Introduction

Using polyhedral oligomeric silsesquioxane (POSS) cores to synthesize dendrimers is attractive because it requires only a minimal number of synthetic steps; in addition, polyhedral structures of the POSS core can produce spherically symmetric dendrimers (even with earlier generations) more successfully than conventional cores can.1, 2, 3, 4 These POSS-cored dendrimers have relatively globular conformations and fewer entangled branches, with a high proportion of terminal functional groups positioned on the external surfaces of the dendrimers, even in earlier generations. Morris and colleagues4, 5 reported that these features were useful for the molecular carriers used as chemical catalysts to obtain the highest number of active sites possible; in these catalysts, the catalytic functionalities are tethered to the exterior surfaces of the dendrimers. Naka et al.6 synthesized a carboxylate-terminated POSS-core dendrimer as the first anionic water-soluble dendrimer based on POSS; they also studied its complexation behavior with Cu2+ ions in an aqueous solution by spectroscopic methods and isothermal titration calorimetry. The conformation of the POSS-core dendrimer causes a decrease in mobility of dendron chains and the sterically hindered structure on which Cu2+ ions are coordinated, even for early-generation dendrimers based on the rigid and symmetrical structure of the POSS core. The POSS-core dendrimer was more successful at entrapping guest molecules compared with a conventional dendrimer with the same terminal number.7 However, only a limited number of examples have been reported; no example has demonstrated the effect of generation numbers on the properties of POSS-core dendrimers.

In the field of dendrimer chemistry, the possibility of tailoring a dendrimer surface arranged with organic molecules that respond to stimuli such as light, solvent, heat and pH is particularly interesting.8, 9, 10, 11 Kono and colleagues8 reported that the lower critical solution temperatures of the dendrimers decreased remarkably as the dendrimer solution increased from G3 to G5. Dendrimers terminated with functional groups have certain remarkable features that the conventional thermosensitive polymers do not, such as a three-dimensional structure and a precisely controlled molecule size. The ability to functionalize surface dendrimers to respond to stimuli has opened new avenues to use these materials in many applications. However, in the case of conventional dendrimers, high-generation dendrimers are required for sharp response to stimuli, because of the densely packed terminal groups. As POSS-core dendrimers are expected to produce spherically symmetric dendrimers in early generations without densely packed terminal groups, POSS-core dendrimers terminated with functional groups are expected to respond sharply to external stimuli. Terminal functional groups on POSS-core dendrimers are expected to show chemical behaviors similar to those of the corresponding monomeric molecules in solution because the terminal function groups are isolated; furthermore, they inhibit the movement of neighboring branches.

Imidazole rings have attracted much attention because of their metal ion binding and buffering properties in the physiological pH range.12 The imidazole ring is a major component of many biochemicals found in nature, such as histidine. Linear polymers grafted with imidazole units showed pH-dependent phase transitions at a pH of 7.0.13, 14 Here, we prepared first- and second-generation imidazole-terminated POSS-core poly(amidoamine)-type dendrimers. We demonstrated that the present POSS-core dendrimers show sharp phase transitions depending on pH, even in an early generation and that the unique coordination behaviors of the metal ions depend on generation of dendrimers.

Experimental procedure

Materials

The first-generation methyl ester-terminated POSS-core poly(amidoamine)-typed dendrimer (POSS-OMe16) prepared according to the literature1, 2 was purified by size-exclusion chromatography (LH-20) using methanol as the eluent. Subsequently, the POSS-OMe16 dendrimer was reacted with excess ethylenediamine to produce the amino-terminated POSS-core dendrimer (POSS-(NH3)16) according to the literature.1 All chemicals were purchased from Wako Pure Chemical Industries Ltd (Osaka, Japan).

Measurements

1H-(300 or 400 MHz), 13C-(75 and 100 MHz) and 29Si-(100 MHz) nuclear magnetic resonance (NMR) spectra were recorded on a Bruker PDX-300 spectrometer (Bruker Biospin GmbH, Rheinstetten, Germany). The ultraviolet-vis spectra were recorded on a Jasco spectrophotometer (model V-670 KKN; Jasco, Tokyo, Japan). MALDI-TOF-MS (matrix assisted laser desorption ionization time-of-flight mass spectrometry) was recorded on a Bruker instrument (Autoflex II from Bruker Daltonics, Billerica, MA, USA). α-Cyano-4-hydroxycinnamic acid was used for MALDI matrices. Elemental analyses were performed on a Yanaco CHN analyzer (model CORDER MT-5; Yanaco, Tokyo, Japan)

POSS-OMe32

Methyl acrylate (33.20 g, 0.386 mol, 6 Equiv. against the primary amine) was added using a syringe to a solution of POSS-(NH3)16 (5.44 g, 2.01 mmol), and the resulting mixture was heated for 6 days at 40 °C. After the volatiles were evaporated, a colorless oil of crude POSS-OMe32 remained. The crude product was purified by preparative size exclusion chromatography (LH-20) using methanol as the eluent to remove the unreacted products. The yield of pure POSS-OMe32 was 82% (8.40 g, 1.65 mmol). 1H-NMR (CDCl3): δ 3.66 (s, 48H), δ 3.27 (m, 32H), δ 2.77 (m, 96H), δ 2.55 (m, 48H), δ 2.44 (m, 64H), δ 2.38 (m, 32H), δ 1.51 (s, 16H), δ 0.59 (s, 16H). MALDI-TOF-MS (m/z, [M+H]+): 5463.78 (calculated for C232N40O92H416Si8); 5464.23 (observed).

POSS-Im16

1-(3-Aminopropyl)imidazole (2.13 g, 0.02 mol) was added to POSS-OMe16 (0.49 g, 0.22 mmol), and the resulting mixture was heated for 18 days at 30 °C under nitrogen. The crude product was purified by preparative size exclusion chromatography (LH-20) using methanol as the eluent to remove the unreacted products. The yield of pure POSS-Im16 was 95% (0.78 g, 0.21 mmol). 1H-NMR (CDCl3): δ 0.42 (s, 16H), δ 1.36 (s, 16H), δ 1.94 (s, 32H), δ 2.31 (s, 32H), δ 2.65 (m, 48H), δ 3.13 (s, 32H), δ 4.00 (s, 32H), δ 6.95 (s, 16H), δ 6.99 (s, 16H), δ 7.52 (s, 16H). 13C-NMR(CDCl3): δ 9.497, 19.767, 30.920, 33.812, 36.173, 44.468, 49.841, 55.480, 119.164, 129.135, 137.269, 172.984. 29Si-NMR: δ 66.45 (s). FT-IR (KBr): 1189 (νSi−O−Si). Anal. calcd for C168N56O28H272Si8 12H2O: C, 50.88; H, 7.22; N, 19.79. Found: C, 50.73; H, 7.35; N, 19.88. MALDI-TOF-MS (m/z, [M+H]+): 3786.176 (calcd for C168N56O28H272Si82H2O); 3785.83 (observed).

POSS-Im32

1-(3-Aminopropyl)imidazole (13.9 g, 0.111 mol) was added to POSS-OMe32 (3.53 g, 0.694 mmol), and the resulting mixture was heated for 10 days at 50 °C under nitrogen. After the volatiles were evaporated, a colorless oil of crude POSS-Im32 remained. The crude product was purified by preparative size exclusion chromatography (LH-20) using methanol as the eluent to remove the unreacted products. The yield of pure POSS-Im32 was 80% (4.70 g, 0.556 mmol). 1H-NMR (CD3OD): δ 7.68 (s, 32H), δ 7.14 (s, 32H), δ 6.95 (s, 32H), δ 4.03 (t, 64H), δ 3.23 (m, 32H), δ 3.15 (m, 64H), δ 2.77 (m, 48H), δ 2.77 (m, 96H), δ 1.94 (m, 64H), δ 1.55 (m, 16H), δ 0.62 (m, 16H). 13C-NMR(CDCl3): δ 174.805, 174.476, 138.619, 129.210, 120.745, 57.178, 53.565, 51.139, 49.836, 45.565, 38.677, 37.449, 34.762, 34.518, 32.021, 21.060, 10.712. 29Si-NMR: δ 66.30. FT-IR (KBr): 1189 (νSi−O−Si). Anal. calcd for C372N124O64H612Si8 18H2O (4 branches unreacted dendrimer): C, 53.22; H, 7.56; N, 20.69. Found: C, 53.10; H, 7.72; N, 20.67. MALDI-TOF-MS (m/z, [M+H]+): 8444.13 (calculated for C392N136O60H640Si8); 8448.12 (observed).

Potentiometric titration

Potentiometric titrations were performed at room temperature using a pH meter (Horiba D-52, Horiba, Kyoto, Japan) and a glass electrode. Titrations were performed with 0.1 M NaOH in 2.0 M NaCl with constant number of imidazole units in the dendrimers. The pH of each aqueous solution of 1-methylimodazole (0.32 mmol), POSS-Im16 (0.02 mmol) and POSS-Im32 (0.01 mmol), which corresponds to 0.32 mmol of imidazole in 60 ml of distilled water with 2.0 M NaCl, was adjusted to 3 with 0.5 M HCl.

Photometric determination

Aqueous CuSO4 solutions were added to aqueous dendrimer solutions that ranged in concentration from 0.0025 to 0.2 mM. The concentration of the CuSO4 solution was 0.2 mM for the POSS-Im16 solution and 0.1 mM for the POSS-Im32 solution. In the continuous variation experiment, mixed solutions were prepared such that the total reagent concentration of Cu2+ ions and dendrimers were maintained at 0.5 mM.

Results and Discussion

Synthesis of imidazole-terminated first- and second-generation POSS-core dendrimers

Imidazole-terminated first- and second-generation POSS-core dendrimers were synthesized with 1-(3-aminopropyl)imidazole from POSS-OMe16 and POSS-OMe32, respectively, by the ester–amide exchange reaction (Scheme 1 and 2). As silsesquioxane units are typically unstable in basic compounds, we examined the amidation condition of POSS-OMe16 with 1-(3-aminopropyl)imidazole by monitoring the presence of the methoxy group in 1H NMR spectra (Figure 1). The rate of the ester–amide exchange reaction increased with increasing reaction temperature, and the silsesquioxane unit in POSS-Im16 decomposed at 40 °C or above according to the 1H NMR analysis. Thus, the reaction temperature should be <40 °C to prepare the dendrimer without decomposing the POSS core. However, no decomposition was observed at 70 °C when using POSS-OMe32. These results suggest that increasing generation numbers inhibits decomposition by basic compounds because they can no longer attack the silsesquioxane core. After the solvent was removed under reduced pressure, both products were purified by preparative size exclusion chromatography to remove unreacted compounds.

Figure 1
figure 1

Time-dependent conversion of the methyl ester of (a) POSS-OMe16 and (b) POSS-OMe32 with 1-(3-aminopropyl)imidazole at different reaction temperatures. (Panel a) : 25 °C, : 30 °C, •: 40 °C, : 50 °C. (Panel b) •: 40 °C, : 50 °C, □: 60 °C, ♦: 70 °C. POSS, poly(amidoamine)-type oligomeric silsesquioxane.

The structures of POSS-Im16 and POSS-Im32 were identified based on multinuclear NMR data (Supplementary Figures S1–S6). The degrees of functionalization of dendrimers with the imidazole unit were estimated using 1H NMR. The areas of the peaks of the aromatic protons on the imidazole ring were calculated and compared against values from the methylene protons on the POSS core and values from the remaining methoxy protons to determine the degrees of functionalization. Both dendrimers had high degrees of functionalization, which were 98 and 90% for POSS-Im16 and POSS-Im32, respectively. The 13C NMR spectra of both dendrimers confirmed their structures. In the 29Si NMR spectra, peaks of the POSS units for POSS-Im16 and POSS-Im32 were observed at −66.45 and −66.30 p.p.m., respectively. No other peaks were detected, indicating that the POSS cores did not decompose.

The MALDI-MS spectrum of POSS-Im16 showed the presence of 3 peaks at 3745.909, 3766.081 and 3785.83 m/z, the first of which is very close to the molecular mass (3749.136 g mol−1) of the fully functionalized dendrimer (Supplementary Figure S7). The latter two molecular ion peaks are closed to the molecular mass of the dendrimers associated with one H2O molecule (3767.152 g mol−1) and two H2O molecules (3785.168 g mol−1). The elemental analysis of POSS-Im16 matched the fully functionalized dendrimer associated with 12 H2O molecules. The MALDI-MS spectrum of POSS-Im32 contained a large number of signals (Supplementary Figure S8). The highest molecular ion peak at 8448.21 m/z is very close to the molecular mass of the fully functionalized dendrimer (8443.12 g mol−1). The peaks at 8271.9 and 8092.4 m/z matched the molecular mass of the dendrimers missing 2 and 4 branches (8256.85 and 8070.58 g mol−1), respectively. The elemental analysis of POSS-Im32 also matched the dendrimer missing 4 branches, which was associated with 18 H2O molecules. These observations agree well with the results on the degrees of functionalization of POSS-Im16 and POSS-Im32 obtained from 1H NMR, as described above.

pH response

Sensitivities of dendrimer phase transitions to pH were evaluated by changes in transmittance intensity of the dendrimer solutions. The initial dendrimer solutions were prepared at pH=6.5 with 0.1 M HCl, and the solution pH values were adjusted by adding 0.1 M NaOH solution. POSS-Im16 with a concentration of 0.02 mmol and POSS-Im32 with a concentration of 0.01 mmol corresponding to 0.32 mmol of the imidazole unit were dissolved in 60 ml of distilled water. Transmittance of the POSS-Im32 solution drastically decreased when the solution pH was >7.2 (Figure 2). This was attributed to the hydrophilic-to-hydrophobic transition arising from deprotonation of the imidazole groups. Figure 3 shows pH dependence of the turbidity of the dendrimer solutions measured at a wavelength of 600 nm. Transmittances of the dendrimer solutions reversibly changed with pH without hysteresis. However, in the case of POSS-Im16, the solution remained transparent in the range of the changing pH. The turbidity point occurred beyond the pH of 8.0 as the amount of POSS-Im16 increased to 0.06 mmol. These results suggest that pH of the turbidity points decreased with increasing generation numbers.

Figure 2
figure 2

Photographs of POSS-Im32 solution (0.17 mmol l−1) with different pH. POSS, poly(amidoamine)-type oligomeric silsesquioxane. A full color version of this figure is available at Polymer Journal online.

Figure 3
figure 3

Dependence of the turbidity of the aqueous POSS-Im32 solutions (0.01 mmol) (□) and the POSS-Im16 solution (0.02 mmol) () on pH. The dendrimers are dissolved in 60 ml of distilled water. The inset is the POSS-Im16 dendrimer (0.06 mmol) dissolved in 60 ml of distilled water. POSS, poly(amidoamine)-type oligomeric silsesquioxane.

The pH titration results of POSS-Im16, POSS-Im32 and 1-methylimidazole are shown in Figure 4. Although the dendrimers were polyelectrolytes, they had well-defined one-step titration curves similar to those of 1-methylimidazole, a monomeric imidazole compound. According to Figure 4, the pKa values of 1-methylimidazole, POSS-Im16 and POSS-Im32 were 7.3, 7.0 and 6.7, respectively. These values decreased as the number of imidazole units per molecule increased, suggesting that deprotonation of the ammonium groups of the later generation is easier than that of the earlier generation. This is caused by electrostatic repulsions between the neighboring imidazole groups due to their close-packed structures.15 Crooks and colleagues16 reported that the slopes of titration curves of imidazole-terminated poly(propyleneimine) dendrimers did not have a sharp break, making it difficult to determine their pKa. They further suggested that the imidazole groups on the surfaces of these dendrimers are communicating with one another and the branched groups. However, the present POSS-core dendrimers had well-defined one-step titration curves. The protonation amount matched the total protonation amount of the surfaces and the branched amines, suggesting that the terminal imidazole groups and the branched amines are isolated; in addition, they inhibited the association of neighboring branches due to the POSS-core structures. Such simple acid–base behavior provides the sharp response to pH shown in Figure 3. The degrees of protonation at the turbidity points for POSS-Im16 and POSS-Im32 are 0.19 and 0.28, respectively. These results indicate that increasing the generation number promotes aggregation in an aqueous solution at lower pH values with a higher degree of protonation.

Figure 4
figure 4

Potentiometric titrations of 1-methylimidazole (), POSS-Im16 () and POSS-Im32 (□) in 2.0 M NaCl solution with 0.1 M NaOH solution. The imidaozole unit of the samples is constant at 0.32 mmol. POSS, poly(amidoamine)-type oligomeric silsesquioxane.

Coordination behavior

The complexing behaviors of POSS-Im16 and POSS-Im32 with Cu2+ ions were studied. The coordination geometry of Cu2+ complexation systems is expected at the position of maximum absorption of the d–d transition.17, 18, 19, 20, 21 A Cu2+–N4 complex involving four tertiary amine groups or amide nitrogens is a possible type of complex when maximum absorption of the d–d transition is observed at 600 nm. A Cu2+–N2O2 complex involving two oxygen groups and two tertiary amine groups is a possible type of complex when the absorption maximum of the d–d transition is >650 nm. Earlier studies that reported the d–d transition 700 nm indicated that the ligand is coordinated with two tertiary nitrogen atoms and two amide oxygen atoms in the equatorial plane and that one amide oxygen atom is axially coordinated.22 To examine the coordination modes of the dendrimers with Cu2+ ions, we performed spectrophotometric titration and kept the amount of Cu2+ ions titrated into the solutions at 0.2 mM for POSS-Im16 and 0.1 mM for POSS-Im32. When the concentration ratio between POSS-Im16 and the Cu2+ ions decreased, the positions of the absorption maxima shifted gradually from 604 nm to a longer wavelength at 763 nm (Figure 5a). The shift in the absorption maxima with decreasing dendrimer amounts indicates that the coordination modes of POSS-Im16 shifted from the Cu2+–N4 complex to the Cu2+–N2O2 complex through intermediate coordination states.19 On the other hand, the absorption maxima remained at 600 nm regardless of the Cu2+ ion and POSS-Im32 concentration (Figure 5b). When the dendrimer concentration was low, that is, [imidazole]/[Cu2+]=1.6 and 0.8, the absorption peak had a maximum at 600 nm with an additional shoulder at 800 nm. This indicates that only the Cu2+–N4 complexation mode exists between Cu2+ ions and POSS-Im32.

Figure 5
figure 5

Spectrophotometric titrations of (a) POSS-Im16 and (b) POSS-Im32 against the Cu2+ ions. The amount of Cu2+ ions titrated into the solutions is 0.2 mM for POSS-Im16 and 0.1 mM for POSS-Im32. [dendrimer]/[Cu2+]=(a) 0.025, (b) 0.05, (c) 0.125, (d) 0.25, (e) 0.50, (f) 0.75, and (g) 1.00. A full color version of this figure is available at Polymer Journal online.

Figure 6 shows the spectrophotometric titration plots. The POSS-Im16 and POSS-Im32 samples were measured spectrophotometrically at the absorption wavelength of 604 nm. The absorption intensities increased as the concentration ratios between POSS-Im16 and Cu2+ ions increased up to a value of 0.25; similarly, the absorption intensity increased with the concentration ratio between POSS-Im32 and the Cu2+ ions up to a ratio of 0.20. The data indicate that the binding capacities of Cu2+ to POSS-Im16 and POSS-Im32 were approximately four and five Cu2+ ions, respectively. Additional studies using the more precise continuous variation method were performed to confirm the binding capacities Cu2+ to the dendrimers (Figure 7 and Supplementary Figure S9). Solutions containing Cu2+ ions and dendrimers in varying concentrations were prepared, and the total reagent concentration ([Cu2+]+[dendrimer]) was maintained at 0.5 mM for the solution with either POSS-Im16 or POSS-Im32. The absorbance of each Cu2+-dendrimer solution was plotted against the molar ratio between the dendrimer and the total concentration of the reagent. The absorptions at 604 nm for POSS-Im16 and POSS-Im32 were used for the Job plots. The Job plot at a longer absorption wavelength of 763 nm for POSS-Im16 was also drawn. The data for POSS-Im16 in Figure 7 show that the inflection points (X) of the Job plots were observed at X=0.15 for 763 nm and at X=0.20 for 604 nm (Table 1), indicating the presence of complexes with 5.7 and 4.0 Cu2+ ions, respectively. Although 16 imidazole nitrogens on the surface and 8 tertiary amines in the branch can potentially coordinate with Cu2+ ions, the present data suggest that the amine units in POSS-Im16 rarely formed Cu2+–N2O2 and Cu2+–N4 complexes because the movement of the terminal groups and branch units was suppressed. Using the inflection points on the plots for the Cu2+–N2O2 and Cu2+–N4 complexes, it was calculated that 2.8 and 4.0 terminal imidazole units coordinated with one Cu2+ ion. According to these data, the dendrimer likely coordinated with most of Cu2+ ions at their imidazole nitrogen, and no tertiary amines in the branches were used for coordination because the interior amines were sterically hindered. For POSS-Im32, the inflection point at the maximum absorbance was 0.15, indicating that 5.7 Cu2+ ions coordinated with one dendrimer and 5.6 terminal imidazole units coordinated with one Cu2+ ion.

Figure 6
figure 6

Spectrophotometric titration plots: absorption at 604 nm as a function of the number of POSS-Im16 () and POSS-Im32 (□) molecules per Cu2+ ions. The concentration of Cu2+ ions is maintained at 0.2 mM for POSS-Im16 and 0.1 mM for POSS-Im32. POSS, poly(amidoamine)-type oligomeric silsesquioxane.

Figure 7
figure 7

Job plots between the dendrimers and CuSO4. [Cu2+] + [dendrimer] is 0.5 mM for POSS-Im16 (: 604 nm; : 763 nm) and 0.5 mM for POSS-Im32 (□: 604 nm). POSS, poly(amidoamine)-type oligomeric silsesquioxane.

Table 1 The Cu2+-to-ligand stoichiometry from the Job plots as shown in Figurea

Conclusions

We have successfully synthesized imidazole-terminated first- and second-generation poly(amidoamine)-type POSS-core dendrimers. The results of pH titrations for POSS-Im16 and POSS-Im32 showed well-defined one-step titration curves identical to the curves of monomeric imidazole compounds, such as 1-methylimidazole. The terminal functional groups on the POSS-core dendrimers showed chemical behaviors similar to those of the corresponding monomeric molecules in the solution. This suggests that the terminal functional groups are isolated, and they inhibit the movement of neighboring branches. Such simple acid–base behavior caused dendrimer phase transitions to be highly sensitive to pH changes. The pKa values of 1-methylimidazole, POSS-Im16 and POSS-Im32 were 7.3, 7.0 and 6.7, respectively. These values decreased with increasing generation numbers, suggesting that deprotonation of the imidazolium and ammonium groups is easier in the later generation than in the earlier generations because of the electrostatic repulsions between the neighboring imidazole groups caused by their close-packed structures. According to the spectrophotometric study, we found that coordination modes of POSS-Im16 changed from the Cu2+–N2O2 complex to the Cu2+–N4 complex as concentrations of the dendrimer increased; in addition, only one complexation mode (Cu2+–N4) existed between Cu2+ ions and POSS-Im32. These results suggest that the terminal functional groups in the first-generation POSS-core dendrimer are isolated and inhibit the movement and association of neighboring branches; the functional groups in the second-generation dendrimers are closely positioned on the external surfaces due to fewer entangled branches. These characteristics of the POSS-core dendrimers could provide unique functions that conventional dendrimers could not. Further studies to reveal additional characteristics of the POSS-core dendrimers are underway.

scheme 1

Synthesis of POSS-Im16.

scheme 2

Synthesis of POSS-Im32.