Control of average size and size distribution in as-grown nanoparticle polymer composites of MSe (M=Cd or Zn) Stephen W. Haggata,a David J. Cole-Hamilton*a and John R. Fryerb aSchool of Chemistry, University of St. Andrews, St. Andrews, Fife, Scotland, UK KY16 9ST bDepartment of Chemistry, Glasgow University, Glasgow, Scotland, UK G12 8QQ The preparation of CdSe and ZnSe semiconductor nanoparticles in the quantum size regime (1–5 nm) within a polymer matrix is described. A carefully controlled reaction temperature and suitable choice of solvent is found to have a dramatic eVect on the size of the particles produced when a soluble pyridyl polymer adduct (a polymer that contains nitrogen and dimethylcadmium or dimethylzinc) is reacted with H2Se in solution.An expected change in colour (absorbance) of the material from black through to yellow is observed for cadmium selenide and dark yellow to light yellow for zinc selenide as the particles decrease in size.TEMs reveal that cadmium selenide particles are evenly distributed in the polymer film with a size distribution wider than analogous sulfide preparations recently reported. The CdSe particles exhibit size quantization and hence a blue shift of the band edge position is observed in the absorbance and photoacoustic spectra as the reaction temperature is decreased or the polymer solubility is improved.ZnSe nanoparticle growth has been found diYcult to control although a change in reaction temperature has an eVect on the particle size similar to that for CdSe. Semiconductor nanosized particles (or QDs) are small molecu- matrix in the past has usually involved the passing of a chalcogenide gas over the surface of a polymer blend or lar clusters in the size range of ca.ten to several hundred a° ngstro�ms in diameter and have attracted considerable co-polymer film which contains either organometallic blocks32,33a,b which have a complex synthesis, or coordinated research interest during the last several years because of their unique quantum eVects observed only at nanosized dimensions metal salts.34 The microphase separation domains in the block copolymer are able to control the particle dispersity and size and which lead to obvious diVerences from bulk macrocrystallites in terms of their electronic, optical and catalytic proper- distribution.However, these methods are carried out in a heterogeneous phase and consequently very low loading of ties.1–16 These materials have possible future applications as photocatalysts in photoreactions,4,13 in electro-luminescent semiconductor particles and varying metal/chalcogenide stoichiometry is observed if the morphology of the polymer film devices (electro-optics)8,9,10 for e.g.the development of flat panel luminescent displays; photoconductive and photovoltaic is poorly controlled. An alternative approach is to mix preprepared particles with suitable polymers and then use the com- devices11,12 in e.g. photocopiers and laser printers; all-optical (non-linear optical)14 devices in e.g. optical switches; and posite to form films.8,9,11,35 Methods in which the particle film composite is formed ‘in situ’ by using a soluble polymer/metal magneto-optics15 in e.g.erasable optical data storage. However, there are important criteria that have to be met before such complex adduct and reacting it with a suitable source of the group 16 element have been less explored, but is the basis for devices can be realised.The first is the preparation of a high concentration of monodisperse particles. The size and shape the successful production of nanoparticulates from aqueous solutions containing polyphosphates.36 In these systems, the of the particles has an influence on the wavelength of their absorbance and light emission which can be ‘tuned’ by altering polymer not only acts as the encapsulant for the particles, but is also capable of controlling the particle size by providing the particle size.Therefore, narrow size distributions are essential if pure colours are to be obtained in luminescence. A slow release of group 12 precursor and by binding to surface metal atoms to terminate the particle growth. Recently,10 the second, vital point is the control of the surface (grain boundary).A substantial percentage of crystallite atoms are surface copolymerisation of styrene and zinc methacrylate to form a soluble zinc containing microgel and subsequent reaction with atoms and these largely account for the chemical and physical properties of the particle. Therefore, in order to prevent higher H2S has produced ZnS nanocrystals in a polymer matrix. How luminescent devices can be made from such composite wavelength emission, surface states must be eradicated and this can be achieved by chemically binding the particles to a materials has also been described.10 Our previous work has described the functionalization of suitable material of higher band gap.4,5 Polymers are able to passivate the material and prevent particle agglomeration polybutadiene with various Lewis-base groups by homogeneous catalysis37–39 to form soluble polymer ligands.Further, whilst maintaining a good spatial distribution of particles. In some cases polymers are able to assist charge transfer in e.g. we have reported that the addition of group 12 metal alkyls forms soluble metal alkyl/polymeric adducts40 and we have photoconductive and electroluminescent devices.Here, it is essential to encapsulate the crystallites in a conducting polymer then observed their reaction with hydrogen sulfide gas41,42 to form semiconductor colloids. Polymer/semiconductor com- in order for charge extraction to proceed.11,12 There are many synthetic methods reported which target surface control and posite films can then be obtained after removing the solvent under vacuum or by filtration which has proved a simple, monodispersity and these include syntheses in colloidal suspensions, 17,18 solutions of single-molecule precursors,19,20 sol– flexible and alternative route to the preparation of nanoparticulate II/VI materials with narrow size dispersion.gels,21 zeolites,22 LB films,23 micelles7,24 and polymer films.8–11 Of these methods, the formation of semiconductor nanosized Particular advantages of this technique include the low reaction temperatures involved, the simplicity of the system, and especi- particles in a polymer medium has received intense interest owing to the ready processibility of the polymer films and ally the ability to obtain polymer nanoparticle composites in which the particle size distribution is narrow without sub- possible future application in device structures.8–11,25–31 The synthesis of nanosized semiconductor particles in a polymer sequent treatment or size fractionation.The average particle J. Mater. Chem., 1997, 7(10), 1969–1975 1969size can be controlled over a wide range by controlling Synthesis of nano-sized CdSe and ZnSe semiconductor particles the reaction parameters, especially temperature.The The preparation procedure for zinc/polymer composites folpolymer/ metal alkyl adducts readily dissociate on heating,43 lows that for cadmium which is now described. A 2pySiPB so they may also be useful for purifying metal alkyls through polymer–dimethyl cadmium solution (3%g cm-3 of polymer) the adduct purification process44 although simpler systems stirred in a flask (250 ml) under nitrogen was gradually exposed based on monomeric Lewis bases are currently preferred to a hydrogen selenide atmosphere until a solid precipitated.because of their lower cost. The transparent yellow polymer adduct solution quickly changed to either a yellow, orange, red, crimson, brown or black coloured suspension of cadmium selenide in polymer Experimental (yellow for zinc selenide), the colour depending on the reaction temperature or choice of solvent.As soon as the precipitate Experiments were carried out under dry oxygen-free argon was observed, the H2Se gas flow was interrupted and the purified by passing through a series of columns consisting of suspension was allowed to settle and then separated by fil- Cr2+ on silica and dry molecular sieveGreaseless joints and tration isolating the coloured polymer composite. taps were employed and manipulations were carried out using standard Schlenk-line and catheter tubing techniques. All the Results solvents were carefully dried by distillation from sodium diphenylketyl. 2-Methylpyridine was purchased from Aldrich Reactions in toluene of polypyridine bound Me2M with H2Se and was distilled prior to use.Butyllithium (1.6 M in hexane) produce composites consisting of nanoparticles embedded and dimethylzinc (2.0 mol dm-3 in toluene) were purchased within the polymer matrix (Fig. 1). A series of diVerent from Aldrich and used as received. Polybutadiene (83% pencoloured CdSe/polymer samples prepared by the method above dant, 17% trans-1,4, Mn=3000) was a commercial product under various synthesis conditions are illustrated in Fig. 2. The (Nippon Soda Company) and was used after pumping for 2 h. size quantization eVect can be seen most dramatically in Me2Cd was prepared by the standard literature method.45 samples Cd1 [-78 °C, Fig. 2(a)], Cd2 [r.t., Fig. 2(b)] and Cd3 Powder X-ray diVraction (PXRD) patterns were recorded [60 °C, Fig. 2(c)] where a darkening in colour indicates an on a Sto� e STADI/P diVractometer using Cu-Ka radiation.increase in particle size and a decrease in band gap energy. Data were collected in transmission mode with a sample The PXRD pattern of Cd3, Fig. 3(c), shows the sharpest mounted in vaseline on a rotating disc and compared with the and most intense peak in all CdSe patterns at the 2h value of standard pattern obtained from the JCPDS database or the 25.5° (002), broader hkl reflections at 42.0° (110) and 50.0° PXRD pattern of a wurtzite sample.(112) as well as very broad (103) reflection at 46.0°. These Transmission electron micrographs (TEMs) were obtained reflections are indicative of the cadmoselite (hexagonal) phase, using a Phillips EM 301 microscope at 80 keV. All samples Fig. 3(e), and not the sphalerite (cubic) phase of CdSe [see the were embedded in an epoxy resin and the sections were then JCPDS pattern of cubic CdSe, Fig. 3(f )]. The hexagonal phase cut on a microtome with a diamond knife. The dried specimen is also confirmed by the lattice spacings in the HRTEM images.sections were put onto a copper grid which had a carbon The broad peak of the polymer is observed at 2h=17°. support film present. Another layer of carbon was then evaporated onto the sample in order to prevent specimen charging. High resolution TEMs (HRTEMs) were obtained by using an ABT 002B microscope at 200 keV. The samples were either prepared as previously described or suspended in acetone and then ultrasonically dispersed and lifted oV onto a graphite grid.Both absorbance and photoacoustic spectroscopy (PAS) were used to measure the band edge of the material. Photoacoustic spectra were obtained using an OAS 400 photoacoustic spectrophotometer as described previously.42 Absorbance spectra were obtained using a Perkin Elmer Lambda 14P UV–VIS spectrophotometer. All samples were scanned from the UV region to the near IR (300–800 nm). The knees of the band edge for both photoacoustic and absorption spectra were recorded and taken as the closest estimation of the band edge value.Of course, the particle size distribution of the material will directly aVect the slope of the band edge and therefore calculated band gap values are indicative of the smallest sized particles only.For commercial CdSe (cadmoselite), the value of the band edge measured in this way is 1.74 eV ( lit. value, 1.74 eV46). Calculations of the band gap when the band edge value is taken from the tail of the spectrum are more inaccurate as the band edge is smeared out by lattice vibrations and falls oV exponentially in accordance with Urbach’s rule.47,48 Synthesis of a polymer adduct with dimethylcadmium or dimethylzinc The polymeric polybutadiene Lewis base containing 2-methylpyridyl (2pySiPB) groups and the subsequent dimethylcadmium or dimethylzinc polymer adduct were synthesized as previously reported.42 In all cases the M/N (M=Zn, Cd) mole Fig. 1 Formation of the 2pySiPB adduct and the reaction of the polymer adduct with hydrogen selenide ratio of metal alkyl to coordinating nitrogen was 0.5. 1970 J. Mater. Chem., 1997, 7(10), 1969–1975Fig. 3 PXRD patterns of CdSe prepared in toluene in the presence of 2pySiPB at (a) -78 °C, run Cd1; (b) 25 °C, run Cd2; (c) 60 °C, run Cd3. Pattern (d) corresponds to a CdSe sample prepared in petrol in the presence of 2pySiPB, run Cd4. Patterns (e) and (f ) correspond to commercial hexagonal (cadmoselite) CdSe and JCPDS pattern for cubic CdSe (sphalerite) respectively.Impurities present within vaseline are denoted as 1. diameter, measurements are therefore calculated from TEMs. The type of solvent chosen for the reaction also has a profound eVect on the crystallite growth. When light petroleum (bp 40–60 °C) is used (see Table 1, Cd4), the PXRD pattern, Fig. 2 CdSe samples showing quantum confinement. Samples (a) to Fig. 3(d), is sharper corresponding to an increase in particle (c) were prepared in toluene in the presence of 2pySiPB at (a) -78 °C, run Cd1; (b) 25 °C, run Cd2; and (c) 60 °C, run Cd3. Sample (d) r.t., diameter. run Cd4, is prepared in light petroleum in the presence of 2pySiPB. The PXRD patterns of ZnSe/polymer composites and the Sample (e) is commercial (bulk) cadmium selenide.JCPDS pattern of hexagonal ZnSe are shown in Fig. 4(a)–(c) where Zn1 [Fig. 4(a)] and Zn2 [Fig. 4(b)] are runs prepared in toluene at -78 °C and r.t. respectively. Again, an increase Estimates of the particle size based on width measurements of the single reflection observed at 42.0° can be made but are not in reaction temperature corresponds with a sharpening of the hkl lines indicating an increase in particle size. The h0l reflection included in this paper because they do not produce accurate values for particles in which distortions and defects occur.(103) at 49.5° is clearly observed for Zn1 and is more intense than other high-angle hkl. At higher temperature, sample Zn2, The reaction temperatures and solvents were varied for a series of experiments in order to correlate the reaction con- all h0l intensities are suppressed.The TEM of run Cd2, Fig. 5, shows an even distribution of ditions with the sizes of the CdSe particles formed and the results are summarized in Table 1. particles within the polymer matrix. A size distribution graph of a TEM obtained by measuring individual particles on a A change in reaction temperature (see Table 1, Cd1–Cd3) has a large eVect on the average particle size as found for TEM image of sample Cd2 (run at r.t.) is illustrated in Fig. 6. TEM studies of samples prepared from runs in toluene and analogous runs for CdS and ZnS.42 PXRD patterns obtained from these samples where the temperature has been varied electron diVraction of all CdSe samples prepared for TEM confirms the crystallite to have lattice spacings indicative of from-78 to 60 °C, Fig. 3(a)–(c), show significant hkl dependent sharpening of lines at higher temperature, i.e. an increase in the Cadmoselite (hexagonal) phase. After two days in an atmosphere of air and after ca. 1 week under nitrogen, the particle size.PXRD studies indicate that the particles exhibit lattice and turbostratic distortions49,50 (see later) and therefore colour of the material obtained from sample Cd1 slowly darkened from a yellow to a light brown colour. However, the calculations of the coherence length of the crystallite using a modification of Scherrer’s formula18 give very poor estimations PXRD patterns taken of the same sample immediately after preparation and then a week later were identical and gave no of crystallite size.For a more reliable indication of crystallite Table 1 Dependence of CdSe particle size on various synthesis conditions av. TEM polymer band particle size run no. Cd/Py ratio solvent T /°C conc. (%) gap/eV (range)/nm bulk CdSe 1.74a Cd1. 0.5 toluene -78 3.0 2.72a 2.3 (1–4.5) Cd2. 0.5 toluene R.T. 3.0 2.17a 2.9 (1.0&nda;5) Cd3. 0.5 toluene 60 3.0 2.15a 3.6 (1.5–6) Cd4. 0.5 light petroleum 25 3.0 2.06b 3.1 (1.5–5.5) Band gap calculated from band edge observed by photoacoustica or absorbanceb spectroscopy. J. Mater. Chem., 1997, 7(10), 1969–1975 1971Fig. 4 PXRD patterns of ZnSe prepared in toluene in the presence of 2pySiPB at (a) -78 °C, run Zn1; (b) 25 °C, run Zn2.Pattern (c) corresponds to the JCPDS pattern for hexagonal ZnSe. Impurities present within vaseline are denoted as 1. Fig. 6 Size distribution graph of CdSe particles, Cd2, prepared in the presence of 2pySiPB in toluene (2.63 eV). However, all the band edges observed for samples of this material are very shallow owing to very wide size distributions of particles and lattice vibrations.This observation suggests that the polymer has little control over ZnSe particle growth and termination, reasons for which are discussed shortly. Many of the composites were investigated for their photoluminescence behaviour. Although blue luminescence was observed for the polymer itself, there was no observable Fig. 5 Low resolution TEM of CdSe particles prepared in run Cd2 luminescence from the nanoparticles.This is probably because low-lying acceptor orbitals on the pyridine quench the emission. sharpening of hkl intensities which would be expected if the particle size increased forcing a change in colour (absorbance). This phenomenon is also observed for ZnSe samples and Discussion possible explanations will be discussed later.The band gap of commercial cadmoselite CdSe, as measured All the PXRD patterns for CdSe samples are consistent with their being of the hexagonal phase. The PXRD pattern of the from photoacoustic spectroscopy is 1.74 eV ( lit. value, 1.70 eV46) and diVers markedly from those of the CdSe/ sample run at high temperature [run Cd3, Fig. 3(c)] suggests that there are turbostratic distortions49,50 present within the polymer composites.Photoacoustic spectra [Fig. 7(a)–(c)] obtained from samples Cd1 (-78 °C), Cd2 (r.t.) and Cd3 crystallites. A perfect crystal has sharp hkl reflections whose structure factors describe the symmetry and contents of the (60 °C) respectively, show a gradual red shift in the band edge indicating an increase in particle size and a decrease in band unit cell.If the layer planes are rotated relative to each other the 00l reflections remain unchanged; the hk0 intensities will gap from 2.72 to 2.15 eV. The UV–VIS absorbance spectrum of the sample prepared in light petroleum [Cd4, Fig. 7(d)] change and the peaks will become broadened because the distortions aVect the unit cell eVectively reducing the crystal gives a calculated band gap value of 2.06 eV and clearly shows a band edge slope indicative of a large size distribution of size.Most aVected will be the hkl reflections, similarly reduced in intensity and broadened. These strong characteristics of particles and exciton relaxation caused by lattice vibrations. In fact, the slopes of the band edges obtained from all samples rotational distortion can be seen in all observed PXRD patterns as the (002) plane is always intense; in most cases the h0l are shallower than e.g.cadmium sulfide samples42 and could be interpreted as a collective result of (i) lattice vibrations planes, (101) at 27°, (102) at 35°, (103) at 46°, and the hkl plane, (112) at 49.5°, are severely broadened. Distortions (phonons); (ii) the faster rate at which cadmium selenide particles are formed compared to the sulfides, thus giving a involving random shifts of lattice planes along the direction of the specified hkl plane of the crystallite give changes in the larger and wider particle size distribution due to slower termination of CdSe particle growth (see later); (iii ) absorbance peak profile exhibiting asymmetry through the peak centre.This eVect is observed in the PXRD pattern of run Cd3, transitions to lower energy bands or trapped ‘surface’ states between the valence and conduction bands which are the result Fig. 3(c), where the (110) peak at 42° has a slightly diVerent rate of intensity decay on the high angle side (r.h.s.) of the peak. of disrupted crystallite/polymer Cd–N interactions present at the grain boundary formed after particle growth; or (iv) traps In preparations where toluene is the reaction solvent there appear to be fewer stacking faults within these crystallites as between the valence and conduction band which pertain to point defects within the bulk of the nanocrystallite. compared to previous CdS or ZnS42 and CdSe samples18 and this is confirmed by the presence of the (103) reflection in Cd3 The band gap for commercial ZnSe as calculated from the photoacoustic measurement is 2.58 eV ( lit. value, 2.58 eV46).[Fig. 3(c)], however, in all cases the (102) reflection is severely broadened and unobserved. While the strong (002) reflection The photoacoustic measurement of the band edge position for the polymer/composite sample Zn2, Fig. 7(e) is 470 nm, could indicate near perfect registry in this particular lattice 1972 J. Mater. Chem., 1997, 7(10), 1969–1975Fig. 8 Propagation and termination of ME (M=Cd, Zn; E=S, Se) particle growth Fig. 4(b)] suggests that there are significant defects within the crystallites, e.g. stacking faults.18 The (002) plane at 27.5° is again the most intense peak in both Zn1 and Zn2 samples.Fig. 7 Photoacoustic spectra of CdSe samples prepared in toluene in the presence of 2pySiPB at (a) -78 °C (run Cd1), (b) r.t. (run Cd2) The ZnSe sample obtained at low temperature [Zn1, Fig. 4(a)] and (c) 60 °C (run Cd3); absorbance spectrum of a CdSe sample shows very broad peaks corresponding to the hexagonal phase prepared in petrol in 2pySiPB at (d) r.t. (run Cd4); photoacoustic of ZnSe at 29.0° (101) and 49.5° (103).The collection of hkl spectrum of a ZnSe sample prepared in toluene in 2pySiPB at (e) r.t. intensities around the (002) region observed in Zn2 is narrower (run Zn1) than expected which could suggest that the number of stacking faults within the crystallite is comparable with the number of planes in perfect hexagonal registry, hence a substantial per- plane of the crystallite, another explanation could be a non-uniformity in the crystallite dimensions suggesting pre- centage of the crystallite co-exists in the form of the cubic phase of ZnSe along with the hexagonal phase.This form of ferred orientation along the direction of this plane. Preparations in light petroleum gave a dark brown colouration polytypism ensures that the suspected (002) reflection therefore more closely resembles the cubic (111) plane of ZnSe and of the reaction mixture indicating the formation of large sized particles.As similarly found in crystallites formed in analogous reinforces the explanation of why the (103) intensity is severely suppressed. reactions for CdS runs, the h0l reflections are noticeably sharper and the (002) reflection broader than found in other The region of quantum confinement for CdSe is very narrow and a slight increase or decrease in particle size would clearly patterns.TEM studies confirm that the particles are agglomerated into clusters with little apparent polymer content but are lead to a change in colour of the material. However, from our sample observations it has been found that although the colour nevertheless comparable in their size distribution (1.5–4.5 nm) to particles observed in sample Cd3.The cadmium selenide/ of the particles (both CdSe and ZnSe) gradually change over a long period of time, the hkl dependent line broadening of polymer adduct is insoluble in non-polar light petroleum and thus cannot disperse the growing nanoparticulates and hence the PXRD patterns remains unchanged indicating no change in particle dimensions. Therefore, we can consider two possibil- allows them to cluster.The suppression of the (103) plane in the PXRD pattern of ities for these observations and firstly, we postulate that on exposure of these samples to moisture a gradual deposition of the ZnSe sample obtained at room temperature [Zn2, J.Mater. Chem., 1997, 7(10), 1969–1975 1973red selenium metal on the surface of the particle occurs as Allen and Dr Douglas F. Foster for useful discussions; and the EPSRC for funding (S.H.). oxygen is a harder base than selenium and O2 thus displaces it forming a colourless, high band gap metal oxide (ZnO, CdO). The second possibility could be the presence of residual hydrogen selenide gas enveloped in the polymer which, on air References exposure, decomposes to elemental selenium in a similar 1 M.G. Bawendi, M. L. Steigerwald and L. E. Brus, Annu. Rev. Phys. manner to selenium alkyls. Chem., 1990, 41, 477. On the basis of the average particle size and the size 2 A. Henglein, Chem. Rev., 1989, 89, 1861. dispersion under a given set of conditions, the eVectiveness 3 Y.Wang and N. Herron, J. Phys. Chem., 1991, 95, 525. 4 H. Weller, Adv.Mater., 1993, 5, 88. of the polymer pyridyl group for controlling the growth 5 H. Weller, Angew. Chem., Int. Ed. Engl., 1993, 32, 41. of nanoparticles appears to decrease in the order 6 J. H. Fendler and F. C. Meldrum, Adv. Mater., 1995, 7, 607. ZnS>CdS>CdSe>ZnSe.This is the same order as we have 7 A. R. Kortan, R. Hull, R. L. Opila, M. G. Bawendi, M. L. observed in the production of nanoparticles from gas phase Stiegerwald, P. J. Carroll and L. E. Brus, J. Am. Chem. Soc., 1990, reactions of Me2M and H2E in the presence of pyridine,51,52 112, 1327. although in those cases, elemental analysis of the particles 8 B. O. Dabbousi, M.G. Bawendi, O. Onitsuka and M. F. Rubner, Appl. Phys. L ett., 1995, 66, 1316. obtained gives further information since, for ZnS and CdS 9 V. L. Colvin, M. C. Schlamp and A. P. Alivisatos, Nature (L ondon), there is approximately one mole of pyridine per surface atom, 1994, 370, 354. whilst for CdSe much less pyridine is present and for ZnSe 10 Y. Yang, J. Huang, S. Liu and J. Shen, J.Mater.Chem., 1997, 7, 131. there is hardly any pyridine present in the deposits. 11 Ying Wang and N. Herron, J. L umin., 1996, 70, 48. We have rationalised these observations in terms of a particle 12 G. Hodes, I. D. J. Howell and L. M. Peter, J. Electrochem. Soc., growth model in which surface metal atoms are bound to one 1992, 139, 3136. 13 K. Kalyanasundaram, in Energy Resources by Photochemistry and methyl group.Since the particle growth is carried out under Catalysis, ed. M. Gra�tzel, Academic Press, London, 1983, p. 217. H2E-rich conditions, these surface metal atoms bind H2E. It 14 E. Wolf, Progress in Optics XXIX, Elsevier Science Publishers, is the relative rates of elimination of methane (to give surface North Holland, 1991, p. 321.HE atoms and propagate growth), and displacement of H2E 15 S. H. Risbud, T he Encyclopedia of Advanced Materials, Cambridge by pyridine (to terminate growth), that determine the eYciency University Press, Cambridge, 1994, p. 2115–2121. with which particle growth is controlled by pyridine. For H2Se, 16 E. Corcoran, Sci. Am., 1990, 263, 74. 17 N. Herron, Y. Wang and H. Eckert, J.Am. Chem. Soc., 1990, which is much more acidic than H2S, loss of methane dominates 112, 1322. and this is more pronounced for ZnSe because the surface 18 C. B. Murray, D. J. Norris and M. G. Bawendi, J. Am. Chem. Soc., ZnMC bonds are more polar (d- on C) than those for Cd. 1993, 115, 8706. Hence control of growth is poor for the selenides, especially 19 J. G. Brennan, T. Siegrist, P.J. Carroll, M. Stuczynski, L. E. Brus ZnSe. For the less acidic H2S, the rate of elimination of and M. L. Steigerwald, J. Am. Chem. Soc., 1989, 111, 4141. methane is low, so that the surface bound H2S adduct has 20 T. Trindade and P. O’Brien, Chem. Mater., 1997, 9, 523. 21 B. Breitscheidel, J. Zieder and U. Schubert, Chem. Mater., 1991, suYcient lifetime for displacement by pyridine, thus terminat- 3, 559.ing growth. The harder nature of Zn than of Cd and of 22 Y. Wang and N. Herron, J. Phys. Chem., 1987, 91, 257. pyridine than of H2S will ensure that the displacement of H2S 23 X. K. Zhao, L. McCormick and J. H. Fendler, Chem. Mater., 1991, by pyridine will be more favored for ZnS than for CdS. We 3, 922. believe that similar arguments qualitatively explain the growth 24 J.N. Robinson and D. J. Cole-Hamilton, Chem. Soc. Rev., 1991, process and its control for the polypyridine–nanoparticle com- 20, 49. 25 Y. Wang, A. Suna, M. Mahler and R. Kasowski, J. Phys. Chem., posites described here. The relevant reaction schemes are 1987, 87, 7315. shown in Fig. 8. 26 M. E. Wozniak, A. Sen and A. L. Rheingold, Chem. Mater., 1992, 4, 753. 27 J. P. Kuczynski, B. H. Milosavljevic and J. K. Thomas, J. Am. Conclusions Chem. Soc., 1986, 108, 2513. 28 Y. Wang and W. Mahler, Opt. Commun., 1987, 61, 233. CdSe semiconductor nanoparticulates in the size range 1–5 nm 29 E. Hilinski, P. Lucas and Y. Wang, J. Chem. Phys., 1988, 89, 3435. with a relatively narrow size distribution can be prepared by 30 Y. Wang, A. Suna, J. McHugh, E.Hilinski, P. Lucas and R. D. reacting a polymer adduct solution with H2Se. When the Johnson, J. Chem. Phys., 1990, 92, 6927. reaction temperature is controlled and toluene is the reaction 31 S. Yanagida, T. Enokida, A. Shihdo, T. Shiragami, T. Ogata, T. Fukumi, T. Sakaguchi, M. Mori and T. Sakata, Chem. L ett., solvent, the size of the growing crystallite and hence band edge 1990, 1773.position can be controlled relatively easily. However, when the 32 V. Sankaran, J. Yue, R. E. Cohen, R. R. Schrock and R. J. Silbey, solvent is light petroleum, generally wide size distributions of Chem. Mater., 1995, 7, 1185. clustered CdSe nanoparticulates are observed, because of the 33 (a) M. MoYtt and A. Eisenberg, Chem. Mater., 1995, 7, 1178; (b) insolubility of the polymer.CdSe nanoparticulates appear to M. MoYtt, L. McMahon, V. Pessel and A. Eisenberg, Chem. be cadmoselite (hexagonal) in phase and in some cases appear Mater., 1995, 7, 1185. 34 Y. Yuan, J. H. Fendler and I. Cabasso, Chem. Mater., 1992, 4, 312. to exhibit partial registry and distortions of the crystallite 35 A. Chevreau, B. Phillips, B. G. Higgins and S. Risbud, J. Mater. lattice.Although the quantum confinement regime observed Chem., 1996, 6, 1643. in CdSe is narrow the diVerent colours exhibited by the 36 L. Spanhel, M. Haase, H. Weller and A. Henglein, J. Am. Chem. material in this region as the particles decrease in size are Soc., 1987, 109, 5649. clearly observed indicating that the particle size distribution is 37 A. Iraqi and D. J. Cole-Hamilton, J.Mater. Chem., 1992, 2, 183 relatively small. The particle size of ZnSe semiconductor and references therein. 38 P. Narayanan, B. Kaye and D. J. Cole-Hamilton, J. Mater. Chem., nanoparticulates is diYcult to control mainly because of the 1993, 3, 19. strong Brønsted acidity of the hydrogen selenide and its rapid 39 A. Iraqi, S. Seth, C. A. Vincent, D. J. Cole-Hamilton, M. D. reaction with surface methyl (ZnMMe) groups which carry Watkinson, I. M. Graham and D. JeVrey, J. Mater. Chem., 1992, more negative charge than cadmium methyl groups. 2, 1057. 40 X. Li, C. M. Lindall, D. F. Foster and D. J. Cole-Hamilton, J.Mater. Chem., 1994, 4, 657. We thank John Mackie at the Bute, St. Andrews, for sample 41 X. Li, J. R. Fryer and D. J. Cole-Hamilton, J. Chem. Soc., Chem. preparation for low resolution TEM. We are also grateful to Commun., 1994, 1715. the Nippon Soda Company for generous gifts of polybuta- 42 S. W. Haggata, X. Li, J. R. Fryer and D. J. Cole-Hamilton, J.Mater. Chem., 1996, 6, 1771. dienes; to Dr Barry Kaye for photography; Professor J. W. 1974 J. Mater. Chem., 1997, 7(10), 1969–197543 X. Li, D. F. Foster and D. J. Cole-Hamilton, Polym. Adv. T echnol., 49 Y. G. Andreev and T. Lundstro� m, J. Appl. Crystallogr., 1995, 28, 534. 50 Y. G. Andreev and T. Lundstro�m, J. Appl. Crystallogr., 1994, 27, 1994, 5, 541. 44 D. J. Cole-Hamilton, Chem. Br., 1990, 26, 852. 767. 51 N. L. Pickett, D. F. Foster, J. R. Fryer and D. J. Cole-Hamilton, 45 D. F. Foster and D. J. Cole-Hamilton, Inorg. Synth., 1997, 31, 29. 46 CRC Handbook of Chemistry and Physics, ed. D. R. Lide, CRC J.Mater. Chem., 1996, 6, 507. 52 N. L. Pickett, D. F. Foster, F. Riddell, J. R. Fryer and D. J. Cole- Press Inc., Boston, MA, 1992, vol. 12, p. 75. I. Pankove, Optical Processes in Semiconductors, Dover Hamilton, J.Mater. Chem., 1997, 7, 1855. Publications, New York, 1975, p.43. 48 F. Urbach, Phys. Rev., 1953, 92, 1324. Paper 7/01943B; Received 19th March, 1997 J. Mater. Chem., 1997, 7(10), 1969–1975 1975