J. Phys. Chem. C 2010, 114, 17311–17317
Highly Efficient Precursors for Direct Synthesis of Tailored CdS Nanocrystals in Organic Polymers Vincenzo Resta,* Anna M. Laera, Emanuela Piscopiello, Monica Schioppa, and Leander Tapfer ENEA, Unita` Tecnica Tecnologie dei Materiali Brindisi (UTTMATB), Strada Statale 7 “Appia” - Km 706, I-72100 Brindisi, Italy ReceiVed: May 5, 2010; ReVised Manuscript ReceiVed: July 19, 2010
An innovative unimolecular precursor structure based on cadmium-bis(benzylthiolates), Cd(SBz)2, has been devised for the preparation of polymer/CdS nanocomposite materials through a simple and inexpensive synthetic route. Cd(SBz)2 and [Cd(SBz)2]2 · MI, MI being methyl imidazole, dispersed in a polymer matrix have been later thermolyzed in very mild conditions in the range of temperatures, TA, 175 °C e TA e 240 °C. Optical absorbance, photoluminescence spectroscopy, and X-ray diffraction show that, for both precursors, CdS nanocrystal (NC) nucleation and growth start well below the literature threshold values and that in the analyzed temperature range the NC’s size is finely tuned in quantum confinement regime conditions. Methyl imidazole added precursor shows highly efficient reactivity at lower temperatures, while the differences become smaller for higher values. Transmission electron microscopy ensures that in the entire range of temperatures, the synthesis of highly dense and homogeneous distributed CdS NCs over the whole extent of the matrix has been achieved with [Cd(SBz)2]2 · MI precursor. The results show that benzylthiolates are really promising molecules as precursors for optoelectronic and/or photovoltaic nanocomposite based devices. Introduction Semiconductor nanocrystals (NCs) are of fundamental interest in material science due to the unique relation between their size and the corresponding physical, chemical, and optical properties. By selection of suitable materials and tailoring of the growth/ postgrowth processing parameters, NCs response can be tuned through the quantum confinement effect when size becomes smaller than the corresponding exciton Bohr radius.1-3 The embedding of semiconductor NCs into organic polymers allows one to obtain organic/inorganic hybrid nanocomposites with extraordinary performances due to the combination of both the NCs and the host properties.4,5 On one hand, organic materials available as matrices are polymers of easy and cheap synthesis techniques for bulklike as well as thin film formation, with high solubility in most of the common solvents and high thermal and mechanical stability.6 On the other hand, they protect inorganic NCs from aggregation, increase their solubility and miscibility, and preserve their chemical and physical properties as electrical and photoconductivity, nonlinear optical response, photoluminescence (PL), and electroluminescence.7 Finally, the combination of semiconductor NCs and inorganic semiconductor matrices,8,9 or conjugated polymers,10,11 with improved power conversion efficiency represents the newest and more interesting solution for high stability, low cost, and easy processing nanocomposite for radiation detection12,13 and light emission devices7,14 and for photovoltaic applications.11,15,16 Systems based on hybrid nanocomposite materials require fine control of the NC’s size inside the polymer matrix in order to achieve a suitable selection of the optical response and, especially for photovoltaic applications, a homogeneous and dense dispersion of the NCs, promoting an efficient polaron pair transfer between polymer and NCs and minimizing charge * To whom correspondence should be addressed. E-mail: vincenzo. [email protected]
recombination.17,18 Hybrid nanocomposites can be obtained, through ex situ approach, by synthesizing separately the organic and inorganic components19,20 and then dispersing the semiconductor NCs in the polymer. But, during the mixing procedure step, phase segregation of inorganic and organic phases can mainly occur because the NCs tend to agglomerate and minimize their surface free energy. On the contrary, in situ synthetic strategies provide better control of stoichiometry and prevent segregation phenomena as the NC’s growth is directly promoted in the polymer matrix.21,22 Such an approach has been devised as a two-step process based on the introduction of metal ions into polymer media and on the subsequent reaction of the samples with suitable solutions or gases.10 However, this technique is not suitable for obtaining nanocomposites with flexible and large-range tunable properties because it is not possible to control the stoichiometry of the semiconductor compound NCs and it is not easy to finely tailor their size and size dispersion. These drawbacks are completely overcome if a unimolecular precursor, containing both the metal and the nonmetal part, is directly introduced into the matrix, and the samples are subsequently processed through annealing and/or laser irradiation in order to induce the precursor decomposition and the NC nucleation and growth.8,11,21,22 CdS is one of the most promising photosensitive materials owing to its large band gap (2,58 eV) and its high electron mobility.23 Indeed, CdS NCs are highly interesting for applications in semiconductor lasers,24 ultrafast all-optical switching,25 and optoelectronic26,27 and biological labeling technologies.28 Cadmium thiolates for instance are promising compounds to be used as unimolecular precursors for CdS NCs,29 but unfortunately they have polymeric structure leading to low volatility and insolubility in most of the common organic solvents and preventing their homogeneous dispersion in polymeric matrices.30
10.1021/jp104097w 2010 American Chemical Society Published on Web 09/23/2010
J. Phys. Chem. C, Vol. 114, No. 41, 2010
Resta et al.
Figure 1. Weight loss (open circles, left-hand side axis) and corresponding DTA (continuous line, right-hand side axis) curves associated with Cd(SBz)2 (a) and [Cd(SBz)2]2 · MI (b). The starting and finishing temperatures for each process recognized in TGA graphs are indicated by increasing cardinals with s and f subscripts, respectively.
In the present work we propose innovative cadmium thiolate precursors suitably modified to obtain soluble cadmium complexes that can be dissolved in solution with polymers.31 Easy deposition techniques, such as spin coating or casting in vacuum, are then available to produce homogeneous and large area films as required for devices fabrication on a large scale. The results obtained with poly(methyl methacrylate) (PMMA) as matrix for CdS NCs are presented as a further improvement with respect to previous works based on polystyrene32 and TOPAS,21 and as a preliminary step before involving a conjugated polymer. PMMA is easy to handle, cheap, and optically transparent in the visible region, also allowing a complete characterization of the NCs fillers. CdS NCs from cadmium thiolate precursors can be synthesized more efficiently with respect to better known and previously used precursors in terms of threshold temperature (below 200 °C) as well as of annealing time (∼30 min), the devised synthetic strategy being particularly suitable for the fabrication of organic/inorganic nanocomposites with low thermal capacity hosts. Experimental Section Precursor Synthesis and Preparation of CdS NCs. All the reagents involved in the synthesis of the precursors and the PMMA matrix were purchased from Aldrich and were used without further purification. The precursor cadmium-bis(benzylthiolate), Cd(SBz)2, was synthesized adapting the protocol described in ref 21. A commercial salt, cadmium nitrate hexahydrate (9 mmol), was dissolved in ethanol (90 mL) before the addition of an aqueous solution of ammonium hydroxide (25%). When benzyl mercaptan (18 mmol) was added, the desired product precipitated as a white powder in reaction medium in quantitative yield. The white precipitate was separated from solution by filtration, washed with absolute ethanol, and finally dried under vacuum. Once Cd(SBz)2 was obtained, a suspension in chloroform with the addition of CH3 imidazole (methyl imidazole, MI) was produced until a completely clear solution was obtained, as the metal thiolates form soluble adducts by incorporation of Lewis bases.33 Upon removal of solvent, the product [Cd(SBz)2]2 · MI was crystallized from toluene cooling the solution to -18 °C. Solutions of [Cd(SBz)2]2 · MI and suspensions of Cd(SBz)2, both with PMMA in chloroform, were prepared maintaining a 20/80 weight ratio: a 0.5 µL solution/suspension containing 20 µg of cadmium thiolate derivative and 80 µg of PMMA was dissolved in 4 µL of chloroform. The precursor-polymer samples were finally produced by casting the solutions and the suspensions on glass plates and on carbon-coated copper grids. A thin film was formed onto the copper grids, while bulklike samples with rough
adhesion to the substrate were obtained when glass plates were used. The thermolysis process was performed on both bulklike and thin film copper grids samples, in nitrogen atmosphere, for annealing temperatures, TA, in the range 175 °C e TA e 240 °C. The processes were carried out with a heating rate of 15 °C min-1, an annealing time of 30 min for the selected temperature, and a cooling rate of 5 °C min-1. The chemical structure of the precursor molecules Cd(SBz)2 and [Cd(SBz)2]2 · MI are shown in the Supporting Information (Figure S1). Characterization Techniques. Thermophysical properties and decomposition of the precursor, i.e., the removal of the organic part and the formation of CdS, were investigated by simultaneous thermogravimetric analysis (TGA) and differential thermal analysis (DTA). Powder compounds of Cd(SBz)2 and [Cd(SBz)2]2 · MI (the precursor only without PMMA) were studied with a Netzsch-Gera¨tebau GmbH STA429 Simultaneous Thermal Analyzer. Samples of about 10 mg were put into an open platinum crucible, and an argon flow of 80 cm3 min-1 was always used to remove gaseous decomposition products. Dynamical TGA measurements were carried out by setting the maximum heating temperature at 350 °C with a scan rate of 10 °C min-1. Static TGA was carried out over 70 min at the fixed temperature of TA ) 185 °C, this value being approached with a scan rate of 15 °C min-1, the same parameter value as for the thermolytic process. The formation and the crystallographic structure of the CdS NCs were analyzed by means of a Philips PW1880 3 kW Bragg-Brentano X-ray diffractometer equipped with a copper target as X-ray source (Cu-KR radiation, λ ) 1.54 Å). For the structural characterization, bulklike samples before and after the thermal annealing were employed. The morphology of CdS NCs was studied by means of a TECNAI F30 transmission electron microscope (TEM) operating at 300 kV with a point-to-point resolution of 0.205 nm. Thin film samples deposited on carbon-coated copper grids were used for TEM observations. UV-visible absorbance spectra of the precursor-doped polymers before and after the thermolytic process were obtained from the transmission (T) measurement, as ln(1/T). The spectra were recorded for chloroform solutions of bulklike samples by using a Xe lamp (LC8 Hamamatsu) and a HR460 monochromator (Jobin Yvon). PL measurements on the same chloroform solutions were performed with a Varian Cary Eclipse Fluorometer by selecting an excitation wavelength λexc ) 330 nm. Results and Discussion Weight loss measured through dynamical TGA analysis and the corresponding DTA signal obtained for Cd(SBz)2 are shown
Synthesis of Tailored CdS Nanocrystals
J. Phys. Chem. C, Vol. 114, No. 41, 2010 17313
in Figure 1a, the signals referring to the left-side and the rightside axis, respectively (as indicated by the dashed arrows). Thermal decomposition of Cd(SBz)2 starts at about 170 °C and finishes at 300 °C (starting and finishing points are labeled as 1s and 1f, respectively, in Figure 1a with a total mass loss of 57.6%. Such a reduction of weight is in good agreement with the theoretical value of 59.7% corresponding to the percent weight of dibenzyl sulfide, as depicted in the Scheme 1 reaction:31 SCHEME 1: CdS Formation from Cd(SBz)2 Thermal Decomposition
The DTA curve demonstrates that thermal decomposition takes place in a multistep process with a main endothermic reaction at 250 °C leading to the total decomposition of Cd(SBz)2. During all the experiments, in the range between 170 and 250 °C all the samples change color from white to bright yellow, consistent with the band gap values associated with CdS NCs.23 Figure 1b shows TGA and DTA curves obtained for the [Cd(SBz)2]2 · MI compound, referred to the left-hand and righthand side axis, respectively (as indicated by the dashed arrows). A 9.1% weight loss observed between 60 and 160 °C (point 1s and 1f in Figure 1b) can be attributed to the cleavage of the CH3-imidazole ligand, the endothermic process peaked at 100 °C being coherent with methyl imidazole thermal cleavage. The measured mass reduction, according to the predicted value of 10.2% as in ref 31, states that the right relation between Cd(SBz)2 and MI molecules is 2:1, confirming that the formula for the Lewis base derived complex is [Cd(SBz)2]2 · MI. At higher temperatures the remaining Cd(SBz)2 decomposes according to the reaction in Scheme 1 reported above. The thermolytic decomposition of the precursor is associated with a multistep process (starting point, 2s, and finishing point, 2f, in Figure 1b) inducing a mass reduction of 54.3% and characterized by the maximum heat absorption at 244 °C. The final product CdS exhibits a bright yellow color as for Cd(SBz)2 compound. The reaction scheme of the thermal decomposition of [Cd(SBz)2]2 · MI and the CdS NC formation can be summarized as the following: SCHEME 2: CdS Formation Due to the Thermal Decomposition of [Cd(SBz)2]2 · MI through an Intermediate Step Characterized by the Cleavage of Methyl Imidazole
The derivatives of static TGA (DTG) data performed on Cd(SBz)2 and [Cd(SBz)2]2 · MI powders are reported in Figure 2 as a function of time. A strong mass reduction for [Cd(SBz)2]2 · MI precursor occurs during the early stage of the annealing process up to 35 min, and it is definitively stabilized after about 40 min. On the contrary, Cd(SBz)2 precursor shows a slight steady weight loss from the beginning of the process that stabilizes at about 70 min. The weight losses for both precursors become comparable after an annealing time of about 80 min (at TA ) 185 °C), as shown in the inset of Figure 2. Thermal analysis clearly indicates that the decomposition of both thiolates starts below 200 °C annealing temperature, the
Figure 2. DTG curves recorded at TA ) 185 °C for Cd(SBz)2 and [Cd(SBz)2]2 · MI, as labeled close to each graph, the corresponding TGA data being reported in the inset.
Figure 3. XRD patterns recorded on CBz (a) and CBz-MI (b) samples before and after the annealing processes at 175, 185, 220, and 240 °C (as labeled in each graph).
[Cd(SBz)2]2 · MI molecule being more efficient in terms of temperature and time at a fixed temperature. X-ray diffraction (XRD) patterns of Cd(SBz)2 and [Cd(SBz)2]2 · MI bulklike samples measured before and after the annealing processes performed at 175, 185, 220, and 240 °C are shown in Figure 3a and Figure 3b, respectively, as a function of the scattering vector.34 For simplicity sake, the samples of Cd(SBz)2 precursor in PMMA will be denoted throughout the work as CBz, and those based on [Cd(SBz)2]2 · MI in PMMA as CBz-MI. XRD patterns before the heat treatment (lowest curve) exhibit both the PMMA amorphous peak (q ) 11 nm-1) and sharp Bragg-like peaks that are attributed to the precursor molecules. For the CBz samples the precursor-related XRD peaks disappear for TA g 220 °C (Figure 3a), and broader diffraction peaks ascribed to the zincblende structure of CdS, indexed as (111), (200), (220), and (311) Bragg peaks, appear (ICDD no. 80-0019).35 On the contrary, for CBz-MI samples annealed at TA g 175 °C, the representative peaks associated with the [Cd(SBz)2]2 · MI precursor (marked by * in the XRD curve recorded before the thermal treatment and reported in the Supporting Information in Figure S2) disappear. The peak at about q ) 20 nm-1, which increases in intensity and becomes
J. Phys. Chem. C, Vol. 114, No. 41, 2010
Resta et al.
TABLE 1: Measured Values of the Band Gap Energy, the PL Peak Position, and the Stokes Shift for CBz and CBz-MI Samples after Annealinga CBz temperature (°C) band gap energy (eV) PL peak (eV) Stokes shift (meV) NC size from XRD (nm) NC size from absorbance (nm) (ref 23)
175 185 220 240 175 185 220 240 3.61 3.41 2.95 3.02 3.08 3.02 2.87 2.97 3.15 3.11 780 600 -
3.03 3.04 2.99 2.96 230 800 400 300 -
1.93 2.38 4.07 3.72 3.45 3.72 4.53 3.96
a The particle size of CdS NCs calculated from XRD,34 and optical absorption,23 data is reported, too.
narrower as the temperature rises, can be regarded as the convolution of (111) and (200) Bragg peaks associated with zincblende CdS, the theoretical positions being q111 ) 18.7 nm-1 and q200 ) 21.6 nm-1, respectively.35 Similarly, the peaks located at 30.6 and 35.8 nm-1 increase in intensity and decrease in breadth as temperature TA raises. These peaks correspond to the (220) and (311) Bragg peaks of the zincblende phase of CdS. The angular position and relative intensity of all the peaks indicate the formation of CdS NCs of zincblende structure, while no indication of CdS wurtzite phase has been found (ICDD no. 80-006).35 XRD patterns obtained for CBz-MI samples exhibit more pronounced CdS-related Bragg peaks with respect to CBz samples. Indeed, CBz XRD patterns for TA e 185 °C show sharp Bragg-like peaks of nondecomposed precursor organic part, while no significant contribution of CdS can be noticed, indicating that the decomposition of Cd(SBz)2 is not yet completed. The average size of CdS NC was calculated by using Scherrer’s formula,34 referring to the measured full-width at halfmaximum of (220) and (311) peaks and excluding the (111) and (200) peaks being convoluted and partially overlapped. The estimated values, as reported in Table 1, for the whole set of CBz-MI samples increase as TA rises, while for Cd(SBz)2-based samples the size was deduced only for TA g 220 °C due to the contribution of the undecomposed precursor molecules. In addition, the experimental patterns were also simulated by using a scattering model for spherical-shaped NCs.21,36 The details of the simulations for the diffraction curves of the CBz and CBzMI samples at 240 °C are reported in the Supporting Information (Figure S3). All the calculated values of the particle size are notably smaller than the Bohr diameter of CdS bulk exciton,23 suggesting that strong quantum confinement regime stands.1,2 The higher efficiency of methyl imidazole added Cd(SBz)2 precursor was also confirmed by higher size of CdS NCs formed at a fixed temperature with respect to CBz samples. Figure 4 shows bright-field TEM images of CBz (Figure 4a) and CBz-MI samples (Figure 4b), annealed at TA ) 185 °C. Despite the weak image contrast between the dark areas, CdS, and the gray background, PMMA, well-defined CdS NCs are observed in both samples. In the CBz-MI sample, a highly homogeneous spatial distribution of CdS NCs within the polymer matrix is observed. On the contrary, without MI group the thermal treatment leads to agglomeration and clustering phenomena of NCs, and the NC density varies from zone to zone, where some areas of the polymer are lacking NCs (for example, see zone I in Figure 4a in contrast with zone II). Highresolution images for CBz and CBz-MI samplessshown in
Figure 4c, d, respectivelysallow reliable estimation of the NCs average diameter, which was determined by statistically analyzing areas with at least 200 NCs. The average particle size is 3.3 ( 0.7 nm for CBz-MI and 2.8 ( 0.3 nm for CBz, respectively, for TA ) 185 °C. These results confirm that for low annealing temperature, CdS NC size is larger for [Cd(SBz)2]2 · MI with respect to CBz precursor. The insets of Figure 4c, d show single CdS NCs exhibiting well-pronounced (111) lattice fringes (fringe distance 0.34 nm) for CBz sample (Figure 4a) and (200) lattice fringes (fringes distance 0.29 nm) for the CBz-MI sample, both belonging to zincblende structure in agreement with lattice theoretical parameter (aCdS(111) ) 0.33 nm, aCdS(200) ) 0.29 nm in ICDD no. 80-0019).35 The high-resolution images demonstrate that most of the CdS NCs are made of single crystals (monocrystals) and only a very small fraction of the NCs exhibits crystalline defects (line defects, dislocations, grain boundaries, etc.), i.e., polycrystals. The optical absorbance spectra of CBz and CBz-MI samples annealed at TA ) 185 °C are shown in Figure 5a. The optical response is characterized by a high-energy absorption peak associated with the first excitonic transition between the ground state and the single electron-hole pair state (1S3/21se),37 and by a long-wavelength absorption tail between 350 and 450 nm. Direct band gap energy for CdS NCs in CBz and CBz-MI samples was calculated by using a linear fit in the region of the lowest absorption edge to a plot of (Rhν)2, R being the absorption coefficient in cm-1 and hν the energy in eV.2,23 The band gap energy evolution as a function of TA is reported in Figure 5b, while the estimated values of NC size are summarized in Table 1. Samples obtained with [Cd(SBz)2]2 · MI precursor, mainly for TA e 185 °C, have a lower band gap energy than CBz samples. As a whole, the band gap energy decreases as TA increases for TA e 185 °C and it saturates at higher temperatures, remaining notably higher than bulk CdS value (dotted line in the inset of Figure 5b),23 irrespective of the precursor. Comparison with literature absorbance spectra suggests that very small NCs (R < 5 nm) were obtained,38 the effective mass approximation being unsuitable for direct size evaluation in such a dimension range.2 The empirical relationship reported in ref 23 was used as an alternative route to tight-binding or pseudopotential methods necessary to account for higher and more realistic confinement regimes. For CBz-MI samples, an average diameter ranging from 3.45 nm up to 4.53 nm was calculated, while for CBz precursor the size range is between 1.93 and 4.07 nm (Table 1). Good agreement in terms of absolute values obtained from TEM analysis was found for CBzMI samples annealed at TA ) 185 °C, and it was confirmed that the NC’s size in CBz-MI samples is always higher than that one in the corresponding CBz samples, as shown qualitatively by XRD and TEM. A more pronounced saturation-like behavior was found when a threshold value is overcome. Optical measurements further confirm that for both Cd(SBz)2 and [Cd(SBz)2]2 · MI precursors, the nucleation produces CdS NCs in the regime of strong quantum confinement. PL spectra measured for CBz-MI (a) and CBz (b) samples annealed at 185 °C, and for CBz sample at 240 °C (c), are shown in Figure 6, respectively (continuous blue line, online version). As labeled in all graphs, the corresponding absorbance spectrum (continuous black line, online version) for each sample is reported. Spectral deconvolution of PL spectra by employing two Gaussian functions was carried out in order to distinguish between the contributions of higher energy band edge emission (dashed light-green, online version) and those of trap state
Synthesis of Tailored CdS Nanocrystals
J. Phys. Chem. C, Vol. 114, No. 41, 2010 17315
Figure 4. Low-magnification bright-field TEM images of CBz (a) and CBz-MI (b) samples, respectively, both annealed at 185 °C. The CBz image shows an area with two zones of different NC density (low density in zone I and high density in zone II). The micrographs (c) and (d) show the corresponding high-resolution images of densely packed and representative CdS NCs obtained in CBz and CBz-MI, respectively, both annealed at 185 °C. The insets shows a single CdS NC exhibiting the lattice fringes of cubic CdS (the marker corresponding to 3 nm).
Figure 5. (a) Absorbance spectra of CBz (black line, color online) and CBz-MI samples (red line, color online) annealed at 185 °C, as labeled close to each graph. (b) Evolution of band gap energy, as a function of TA for CBz (squares) and CBz-MI samples (circles).
emission (dashed dark-green line, online version).21,22,39 The corresponding theoretical overall emission for CBz-MI (Figure 6a, circles), CBz annealed at 185 °C (Figure 6b, squares), and CBz annealed at 240 °C (Figure 6c, triangles) is also shown. For [Cd(SBz)2]2 · MI precursor PL is dominated by CdS NCs band edge emission (2.99 eV at TA ) 185 °C, Figure 6a) that is notably larger than trap state emission (2.60 eV at TA ) 185 °C, Figure 6a), such a behavior being accomplished in the whole range 175 °C e TA e 240 °C. On the contrary, for CBz samples annealed at TA e 185 °C, the spectra are dominated by the surface defects band generated by anionic vacancies, almost completely screening the band edge emission (Figure 6b). When the annealing of CBz samples is performed at TA g 220 °C, defects band emission shrinks and CdS band edge emission
dominates the PL spectrum (Figure 6c for TA ) 240 °C). The latter is in a spectral position close to that of CBz-MI sample annealed at TA ) 185 °C (reinforcing the idea of a saturationlike behavior). The evolution of measured band edge emission and of the corresponding Stokes shift as a function of TA is shown in Figure 6d and in its inset, respectively, and the values are summarized in Table 1. Irrespective of the precursor, when the temperature increases, a red shift of CdS NCs emission was observed and a saturation-like behavior was reached close to TA g 220 °C, similar to the results observed for the absorption band (Figure 5b). Band edge emission is normally red-shifted in CBz-MI samples with respect to the corresponding CBz ones, of a constant value of ∼100 meV in the annealing temperature range 175 °C e TA e 240 °C. According to XRD and TEM results, also PL states that for low annealing temperatures (TA e 185 °C) bigger CdS NCs are produced when [Cd(SBz)2]2 · MI precursor is used, while at higher temperatures (when a threshold value is overcome) the NC’s size becomes similar for both precursor used. Trap state emission is not affected by the temperature variation as expected and lies around 480 nm irrespective of the precursor used. Finally, the shift between the absorbance band maximum and the emission is found to be bigger for CBz-MI samples than the corresponding CBz precursor samples, and it ranges between 400 and 640 meV for the CBz-MI and between 230 and 790 meV for the CBz, respectively. The evolution of the Stokes shift does not precisely recover the evolution of PL emission,
J. Phys. Chem. C, Vol. 114, No. 41, 2010
Figure 6. PL emission spectra (continuous blue line, color online) of CBz-MI (a) and CBz (b) sample annealed at 185 °C and CBz annealed at 240 °C (c). The corresponding absorbance spectrum (continuous black line, color online) is reported in each graph. Gaussian functions calculated for band edge emission (dashed lightgreen line, color online) and defect band emission (dashed darkgreen line, color online) are reported. The corresponding theoretical overall emission for CBz-MI annealed at 185 °C (circles) and CBz annealed at 185 °C (squares) and 240 °C (triangles) is presented. (d) Summary of PL peak positions as a function of the annealing temperature for samples CBz (squares) and CBz-MI (circles). The Stokes shift estimated for CBz (squares) and CBz-MI (circles) is reported in the inset.
especially for CBz-MI samples, suggesting that it is not only related to the different light-emitting states (the fine structure) of CdS NCs, but it is also affected by exciton-acoustic phonon coupling, and for this reason the Stokes shift does not monotonically decrease as the NC’s size increases.40 The results presented demonstrate that cadmium-bis(benzylthiolate) efficiently works as unimolecular precursor for CdS NCs synthesis. The precursor decomposition starts at low annealing temperature, TA ∼ 175 °C, especially for [Cd(SBz)2]2 · MI, MI easing the formation of the initial CdS molecules and then of the NCs nucleation and growth. The threshold of the thermolytic process is considerably lowered with respect to previous works as can be noticed in refs 21 and 32. When compared to older unimolecular precursors such as Cd(SC12H25)2 or Cd(SC18H37)2, the annealing temperature is reduced about 70 °C to obtain ∼2 nm size NCs with Cd(SBz)2 precursor and to synthesize ∼4 nm size NCs with [Cd(SBz)2]2 · MI; for the latter, the absolute temperature is reduced more than 100 °C for smaller sizes.21 The time of annealing process is also considerably reduced.21 Reducing both the temperature and the thermal treatment duration, the annealing process becomes suitable also for hosting materials with thermal capacity smaller than PMMA, e.g., such as for conjugated polymers, that are also of great scientific and technological interest in energy conversion devices (e.g., electroluminescent, photovoltaic, and sensor devices). TGA analysis clearly shows that thermolysis of the precursors is quite similar for both Cd(SBz)2 and [Cd(SBz)2]2 · MI except, in the latter case, for the early stage removal of methyl imidazole, due to its higher volatility. The DTG graph clarified that in terms of time the molecule with methyl imidazole is extremely more reactive with a speed twice that of Cd(SBz)2. These data are in accordance with morphological, optical, and
Resta et al. structural analysis results that revealed a much higher growth efficiency when [Cd(SBz)2]2 · MI precursor is employed. Such a behavior suggests that when the precursor decomposition starts, the nuclei formation is favored by the presence of methyl imidazole due to different bond dissociation energy and the spatial arrangement (dispersion) in the polymer matrix with respect to Cd(SBz)2. Indeed, XRD measurements of nonannealed samples (last graphs in Figure 3a, b) show different molecular orderings for Cd(SBz)2 and [Cd(SBz)2]2 · MI precursors, the former exhibiting a spatial ordering, in accordance with a primitive cubic lattice,41 and with lattice constant of 1.37 nm, while the latter is characterized by a lamellar structure (superlattice period 1.65 nm).21 High-order Bragg peaks in the XRD pattern of nonannealed CBz sample (Figure 3a) suggests the formation of large “macromolecules”; thus, the molecules precursors are not homogeneously dispersed within the polymer matrix but are spatially localized forming large domains. Differently, the ordering of the [Cd(SBz)2]2 · MI is less pronounced and the domain size (lamellae) is much smaller, and such a molecular structure, corresponding ordering, and spatial arrangement within the polymer determine a lower decomposition temperature and a higher decomposition velocity. The details of the XRD pattern analysis of the CBz and CBz-MI samples prior annealing are reported in the Supporting Information (Figure S2). The possible spatial ordering and geometrical configuration of the Cd(SBz)2 and [Cd(SBz)2]2 · MI precursor molecules within the PMMA polymer matrix are shown in Figures S4 and S5, respectively, of the Supporting Information. These results are in very good agreement with TEM observations (Figure 4), as well as absorbance (Figure 5) and PL measurements (Figure 6). The differences in the growth evolution are partially overcome when annealing temperature is increased to TA g 220 °C. Here, the higher energy injection induces an increase in the precursor bond breaking and promotes the corresponding molecule mobility. Thus, the total amount of the precursor is involved, and, irrespective of the dispersion in the polymer, the final dimensions of the NCs depend only on the quantity of precursor used. Then, such a NCs formation and growth dynamic model explains not only the higher growth efficiency in CBz-MI samples at lower temperatures but also the saturation-like behavior observed for both precursors when the temperature TA is increased. In fact, in PL spectra of CBz samples annealed at low temperature (below 200 °C), the low-energy band emission dominates over band edge emission. This fact occurs mainly due to two reasons: (i) on one hand, the smaller size of the NCs and the higher density of trap state in their surface; (ii) on the other hand, the reduced intensity of CdS emission indicates a reduced number of defined NCs grown. Finally, the band edge emissions in CBz-MI and CBz samples become similar when TA ) 240 °C. A final consideration about the synthesis procedure follows. The results obtained in terms of optical response of the samples as well as the data obtained on morphology and structure of the CdS NCs bear out the effectiveness of the devised synthetic route for the formation of hybrid nanocomposite materials based on polymers doped with semiconductor NCs. The wide range of temperatures and the fine size tuning prove the accuracy on the growth control of CdS NCs below the exciton Bohr radius threshold. Furthermore, the results obtained with [Cd(SBz)2]2 · MI show that the addition of methyl imidazole promotes the solubility of the precursor in most of the common solvents and allows a homogeneous dispersion of the molecules in the hosting polymer, the latter being a fundamental prerequisite to obtain
Synthesis of Tailored CdS Nanocrystals nanocomposites with homogeneously distributed NCs, without particle aggregation and high-density fluctuation within the matrix. This statement is clearly demonstrated by Figure 4b where a representative low-magnification image shows that a homogeneous synthesis of narrow distributed CdS NCs was obtained over the whole extent of the matrix, thus promoting [Cd(SBz)2]2 · MI precursor as a promising material for the synthesis of hybrid organic/inorganic nanocomposite materials for fluorescent, electroluminescent, and photovoltaic applications. Conclusions New benzylthiolate-based unimolecular precursors for in situ synthesis of CdS NCs in polymer matrices were successfully synthesized and their thermophysical properties analyzed. The high efficiency of such innovative molecules was demonstrated in terms of a noteworthy lowering of the absolute temperature required for the production of semiconductor NCs in quantum confinement regime. At the same time the process duration was greatly reduced with respect to other unimolecular precursors previously investigated. In the range of annealing temperatures examined, 175 °C e TA e 240 °C, the size of CdS NCs was successfully tailored obtaining hybrid organic/inorganic nanocomposites with tunable optical properties in terms of absorption and light emission response in all the cases studied, while the PMMA polymer remained macroscopically/microscopically unaffected. These achievements suggest the possibility to extend this in situ synthesis route to fabricate hybrid nanocomposites to other polymer matrices even with notably lower heat capacity than in the present case, i.e., conjugated polymer. Furthermore, when methyl imidazole is added to cadmiumbis(benzylthiolates), the effectiveness of the synthesis route considerably increases in term of temperature and process duration, a uniform and homogeneous spatial precursor distribution within the polymer also being obtained, and consequently a high homogeneous nucleation and growth of CdS NCs in large areas/volumes of the polymer matrix. The latter result especially indicates [Cd(SBz)2]2 · MI precursor as an ideal molecule for the production of dense, well-dispersed, stoichiometrically, and morphologically controlled CdS NCs in organic matrices for optoelectronic and photovoltaic applications. Acknowledgment. The authors thank A. Cappello, T. Nocco, and M. Palmisano for their valuable technical assistance during the optical spectroscopy, XRD measurements, and TEM analysis. This work is supported by the Regione Puglia (Bari, Italy)sProject PONAMAT (PS_016). Supporting Information Available: Chemical structure of the precursor molecules (Figure S1); analysis of the XRD patterns of the samples with the precursors dispersed within the PMMA polymer prior to the thermal annealing process (Figure S2); details of the simulations of the XRD patterns obtained for samples annealed at T ) 240 °C (Figure S3); molecular arrangement (ordering) and configuration of the Cd(SBz)2 and [Cd(SBz)2]2 · MI precursor molecules within the polymer. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Burda, C.; Chen, X.; Narayanan, R.; El-Sayed, M. A. Chem. ReV. 2005, 105, 1025.
J. Phys. Chem. C, Vol. 114, No. 41, 2010 17317 (2) Trindade, Y.; O’Brien, P.; Pickett, N. L. Chem. Mater. 2001, 13, 3843–3858. (3) Brus, L. E. J. Chem. Phys. 1984, 80, 4464. (4) Colvin, V. L.; Schlamp, M. C.; Alivisatos, A. P. Nature 1994, 370, 354. (5) Mondal, S. P.; Reddy, V. S.; Das, S.; Dhar, A.; Ray, S. K. Nanotechnology 2008, 19, 215306. (6) Gross, S.; Camozzo, D.; Di Noto, V.; Armelao, L.; Tondello, E. Eur. Polym. J. 2007, 143, 673. (7) Chai, R.; Lian, H.; Li, C.; Cheng, Z.; Hou, Z.; Huang, S.; Lin, J. J. Phys. Chem. C 2009, 113, 8070. (8) Resta, V.; Laera, A. M.; Piscopiello, E.; Capodieci, L.; Ferrara, M. C.; Tapfer, L. Phys. Status Solidi A 2010, 207, 1631. (9) Liu, Z.; Miyauchi, M.; Uemura, Y.; Cui, Y.; Hara, K.; Zhao, Z.; Sunahara, K.; Furube, A. Appl. Phys. Lett. 2010, 96, 233107. (10) Review of recent works on in situ two-step synthesis: Dong, Y.; Jianmei, L.; Xu, Q. Polym. Composite 2009, 10, 723. Keeble, D. J.; Thomsen, E. A.; Stavrinadis, A.; Samuel, I. D. W.; Smith, J. M.; Watt, A. A. R. J. Phys. Chem. C 2009, 113, 17306. Chen, L.; Wang, C.; Li, Q.; Shengyang, Y.; Hou, L.; Chen, S. J. Mater. Sci. 2009, 44, 3413. Liao, H.C.; Chen, S.-Y.; Liu, D.-M. Macromolecules 2009, 42, 6558. Zhang, H.; Han, J.; Yang, B. AdV. Funct. Mater. 2010, 20, 1533. (11) Leventis, H. C.; King, S. P.; Sudlow, A.; Hill, M. S.; Molloy, K. C.; Haque, S. A. Nano Lett. 2010, 10, 1253. (12) Campbell, I. H.; Crone, B. K. AdV. Mater. 2006, 18 (1), 77. (13) Sto¨ferle, T.; Scherf, U.; Mahrt, R. F. Nano Lett. 2009, 9, 453. (14) Zorn, M.; Bae, W. K.; Kwak, J.; Lee, H.; Lee, C.; Zentel, R.; Char, K. ACS Nano 2009, 3, 1063. (15) Yun, D.; Feng, W.; Wu, H.; Yoshino, K. Sol. Energy Mater. Sol. Cells 2009, 93, 1208. (16) Xi, L.; Wen Tan, W. X.; Chua, K. S.; Boothroyd, C.; Lam, Y. M. Thin Solid Films 2009, 517, 6430. (17) Greenham, N. C.; Peng, X.; Alivisatos, A. P. Phys. ReV. B 1996, 54, 17628. (18) Deibel, C. Phys. Status Solidi A 2009, 206 (12), 2731. (19) Peng, X.; Manna, L.; Yang, W.; Wickham, J.; Scher, E.; Kadavanich, A.; Alivisatos, A. P. Nature 2000, 404, 59. (20) Rogach, A. L.; Kornowski, A.; Gao, M.; Eychmueller, A.; Weller, H. J. Phys. Chem. B 1999, 103, 3065. (21) Di Luccio, T.; Laera, A. M.; Tapfer, L.; Kempter, S.; Kraus, R.; Nickel, B. J. Phys. Chem. B 2006, 110, 12603. (22) Fragouli, D.; Resta, V.; Pompa, P. P.; Laera, A. M.; Caputo, G.; Tapfer, L.; Cingolani, R.; Athanassiou, A. Nanotechnology 2009, 20, 155302. (23) Landolt-Boernstein, Semiconductors Quantum Structures, Group III, Vol. 34, Sub-Volume C, Optical Properties, Part 2sSec. 5.5.13; Springer: Berlin, 2004. (24) Chan, Y.; Steckel, J. S.; Snee, P. T.; Caruge, J.-M.; Hodgkiss, J. M.; Nocera, D. G.; Bawendib, M. G. Appl. Phys. Lett. 2005, 86, 073102. (25) He, J.; Ji, W.; Ma, G. H.; Tang, S. H.; Kong, E. S. W.; Chow, S. Y.; Zhang, X. H.; Hua, Z. L.; Shi, J. L. J. Phys. Chem. B 2005, 109, 4373. (26) Menon, V. M.; Luberto, M.; Valappi, N. V.; Chatterjee, S. Opt. Express 2008, 16, 19535. (27) Khanna, P. K.; Singh, N. J. Lumin. 2007, 127, 474. (28) Bruchez, M., Jr.; Moronne, M.; Gin, P.; Weiss, S.; Alivisatos, A. P. Science 1998, 281, 2013. (29) Nair, P. S.; Radhakrishnan, T.; Revaprasadu, N.; Van Sittert, C. G. C. E.; Djokovic, V.; Luyt, A. S. Mater. Lett. 2004, 58, 361. (30) Dance, I. G. Polyhedron 1986, 5, 1037. (31) Rees, W. S., Jr.; Krauter, G. J. Mater. Res. 1996, 11, 3005. (32) Antolini, F.; Pentimalli, M.; Di Luccio, T.; Terzi, R.; Schioppa, M.; Re, M.; Mirenghi, L.; Tapfer, L. Mater. Lett. 2005, 59, 3181. (33) Basu, S.; Mondal, S.; Chatterjee, U.; Mandal, D. Mater. Chem. Phys. 2009, 116, 578. (34) Warren, B. E. X-Ray Diffraction; Dover Publications: New York, 1990. (35) PCPDFWIN v. 2.1, ICDD_JCPDS-International Centre for Diffraction Data, 2000. (36) Ida, T.; Shimazaki, S.; Hibini, H.; Toraya, H. J. Appl. Crystallogr. 2003, 36, 1107. (37) Schmidt, H. M.; Weller, H. Chem. Phys. Lett. 1986, 129, 615. (38) Efros, A. L.; Rosen, M. Annu. ReV. Mater. Sci. 2000, 30, 475. (39) Wang, Y.; Herron, N. J. Phys. Chem. 1988, 92, 4988. (40) Liptay, T. J.; Marshall, L. F.; Rao, P. S.; Ram, R. J.; Bawendi, M. G. Phys. ReV. B 2007, 76, 155314. (41) Suryanarayana, C.; Grant Norton, M. X-Ray Diffraction: A Practical Approach; Springer Verlag: New York, 1998; Part II, pp 97-105.