The synthesis and characterization of a new family of polyamide dendrimers

Article abstract of DOI:10.1002/(SICI)1521-3765(19980515)4:5<781::AID-CHEM781>3.0.CO;2-3

A new family of dendrimers has been constructed with 5-hydroxy-isophthalic acid and diethanolamine as the sources of the branching units. The design of a second-generation building block in the form of an orthogonally-protected aminotetraacid (a single diethylphosphoramide-protected amine and four methyl-ester-protected carboxylic acid groups) has been established by a series of logical developments involving the synthesis of various dendrimer prototypes. This building block has been utilized in both convergent and divergent methods in the synthesis of monodisperse dendrimers up to the fourth generation and subsequently polydisperse sixth-generation dendrimers at the level of a mixture. One-, two-, and three-directional dendrimers of the second and fourth generation were synthesized in order to carry out detailed comparisons by GPC and by ¹H NMR spectroscopy. The GPC results suggest that the fourth-generation dendrimers adopt globular shapes, while the variable-temperature NMR spectroscopic investigations demonstrate that the periphery of the fourth-generation dendrimer is less mobile than either of its internal regions or the analogous portions of the second-generation dendrimer. Molecular modeling suggests highly globular shapes for the larger dendrimers and gives values for molecular radii in very close agreement with those obtained from analysis of the GPC results.

Full text of DOI:10.1002/(SICI)1521-3765(19980515)4:5<781::AID-CHEM781>3.0.CO;2-3

FULL PAPER
The Synthesis and Characterization of a New Family of
Polyamide Dendrimers
Peter R. Ashton, Derek W. Anderson, Christopher L. Brown, Andrew N. Shipway,
J. Fraser Stoddart,* and Malcolm S. Tolley
Abstract: A new family of dendrimers
has been constructed with 5-hydroxy-
isophthalic acid and diethanolamine as
the sources of the branching units. The
design of a second-generation building
block in the form of an orthogonally-
protected aminotetraacid (a single di-
ethylphosphoramide-protected amine
and four methyl-ester-protected carbox-
ylic acid groups) has been established by
a series of logical developments involv-
ing the synthesis of various dendrimer
prototypes. This building block has been
utilized in both convergent and diver-
gent methods in the synthesis of mono-
disperse dendrimers up to the fourth
generation and subsequently polydis-
perse sixth-generation dendrimers at
the level of a mixture. One-, two-, and
three-directional dendrimers of the sec-
ond and fourth generation were synthe-
sized in order to carry out detailed
comparisons by GPC and by H NMR
spectroscopy. The GPC results suggest
that the fourth-generation dendrimers
adopt globular shapes, while the varia-
ble-temperature NMR spectroscopic in-
vestigations demonstrate that the pe-
riphery of the fourth-generation den-
drimer is less mobile than either of its
internal regions or the analogous por-
tions of the second-generation dendrim-
er. Molecular modeling suggests highly
globular shapes for the larger dendrim-
ers and gives values for molecular radii
in very close agreement with those
obtained from analysis of the GPC
results.
1
Keywords: amino
polyacids
´
branched monomers ´ copolymer-
izations ´ dendrimers ´ polymers
shapes of these cavities and their containment characteristics
are expected[8] to confer useful properties. Although den-
drimers have been evaluated by many different methods,
including electron microscopy,[9] ESR spectrometry,[10] 13C
nuclear relaxation measurements[11] and solvatochromic stud-
ies,[12] the direct observation and assessment of dendritic
topology is difficult. Computer modeling studies[13] have
probably provided the best graphic displays of dendrimers
so far. In our own recent research, reported in this paper, we
have used the technique of variable-temperature NMR
spectroscopy to study the dynamic processes within the
dendritic structures and to establish the relationship between
these processes and the size of the dendrimer. Further, we
have related our findings to results obtained from both gel
permeation chromatography and molecular modeling studies.
The construction of dendrimers has been dominated by two
antithetical philosophies, the convergent[14] and the diver-
gent[15] ones. While the divergent approach involves building
the dendrimer outwards starting from a core, the convergent
approach begins with what will ultimately be the periphery of
the dendrimer, and entails building inwards to the core. Both
methods require two steps for the growth of each generation:
the activation of the growing dendritic fragment, and the
addition of new monomer units; both methods have their
Introduction
Recently, dendrimers have attracted a lot of attention from
researchers[1] for a variety of reasons. Some dendrimers are
being produced commercially[2] in multikilogram quantities,
while others are showing promise in fields as diverse as
enantioselective catalyses[3] and drug delivery systems.[4] Their
potential applications[5] rely, to a large extent, on their unique
topologies[6] and on the properties[7] dictated by their highly
branched structures. Medium- to large-sized dendrimers
adopt globular conformations in which sizable cavities are
located beneath their tightly packed surfaces. The specific
[*] Prof J. F. Stoddart,[‡] P. R. Ashton, Dr C. L. Brown, Dr A. N. Shipway,
M. S. Tolley
School of Chemistry, The University of Birmingham
Edgbaston, Birmingham B15 2TT (UK)
Dr D. W. Anderson
Chemical and Biological Defence Establishment
Porton Down, Wiltshire SP4 0JQ (UK)
‡
[ ] Current address: Department of Chemistry and Biochemistry
University of California at Los Angeles
405 Hilgard Avenue, Los Angeles, CA90095 (USA)
Fax : (‡ 1)310-206-1843
E-mail: stoddart@chem.ucla.edu
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0781 $ 17.50+.25/0
781

J. F. Stoddart et al.
FULL PAPER
strengths and weaknesses. The possibly more widely used
method, the divergent approach, has been dogged by poly-
dispersity,[16] which arises from incomplete reaction and
purification problems. While polydispersity does not consti-
tute a significant setback in a materials context for applica-
tions, it is associated with inseparable mixtures of macro-
molecules that become almost impossible to modify chemi-
Results and Discussion
Much of any dendritic structure is composed of branching
moieties, and so considerable care has to be taken in their
choice. Branching is usually[26] incorporated into the mono-
mer unit, although it can be achieved[27] during the dendrimer
growth step from a nonbranched monomer. We made an
early decision to choose branching monomers which are
linked by a nonbranching reaction, as this gave us with a
wider choice of reactions for joining the units. There has been
a great deal of success[28] in the synthesis of application-ori-
ented dendrimers designed around specific branching units.
For our purposes, we were interested ultimately in the
dendrimer functioning as a host molecule, and so we chose
branching units which might provide sites for noncovalent
bonding interactions. We identified two starting materialsÐ
namely, diethanolamine (1) and 5-hydroxyisophthalic acid (2,
Figure 1)Ðwhich could be used to introduce branching into
cally in
a highly controlled manner. The alternative
convergent approach is preferred for the preparation of
monodisperse samples, and has been utilized[17] in the
construction of dendrimers that have subsequently been
subjected to precisely controlled chemical modifications.
Convergent syntheses can, however, suffer[18] from low yields
that are often attributed to steric hindrance at the focal points
of large wedges. In addition, they are less easily scaled up,
since even a quantitative yield in a generation-adding step
adds very little to the sample mass, whereas a divergently-
added generation more than doubles the mass of a dendrimer.
While the convergent and divergent approaches have been
used extensively[19] to incorporate a wide variety of structural
entities into dendrimers, the problem remains that each
synthesis is only specific to one particular dendrimer. Better
methodologies for the accelerated construction of constitu-
tionally pure dendrimers have been developed in response to
the need for more time-efficient syntheses. Although method-
ologies such as the use of hypercores,[20] branched mono-
mers,[21] and a two-step approach[22] have appeared, they offer
limited opportunities with respect to flexibility of constitu-
tional control. What is required is a more general synthetic
strategy that affords maximum flexibility in relation to the
constitution, without leading to an increase in the number of
synthetic steps.
Moore[23] has described a double-exponential strategy in
which the construction of dendrimers is actually more
convergent in a synthetic sense. In this strategy, orthogonal
protecting groups on either end of the monomer are removed
selectively, and the two different products that result are
reacted subsequently with each other to produce a new, larger
branching unit. This approach gives access to macromolecular
structures in just a few steps and, after each growth step, the
dendritic product is an analogue of the original monomer.
Any dendritic system relies essentially[24] on just one
reaction for the linking of its branching units. This fact not
only emphasizes just how important the choice of that
reaction is, but also has serious implications if one is interested
in subtle engineering[25] of the dendritic architecture.
OH
CO2H
HO
HN
1
2
CO2H
OH
Figure 1. Branched starting materials for the construction of a dendritic
building block.
the structure. The units incorporated possible hydrogen-
bonding sites together with a p-electron aromatic system. By
creating an oligomer composed of two units, we anticipate the
emergence of a method with several advantages over more
traditional ones (which involve only one branching moiety).
Firstly, the basic building block would be a branched mono-
mer[29] of the second generation, that is, it branches twice, and
would add two generations to a growing dendritic fragment.
Very fast growth should be achievable, thus countering one of
the much-cited[30] weaknesses of many dendrimer syntheses.
Secondly, we would create a dendrimer with an interesting
topology. Internal functionality would be more diverse than is
the case in many dendrimers, and the dendrimer would consist
of alternating shells, a structure described[31] as a dendritic
layer± block copolymer. Following the synthesis of an appro-
priate dendritic building block, the linking of these units to
create larger structures then had to be addressedÐand was.
We present here the development of a building block for
the construction of dendrimers which offers a high degree of
versatility. By means of well-known protection/deprotection
reactions and an amide-bond-forming growth reaction, this
unit is capable of growth in either the convergent or the
divergent direction, and can thus be considered a double
exponential building block. The basic building block is also of
interest due to its doubly branched structure. It is a second-
generation monomer as its reaction with a dendritic fragment
increases the number of generations by two. Furthermore, the
two generations are of entirely different constitutions, leading
to a final dendrimer with ªalternating shellº topology and
diverse functionality.
Synthesis of a double-branched unit: The building block 6 was
synthesized in four steps from diethanolamine and 5-hydroxy-
isophthalic acid (Scheme 1). This unit has a reactive func-
tionality (an amine) at its focal point and relatively passive
benzyl alcohols on its surface. The synthesis of the precursor 5
was achieved readily, requiring no column chromatography;
these reactions were performed easily on a 100 mmol scale.
The Williamson ether coupling of 3 with 4 was found to give
only a 14% yield of compound 5 when two molar equivalents
of the phenol were used. This yield was raised to 86% by the
use of four molar equivalents of 4, the excess of which could
be recovered easily. Since the reduction of 5 with lithium
aluminum hydride in THF was unsuccessful, owing to
782
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0782 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
OH
OTs
Ts N
OTs
i) (EtO)2P(O)H / CCl4
K2CO3 / KHCO3
Bu4NI / THF
TsCl
C5H5N
OH
HN
OH
OTs
O
HN
CH2Cl2
EtO P N
EtO
ii) TsCl
CH2Cl2 / Et3N
K2CO3
MeCN
OH
1
K2CO3
18-crown-6
85%
OTs
3
42%
1
7
∆
EtOH
H2SO4
MeCN
∆
CO2H
CO2H
CO2Et
CO2Et
i) MeOH / H2SO4
ii) LiAlH4 / THF
CO2H
OAc
OAc
HO
HO
∆
HO
AcO
iii)
Ac2O / C5H5N
100%
86%
CO2H
2
4
92%
23%
2
10
CO2Et
OH
OAc
OAc
O
O
Red-Al
Dioxane
O
O
CO2Et
CO2Et
OH
OH
HCl
O
Ts N
HN
OAc
OAc
OAc
OAc
∆
EtO P N
EtO
HN
THF
O
O
O
O
CO2Et
OH
14%
6
5
OAc
OAc
31%
12
11
Scheme 1. Synthesis of the double-branching unit 6 from diethanolamine
Scheme 2. Synthesis of the improved double-branching unit 12.
and 5-hydroxyisophthalic acid.
to mask the four hydroxyl groups with protecting groups that
could be readily removed when the free alcohols were
required. Scheme 2 shows the successful route to the acyl-
protected monomer 12.
solubility difficulties and the low reactivity of the sulfona-
mide, the reducing agent was changed to sodium bis(2-
methoxyethoxy)aluminum hydride[32] (Red-Al), and the sol-
vent to 1,4-dioxane. While these conditions were found to
cleave the phenyl ethers, reaction with the tosyl protecting
group was even faster, allowing isolation of the branching unit
6 in 14% yield when the reaction was quenched after
20 hours. Purification of the highly polar product (6) from
the reaction mixture proved extremely difficult and an
analytically pure sample was not isolated. It seems that the
low yield can be attributed in part to the technical problems
associated with isolating the product from the crude reaction
mixture. Alternative deprotections[33] of the tosyl protecting
group were found to be unsatisfactory.
The low yield obtained in the synthesis of 6 served to cast
serious doubts over its use in the construction of dendrimers.
Its preparation, which requires column chromatography and
suffers from a low yield, would be a serious handicap even
before dendrimer construction began. In addition, 6 is only
sparingly soluble in aprotic solvents, leading to difficulty in its
use and similar properties in the dendrimers we could
construct from it. We have shown, however, that a diethanol-
amine/5-hydroxyisophthalic acid-based dendritic branching
unit can be prepared, and we hope that small design changes
in this prototype might permit the development of a more
feasible system for the production of dendrimers on a useful
scale.
The choice of the amino protecting group turned out to be
particularly crucial with respect to decomposition[34] and
elimination reactions in the ether-forming step. The diethyl-
phosphoryl protecting group, used[35] in the construction of
aza-crown ethers, was chosen after an investigation of more
traditional candidates. The phosphoramide was formed under
basic conditions and was removed with anhydrous hydrogen
chloride. Protection of diethanolamine[36] (1) using diethyl-
phosphite followed a literature method, and the crude
product was tosylated to afford 7 in a yield of 42% for the
two steps. The triacetate 10 was obtained from 5-hydroxy-
isophthalic acid (2), via 8 and 9 in an overall yield of 92% over
three steps. The reaction of 10 with 7 to give the double-
branched unit 11 occurred in only a 23% yield. Although it is
known[37] that that carbonate ion in refluxing THF or MeCN
can deprotect acylated phenols, generating a phenoxide ion,
this reaction is slow, and by-products from the desired ether
synthesis are formed as a result of the elimination of TsOH
and the nucleophilic attack[38] of acetate on 7. Deprotection of
the diethylphosphite-protected branching unit 11 occurred in
a 31% yield to give the new branching unit 12 as a thick, color-
less oil. By-products from the deprotection can be attributed
to cleavage of the phenyl ethers, and so the reaction must be
followed carefully in order to maximize the yield.
The synthesis of a second-generation dendrimer was
performed as a demonstration of the use of the new branching
unit 12 in dendrimer synthesis (Scheme 3). Benzene 1,3,5-
tricarbonyl trichloride was utilized as a three-directional core
with which the (amine) focal points of three branching units
could form an amide bond. The acyl-protected dendrimer 13
was obtained in 64% yield after column chromatography, and
the subsequent deprotection of all twelve benzylic hydroxyl
groups with NaOMe in MeOH gave an 86% yield of the
The production of a small dendrimer: The problems associ-
ated with the branching unit 6 are dominated by its poor
solubility as well as by the difficulty encountered in the
deprotection of the amino function. We therefore aimed to
make a more lipophilic analogue which did not require such a
demanding deprotection. Our solution to these problems
(Scheme 2) was to utilize a more labile nitrogen-protecting
group and to perform the reduction of the ester groups much
earlier in the synthesis. The latter modification also allowed us
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0783 $ 17.50+.25/0
783

J. F. Stoddart et al.
FULL PAPER
The double-branching unit 15 can be considered an
orthogonally-protected amino tetraacid, ideal for use as a
versatile dendritic building block. Reactions involving the
formation of amide bonds from amines and acids are
extremely well known as a consequence of their importance
in nature, and they have also been used in previous dendrimer
syntheses. Very early in the history of dendrimers, lysine was
used as a branching unit by Denkewalter,[39] and more
recently, glutamic acid[40] has also been utilized in the syn-
thesis of dendrimers. Unnatural amino polyacids have also
AcO
O
OAc
Cl
AcO
AcO
Cl
O
O
OAc
O
O
N
AcO
O
OAc
OAc
Cl
N
O
N
Pyridine
O
OAc
OAc
O
O
O
O
HN
OAc
THF
O
AcO
O
64%
OAc
AcO
AcO
13
12
AcO
HO
OH
been used as branching units, notably by Shinkai,[41] Fre-
Â
HO
HO
O
chet,[42] and Diederich.[43]
The versatility of 15 was established by its use in a double
exponential growth step (Scheme 5). The amino function can
be deprotected with anhydrous hydrogen chloride in THF,
O
N
HO
O
OH
OH
NaOMe
MeOH
N
N
O
O
O
O
O
86%
OH
HO
O
HO
HO
14
CO2Me
O
HO
HCl / THF
Scheme 3. Synthesis of the deprotected second-generation dendrimer 14
from the improved double-branching unit.
CO2Me
N
H
17
84%
CO2Me
CO2Me
O
O
three-directional second-generation dendrimer 14. The den-
drimer was a glassy solid, with a molecular mass of 1287 Da,
and was soluble in polar solvents. The 1H NMR spectra
recorded at 273 K and 300 MHz for both the protected and
the deprotected dendrimers were broad as a consequence of
hindered rotation about the amide bond. However, at 373 K
CO2Me
CO2H
O
EtO P
EtO
CO2Me
CO2Me
N
15
O
NaOH
THF
H2O
O
CO2Me
O
CO2H
CO2H
N
EtO P
EtO
16
1
and 400 MHz, amide bond rotation is fast on the H NMR
timescale, leading to sharp spectra and hence making
characterization easier.
100%
O
CO2Me
CO2H
CO2Me
CO2Me
O
CO2Me
CO2Me
The synthesis of a new family of dendrimers: Although we
had been able to demonstrate the use of 11 as a branching
unit, the dendritic system that was developed had little
potential for the construction of large, highly designed
macromolecules. Our next objective was to develop the
system to allow more flexibility in dendrimer construction.
Considering our earlier experiences, we were struck by the
possibility that we might be able to create a double
exponential-type building block from two branching synthons
that we had synthesized previously. To this end, reaction of the
phenol 8 with the bistosylate 7 afforded the diethylphosphite-
protected tetramethyl ester 15 (Scheme 4) in a yield of 84%.
N
O
O
N
O
DCC
HOBt
THF
CO2Me
CO2Me
O
O
O
O
O
EtO P
EtO
CO2Me
CO2Me
N
18
O
93%
N
O
CO2Me
CO2Me
O
O
CO2Me
O
N
CO2Me
CO2Me
O
CO2Me
CO2Me
OTs
O
EtO P N
EtO
Scheme 5. Use of the building block 15 in a double exponential growth
step to form the fourth-generation building block 18.
CO2Me
O
OTs
K2CO3 / MeCN
O
7
CO2Me
CO2Me
and the carboxylic acids with NaOH in refluxing H2O/THF.
Both reactions proceeded in high yields (84% and 100%,
respectively), giving pure products 17 and 16, respectively,
after a single precipitation. In each case, the integrity of the
other protecting group is maintained. The amine 17 and
tetraacid 16 were coupled in the presence of dicyclohexylcar-
bodiimide (DCC) and 1-hydroxybenzotriazole (HOBt) to
give the fourth-generation wedge 18. It has a single protected
EtO P N
EtO
∆
O
CO2Me
84%
CO2Me
HO
15
CO2Me
8
Scheme 4. Synthesis of the orthogonally protected dendrimer building
block 15.
784
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0784 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
amine at its focal point, sixteen protected carboxylic acids on
its surface, and a molecular mass of ca. 2500 Da. The amine 17
was also treated with trimesic acid (20) to give the second-
generation dendrimer 19 in 70% yield (Scheme 6). Here, the
structures to be drawn concisely and help to clarify subse-
quent synthetic schemes and diagrams.
By way of introduction, the four reactions shown in detail in
Scheme 5 and Scheme 6 are depicted in cartoon form in
Scheme 7, along with some further reactions. This depiction
clarifies the way in which a controlled structure can be built up
in a highly efficient manner from the monomer 15 and the
core unit 20. The second-generation two-directional dendrim-
er 21 was isolated as a by-product of the three-directional
dendrimer-forming reaction to give 19. This scheme shows us
how much larger structures may be built up in very few steps.
The construction of two dendrimers from the small dendrimer
19 and the fourth-generation wedge 18 are also shown. The
first step in each case is the focal-point deprotection of 18.
This reaction was carried out with anhydrous hydrogen
chloride in THF, giving the amine 24 in a yield of 72%.
Although side reactions, resulting from ether and amide
cleavage, are slow under these conditions, the starting
material and the product both contain a very large number
of these somewhat reactive bonds. These cleavage reactions
can therefore cause a significant loss of yield if the reaction is
allowed to run for any longer than is absolutely necessary.
The fourth-generation dendrimer 26 was synthesized by the
reaction of the activated wedge 24 with trimesic acid (20)
under DCC/HOBt coupling conditions (Scheme 7). The
product has 48 terminal ester functions, two alternating
generation moieties, and a nominal mass of 7108 Da. The
yield (13%) was considerably lower than those obtained in
previous DCC ± HOBt couplings, and a significant quantity
(32%) of the disubstituted trimesic acid 25, a two-directional
dendrimer, was isolated. This decrease in yield suggests that
steric hindrance plays a significant role in the reactionÐan
encouraging sign that an interesting topology is being created.
An analytically pure sample of the three-directional den-
drimer 26 was not isolated from this reaction mixture, since it
was readily prepared by the reaction of the fourth-generation
wedge 24 with 1,3,5-benzenetricarbonyl trichloride in THF
(reaction not shown), which proceeds in the remarkable yield
of 63%.
O
OH
HO
CO2Me
O
O
O
HO
CO2Me
CO2Me
DCC HOBt THF
HN
17
70%
O
CO2Me
CO2Me
CO2Me
MeO2C
CO2Me
O
CO2Me
CO2Me
O
O
N
MeO2C
N
O
O
19
O
O
O
CO2Me
N
MeO2C
O
CO2Me
MeO2C
CO2Me
Scheme 6. Synthesis of the second-generation dendrimer 19.
activated wedge 17 was used as a second-generation wedge to
attach to the three-directional core. Dendrimer 19 has a
molecular mass of just over 1500 Da, and is terminated with
12 methyl ester functions on its surface.
From here on, in order to illustrate macromolecular
structures on the printed page, it will become necessary to
represent them as cartoons (Figure 2), in which each of the
The construction of the sixth-generation dendrimer 27
would be a true demonstration of the versatility of the
building block 15. While a more traditional synthesis would
take 12 steps from the monomer to create such a dendrimer,
the synthesis starting from 15 requires only seven. Here, the
second-generation dendrimer 19 was converted into a second-
generation core 22 by the hydrolysis of its 12 surface ester
groups. This core was then coupled with the fourth-generation
wedge 24 to give sixth-generation products. Since the
formation of dendrimer 27 (29000 Da) would require the
complete reaction of all 12 core functions, it is perhaps hardly
surprising that it does not go to completion. The hydrolysis
of dendrimer 19 to the dodecaacid 22 proceeds in excellent
yield.
We embarked on this sixth-generation dendrimer-forming
reaction not expecting to produce a perfect dendrimer, but to
see how far the reaction would proceed and to study the
properties of any products that we could isolate. Since the
crude reaction mixture defied purification by any form of
adsorption chromatography, we decided to study it by gel
Figure 2. The key to the schematic representations of dendrimers used in
this paper.
dendritic branching units is represented by truncated wedge
shapes with different shadings. Reactive functionalities are
shown as dark shadings, while passive functionalities are
shown as light ones. These illustrations allow very large
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0785 $ 17.50+.25/0
785

J. F. Stoddart et al.
FULL PAPER
Scheme 7. Cartoon representation of the synthesis of dendrimers from the building block 15.
permeation chromatography (GPC). This technique was able
to resolve a narrow peak corresponding to high molecular
weight products. A measured amount of the reaction mixture
was subjected to preparative GPC.[44] The high molecular
weight peak was isolated and found to account for approx-
imately 32% of the dendritic starting materials, indicating
that, on average, reaction of about one-third of the 12 active
sites on the core had occurred. Clearly, steric hindrance is of
extreme importance in this reaction. This conclusion is
consistent with the observation of a low yield in the formation
of the fourth-generation dendrimer 26.
In order to complete the family of dendrimers, the one-
directional analogues 29 and 30 of the second- and fourth-
generation dendrimers 19 and 26 were formed (Scheme 8).
The second- and fourth-generation wedges 17 and 24 were
each coupled with benzoic acid (28), a one-directional version
of the trimesic acid core that was used to make the three-
directional dendrimers.
Scheme 8. Cartoon representation of the synthesis of one-directional
dendrimers 29 and 30.
Mass spectrometry: The mass spectrometry (MS) of den-
drimers is a much studied[45] subject. All the dendrimers in our
series were studied by FABMS and MALDI-TOFMS. The
peaks, which were obtained for pseudomolecular ions, are
786
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0786 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
Table 1. Molecular ions obtained for the second- and fourth-generation dendrimers and wedges observed by FABMS and MALDI-TOFMS.
Wedges[a]
amine wedge
Dendrimers[b]
protected wedge
Nc ˆ 1
Nc ˆ 2
Nc ˆ 3
‡
‡
‡
‡
‡
Second generation 626 ˆ [M‡H] (FAB) 490 ˆ [M‡H] (FAB) 594 ˆ [M‡H] (FAB) 1167 ˆ [M‡H] (FAB)
1624 ˆ [M‡H] (FAB)
‡
‡
‡
‡
‡
Fourth generation 2456 ˆ [M‡H] (FAB) 2320 ˆ [M‡H] (FAB) 2446 ˆ [M‡H] (FAB) 4827 ˆ [M‡H] (FAB)
7113 ˆ [M‡H] (FAB)
4848,[c] 4847[d] (MALDI-TOF) 7144,[c] 7146[d] (MALDI-TOF)
[a] Protected wedge refers to the diethylphosphoramide-protected wedges 15 and 18. Amine wedge refers to the deprotected wedges 17 and 24. [b] Nc
denotes the core multiplicity of the dendrimer. [c] Performed with a gentisic acid matrix. [d] Performed with a retinolic acid matrix.
tabulated in Table 1. Although we were able to generate
‡
[M‡H] ions for all the pure dendrimers by FABMS, the
fourth-generation three-directional dendrimer 26 gave a peak
of very low intensity, and the mixture of sixth-generation
products showed no high molecular weight peaks at all.
MALDI-TOFMS was carried out on the two- and three-
directional fourth-generation dendrimers. This technique is
extremely sensitive to the matrix, and initial experiments with
sinnapinic acid and indoleacrylic acid matrices were unsuc-
cessful in affording any satisfactory results. A gentisic acid
matrix revealed molecular ions for fourth-generation den-
drimers, but strong spectra were only obtained with a retinolic
acid matrix in which the dendrimers could be ionized with a
relatively low laser power, leading to much less fragmentation
and baseline noise than was observed for other matrices.
Molecular ions obtained by MALDI-TOFMS were in the
‡
mass region expected for [M‡Na] ions. Analysis of the
mixture of sixth-generation products by MALDI-TOFMS
was unsuccessful.
Figure 3. The composite chromatograms of: A) the second-generation
dendrimers 19, 21, and 29; B) the fourth-generation dendrimers 23, 26, and
30.
Gel permeation chromatography: GPC was carried out on the
entire family of dendrimers in order to examine the structural
changes that take place with varying generation and core
directionality. Elution times were measured for:
as an intrinsic property of medium-to-high-generation den-
drimers. As the dendrimer grows, the number of surface
groups increases exponentially, eventually leading[46] to a
globular, almost spherical shape. The diameter of this shape
depends very largely on the length of the arms of the
dendrimer from its core to the surface. In the case of the two-
and three-directional fourth-generation dendrimers, this
length is exactly the same. We are led to conclude that the
fourth-generation dendrimer 26 is forced to adopt a highly
globular conformation in solution. The second-generation
dendrimer 19 does not have a greater surface area per surface
group[47] than 26 and yet does not exhibit this unusual feature.
The globular conformation of 26 is thus partly a result of its
large number of end groups giving rise to a high probability of
spherical topology, that is, the shape is entropy-driven rather
than sterically driven as is the case with the spherical
architecture of a dendrimer close to its dense-packing limit.
The composite GP chromatogram given in Figure 4 shows
the series of three-directional dendrimers from generation
zero to the mixture of sixth-generation products. While we can
observe that increasing generation number leads to increasing
size, we cannot gain quantitative knowledge from the data in
this form. In order to correlate generation number with a
practical quantity, the GPC column was calibrated with
respect to the hydrodynamic radius of the analyte.[48] Figure 5
shows a graph of dendrimer generation against calculated
the two three-directional dendrimers 19 and 26;
the two two-directional dendrimers 21 and 25, formed as
by-products in the synthesis of 19 and 26;
the two one-directional dendrimers 29 and 30;
the high molecular weight material from the sixth-gener-
ation dendrimer-forming reaction;
the zero-generation dendrimer analogue trimethyl trimes-
ylate (31).
Comparison of the composite GP chromatograms A and B
(Figure 3) sheds some interesting light on the relative
structures of the dendrimers. Chromatogram A shows a series
of second-generation dendrimers. It is clear that, in this
instance, raising the directionality of the core lowers the
elution time, demonstrating that molecules with higher
directionalities (and thus higher masses) occupy greater
volumes, as would be expected for linear polymers. This
trend is not nearly so pronounced for the corresponding
series of fourth-generation dendrimers (see chromatogram
B). While the one-directional dendrimer is observed to be
significantly smaller than either of its two- or three-direc-
tional derivatives, the latter have almost identical elution
volumes. This result suggests that these molecules occupy
almost the same volume despite the fact that the three-
directional dendrimer has 1.5 times the mass of the two-di-
rectional derivative. Such an observation can be interpreted
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0787 $ 17.50+.25/0
787

J. F. Stoddart et al.
FULL PAPER
visualization of the data not only reveals how similar the
spectra are, but also shows an interesting trend in chemical
shifts for the 1H nuclei near the surface of the dendrimer. As
the number of terminal groups increases (moving down the
table), the signals for these surface protons resonate slightly
further upfield. These shifts (0.1 ± 0.3 ppm) are small, and may
be the consequence of a microenvironmental effect. Similar
upfield shifts of these resonances are observed as the
dendrimers are warmed up. The internal moieties of the
larger dendrimers do not exhibit this behavior. In fact,
increasing the core directionality of the dendrimers causes a
Figure 4. The composite chromatogram for the three-directional dendrim-
ers 19, 26, 27, and 31. The small peak at 20 min is a consequence of the
(butylated hydroxytoluene) stabilizer present in the GPC grade THF that
was used in the purification of the sixth-generation product mixture.
1
very slight downfield movement of the aromatic H nuclei.
These trends are shown in graph form in Figure 6.
diameter. Although the dis-
Inner
unit
Outer
unit
Compound
tinctive linear relationship
that is an accepted feature[18]
of dendrimers is undeniably
evident, there is a slight
deviation observed for the
31
19
26
27
Second
generation
2.0
Fourth
generation
δ
8.0
1.0
0
sixth-generation
mixture.
7.5
7.0
0
2
4
6
This nonlinearity could be
explained by the close pack-
ing of surface groups. As the
generation number increases
and surface congestion be-
comes important, we might
expect that the internal
Generation
Figure 5. A plot of dendrimer
generation against hydrodynam-
ic radius for the three-direction-
al dendrimers 19, 26, 27, and 31.
10
100
Number of end groups
Figure 6. A plot showing the variation of chemical shift of aromatic 1H
nuclei with the number of surface (methyl ester) groups on the dendrimer,
on a logarithmic scale. The data is listed in Table 2.
structures develop a much more extended conformation.
The result would be a plot of generation against diameter
which deviates from linearity at high values.
The 1H NMR spectra for the one-, two-, and three-
directional second-generation dendrimers 29, 21, and 19
appear to be very similar with the exception of the resonances
for the aromatic core. Spectra for the three-directional
dendrimer 19 in CD3SOCD3 are shown in Figure 7. At
NMR spectroscopy: There is great interest in the internal
structure and dynamics of dendritic systems, since these
properties dictate many of their unusual characteristics.
Although many probes of the dendritic microenvironment
have been developed, and molecular dynamics simulations
have also been utilized, NMR spectroscopy provides a good
way of determining the internal dynamics of dendrimers. For
instance, 13C T1 relaxation times have been used[49] to gain
insight into the mobility of different parts of dendrimers.
The 1H NMR chemical shift data for the second- and
fourth-generation dendrimers are listed in Table 2. This
Table 2. 1H Chemical shift data for the second- and fourth-generation
dendrimers 19, 21, 25, 26, 29, and 30.
[a]
d at 373 K at 400 MHz in CD3SOCD3
aliphatic
CH2O
inner[f] Ar outer[f] Ar
g[b] nc[c] NCH2
2H
H
2H
H
Me
Figure 7. The 1H NMR spectra of the three-directional second-generation
dendrimer 19 at 304 and 373 K recorded in CD3SOCD3 at 400 MHz.
29
21
19
30
25
26
2
2
2
4
4
4
1
2
3
1
2
3
4.06
3.88
3.91
3.85 ± 3.9
3.8 ± 3.9
4.40
4.31
4.29
±
±
±
±
±
±
7.63 8.10 3.90
7.53 7.97 3.85
7.49 7.92 3.84
297 K, apart from the core and surface protons, every signal
appears as two broad peaks, a consequence of the slow
rotation of amide bonds on the 1H NMR timescale. When the
temperature is raised to 397 K, these broad signals collapse
into A2B2 and AB2 systems for aliphatic and aromatic protons,
respectively, in line with amide bond rotation becoming fast
on the 1H NMR timescale. Energy barriers for these
degenerate processes were calculated[50] (Table 3) at the
4.28[d] 4.24[e] 7.06 7.20 7.50 7.94 3.83
4.2 ± 4.3
7.09 7.21 7.45 7.88 3.80
3.90[d] 3.83[e] 4.30[d] 4.20[e] 7.10 7.23 7.40 7.83 3.77
[a] Resonances for protons of the core moieties have not been tabulated.
[b] g represents the generation number of the dendrimer. [c] nc represents
the core multiplicity of the dendrimer. [d] Inner moiety. [e] Outer moeity.
[f] Where the dendrimer has only one aromatic branching unit, this is
considered to be the outer unit.
788
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0788 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
Table 3. Kinetic and thermodynamic parameters for amide bond rotations
of the one-, two-, and three-directional second-generation dendrimers 29,
21, and 19 in CD3SOCD3 and at 400 MHz determined by variable-temper-
building blocks are separated into two signals as a conse-
quence of the hindered rotation about the inner amide bonds.
The signals for protons in the outer building blocks, however,
can be affected by both the inner and the outer amide bonds,
leading to the observed four signals for proton z. As the
sample is warmed up, pairs of the four signals coalesce,
leaving two peaks at 306 K. If the process we are observing
emanates from the inner amide bond, then the two coales-
cences describe the same process, and will have the same
energy. However, if this process is a consequence of the outer
amide bonds, then the two coalescences describe different
processes, and may exhibit different energy barriers. When
the sample is warmed up further, the peaks finally collapse
into a single peak, as all the amide bond rotations become fast
on the 1H NMR timescale. Free energies of activation for the
processes described above were calculated from VT-NMR
coalescence studies on the dendrimer. The DG= values are
presented in Table 4. We might expect that the smaller
1
ature H NMR spectroscopy by the coalescence method (see ref. [23]).
1
Probe protons[b] DnÄ (Hz) kc (s
)
Tc (K) DG=
(kcalmol
1
)
29 (nc ˆ 1)[a] NCH2 {b}
78.8
45.2
8.0
36.4
16.4
64.0
175
100
18
81
36
318
313
304
315
313
324
15.4
29 (nc ˆ 1)
21 (nc ˆ 2)
21 (nc ˆ 2)
19 (nc ˆ 3)
19 (nc ˆ 3)
CH2O {c}
15.5
16.1
15.7
16.1
15.8
Ar (2H) {d}
Ar (H) {e}
Ar (2H) {d}
Ar (H) {e}
142
[a] nc refers to the core multiplicity of the dendrimer. [b] The letter in curly
brackets refers to the analagous proton of 19, depicted in Figure 7.
coalescence temperatures for the aromatic protons, measured
by variable-temperature (VT) NMR spectroscopy. In the case
of the one-directional dendrimer 29, the coalescence temper-
ature of the aromatic proton resonances was too close to the
freezing point of the solvent to be accessible and so the
coalescences of the aliphatic resonances were measured.
Although there appears to be a slight increase in the free
energy of activation for the amide bond rotation process as
the dendrimer acquires more directionality, the results for the
three dendrimers lie within experimental error of each other.
In essence, we are unable to suggest that the three-directional
dendrimer 19 is significantly more rigid than the one-direc-
tional dendrimer 29. This result is consistent with the GPC
finding that the second-generation dendrimer series behave
much like linear polymers.
Table 4. Kinetic and thermodynamic parameters for amide bond rotations
of the 2-directional second-generation dendrimer 25 in CD3SOCD3 and at
400 MHz as determined by the NMR coalescence method.
1
Process[a] Probe
protons[b]
DnÄ (Hz)
kc (s
)
Tc (K)
DG=
(kcalmol
1
)
outer
outer
inner
inner
inner
y[c,d]
25.2
56
11.6
14.8
16.0
56
124
26
33
36
321
326
307
307
307
16.3
16.0
16.0
15.8
15.8
z[c]
w
z[c]
z[c]
[a] Inner refers to the rotation of the amide bond closer to the core, whilst
outer refers to the rotation of the amide bond closer to the suface. [b] The
protons w, x, y, z are defined in Figure 8. The resonance for proton x is
obscured. [c] Inner and outer assignments are tentative. [d] Proton y cannot
be used to measure other processes under these conditions as a temper-
ature below the freezing point of the solvent is required.
The two-directional fourth-generation dendrimer 25 is the
largest dendrimer we have prepared in sufficient purity for
VT-NMR spectroscopic analysis. The aromatic regions of H
NMR spectra, recorded at various temperatures, are shown in
Figure 8. At 292 K, the resonances for the protons of the inner
1
separations (two signals into four) of the protons z reflect
hindered rotation about the more distant inner amide bonds.
The results of the coalescence studies are consistent with this
hypothesisÐthe two coalescences describe a process with the
same free energy of activation. Comparisons of these results
with those from the analyses of the second-generation
dendrimers can be made (Figure 8). The outer amide bond
rotations probed by protons y and z in the fourth-generation
dendrimer are associated with slightly higher energy barriers
than the analogous processes (probed by protons d and e) in
the second-generation dendrimers. This observation suggests
that the larger dendrimer is more sterically crowded on the
surface of the molecule. Indeed, this conclusion has already
been reached on the basis of GPC analyses and comparisons
of yields.
Molecular modeling: Molecular modeling studies have been
used extensively in the analysis and visualization of dendrim-
ers.[51] Of particular interest are their shapes and diameters,
the existence of voids and channels, and the positions of the
chain ends. These topological features are likely to dictate the
properties, unique to dendrimers, that we wish to exploit.
Molecular dynamics simulations were conducted[52] on the
Figure 8. The 1H NMR spectra of the two-directional fourth-generation
dendrimer 23 at 292, 307, and 373 K recorded in CD3SOCD3 at 400 MHz.
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0789 $ 17.50+.25/0
789

J. F. Stoddart et al.
FULL PAPER
dendrimers 19, 26, and 27, utilizing the AMBER* forcefield
resident within the Macromodel[53] package. Figure 9 shows
structures of the second-generation three-directional den-
drimer 19, taken from a stochastic molecular dynamics run
Finally, a structure for the sixth-generation three-direc-
tional dendrimer 27 is shown in Figure 11. Analysis of the
energetics of this structure reveals that none of the energy
Figure 9. Two room-temperature samples (A and B) of conformations
available to the three-directional second-generation dendrimer 19 at 300 K
as determined by stochastic molecular dynamics.
after equilibration at 300 K. A high degree of conformational
freedom is evident, reiterating the similarity observed be-
tween this dendrimer and a linear polymer. The models
measure an average of 1.1 nm in radius, in very close
agreement with the results (radius of 1.0 nm) of the GPC
analysis. The structure of the fourth-generation three-direc-
tional dendrimer 26 (Figure 10) is strikingly different. Much
Figure 11. One transient conformation of the sixth-generation dendrimer
27 at 300 K as determined by stochastic molecular dynamics.
terms are unrealistically high, suggesting that the structure is
certainly a feasible synthetic target. This is not surprising,
since the theoretical volume available to monomer units in the
sixth-generation dendrimer is little different from that in the
second-generation dendrimer 19.[54] All of the observed
conformations at 300 K are highly globular, and channels
and voids are apparent in the structure. As with the fourth-
generation dendrimer, the chain ends reside predominantly at
the surface of the molecule, and once again, the radius of the
model (2.4 nm) compares remarkably well with the value
obtained experimentally (2.3 nm) from GPC of the sixth-
generation product mixture.
Conclusions
The construction of a family of dendrimers derived from
diethanolamine and 5-hydroxyisophthalic acid branching
units has been accomplished. The family consists of one-,
two-, and three-directional second- and fourth-generation
dendrimers (the zero-generation analogue) and a mixture of
sixth-generation dendritic products. These dendrimers have
been synthesized from a second-generation building block in
the form of an orthogonally protected amino tetraacid. We
believe that the simplicity and versatility of this dendritic
Figure 10. A possible conformation of the fourth-generation dendrimer 25
at 300 K as determined by stochastic molecular dynamics.
less conformational freedom is observed, and a more globular
structure results from the steric congestion of surface groups.
The model has an average radius of 1.6 nm, which is once
again consistent with the GPC analysis where a radius of
1.5 nm was deduced.
790
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0790 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
reach ambient temperature, and was then stirred overnight. The reaction
mixture was poured onto ice (50 g) and then neutralized by the addition of
concentrated HCl. The organic layer was added to extractions (CH2Cl2,
2 Â 50 mL) of the aqueous layer, before being dried (CaCl2). The solvent
was removed in vacuo to give a solid which was recrystallized twice from
MeOH to give the sulfonamide 3 (45.8 g, 85%), white solid, m.p. 95 ± 968C.
1H NMR (300 MHz, CDCl3, 258C): d ˆ 7.76 (m, 4H; Ar), 7.62 (m, 2H; Ar),
7.37 (m, 4H; Ar), 7.30 (m, 2H; Ar), 4.11 (t, J ˆ 6 Hz, 4H; CH2O), 3.37 (t,
J ˆ 6 Hz, 4H; NCH2), 2.47 (s, 6H; CH3), 2.43 (s, 3H; CH3); 13C NMR
(75 MHz, CDCl3, 258C): d ˆ 145, 144, 135, 132, 130, 128, 127, 68, 48, 21; MS
building block gives us the potential to incorporate other units
within the dendritic structure at will; any suitably protected
amino polyacid could be used as a monomer in any
convergent or divergent generation-adding step.
The internal structure of the dendrimers has been probed
by GPC, VT-NMR spectroscopy and molecular modeling
experiments. The results of these studies are highly consistent
with each other. They show that the higher generation
dendrimers adopt the globular, void-containing conforma-
tions associated with the properties that are peculiar to
dendrimers. This observation is particularly interesting con-
sidering that these higher generation dendrimers are not close
to their starburst limit. We are led to the conclusion that these
structures and their analogues could harbor the potential to
act as host molecules and exhibit other novel properties.
However, caution must be exercised when elevating dendritic
structures to supramolecular status. The dendritic structure is
a dynamic system, and, as such, lacks the degree of preorga-
nization that is necessary for highly specific interactions with
substrates.
‡
‡
(CI ‡ ): m/z (%) ˆ 585 (11%) [M‡NH3] , 567 (1%) [M] , 242 (100%)
‡
[M 2OTs‡H] ; C25H29NO8S3: calcd C 52.89, H 5.15, N 2.47; found C
53.19, H 5.14, N 2.36.
Diethyl 5-hydroxyisophthalate (4): 5-Hydroxyisophthalic acid (2, 23.5 g,
129 mmol) was dissolved in EtOH (400 mL), and concentrated H2SO4 was
added (20 mL). The reaction mixture was stirred under reflux for 5 d. After
cooling, saturated NaHCO3 solution (ca. 400 mL) was added carefully to
neutralize the acid. The solution was reduced in volume to ca. 400 mL in
vacuo, more H2O (200 mL) was added, and the product was obtained by
filtration, before being washed with H2O (50 mL) to give the diester 4
(26.3 g, 86%) as a colorless solid, m.p. 105 ± 1068C. 1H NMR (300 MHz,
CD3COCD3, 258C): d ˆ 9.06 (s, 1H; OH), 8.12 (t, J ˆ 1.5 Hz, 1H; Ar), 7.70
(d, J ˆ 1.5 Hz, 2H; Ar), 4.37 (q, J ˆ 7 Hz, 4H; CH2), 1.38 (t, J ˆ 7 Hz, 6H;
CH3); 13C NMR (75 MHz, CD3COCD3, 258C): d ˆ 165.9, 158.7, 133.0,
‡
121.8, 121.0, 61.7, 14.4; MS (CI ‡ ): m/z (%) ˆ 239 (90%) [M‡H] , 256
‡
(100%) [M‡NH4] ; C12H14O5: calcd C 60.50, H 5.92; found C 60.76, H 5.97.
N,N-Bis{2-[3,5-bis(ethoxycarbonyl)phenoxy]ethyl}-p-toluene sulfonamide
(5): A suspension of K2CO3 (38 g, 275 mmol) in MeCN (1500 mL) was
degassed. The sulfonamide 3 (18.3 g, 32.3 mmol) and the diester 4 (26.7 g,
112 mmol) were dissolved in the solvent, and the reaction mixture was then
stirred under reflux for 7 d. The reaction mixture was cooled down, filtered
at the pump, and the solid was washed with MeCN (ca. 50 mL). The solvent
was removed in vacuo to leave a thick yellow syrup, which crystallized from
MeOH/H2O to give the pure sulfonamide 5 (18.7 g, 83%) as a white solid,
m.p. 668C. 1H NMR (300 MHz, CDCl3, 258C): d ˆ 8.26 (t, J ˆ 1.5 Hz, 2H;
Ar), 7.75 (m, 2H; Ar), 7.62 (d, J ˆ 1.5 Hz, 4H; Ar), 7.26 (obscured by
solvent, m, Ar), 4.39 (q, J ˆ 7 Hz, 8H; CO2CH2), 4.27 (t, J ˆ 6 Hz, 4H;
CH2O), 3.75 (t, J ˆ 6 Hz, 4H; NCH2), 2.38 (s, 3H; CH3), 1.39 (t, J ˆ 7 Hz,
12H; CH3); 13C NMR (75 MHz, CDCl3, 258C): d ˆ 165.5, 158.2, 143.8,
136.6, 132.2, 129.8, 127.2, 123.4, 119.5, 67.6, 61.4, 48.7, 21.5, 14.3; MS
Experimental Section
General procedures and instrumentation: Chemicals were purchased from
Aldrich and were used without further purification. Solvents were dried
according to literature methods. Column chromatography was performed
on silica gel 60 (particle size 0.040 ± 0.063 mm, Merck 9385) and thin-layer
chromatography (TLC) on Merck aluminum-backed plates coated with
silica gel 60 F254 (layer thickness 0.2 mm). Once eluted, plates were dried
with hot air and scrutinized under a UV lamp or developed using a
ninhydrin[55] or cerium[56] stain. High performance liquid chromatography
(HPLC) was performed on a Phenomenex IB-Sil C-18 column (250 Â
10 mm), eluted on Gilson 305 and 306 HPLC pumps. The pumps were
controlled by external Gilson 715 software running on a 486 PC; detection
was achieved with a Dynamax UV-1 ultraviolet detector set at 254 nm.
GPC was carried out on two 300 mm Phenomenex Phenogel styrene/
divinylbenzene columns with pore sizes of 500 and 50 Š. The columns were
linked in series and were calibrated with polystyrene standards purchased
from Polymer Laboratories Limited. The columns were eluted with pure
GPC grade THF (stabilized with BHT) on the HPLC system. Melting
points were determined on an Electrothermal 9200 melting point apparatus
and are uncorrected. Routine NMR spectra were recorded on a Bruker
AC300 (300 MHz for 1H, 75 MHz for 13C) spectrometer. Variable-temper-
ature spectra were obtained on a Bruker AMX400 (400 MHz for 1H,
100 MHz for 13C) spectrometer. With both spectrometers, the solvent
reference was used as internal standard. Electron impact ± chemical
ionization mass spectrometry (EI-CIMS) was carried out on a Kratos
Profile mass spectrometer and liquid secondary ion mass spectrometry
(LSIMS) on a VG ZabSpec instrument, equipped with a cesium ion source
and using 3-nitrobenzyl alcohol as the matrix. Matrix-assisted laser
desorption ± ionization (time of flight) mass spectrometry (MALDI-
TOFMS) was performed on a Kratos Kompact instrument with either
gentisic acid (2,5-dihydroxybenzoic acid), sinnapinic acid (3,5-dimethoxy-
4-hydroxycinnamic acid), 3-b-indoleacrylic acid, or all trans-retinoic acid as
the matrix, and was calibrated with insulin. Microanalyses were performed
by either the University of Birmingham Microanalytical Service, the
University of Sheffield Microanalytical Service, or the University of North
London Microanalytical Service.
‡
‡
(EIMS ‡ ): m/z (%) ˆ 700 (16%) [M‡H] , 654 (100%) [M OEt] ;
C35H41NO10S: calcd C 60.07, H 5.74, N 2.00; found C 59.97, H 5.74, N 2.07.
Bis{2-[3,5-bis(hydroxymethyl)phenoxy]ethyl}amine (6): Red-Al (65% in
PhMe, 7 mL) was added to dry dioxane (100 mL) with vigorous stirring
under nitrogen. A solution of sulfonamide 5 (0.89 g, 1.27 mmol) in dry
dioxane (50 mL) was added dropwise over a period of 30 min using a
pressure-equalized dropping funnel. The bright yellow reaction mixture
was brought to reflux. The color changed to green and then to reddish-
orange. The reaction mixture was then heated under reflux for 18 h, after
which time the mixture resembled mud. On cooling, the excess of reagent
was decomposed by the dropwise addition of 10% aqueous NaOH solution
(15 mL). This mixture was stirred for 30 min before the solvents were
removed in vacuo. The residue was dissolved in H2O (50 mL) and the
aqueous solution was extracted with EtOAc (5 Â 50 mL). These extracts
were combined, dried (Na2SO4), filtered, and the solvent was removed in
vacuo to leave a thick yellow syrup. This crude product was purified by
column chromatography on silica gel, eluting with a solvent gradient from
pure EtOAc to 6:3:1 EtOAc:Et3N:MeOH to give the amine 6 (68 mg,
1
14%) as a thick colorless syrup. H NMR (300 MHz, CD3COCD3, 258C):
d ˆ 6.90 (s, 2H; Ar), 6.83 (s, 4H; Ar), 4.56 (s, 8H CH2OH), 4.07 (t, J ˆ 6 Hz,
4H; CH2OAr), 3.5 (br, 4H; OH), 3.02 (t, J ˆ 6 Hz, 4H; NCH2); 13C NMR
(75 MHz, CD3COCD3, 258C): d ˆ 160.0, 144.7, 117.7, 111.9, 68.4, 64.4, 49.3;
‡
MS (EI ‡ ): m/z (%) ˆ 210 (84%) [M CH2OC6H3(CH2OH)2] .
N,N-Bis[2-(4-toluene)sulfonyloxy]ethyl(diethyl)phosphoramide (7): Di-
ethanolamine (1, 25.4 g, 242 mmol) was added to a stirred mixture of
KHCO3 (49.5 g, 495 mmol), K2CO3 (70.4 g, 510 mmol) and TBAB (4.32 g,
12 mmol) in THF (250 mL). The mixture was cooled to 58C under a
nitrogen atmosphere, after which a solution of diethylphosphite (36.3 g,
265 mmol) in CCl4 (50 mL) was added dropwise. The reaction mixture was
allowed to reach ambient temperature and then stirred for a further 1 h.
The mixture was filtered and the filtrate was combined with a washing of
N,N-Bis[2-(p-toluenesulfonyloxy)ethyl]-p-toluenesulfonamide (3): Dry
Et3N (80 mL), diethanolamine (1, 10.0 g, 95 mmol), and a catalytic quantity
of DMAP were dissolved in dry CH2Cl2 (400 mL) in a 1 L flask. The
solution was cooled to 08C, then p-toluenesulfonyl chloride (65.1 g,
341 mmol) in dry CH2Cl2 (100 mL) was added slowly through a pressure-
equalized dropping funnel, while the the temperature was maintained
below 58C with constant stirring. The reaction mixture was allowed to
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0791 $ 17.50+.25/0
791

J. F. Stoddart et al.
FULL PAPER
the solid with CH2Cl2 (50 mL). Finally, the solvents were removed in vacuo.
The remaining thick oil (64.7 g) was dissolved in CH2Cl2 (200 mL); C5H5N
(45 mL) and a catalytic quantity of DMAP were added. The reaction
mixture was cooled to 08C, while stirring was maintained under a nitrogen
atmosphere. A solution of p-toluenesulfonyl chloride (106.7 g, 562 mmol)
in CH2Cl2 (200 mL) was added carefully without allowing the temperature
to exceed 08C. Then, the reaction mixture was allowed to warm up while it
was stirred for a further 4 h. H2O (150 mL) was added and the reaction was
stirred vigorously for a further 30 min, before more H2O (150 mL) was
added. The organic layer was combined with one extraction (50 mL) of the
aqueous phase with CH2Cl2. The crude material was recovered by
evaporation of the organic solvents after the solution had been dried
(CaCl2). The pure product was easily obtained by flash column chroma-
tography on a short silica gel column, eluting with a gradient from pure
CH2Cl2 to pure EtOAc to give the phosphoramide 7 (55.3 g, 42%) as a
thick colorless oil. 1H NMR (300 MHz, CDCl3, 258C): d ˆ 7.76 (d, J ˆ
8.1 Hz, 4H; Ar), 7.34 (d, J ˆ 8.1 Hz, 4H; Ar), 4.04 (t, J ˆ 5.5 Hz, 4H;
CH2OTs), 3.9 (m. 4H; CH2OP), 3.27 (m. 4H; NCH2), 2.43 (s. 6H; ArCH3),
1.21 (dt, J ˆ 0.95, 7.1 Hz, 6H; CH3); 13C NMR (75 MHz, CDCl3, 258C): d ˆ
145.0, 132.6, 129.9, 127.9, 68.8, 62.5 (d, J ˆ 5.6 Hz), 46.2 (d, J ˆ 4.5 Hz),
21.6, 16.0 (d, J ˆ 7.3 Hz); 31P NMR (162 MHz, CDCl3, 258C): d ˆ 9.20;
colorless oil. 1H NMR (300 MHz, CD3COCD3, 258C): d ˆ 7.00 (s, 2H; Ar),
6.93 (s, 4H; Ar), 5.06 (s, 8H; PhCH2O), 4.21 (t, J ˆ 6 Hz, 4H; CH2O), 4.05
(m, 4H; CH2OP), 3.57 (m, 4H; NCH2), 2.06 (s, 12H; COCH3), 1.28 (t, J ˆ
7 Hz, 6H, CH3); 13C NMR (75 MHz, CD3COCD3, 258C): d ˆ 170.7, 159.9,
139.2, 120.5, 114.3, 67.8, 66.0, 62.5, 62.4, 47.3, 20.7, 16.5, 16.4; 31P NMR
(162 MHz, CDCl3, 258C): d ˆ 10.25; MS (FAB ‡ ): m/z (%) ˆ 682 (100%)
‡
[M‡H] ; C32H44NO13P: calcd C 56.38, H 6.51, N 2.05; found C 56.33, H
6.64, N 2.13.
Bis{[2-(3,5-bisacetoxymethyl)phenoxy]ethyl}amine (12): A solution of the
phosphoramide 11 (2.01 g, 2.95 mmol) in freshly distilled THF (50 mL) was
cooled to <58C before it was saturated with anhydrous HCl. The reaction
mixture was allowed to warm up to ambient temperature and the reaction
was followed by TLC until the starting material had almost completely
disappeared (ca. 2 h). The volatiles were removed in vacuo and then the
residue was shaken with Et2O (50 mL). After decanting the liquid, the
remaining material was dissolved in H2O (50 mL) before being basified
with 15% NaOH solution (2.5 mL). The mixture was extracted with
CH2Cl2 (2 Â 15 mL) and the extractions were combined, dried (MgSO4),
and concentrated to a residue. This crude product was purified by column
chromatography on silica gel, eluting with a gradient from EtOAc:MeOH
(9:1) to EtOAc:Et3N:MeOH (8:1:1) to give the amine 12 (498 mg, 31%) as
‡
‡
MS (FAB ‡ ): m/z (%) ˆ 378 (88) [M OTs] , 550 (100) [M‡H] ;
1
a thick oil. H NMR (300 MHz, CD3COCD3, 258C): d ˆ 6.99 (s, 2H; Ar),
C22H32NO9PS2: calcd C 48.08, H 5.87, N 2.55; found C 47.60, H 6.07, N 2.60.
6.93 (s, 4H; Ar), 5.05 (s, 8H; PhCH2O), 4.16 (t, J ˆ 6 Hz, 4H; CH2O), 3.10
(t, J ˆ 6 Hz, 4H; NCH2), 2.06 (s, 12H; CH3); 13C NMR (75 MHz,
CD3COCD3, 258C): d ˆ 170.7, 160.1, 139.2, 120.4, 114.4, 68.5, 66.0, 20.7;
Dimethyl 5-hydroxyisophthalate (8): 5-Hydroxyisophthalic acid (2, 120 g,
659 mmol) and H2SO4 (60 mL) were dissolved in MeOH (1000 mL) and the
solution was stirred under reflux for 20 h. The reaction mixture was allowed
to cool down to ambient temperature and then H2O (1000 mL) was added.
The product was isolated by filtration from the crude reaction mixture. It
was then washed with H2O (100 mL) to give the ester 8 (135 g, 97%) as a
colorless solid, m.p. 1658C. 1H NMR (300 MHz, CD3COCD3, 258C): d ˆ
9.10 (br, 1H; OH), 8.11 (t, J ˆ 1 Hz, 1H; Ar), 7.68 (d, J ˆ 1 Hz, 2H, Ar),
3.91 (s, 6H; CH3); 13C NMR (75 MHz, CD3COCD3, 258C): d ˆ 166.27,
‡
MS (FAB ‡ ): m/z (%) ˆ 546 (3%) [M‡H] ; C28H35NO10: calcd C 61.64,
6.47, N 2.57; found C 61.65, H 6.48, N 2.49.
1,3,5-Tris{N,N-bis[2-(3,5-bis(acetoxymethyl)phenoxy)ethyl]aminocarbo-
nyl}benzene (13): A solution of 1,3,5-benzenetricarbonyl chloride (69.1 mg,
0.26 mmol) in THF (3 mL) was added dropwise with stirring to a mixture of
the amine 12 (423 mg, 0.77 mmol) and C5H5N (80 mL, 1 mmol) in THF
(20 mL). The reaction mixture was stirred at ambient temperature for
10 min. The solvent was removed in vacuo and then the crude product was
subjected to column chromatography on silica gel, eluting with EtOAc to
give the protected dendrimer 13 (298 mg, 64%) as a thick oil. 1H NMR
(400 MHz, CD3SOCD3, 1008C): d ˆ 7.63 (s, 3H; Ar[core]), 6.90 (s, 6H;
Ar), 6.82 (s, 12H; Ar), 4.99 (s, 24H; PhCH2O), 4.18 (br, 12H; CH2O), 3.84
(br, 12H; NCH2), 2.02 (s, 36H; CH3); MS (FAB ‡ ): m/z (%) ˆ 1815
‡
158.48, 132.84, 122.05, 121.00, 52.48; MS (EI ‡ ): m/z (%) ˆ 210 (36) [M] .
3,5-Bishydroxymethylphenol (9): A solution of dimethyl 5-hydroxyisoph-
thalate (8, 30.0 g, 143 mmol) in dry THF (150 mL) was added slowly under
nitrogen to a stirred suspension of LiAlH4 (10.0 g, 263 mmol) in THF
(500 mL). The reaction was stirred under reflux for 3 h before being left to
cool down. The mixture was acidified by the addition of 10% H2SO4
(200 mL) and then the THF was removed by distillation in vacuo. The
resulting solution was extracted with EtOAc (5 Â 250 mL) and the extracts
were combined and dried (MgSO4). Filtration and evaporation of the
solvent gave the pure phenol 9 (22.2 g, 100%) as a colorless solid, m.p. 73 ±
748C. 1H NMR (300 MHz, CD3COCD3, 258C): d ˆ 8.17 (s, 1H; ArOH),
6.81 (s, 1H; Ar), 6.74 (s, 2H; Ar), 4.55 (d, J ˆ 5 Hz, 4H; CH2), 4.10 (t, J ˆ
5 Hz, 2H; OH); 13C NMR (75 MHz, CD3COCD3, 258C): d ˆ 158.2, 144.7,
‡
(100%) [M‡Na] ; C93H105N3O33: HRMS calcd 1814.652 [M‡Na], 1815.656
[M‡Na]; found 1814.660, 1815.667.
1,3,5-Tris{N,N-bis[2-(3,5-bis(hydroxymethyl)phenoxy)ethyl]aminocarbo-
nyl}benzene (14): Na metal (60 mg, 2.6 mmol) was dissolved in dry MeOH
(10 mL) at 08C. A solution of the protected dendrimer 13 (166 mg,
0.87 mmol) in CH2Cl2 (1 mL) was added and then the reaction mixture was
stirred at ambient temperature for 2 h. The mixture was neutralized
carefully with 10% H2SO4, filtered, and the solvents were removed in
vacuo to leave a product, which was purified by HPLC (Phenomenex IB-
SIL C-18 column, H2O) to give the dendrimer 14 (98 mg, 86%) as a thick
oil. 1H NMR (400 MHz, CD3SOCD3, 1008C): d ˆ 7.60 (s, 3H; Ar[core]),
6.88 (s, 6H; Ar), 6.73 (s, 12H; Ar), 4.70 (brt, J ˆ 5 Hz, 12H; OH), 4.44 (d,
J ˆ 5 Hz, 24H; PhCH2O), 4.15 (brt, J ˆ 5 Hz, 12H; CH2O), 3.81 (brt, J ˆ
5 Hz, 12H; NCH2); 13C NMR (100 MHz, CD3SOCD3, 1008C): d ˆ 169.6,
157.8, 143.5, 136.7, 125.5, 116.9, 110.8, 65.1, 62.6; MS (FAB ‡ ): m/z (%) ˆ
‡
116.5, 112.6, 64.6; MS (CI ‡ ): m/z (%) ˆ 154 (100%) [M] ; C8H10O3: calcd
C 62.34, H 6.49; found C 62.32, H 6.69.
3,5-Bis(acetoxymethyl)acetoxybenzene (10): Ac2O (250 mL) was added
with stirring to the phenol 9 (22.0 g, 142 mmol) in dry C5H5N (1000 mL)
under nitrogen. The reaction mixture was stirred at ambient temperature
for 24 h and then the solvents were removed in vacuo. The residue was
dissolved in EtOAc (100 mL) and purified by filtration through silica,
which was washed with further EtOAc (100 mL). The EtOAc was removed
to give the pure triacetate 10 (38.0 g, 95%) as a thick colorless oil, b.p.
‡
1288 (59%) [M‡H] ; C69H82N3O21: HRMS calcd 1288.544 [M‡H]; found
1
1158C (0.01 mbar, decomp.). H NMR (300 MHz, CDCl3, 258C): d ˆ 7.16
1288.545.
(brs, 1H; Ar), 7.00 (brs, 2H; Ar), 5.04 (s, 4H; CH2), 2.24 (s, 3H; CH3), 2.06
(s, 6H; CH3); 13C NMR (75 MHz, CDCl3, 258C): d ˆ 170.2, 169.2, 150.6,
138.2, 124.6, 120.8, 64.7, 20.9, 20.7; MS (CI ‡ ): m/z (%) ˆ 298 (43%)
N,N-Bis{2-[(3,5-bismethoxycarbonyl)phenoxy]ethyl}(diethyl)phosphora-
mide (15): A mixture of the phosphoramide 7 (8.44 g, 15.4 mmol), the ester
8 (10.59 g, 50.4 mmol), and K2CO3 (30.0 g, 217 mmol) in MeCN (100 mL)
was stirred under reflux in a nitrogen atmosphere for 7 h. The cooled
reaction was filtered and the solid was washed with MeCN (10 mL). The
solvent was removed in vacuo and the product was crystallized from
MeOH/H2O. Final traces of the phenolic starting material were removed by
‡
[M‡NH4] ; C14H16O6: calcd C 60.00, H 5.75; found C 59.66, H 6.00.
N,N-Bis[2-(3,5-bisacetoxymethyl)phenoxy]ethyl(diethyl)phosphoramide
(11): The triacetate 10 (2.92 g, 104 mmol), 18-crown-6 (315 mg, 1.2 mmol),
K2CO3 (5.15 g, 37.3 mmol), and MeCN (dry, 50 mL) were placed in a flask
under nitrogen and the reaction mixture was sonicated for 5 min. The
mixture was stirred under reflux for 1 h before it was allowed to cool down
to ambient temperature, at which point a solution of the phosphoramide 7
(1.69 g, 3.8 mmol) in MeCN (dry, 10 mL) was added through a pressure-
equalized dropping funnel. The reaction mixture was reheated and stirred
under reflux for 2 d. After cooling, the mixture was filtered and the solid
washed with MeCN (10 mL). The solvent was removed to leave a crude
product, which was subjected to column chromatography on silica gel,
eluting with EtOAc to give the phosphoramide 11 (517 mg, 23%) as a thick
drying
a CH2Cl2 solution of the product over K2CO3 to give the
phosphoramide 15 (7.80 g, 84%) as a colorless solid, m.p. 80 ± 818C. 1H
NMR (300 MHz, CDCl3, 258C): d ˆ 8.23 (t, J ˆ 1.3 Hz, 2H; Ar), 7.66 (d,
J ˆ 1.5 Hz, 4H; Ar), 4.19 (t, J ˆ 14.8 Hz, 4H; CH2OAr), 4.04 (q, J ˆ 7.1 Hz,
4H; CH2OP), 3.89 (s. 12H; CO2Me), 3.58 (m. 6H; CH3), 1.27 (t, J ˆ 7.0 Hz,
4H; NCH2); 13C NMR (75 MHz, CDCl3, 258C): d ˆ 165.9, 158.6, 131.8,
123.2, 119.6, 67.9, 62.5, 62.4, 52.4, 46.9, 46.8, 16.2; 31P NMR (162 MHz,
‡
CDCl3, 258C): d ˆ 10.16; MS (FAB ‡ ): m/z (%) ˆ 626 (100) [M‡H] ;
C28H36NO13P: calcd C 53.76, H 5.80, N 2.24; found C 53.57, H 5.89, N 2.12.
792
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0792 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
N,N-Bis{2-[3,5-bis(hydroxycarbonyl)phenoxy]ethyl}(diethyl)phosphora-
solvent was removed in vacuo. The product was isolated by flash column
chromatography on silica gel, eluting with EtOAc to give the dendrimer 21
(122 mg, 34%) as an amorphous solid. 1H NMR (400 MHz, CD3SOCD3,
1008C): d ˆ 8.04 (d, J ˆ 1.6 Hz, 2H; Ar[core]), 7.97 (t, J ˆ 1.4 Hz, 8H; Ar),
7.84 (t, J ˆ 1.6 Hz, 1H; Ar[core]), 7.53 (d, J ˆ 1.4 Hz, 4H; Ar), 4.31 (t, J ˆ
10 Hz, 8H; CH2O), 3.88 (t, obscured; NCH2), 3.85 (s, 27H; CH3); 13C NMR
(100 MHz, CD3SOCD3, 1008C): d ˆ 169.4, 164.5, 158.0, 136.8, 131.2, 129.9,
129.0, 127.8, 124.3, 121.6, 118.8, 65.9, 51.6, 45.5; MS (FAB ‡ ): m/z (%) ˆ
mide (16): 15% NaOH (10 mL) was added to
a solution of the
phosphoramide 7 (1.083 g, 1.7 mmol) in THF (40 mL). The mixture was
stirred under reflux for 2 h and then it was allowed to cool down to room
temperature. The THF was removed in vacuo, after which the product was
precipitated by the addition of H2O (20 mL) and 10% H2SO4 (20 mL). The
precipitate was collected by filtration, and was washed with H2O (5 mL) to
give the phosphoramide 16 (989 mg, 100%) as a colorless solid, decomp.
>2508C. 1H NMR (300 MHz, CD3SOCD3, 258C): d ˆ 13.4 (br, 4H;
CO2 H), 8.06 (brt, 2H; Ar), 7.61 (d, J ˆ 1.1 Hz, 4H; Ar), 4.22 (br, 4H;
CH2O), 3.95 (m, 4H; CH2OP), 3.48 (brm, 4H, NCH2), 1.20 (t, J ˆ 7.2 Hz,
6H; CH3); 13C NMR (75 MHz, CD3SOCD3, 258C): d ˆ 166.7, 158.8, 133.5,
122.6, 119.0, 66.9, 61.7, 45.6, 16.2; 31P NMR (162 MHz, CD3SOCD3, 258C):
‡
1167 (85%) [M‡H] ; C58H58N2O24: calcd C 59.69, H 5.01, N 2.40; found C
59.6, H 4.91, N 2.25.
1,3,5-Tris{N,N-bis[2-(3,5-bis(hydroxycarbonyl)phenoxy)ethyl]aminocarbo-
nyl}benzene (22): A 7.5% aqueous solution of NaOH (1.5 mL) was added
to a stirred solution of the dendrimer 19 (665 mg, 0.41 mmol) in THF
(15 mL). The reaction mixture was stirred under reflux for 1 h, after which
time 15% aqueous NaOH solution (1.5 mL) was added. The reaction
mixture was stirred under reflux for a further 1 h and then the THF was
removed in vacuo. The product was precipitated by the addition of a 10%
aqueous solution (10 mL) of H2SO4. It was collected by filtration and then
washed with H2O (10 mL) and dried in a desiccator to give the dodecaacid
‡
d ˆ 10.51; MS (FAB ): m/z (%) ˆ 568 (100) [M H] ; C24H28NO13P:
calcd C 50.62, H 4.96, N 2.46; found C 50.48, H 4.94, N 2.26.
Bis{2-[3,5-bis(methoxycarbonyl)phenoxy]ethyl}amine (17): A solution of
the phosphoramide 7 (2.28 g, 3.6 mmol) in THF (50 mL) was cooled down
to below 58C. HCl gas was bubbled through the solution for 30 min, after
which time the cold bath was removed and the reaction mixture was left to
stand for 3 h. The volatiles were removed in vacuo, and then Et2O (50 mL)
was added. The crude HCl salt was isolated by filtration, and was washed
with Et2O (10 mL). The HCl salt was dissolved in H2O/MeOH (ca. 10 mL)
and the free amine was precipitated in pure form by the addition of 15%
NaOH solution (2.5 mL). The product was isolated by filtration, and then
washed with H2O (2 mL). Finally, the pure material was dried in a
desiccator to give the amine 17 (1.49 g, 84%) as a colorless solid, m.p. 131 ±
1328C. 1H NMR (300 MHz, CD3COCD3, 258C): d ˆ 8.61 (t, J ˆ 1.5 Hz,
2H; Ar), 8.18 (d, J ˆ 1.5 Hz, 4H; Ar), 4.69 (t, J ˆ 5.5 Hz, 4H; CH2OAr),
4.37 (s, 12H; CH3), 3.59 (t, J ˆ 5.3 Hz, 4H; NCH2); 13C NMR (75 MHz,
CD3COCD3, 258C): d ˆ 166.2, 160.2, 132.9, 132.0, 120.3, 69.5, 52.7, 49.1; MS
1
22 (630 mg, 100%) as a white powder. H NMR (400 MHz, CD3SOCD3,
908C): d ˆ 8.02 (6H; Ar[surface]), 7.74 (s, 3H; Ar[core]), 7.59 (12H;
Ar[surface]), 4.30 (brt, 12H; CH2O), 3.91 (brt, 12H; NCH2); 13C NMR
(100 MHz, CD3SOCD3, 908C): d ˆ 169.6, 165.6, 157.8, 136.5, 132.2, 125.8,
122.2, 118.7, 65.6; MS (FAB ): m/z (%) ˆ 1455 (100%) [M H] .
Bis{2-[3,5-bis(N,N-bis(2-(3,5-bis(methoxycarbonyl)phenoxy)ethyl)amino-
carbonyl)phenoxy]ethyl}amine (24): A solution of the phosphoramide 18
(5.188 g, 2.11 mmol) in THF (100 mL) was cooled down to below 58C. HCl
gas was bubbled through the solution for 30 min, after which time the cold
bath was removed and the reaction was left to stand for 2 h while the
reaction was monitored by TLC. The volatiles were removed in vacuo and
then Et2O (100 mL) was added. The Et2O was decanted and the solid was
dissolved in the minimum volume of Me2CO. The free amine was
precipitated by the addition of a 15% NaOH solution (2 mL) and H2O
(100 mL). The crude product was isolated by filtration, then purified by
flash column chromatography on silica gel, eluting with a gradient from
10% MeOH in EtOAc to 1:1:8 Et3N:MeOH:EtOAc to give the amine 24
‡
(FAB ‡ ): m/z (%) ˆ 490 (100) [M‡H] ; C24H27NO10: calcd C 61.00, H 6.26,
N 2.63; found C 61.26, H 6.27, 2.61.
N,N-Bis{2-[3,5-bis(N,N-bis(2-(3,5-bis(methoxycarbonyl)phenoxy)ethyl)-
aminocarbonyl)phenoxy]ethyl}(diethyl)phosphoramide (18): A solution of
the amine 17 (6.545 g, 13.4 mmol) and the phosphoramide 16 (1.901 g,
3.34 mmol) in THF (150 mL) was cooled down to 08C. DCC (2.85 g,
13.8 mmol) and HOBt (1.82 g, 13.5 mmol) were added, the ice-bath was
removed, and the reaction mixture was stirred for 4 h. The solvent was
removed in vacuo and the product was isolated by flash column
chromatography on silica gel, eluting with a gradient from EtOAc to
10% MeOH in EtOAc to give the phosphoramide 18 (7.60 g, 93%) as a
(3.52 g, 72%) as
a
colorless amorphous solid. 1H NMR (400 MHz,
CD3SOCD3, 1008C): d ˆ 7.95 (t, J ˆ 1.5 Hz, 8H; Ar[outer]), 7.52 (d, J ˆ
1.3 Hz, 16H; Ar[outer]), 7.17 (s, 2H; Ar[inner]), 7.06 (s, 4H; Ar[inner]),
4.29 (t, J ˆ 5.0 Hz, 16H; CH2O[outer]), 4.10 (t, J ˆ 5.6 Hz, 4H; CH2O[in-
ner]), 3.87 (obscured, NCH2[outer]), 3.84 (s, 48H; CH3), 3.01 (t, J ˆ 5.6 Hz,
4H; NCH2[inner]); 13C NMR (75 MHz, CD3COCD3, 258C): d ˆ 171.7,
165.9, 159.7, 159.2, 139.1, 132.4, 123.0, 119.8, 118.8, 115.0, 69.0, 67.4, 52.6,
1
colorless, amorphous solid. H NMR (400 MHz, CD3SOCD3, 1008C): d ˆ
7.94 (t, J ˆ 1.3 Hz, 8H; Ar), 7.50 (d, J ˆ 1.3 Hz, 16H; Ar), 7.19 (brt, 2H; Ar),
7.06 (d, J ˆ 1.2 Hz, 4H, Ar), 4.28 (t, J ˆ 5.1 Hz, 16H; CH2OAr), 4.18 (t, J ˆ
5.7 Hz, 4H; CH2OAr), 3.9 ± 3.8 (m, 68H; NCH2, CH2OP, CO2CH3), 3.44
(m, 4H; PNCH2), 1.14 (t, J ˆ 7.0 Hz, 6H; CH3); 13C NMR (75 MHz,
CD3SOCD3, 258C): d ˆ 171.6, 165.8, 159.4, 159.1, 139.2, 132.3, 123.0, 119.8,
118.9, 114.8, 62.4, 62.1, 52.5, 16.4, 16.3; 31P NMR (162 MHz, CD3SOCD3,
‡
50.1, 49.0, 46.2; MS (FAB ‡ ): m/z (%) ˆ 2320 (70%) [M‡H] ;
C108H119N5O46: calcd C 60.08, H 5.17, N 3.02; found C 60.54, H 5.18, N 3.16.
1,3,5-Tris{N,N-bis[2-(3,5-bis(N,N-bis(2-(3,5-bis(methoxycarbonyl)phen-
oxy)ethyl)aminocarbonyl)phenoxy)ethyl]aminocarbonyl}benzene
(26),
Method A: A solution of the amine 24 (142 mg, 0.061 mmol) in THF
(15 mL) was cooled down to < 58C. A solution of 1,3,5-benzenetricarbonyl
trichloride (5.0 mg, 0.053 mmol) in THF (160 mL) was added slowly using a
syringe, while the reaction mixture was stirred and allowed to warm up to
ambient temperature. After stirring for a further 1 h, the solvent was
removed in vacuo and the product was isolated by column chromatography
on silica gel, eluting with a gradient from EtOAc to 10% MeOH in EtOAc
to give the three-directional dendrimer 26 (91 mg, 63%) as a colorless
amorphous solid. 1H NMR (400 MHz, CD3SOCD3, 1008C): d ˆ 7.83 (t, J ˆ
1.3 Hz, 24H; Ar[outer]), 7.76 (s, 3H; Ar[core]), 7.40 (d, J ˆ 1.7 Hz, 48H;
Ar[outer]), 7.23 (brt, 6H; Ar[inner]), 7.10 (d, J ˆ 1.4 Hz, 12H; Ar[inner])
4.30 (brt, 12H; CH2O[inner]), 4.20 (brt, 48H; CH2O[outer]), 3.90 (brt,
12H; NCH2[inner]), 3.83 (brt, 48H; NCH2[outer]), 3.77 (s, 144H; CH3);
13C NMR (100 MHz, CD3SOCD3, 1008C): d ˆ 169.9, 169.7, 164.4, 157.8,
137.6, 136.4, 131.0, 127, 121.7, 121.5, 118.7, 117.5, 117.3, 113.8, 113.6, 66.5,
‡
258C): d ˆ 10.38; MS (FAB ‡ ): m/z (%) ˆ 2456 (100) [M‡H] ;
C120H128N5O49P: calcd C 50.70, H 5.25, N 2.85; found C 50.85, H 5.31, N 2.79.
1,3,5-Tris{N,N-bis[2-(3,5-bismethoxycarbonyl)phenoxy]ethyl}aminocarbo-
nylbenzene (19): A solution of the amine 17 (765 mg, 1.56 mmol) and
trimesic acid (109 mg, 0.52 mmol) in THF (75 mL) was cooled down to 08C.
DCC (322 mg, 1.56 mmol) and HOBt (199 mg, 1.47 mmol) were added, the
ice bath was removed, and the reaction mixture was stirred for 4 h. The
solvent was removed in vacuo and the product was isolated by flash column
chromatography on silica gel, eluting with a gradient from EtOAc to 10%
MeOH in EtOAc to give the dendrimer 19 (590 mg, 70%) as a colorless,
amorphous solid. 1H NMR (400 MHz, CD3SOCD3, 1008C): d ˆ 7.92 (t, J ˆ
1.2 Hz, 6H; Ar), 7.74 (s, 3H; core Ar), 7.49 (d, J ˆ 1.3 Hz, 12H; Ar), 4.29 (t,
J ˆ 5.1 Hz, 12H; CH2OAr), 3.91 (t, J ˆ 5.1 Hz, 12H, NCH2), 3.84 (s, 36H;
CH3); 13C NMR (100 MHz, CD3SOCD3, 1008C): d ˆ 169.7, 164.5, 158.0,
136.6, 131.2, 125.9, 121.7, 118.8, 66.1, 51.6, 46.7; MS (FAB ‡ ): m/z (%) ˆ
‡
66.1, 65.7, 52.2, 51.5, 50.8, 46.6; MS (FAB ‡ ): m/z (%) ˆ 7113 (3) [M‡H] ;
‡
‡
1414 (96) [M OPh(CO2Me)2] , 1624 (100) [M‡H] ; C81H81N3O33: calcd C
C357H357N15O141: calcd C 60.28, H 5.06, N 2.95; found C 59.97, H 5.20, N 2.85.
59.89, H 5.03, N 2.59; found C 59.84, H 5.17, N 2.63.
Dendrimer 26 (Method B) and 1-methoxycarbonyl-3,5-tris{N,N-bis[2-(3,5-
bis(N,N-bis(2-(3,5-bis(methoxycarbonyl)phenoxy)ethyl)aminocarbonyl)-
phenoxy)ethyl]aminocarbonyl}benzene (25): A solution of the amine 24
(1.09 g, 0.47 mmol) and trimesic acid (32.7 mg, 0.156 mmol) in THF
(20 mL) was cooled down to 08C. DCC (99 mg, 0.48 mmol) and HOBt
(74 mg, 0.48 mmol) were added, the ice-bath was removed, and the reaction
was stirred for 12 h. The solvent was removed in vacuo and then the
1-Methoxycarbonyl-3,5-bis{N,N-bis[2-(3,5-bis(methoxycarbonyl)phen-
oxy)ethyl]aminocarbonyl}benzene (21):
A solution of the amine 17
(452 mg, 0.92 mmol) and trimesic acid (64 mg, 0.30 mmol) in MeCN
(20 mL) was cooled down to 08C. DCC (196 mg, 0.95 mmol), HOBt
(128 mg, 0.95 mmol), and THF (5 mL) were added and the ice-bath was
removed. The reaction mixture was stirred for 3 h, after which time the
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0793 $ 17.50+.25/0
793

J. F. Stoddart et al.
FULL PAPER
products were isolated by flash column chromatography on silica gel,
eluting with a gradient from EtOAc to 10% MeOH in EtOAc to give the
three-directional dendrimer 26 (144 mg, 13%) and the two-directional
dendrimer 25 (237 mg, 32%) as a colorless, amorphous solid. 1H NMR
(400 MHz, CD3SOCD3, 1008C): d ˆ 8.05 (d, J ˆ 1.5 Hz, 2H; Ar[core]), 7.88
(t, J ˆ 1.3 Hz, 16H; Ar[outer]), 7.79 (t, J ˆ 1.3 Hz, 1H; Ar[core]), 7.45 (d,
J ˆ 1.4 Hz, 32H; Ar[outer]), 7.21 (brt, 4H; Ar[inner]), 7.09 (d, J ˆ 1.2 Hz,
8H; Ar[inner]) 4.25 (m, 40H; CH2O[inner‡ outer]), 3.85 (m, 40H;
NCH2[inner‡ outer]), 3.80 (s, 99H; CH3[core ‡ surface]); 13C NMR
(100 MHz, CD3SOCD3, 908C): d ˆ 169.9, 169.3, 164.4, 157.8, 137.6, 136.8,
131.0, 129.5, 127.7, 121.5, 118.7, 117.4, 113.6, 66.0, 65.4, 51.5; MS (FAB ‡ ):
[1] A search of the Science Citation Index with the keyword dendrimer
reveals hits for the last five years as follows: 1991, 3; 1992, 11; 1993, 19;
1994, 30; 1995, 36; 1996, 60. For recent reviews concerning dendrim-
ers, see: a) H. B. Mekelburger, W. Jaworek, F. Vögtle, Angew. Chem.
1992, 104, 1609 ± 1614; Angew. Chem. Int. Ed. Engl. 1992, 31, 1571 ±
1576; b) G. R. Newkome, C. N. Moorefield, G. R. Baker, Aldrichimica
Acta 1992, 25, 31 ± 38; c) D. A. Tomalia, Adv. Mater. 1994, 6, 529 ± 539;
d) G. R. Newkome, C. N. Moorefield in Comprehensive Supramolec-
ular Chemistry, Vol. 10 (Ed.: D. N. Reinhoudt), Pergamon, Oxford,
Â
UK, 1996, 777 ± 832; e) J. M. J. Frechet, C. J. Hawker in Comprehen-
sive Polym. Sci. (Eds.: S. L. Aggarwal, S. Russo), 1996, 140 ± 201;
f) G. R. Newkome, C. N. Moorefield, F. Vögtle in Dendritic Molecules,
VCH, Weinheim, Germany, 1996; g) O. A. Matthews, A. N. Shipway,
J. F. Stoddart, Prog. Polym. Sci. 1998, 23, 1 ± 56.
‡
m/z (%) ˆ 4827 (27) [M‡H] ; C242H242N10O96: calcd C 60.22, H 5.05, N
2.90; found C 60.30, H 5.19, N 2.88.
1,3,5-Tris{N,N-bis[2-(3,5-bis(N,N-bis(2-(3,5-bis(N,N-bis(2-(3,5-bis(meth-
oxycarbonyl)phenoxy)ethyl)aminocarbonyl)phenoxy)ethyl)aminocarbo-
nyl)phenoxy)ethyl]aminocarbonyl}benzene (27): A solution of the amine
24 (1.09 g, 0.47 mmol) and the dodecaacid 22 (56.2 mg, 0.039 mmol) in THF
(25 mL) was cooled down to 08C. DCC (110 mg, 0.53 mmol) and HOBt
(77 mg, 0.50 mmol) were added, the ice-bath was removed, and the reaction
mixture was stirred overnight. More DCC (90 mg, 0.44 mmol) was added and
the reaction mixture was stirred for a further day. The solvent was removed in
vacuo to give a crude mixture. Quantitative GPC of an aliquot of this mixture
showed that it contained products with a narrow hydrodynamic radius
range of around 2.4 nm and a yield of 32% with respect to the wedge.
[2] D. A. OꢁSullivan, Chem. Eng. News 1993, August 16, 20 ± 23; the
PAMAM dendrimers are commercially available through Aldrich.
[3] H. Brunner, J. Organomet. Chem. 1995, 500, 39 ± 46.
[4] a) J. R. Robinson, G. M. Mlynek, Adv. Drug Deliv. Rev. 1995, 16, 45 ±
50; b) A. E. Beezer, J. C. Mitchell, R. M. Colegate, D. J. Scally, L. J.
Twyman, R. J. Willson, Thermochim. Acta. 1995, 277 ± 283; c) K.
Â
Sheldon, D. Liu, J. Ferguson, J. Gariepy, Proc. Natl. Acad. Sci. USA
1995, 92, 2056 ± 2060.
[5] a) D. A. Tomalia, P. R. Dvornic, Nature 1994, 372, 617 ± 618; b) J.
Alper, Science 1991, 251, 1562 ± 1564.
[6] a) D. A. Tomalia, M. Hall, D. M. Hedstrand, J. Am. Chem. Soc. 1987,
109, 1601 ± 1603; b) A. M. Naylor, W. A. Goddard III, G. E. Kiefer,
D. A. Tomalia, ibid. 1989, 111, 2339 ± 2341.
N,N-Bis{2-[3,5-bis(methoxycarbonyl)phenoxy]ethyl}benzamide (29):
A
solution of the amine 17 (266 mg, 0.544 mmol) and benzoic acid (72 mg,
0.59 mmol) in THF (20 mL) was cooled down to 08C. DCC (122 mg,
0.59 mmol) and HOBt (81 mg, 0.60 mmol) were added, the ice-bath was
removed, and the reaction mixture was stirred for 5 h. The solvent was
removed in vacuo and the product was isolated by flash column
chromatography on silica gel, eluting with a gradient from EtOAc to
10% MeOH in EtOAc to give the amide 29 (252 mg, 78%) as a colorless,
amorphous solid. 1H NMR (300 MHz, CD3COCD3, 258C): d ˆ 8.10 (t, J ˆ
3 Hz, 2H; Ar[branching unit]), 7.63 (d, J ˆ 3 Hz, 4H; Ar[branching unit]),
7.55 (m, 2H; Ar[core]), 7.45 (m, 3H; Ar[core]), 4.40 (br, 4H; CH2O), 4.06
(br, 4H; NCH2), 3.90 (s, 12H; CH3); 13C NMR (100 MHz, CD3SOCD3,
908C): d ˆ 171.3, 164.8, 158.2, 136.3, 131.4, 128.8, 127.8, 126.4, 121.7, 119.0,
[7] a) J. F. G. A. Jansen, R. A. J. Janssen, E. M. M. de Brabander-van -
den Berg, E. W. Meijer, Adv. Mater. 1995, 7, 561 ± 564; b) K. L.
Â
Wooley, C. J. Hawker, J. M. Pochan, J. M. J. Frechet, Macromolecules
1993, 26, 1514 ± 1519; c) G. R. Newkome, C. N. Moorefield, J. M.
Keith, G. R. Baker, G. H. Escamilla, Angew. Chem. 1994, 106, 701 ±
703; Angew. Chem. Int. Ed. Engl. 1994, 33, 666 ± 668; d) G. R.
Newkome, C. N. Moorefield, G. R. Baker, R. K. Behera, G. H.
Escamilla, M. J. Sanders, ibid. 1992, 104, 901 ± 903 and 1992, 31,
917 ± 919; e) D. A. Tomalia, A. M. Naylor, W. A. Goddard III, ibid.
1990, 102, 119 ± 157 and 1990, 29, 138 ± 175.
[8] R. F. Service, Science 1995, 267, 458 ± 459.
‡
66.3, 51.8; MS (FAB ‡ ): m/z (%) ˆ 594 (43%) [M‡H] ; C31H31NO11: calcd
[9] G. R. Newkome, C. N. Moorefield, G. R. Baker, M. J. Saunders, S. H.
Grossman, Angew. Chem. 1991, 103, 1207 ± 1209; Angew. Chem. Int.
Ed. Engl. 1991, 30, 1178 ± 1180.
[10] M. F. Ottaviani, S. Bossmann, N. J. Turro, D. A. Tomalia, J. Am. Chem.
Soc. 1994, 116, 661 ± 671.
C 62.73, H 5.26, N 2.36; found C 62.89, H 5.19, N 2.19.
N,N-Bis{2-[3,5-bis(N,N-bis(2-(3,5-bis(methoxycarbonyl)phenoxy)ethyl)-
aminocarbonyl)phenoxy]ethyl}aminocarbonylbenzamide (30): A solution
of the amine 24 (126 mg, 0.054 mmol) and benzoic acid (7.3 mg,
0.060 mmol) in THF (2 mL) was cooled down to 08C. DCC (12.7 mg,
0.62 mmol) and HOBt (8.1 mg, 0.060 mmol) were added, the ice-bath was
removed, and the reaction mixture was stirred overnight. The solvent was
removed in vacuo and then the product was isolated by flash column
chromatography on silica gel, eluting with a gradient from EtOAc to 10%
MeOH in EtOAc to give the amide 30 (125 mg, 95%) as a colorless,
amorphous solid. 1H NMR (400 MHz, CD3SOCD3, 1008C): d ˆ 7.94 (t, J ˆ
1.4 Hz, 8H; Ar[outer]), 7.50 (d, J ˆ 1.5 Hz, 16H; Ar[outer]), 7.29 (m, 5H;
Ar[core]), 7.20 (t, J ˆ 1.2 Hz, 2H; Ar[inner]), 7.06 (d, J ˆ 1.2 Hz, 4H;
Ar[inner]), 4.28 (t, J ˆ 5.1 Hz, 16H; CH2O[outer]), 4.24 (t, J ˆ 5.6 Hz, 4H;
CH2O[inner]), 3.88 (m, 20H; NCH2[inner‡ outer]), 3.83 (s, 48H; CH3);
13C NMR (100 MHz, CD3SOCD3, 1008C): d ˆ 170.0, 164.6, 158.0, 137.8,
136.2, 131.2, 128.5, 127.6, 126.1, 124.3, 121.7, 118.8, 117.4, 113.7, 66.1, 65.7,
[11] D. A. Tomalia, H. Baker, J. Dewald, M. Hall, G. Kallos, S. Martin, J.
Roeck, J. Ryder, P. Smith, Polym. J. 1985, 17, 117 ± 132.
Â
[12] C. J. Hawker, K. L. Wooley, J. M. J. Frechet, J. Am. Chem. Soc. 1993,
115, 4375 ± 4376.
[13] M. L. Mansfield, L. I. Klushin, Macromolecules 1993, 26, 4262 ± 4268.
Â
[14] C. J. Hawker, J. M. J. Frechet, J. Am. Chem. Soc. 1990, 112, 7638 ±
7647.
[15] Many dendrimer syntheses rely on the so-called divergent approach
and date back to the seminal paper demonstrating the feasibility of
dendrimer synthesis. See: E. Buhleier, W. Wehner, F. Vögtle, Synthesis
1978, 155 ± 158.
[16] M. L. Mansfield, Macromolecules 1993, 26, 3811 ± 3814.
Â
[17] K. L. Wooley, C. J. Hawker, J. M. J. Frechet, J. Chem. Soc. Perkin
‡
51.6, 46.1 (br); MS (FAB ‡ ): m/z (%) ˆ 2446 (100%) [M‡H] ;
Trans. 1 1991, 1059 ± 1076.
C123H123N5O47: calcd C 60.96, H 5.12, N 2.89; found C 61.08, H 5.17, N 2.93.
Â
[18] J. M. J. Frechet, C. J. Hawker, K. L. Wooley, J. Macromol. Sci. Pure
Appl. Chem. 1994, A31, 1627 ± 1645.
1,3,5-Tris(methoxycarbonyl)benzene (31):
A solution of trimesic acid
[19] a) S. Campagna, G. Denti, S. Serroni, A. Juris, M. Venturi, V.
Ricevuto, V. Balzani, Chem. Eur. J. 1995, 1, 211 ± 220; b) H.-F. Chow,
C. C. Mak, J. Chem. Soc. Perkin Trans. 1 1994, 2223 ± 2228; c) N.
Launay, A.-M. Caminade, J.-P. Majoral, J. Am. Chem. Soc. 1995, 117,
3282 ± 3283; d) A. Rajca, S. Utamapanya, ibid. 1993, 115, 10688 ±
10694; e) S. Achar, R. J. Puddephatt, J. Chem .Soc. Chem. Commun.
1994, 1895 ± 1896; f) R. H. E. Hudson, M. J. Damha, J. Am. Chem. Soc.
1993, 115, 2119 ± 2124; g) D. Seyferth, D. Y. Son, A. L. Rheingold,
R. L. Ostrander, Organometallics 1994, 13, 2682 ± 2690.
(24.5 g, 97 mmol) and H2SO4 (25 mL) in MeOH (250 mL) was stirred
under reflux for 18 h. The reaction mixture was cooled down to 58C and
H2O (150 mL) was added. The precipitated product was isolated by
filtration and was washed with H2O (50 mL) to give the pure triester 31
(28.5 g, 97%) as a colorless solid, m.p. 145 ± 1468C; Lit. 145 ± 1478C
(Aldrich). 1H NMR (300 MHz, CD3COCD3, 258C): d ˆ 8.34 (s, 3H; Ar),
3.98 (s, 9H; CH3).
Acknowledgments: This research was supported by the Chemical and
Biological Defence Establishment of the Defence Evaluation and Re-
search Agency of the Ministry of Defence in the United Kingdom.
Â
[20] K. L. Wooley, C. J. Hawker, J. M. J. Frechet, J. Am. Chem. Soc. 1991,
113, 4252 ± 4261.
Â
[21] K. L. Wooley, C. J. Hawker, J. M. J. Frechet, Angew. Chem. Int. 1994,
106, 123 ± 126; Angew. Chem. Int. Ed. Engl. 1994, 33, 82 ± 85.
Received: September 15, 1997 [F821]
794
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0794 $ 17.50+.50/0
Chem. Eur. J. 1998, 4, No. 5

Polyamide Dendrimers
781±795
[22] R. Spindler, J. Chem. Soc. Perkin Trans. 1 1993, 913 ± 918.
[23] T. Kawaguchi, K. L. Walker, C. L. Wilkins, J. S. Moore, J. Am. Chem.
Soc. 1995, 117, 2159 ± 2165.
area) of its second-generation analogue. The transformation from a
free to a globular dendrimer conformation is usually associated with
an increase in the end group:surface area ratio.
Â
[24] For an exception to this rule, see: R. Spindler, J. M. J. Frechet, J.
[48] The calibration data was obtained by examining a set of nine low
polydispersity polystyrene standards and the average hydrodynamic
radii of the samples was calculated from the equation d ˆ (240/pN)1/3
(M[h])1/3, where d is the average hydrodynamic diameter, M is the
molecular weight, [h] is the intrinsic viscosity of the analyte, and N is a
constant. Standard GPC calibration curves for molecular size (R. D.
Hester, P. H. Mitchell, J. Polym. Sci. Chem. Ed. 1980, 18, 1727 ± 1738)
were found to be unsatisfactory. This observation may be a conse-
quence of a GPC system consisting of two columns of different pore
sizes, linked in series. An excellent fit (R2 ˆ 0.999) was found for the
relationship rˆ aVeb (without a theoretical basis), where r is the
average hydrodynamic radius, Ve is the elution volume, and a and b are
constants.
Chem. Soc. Perkin Trans. 1 1993, 913 ± 918.
Â
[25] C. J. Hawker, J. M. J. Frechet, J. Chem. Soc. Perkin Trans. 1 1992,
2459 ± 2469.
[26] a) H.-F. Chow, C. C. Mak, J. Chem. Soc. Perkin Trans. 1 1994, 2223 ±
2228; b) S. Achar, R. J. Puddephatt, Organometallics 1995, 14, 1681 ±
1687; c) D. Seyferth, D. Y. Son, ibid. 1994, 13, 2682 ± 2690; d) K.
Rengan, R. Engel, J. Chem. Soc. Perkin Trans. 1 1991, 987 ± 990.
[27] a) E. M. M. de Brabander-van den Berg, E. W. Meijer, Angew. Chem.
1993, 105, 1370 ± 1372; Angew. Chem. Int. Ed. Engl. 1993, 32, 1308 ±
1311; b) N. Launay, A. M. Caminade, J.-P. Majoral, J. Am. Chem. Soc.
1995, 117, 3282 ± 3283.
[28] a) G. R. Newkome, V. V. Narayanan, A. K. Patri, J. Gross, C. N.
Moorefield, G. R. Baker, Proc. ACS Division of Polymeric Materials:
Sci. Eng. 1995, 73, 222 ± 223; b) A. Rajca, S. Utamapanya, J. Am.
Chem. Soc. 1993, 115, 10688 ± 10694; c) T. Nagasaki, O. Kimura, M.
Ukon, S. Arimori, I. Hamachi, S. Shinkai, J. Chem. Soc. Perkin Trans. 1
1994, 75 ± 81; d) H. Uchida, Y. Kabe, K. Yoshino, A. Kawamata, T.
Tsumuraya, S. Masumune, J. Am. Chem. Soc. 1990, 112, 7077 ± 7079.
[49] D. A. Tomalia, V. B. M. Hall, D. M. Hedstrand, Macromolecules 1987,
20, 1167 ± 1169.
[50] The kinetic data were obtained by the coalescence method, where
values of the rate constant (kc) at the coalescence temperature (Tc)
were obtained (I. O. Sutherland, Annu. Rep. NMR Spectrosc. 1971, 4,
71 ± 235) from the approximate expression kc ˆ p(DnÄ)/(2)1/2, where DnÄ
is the limiting chemical shift difference (in Hz) between the exchang-
ing proton resonances. The Eyring equation was then used to calculate
Â
[29] K. L. Wooley, C. J. Hawker, J. M. J. Frechet, Angew. Chem. Int. Ed.
Engl. 1994, 33, 82 ± 85.
ˆ
[30] a) T. Kawaguchi, K. L. Walker, C. L. Wilkins, J. S. Moore, J. Am.
Chem. Soc. 1995, 117, 2159 ± 2165; b) A. Miedaner, C. J. Curtis, R. M.
Barkley, D. L. DuBois, Inorg. Chem. 1994, 33, 5482 ± 5490.
DGc from kc at Tc.
[51] a) P. R. Ashton, S. E. Boyd, C. L. Brown, N. Jayaraman, S. A.
Nepogodiev, J. F. Stoddart, Chem. Eur. J. 1996, 2, 1115 ± 1128; b) M.
Murat, G. S. Grest, Macromolecules 1996, 29, 1278 ± 1285.
Â
[31] C. J. Hawker, J. M. J. Frechet, J. Am. Chem. Soc. 1992, 114, 8405±8413.
[32] E. H. Gold, E. Babad, J. Org. Chem. 1972, 37, 2208 ± 2210.
[33] a) E. Amble, J. Dale, Acta Chem. Scand. 1979, 33, 698 ± 700; b) P. J.
Maurer, H. Takahata, H. Rapoport, J. Am. Chem. Soc. 1984, 106,
1095 ± 1098; c) T. Hamada, J. Am. Chem. Soc. 1986, 108, 140 ± 145;
d) W. Yuan, K. Fearon, M. H. Gelb, J. Org. Chem. 1989, 54, 906 ± 910.
[34] a) S. J. Blarer, D. Seebach, Chem. Ber. 1983, 116, 2250; b) T. P. Curran,
M. P. Pollastri, S. M. Abelleira, R. J. Messier, T. A. McCollum, C. G.
Rowe, Tetrahedron Lett. 1994, 35, 5409 ± 5412.
[35] L. Qian, Z. Sun, M. P. Mertes, K. B. Mertes, J. Org. Chem. 1991, 56,
4904 ± 4907.
[36] A. Zwierak, K. Osowska, Synthesis 1984, 223 ± 224.
[37] a) S. K. Banerjee, B. D. Gupta, K. Singh, J. Chem. Soc. Chem.
Commun. 1982, 815 ± 816; b) H. E. Zaugg, J. Org. Chem. 1976, 41,
3419 ± 3421.
[52] Molecular simulations were carried out using the AMBER forcefield,
as implemented in Macromodel (V.5.0),[53] running on a Silicon
Graphics Indigo2 Workstation. The dendrons were assembled within
the Macromodel INPUT submode and then fully minimized (final
gradient <0.5 kJŠ 1) by means of the Polak Ribiere Conjugate
Gradient (PRCG) algorithm with extended cut-offs. Solvation was
included in the form of the GB/SA solvation model for CHCl3. The
individual dendron units were then attached to the central core and
the angle between the dendron and the core was adjusted manually to
minimize steric clashes between separate dendrons. The whole
assembly was then fully minimized by the above method (final
gradient <0.5 kJŠ 1). Simulated annealing, using stochastic molec-
ular dynamics (SD) was performed within Macromodel. The system
was allowed 20 ps of equilibration time at 300 K (timestep 1.5 fs),
before being cooled to 50 K over 100 ps to afford a structure different
from the starting geometry. This structure was then fully minimized
[38] A. N. Shipway, J. F. Stoddart, unpublished results.
[39] R. G. Denkewalter, J. Kolc, W. J. Luskasavage, US Patent 4289872,
1981.
[40] L. J. Twyman, A. E. Beezer, J. C. Mitchell, Tetrahedron Lett. 1994, 35,
4423 ± 4424.
(AMBER*, PRCG, extended cut-offs, GB/SA solvation) until the
1
RMS deviation was <0.5 kcalŠ
.
Structures used to describe
conformations of the dendrimers at 300 K were taken from the end
of the 20 ps equilibration period. Radii of the dendrimers were
measured from within CSC Chem3D pro by the averaging of about
ten core-to-surface distances.
[41] T. Nagasaki, O. Kimura, M. Ukon, I. Hamachi, S. Shinkai, J. Chem.
Soc. Perkin Trans. 1 1994, 75 ± 81.
Â
[42] K. E. Uhrich, J. M. J. Frechet, J. Chem. Soc. Perkin Trans. 1 1992,
1623 ± 1630.
[53] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M. Lipton,
C. Caufield, G. Chang, T. Hendrickson, W. C. Still, J. Comp. Chem.
1990, 11, 440 ± 467.
[54] The sixth-generation contains 63 double-branched building blocks, 21
times the number (3) in the second-generation dendrimer. The sixth-
generation dendrimer has three times the radius of the second-
generation dendrimer, however, giving it 27 times the available
volume in which to accommodate only 21 times the number of
moieties.
[43] P. J. Dandliker, F. Diederich, M. Gross, C. B. Knobler, A. Louati,
E. M. Sanford, Angew. Chem. 1994, 106, 1821 ± 1824; Angew. Chem.
Int. Ed. Engl. 1994, 33, 1739 ± 1742.
[44] An aliquot (39 mg) of the crude reaction mixture, which contained
34 mg of dendritic materials with respect to the starting amine, was
estimated by GPC to contain 10.9 mg of sixth-generation products.
[45] a) B. L. Schwartz, A. L. Rockwood, R. D. Smith, D. A. Tomalia, R.
Spindler, Rapid Commun. Mass Spectrom. 1995, 9, 1552 ± 1555; b) P. R.
Dvornic, D. A. Tomalia, Macromol. Symp. 1995, 98, 403 ± 428; c) J. W.
[55] The ninhydrin reagent was prepared with ninhydrin (0.3 g) and glacial
acetic acid (3 mL) in n-butanol (100 mL).
Â
Leon, J. M. J. Frechet, Polym. Bull. 1995, 35, 449 ± 455.
[46] M. J. Maciejewski, J. Macromol. Sci. Chem. 1982, A17(4), 689 ± 703.
[47] The fourth-generation dendrimer carries four times the number of end
groups and has twice the arm length (hence four times the surface
[56] The cerium reagent was prepared with phosphomolybdic acid (2.5 g),
cerium(iv) sulfate (1.0 g), and sulfuric acid (6 mL) in water
(100 mL).
Chem. Eur. J. 1998, 4, No. 5
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
0947-6539/98/0405-0795 $ 17.50+.25/0
795