Heterometal assembly in dendritic polyphenylazomethines

Article abstract of DOI:10.1246/bcsj.80.1563

Fine-controlled metal-organic hybridization is strongly desired for advanced nano-materials. Dendrimers are highly branched organic macromolecules with successive layers or generations of branch units surrounding a central core. Metal-organic hybrid versions have also been produced by trapping metal ions or metal clusters within the voids of dendrimers. Here, we report a technique to form a dendritic organic structure, which includes various metals, with fine control over the number and location of the metal salts through a radial stepwise complexation. FeCl₃, GaCl₃,VCl ₃, and SnCl₂ were decorated on a phenylazomethine dendrimer with 4th generations at the first, second, third, and forth layers, respectively. These stepwise complexation and heterometal assembling were also supported by TEM observation, ⁵⁷Fe Moessbauer spectroscopy, XPS measurement, and NMR spectroscopy.

Full text of DOI:10.1246/bcsj.80.1563

ꢀ 2007 The Chemical Society of Japan
Bull. Chem. Soc. Jpn. Vol. 80, No. 8, 1563–1572 (2007) 1563
Heterometal Assembly in Dendritic Polyphenylazomethines
Kensaku Takanashi, Atsunobu Fujii, Reina Nakajima, Hiroshi Chiba,
ꢀ
Masayoshi Higuchi, Yasuaki Einaga, and Kimihisa Yamamoto
Department of Chemistry, Faculty of Science and Technology, Keio University,
3-14-1 Hiyoshi, Kouhoku-ku, Yokohama 223-8522
Received January 12, 2007; E-mail: yamamoto@chem.keio.ac.jp
Fine-controlled metal–organic hybridization is strongly desired for advanced nano-materials. Dendrimers are
highly branched organic macromolecules with successive layers or ‘‘generations’’ of branch units surrounding a central
core. Metal–organic hybrid versions have also been produced by trapping metal ions or metal clusters within the voids of
dendrimers. Here, we report a technique to form a dendritic organic structure, which includes various metals, with fine
control over the number and location of the metal salts through a radial stepwise complexation. FeCl3, GaCl3, VCl3, and
SnCl2 were decorated on a phenylazomethine dendrimer with 4th generations at the first, second, third, and forth layers,
respectively. These stepwise complexation and heterometal assembling were also supported by TEM observation, 57Fe
Mossbauer spectroscopy, XPS measurement, and NMR spectroscopy.
¨
Recently, with the development of nanotechnology, fine-
controlled metal–organic hybridization is strongly desired for
advanced nano-materials. For instance, in the research of metal
complexes, a catalyst system in which the metal is precisely
arranged has been examined. One of the ultimate shapes for
this system is a protein in which the number and location of
metal ions are precisely controlled.1–3
(a)
N
N
N
N
N
N
N
N
N
N
N
N
N
The use of dendrimers4 has attracted much attention for
catalyst5 and cluster synthesis.6 Dendrimers have a single mo-
lecular weight and three-dimensional spreading structure, in
which the density of the branches increases radially outward.
Due to these structural properties,4,7 dendrimers are expected
to be useful as templates for the generation of clusters with nar-
row dispersity or fine-controlled macromolecular complexes.
The present authors have reported the stepwise complexa-
tion of dendritic polyphenylazomethine (DPA, Fig. 1) with
Sn8,9 and Fe10,11 ions by exploiting the basicity gradient of
the constituent imines. It has been demonstrated that, through
stepwise radial complexation, it is possible to control the posi-
tion of complexation and the order of release in the molecule9
and the complexation order can be controlled by the substitu-
ents chosen for the core phenyl.12
N
N
N
2
3
4
1
N
N
N
N
N
N
N
N
N
N
N
N
N
N
(b)
4 equiv.
2 equiv.
metal salt:
Based on this stepwise complexation, several kinds of metal
ions might be decorated on the molecules with their number
and location precisely controlled. The precise control of several
kinds of metals on the molecules is expected to have novel
effects on the catalysis system and synthesis of the heterometal
clusters.
DPA G4
8 equiv.
16 equiv.
Fig. 1. (a) Dendritic polyphenylazomethine G4 (DPA G4).
(b) Schematic representation of stepwise complexation of
DPA G4 with metal salts.
We now report the stepwise complexation of DPA with
GaCl3 and VCl3. Further, we demonstrate the selective layer-
ing of 2–4 kinds of metals on DPA.
Results and Discussion
UV–vis titration and layer-selective reduction.8–10 In the in-
vestigation of other metal salts that complexed with DPAs,
GaCl3, and VCl3 were identified for further analysis.
Stepwise Complexation of DPA with GaCl3 and VCl3.
The stepwise complexation of DPA with SnCl2 and FeCl3 was
confirmed based on the shift in the isosbestic points during
The titrations of DPA G4 with metals were investigated by
Published on the web August 10, 2007; doi:10.1246/bcsj.80.1563

1564 Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007)
Heterometal Assembly in Dendritic Polyphenylazomethines
(b)
(b)
2 VCl3
4 VCl3
2 GaCl3
4 GaCl3
366 nm
367 nm
VCl3:
362 nm
370 nm
GaCl3:
8 VCl3
16 VCl3
358 nm
8 GaCl3
364 nm
16 GaCl3
362 nm
359 nm
Fig. 3. (a) UV–vis titration of DPA G4 (5:0 ꢁ 10ꢂ6 M,
acetonitrile/chloroform = 1:1) with VCl3 (5:0 ꢁ 10ꢂ3 M,
acetonitrile) showing the detailed spectra around the
isosbestic points (inset). (b) Schematic representation of
stepwise complexation of DPA G4 with VCl3.
Fig. 2. (a) UV–vis titration of DPA G4 (5:0 ꢁ 10ꢂ6 M,
acetonitrile/chloroform = 1:1) with GaCl3 (5:0 ꢁ 10ꢂ3 M,
acetonitrile) showing the detailed spectra around the
isosbestic points (inset). (b) Schematic representation of
stepwise complexation of DPA G4 with GaCl3.
UV–vis spectroscopy. In the case of GaCl3 (Fig. 2), the color
of the DPA solution (5:0 ꢁ 10ꢂ6 M, acetonitrile/chloro-
form = 1:1) changed with the addition of GaCl3 (5:0 ꢁ
10ꢂ3 M, acetonitrile) based on the complexation of imine
and metal. With the addition of 0–2 equiv of GaCl3, the isos-
bestic point was identified at 370 nm. However, the isosbestic
points shifted to lower values with further addition of GaCl3 to
366 nm at 2–6 equiv, 364 nm at 6–14 equiv, and 362 nm at 14–
30 equiv. The observation of four discrete isosbestic points
suggests that four different types of complexation occurred
in this process. This behavior is similar to that for SnCl2 and
FeCl3.8–10 The equivalent amounts for each step were 2, 4, 8,
and 16, corresponding to the number of each layer of imine
sites. This demonstrates that GaCl3 can complex with DPA
in the same stepwise fashion as SnCl2 and FeCl3. This step-
wise complexation was also observed for DPA G2 and G3
(Fig. S1), where two and three isosbestic points were observed
during the titration, respectively.
Further, in the case of VCl3 (Fig. 3), the isosbestic point
shifted by 4 points (367 nm at 0–2 equiv, 362 nm at 2–6 equiv,
359 nm at 6–14 equiv, 358 nm at 14–30 equiv). The agreement
of the equivalent amounts for each step and the number of each
layer of imine sites shows a stepwise radial complexation
from the inner to outer layers. These were also observed
for DPA G2 (Fig. S2). Moreover, the Job-plot13–15 was exam-
ined for GaCl3 with imine on DPA G0 (Fig. S3). The maxi-
mum value in the Job plot appeared at a mole fraction of 0.5
indicating the formation of a 1:1 complex. This result supports
quantitative complexation under titration conditions.
The complexation abilities of four metal salts, that is, FeCl3,
GaCl3, VCl3, and SnCl2, were evaluated in a 1:1 acetonitrile/
tetrahydrofuran16 solution of DPA G1 using curve-fitting of a
theoretical simulation for the titration17 (Fig. 4). The results
indicated that the order of complexation ability was as follows:
FeCl3 > GaCl3 > VCl3 > SnCl2. FeCl3 showed the largest
equilibrium constant for the complexation (K) between metal
salt and imine, which was determined to be 105 Mꢂ1. SnCl2
showed the smallest one (K ¼ 103 Mꢂ1). The difference in K
was found to be an important factor in heterometal assembly.
Heterometal Assembling on DPA.
The above results
demonstrate the stepwise radial complexation of DPA with
GaCl3 and VCl3. The complexation ability decreased in the
order FeCl3 > GaCl3 > VCl3 > SnCl2. As an extension of
the previous work on the assembly of a single metal salt on
DPA, the assembly of two kinds of metals was attempted.
DPA undergoes stepwise complexation with metals by com-
plexing with imines sequentially from the inner layer to the
outer layer following the gradients of electric charge densi-
ty.8–10 The coordination of different metal salts will occur in
a sequence determined by the coordination constants between
the metal salts and the imine layers. First, layer-selective het-
erometal assembly18–20 on DPAs was examined in this study

K. Takanashi et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007) 1565
with a combination of FeCl3 and SnCl2, which have different
coordination constants (KFeCl3 ¼ 105 Mꢂ1, KSnCl2 ¼ 103 Mꢂ1).
We first performed the UV–vis titration by the addition of
FeCl3 followed by SnCl2 using DPA G4 (5:0 ꢁ 10ꢂ6 M, aceto-
nitrile/chloroform = 1:1, Fig. 5).10 After adding 6 equiv of
FeCl3, the isosbestic points were observed at 371 nm for 0–2
equiv of FeCl3 and 369 nm for 2–6 equiv of FeCl3, which cor-
respond to the complexation of 1–2 layers. After adding SnCl2,
the isosbestic points shifted to 364 nm for 0–8 equiv of SnCl2
and to 363 nm for 8–24 equiv of SnCl2, which correspond to
the complexation of 3–4 layers (Figs. 5a and 5b-(2)). Similar-
ly, adding 2 equiv of FeCl3 followed by 28 equiv of SnCl2, the
isosbestic points shifted an equivalent amount for each step,
corresponding to the number of 2–4 layers of imine sites for
SnCl2 (Fig. 5b-(1), Fig. S4). Adding FeCl3 to the third layer
followed by SnCl2, the isosbestic point shifted one point dur-
ing addition of SnCl2, which means SnCl2 accumulated on the
fourth layer (Fig. 5b-(3), Fig. S4). These results indicate that
SnCl2 accumulated on the outer layer and did not exchange
with FeCl3 at complex sites (Fig. 5 and Fig. S4). Another
combination of metals was examined and similar results were
obtained (Fig. S5 and Table 1).
But, adding 6 equiv of SnCl2 to the second layer, followed
by 6 equiv of FeCl3, the isosbestic points during addition of
FeCl3 were observed at 361 nm for 0–2 equiv of FeCl3 and
359 nm for 2–6 equiv of FeCl3. The clear change in the isos-
bestic point therefore indicates that the heterometal complex-
ation is not governed by the order of addition of the metals.
Equivalent values for each isosbestic point during FeCl3 were
However, in case of the addition of SnCl2 followed by
FeCl3, the shift in the isosbestic points was different from
the case of FeCl3 followed by SnCl2 (Fig. 6). If the metal salts
were assembled according to the order of addition of the metal
salts, FeCl3 would be present in the third layer, and the isos-
bestic point during the addition of FeCl3 would not change.
(b)
(1)
4 SnCl2
8 SnCl2
364 nm
16 SnCl2
363.5 nm
366.5 nm
4 FeCl3
369 nm
(2)
8 SnCl2
364 nm
16 SnCl2
363 nm
4 FeCl3
370 nm
(3)
8 FeCl3
366 nm
16 SnCl2
363 nm
Fig. 5. (a) Heterometal assembly of DPA G4 (5:0 ꢁ 10ꢂ6
M, acetonitrile/chloroform = 1:1) by the addition of 6
equiv of FeCl3 followed by 24 equiv of SnCl2. Detailed
spectra of the isosbestic points are inset. These spectra
correspond to the schematic representation of the pattern
(1). The other spectra which correspond to the schematic
representation of the pattern (2) and (3) are shown in
Fig. S4. (b) Schematic representation of heterometal
assembly on DPA G4 by the addition of FeCl3 followed
by SnCl2.
Fig. 4. Titration curves of DPA G1 (5:0 ꢁ 10ꢂ5 M, tetra-
hydrofuran/acetonitrile = 1:1) complexed with FeCl3,
GaCl3, VCl3, and SnCl2. Optical difference at the band
attributed to the complexation during the titration was nor-
malized as ꢀA.29 The complexation abilities of four metal
salts were derived by curve-fitting a theoretical simulation
to the titration.
Table 1. Isosbestic Points during the Titration of DPA G4 with Two Metals (FeCl3, GaCl3, and SnCl2)
2Fe !
28Ga
/nm
6Fe !
24Ga
/nm
14Fe !
16Ga
/nm
2Fe !
28Sn
/nm
6Fe !
24Sn
/nm
14Fe !
16Sn
/nm
2Ga !
28Sn
/nm
6Ga ! 14Ga !
24Sn
/nm
16Sn
/nm
1st (2 equiv)
2nd (4 equiv)
3rd (8 equiv)
370.5(Fe) 370.5(Fe) 371(Fe)
369(Fe)
366(Fe)
371(Fe)
371(Fe) 370.5(Fe) 370(Ga) 370(Ga) 370(Ga)
368(Ga)
365.5(Ga) 368(Ga)
369(Fe)
366.5(Sn) 369(Fe) 369(Fe)
364(Sn) 364(Sn) 366(Fe)
366(Sn) 366(Ga) 366(Ga)
365(Sn) 364(Sn) 364(Ga)
363(Sn) 363(Sn) 362(Sn)
4th (16 equiv) 363.5(Ga) 365.5(Ga) 370(Ga) 363.5(Sn) 363(Sn) 363(Sn)

1566 Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007)
Heterometal Assembly in Dendritic Polyphenylazomethines
(b)
(b)
2 FeCl3
361 nm
6 SnCl2
2 FeCl3
371 nm
2 FeCl3
369 nm
2 SnCl2
366 nm
6 FeCl3
4 FeCl3
367 nm
4 SnCl2
364 nm
16 SnCl2
363 nm
4 FeCl3
359 nm
2 SnCl2
363 nm
8 FeCl3
8 FeCl3
364 nm
8 SnCl2
362 nm
Fig. 6. (a) Heterometal assembly of DPA G4 (5:0 ꢁ 10ꢂ6
M, acetonitrile/chloroform = 1:1) by the addition of 6
equiv of SnCl2 followed by 6 equiv of FeCl3 and 2 equiv
of SnCl2. Detailed spectra of the isosbestic points are in-
set. (b) Schematic representation of heterometal assembly
on DPA G4 by the addition of SnCl2 followed by FeCl3.
Fig. 7. (a) Heterometal assembly of DPA G4 (5:0 ꢁ 10ꢂ6
M, acetonitrile/chloroform = 1:1) by the addition of 10
equiv of FeCl3 followed by 4 equiv of SnCl2. Detailed
spectra of the isosbestic points are inset. The other patterns
are shown in Fig. S7. (b) Schematic representation of
heterometal assembly on the same layer of DPA G4 with
the addition of FeCl3 followed by SnCl2.
2 and 4, which correspond to the number of first and second
layers of imine sites (Fig. 6); therefore SnCl2 on the first and
second layer moved to the third layer with the addition of
FeCl3. We also examined the pattern of SnCl2 followed by
GaCl3, and a similar result was obtained (Fig. S6).
Moreover, we examined the accumulation of two kinds of
metals on the same layer. Upon adding FeCl3 followed by
SnCl2 to the same layer, two isosbestic points were observed,
one derived from FeCl3 and the other from SnCl2 (Fig. 7 and
Fig. S7). A similar pattern occurred even if the combination of
metal salts was changed (Fig. S8). Therefore, different isos-
bestic points appeared when different metal salts were mixed
in the same DPA layer.
On the other hand, adding 8 equiv of SnCl2 followed by 6
equiv of FeCl3 to the same layer, two isosbestic points were
observed at 360 nm for 0–2 equiv of FeCl3 and 362 nm for
2–6 equiv of FeCl3. This indicates that FeCl3 drives out SnCl2
on both the first layer and the second layer (Fig. 8). In addi-
tion, the location and type of metals was confirmed in the
UV–vis spectra, because two isosbestic points were observed
when two metals filled in the layer.
We used UV–vis spectroscopy to confirm that the complex
reaches the same final state regardless of the order of addition.
We investigated the final states of the complexes incorporating
6 equiv of FeCl3 and 8 equiv of SnCl2 in DPA G4 (5:0 ꢁ 10ꢂ6
M, acetonitrile/chloroform = 1:1) with these three patterns:
[6FeCl3 ! 8SnCl2], [6SnCl2 ! 6FeCl3 ! 2SnCl2], and
[8SnCl2 ! 6FeCl3]. The resulting spectra were in agreement
(ꢀmax ¼ 397 nm, " ¼ 2:5 ꢁ 105 dm3 molꢂ1 cmꢂ1) (Fig. 9). In
addition, the spectra agreed for the combination of 6 equiv
of GaCl3 and 8 equiv of SnCl2, (Fig. S9). For heterometal
assembly in DPA using FeCl3 and SnCl2, regardless of the
order of addition of the metals, FeCl3 complexed in the interior
layer of DPA G4, whereas SnCl2 complexed in the exterior
layer. This result indicates that the complex reaches a thermo-
dynamically stable condition regardless of the order of addi-
tion. This result can be attributed to the difference in K be-
tween FeCl3 and SnCl2, where the metal with the larger K
complexes with the inner layer imines, and the metal with the
smaller K coordinates with the outer layer imines.
Based on this principle, heterometal assembly with FeCl3,
GaCl3, VCl3, and SnCl2 was carried out. As the titration was
carried out in the order of complexation ability, four shifts in
isosbestic points (371 nm ! 368 nm ! 356 nm ! 364 nm)
were observed (Fig. 10). This indicates that the complexation
is in a stepwise radial fashion, because the added metal equiva-
lents, 2, 4, 8, and 16 corresponded to the number of complex-
ation sites in each layer of DPA (Fig. 10). Various patterns of

K. Takanashi et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007) 1567
(b)
(b)
2 FeCl3
371 nm
4 GaCl3
368 nm
8 SnCl2
2 FeCl3
360 nm
4 FeCl3
362 nm
8 VCl3
16 SnCl2
364 nm
356 nm
Fig. 8. (a) Heterometal assembly of DPA G4 (5:0 ꢁ 10ꢂ6
M, acetonitrile/chloroform = 1:1) by the addition of 8
equiv of SnCl2 followed by 6 equiv of FeCl3. Detailed
spectra of the isosbestic points are inset. (b) Schematic
representation of heterometal assembly of DPA G4 by
the addition of SnCl2 followed by FeCl3.
Fig. 10. (a) UV–vis titration of DPA G4 (5:0 ꢁ 10ꢂ6 M,
acetonitrile/chloroform = 1:1) with FeCl3, GaCl3, VCl3,
and SnCl2. (b) Schematic representation of heterometal
assembly of DPA G4 by FeCl3, GaCl3, VCl3, and SnCl2.
Table 2. Isosbestic Points during the Titration of DPA G4
with Three Different Metals (FeCl3, GaCl3, and SnCl2)
Fe, Fe,
Ga, Sn
/nm
Fe, Ga,
Ga, Sn
/nm
Fe, Ga,
Sn, Sn
/nm
1 ! 2 !
3 ! 4
8 SnCl2
6 FeCl3
1st (2 equiv)
2nd (4 equiv)
3rd (8 equiv)
4th (16 equiv)
370(Fe)
370.5(Fe)
366.0(Ga)
364.5(Ga)
363.5(Sn)
370(Fe)
366.5(Fe)
364.5(Ga)
363.5(Sn)
365.5(Ga)
364.0(Sn)
362.5(Sn)
6 FeCl3
metal in the inner layer and the weaker coordinating metal in
the outer layer.
Fig. 9. Schematics of the addition of the metal salts in the
following order: [6FeCl3 ! 8SnCl2; ꢀmax = 397 nm,
" ¼ 2:5 ꢁ 105 dm3 molꢂ1 cmꢂ1 (based on the concentra-
tion of DPA G4 (5:0 ꢁ 10ꢂ6 M, acetonitrile/chloroform =
1:1)], [8SnCl2 ! 6FeCl3; ꢀmax ¼ 397 nm, " ¼ 2:5 ꢁ 105
TEM Observation of Dendrimer Metal Assemblies.
Since TEM can observe only metal assemblies, it is a useful
technique for studying our heterometal dendrimer assemblies.
Since metals are assembled stepwise from the interior of a den-
drimer, the assembly areas should be different depending on
the number of metals assembled. Generally in TEM observa-
tion, organic substances are dyed with RuO4 vapor to intensify
the visual contrast. However, tin and some other metals assem-
bled in an organic substance are observable without dyeing or
by dyeing for a very short time. We began with the TEM
observation of DPA G4 assembled with 14 equiv FeCl3
((FeCl3)14@DPA G4) and 14 equiv FeCl3 and 16 equiv SnCl2
((FeCl3)14(SnCl2)16@DPA G4) (Fig. 11).8,9 We measured the
dm3 molꢂ1 cmꢂ1
]
and [6SnCl2 ! 6FeCl3 ! 2SnCl2;
ꢀmax ¼ 397 nm, " ¼ 2:5 ꢁ 105 dm3 molꢂ1 cmꢂ1].
metals were examined (Table 2). We succeeded in developing
a method of heterometal assembly incorporating different
metal salts in each of four layers. From these results, hetero-
metal assembling on DPA appears to be governed by the com-
plexation ability of metal salts, with the stronger coordinating

1568 Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007)
Heterometal Assembly in Dendritic Polyphenylazomethines
Fig. 11. TEM observation of dendrimer metal assemblies.
(a) (FeCl3)14@DPA G4 was dyed for a short time,
and the metal assembly was observed. (b) (FeCl3)14-
(SnCl2)16@DPA G4 was dyed for a short time, and the
metal assembly was observed.
sizes of the metal assemblies in (FeCl3)14@DPA G4 and
estimated their average diameter to be 2.0 nm, and that in
(FeCl3)14(SnCl2)16@DPA G4 to be 2.4 nm. As the number of
metals increased, the assembly areas expanded. The size of the
metal assembly in (FeCl3)14@DPA G4 matched that in
(SnCl2)14@DPA G4, which we have reported earlier.8,9 This
size covers the metal assemblies up to the third layer inside
the dendrimer and suggests that the 14 metals are assembled
in the first through the third layers from the center of the den-
drimer, and not randomly over the four layers. On the other
hand, the size of the metal assembly in (FeCl3)14(SnCl2)16@
DPA G4 was 0.4 nm larger than that in (FeCl3)14@DPA G4,
which suggests that some tin filled the fourth layer.21
Next, we dyed (FeCl3)14@DPA G4 and (FeCl3)14(SnCl2)16@
DPA G4 for about 15 min and observed the sizes of the whole
dendrimer molecules (Fig. S10). We estimated the sizes to be
2.5 nm for the (FeCl3)14@DPA G4 molecule and 2.7 nm for
(FeCl3)14(SnCl2)16@DPA G4. These values are larger than
the metal areas and the sizes of assemblies in (FeCl3)14@DPA
G4 and (FeCl3)14(SnCl2)16@DPA G4 are correct.8,9
Fig. 12. 57Fe Mossbauer spectra of: (a) (FeCl ) @DPA
¨
3
2
G4 (IS = 0.30, QS = 0.80), (b) (FeCl3)2(SnCl2)4@DPA
G4 (IS = 0.30, QS = 0.80), and (c) (FeCl3)6@DPA G4
(IS = 0.30, QS = 0.80, IS = 0.34, QS = 0.85). (FeCl3)2
@DPA G4 and (FeCl3)2(SnCl2)4@DPA G4 exhibit very
similar spectra, which indicates that the Fe environments
are similar. On the other hand, the peak for (FeCl3)6@
DPA G4 is flat (see Fig. S15), indicating that it is a mix
of two components.
(FeCl3)6@DPA G4, a two-doublet fitting was needed (IS =
0.30, QS = 0.80, IS = 0.34, QS = 0.85). These results sug-
gest that Fe is attached only to the core of (FeCl3)2@DPA
G4. Results for (FeCl3)6@DPA indicate the existence of a
component coordinated in the first layer and another compo-
nent in the second layer. This conclusion is also supported
by the fact that two-component fitting was certainly possible
in the analysis of the spectrum of (FeCl3)6@DPA G4 based
on the parameters for (FeCl3)2@DPA G4, with a 1:2 ratio
between the two components.
(FeCl3)2(SnCl2)4@DPA G4 exhibits a spectrum very simi-
lar to that of (FeCl3)2@DPA G4. (FeCl3)2(SnCl2)4@DPA G4
is also a single doublet, and we were able to fit it by using pa-
rameters from (FeCl3)2@DPA G4. This result shows that the
environment for the Fe is the same as that in (FeCl3)2@DPA
G4 and indicates that the Fe in (FeCl3)2(SnCl2)4@DPA G4
is not randomly placed, but attached to the core.
The results described above demonstrate that the FeCl3
complex forms stepwise from the center, and that the hetero
assembly consists of precisely located Fe and Sn atoms.
X-ray Photoelectron Spectra of Dendrimer Complexes.
In XPS observation of a metal complex, the more intense the
electron donation from the ligand, the lower the eV at which
the metal ion’s spectrum peak is observed.24 When the position
of the metal located in the DPA changes, the spectra also
changes corresponding to the difference in the electric charge
As a result of the above, we have shown that the size of the
metal-coordinated sections as well as the molecules created in-
creases as a result of metal assembly in dendrimers, even when
different kinds of metals are assembled. This fact suggests that
the metals complex stepwise from the inside of the dendrimer.
57
¨
Fe Mossbauer Spectra of DPA Complexes. In order
to clarify the precise structure of the heterometal assembly,
we measured Mossbauer spectra of 57Fe (Fig. 12). To obtain
¨
(57FeCl3)2@DPA G4, we formed a complex of 57FeCl3 and
DPA G4, brought it into a state of equilibrium, and then
precipitated it into hexane. Figure 12 shows the Mossbauer
¨
spectra of the powdered solids that we obtained by treating
(57FeCl3)2@DPA G4, (57FeCl3)6@DPA G4, and (57FeCl3)2-
(SnCl2)4@DPA G4 in a similar way. (In Fig. S11, the spectra
were normalized with respect to intensity.) (FeCl3)2@DPA G4
shows a spectrum different from that of FeCl3 alone (IS =
0.436, QS = 0 at RT).22 This is due to the complex formation
of FeCl3 with imine. The (FeCl3)2@DPA G4 spectra shows no
temperature dependence (Fig. S12). In addition, the IS and QS
values indicate the high-spin state of Fe.23
A single doublet-fitting was used with the spectrum of
(FeCl3)2@DPA G4 (IS = 0.30, QS = 0.80) and indicates the
existence of a single component, Fe. On the other hand, for

K. Takanashi et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007) 1569
(b)
2 SnCl2
2 SnCl2
Fig. 13. Ga3d3/2 XPS spectra of (GaCl3)2@DPA G4
(20.75 eV), (GaCl3)6@DPA G4 (20.90 eV), (GaCl3)14@
DPA G4 (21.05 eV), and (GaCl3)30@DPA G4 (21.15
eV). As GaCl3 increased, the spectral peaks shifted to a
higher electronvolts.
2 SnCl2
2 BF3
Fig. 14. (a) Chart showing the relationship between the
chemical shift of the fluorine peak in the 19F NMR spec-
trum and the quantity of BF3 added (indicated as ). Also
shown is the relationship in the case of adding SnCl2
to (BF3)2@DPA G4, with the vertical axis denoting the
chemical shift and the horizontal axis the quantity of
SnCl2 added (indicated as ). Similarly, a relationship
is also shown for the case of adding SnCl2 after adding
two equivalent amounts of BF3 to (SnCl2)2@DPA G4
(indicated as ). (b) Schematic representation of adding
SnCl2 to (BF3)2@DPA G4 and 2 equivalent amounts of
BF3 to (SnCl2)2@DPA G4.
density. Due to its detection sensitivity, GaCl3[Ga(3d3=2)] was
the metal of choice, and its assemblies were produced in DPA
(Fig. 13). The UV–vis spectrum of GaCl3 changes stepwise.
Based on this change, we synthesized (GaCl3)2@DPA G4,
(GaCl3)6@DPA G4, (GaCl3)14@DPA G4, and (GaCl3)30@
DPA G4 and conducted XPS measurements on the assemblies.
The top peak in the spectrum of Ga(3d3=2) shifted upwards
from 20.75 to 20.90, 21.05, and 21.15 eV as the metal assem-
bly increased from 2 to 6, 14, and 30, respectively. Simulations
were performed on how the assembly takes place and how the
spectrum changes (Fig. S13). If the metals assemble randomly
in the different layers of imine sites, the spectrum shows no
change even when the number of metals assembled increases.
Only when different layers of imine sites have different elec-
tron-donating properties and the metals assemble stepwise
from the inside do the peaks shift to higher energy. Since
this agrees with our experimental results, it is evident that the
metal assembly had some regularity. The spectral peak shifted
towards higher energy as the assembly increased, which indi-
cates a decrease in average electric charge density of the metal
captured into the dendrimer. In more specific terms, this fact
indicates that as metals add, they firstly assembled in the inner
layer of imines, where the charge density is high, followed
by assembly into the outer layer where the charge density
is lower. We observed similar shifts in peaks in the XPS spec-
tra of (SnCl2)n@DPA G4 [Fig. S14 for (SnCl2)2@DPA G4
(487.10 eV), (SnCl2)6@DPA G4 (487.25 eV), (SnCl2)14@DPA
G4 (487.30 eV), and (SnCl2)30@DPA G4 (487.35 eV)]. These
results also prove that the metals form a complex from the
inner layer.
the DPA.25 At NMR concentrations, a complex formed more
than 99% of the time.
Spectra were collected at room temperature. We added the
BF3 solution dropwise (0.25 equiv for each drop) into the
DPA G4 solution (acetonitrile/chloroform = 1:1, 1:0 ꢁ 10ꢂ4
M) and measured the NMR spectra with trifluoroacetic acid
as an external standard (ꢂ76:5 ppm).
With BF3 alone, the peak appears at ꢂ150:5 ppm. With 0.25
equivalents of BF3 added to the DPA G4 solution, the peak
appeared at ꢂ152:7 ppm, because of its coordination on the
imine. This coordination to the imine suggests that the electric
charge density of F increases. With more BF3 added, the peak
shifted to a higher chemical shift (Fig. 14, ꢂ152:7 to ꢂ151:5
ppm). During the titration from 0 to 14 equivalent amounts,
up to the third layer, we observed only one peak. Though we
saw no peak reflecting each of the layers, we think that this is
because the coordination kinetics and the time scale of the
NMR measurement are comparable, and averaged peaks were
observed as expected. What is important here is that since BF3
was almost completely incorporated into the complex, no shift
in the F peak would have been seen if the chemical environ-
ment within the dendrimer had been consistent throughout or
if BF3 had been randomly placed (Fig. S20). The shift in the
NMR Spectroscopy. BF3 assembled stepwise in DPA G4
in four layers, beginning with the innermost one. Figure S15
shows the UV–vis titrations. We estimated K to be equal to
or larger than that of FeCl3 (K > 108 Mꢂ1, Figs. S16–S19).
We measured the 19F NMR spectrum of BF3 introduced into

1570 Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007)
Heterometal Assembly in Dendritic Polyphenylazomethines
GaCl3 (acetonitrile, 5:0 ꢁ 10ꢂ3 M), SnCl2 (acetonitrile, 5:0 ꢁ
10ꢂ3 M), and VCl3 (acetonitrile, 5:0 ꢁ 10ꢂ3 M) were prepared
in volumetric flasks under nitrogen. To a quartz cell was added
3.0 mL of the DPA G4 solution, followed by the addition of
3 mL of metal salt solution (i.e., 1 equivalent for DPA G4). The
UV–vis spectra were measured until agreement was obtained with
the one immediately before. Measurements were repeated after
each addition of 3 mL of metal salt solution to achieve UV–vis
titration. In the case of FeCl3, the absorbance of FeCl3 was
subtracted.10
F peak take place because the F atoms exists in different envi-
ronments and are regularly assembled. The peak shifted up-
field with addition of BF3, suggesting that the electron density
of F increases (more shielding). We propose that, since the
electron density of the imine is lower in the outer layers, the
downfield F peak shift (less shielding) indicates more BF3 is
coordinated in the outer layers. This result suggests the exis-
tence of an electron gradient of nitrogen from the inner to the
outer layers.
BF3 has the largest complexation constant of all the metals
that gradually form a complex with DPA. Therefore, BF3 re-
mained in the inner layers even after mixing with another
metal. Despite the addition of a heterometal, the chemical shift
remains the same. We added some Sn after adding 2 equiva-
lents of BF3 and then observed the chemical shift. We found
the chemical shift remained the same, regardless of the
quantity of Sn added (Fig. 14). Similarly, we found that the
chemical shift was the same for (SnCl2)2@DPA G4 with 2
equivalent amounts of BF3 added and after adding SnCl2 to the
mixture (Fig. 14). These results also suggest that in the assem-
blies of heterometals, BF3 assembles in the inner layers and
SnCl2 in the outer layers. In other words, the assembly struc-
ture is metals with a higher K in the inner layers and metals
with a lower K in the outer layers.
Job Plot of DPA G0 Imine and GaCl3. A Job plot of the
reaction between GaCl3 and DPA G0 was done. FðxÞ ¼ Abs.=
ðCG0 þ CGaCl3Þ ꢂ ð"G0 ꢂ "GaCl3Þx ꢂ "GaCl3
,
x ¼ CG0=ðCG0 þ
CGaCl3Þ; mole fraction of DPA G0. Solutions of DPA G0 (aceto-
nitrile/chloroform = 1:1, 2:5 ꢁ 10ꢂ4 Mꢂ1) and GaCl3 (acetoni-
trile/chloroform = 1:1, 2:5 ꢁ 10ꢂ4 Mꢂ1) were prepared in volu-
metric flasks, the latter under nitrogen. To a quartz cell was added
2.0 mL of the DPA G0 solution followed by the sequential addi-
tion of 0.2 mL aliquots of the GaCl3 solution up to a total of 2.2
mL to obtained solutions mixed in various proportions. The UV–
vis spectra were recorded between each step. The same operation
was also performed by the addition of DPA G0 solution to the
GaCl3 solution. The plot shows a maximum at 0.5 mole fraction
of DPA G0, which means that the imine forms a 1:1 complex with
GaCl3. The value of K, which was found by curve-fitting a theo-
retical simulation to the experimental data was more than 108 Mꢂ1
(in acetonitrile/chloroform).
Conclusion
Titration of DPA G1 with Metals in Acetonitrile/Tetra-
We demonstrated the stepwise complexation of GaCl3 and
VCl3 on DPA by the shifts of isosbestic points of UV–vis spec-
tra. Further, we developed a method of precise heterometal
assembly, which makes use of these different complexation
abilities (FeCl3 > GaCl3 > VCl3 > SnCl2). The stepwise com-
plexation and heterometal assembling were also supported
hydrofuran = 1:1 Solvent.
Solutions of DPA G1 (aceto-
nitrile/tetrahydrofuran = 1:1, 5:0 ꢁ 10ꢂ5 M), FeCl3 (acetonitrile,
1:0 ꢁ 10ꢂ2 M), GaCl3 (acetonitrile, 1:0 ꢁ 10ꢂ2 M), SnCl2 (aceto-
nitrile, 1:0 ꢁ 10ꢂ2 M), and VCl3 (acetonitrile, 1:0 ꢁ 10ꢂ2 M) were
prepared in volumetric flasks under nitrogen. To a quartz cell was
added 3.0 mL of the DPA G4 solution, followed by the addition of
3 mL of metal salt solution (i.e., 0.2 equivalent for DPA G1). The
measurement of the spectra and the addition of the metal salt solu-
tion continued until the spectral change stopped.
by TEM, 57Fe Mossbauer spectroscopy, XPS measurements,
¨
and NMR spectroscopy. These methods may be used to gene-
rate new materials and could be applied to many fields, such as
catalysts and electronics.
XPS Measurements of DPA Complexes. To a 1 mL solution
of DPA G4 (acetonitrile/chloroform = 1:1, 5 ꢁ 10ꢂ4 M), 2, 6, 14,
or 30 equiv GaCl3 solution (acetonitrile, 2:0 ꢁ 10ꢂ2 M) was add-
ed. This GaCl3@DPA G4 solution (10 mL) was then applied to
the Au plates (7 ꢁ 7 mm2) and dried in vacuo. The samples were
characterized by the Ga(3d) peak. The Au(3f7=2) peak was used as
an internal standard (83.8 eV).
TEM Observation of DPA Complexes. TEM samples were
prepared by applying 1.5 mL of the solution of (FeCl3)14@DPA
G4 or (FeCl3)14(SnCl2)16@DPA G4 (chloroform, 1 ꢁ 10ꢂ5 M) to
two carbon-coated TEM grids and dried in vacuo. One plate was
treated by RuO4 vapor for 1 min and the other was treated for
15 min. The TEM images were obtained at 200 kV and the size
distribution of particles was determined by counting about 100
Experimental
Materials. DPAs were prepared according to the reported
method.26,27 Dehydrated chloroform and tin(II) chloride were pur-
chased from Wako Pure Chemical industries (Japan), dehydrated
acetonitrile and BF3 OEt2 were purchased from Kanto Kagaku
ꢃ
(Japan), iron(III) chloride was purchased from Merck, and gal-
lium(III) chloride and vanadium(III) chloride were purchased
from Sigma-Aldrich. 57Fe was purchased from TANGO OVER-
SEAS (Japan) and 57FeCl3 was prepared according to the litera-
ture method.28
General Method.
Shimadzu UV-3150, UV-3100PC, and UV-2400PC spectrophoto-
UV–vis spectra were measured using
meters. The 57Fe Mossbauer spectra were measured using a
particles.
57
¨
¨
Topologic systems model 222 constant-acceleration spectrometer
with a 57Co/Rh source in the transmission mode. The measure-
ments at low temperature were performed with a closed-cycle he-
lium refrigerator (Daikin). The XPS measurements were performed
using a JEOL JPS-9000MC spectrometer with a Mg Kꢁ radiation.
The TEM images were obtained using a TECNAI F20 field-emis-
sion electron microscope operated at 200 kV. The NMR spectra
were measured using a JEOL JNM-A400 spectrometer.
Fe Mossbauer Spetra. All samples were prepared using
57FeCl3. The samples of (FeCl3)2@DPA G4, (FeCl3)2(SnCl2)4
@DPA G4, and (FeCl3)6@DPA G4 were prepared by reprecipita-
tion from hexane. A typical preparation of (FeCl3)2@DPA G4 was
as follows. 7.0 mg of 57FeCl3 (7.0 mg, 43.5 mmol) was dissolved in
(1.0 mL) and 0.128 mL of this solution was added to the solution
of DPA G4 (15 mg, 2.75 mmol, chloroform:acetonitrile = 1:1, 3
mL). This Fe–DPA complex solution was poured into 300 mL
of hexane with sonication, and the precipitate was collected by
membrane filtration. The samples were dried in vacuo and wrap-
UV–Vis Titration. Solutions of DPA G4 (acetonitrile/chloro-
form = 1:1, 5:0 ꢁ 10ꢂ6 M), FeCl3 (acetonitrile, 5:0 ꢁ 10ꢂ3 M),

K. Takanashi et al.
Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007) 1571
ped in aluminum foil to perform the measurements. RT data for
(FeCl3)2@DPA G4 were also collected. The spectra were ana-
lyzed by fitting a Lorentzian line shape using MOSSWINN ver-
sion 3 software.
Aranzaes, Angew. Chem., Int. Ed. 2005, 44, 7852. c) M. Kimura,
T. Shiba, M. Yamazaki, K. Hanabusa, H. Shirai, N. Kobayashi,
J. Am. Chem. Soc. 2001, 123, 5636. d) P. Weyermann, J.-P.
Gisselbrecht, C. Boudon, F. Diederich, M. Gross, Angew. Chem.,
Int. Ed. 1999, 38, 3215. e) H. Zeng, G. R. Newkome, C. L. Hill,
Angew. Chem., Int. Ed. 2000, 39, 1771. f) G. E. Oosterom, J. N. H.
Reek, P. C. J. Kamer, P. W. N. M. van Leeuwen, Angew. Chem.,
Int. Ed. 2001, 40, 1828. g) M. Uyemura, T. Aida, J. Am. Chem.
Soc. 2002, 124, 11392. h) J.-L. Zhang, H.-B. Zhou, J.-S. Huang,
C.-M. Che, Chem. Eur. J. 2002, 8, 1554. i) L. Plault, A. Hauseler,
S. Nlate, D. Astruc, J. Ruiz, S. Gatard, R. Neumann, Angew.
19F NMR Spectra. Spectra were collected at room tempera-
ture. We adjusted the BF3 solution by diluting BF3 OEt2 with
ꢃ
acetonitrile. We added this solution dropwise (0.25 equiv for each
drop) into a DPA solution (acetonitrile/chloroform = 1:1, 1:0 ꢁ
10ꢂ4 M) and measured the NMR spectra with trifluoroacetic acid
as an external standard (ꢂ76:5 ppm). Similarly, we added SnCl2
solution dropwise into a (BF3)2@DPA solution and measured
the NMR specra. The case of adding SnCl2 after adding two
equivalent amounts of BF3 to (SnCl2)2@DPA G4 was examined
in the same way.
´
´
Chem., Int. Ed. 2004, 43, 2924. j) K. Heuze, D. Mery, D. Gauss,
J.-C. Blais, D. Astruc, Chem. Eur. J. 2004, 10, 3936. k) S. A.
Chavan, W. Maes, L. E. M. Gevers, J. Wahlen, I. F. J.
Vankelecom, P. A. Jacobs, W. Dehaen, D. E. de Vos, Chem.
Eur. J. 2005, 11, 6754.
This work was partially supported by the CREST
project from JST and Grants-in-Aid for Scientific Research
(Nos. 19205020 and 178146) from the Ministry of Education,
Culture, Sports, Science and Technology.
6
Cluster synthesis using dendrimers, see for example:
a) R. W. J. Scott, O. M. Wilson, R. M. Crooks, J. Phys. Chem.
B 2005, 109, 692. b) Y. Niu, L. K. Yeung, R. M. Crooks,
J. Am. Chem. Soc. 2001, 123, 6840. c) R. W. J. Scott, A. K. Datye,
R. M. Crooks, J. Am. Chem. Soc. 2003, 125, 3708. d) R. W. J.
Scott, O. M. Wilson, S.-K. Oh, E. A. Kenik, R. M. Crooks,
J. Am. Chem. Soc. 2004, 126, 15583. e) J. C. Garcia-Martinez,
R. M. Crooks, J. Am. Chem. Soc. 2004, 126, 16170. f) Y.-G.
Kim, S.-K. Oh, R. M. Crooks, Chem. Mater. 2004, 16, 167.
g) C. Wang, G. Zhu, J. Li, X. Cai, Y. Wei, D. Zhang, S. Qiu,
Chem. Eur. J. 2005, 11, 4975.
Supporting Information
UV–vis titration, Job-plot, TEM images, 57Fe Mossbauer spec-
¨
tra, XPS spectra, and 19F NMR chart. This material is available
free of charge on the Web at: http://www.csj.jp/journals/bcsj/.
References
1
a) R. E. Blankenship, Molecular Mechanisms of Photosyn-
7
a) M. Ballauff, C. N. Likos, Angew. Chem., Int. Ed. 2004,
´
thesis, Blackwell Science, Oxford, U.K., 2002. b) A. E. Shilov,
Metal Complexes in Biomimetic Chemical Reactions, Springer-
Verlag, Berlin, Germany, 1997. c) Enzyme Catalysis in Organic
Synthesis, ed. by K. Drauz, H. Waldmann, VCH, Weinheim,
Germany, 1995.
43, 2998. b) C. J. Hawker, J. M. J. Frechet, J. Am. Chem. Soc.
1992, 114, 8405. c) V. Maraval, R. Laurent, B. Donnadieu, M.
Mauzac, A.-M. Caminade, J.-P. Majoral, J. Am. Chem. Soc.
2000, 122, 2499.
8
Nature 2002, 415, 509.
K. Yamamoto, M. Higuchi, S. Shiki, M. Tsuruta, H. Chiba,
2
a) Macromolecular Metal Complexes, ed. by F. Ciardelli,
E. Tsuchida, D. Worle, Springer-Verlag, Berlin, Germany, 1995.
b) Macromolecular Complexes, ed. by E. Tsuchida, VCH, New
York, 1991. c) Metal Containing Polymeric Materials, ed. by
C. U. Pittman, B. M. Cullberston, J. E. Sheet, ACS Symposium
Series, Plenum, New York, 1996.
9
J. Am. Chem. Soc. 2003, 125, 9988.
10 R. Nakajima, M. Tsuruta, M. Higuchi, K. Yamamoto,
J. Am. Chem. Soc. 2004, 126, 1630.
11 O. Enoki, T. Imaoka, K. Yamamoto, Bull. Chem. Soc. Jpn.
2006, 79, 621.
M. Higuchi, M. Tsuruta, H. Chiba, S. Shiki, K. Yamamoto,
3
a) D. Astruc, F. Chardac, Chem. Rev. 2001, 101, 2991.
b) R. M. Crooks, M. Zhao, L. Sun, V. Chechik, L. K. Yeung,
Acc. Chem. Res. 2001, 34, 181. c) R. van Heerbeek, P. C. J.
Kamer, P. W. N. M. van Leeuwen, J. N. H. Reek, Chem. Rev.
2002, 102, 3717. d) M. Ooe, M. Murata, T. Mizugaki, K. Ebitani,
K. Kaneda, J. Am. Chem. Soc. 2004, 126, 1604.
12 K. Yamamoto, M. Higuchi, A. Kimoto, T. Imaoka, K.
Masachika, Bull. Chem. Soc. Jpn. 2005, 78, 349.
13 P. Job, Ann. Chim. 1928, 9, 113.
14 W. C. Vosburgh, G. R. Cooper, J. Am. Chem. Soc. 1941,
63, 437.
4
Reviews of dendrimer, see for example: a) F. Zeng, S. C.
15 Z. D. Hill, P. MacCarthy, J. Chem. Educ. 1986, 63, 162.
16 Because the complexation ability of Fe and Ga in aceto-
nitrile/chloroform was too high to compare, the experiments were
performed in acetonitrile/tetrahydrofuran. Generally, the com-
plexation ability decreases in polar solvent, so tetrahydrofuran
was a good solvent for comparing the complexation ability.
17 The absorbance (Abs.) of the solution and K between
metal salt and DPA imine are represented as follows;
Zimmerman, Chem. Rev. 1997, 97, 1681. b) A. W. Bosman,
H. M. Janssen, E. W. Meijer, Chem. Rev. 1999, 99, 1665. c) M.
Fischer, F. Vogtle, Angew. Chem., Int. Ed. 1999, 38, 884.
¨
d) A. J. Berresheim, M. Muller, K. Mullen, Chem. Rev. 1999,
¨
99, 1747. e) F. Vogtle, S. Gestermann, R. Hesse, H. Schwierz,
¨
¨
B. Windisch, Prog. Polym. Sci. 2000, 25, 987. f) M. D. Watson,
A. Fechtenkotter, K. Mullen, Chem. Rev. 2001, 101, 1267.
¨
g) S. M. Grayson, J. M. J. Frechet, Chem. Rev. 2001, 101, 3819.
¨
´
h) G. R. Newkome, C. N. Moorefield, F. Vogtle, Dendrimers
K ¼ Ccomplex=ðCmetalCimineÞ, Abs. ¼ "metalCmetal þ "imineCimine
complexCcomplex. The replacement of Ccomplex as x gives the
equations; K ¼ x=fðM0 ꢂ xÞðL0 ꢂ xÞg, Abs. ¼ "metalðM0 ꢂ xÞ þ
þ
"
¨
and Dendrons: Concepts, Syntheses, Applications, VCH,
´
Weinheim, Germany, 2001. i) J. M. J. Frechet, D. A. Tomalia,
Dendrimers and Other Dendritic Polymers, Wiley, Chichester,
U.K., 2002.
5 Catalysts using dendrimers, see for example: a) R. van
Heerbeek, P. C. J. Kamer, P. W. N. M. van Leeuwen, J. N. H.
Reek, Chem. Rev. 2002, 102, 3717. b) D. Astruc, F. Lu, J. R.
"imineðL0 ꢂ xÞ þ "complexx, where M0 is the concentration of the
amount of the added metal salts and L0 is the initial concentration
of the imine on DPA. Curve fitting of the experimental data
gave the coordination constant (K) between metal salt and imine
on DPA.
18 N. Toshima, T. Yonezawa, New J. Chem. 1998, 22, 1179.

1572 Bull. Chem. Soc. Jpn. Vol. 80, No. 8 (2007)
Heterometal Assembly in Dendritic Polyphenylazomethines
19 S. Sun, C. B. Murray, D. Weller, L. Folks, A. Moser, Sci-
ence 2000, 287, 1989.
Am. Chem. Soc. 1971, 93, 1904. c) H. Inoue, T. Kobayashi, E.
Fluck, Hyperfine Interact., 1994, 68, 193. d) M. Endo, Z. Jin, S.
Kasai, H. Hasegawa, Jpn. J. Appl. Phys. 2002, 41, 2689.
25 Though 11B is NMR active as well, it is not good in this
case since its BF3 complex is insensitive to a change in the envi-
ronment and therefore not expected to cause a significant change
in the chemical shift.
26 M. Higuchi, S. Shiki, K. Yamamoto, Org. Lett. 2000, 2,
3079.
27 K. Takanashi, H. Chiba, M. Higuchi, K. Yamamoto,
Org. Lett. 2004, 6, 1709.
28 A. R. Pray, Inorg. Synth. 1957, 5, 153.
29 ꢀA is the absorption at x equiv of metal normalized by the
absorption of excess amount of metal salt. In case of FeCl3 and
GaCl3, 4 equiv of metal salts is the excess amount of the metals,
because the change in the UV–vis spectra was saturated. In the
same way, VCl3 is 5 equiv and SnCl2 is 12 equiv.
20 S. Zhou, B. Varughese, B. Eichhorn, G. Jackson, K.
McIlwrath, Angew. Chem., Int. Ed. 2005, 44, 4539.
21 Calculation based on the occupied volume at 14 equivalent
amounts as 2.0 nm gave us an estimate that the size of the whole
dendrimer complex will be 2.5 to 2.6 nm with the molecular size
being 1.95 to 2.04 nm.
22 N. N. Greenwood, T. C. Gibb, Mossbauer Spectroscopy,
Chapman and Hall, London, U.K., 1971.
¨
23 a) F. Genoud, I. Kulszewicz-Bajer, A. Bedel, J. L. Oddou,
C. Jeandey, A. Pron, Chem. Mater. 2000, 12, 744. b) I.
Kulszewicz-Bajer, J. Suwalski, Synth. Met. 2000, 12, 744. c) V.
Arora, Rajesh, P. Mathur, S. R. Vadera, Transition Met. Chem.
1999, 24, 92.
24 a) R. D. Feltham, P. Brant, J. Am. Chem. Soc. 1982, 104,
641. b) C. D. Cook, K. Y. Wan, U. Gelius, K. Hamrin, G.
Johansson, E. Olsson, H. Siegbahn, C. Nordling, K. Siegbahn, J.