis showed
a complete disappearance of the stretching bands corre-
sponding to the stretching Carom-H bonds, and a similar
single crystal of the compounds studied here has proven
to be very difficult due to the great number of isomers.
Unfortunately, this fact precludes a complete study of the
inclusion compounds formed.
Syn th esis a n d Ch a r a cter iza tion of th e P olyr a d i-
ca ls. The synthesis of the polyradical species 12-15 was
undertaken by treatment of the polyarylmethane 8 with
+
-
4
a solution of aqueous n-Bu N OH by using different
substrate/base ratios and reaction times in THF followed
by oxidation of the resulting polycarbanions with p-
chloranil.
1
behavior was observed for the corresponding H NMR
signals. The elemental analysis agrees with the spec-
troscopic studies confirming that the chlorination was
smoothly completed without any fragmentation yielding
the dendrimer 8 which contains 42 Cl atoms. By FAB
mass spectroscopy, using positive and negative ions and
various matrixes, we did not observe with 8 any signifi-
cant parent peak. Nevertheless, by MALDI/TOF tech-
nique in negative mode and using 2,5-dihydroxybenzoic
acid as a laser absorbing matrix, the molecular peak is
seen with a series of consecutive losses of 35 units
corresponding to the loss of chlorine atom from the parent
peak.
Our previous experience with the radicals of series I
was that controlling the excess of the base and/or reaction
times in the first step allows the control of the formation
of the corresponding mono-, di-, and tricarbanions and,
therefore, of the desired polyradical. Nevertheless, due
to the lack of knowledge of the precise experimental
conditions required to generate the different carbanionic
species from 8, a preliminary experiment was designed.
An aliquot of 8 was treated with an excess of base, and
the evolution of the reaction was followed by UV-vis
spectroscopy. It can be anticipated that the intensity of
the 518 nm band, characteristic of the carbanionic
Characterization of the rest of synthesized compounds
1
was achieved by H NMR, IR and UV-vis spectroscopies
as well as by elemental analysis and thermal analysis.
(15) (a) Miller, T. M.; Neenan, T. X. Macromolecules 1992, 25, 3143.
b) Miller, T. M.; Neenan, T. X. Chem. Mater. 1990, 2, 346.
(
(
16) (a) Hartley, H.; Thomas, N. G. J . Chem. Soc. 1906, 1013. (b)
(
13) (a) Ballester, M.; Molinet, C.; Casta n˜ er, J . J . Am. Chem. Soc.
Driver, J . E.; Mok, S. F. J . Chem. Soc. 1955, 3914. (c) Driver, J . E.;
Lui, T. F. J . Chem. Soc. 1958, 3219. (d) Barlow, G. B.; Clamp, A. C. J .
Chem. Soc. 1961, 393. (e) Recca, A.; Bottino, F. A.; Libertini, E.;
Finocchiaro, P. Gazz. Chim. Ital. 1979, 109, 213.
1
960, 82, 4254. (b) Fieser, L. F.; Fieser, M. Reagents for Organic
Synthesis; J ohn Wiley & Sons: New York, 1967; p 1131. (c) Ballester,
M.; Riera, J .; Casta n˜ er, J .; Badia, C.; Monso, J . M. J . Am. Chem. Soc.
1
971, 93, 2215.
14) Ballester, M.; Olivella, S. Polychloroaromatic Compounds;
Suschitzy, C., Ed.; Plenum Press: New York, 1974; p 126.
(17) (a) Veciana, J .; Carilla, J .; Miravitlles, C.; Molins, E. J . Chem.
Soc. Chem. Commun. 1987, 811. (b) Veciana, J .; Carilla, J .; Miravitlles,
C.; Molins, E. J . Inclus. Phenom. 1990, 8, 395.
(

Synthesis of New Dendritic Polyalkylaromatic Polyradicals
J . Org. Chem., Vol. 62, No. 26, 1997 9013
Ch a r t 2
Ta ble 2. Differ en t Rea ction Con d ition s Used for th e
Gen er a tion of th e P olya n ion s Der ived fr om 8
reaction
base/8a
reaction times (days)
1
2
3
4
5
1/1
2/1
3/1
100/1
100/1
0.2
0.2
0.2
0.3
28
F igu r e 1. HPLC of polyarylmethane 8 (a), and samples A
(b), B (c), C (d), D (e), E (f) on an ODS column operating at
a
Molar ratio.
263 K and using CH CN/THF mixtures as eluent (55/45).
3
species, will increase until a maximum is reached where
the formation of the tetracarbanionic species should be
assumed. As expected the UV-vis spectra recorded 24
h after the reaction was initiated revealed a relatively
high intensity of the 518 nm band, confirming that the
first carbanion species can be obtained in short periods
of time. From here on the evolution was slower as
expected from an increase of the intramolecular Coulom-
bic repulsions on increasing the number of charges. After
one month of reaction, the intensity of the band keeps
constant for more than two weeks, confirming that the
reaction was concluded.
Taking into consideration the above-mentioned experi-
ment, five different reactions using different substrate/
base ratios and reaction times were carried out and the
resulting carbanionic species were oxidized with an
excess of p-chloranil. The experimental conditions used
in the first step of each of these reactions are summarized
in Table 2.
UV-vis spectra and HPLC experiments give qualita-
tive evidence of the increase of the radical character of
the samples when increasing the base proportion and the
reaction times. Thus, the molar extinction coefficients
6
of the characteristic radical bands at 568 and 389 nm
increase progressively up to the last sample (Table 3) as
expected from an increase of the number of radical
centers in the final products.
On the other hand, the complexity of the chromato-
grams diminishes from samples A to E (see Figure 1),
as expected by the higher number of atropisomers and
the lower energy barriers when the number of Ar
3
CH
•
units is larger than the number of Ar
3
C in a given
molecule.18
Electr och em istr y. Electrochemical studies in CH
2
-
Cl
2
with tetrabutylammonium hexafluorophosphate as
supporting electrolyte were done at room temperature.
In the cyclic voltammogram (CV) of sample A only one
reversible wave, characteristic of a monoradical species,
The samples obtained from the experiments 1-5, that
will be referred to from now on as samples A-E,
respectively, were isolated as deep red-purple solids,
completely stable in ambient conditions after their
chromatographic purification. Finally the samples were
characterized by electrochemistry, UV-vis and ESR
spectroscopies, HPLC, and magnetic measurements.
(
19) Zero-field splitting parameters are much lower than those of
the quartet of series I as well as its corresponding triplet [see ref 6c].
This fact can be due to the location of the radical centers on two or
three electronically isolated external Ar
3
location on neighboring Ar C units that are electronically connected
cannot be excluded since a large torsion of these units, due to the high
steric congestion, could lead also to a large decrease of the zero-field
splitting.
•
3
C
units. Nevertheless, the
•
(
20) Experimental magnetization data, M, were fitted to eq 1,
n
(
18) It has been shown that the different atropisomers of these kind
M
)
f g S B (ø )
(1)
of polyradicals display striking differences in their overall sizes, shapes,
and surfaces, allowing these stereoisomers to be chromatographically
separated. The interconversion between the stereoisomers of these
polyradicals is relatively low due to the bulky o-chlorine atoms that
hinder the correlated rotation of the aryl groups [see: ref 7 and
Ventosa, N.; Ruiz, D.; Tomas, X.; Bieber, A.; Andr e´ , J .-J .; Rovira, C.;
Veciana, J . J . Am. Chem. Soc., submitted.], and, therefore, it can be
quenched by lowering the temperature. Nevertheless, the intercon-
∑
i
i
i
S
i
i
N µ
A
B
i)1
where B i
with g and an effective spin number of S
fraction of f in the mixture of n paramagnetic species. The Brillouin
S
(ø
i
) is the Brillouin function for a paramagnetic species, i,
i
i
which is present with a molar
i
function is given by eq 2,
2S + 1
2S + 1
1
Si
i
i
1
2
1
version of the stereogenic Ar
faster that that of the stereogenic Ar
3
CH centers in these polyradicals is always
BS (øi) )
[
cot gh
(
øi
)
-
cot gh(2øi
)
]
(2)
i
2
2
•
3
C centers [see: Crespo, M. I.
Ph.D. Thesis. Institut Qu ´ı mic de Sarri a´ , 1991, Barcelona, Spain.]. The
stereochemical analysis of tetraradical 15, based on previous consid-
erations given by Mislow et al. [see: Gust, D.; Mislow, K. J . Am. Chem.
Soc. 1973, 95, 1535 and Mislow, K. Acc. Chem. Res. 1976, 9, 26.] and
on energy minimizations performed by empirical calculations with the
Drieding force-field, predicts 12 different enantiomeric pairs for which
different relative abundancies are expected. For the rest of the
polyradicals, a larger number of stereoisomers is expected since the
being the argument, ø , of this function as:
i
giµBH
øi )
(3)
kTeff
where Teff is the effective temperature, Teff ) T - θ, which takes into
account in a first approximation the averaged weak intermolecular
magnetic interactions between the species. For a rigorous treatment
of such intermolecular interactions, a generalization of the molecular
mean-field treatment is required; see ref 6c.
presence of Ar
3
CH moieties introduces additional stereogenic elements.
Thus, 24 different enantiomeric pairs are expected for polyradical 14.

9
014 J . Org. Chem., Vol. 62, No. 26, 1997
Ruiz-Molina et al.
F igu r e 2. Cyclic voltammograms of samples A (a), C (b), E
+
6-
(
2 2 4
c) in CH Cl , with 0.1 M Bu N PF , Pt wire (working and
auxiliary), and Ag/AgCl (reference) electrodes.
Ta ble 3. Mola r Extin ction Coefficien ts (in L cm -1 m ol-1
of th e Tw o Ch a r a cter istic Ra d ica l Absor p tion Ma xim a
for Sa m p les A-E
)
F igu r e 3. First derivative ESR spectra at 165 K in glassy
toluene of samples D (a), E (b). Insets show the observed
signals corresponding to ∆M ) (2 forbidden transitions.
s
sample
band, λ ) 568 nm
band, λ ) 389 nm
A
B
C
D
E
800
1800
2600
3300
4450
22000
48000
68000
77500
102500
trolysis experiments confirm an increase of the radical
character of the samples when increasing the substrate/
base ratio and reaction times used in their synthesis as
well as the formation of mixtures of different polyradicals
in each case. It must also be emphasized that the
presence of only three reversible waves in the CV, instead
of the four waves expected for a tetraradical, and the
consumption of 2.9 Faraday/mol of sample E seem to
point out that we only have been able to synthesize up
to the triradical species 14 and not to the desired
tetraradical 15. Such results were confirmed later on by
magnetic measurements.
Magn etic P r oper ties. X-band ESR spectra of samples
A-E in frozen toluene matrix consists of the superim-
position of two types of lines: a symmetric broad central
line and a few symmetrical lines besides the central one
that increases in complexity on going from sample A to
sample E (see Figure 3). The central line is due to
monoradical species and polyradical species with very
small spin-spin dipolar interactions, whereas the other
set of symmetrical lines are attributed to polyradical
species with larger dipolar interactions.
was observed at E1/2 ) -0.18 V (vs the standard Ag/AgCl
electrode). CV’s of the rest of samples revealed a qualita-
tive increase of the number of electrochemical processes
as expected from the successive replacement of the Ar -
3
3
CH by Ar C units. However, while the analogous
polyradical series derived from the first generation of
series I show well-defined stepwise one-electron reduction
processes, the present series of polyradicals display CV’s
where the different waves can only be differentiated as
a shoulders due to the small peak-to-peak separations
•
(
60-80 mV or even smaller). Representative voltamme-
tries of samples A, C, and E are depicted in Figure 2.
The complexity of these CV’s only allows a qualitative
estimation of the nature of polyradical species formed in
each experiment. A detailed inspection of the CV of
sample E revealed that only three reversible waves are
present. Quantitative results were obtained by bulk
electrolysis experiments. In this way the chronoampero-
metric limiting currents of samples A-E were in agree-
ment with the consumption of 0.6, 1.3, 1.5, 1.8, and 2.9
Faraday/mol, respectively.
Due to the complexity of those signals, any assignment
to a specific species cannot be performed, showing only
that samples consist of complex mixtures of polyradicals.
The only data that can be qualitatively extracted from
the ESR spectra is that the zero-field splitting param-
eters of polyradical species are small, indicating weak
The increase of the number of waves in the CV’s and
the number of Faraday/mol consumed in the bulk elec-

Synthesis of New Dendritic Polyalkylaromatic Polyradicals
J . Org. Chem., Vol. 62, No. 26, 1997 9015
the large number of parameters to be adjusted. There-
fore, to improve the accuracy of the method, we imposed
that the possible solutions should also agree with the
values previously obtained in the chronoamperometric
experiments and in the magnetic susceptibility measure-
ments.
From the results of the fitting shown in Table 4, two
main conclusions can be extracted: (1) Despite the large
excess of base and the long reaction times used, the
tetraradical species 15 remains elusive since sample E
is mostly formed by the triradical species 14. (2) The
control of the formation of a specific polyradical is not
possible by changing the substrate/base ratio and the
reaction times since complex mixtures of polyradicals
were always obtained. Such results suggest that the high
steric congestion of polyarylmethane 8 hinders that the
reagent (base) at all the reactive centers (acidic protons)
of 8, thus precluding the formation of tetraradical 15.
F igu r e 4. Temperature dependence of the product øT for
samples B (∆), C (O), and E (0). Continuous lines are the
magnetic behaviors expected for pure paramagnetic systems
with weak antiferromagnetic intermolecular (θ ) 2 K) having
the indicated spin values.
Con clu sion s
The high steric congestion in the dendritic series II and
III originates that the limit generation is reached in early
synthetic steps. Such a situation prevents us from
obtaining the precursor of the first generation of series
II. By contrast, thanks to the lower branching degree of
series III, we have been able to synthesize the highly
chlorinated hydrocarbon 8, the precursor of the first
generation of series III. However, the large overcrowding
present in this dendritic series and the strong experi-
mental conditions required in obtaining 8 made futile any
effort to produce higher generations of series III without
structural defects.
It has also been shown that the steric congestion does
not only affect the synthesis of the highly chlorinated
hydrocarbon precursors but also the preparation of the
polyradicals series derived from 8. Despite the large
excess of base and the long reaction times used in the
synthesis of such polyradicals, the preparation of the first
target compound, the tetraradical species 15 is elusive.
In addition, it is very difficult to control the formation of
a pure polyradical, always obtaining mixtures of different
species.
spin-spin dipolar interactions between the unpaired
electrons.19 Only samples D and E display significant
signals in the g ≈ 4 region (see insets of Figure 5)
corresponding to forbidden ∆M ) (2 transitions, indi-
S
cating the presence of a larger proportion of high-spin
species.
The magnetic susceptibility was measured in the
temperature range 2-250 K for samples A-E. The øT
vs T plots of such samples give straight lines down to 30
K where small downward deviations, due to the presence
of weak intermolecular antiferromagnetic interactions,
are observed (see Figure 4). The magnetic effective
moments found for samples A-E at room temperature
-
1
were 0.23, 0.45, 0.62, 0.80, and 1.15 emu K mol
,
respectively, showing that the samples are in fact mix-
tures of radical species with different effective spin
numbers S.
Magnetization measurements of samples A-E were
performed at 1.8 K and 3 K as a function of the external
magnetic field. The plots of the magnetization per mol
Exp er im en ta l Section
B
(M/Nµ ) versus the ratio of the magnetic field over the
absolute temperature for the most representative samples
are shown in Figure 5.
To determine the average effective spin numbers and,
All solvents were reagent grade from SDS and were used
as received unless otherwise indicated. All reagents, organic
and inorganic, were of high purity grade and obtained from
E. Merck, Fluka Chemie, and Aldrich Chemical Co. The
Friedel-Crafts condensations were carried out in a glass
pressure reactor. Elemental analyses were obtained in the
Servei de Microanalisi del CID (CSIC), Barcelona. NMR
spectra were recorded on a Bruker AC-400 and on a Bruker
â
therefore, the composition of each sample the M/ Nµ vs
H/T data were fitted to a magnetic model considering the
possible presence of mixtures of the precursor 8, the
monoradical 12 (S ) 1/2), and a triplet (S ) 1) and a
quartet (S ) 3/2) as well as a biradical (2‚S ) 1/2) and
triradical (3‚S ) 1/2) species, both with magnetically
independent radical centers.20 Such a combination of
species makes sense provided that two possible pathways
are expected for the reaction of 8 with the base. The first
pathway initially assumes the succesive generation of the
2
50. The MALDI mass spectrum was recorded using a
KRATOS MALDI spectrometer in Linear High Power negative
mode.
Tr is(2,3,4,6-tetr a ch lor op h en yl)m eth a n e (6). A mixture
of anhydrous chloroform (0.6 mL; 7.4 mmol), aluminum
trichloride (3.95 g; 51.5 mmol), and 1,2,3,5-tetrachlorobenzene
(13.26 g; 61.4 mmol) was heated at 110 °C with magnetic
-
-
3 3
external Ar C units and finally of the central Ar C one.
stirring in a glass pressure vessel during 3.5 h. The mixture
was poured into cracked ice/hydrochloric acid and extracted
with chloroform. The organic layer was washed with aqueous
This pathway is the most likely since it minimizes
intramolecular Coulombic repulsions among the negative
-
charges of the Ar
3
C
units and will provide mixtures of
NaHCO
chromatography (silica gel, hexane) afforded 4.1 g (84%) of 6
3
and with water, dried, and evaporated. Flash
monoradical (S ) 1/2), biradical (2‚S ) 1/2), and triradical
1
2 2
as a spongy white solid: mp 241-243 °C; H NMR (CD Cl ) δ
(
3‚S ) 1/2) species. The second pathway assumes the
-
-
7.00-7.18 (m, 1H), 7.61-7.75 (m, 3H); IR (KBr): ν 3123, 3072,
3 3
formation of the central Ar C and external Ar C units
2
1
956, 2924, 1560, 1532, 1406, 1363, 1335, 1300, 1237, 1181,
117, 948, 854, 811, 718, 643, 618, 575, 532 cm ; UV-vis
simultaneously to obtain, after oxidation, mixtures of
triplet and quartet species. The main drawback of the
-1
6
(C H12) λmax nm (log ꢀ) 219(4.98), 247(4.48), 279(2.92), 289(3.12),
298(3.12). Anal. Calcd for C H Cl : C, 34.70; H, 0.61; Cl,
19 4 12
fitting of M/ Nµ
â
vs H/ T data to a proper magnetic model
for mixtures of species with different S values relies on
64.69. Found: C, 34.61; H, 0.62; Cl, 64.70.

9
016 J . Org. Chem., Vol. 62, No. 26, 1997
Ruiz-Molina et al.
F igu r e 5. Field strength dependence of the magnetization of samples B (a), C (b), D (c), E (d) at (∆) 1.8, and (O) 3 K. Solid lines
are the calculated molar magnetization curves for pure paramagnets in systems with weak antiferromagnetic intermolecular
interactions (θ ) 2 K) having the indicated spin values.
vacuum. 1H NMR (CD
(m, 3H); IR (KBr): ν 3041, 2924, 1532, 1370, 1272, 1251, 1223,
Ta ble 4. Mola r Ra tio of th e Differ en t P olyr a d ica l
Sp ecies F ou n d fr om th e F ittin g of Ma gn etiza tion Da ta
for Sa m p les I-V
2
2
Cl ) δ 6.96-7.20 (m, 1H), 7.60-7.76
1
4
117, 1033, 963, 921, 871, 810, 773, 715, 673, 636, 615, 536,
78, 447 cm ; UV-vis (C H12) λmax nm (log ꢀ) 230(5.13),
6
-
1
θ
8
S )
2‚S )
3‚S )
S )
S )
sample
(K) (%) 1/2 (%) 1/2 (%) 1/2 (%) 1 (%) 3/2 (%)
250(4.72), 286(3.25), 298(3.47), 309(3.54). Anal. Calcd for
Cl18: C, 29.15; H, 0.44; Cl, 70.40. Found: C, 28.92; H,
22 4
C H
A
B
C
D
E
-0.2 45
50
82
34
12
0
5
0
0
0
43
80
0
0
0
5
15
0
0
0
0
5
0
.33; Cl, 70.55.
-0.6
-1.1
-2.0
-2.3
0
0
0
0
18
66
40
0
Tr is(r,r-diph en yl-2,4,5,6-tetr ach lor otolyl)m eth an e (11).
A solution of 7 (0.442 g; 0.5 mmol) in anhydrous benzene (17.7
mL; 200 mmol) containing aluminum trichloride (0.463 g; 3.5
mmol) was refluxed in an inert atmosphere during 30 min and
then poured into cracked ice/hydrochloric acid. The organic
Tr is(r,r,2,4,5,6-h exa ch lor otolyl)m eth a n e (7). A solu-
tion of compound 6 (0.390 g; 0.6 mmol) in chloroform (25 mL;
10 mmol) containing aluminum trichloride (0.095 g; 0.7 mmol)
was refluxed in an argon atmosphere for 3 h. The resulting
brown solution was poured into cracked ice/hydrochloric acid.
The organic phase was separated and washed with aqueous
3
phase was separated and washed with aqueous NaHCO and
with water. After drying over anhydrous sodium sulfate, the
solvent was evaporated and the residue flash chromatographed
3
3
(silica gel, hexane/CHCl 4:1 f 2:1). The resulting main
product was recrystallized from hexane and heated at 90 °C
under vacuum in order to eliminate the included solvent,
NaHCO
sulfate, the solvent was evaporated and the brown residue
flash chromatographed (silica gel, hexane/CHCl 3:1). The
3
and with water. After drying over anhydrous sodium
yielding 11 as a white solid (0.521 g; 89%): mp 245-247 °C,
3
1
2 2
H NMR (CD Cl ) δ 6.38-6.52 (m, 3H), 6.84-7.25 (m, 32H);
white spongy solid obtained was recrystallized from benzene,
and a clahtrate of 7 with benzene in a 2:1 proportion was
isolated (0.545 g, 96%): mp 198-200 °C. Anal. Calcd for
IR (KBr): ν 3091, 3063, 3028, 2924, 1950, 1887, 1810, 1754,
1600, 1537, 1495, 1453, 1376, 1348, 1278, 1187, 1159, 1110,
1075, 1033, 921, 837, 816, 747, 719, 698, 621, 600, 572, 501,
C
3
22
H
4
Cl18‚0.5C
6 6
H : C, 34.70; H, 0.61; Cl, 64.69. Found: C,
-
1
4.61; H, 0.62; Cl, 64.70. After heating at 100 °C under
6
473 cm ; UV-vis (C H12) λmax nm (log ꢀ) 252(4.67), 292(3.07),

Synthesis of New Dendritic Polyalkylaromatic Polyradicals
J . Org. Chem., Vol. 62, No. 26, 1997 9017
3
hexane/CHCl 3:1). Compound 2 eluted first and was recrys-
3
3
02(2.99). Anal. Calcd for C58
6.79. Found: C, 60.07; H, 3.27; Cl, 36.58.
H
34Cl12: C, 60.24; H, 2.96; Cl,
tallized from benzene yielding a clahtrate with benzene in a
1
Tr is(r,r-b is(p en t a ch lor op h en yl)-2,4,5,6-t et r a ch lor o-
tolyl)m eth a n e (8). A solution of compound 11 (0.100 g; 0.086
2:1 proportion (0.245 g, 45%): mp 250 °C dec; H NMR (CD
Cl ) δ 6.92-7.21 (m, 1H), 7.35 (s, 3H), 7.60-7.81 (m, 6H); IR
(KBr): 3044, 1532, 1377, 1272, 1223, 1040, 1005, 787, 722,
2
-
2
mmol) and S
mmol) and the resulting dark red solution was refluxed during
h. Then the solution was concentrated to about one-third of
this initial volume and refluxed an additional 1 h adding,
occasionally, small amounts of SO Cl in order to keep the
volume of the solution constant. Finally, ice was added to the
reaction mixture, and then solid NaHCO was added until no
2 2 2 2
Cl (0.022 g;0.16 mmol) in SO Cl (3.5 mL; 43.2
-
1
790, 582, 442 cm ; UV-vis (C
6
H
12) λmax nm (log ꢀ) 290(3.11),
3
299(3.28), 309(3.27). Anal. Calcd for C25
H
7
6 6
Cl21‚0.5 C H
: C,
30.83; H, 0.92; Cl, 68.25. Found: C, 30.06; H, 0.98; Cl, 68.36.
2
2
Syn th esis of th e P olyr a d ica l Sp ecies. Tetrabutylam-
monium hydroxide (40% in water) was added in the dark to a
suspension of 8 in THF, and the resulting mixture was stirred
at room temperature. After a specific period of time, an excess
of p-chloranil was added and the stirring continued an ad-
ditional 3-4 h. Elimination of the solvent gave a residue
which was passed through silica gel (CCl ) to gave the
4
polyradical sample as a deep-red purple solid which is stable
in contact with the atmosphere and at temperatures up to 150
3
more gas evolution took place. The resulting mixture was
heated on a steam bath, cooled, and finally, acidified with
aqueous HCl. The insoluble material was filtered, washed
with water and dried. By flash chromatography (silica gel,
first hexane, then hexane/chloroform 5:1) and recrystallization
from benzene/methanol, compound 8 was obtained (0.086 gr,
1
4
5%): mp: 225-227 °C; H NMR (CD
2
Cl
2
) δ 6.5-7.2 (m, 4H);
°C. By recrystallization from benzene, polycrystalline samples
IR (KBr): ν 2924, 1532, 1370, 1342, 1244, 1188, 1117, 1040,
were obtained. Elemental analyses of all samples are almost
identical with the starting precursor and agree with the
calculated values.
-
1
8
78, 808, 716, 681, 646, 527, 470 cm ; UV-vis (C
6
H12) λmax
nm (log ꢀ) 223(5.68), 295(3.59), 305(3.51); MS (MALDI-TOF):
-
2
6
4
189 (M ). Anal. Calcd for C58H Cl42: C, 31.81; H, 0.18; Cl,
8.00. Found: C, 32.04; H, 0.21; Cl, 67.81.
Ack n ow led gm en t. This work has been supported
by CICYT (MAT94-0797), DGCYT (PB96-0872-C02-01)
and by the Generalitat de Catalunya (SGR95-0057).
D.R. thanks the “Comissionat per a Universitat i
Recerca de la Generalitat de Catalunya (CIRIT)” for a
fellowship. We thank Amelia J ackson of KRATOS Ltd.,
Manchester, U.K., for recording the mass spectrum.
Tr is(r,r,r′,r′,2,4,6-h ep t a ch lor oxylyl)m et h a n e (2).
A
solution of tris(2,4,6-trichlorophenyl)methane (1) (0.3 g; 0.5
mmol) in chloroform (25 mL; 280 mmol), containing aluminum
trichloride (0.11 g; 0.8 mmol), was refluxed in an argon
atmosphere for 21 h. The resulting blue solution was poured
into cracked ice/hydrochloric acid, and the organic phase was
3
separated and washed with aqueous NaHCO and with water.
After drying over anhydrous sodium sulfate, the solvent was
evaporated and the residue flash chromatographed (silica gel,
J O9622385