Positive dendritic effects on the electron-donating potencies of poly(propylene imine) dendrimers

Article abstract of DOI:10.1021/ol0500564

(Chemical Equation Presented) Two series of poly(propylene imine), PPI, dendrimers terminated with a redox-active donor, 4-dimethylaminobenzyl (4-DMAB), including their respective nondendronized model compounds, are reported. In these two series, a positive dendritic effect was observed for the formation of charge-transfer (CT) complexes between the dendrimers and 7,7,8,8- tetracyanoquinodimethane (TCNQ). However, the nondendronized compounds did not form CT complexes with TCNQ, even though their redox potentials are similar to those of the 4-DMAB units attached to the dendrimers.

Full text of DOI:10.1021/ol0500564

ORGANIC
LETTERS
2005
Vol. 7, No. 7
1287-1290
Positive Dendritic Effects on the
Electron-Donating Potencies of
Poly(propylene imine) Dendrimers
Winston Ong and Robin L. McCarley*
Department of Chemistry, Louisiana State UniVersity, Baton Rouge, Louisiana 70803
tunnel@lsu.edu
Received January 11, 2005
ABSTRACT
Two series of poly(propylene imine), PPI, dendrimers terminated with a redox-active donor, 4-dimethylaminobenzyl (4-DMAB), including their
respective nondendronized model compounds, are reported. In these two series, a positive dendritic effect was observed for the formation of
charge-transfer (CT) complexes between the dendrimers and 7,7,8,8-tetracyanoquinodimethane (TCNQ). However, the nondendronized compounds
did not form CT complexes with TCNQ, even though their redox potentials are similar to those of the 4-DMAB units attached to the dendrimers.
Despite the large sizes and complexities of functionalized
dendrimers,1 interest in these macromolecules continues to
increase because their well-defined structures allow chemists
to precisely tailor them with different property-responsive
groups. For example, unsymmetric-type dendrimers (like
those pioneered by the groups of Freche´t and Newkome2)
provide motifs for core functionalization, while the sym-
metric types such as the commercially available poly(amido
amine) (PAMAM), cyclotriphosphazene (PMMH), and poly-
(propylene imine) (PPI) dendrimers are generally the most
widely used in studies of periphery modification.3
The current literature contains no report that straightfor-
wardly correlates the voltammetric potentials of redox-
responsive dendrimers with the strengths of their charge-
transfer (CT) complexes in bulk solutions. A comprehensive
understanding of this phenomena involving dendrimers needs
to be addressed, because its fundamental and noncovalent
nature will, for example, benefit the development of dendritic
materials that contain platforms based on supramolecular
interactions.4
We commence our research efforts in the area of su-
pramolecular dendrimers by focusing our initial studies on
PPI dendrimers functionalized with multiple redox-respon-
sive groups at their peripheries. Though any substituent can
be covalently attached to these dendrimers, whose native
structures feature multiple tiers of electron-rich tertiary
amines, the desired CT activities of the resulting dendrimers
can be further enhanced if the proper groups are selected.
Specifically, if the oxidation potentials of the added groups
are less anodic than those of the interior tertiary amines of
the PPI dendrimers, the electron-donating potencies of the
resulting dendrimers are expected to be improved. We show
that this is the case when 7,7,8,8-tetracyanoquinodimethane
(TCNQ) is utilized as the acceptor and the end groups of
the PPI dendrimers are modified with the redox-responsive
4-dimethylamino (4-DMAB) moieties.
(1) Fischer, M.; Vo¨gtle, F. Angew. Chem., Int. Ed. 1999, 38, 885.
(2) (a) Hawker, C.; Freche´t, J. M. J. J. Am. Chem. Soc. 1990, 112, 7638.
(b) Newkome, G. R.; Weis, C. D.; Childs, B. J. Des. Monomers Polym.
1998, 1, 3.
(3) For reviews, see: (a) Grayson, S. M.; Freche´t, J. M. J. Chem. ReV.
2001, 101, 3819. (b) Bosman, A. W.; Janssen, H. M.; Meijer, E. W. Chem.
ReV. 1999, 38, 885.
(4) For recent examples, see: (a) Ornatska, M.; Bergman, K. N.; Rybak,
B.; Peleshanko, S.; Tsukruk, V. V. Angew. Chem., Int. Ed. 2004, 43, 5246.
(b) Ooe, M.; Murata, M.; Mizugaki, T.; Ebitani, K.; Kaneda, K. J. Am.
Chem. Soc. 2004, 126, 1604.
10.1021/ol0500564 CCC: $30.25
© 2005 American Chemical Society
Published on Web 03/09/2005

series, respectively, in satisfactory yields (65-87%).6 The
crude reaction solutions were purified with minimal exposure
to ambient light because the title dendrimers and model
compounds are photolabile. In addition to the high symmetry
Scheme 1. Synthesis of the (4-DMAB)- and
Phenyl-Terminated PPI Dendrimers and Their Model
Compounds6
1
expected for the H and 13C NMR resonances (Figures S1
and S2, Supporting Information) of these compounds,
MALDI-TOF MS analyses confirmed that the dendrimers
were completely terminated with 4-DMAB or benzyl units.
Surprisingly, dendrimers 1u and 2u are sparingly soluble in
CH2Cl2 and CHCl3 (<1 × 10-4 M in end group [4-DMAB])
considering that their higher generation homologues 3u-
5u are soluble at similar end group concentration levels in
the same solvents.
Cyclic voltammetry (CV) experiments in CH2Cl2 reveal
that the dendrimers in the urea and amide series, as well as
model compounds 0a and 0u, can be oxidized at accessible
potentials (Figure S3, Supporting Information).6 The model
compound 0a displays a quasi-reversible behavior with an
oxidation peak potential (Epa) of +1.17 V vs Me10Fc0/+, while
the dendritic analogues 1a-5a show an irreversible voltam-
metric behavior, i.e., no voltammetric responses are observed
in the cathodic scans. On the other hand, while the CV of
0u is reversible (Epa ) +0.65 V vs Me10Fc0/+), those of its
dendritic analogues 1u-5u are severely distorted by pre-
cipitation effects upon oxidation. For both (4-DMAB)-
terminated series, the Epa values for the 4-DMAB units
remain constant within a series (Epa(amide) ) +1.09 V and
Epa(urea) ) +0.56 V vs Me10Fc0/+), which means that the
thermodynamic demand for heterogeneous electron transfer
(ET) in these dendrimers is unaffected by dendritic growth.7
For the control series 1c-5c, only weak voltammetric
responses between +0.7 and +1.0 V (vs Me10Fc0/+), arising
from the interior branching tertiary amines, are observed.
This is a clear indication that the heterogeneous ET events
emanating from the inner tertiary amines of the PPI den-
drimers are kinetically too slow relative to both the time scale
of the voltammetric experiments (scan rate ) 0.1 V/s) and
the ET rates observed for the 4-DMAB urea and amide series
(representative CVs in Figure S4, Supporting Information).6
The Epa values for the urea and amide series are noticeably
distinct from each other. The difference of these values, ∆Eox,
arbitrarily defined here as Epa(amide) - Epa(urea), is ap-
proximately +0.53 V in CH2Cl2 (corresponding ∆(∆Gox) of
+12 kcal mol-1), indicating that the urea series is thermo-
dynamically much easier to oxidize than the amide series.
The rationale behind this observation is that the lone electron
pair of the additional N atom in the urea series (the atom
between the urea carbonyl and 4-DMAB) facilitates oxidation
by providing further stability to the oxidized 4-DMAB+ units
through resonance interactions.
Scheme 1 shows the synthetic route for the five PPI
dendrimer generations having either urea (1u-5u) or amide5
(1a-5a) connectivities between the native dendrimer struc-
tures and the 4-DMAB units and for the corresponding model
compounds (0u and 0a, respectively) containing only an
n-heptyl group in place of the dendrimers (representative
structures shown in Figure 1). We also prepared a control
Figure 1. Structures of 5 and TCNQ.
series (1c-5c), composed of dendrimers terminated with
redox-inactive phenyl units, to provide baseline activity
profiles for the two (4-DMAB)-terminated dendrimer series.
Briefly, these molecules were prepared in one step by
condensing the toluene-stripped parent PPI dendrimers with
either 4-(dimethylamino)benzoyl chloride, 4-(dimethylamino)-
phenyl isocyanate, or benzoyl chloride in dry CH2Cl2
overnight to yield the corresponding amide, urea, or control
The aforementioned electrochemical data suggest that, in
purely voltammetric terms, the urea series should be a better
electron donor than the amide series. To gain further insight
to this dissimilarity, or lack thereof, we probed the electron-
donating abilities of 1u-5u and 1a-5a by investigating their
efficiencies in reducing the well-known electron acceptor
(5) Amide dendrimer series has been reported previously (for details,
see: Put, R. J. H.; Clays, K.; Persoons, A.; Biemans, H. A. M.; Luijkx, C.
P. M.; Meijer, E. W. Chem. Phys. Lett. 1996, 260, 136). However, in the
reported approach, the unreacted 4-(dimethylamino)benzoyl residues could
not be removed from 5a. In our protocol,6 we did not encounter this situation.
(6) See Supporting Information for details.
(7) This is commonly observed with other symmetric-type dendrimers
terminated with multiple redox sites. For example, see: Casado, C. M.;
Cuadrado, I.; Mora´n, M.; Alonso, B.; Garc´ıa, B.; Gonza´lez, B.; Losada, J.
Coord. Chem. ReV. 1999, 185-186, 53.
1288
Org. Lett., Vol. 7, No. 7, 2005

TCNQ. This acceptor, which has a reduction potential of
approximately +0.30 V vs Me10Fc0/+ (CH2Cl2/0.20 M
Bu4NPF6) for the formation of its radical anion, is known to
undergo a one-electron reduction process in the presence of
PPI dendrimers to produce the corresponding TCNQ‚-
ammonium noncovalent CT pairs.8
Table 1. Conversion of TCNQ to Its Radical Anion Form at
Different Times After the Addition of the Dendritic (1-5) and
Model (0) Donors
% conversion
250 min 600 min
% conversion
donor
donor
250 min
600 min
Figure 2 shows that the absorption spectrum of TCNQ in
0a
0c
1a
1c
2a
2c
3a
3u
0
0
17
4.1
63
31
0
0
23
6.8
76
40
3c
4a
4u
4c
5a
5u
5c
36
66
96
55
66
92
61
45
77
>99
66
77
99
73
>99
87
>99
72
the dendrimers, especially by the higher-generations belong-
ing to the (4-DMAB)-terminated series. Both model mol-
ecules 0u and 0a, which haVe indiVidual Epa Values only
slightly more anodic than their respectiVe dendrimer homo-
logues (e+0.09 V, ∆(∆Gox) e 2 kcal mol-1), did not reduce
TCNQ at all after 600 min! This is unexpected considering
that, for example, 3u, utilized at the same effective [4-DMAB]
as 0u, completely (>99% conversion) reduced TCNQ in less
than 250 min. Even 1c, a donor that is not as thermodynami-
cally potent as 0u (on the basis of the respective Epa values),
reduced TCNQ to some extent after 600 min (6.8%).
Additional experiments utilizing a 10-fold excess of 0u or
0a (relative to the concentrations described in Figure 3) only
led to reduction of TCNQ by 1% after 1500 min (data not
shown). In all of the experiments in Figure 3, the analytical
concentrations of the dendrimers were adjusted in order to
normalize the effective concentrations of the 4-DMAB
units.12 Thus, any enhancement in the reduction efficiency
of TCNQ by using dendrimers is the result of a positiVe
dendritic effect.13 This effect is surprising from a thermo-
dynamic standpoint, especially for the (4-DMAB)-terminated
Figure 2. Absorption spectra of TCNQ (1.66 × 10-5 M) before
and after the addition of 5a (end group [4-DMAB] ) 1.91 × 10-4
M) and of Bu4N‚TCNQ in degassed CH2Cl2. Inset shows the ESR
spectrum of a TCNQ solution obtained 720 min after the addition
of 5a.
degassed CH2Cl2 displays an absorption band centered at
401 nm. Upon addition of 5a, this band decreases in
magnitude as new bands progressively emerge from 650 to
900 nm, indicating the formation of the TCNQ radical anion.
The ESR spectrum of this solution, taken 720 min after the
addition of 5a, is consistent with the signature resonances
of TCNQ- (aN ) 1.01 G, aH ) 1.42 G).9 The absorption
spectrum of the radical anion TCNQ•-, obtained from its
tetrabutylammonium form (Bu4N‚TCNQ), is also shown for
comparison. The rest of the TCNQ‚dendrimer solutions also
exhibit similar spectral features (ESR and absorption) as that
of the 5a‚TCNQ complex.
Table 1 and Figure 3 describe the loss of TCNQ as it is
being reduced by the dendrimers. The complex nature of
the CT interactions10 may have prevented us from extracting
any kinetic parameters.11 Nonetheless, it is very clear that
the urea series is more efficient than the amide and control
series in reducing TCNQ. For example, among third-
generation dendrimers, 3u quantitatively reduced TCNQ,
while 3a and 3c converted it by only 73 and 36%,
respectively, after 250 min.
Perhaps, the most dramatic observation in Figure 3 and
Table 1 is the amplification of the reduction of TCNQ by
(8) Jansen, J. F. G. A.; de Brabander-van den Berg, E. M. M.; Meijer,
E. W. Science 1994, 265, 1226.
Figure 3. Kinetic profile for the decay of the corrected Abs401nm
value of TCNQ in the presence of the urea, amide, and control
dendrimer series in CH2Cl2. In all cases, the [TCNQ] and end-
group [4-DMAB] (0u, 0a, and urea and amide series) or [phenyl]
(control series) were fixed at 1.66 × 10-5 and 1.90 × 10-4 M,
respectively. The dotted lines at 250 and 600 min represent the
times of the conversion profiles depicted in Table 1. Measurements
for 1u and 2u were omitted due to the limited solubilities of these
dendrimers in CH2Cl2.
(9) Simulation of the TCNQ‚- spin system using these aN and aH values,
g ) 2.003, and line broadening ) 0.19 yielded the best fits to the
experimental spectrum. For the literature aN and aH values of TCNQ-, see:
Jones, M. T.; Hertler, W. R. J. Am. Chem. Soc. 1964, 86, 1881.
(10) Dendrimers in the amide and urea series contain two thermodynami-
cally distinct types of donors (interior PPI amines and 4-DMAB units) that
can compete for CT interactions with TCNQ.
(11) We could not fit the corrected absorption changes into the standard
decay forms. Thus, the profiles in Figure 3 are arbitrarily plotted in the
first-order form for ease of comparison.
Org. Lett., Vol. 7, No. 7, 2005
1289

series, because even though the oxidation potential of the
4-DMAB units in a given series is unaffected by dendritic
growth, their reduction potencies toward TCNQ are affected
(Table 1 and Figure 3). Most likely, the increased local
concentrations of the 4-DMAB residues in dendrimers,
relative to the model compounds, allow the dendronized
4-DMAB units to intramolecularly cooperate in “chelating”
the TCNQ guests. At the concentration levels used in this
work, the most effective CT partners for TCNQ in the urea
and amide series are the third generations 3u and 3a,
respectively. Perhaps, the size of a third-generation PPI
dendrimer offers a geometric dimension such that the stability
of the predominant TCNQ‚4-DMAB CT pairs is maximized.
Dendrimers belonging to the (4-DMAB)-terminated series
contain two electron-donating sites: interior branching
amines and peripheral 4-DMAB units. Thus, the incoming
TCNQ molecules can distribute themselves at these two
competing sites, the degree of which will be governed by
the physical accessibilities and thermodynamic requirements
of the hosting sites. For the urea dendrimers 3u-5u, CT
formation dominates at the peripheral 4-DMAB units of the
dendrimers because of their assumed exterior location
(Scheme 2), less anodic oxidation potentials (thermody-
allows the interior amines of the dendrimers to substantially
compete as electron donors in the overall CT formation
process. Thus, it is not too surprising to realize from Figure
3 and Table 1 that the presence of the relatively dormant
4-DMAB groups in 1a-5a only makes the amide series a
slightly better donor than the control series.14
The residence site of most of the CT pairs in the urea
series is most likely at the periphery of the dendrimers
(Scheme 2). In going from the urea to the amide series, the
Epa values of the 4-DMAB units uniformly shift to more
anodic values by approximately +0.53 V. In accompanying
this thermodynamic transition, the CT sites subsequently and
partially migrate from the peripheries to the interior branch-
ing amines of the dendrimers. Taken together, this transition
will effectively result in the formation of CT pairs wherein
the TCNQ anions are partially encapsulated by the den-
drimers in the amide series.15 Thus, in the case of the amide,
and hence the control series as well, this internal location
can pose as a kinetic barrier in the reduction process because
the diffusing TCNQ molecules must penetrate the outer
layers of the dendrimers before docking at the internal
hosting sites.
The role of the PPI interior amines toward the reduction
of TCNQ is very apparent in the case of the control
dendrimers 1c-5c. This series, which does not contain any
other competing sites for CT pair formation with TCNQ other
than the interior amines of the dendrimers, convincingly
shows that the reduction efficiencies of the PPI dendrimers
steadily increase with increasing generation (from 6.8% for
1c up to 72% for 5c after 600 min).
Scheme 2. Reduction of TCNQ by the (4-DMAB)-Terminated
Urea Dendrimers
To the best of our knowledge, the study presented here is
the first report to straightforwardly correlate the reducing
efficiencies of dendrimers with the oxidation potentials of
their substituents. We are also currently expanding this work
to other solvent systems because the structures of the title
dendrimers contain critical solvent-dependent sites (e.g.,
ureas for mutual H-bonding) that can potentially alter the
course and rate of reduction of TCNQ and other acceptors.
In conclusion, judging from the data in Figure 3 and Table
1, we have shown that the absence of the PPI backbone
muted the reducing activity of 4-DMAB. On the other hand,
the positiVe dendritic effect is reflected in the two series of
(4-DMAB)-terminated dendrimers: the dendrimers in these
two series, which have structurally similar reducing units
yet markedly different Epa values, showed improved reducing
activities toward TCNQ at rates corresponding to the Epa
values of the 4-DMAB units.
namic), and more efficient ET rates (kinetic), as suggested
by the CV data, relative to those of the interior amines. On
the other hand, with all other parameters being similar to
the urea series except for the +12 kcal mol-1 gap of the
individual 4-DMAB units, the reducing efficiencies of 1a-
5a are retarded relative to 3u-5u. In this case, the lack of
ET activity from the 4-DMAB units of the amide series
(12) Even though the effective [4-DMAB] has been constrained, the total
number of donor groups (4-DMAB and interior tertiary amines) still slightly
varies between generations but not to a significant extent. See ref 14 for
details.
(13) For recent examples of positiVe dendritic effects, see: (a) Delort,
E.; Darbre, T.; Reymond, J. L. J. Am. Chem. Soc. 2004, 126, 15642. (b)
Dahan, A.; Portnoy, M. Org. Lett. 2003, 5, 1197.
Acknowledgment. The authors are grateful to the Na-
tional Science Foundation for the generous support of this
work (to R.L.M., CHE-0108961). We also thank Professor
Brian Hales for assistance with the ESR measurements and
Dr. Tracy Donovan McCarley for the MALDI-TOF MS
experiments.
(14) One must realize that the total number of donor sites for the
dendrimers in the (4-DMAB)-terminated series is almost doubled relative
to those in the control series. As an example, the theoretical number of
donor sites in the amide series 1a-5a, ordered as 4-DMAB and interior
tertiary amine units, respectively, is as follows: 1a, 4 and 2; 2a, 8 and 6;
3a, 16 and 14; 4a, 32 and 30; 5a, 64 and 62.
Supporting Information Available: Synthesis of den-
1
drimers 1-5 and model compounds 0, including their H
and 13C NMR spectra, and CV data of 1a-5a, 1u-5u, and
1c. This material is available free of charge via the Internet
at http://pubs.acs.org.
(15) The “dendritic box”,8 a t-Boc-L-Phe-terminated, fifth-generation PPI
dendrimer, encapsulates TCNQ‚- in its dimeric form (see: Bosman, A.
W.; Jansen, J. F. G. A.; Janssen, R. A. J.; Meijer, E. W. Polym. Mater. Sci.
Eng. 1995, 73, 340). In all of our absorption measurements, we did not
2-
observe any spectral features associated with (TCNQ)2 formation.
OL0500564
1290
Org. Lett., Vol. 7, No. 7, 2005