Viologen-based benzylic dendrimers: selective synthesis of 3,5-bis(hydroxymethyl)benzylbromide and conformational analysis of the corresponding viologen dendrimer subunit

Article abstract of DOI:10.1016/j.tetlet.2010.02.097

Convergent and divergent strategies for the synthesis of viologen dendrimers with 1,3,5-tri-methylene-branching units are discussed. The title compound is easily transformed into 1-[3,5-bis(hydroxymethyl)benzyl]-4-(pyridin-4-yl)pyridinium hexafluorophosphate, which is used in sequential growth and activation steps as a CB₂ compound in the cascade-type dendrimer synthesis (B = -OH, activation = -OH → Br). Analysis of the dendrimer structure reveals that three torsional angles, that is, τ₁ between the two pyridinium units, τ₂ between the methylene and pyridinium and τ₃ between the methylene and phenyl, determine the conformational space of the dendrimers. We report here the crystal structure of 1-[3,5-bis(hydroxymethyl)benzyl]-4-(pyridin-4-yl)pyridinium as PF₆^- salt which represents the smallest subunit of the dendrimer that shows the same three torsional angles. The crystal structure together with the results from PM3 calculations opens an avenue to judge the structure of benzylic viologen-based dendrimers.

Full text of DOI:10.1016/j.tetlet.2010.02.097

Tetrahedron Letters 51 (2010) 2188–2192
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Viologen-based benzylic dendrimers: selective synthesis of 3,5-bis(hydroxy-
methyl) benzylbromide and conformational analysis of the corresponding
viologen dendrimer subunit
b
b
Murugavel Kathiresan a, Lorenz Walder a, , Fei Ye , Hans Reuter
*
a Institute of Chemistry, OC II, University of Osnabrueck, Barbarastrasse 7, D-49069 Osnabrueck, Germany
b Institute of Chemistry, AC II, University of Osnabrueck, Barbarastrasse 7, D-49069 Osnabrueck, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Convergent and divergent strategies for the synthesis of viologen dendrimers with 1,3,5-tri-methylene-
branching units are discussed. The title compound is easily transformed into 1-[3,5-bis(hydroxy-
methyl)benzyl]-4-(pyridin-4-yl)pyridinium hexafluorophosphate, which is used in sequential growth
and activation steps as a CB2 compound in the cascade-type dendrimer synthesis (B = –OH, activa-
tion = –OH ? Br). Analysis of the dendrimer structure reveals that three torsional angles, that is, s1
Received 26 January 2010
Revised 15 February 2010
Accepted 17 February 2010
Available online 21 February 2010
between the two pyridinium units, s2 between the methylene and pyridinium and s3 between the meth-
Keywords:
Dendrimer
Viologen
ylene and phenyl, determine the conformational space of the dendrimers. We report here the crystal
structure of 1-[3,5-bis(hydroxymethyl)benzyl]-4-(pyridin-4-yl)pyridinium as PF6ꢀ salt which represents
the smallest subunit of the dendrimer that shows the same three torsional angles. The crystal structure
together with the results from PM3 calculations opens an avenue to judge the structure of benzylic viol-
ogen-based dendrimers.
Synthesis
Benzylic bromides
PM3 simulation
X-ray
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
Balzani and co-workers4 (Scheme 1). The divergent approach of
the viologen dendrimers3 emanates from the 1,3,5-tris(bromo-
In the last three decades, a number of methods have been re-
ported concerning the cascade synthesis of dendrimers with trifold
core using different branching units. These syntheses require a tri-
functional CB2- or BC2-type building block as the branching units
with B representing a latent functionality and C representing a
reactive functionality which reacts with A, where A is available
by the activation of B. Depending on the synthesis (divergent or
convergent) either CB2- or BC2-branching units are required. A rep-
resentative example for the divergent approach is the synthesis of
the propyl amine dendrimer from acrylonitrile (CB2-type branch-
ing unit) with –C@C– representing C and the nitrile function repre-
senting B2 which can be activated to A2, A representing an amine
nitrogen which can undergo double alkylation.1 A representative
example for the convergent approach is the synthesis of the poly-
(benzyl ether) dendrimer from 3,5-dihydroxybenzyl alcohol (C2B-
type branching unit) with B representing the benzylic alcohol that
can be transformed into the corresponding bromide (A) and the
phenolic OH groups representing C2.2 A decade ago, we have intro-
duced the synthesis of viologen dendrimers3 via a divergent ap-
proach using CB2-type branching units, the corresponding
convergent approach using BC2-branching units was reported by
methyl)benzene core and the step-wise use of CB2-branching units,
where C represents the reactive pyridyl nitrogen and B2 represents
two OH groups which can be transformed into the benzylic bro-
mide (A). The corresponding BC2 dicationic compound is a poten-
tial branching unit for the convergent approach.
The selective monoalkylation of 4,40-bipyridine with 1,3,5-
tris(bromomethyl)benzene is possible but its further transformation
to the corresponding bis(hydroxymethyl)-4-(pyridine-4yl)pyridini-
um salt cannot be achieved because under the typical basic condition
necessary for this reaction, the bipyridinium is irreversibly attacked
by OHꢀ. Selective double substitution of 1,3,5-tris(bromo-
methyl)benzene by two bipyridines is not possible as the product is
prone to polymerize. Thus both ‘viologen’-branching units CB2 and
BC2 require the specific synthesis of the corresponding benzylic bro-
mides 1c and 1d, respectively. The procedures have been reported
but the yields are not satisfactory.5,6
2. Results and discussion
Possible routes for the synthesis of 1c are presented in Scheme
2. These include (I) the hydrolysis approach using 1,3,5-tris(bro-
momethyl)benzene 1e, (II) the Appel method, (III) the Frechet-type
synthesis of 2d followed by reduction and (IV) the bromination of
diethyl 5-methylisophthalate.
* Corresponding author. Tel.: +49 451 9692495; fax: +49 541 9693308.
E-mail address: lowalder@uos.de (L. Walder).
0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2010.02.097

M. Kathiresan et al. / Tetrahedron Letters 51 (2010) 2188–2192
2189
Scheme 1. Divergent and convergent strategies for benzylic viologen dendrimers.
(I) Hydrolysis of 1,3,5-tris(bromomethyl)benzene using different
equivalents of strong or weak bases was not successful as the reac-
tion yielded either mixture of ethers along with hydroxyl derivatives
or completely hydrolyzed products, as reported for other benzylic
bromide derivatives using different basic conditions.7–9 (II) Since
our first report on viologen dendrimer synthesis,3 we use the Appel
method to synthesize 1c and 1d. The synthesis starts with the con-
version of trimesic acid to trimethyl 1a or triethyl ester 2a (90% in
both cases), followed by the reduction of the triester 1a with LiAlH4
to yield the triol (1b) in 76% yield. The triol is then brominated using
1 equiv of Appel reagent to yield a mixture of the products 1c, 1d and
1e which must be separated by tedious column chromatography [1c
(36%) and 1d (12%)]. There are two other methods tailored for higher
yields of 1d10 (a well-known linker possessing two reactive and a la-
tent site) by Stoddart and co-workers6 and by Diez-Barra et al.5
where the former used 3 equiv of the Appel reagent in THF and the
latter used 1.1 equiv of Appel reagent in CH3CN with 43% and 52%
yields, respectively, after column chromatography. No reports are
available on Appel conditions favouring high yields of 1c. The limita-
tions of Appel method are obvious: the reaction is not specific and
the necessity of separation using column chromatography which
limits the work-up to few hundred milligram scale. In order to over-
come these problems and remembering that DIBAL-H is known to
selectively reduce the ester functionality in the presence of primary
or benzylic halides,11,12 we propose route IV involving 5-bromo-
methyl diethylisophthalate13 as the key intermediate. Its reduction
with 1 M DIBAL-H in DCM gave 1c in ca. 70% without the need of
any column separation. We carried out the same reduction using
LiAlH4 at 0 °C as it is reported that LiAlH4 selectively reduces ester
functionality at lower temperature without reducing alkyl
halides,14–16 however, this reaction does not give reproducible re-
sults. Path III follows a route described earlier by Frechet13 (2a–d),
we followed the same route with some modifications. The method
is based on the selective hydrolysis of 1,3,5-triethyltrimesic ester
2a to 5-carboxy diethylisophthalate 2b in 75% yield. The monocarb-
oxydiesteris thenreduced to 5-hydroxymethyl-diethylisophthalate
2c in 90% yield using 1 M BH3–THF complex (lit.13 78% using 1 M
BH3–(CH3)2S). Bromination of the hydroxyl precursor using 5.6 M
HBr in HOAc gave 5-(bromomethyl)diethylisophthalate 2d in 95%
(lit.13 90%, PBr3 as brominating agent) which is then reduced to
3,5-bis(hydroxymethyl)benzylbromide 1c 70% using 1 M DIBAL-H
in DCM following the reported procedure.17 Notably, all these steps

2190
M. Kathiresan et al. / Tetrahedron Letters 51 (2010) 2188–2192
Scheme 2. Synthesis of hydroxymethyl- and bis(hydroxymethyl)benzylic bromides—the precursors of CB2- and BC2 viologen-branching units. Reagents and conditions: (a)
hydrolysis under different basic conditions; (b) ROH, H2SO4, reflux, 24 h; (c) LAH, THF, reflux, 24 h; (d) CBr4/PPh3, THF, 0 °C to rt, 3 h; (e) KOH, EtOH–THF, reflux, 24 h; (f) 1 M
BH3–THF, 0 °C to rt, 6 h; (g) 5.6 M HBr–HOAc, rt, 36 h; (h) 1 M DIBAL-H/DCM, 0 °C to rt, 6 h; (i) NBS, Bz2O2, CCl4, reflux; (j) (1) 4,40-bipyridine, CH3CN, 80 °C, 24 h; (2) 3 M
NH4PF6/H2O.
Figure 1. (a) X-ray structure of 3a with torsions
X-ray conformation); (c) section of dendrimer consisting of 3a subunits.
s1 (37.8° (C4–C3–C6–C10)), s2 (87.3° (C8–N2–C11–C12)) and s3 (ꢀ80.8° (N2–C11–C12–C13)); (b) dendrimer subunit (based on 3a

M. Kathiresan et al. / Tetrahedron Letters 51 (2010) 2188–2192
2191
ꢀ
lead to the desired intermediates in good yield using simple recrys-
tallization steps without column purifications. (For detailed syn-
thetic procedures, see Supplementary data). (IV) The intermediate
2d canalso be synthesizedby brominating 5-methyldimethylisophth-
alate using NBS as shown in route IV but this reaction proceeds with
low yield because of the electron-withdrawing effects of the two ester
groups.18
delivering reasonable values. If the PF6 counter ion is omitted in
the same calculation, the lowest energy torsional angles s1–3 do
not change, indicating that the counter ion is not governing the tor-
sional angles, rather the organic structure governs the position of
PF6. The torsional energy barriers related to a 360° rotation around
s1
,s
2 and
at 10° increments without further geometry optimization (see Sup-
plementary data). For a 360° rotation around 1 (with the 2 and s3
s3 have been judged from single point energy calculations
The mono- or di-alkylation of 4,40-bipyridine is highly influ-
enced by the solubility of the products in the solvent media. Dial-
kylation can be prevented by choosing a solvent in which the
starting materials are soluble and the monoalkylated product is
insoluble. In our case, we suppress the double alkylation by adding
1 equiv of 1c slowly to 5 equiv excess of 4,40-bipyridine in CH3CN.
The precipitated product is then filtered, washed with CH3CN, dis-
solved in water and precipitated as hexafluorophosphate salt. The
pale yellow powder thus obtained is again dissolved in water,
heated to 80 °C and then cooled to yield pale yellow crystals.
(Note: the pale yellow crystals became light green upon exposure
to air). 3b is synthesized using the same procedure, that is, 1 equiv
of 1d is added slowly to 10 equiv of 4,40-bipyridine in CH3CN and
the resulting product is precipitated as hexafluorophosphate. The
detailed synthetic procedures, CV and UV–vis characterizations
are given in the Supplementary data.
s
s
values fixed at their X-ray values) a fourfold barrier with a height
below 1 kcal/mol was found, with one of the minima identical to
the X-ray value (37.9°). PM3-based Wiberg atom–atom bond order
calculations34 reveal a bond order of only 1.02 between C3 and C6
for the minimum conformation in agreement with the observed
low-energy torsional profile.
Rotation along s2 (with the s1 and s3 values fixed at their X-ray
values) shows again a low-energy barrier (2–3 kcal/mol) but two-
fold, as expected for the methylene–pyridinium interaction. Finally
for the rotation around s3 showed a low-energy barrier but twice
as that of s2 (5–6 kcal/mol) and twofold as expected for the meth-
ylene–phenyl interaction.
1H NMR spectra of the viologen dendrimers with the 3a subunit
show high symmetry indicating fast conformational changes.3 Fur-
thermore, cyclic voltammetry studies on 3a and 3b show slow fast
electron transfer rates, which is again typical for redox couples
with low activation barriers.
In summary, the conformational analysis of 3a, based on X-ray,
cyclic voltammetry and NMR data, representing a structural sub-
unit of benzylic viologen dendrimers, reveals high flexibility with
respect to the three torsional angles that play a role for the dendri-
mers shape.
3. X-ray crystallography, cyclic voltammetry and modelling
Viologen dendrimers3,4,19 and dendrons20 have been prepared
extensively from 4,40-bipyridine and 3a using sequential substitu-
tion and activation reactions discussed in the prior paragraph. The
geometry of the resulting dendritic structure has a large impact on
the pimerization of viologen subunits,20 the diffusion coefficient of
the dendrimer, which in turn depends on the hydrodynamic ra-
dius,3,19 and the size of the internal voids which is responsible
for the pickup of large counter ions.4,21
4. Conclusions
The synthesis of viologen dendrimers with 1,3,5-trimethylene-
branching units requires facile access to the CB2-type synthon,
1-[3,5-bis(hydroxymethyl)benzyl]-4-(pyridin-4-yl)pyridinium
hexafluorophosphate (Scheme 1). It is available according to
Scheme 2 from 1c and 4,40-bipyridine. So far the synthesis of 1c fol-
lowed route II (Scheme 2) involving tedious chromatographic sep-
aration. The synthetic route III given in Scheme 2, explored in this
work, allows the production of 1c without the need of chromatog-
raphy. The X-ray analysis of 1-[3,5-bis(hydroxymethyl)benzyl]-4-
(pyridin-4-yl)pyridinium hexafluorophosphate combined with
PM3-modelling studies gives a first time access to the sound esti-
mate of the viologen dendrimer conformation.
A closer look at the dendrimer structure III shown in Figure 1
reveals that the overall conformational space available for the
branches is given by only three torsional angles, that is, s1
and 3, assuming that all the bending angles do not deviate much
from their equilibrium value. The three torsional angles are located
between the two pyridine moieties ( 1), the methylene H and the
pyridine ( 2) and between the same methylene and the phenyl
group ( 3) (II in Fig. 1). Obviously, the salt 3a (I in Fig. 1) exhibits
, s2
s
s
s
s
the same set of angles. We were able to grow crystals of 3a and
to resolve its structure by X-ray analysis22 (Fig. 1 I). The X-ray
structure reveals that there are two molecules in a triclinic unit cell
and the existence of an intermolecular hydrogen bond between
O(1A)–H(1A) and N(1) (see Supplementary data). The torsional an-
gles found in 3a can be used as a reasonable starting point for the
discussion of the dendrimer conformation. The torsional angles s1
in other bipyridinium systems cover the range from 20 to 50° in
case of mono- and dialkylated viologens6,23–29 with no large influ-
ence of the counter ion. However, the oxidation state is of impor-
Acknowledgements
M.K. thanks the Graduate College 612 funded by DFG for finan-
cial support. The discussion with D. H. Taffa on the interpretation
of the electrochemical and UV–vis data is greatly appreciated.
tance, thus a diphenyl viologen shows dihedral angle (s1) of 37°
Supplementary data
and 1° for the dicationic and the radical cationic state, respectively.
The other two torsions s2 and s3 have been reported for dibenzyl
viologen by Inoue et al.30 and by Garcia et al.31 The former found
three different s3 and s2 within a single crystallographic cell, the
Supplementary data (detailed experimental procedures, 1H and
13C NMR spectra of 1c, 3a and 3b, cyclic voltammograms, UV–vis
spectra of 3a and 3b and PM3 calculations) associated with this
article can be found, in the online version, at doi:10.1016/
j.tetlet.2010.02.097.
latter found
torsional angle preference, that is, a low torsional energy profile
for s1 s2 and s3 planes.
s3 = -88.2° and s2 = 168.8°. These findings indicate no
,
Semi-empirical PM3 calculations were performed using Argu-
slab 4.0.132 and Hyperchem 8.0.6.33 When the X-ray structure is
used as a starting point for geometry optimization, a local mini-
mum is found with only minor deviation from the solid state struc-
ture (except for a lateral shift of PF6ꢀ), indicating that PM3 is
References and notes
1. Buhleier, E.; Wehner, W.; Voegtle, F. Synthesis 1978, 155.
2. Hawker, C. J.; Frechet, J. M. J. J. Am. Chem. Soc. 1990, 112, 7638.
3. Heinen, S.; Walder, L. Angew. Chem., Int. Ed. 2000, 39, 806.

2192
M. Kathiresan et al. / Tetrahedron Letters 51 (2010) 2188–2192
4. Marchioni, F.; Venturi, M.; Credi, A.; Balzani, V.; Belohradsky, M.; Elizarov, A.
M.; Tseng, H. R.; Stoddart, J. F. J. Am. Chem. Soc. 2004, 126, 568.
5. Diez-Barra, E.; Garcia-Martinez, J. C.; Merino, S.; del Rey, R.; Rodriguez-Lopez,
J.; Sanchez-Verdu, P.; Tejeda, J. J. Org. Chem. 2001, 66, 5664.
6. Menzer, S.; White, A. J. P.; Williams, D. J.; Belohradsky, M.; Hamers, C.; Raymo,
F. M.; Shipway, A. N.; Stoddart, J. F. Macromolecules 1998, 31, 295.
7. Sivakumar, C.; Nasar, A. S. J. Polym. Sci., Part A: Polym. Chem. 2009, 47, 3337.
8. Tran, H.-A.; Collins, J.; Georghiou, P. E. New J. Chem. 2008, 32, 1175.
9. Shimizu, S.; Suzuki, T.; Shirakawa, S.; Sasaki, Y.; Hirai, C. Adv. Synth. Catal. 2002,
344, 370.
10. Arendt, M.; Sun, W.; Thomann, J.; Xie, X.; Schrader, T. Chem. Asian J. 2006, 1,
544.
11. Esser, B.; Bandyopadhyay, A.; Rominger, F.; Gleiter, R. Chem. Eur. J. 2009, 15,
3368.
12. Allard, E.; Delaunay, J.; Cousseau, J. Org. Lett. 2003, 5, 2239.
13. Leon, J. W.; Kawa, M.; Frechet, J. M. J. J. Am. Chem. Soc. 1996, 118, 8847.
14. Komissarov, V. V.; Panova, N. G.; Kritzyn, A. M. Russ. J. Bioorg. Chem. 2008, 34,
67.
21. Schon, P.; Degefa, T. H.; Asaftei, S.; Meyer, W.; Walder, L. J. Am. Chem. Soc. 2005,
127, 11486.
22. CCDC-756509 (for 3a) contains the Supplementary data for this Letter. These
data can be obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
23. Dickson, S. J.; Wallace, E. V. B.; Swinburne, A. N.; Paterson, M. J.; Lloyd, G. O.;
Beeby, A.; Belcher, W. J.; Steed, J. W. New J. Chem. 2008, 32, 786.
24. Li, S. J.; Liu, M.; Zheng, B.; Zhu, K. L.; Wang, F.; Li, N.; Zhao, X. L.; Huang, F. H.
Org. Lett. 2009, 11, 3350.
25. Porter, W. W.; Vaid, T. P. J. Org. Chem. 2005, 70, 5028.
26. Scheytza, H.; Rademacher, O.; Reissig, H. U. Eur. J. Org. Chem. 1999, 2373.
27. Soleimannejad, J.; Aghabozorg, H.; Hooshmand, S. Acta Crystallogr. Sect.
E. -Struct. Rep. Online 2008, 64, M564.
28. Spruell, J. M.; Paxton, W. F.; Olsen, J. C.; Benitez, D.; Tkatchouk, E.; Stern, C. L.;
Trabolsi, A.; Friedman, D. C.; Goddard, W. A.; Stoddart, J. F. J. Am. Chem. Soc.
2009, 131, 11571.
29. Zhao, L.; Dong, Y. R.; Xie, H. Z. Acta Crystallogr., Sect. E.-Struct. Rep. Online 2009,
65, M450.
15. Drescher, S.; Meister, A.; Blume, A.; Karlsson, G.; Almgren, M.; Dobner, B. Chem.
Eur. J. 2007, 13, 5300.
16. Snider, B. B.; Lu, Q. J. Org. Chem. 1996, 61, 2839.
30. Inoue, M. B.; Inoue, M.; Machi, L.; Brown, F.; Fernando, Q. Inorg. Chim. Acta
1995, 230, 145.
31. Garcia, M. D.; Blanco, V. c.; Platas-Iglesias, C.; Peinador, C.; Quintela, J. M. Cryst.
´
17. Cooke, G.; Woisel, P.; Bria, M.; Delattre, F.; Garety, J. F.; Hewage, S. G.; Rabani,
G.; Rosair, G. M. Org. Lett. 2006, 8, 1423.
18. Liu, P. N.; Chen, Y. C.; Deng, J. G.; Tu, Y. Q. Synthesis 2001, 2078.
19. Heinen, S.; Meyer, W.; Walder, L. J. Electroanal. Chem. 2001, 498, 34.
20. Felderhoff, M.; Heinen, S.; Mulisho, N.; Webersinn, S.; Walder, L. Helv. Chim.
Acta 2000, 83, 181.
Growth Des. 2009, 9, 5009.
32. ArgusLab 4.0.1 Mark A. Thompson Planaria Software LLC‘, Seattle‘, WA http://
www.arguslab.com.
33. HyperChem(TM)‘, Hypercube‘, Inc.‘, 1115 NW 4th Street‘, Gainesville‘, Florida
32601‘, USA.
34. Stewart, J. J. P. J. Comput. Chem. 1989, 10, 221.