Self-assembly of hyperbranched spheres Wilhelm T. S. Huck, Frank C. J. M. van Veggel* and David N. Reinhoudt* L aboratory of Supramolecular Chemistry and T echnology, T wente University, P.O. Box 217, 7500 AE Enschede, T he Netherlands A new type of building block with two coordinatively unsaturated palladium centres has been described that self-assembles in nitromethane solution and disassembles when acetonitrile is added.The resulting hyperbranched, organopalladium spheres have a remarkably narrow size distribution as was evidenced by light-scattering, AFM and TEM measurements. Variation of the structure of the building blocks showed the possibility to vary the size of the self-assembled spheres between 100 and 400 nm. There is an ongoing demand for the development of nanosize tures have been synthesized and the resulting assemblies have been studied by quasi-elastic light scattering (QELS), atomic materials.Especially, the electronics industry is searching for force microscopy (AFM) and transmission electron microscopy nanosize architectures permitting increased processing speeds, (TEM). In this paper we present a general methodology to and higher levels of integration.Current nanophysical ‘enginsynthesize spherical particles in the range 100–400 nm. eering down’ approaches can routinely fabricate structures in the range of 0.1–0.2 mm. Alternatively, chemists exploit a ‘bottom up’ approach to synthesize nanosize devices with Experimental molecular precision. In this respect, synthetic organic chemistry Melting points were determined with a Reichert melting point oers a powerful and versatile tool for the synthesis of large apparatus and are uncorrected. 1H NMR and 13C NMR and complex molecules with a variety of functions and shapes. spectra were recorded in CDCl3 (unless indicated otherwise) However, the synthesis of very large molecules using only with Me4Si as internal standard on a Bruker AC 250 spec- covalent bonds is less practical.Non-covalent interactions can trometer. Mass spectra were recorded with a Finnigan MAT be exploited to connect building blocks in the construction of 90 spectrometer using m-nitrobenzyl alcohol as a matrix. THF much larger architectures. The rational use of non-covalent was freshly distilled from Na/benzophenone, hexane (referring interactions requires the understanding of the principles of to light petroleum with bp 60–80 °C) and CH2Cl2 fromK2CO3.self-assembly.1 Our current research aims at the development Nitromethane was washed with 1 M HCl and water and of new methods of self-assembly that will lead to large struc- distilled from CaCl2. NaH was a 50% dispersion in mineral tures.2 Dendrimers represent a class of polymers that combine oil and was used after washing with hexane.Other chemicals a high molecular mass with a well defined spherical shape of were of reagent grade and were used as received. Column nanometre dimensions. Sequential synthesis, either divergent3 chromatography was performed with silica gel 60H or convergent,4 can be used to grow dendrimers stepwise, (0.005–0.040 mm) from Merck.[Pd(MeCN)4 ][BF4]2,8 5- generation after generation. Alternatively, instead of using this hydroxyisophthalic acid dimethyl ester,9 and a,a¾,a-tribromo- laborious method a one-pot polymerization of suitable building mesitylene10 were prepared according to literature procedures. blocks yields less regular hyperbranched organic polymers. QELS was performed with a Malvern PCS 100 goniometer Recently, Fre� chet et al.used branched monomers that contain (Malvern Instruments, Malvern, England) at an angle of 90°, polymerizable groups at both ends, so-called AB2-type mon- with an Adlas Model DPY 305 II, 50 mW, continuous wave omers, that yield hyperbranched polymers with a relatively narrow size distribution and high molecular masses.5 Recently, several groups have used transition metals to build small metallodendrimers.6 Our approach is to design AB2-type building blocks for the self-assembly of hyperbranched coordination polymers.7 These building blocks should contain all information necessary to form well defined three-dimensional hyperbranched structures.In each building block 1, two kinetically inert Pd centres and one kinetically labile CH2CN ligand, are present.In a coordinating solvent, like MeCN, a solvent molecule is weakly coordinating to the fourth coordination site and hence the building blocks remain monomeric. Removal of this solvent and introduction of a non-coordinating solvent, such as MeNO2, initiates intermolecular coordination of the labile ligands. This means that a new building block is added to the starting nucleus generating two new sites for further growth.Schematically the self-assembly process is shown in Fig. 1. In a dendritic model, the centre of the self-assembled spheres will be less dense than the surface. Therefore, the size of the 1 building blocks should be related to the size of the assemblies Fig. 1 Schematic representation of the self-assembly of hyperbranched polymers formed.For this reason, building blocks with dierent struc- J. Mater. Chem., 1997, 7(7), 1213–1219 1213diode pumped YAG laser, wavelength 532 nm (Adlas, Lu�beck, layer was concentrated under reduced pressure. The TBDMS groups were then removed by dissolving the thioether in THF Germany), and an ALV 5000 multiple tau digital correlator (ALV, Langen, Germany).The correlation function was trans- (100 ml) and adding 1 equiv. of CsF (2.0 g, 0.013 mol). After stirring overnight at 50°C, THF was evaporated and the formed into a diameter distribution with the CONTIN program. The instrument used for atomic force microscopy (AFM) residue dissolved in dichloromethane followed by washing with brine, drying over MgSO4 and concentration in vacuo.Column was a Nanoscope III operating in constant force (ca. 50 nN) mode with cantilever force constants of 0.58 N m-1 and a chromatography (silica gel, eluent CH2Cl2) gave pincer ligand 6a as a colourless oil. Yield 3.38 g (77%); 1H NMR d, 7.29–7.16 home-made supertip. Samples for transmission electron microscopy (TEM) were prepared by slow evaporation of a (m, 10H, SPh), 6.79 (s, 1H, ArH), 6.64 (s, 2H, ArH), 4.89 (br s, 1H, OH), 4.00 (s, 4H, CH2S); 13C NMR d, 155.7, 139.5, nitromethane solution on a carbon-coated copper grid. 131.2, 129.9, 128.9, 128.6, 126.4, 121.8, 114.6, 38.7; EIMS m/z, 5-tert-Butyldimethylsilyloxyisophthalic acid dimethyl ester 3 338.078 (M+, calc. for C20H18OS2: 338.080). TBDMSiCl (16.0 g, 0.106 mol) was dissolved inCH2Cl2 (50 ml) 3,5-Bis(-1-naphthylthiamethyl)phenol 6b.White solid. Yield and was slowly added to a solution of 5-hydroxyisophthalic 76%, mp 75–77°C; 1H NMR d, 7.88–7.68 (m, 8H, Snaphthyl), acid dimethyl ester 2 (11.18 g, 0.053 mol), Et3N (6.99 g, 7.48–7.30 (m, 6H, Snaphthyl), 6.90 (s, 2H, ArH), 6.72 (s, 1H, 0.069 mol), and DMAP (catalytic amount) in CH2Cl2 (150 ml) ArH), 4.09 (s, 4H, CH2S) 13C NMR d, 155.8, 139.4, 133.7, at 0°C.After stirring overnight at room temperature, the 131.9, 128.4–125.8, 122.0, 114.8, 38.5; EIMS m/z, 438.402 (M+, mixture was washed with 1 M HCl, a saturated aqueous calc. for C28H22OS2 : 438.601). solution of NaHCO3, and brine. After drying (MgSO4), the solvent was evaporated and 3 was obtained as a white solid 3,5-Bis(tert-butylthiamethyl)phenol 6c.Colourless oil, solidi- upon drying under high vacuum. Yield 15.3 g (89%), mp fied slowly upon standing. Yield 47%, mp 80–82 °C; 1H NMR 68–70°C; 1H NMR d, 8.29 (s, 1H, ArH), 7.68 (s, 2H, ArH), d, 6.89 (s, 1H, ArH), 6.70 (s, 2H, ArH), 4.68 (bs, 1H, OH), 3.67 3.93 (s, 6H, OCH3), 1.00 (s, 9H, But), 0.23 (s, 6H, SiCH3); 13C (s, 4H, CH2S), 1.30 (s, 18H, CH3); 13C NMR d, 155.0, 140.5, NMR d, 166.1, 156.0, 131.8, 125.4, 123.7, 52.6, 25.6, 18.2, -4.5; 122.0, 114.5, 33.2, 30.9, 23.6; EIMS m/z, 298.142 (M+, calc.for FAB MS m/z, 324.1 (M+, calc. 324.1); Anal. Calc. for C16H26OS2 : 298.143). C16H24O5Si: C, 59.23; H, 7.46. Found: C, 59.55; H, 7.60%. 3,5-Bis(ethylthiamethyl)phenol 6d. Pale yellow oil. Yield 3,5-Bis(hydroxymethyl )phenol tert-butyldimethylsilther 4 23%; 1H NMR d, 6.83 (s, 1H, ArH), 6.69 (s, 2H, ArH), 5.21 LiAlH4 (4.8 g, 0.12 mol) was suspended in dry THF (200 ml) (br s, 1H, OH), 3.62 (s, 4H, CH2S), 2.44 (q, 4H, J 8.4 Hz, and a solution of diester 3 (20.0 g, 0.062 mol) in THF (100 ml) SCH2CH3), 1.23 (t, 6H, J 8.4 Hz, CH3); 13C NMR d, 155.8, was slowly added.The mixture was stirred overnight at room 140.5, 121.8, 114.4, 35.6, 25.6, 14.4; EIMS m/z, 242.080 (M+, temperature after which THF was evaporated.The resulting calc. for C12H18OS2 : 242.080). paste was dissolved in dichloromethane (200 ml) and cooled to 0°C and 2 M HCl (100 ml) was added, after which the layers a,a¾-Dibromo-a-cyanomesitylene 7. To a solution of a,a¾,a- were separated. Extraction of the aqueous layer with dichloro- tribromomesitylene (5.0 g, 0.014 mol) in MeCN (200 ml) were methane (3×100 ml) and drying of the combined organic added powdered KCN (0.91 g, 0.014 mol), 18-crown-6 (0.25 g, layers gave, after removal of the solvent, pure 4 as a white 0.95 mmol), and water (0.5 ml).The mixture was refluxed for solid. Yield 15.3 g (93%), mp 99–100 °C; 1H NMR d, 6.94 (s, 48 h after which the solvent was evaporated and the residue 1H, ArH), 6.77 (s, 2H, ArH), 4.62 (s, 4H, CH2O), 0.99 (s, 9H, was taken up in CH2Cl2 (100 ml). After washing with brine But), 0.20 (s, 6H, SiCH3); 13C NMR d, 156.0, 142.8, 118.1, 64.9, and drying over MgSO4 the crude reaction mixture was 25.7, 18.2, -4.4; FAB MS m/z, 268.0 (M+, calc. 268.1); Anal. concentrated in vacuo and purified by column chromatography Calc.for C14H24O3Si: C, 62.64; H, 9.01. Found: C, 62.59; (SiO2, eluent CH2Cl2). This yielded pure 7 as a colourless oil H, 9.19%. which solidified upon standing. Yield 0.85 g (20%), mp 67–69°C; 1H NMR d, 7.40 (s, 1H, ArH), 7.32 (s, 2H, ArH), 3,5-Bis(chloromethyl ) phenol tert-butyldimethylsilyl ether 5 4.47 (s, 4H, CH2Br), 3.72 (s, 2H, CH2CN); 13C NMR d, 139.6, 131.3, 129.3, 128.5, 117.3, 32.0, 23.4; EIMS m/z, 302.9 (M+, Diol 4 (4.0 g, 0.015 mol) and Et3N (6.2 g, 0.045 mol) were calc. 303.0); IR (KBr) 2252 cm-1 (CON); Anal. Calc. for dissolved in dry CHCl3 and cooled to 0°C. Mesityl chloride C10H9NBr2: C, 39.64; H, 2.99; N, 4.62. Found: C, 39.69; H, (MsCl) (3.44 g, 0.045 mol) was slowly added at 0°C after 2.94; N, 4.50%. which the reaction mixture was slowly heated to 50°C.Stirring overnight and subsequent washing with 1 M NaOH and 1 M General procedure for the synthesis of thioether ligands 8a–d HCl, drying over MgSO4 and evaporation of the solvent gave 5 as a colourless oil. Yield 4.2 g (92%); 1H NMR d, 6.80 (s, a,a¾-Bis[3,5-bis(phenylthiamethyl)phenyloxy]-a-cyano- 1H, ArH), 6.62 (s, 2H, ArH), 4.32 (s, 4H, CH2Cl), 0.78 (s, 9H, mesitylene 8a.Compound 6a (0.3 g, 0.75 mmol) was stirred But), 0.00 (s, 6H, SiCH3); 13C NMR d, 156.1, 139.2, 121.4, with K2CO3 (0.13 g, 0.94 mmol) in MeCN (50 ml) for 1 h at 118.9, 45.7, 26.6, 24.7, -3.5; EIMS m/z, 304.081 (M+, calc. for room temperature (r.t.). Spacer 7 (0.13 g, 0.38 mmol) was added C14H22SiCl2O: 304.081). and the reaction mixture was stirred for 4 days at r.t. The reaction was concentrated under reduced pressure and General procedure for the synthesis of thioethers 6a–d dichloromethane (50 ml) was added to the residue.The organic layer was washed with brine, dried over MgSO4 and concen- 3,5-Bis(phenylthiamethyl)phenol 6a. Thiophenol (5.77 g, 0.052 mol) was added to a stirred suspension of NaH (2.52 g, trated. Column chromatography (silica gel, CH2Cl2–hexanes 80520) gave pure 8a as a colourless oil.Yield 0.26 g (72%); 0.104 mol) in THF (200 ml) and the mixture was stirred for 1 h to allow formation of the sodium thiophenolate salt. To 1H NMR d, 7.31 (s, 1H, ArCNH), 7.29–7.16 (m, 22H, SPh+ArCNH), 6.85 (s, 2H, ArOH), 6.79 (s, 4H, ArOH), 4.94 (s, the resulting milky solution dichloride 5 (4.0 g, 0.013 mol) was added and the reaction mixture was stirred overnight at 50°C. 4H, CH2O), 4.03 (s, 8H, CH2S), 3.73 (s, 2H, CH2CN); 13C NMR d, 158.6, 139.3, 138.4, 136.1, 130.7, 129.9, 128.9, 126.4, After removal of the solvent the crude reaction mixture was dissolved in dichloromethane (200 ml) and washed with brine 126.0, 122.2, 117.6, 114.0, 69.2, 38.9, 23.5; EIMS m/z, 817.216 (M+, calc. for C50H43NO2S4 : 817.218); IR (KBr) 2251 cm-1 (100 ml).To remove excess of thiol the organic layer was washed with 2 M NaOH (100 ml). After drying, the organic (CON). 1214 J. Mater. Chem., 1997, 7(7), 1213–1219a,a¾-Bis[3,5-bis(1-naphthylthiamethyl)phenyloxy]-a-cyano- 0.14 g (20%), mp 131–132 °C; 1H NMR d, 7.38 (s, 1H, ArCNH), 7.32 (s, 2H, ArCNH), 6.68 (s, 4H, ArPdH), 4.97 (s, 4H, CH2O), mesitylene 8b.Colourless oil. Yield 70%; 1H NMR d, 7.80–7.59 (m, 16H, Snaphthyl), 7.47–7.32 (m, 12H, Snaphthyl), 7.30 (s, 4.10 (br s, 8H, CH2S), 3.81 (s, 2H, CH2CN), 1.69 (s, 36H, But); 13C NMR d, 156.2, 150.6, 138.5, 130.8, 126.5, 117.6, 108.4, 69.5, 1H, ArCNH), 7.21 (s, 2H, ArCNH), 6.95 (s, 2H, ArOH), 6.75 (s, 4H, ArOH), 4.74 (s, 4H, CH2O), 4.12 (s, 8H, CH2S), 3.62 (s, 52.0, 42.7, 30.6, 23.6; FAB MS m/z 984.3 [(M-Cl)+, calc. 983.6]; Anal. Calc. for C50H41NO2S4Pd2Cl2·H2O: C, 46.97; H, 2H, CH2CN); 13C NMR d, 158.7, 139.2, 138.3, 133.6, 131.9, 130.5, 128.8, 128.0, 127.8, 127.7, 126.5, 126.3, 125.9, 125.8, 122.3, 5.36; N, 1.38. Found: C, 47.58; H, 5.36; N, 1.38%. 114.2, 69.2, 38.8; EIMS m/z, 1017.280 (M+, calc. for C66H51NO2S4: 1017.280). Bis (PdMCl) complex 9d.Brownish solid. Yield 16%, mp 106–107 °C; 1H NMR d, 7.39 (s, 1H, ArCNH), 7.36 (s, 2H, a,a¾-Bis[3,5-bis(tert-butylthiamethyl)phenyloxy]-a-cyano- ArCNH), 6.68 (s, 4H, ArPdH), 4.98 (s, 4H, CH2O), 4.2 (br s, 8H, mesitylene 8c. Slightly yellow oil. Yield 25%. 1H NMR d, 7.46 CH2S), 3.81 (s, 2H, CH2CN), 3.19 (q, 8H, J 8.4 Hz, SCH2CH3 ), (s, 1H, ArCNH), 7.37 (s, 2H, ArCNH), 6.96 (s, 2H, ArOH), 6.85 1.67 (t, 12H, J 8.4 Hz, CH3); FAB MS m/z 871.8 [(M-Cl)+, (s, 4H, ArOH), 5.08 (s, 4H, CH2O), 3.80 (s, 2H, CH2CN), 3.71 calc. 871.3]. Anal. Calc. for C34H41NO2S4Pd2Cl2·0.5C6H14: C, (s, 4H, CH2S), 1.33 (s, 36H, CH3); 13C NMR d, 158.7, 140.4, 46.74; H, 5.09; N, 1.47. Found: C, 46.33; H, 4.96; N, 1.66%. 138.6, 130.6, 126.4, 122.5, 113.9, 69.3, 42.9, 33.4, 30.9, 23.6; EIMS m/z, 737.339 (M+, calc.for C42H59NO2S4: 737.343). Results and Discussion a,a¾-Bis[3,5-bis(ethylthiamethyl)phenyloxy]-a-cyano- Synthesis of building blocks mesitylene 8d. Colourless oil. Yield 48%; 1H NMR d, 7.47 (s, A convergent synthesis route has been exploited to connect 1H, ArCNH), 7.37 (s, 2H, ArCNH), 6.88 (s, 2H, ArOH), 6.85 (s, two pincer ligands to a spacer containing the kinetically labile 4H, ArOH), 5.06 (s, 4H, CH2O), 3.77 (s, 2H, CH2CN), 3.67 (s, (cyano) ligand.Key intermediates 6 were synthesized in six 8H, CH2S), 2.44 (q, 8H, J 8.4 Hz, SCH2CH3 ), 1.22 (t, 12H, J steps from 5-hydroxyisophthalic acid as shown in Scheme 1. 8.4 Hz, CH3); 13C NMR d, 158.7, 141.3, 138.5, 130.7, 127.8, The first step is the esterification of 5-hydroxyisophthalic acid 126.1, 122.2, 120.7, 117.6, 113.7, 69.2, 53.5, 37.5, 25.4, 23.5, 14.4; according to literature procedures.8 The phenolic group was FAB MS m/z, 625.620 (M+, calc.for C34H43NO2S4: 625.959). protected with TBDMSiCl in 95% yield and subsequently the esters were reduced with LiAlH4 to the diol 4 in 70% yield. General procedure for the cyclopalladation of the pincer ligands The diol was stirred overnight with MsCl–Et3N at 50°C which 8a–d and conversion into the chloride complexes 9a,b and d gave complete conversion into the dichloride 5.The thioether Bis(PdMCl ) complex 9a. Ligand 8a (0.50 g, 0.62 mmol) was functions were introduced by stirring the appropriate thiol dissolved in acetonitrile (150 ml) and placed under an Ar with NaH in THF to generate the sodium thiolate, and atmosphere.Solid [Pd(MeCN)4 ][BF4 ]27 (0.54 g, 1.22 mmol) subsequent addition of the dichloride 5. This reaction yielded was added in one portion. The orange solution was warmed the SCS pincer ligands with the p-hydroxy functions protected. to 40°C and stirred until the colour changed to pale yellow. The TBDMSi ether was deprotected with CsF to give the SCS After cooling to r.t.and evaporation of the solvent the yellow pincer type ligand 6 in 10% overall yield from 5-hydroxyisoph- cyclopalladated product was obtained in quantitative yield. thalic acid. The thioethers 6 were coupled to spacer 7 which This product was dissolved in CH2Cl2–MeCN (351, 100 ml) introduces the weakly coordinating cyano group. The spacer and the reaction mixture was stirred vigorously for 30 min 7 was prepared from a,a¾,a-tribromomesitylene by refluxing with brine (100 ml).The layers were separated and the organic in acetonitrile for two days with powdered KCN. From the layer was washed with water (100 ml) and evaporated to resulting mixture 7 could be isolated in 20% yield. The dryness. Purification by column chromatography (silica gel, formation of the benzylic ethers proceeded in rather poor CH2Cl2–MeOH, 9555) gave 9a as a yellow solid (1.18 g, 50%), yields because 7 slowly decomposed.Ligands 8a–d were cyclo- mp 132–133 °C; 1H NMR d, 7.81–7.74 (m, 8H, SPh), 7.39 (s, palladated with [Pd(MeCN)4][BF4]2 in acetonitrile in high 1H, ArCNH), 7.35–7.29 (m, 14H, SPh+ArCNH), 6.64 (s, 4H, yields. Prior to self-assembly studies, the cationic solvento ArPdH), 4.95 (s, 4H, CH2O), 4.5 (bs, 8H, CH2S), 3.76 (s, 2H, complexes were converted into chloropalladium complexes by CH2CN); 13C NMR d, 156.4, 152.3, 150.2, 138.4, 132.3, 131.4, stirring a solution in CH2Cl2 and MeCN with brine.This 130.9, 129.8, 129.7, 126.6, 126.0, 117.7, 109.2, 69.5, 51.7, 23.6; made purification easier and allowed the simple introduction FAB MS m/z, 1064.3 [(M-Cl)+, calc. 1064.3]; IR (KBr), of various non-coordinating anions (vide infra). The tert-butyl 2252 cm-1 (CON). Anal. Calc. for C50H41NO2S4Pd2Cl2·H2O: derivative 9c was prepared by stirring the cationic palladium C, 53.72; H, 3.88; N, 1.25. Found: C, 53.65; H, 3.73; N, 1.64%. complex from 8c with NMe4Cl, as reaction with brine only gave intractable polymeric products.Bis(PdMCl ) complex 9b. Orange solid. Yield 56%, mp The 1H NMR spectra of 9a–d in CD3CN show a broad 160–162 °C; 1H NMR d, 8.30 (s, 4H, Snaphthyl), 7.94–7.76 (m, singlet for the CH2SR protons at d 4.6 because of the slow 16H, Snaphthyl), 7.52–7.49 (m, 8H, Snaphthyl), 7.40 (s, 1H, conformational interconversion of the palladium(II)-containing ArCNH), 7.34 (s, 2H, ArCNH), 6.67 (s, 4H, ArPdH), 4.99 (s, 4H, five-membered rings.11 The signal for the protons ortho to CH2O), 4.63 (br s, 8H, CH2S), 3.76 (s, 2H, CH2CN); 13C both donor atoms in 9a–d at d 6.85 is absent, indicating NMR d, 157.2, 150.3, 138.4, 133.4, 133.2, 131.2, 129.7, 129.3, complete cyclopalladation.The 1H NMR spectrum clearly 128.2, 127.8, 127.6, 127.1, 109.2, 69.5, 52.4, 23.1; FAB MS showed that no cyclopalladation had occurred at other m/z 1264.3 ([M-Cl]+, calc. 1264.6). Anal. Calc. for aromatic positions. C66H49NO2S4Pd2Cl2 ·2H2O: C, 59.33; H, 3.99; N, 1.05; S, 9.60. Self-assembly takes place when the CH2CN group of one Found: C, 58.93; H, 3.73; N, 1.18; S, 9.46%. building block coordinates intermolecularly to the Pd centre of another. Therefore the Pd centres were first activated by Bis(PdMCl ) complex 9c.The cyclopalladation of ligand 8c the replacement of the chloride by dierent non-coordinating (0.50 g, 0.68 mmol) was performed as described for 9a. The anions. Addition of one equivalent of the appropriate silver crude palladium complex was dissolved in CH2Cl2–MeCN salt activates the Pd centre by precipitation of AgCl in a fast (151, 50 ml) and excess NMe4Cl (0.50 g, 4.57 mmol) was added and quantitative reaction.† In acetonitrile solution a bis- in one portion.The reaction mixture was stirred overnight. After removal of the salts, the filtrate was evaporated to dryness. Purification by column chromatography (silica gel, † This has been demonstrated using 31P NMR spectroscopy in an analogous PCP pincer complex; unpublished results.CH2Cl2–MeOH, 9555) gave 9c as a brownish solid. Yield J. Mater. Chem., 1997, 7(7), 1213–1219 1215Scheme 1 Scheme 2 X-=BF4-, ClO4-, PF6-, triflate, tosylate, BPh4- acetonitrile complex·2X- (X-=BF4-, ClO4-, PF6-, triflate, indicating large structures. When small amounts of acetonitrile were added to these solutions all signals became sharp. This tosylate or BPh4-) is formed (Scheme 2).proves that the self-assembly of 1a–d and disassembly of the hyperbranched polymer is a reversible process. In order to Self-assembly and characterization obtain information on the size and shape the resulting assembl- The self-assembly process of 1a–d was initiated by elimination ies were further characterized by QELS, TEM and AFM.‡ of the acetonitrile ligands.After removal of the acetonitrile QELS measurements of nitromethane solutions of assemblies (the 1H NMR spectrum in CD3NO2 showed no acetonitrile), of 1a (X-=BF4-) showed particles with an average hydrody- the CH2CN groups of the bis-palladium complexes occupy the namic diameter of 180 nm using the CONTIN curve-fitting fourth coordination site. In the IR spectrum the coordination of the cyano group was confirmed by the characteristic shift of the CON stretch vibration from 2250 cm-1 for the mon- ‡ Gel permeation chromatography (GPC) was not successful.Probably omeric building blocks to 2290 cm-1 upon coordination.12 the highly charged assemblies have a strong interaction with the column material and are therefore dicult to characterize using GPC.The 1HNMRspectra of 1a–d showed broad peaks in CD3NO2, 1216 J. Mater. Chem., 1997, 7(7), 1213–1219aggregates than expected, but this might be explained by solvent or water coordination to the anion even inside the spheres, making the apparent size of this anion larger. Increasing the thioether bulk from ethyl to phenyl groups (X-=tosylate) gave smaller aggregates.The naphthyl thioether building blocks showed a similar trend as the phenyl thioethers. The larger naphthyl groups gave smaller spheres in all cases. The combination of naphthyl thioethers with triflate anions formed rather small aggregates. The general conclusion from these data is that larger anions and/or thioether groups give smaller assemblies. The size of the aggregates was also measured by contact 400 300 200 100 0 25 125 225 320 – – – anion volume/Å3 diameter/nm Fig. 2 Relationship between the sphere size and monomer or anion mode AFM. Samples were prepared by evaporation of a volume as measured by light-scattering (+, Et; *, Ph; l, But; nitromethane solution of self-assembled spheres 1a–d on a $, naphthyl) clean gold surface. A representative part of these surfaces is shown in Fig. 3(a)–( f ). In all cases, large, spherical objects Fig. 3 AFM pictures of self-assembled spheres of (a) 9a (X-=BF4-); (b) 9b (X-=BF4-); (c) 9b (X-=OTf-); (d) 9d (X-=BF4-); (e) 9a (X-=Tos); ( f ) 9a (X-=PF6-) program. This size is independent of the sample concentration were observed that correspond quite well with the dimensions determined by light-scattering.The average diameter for indicating that these are single particles and no clusters. From the peak width a standard deviation of approximately 30 nm assemblies of 1a (X-=BF4-), as found by the grain size analysis routine of the instrument software of these aggregates was estimated. By changing the anion from ClO4- to BPh4- the size of the spheres as measured by QELS decreased from is 205 nm, with a standard deviation of 30 nm.The aggregates seem to have a disc-like shape when studied with AFM. The 400 to 180 nm in diameter (Fig. 2). The sizes of the anions were calculated using the Connoly flattening might be caused by spreading of the spheres on the surface or by interaction of the sample with the AFM tip. Surfaces option implemented in the Cerius2 program package. 13 This gave arbitrary volumes which should be used only Repetitive scans of one spot ‘wiped’ the surface clean, which indicates that the organic material has a soft constitution. in a relative sense. The BF4- anion gave much smaller J. Mater. Chem., 1997, 7(7), 1213–1219 1217When a glass substrate was used instead of gold, the same A linear, non-branched palladium(II) complex containing one pincer complex and one cyanomethyl group was treated spheres were observed of roughly the same size.This indicates that the possible interaction of sulfur atoms in the building according to the same self-assembly procedures but AFM and TEM measurements did not show globular structures. This blocks and the gold surface did not significantly alter the morphology of the spheres.Grazing-angle FTIR spectroscopy supports our concept in which branching is essential. Indications for a dendritic structure, which results in a dense on a gold surface covered with the spheres showed the characteristic CON signal of coordinated cyano groups at 2289 cm-1 outer sphere, came from additional disassembly experiments. Light-scattering experiments show that when ca. 20 equiv. of in agreement with the bulk spectra. From AFM data sizes of 160 (±20) nm for 1b (X-=BF4-) and 110(±20) nm for 1b acetonitrile per building block are added to a nitromethane solution of the spheres, they slowly disassemble in the course (X-=OTf-) were measured respectively [Fig. 3(b) and (c)]. This is in good agreement with the QELS data (140 and of 10–15 min.The slow rate of disassembly indicates that acetonitrile cannot easily penetrate the outer shell of building 104 nm, respectively). In the case of 1a (X-=Tos) and 1a (X-=PF6-) extensive clustering made size determination blocks. When a larger nitrile like benzonitrile is used, disassembly is hardly observed, even after heating to 70°C unreliable [Fig. 3(e)–( f )]. The samples for TEM were used without additional shading for 15 min.Dendrimers with a similar dense shell with a solid-phase character have been reported by Meijer and or staining and the contrast results from the Pd centres that are distributed throughout the spherical assembly. In all pic- co-workers.15 The explanation for the formation of rather well defined tures, the only structures that could be observed had spherical shapes in the expected size range.Fig. 4(a) and (b) clearly show spheres and the relationship between the size of the spheres and the structure of the monomers and the non-coordinating globular aggregates in the range 150–200 nm for 1a and 1c (both with BF4- counter anions). This is in good agreement anions, remains speculative. In our dendritic model the anions will occupy the voids created by the branching of the mon- with the diameter measured with QELS.Fig. 4(a) shows globular structures with a light shell and a little darker nucleus. omers. When these cavities become too small the anions are forced out of the sphere and will occupy the surface, thereby This indicates that the structures are a little thicker in the middle than at the edges, as expected for spherical assemblies.blocking the assembly process. The size at which this occurs, is apparently dependent on both the bulk of the bis-palladium Energy dispersive X-ray spectrometry (EDX) revealed the presence of the elements Pd and S in these aggregates. The complexes as well as the size of the anion. TEM image for 1a (X-=Tos) shows assemblies of ca. 225 nm [Fig. 4(c)], which is also in good agreement with QELS data Conclusions (220 nm). As shown in Fig. 4(d), strong clustering is observed for 1d (X-=BF4-). By dividing the radius of the spheres by Building blocks that contain all information necessary for self- the size of the building blocks we estimate a number of roughly assembly give regular assemblies with a (relatively) small 50 ‘generations’.The outer layer of Pd complexes is not polydispersity. The self-assembled structures are held together occupied by MeCN ligands, as these are not observed in the via coordinative bonds. Introduction of bulky groups in the 1H NMR spectrum. Probably, water present in nitromethane building blocks and/or dierent sizes of non-coordinating is coordinating to these Pd centres.14 counter anions influences the outcome of the self-assembly process.In this way spherical assemblies are obtained ranging from 100 to 400 nm in diameter. We are grateful to Dr. E. G. Keim (Center for Materials Research, University of Twente) for TEM measurements and Dr. J. W. Th. Lichtenbelt (Akzo Nobel Central Research) for assistance with QELS measurements.We thank the Dutch Foundation for Chemical Research (SON) for financial support. References 1 D. Philp and J. F. Stoddart, Angew. Chem., Int. Ed. Engl., 1996, 35, 1155. 2 R. H. Vreekamp, J. P. M. van Duynhoven, M. Hubert, W. Verboom and D. N. Reinhoudt, Angew. Chem., Int. Ed. Engl., 1996, 35, 1215; W. T. S. Huck, F. C. J. M. van Veggel and D. N. Reinhoudt, Angew. Chem., Int. Ed.Engl., 1996, 35, 1213. 3 D. A. Tomalia, A. Naylor and W. A. Goddard III, Angew. Chem., Int. Ed. Engl., 1990, 29, 138. 4 C. J. Hawker and J. M. J. Fre� chet, J. Am. Chem. Soc., 1990, 112, 7638. 5 J.M.J.Fre�chet, M. Henmi, I. Gitsov, S. Aoshima, M. R. Leduc and R. B. Grubbs, Science, 1995, 269, 1080. 6 S. Campagna, G. Denti, S. Serroni, A. Juris, M. Venturi, V. Ricevuto and V. Balzani, Chem. Eur. J., 1995, 1, 211; S. Achar and R. J. Puddephatt, Angew. Chem., Int. Ed. Engl., 1994, 33, 847. 7 W. T. S. Huck, F. C. J. M. van Veggel, B. L. Kropman, D. H. A. Blank, E. G. Keim, M. M. A. Smithers and D. N. Reinhoudt, J. Am. Chem. Soc., 1995, 117, 8293; W. T. S. Huck, B. H. M. Snellink-Rue�l, J. W. Th. Lichtenbelt, F. C. J. M. van Veggel and D. N. Reinhoudt, Chem. Commun., 1997, 9. 8 A. Sen and L. Ta-Wang, J. Am. Chem. Soc., 1981, 103, 4627. 9 A. H. van Oijen, N. P. M. Huck, J. A. W. Kruijtzer, C. Erkelens, Fig. 4 TEM pictures of (a) 9a (X-=BF4-); (b) 9c (X-=BF4-); J. H. van Boom and R. J. M. Liskamp, J. Org. Chem., 1994, 59, 2399. (c) 9a (X-=Tos); (d) 9d (X-=BF4-) (bar represents 100 nm) 1218 J. Mater. Chem., 1997, 7(7), 1213–121910 F. Vo� gtle, M. Zuber and R. Lichtenhaler, Chem. Ber., 1973, 106, and H. J. C. Ubbels, J. Am. Chem. Soc., 1982, 104, 6609. 15 J. F. G. A. Jansen, E. M. M. de Brabander-van den Berg and 717. 11 A. J. Canty and N. J. Minchin, J. Chem. Soc., Dalton T rans., 1987, E. W. Meijer, Science, 1994, 266, 1226. 1477. 12 B. N. Storho and H. C. Lewis, Coord. Chem. Rev., 1977, 23, 1. Paper 6/08577F; Received 23rd December, 1996 13 Cerius2, Molecular Simulations Inc,Waltham, MA. 14 D. M. Grove, G. van Koten, J. N. Louwen, J. G. Noltes, A. L. Spek J. Mater. Chem., 1997, 7(7), 1213–1