Enhancement of photoluminescence in manganese-doped ZnS nanoparticles due to a silica shell

JOURNAL OF CHEMICAL PHYSICS

VOLUME 118, NUMBER 19

15 MAY 2003

Enhancement of photoluminescence in manganese-doped ZnS nanoparticles due to a silica shell Anita S. Ethiraj, Neha Hebalkar, and S. K. Kulkarnia) Department of Physics, University of Pune, Pune-411007, India

Renu Pasricha National Chemical Laboratory, Pashan Road, Pune-41108, India

J. Urban Fritz Haber Institut der Max Planck Gesellshaft, Faradayweg 4-6, D-14195 Berlin, Germany

C. Dem, M. Schmitt, and W. Kiefer Institut fu¨r Physikalische Chemie, Universita¨t Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany

L. Weinhardt, S. Joshi, R. Fink, C. Heske, C. Kumpf, and E. Umbach Experimentelle Physik II, Universita¨t Wvrzburg, Am Hubland, D-97074 Wu¨rzberg, Germany

共Received 17 September 2002; accepted 19 February 2003兲 Zinc sulphide nanoparticles doped with manganese 共ZnS:Mn兲 have been stabilized using thioglycerol 关 HSCH2CH共OH兲CH2OH兴 molecules. The nanoparticles 共⬃1.7 nm兲 are highly stable and exhibit photoluminescence at ⬃600 nm when excited with ultraviolet light. For increasing luminescence and stability the particles are further treated with tetraethylorthosilicate 共TEOS兲 关 Si共C2H5O兲4兴 in an aqueous medium, yielding either a disordered silica matrix or spherical core-shell particles of up to ⬃900 nm size with strongly enhanced luminescence under certain conditions. Photoluminescence, excitation spectroscopy, transmission electron microscopy, energy dispersive analysis of x-rays, x-ray diffraction, Raman spectroscopy, and x-ray photoelectron spectroscopy measurements have been performed for the characterization of the ZnS:Mn nanoparticles alone, in the silica matrix as well as in spherical silica shells. Among other things, the analysis indicates that the thioglycerol capping has been affected by the coating neither in the silica matrix nor in core-shell particles. © 2003 American Institute of Physics. 关DOI: 10.1063/1.1566932兴

I. INTRODUCTION

than conventional fluorophores is that particles of the same material but of different sizes fluoresce at different wavelengths, but can be simultaneously excited with a single wavelength in the UV range, also reducing the photobleaching. Semiconductor nanoparticles can be doped with various transition or rare-earth ions20–25 fluorescing at different wavelengths depending upon the dopant. Such doped particles can also produce photoluminescence at different wavelengths depending upon the dopant, using a single wavelength in the UV range for excitation. Manganese-doped ZnS 共ZnS:Mn兲 and CdS 共CdS:Mn兲 nanoparticles have been widely investigated24 –28 to probe the possible enhancement and quantum size effects on fluorescence. We also have synthesized ZnS nanoparticles doped with manganese and other elements.21–23 However, it is a common experience that the nanoparticle surfaces, which are the more important the smaller the particles are, are strained, have dangling bonds, are susceptible to oxidation, coalescence, or are prone to other instabilities.29,30 Therefore the nanoparticles are often surface passivated14 –16 using some organic or inorganic molecules. In spite of this, particles tend to agglomerate3 and/or ‘‘age’’ because of oxidation, photolytic decomposition of the capping material, etc. In order to overcome this drawback nanoparticles can be either coated with a large-band-gap semiconductor material, a polymer, or

Currently, there is very much interest in core-shell particles1–7 as they have properties which may differ from either those of the core or of the shell material. Core-shell particles, in which the core is a metal and the shell is silica 共or a polymer兲 or vice versa, have shown changes of their optical properties. Such particles are being investigated as possible candidates for sensor materials, pigments, precursors for photonic crystals, and so on. Although there are many attempts to make metal core and silica shell particles 共or vice versa兲, so far there are very few reports on semiconductor core and silica shell particles. It is indeed possible to make semiconductor–silica core shell particles8 –10 which may have better properties than nanoparticles in some respects. Semiconductor nanoparticles have advantageous optical properties like size-dependent absorption and fluorescence,11–16 which can be correlated to size quantization effects. In some cases the fluorescence yield increases in nanoparticles as compared to the bulk material. This leads to applications like biological labels17,18 or bar codes.4,19 One advantage of using nanoparticles as fluorescent objects rather a兲

Author to whom correspondence should be addressed. Electronic mail: [email protected]

0021-9606/2003/118(19)/8945/9/$20.00

8945

© 2003 American Institute of Physics

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

8946

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

with insulators like silica, titania, etc., leading to core-shell particles. In this paper we discuss some interesting experiments in which Mn-doped ZnS 共ZnS:Mn兲 nanoparticles capped with thioglycerol 共TG-capped ZnS:Mn兲 are synthesized using the method of size selective precipitation. Fairly well fluorescing particles are produced by this procedure. The nanoparticles are then treated with an aqueous solution of silica (SiO2) particles producing liquid. 关It should be mentioned that no attempt has been made to determine the exact Si:O ratio in the different samples synthesized here, although they are referred to as silica- (SiO2-) coated particles, as both x-ray photoemission spectroscopy 共XPS兲 and energy dispersive analysis of x-rays 共EDAX兲 analysis showed an O:Si ratio larger than 2.兴 The silica treatment results in highly luminescent particles referred to as ZnS:Mn@SiO2. Upon reaction with the silica-particle producing solution the ZnS:Mn nanoparticles undergo nucleation and growth until macroscopic, spherical silica particles of about 900 nm size are produced in which the ZnS:Mn nanoparticles are entrapped. For the analysis of the ZnS:Mn@SiO2 particles at the various stages of growth a variety of techniques like photoluminescence 共PL兲, photoluminescence excitation spectroscopy 共PLE兲, x-ray diffraction 共XRD兲, XPS, Raman spectroscopy 共RS兲, transmission electron microscopy 共TEM兲, and EDAX were employed. It was observed as discussed below that the thioglycerol capping was retained inside the silica coating and that the luminescence for the ZnS:Mn@SiO2 as well as for the ZnS:Mn nanoparticles dispersed in a disordered silica matrix is strongly enhanced. II. EXPERIMENT

The synthesis of ZnS:Mn@SiO2 samples using the sizeselective precipitation method16 was carried out by first producing ZnS samples doped with manganese and then reacting them in a silica-growing solution. For ZnS:Mn synthesis the following steps were taken 共Fig. 1兲. First, equimolar 共2.2 mM兲 aqueous solutions of ZnCl2 and MnCl2 were stirred together at room temperature. Thioglycerol 共3.5 mM兲 was added while the stirring was continued. Anhydrous Na2S 共0.5 mM兲 aqueous solution was then added and the solution was heated to 80 °C for 22 h while stirring. The solution was allowed to cool to room temperature and concentrated in a rotary evaporator. Acetone was added until flocculation occurred. Centrifugation was carried out, and the supernatant was removed to obtain the precipitate. The latter was washed thoroughly in methanol followed by washing in diethylether and drying in vacuum. Washing alternately in methanol and diethylether was continued until no change was observed in the UV absorption spectra. An excitonic peak was observed at ⬃280 nm, with a peak shape similar to that reported in our earlier work21 for ZnS:Mn. It should also be noted that small amounts of Mn doping less than about 1% do not affect the ZnS absorption spectra which are very similar to undoped nanoparticle spectra. Nearly monodispersed silica particles 共⬍5%兲 were routinely produced by a procedure first suggested by Sto¨ber et al.,31 adjusting the concentrations of ethanol:water:ammonium hydroxide:TEOS „tetraethylorthosilicate 关 Si共C2H5O兲4兴 …

Ethiraj et al.

and the time for which they react. As the reaction time increases, the particles of silica grow from tens of nanometers to hundreds of nanometers in size. A solution of ethanol:water:ammonium hydroxide:49:31:4 was prepared and stirred vigorously for 30 min at room temperature. The ZnS:Mn powder was added to the solution and further stirred for 60 min. TEOS was added to this solution and stirred further for various timings to obtain different samples. The solution was put in a centrifuge to obtain the precipitate, which was then washed in doubledistilled water and dried at room temperature. The samples so obtained were completely water soluble. Figure 1 gives the flow chart for producing SiO2, TG-capped ZnS:Mn, and ZnS:Mn@SiO2 particles. The photoluminescence measurements were carried out using a Perkin–Elmer model LS-50 and a xenon flash lamp as the source of excitation. Excitation and emission spectra in the range 200–900 nm could be recorded as follows. Initially, the wavelength of the excitonic peak for ZnS:Mn, observed in UV absorption at ⬃280 nm, was chosen as fixed wavelength for excitation. With a filter at 390 nm the emission 共photoluminescence兲 spectrum was recorded in the range from 400 to 900 nm. Using the wavelength of the maximum PL intensity, the excitation spectrum was recorded in the 250–350 nm 共PLE兲 range. The peak of maximum intensity was again chosen to record the emission spectrum 共PL兲. By repeating the procedure a couple of times a consistent set of excitation-emission spectra was obtained such that the photoluminescence spectra were recorded by the wavelength of maximum intensity in the excitation spectra and the excitation spectra with the maximum intensity peak of the PL spectra. Thus an excitation peak was observed at ⬃315 nm and a photoluminescence peak at ⬃600 nm. The PLE and PL results will be discussed in Sec. III. For the XRD experiment a Philips powder diffractometer 共model PW 1840兲 with a stationary anode x-ray source 关wavelength 1.54 Å 共Cu K ␣ )] was used. The experiment was performed in a geometry similar to the Debye–Scherrer geometry, but with a modified sample holder system. The powder was prepared as a layer of approximately 1 mm thickness on a glass sample holder and measured in reflection geometry with respect to the holder. Thus diffuse background scattering from the glass holder could be avoided to a large extent. Anomalous XRD was performed at beamline BW2 at the DORIS synchrotron of the Hamburger Synchrotronstrahlungslabor HASYLAB in Hamburg, Germany, in a common Debye–Scherrer geometry. A glass capillary was used as a sample holder in this case. The Raman spectra were taken with a micro-Raman setup 共LabRam, Jobin-Yvon-Horiba兲. The spectrometer has a focal length of 300 mm and is equipped with a 950-lines/mm grating. The spectral resolution was about 6 cm⫺1. As excitation wavelength the 632.8-nm line of a He:Ne laser with a power of about 1 mW was used. The scattered light was detected by a CCD camera operating at 220 K. An Olympus MLPlanFL 50 objective focused the laser light onto the powder samples. Raman spectra were recorded in the 150– 4000 cm⫺1 wave number region. X-ray photoelectron spectroscopy was performed using

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

Enhancement of photoluminescence in manganese-doped ZnS

8947

FIG. 1. Flow chart of synthesis (SiO2, ZnS:Mn, ZnS:Mn@SiO2).

a Thermo-VG Multilab machine equipped with a CLAM4 analyzer and a Mg/Al twin-anode x-ray source. The survey spectra discussed below were recorded using the Mg source with 300 W and the analyzer with 50 eV pass energy, resulting in 2.5 min recording time for each spectrum. Charging was reduced by pressing the sample powder into an In foil resulting only in a charging shift of less than 2 eV with no significant additional peak broadening. Close-up spectra around all predominant peaks were also recorded using 20 eV pass energy for checking details, but these are not shown here since they do not add significant information. The spectra were plotted as recorded; i.e., no additional normalization was tried due to the lack of suitable normalization procedures. Rough relative intensities can be derived from peak heights since the peak shapes remained constant for the various samples. Transmission Electron Microscopy was carried out using

a Philips CM200 FEG microscope equipped with a field emission gun. Energy dispersive analysis of x rays also could be performed in reflection geometry in the same microscope. The particles were dispersed in water, and a drop of solution was allowed to dry on a 5-nm-thick carbon film deposited on a copper grid. The copper grid was transferred into the microscope after the liquid was evaporated. The images were recorded using 200 keV. A resolution of ⬃0.2 nm could be achieved. III. RESULTS AND DISCUSSIONS

In this investigation we have synthesized ZnS:Mn nanoparticles capped with thioglycerol having an average size of ⬃1.7 nm as determined by XRD analysis as discussed below. Capping with thioglycerol molecules avoids further growth or agglomeration of the particles. Atomic absorption mea-

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

8948

Ethiraj et al.

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

surements of ZnS:Mn-doped particles have shown that there is 0.056 wt % of Mn. This corresponds to about one Mn atom per ZnS nanoparticle.25 The optical properties of Mndoped ZnS nanoparticles have been reported earlier.21,24,28 Although the initial suggestion of Bhargava et al.20,24 that the doped nanoparticles of semiconductors form a new class of luminescent materials has been questioned in recent years,28,32,33 it has been found for nanoparticles of CdS and ZnS doped with manganese that magnetic interactions strongly increase with the reduction in particle size, giving rise to giant internal magnetic fields.34,35 These effects, though very important, are not addressed here, but we concentrate on the aspects related to the luminescence, structure, and other properties of Mn-doped ZnS nanoparticles without and with silica shells. The effect of quantum confinement on the luminescence efficiency in the case of Mn-doped ZnS nanoparticles is a point of debate. Nevertheless, strong photoluminescence is observable from Mn-doped ZnS nanoparticles due to the occupancy of Zn2⫹ sites by Mn2⫹ in the ZnS lattice. This produces localized levels ( 4 T 1 and 6 A 1 ) in the energy gap of the semiconductor ZnS nanoparticle due to crystal field effects24 similar to that in the bulk ZnS:Mn. An electron can undergo photoexcitation process in the host ZnS lattice of nanoparticles and subsequently decay via a nonradiative transition to the 4 T 1 level from which it then makes a radiative transition to the 6 A 1 level. It should be mentioned that this transition from the 4 T 1 level to the 6 A 1 level is the same as in bulk ZnS:Mn. However, in the semiconductor nanoparticles the energy gap is larger than that in the bulk11–14 and an excitonic peak is observed in the optical absorption spectra. In our case the optical absorption peak occurs at 280 nm 共not shown兲 occurs at ⬃280 nm. One would have expected a sharp peak at ⬃280 nm in the photoluminescence excitation spectrum also in the ZnS:Mn nanoparticle samples here, provided all the photoexcited electrons were directly transferred to the 4 T 1 level. However, the appearance of a peak at a longer wavelength ⬃315 nm in the photoexcitation spectra suggests that some levels 共defect levels兲 do exist in the energy gap which participate in the photoluminescence process. In other words, defect states of the host ZnS lattice either get excited directly or photoexcited 共at least some兲 electrons are first transferred to the defect states nonradiatively and subsequently transferred to 4 T 1 . It has been reported by various groups21,24,28 that a transition from 4 T 1 to 6 A 1 gives a photoluminescence peak between 580 and 600 nm and does not seem to be much influenced by the nanoparticle size. There is no size effect reported so far on the peak position of this radiative transition. In the present case we also observe a strong photoluminescence for the ZnS:Mn nanoparticles at ⬃600 nm. Following Sto¨ber et al.,31 we have established in our laboratory a procedure for synthesizing monodispersed large silica particles. In Fig. 2 a TEM photograph of SiO2 particles is given as an example. It can be noticed that the particles are uniform and ⬃250 nm in size 共⬍5% dispersed兲. It required 180 min reaction time to get particles of this size. The particles are water soluble and highly stable. They could be stored as dry powder and dissolved in water at a later time.

FIG. 2. TEM photograph of SiO2 particles.

The ZnS:Mn@SiO2 samples resulted from a reaction of ZnS:Mn nanoparticles with a silica particle-producing solution as discussed in the experimental section 共see also Fig. 1兲. The solution was stirred up to 180 min, and some fraction of the solution was removed at various times, producing different samples. Interestingly, even the samples reacted as shortly as 5 or 15 min showed chemical and morphological stability, giving reproducible results even after 2 months. In Fig. 3共a兲, on the right-hand side photoluminescence spectra are shown for TG-capped ZnS:Mn nanoparticles along with those for nanoparticles treated in silica solution for 5 min and for pure SiO2 particles. On the left, excitation spectra 共PLE兲 recorded for ⬃600 nm emission are displayed. It can be seen that a continuous range of wavelengths can be employed to excite the 600 nm photoluminescence line. In fact, this feature of nanoparticles viz., the ability to excite visible photoluminescence with a broad range of wavelengths—is considered to be very important.17 This was also discussed in Sec. I. Here the PL spectra that could be excited by any wavelength in the UV range in nanoparticles show a broad peak around 600 nm emission which is typical for the Mn 4 T 1 to 6 A 1 transition even in the bulk, as discussed above. The excitation spectrum of ZnS:Mn and in particular that of ZnS:Mn@SiO2 displays a broad peak around 315 nm and additional oscillator strength at shorter wavelengths. Note that the luminescence at ⬃600 nm occurs after the transfer of the excitation from the host ZnS lattice to the Mn ion. Of course, pure SiO2 particles have negligible

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

FIG. 3. 共a兲 Left: excitation spectra of the various samples indicated in the figure using the maximum of the PL spectra as detection energy. Right: photoluminescence spectra of the same samples excited at the maximum of the excitation spectrum. 共b兲 Photoluminescence intensity 共recorded at 600 nm兲 plotted versus SiO2 reaction time of ZnS:Mn nanoparticles.

fluorescence in the visible wavelength region and also no excitation resonance peak. For all ZnS:Mn@SiO2 samples investigated here the PL emission occurs at ⬃600 nm. However, the intensity varies with the reaction time of the ZnS:Mn particles with the silica-producing solution. In Fig. 3共b兲 the intensity of the PL peak at 600 nm is plotted against the reaction time. It is clearly seen that initially, i.e., from the bare, TG-capped ZnS particle, the intensity increases by a factor of 3 at 5–15 min reaction time, but then decreases for 30 and 180 min. Note that, despite the SiO2 shell around the luminescent ZnS core with its attenuation effect, there was more PL intensity for all SiO2-treated samples than for the bare, TG-capped ZnS:Mn particles. Even the very large spherical particles formed after 180 min reaction time 共see TEM below兲 showed a PL intensity significantly larger than that from the untreated particles. This is quite interesting as—unlike in metallic samples— there is no shift in absorption or luminescence peak position, but the intensity changes are prominent. We interpret this enhancement of the photoluminescence intensity as removal of electron capture centers on the TG-capped surface of ZnS:Mn particles and/or as removal of nonradiative decay channels due to the silica shell. Moreover, the silica shell around the ZnS:Mn particles also can protect the particles from the ambient. In order to understand where the ZnS

Enhancement of photoluminescence in manganese-doped ZnS

8949

nanoparticles are located with respect to silica—viz., inside the silica, on the silica, or attached to the silica—various other experiments using different techniques were performed. Figure 4共a兲 presents a low-resolution TEM photograph of ZnS:Mn@SiO2 for 15 min. It can be seen that there is a considerable disorder in general. However, magnified images of various regions of the same sample show that some regular particles have started growing. An example is given in Fig. 4共b兲. The TEM photograph of Fig. 4共c兲, recorded for ZnS:Mn@SiO2 after 180 min reaction, shows the formation of spherical particles of ⬃900 nm diameter. The EDAX analysis indicates the presence of Zn, S, Si, O, and a small amount of C in ZnS:Mn@SiO2 for 5- and 15-min-grown samples, but no Zn or S for the 180-min-grown sample. The latter finding is no surprise as EDAX was recorded in the backreflection mode and the 1.7-nm ZnS cores are inside 900-nm large particles. However, the fact that one can still observe larger photoluminescence intensity 共see Fig. 3兲 compared to the bare TG-capped ZnS:Mn nanoparticles and that also Raman peaks 共Fig. 7兲 from the ZnS core and the TG capping are still visible in this case while the corresponding 共surface-sensitive兲 XPS spectrum 共Fig. 8兲 does not contain Zn and S peaks is a clear indication that ZnS:Mn particles are indeed encapsulated by silica coating which is transparent to visible and UV radiation. In Fig. 5 six different XRD measurements for samples with different SiO2 reaction times 共5, 15, 30, and 180 min兲 and, for comparison, ZnS:Mn nanoparticles without SiO2 coating and pure SiO2 particles are shown. The diffraction patterns for the pure ZnS:Mn and the 5-, 15-, and 30-min SiO2-reacted samples show three wide Bragg reflection peaks at 2⌰⬇28°, 48°, and 55° for each sample which are significantly broadened by the small particle size. These patterns, and especially the equal intensity of the Bragg peaks at 48° and 55°, are a typical fingerprint of the zinc-blende structure, which is in agreement with previous results on mercaptoethanol-capped ZnS nanoparticles.36 An additional feature which is clearly visible in the diffraction patterns of the ZnS:Mn@SiO2 samples is a shoulder at the low-angle tail of the peak at 28° which stems from SiO2 共see also the peak at 2⌰⬇24° for pure SiO2). This shoulder becomes more pronounced with increasing SiO2-reaction time 共an exception is the curve for 15 min reaction time which shows sharper peaks possibly due to some agglomeration兲. After 180 min reaction time the intensity of the ZnS Bragg peaks has almost vanished, indicating that the ZnS contribution to the sampled volume became too small compared to that of SiO2. Although an exact determination of the number of ZnS nanoparticles inside the SiO2 shell is not possible from the experiments carried out so far, we estimate that there are ⬃2.5⫻109 particles of ZnS in ⬃2.9⫻1012 SiO2 molecules or inside a 900-nm SiO2 particle. The estimation is based on the EDAX results and on the relative amounts of precursor chemicals used during synthesis. One possibility to separate the ZnS:Mn signal from the SiO2 signal is using anomalous XRD.37 Figure 6 shows a first test experiment with the 15-min sample performed at the

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

8950

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

Ethiraj et al.

FIG. 4. 共a兲 Low-resolution TEM of ZnS:Mn@SiO2, 15 min. 共b兲 High-resolution TEM of ZnS:Mn@SiO2, 15 min. 共c兲 TEM of ZnS:Mn@SiO2, 180 min.

Zn K-absorption edge (h v ⫽9.655 keV兲 and slightly below (h v ⫽9.630 keV兲. At these two photon energies the atomic form factor for all atoms does not vary significantly, except for the Zn atoms. For those the difference is large enough to be detected 关 f Zn(9.655 keV)⫽Z⫺7.73, f Zn(9.63 keV)⫽Z ⫺5.53, with Z⫽30], if the counting time provides sufficient statistics. When the two measurements are subtracted from each other all signal contributions, except for those coming from the Zn atoms, cancel out. This is nicely demonstrated in Fig. 6, which shows the measurements at both energies and the difference pattern. In the two upper curves the shoulder at q⫽1.6 Å⫺1 representing the SiO2 feature is clearly visible, whereas in the difference curve it has almost disappeared. An easy way to obtain an approximate average size of the ZnS:Mn particles from this diffraction pattern is the Scherrer equation d⫽0.94␭/FWHM(2⌰)cos ⌰, where d is the diameter of the crystalline core of the particles, ⌰ is the Bragg angle of the peak considered, and FWHM共2⌰兲 is its full width at half maximum in 2⌰. Using this simple approach, an average diameter of 1.7 nm is found. Note that this number is a crude estimate, since in this range the Scherrer equation represents an extrapolation. However, all other

FIG. 5. X-ray diffraction patterns for various ZnS:Mn samples with different SiO2 reaction times and for pure SiO2. Curves are smoothed for clarity. The error bar indicates the noise level.

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

Enhancement of photoluminescence in manganese-doped ZnS

8951

FIG. 6. Anomalous x-ray diffraction patterns for the ZnS:Mn@SiO2 sample after 15 min reaction time measured at two different energies close to the Zn K-absorption edge. The intensity is plotted vs q⫽4 ␲ sin ␪ /␭. The difference pattern represents the predominant scattering from the Zn-containing nanoparticle.

methods for size determination also provide only rough estimates as, e.g., the size determination from UV–VIS absorption data4,11–16 via an effective mass approximation model. In our case, this approach yields a particle size of ⬃1.4 nm 共UV absorption peak at ⬃280 nm兲. We obtain using the effective mass approximation11 and using the energy gap determined by the UV absorption peak at ⬃280 nm for ZnS:Mn particles used in this work, particles size ⬃1.4 nm close to that determined by x-ray diffraction. Our present results thus represent an alternative approach and hence contribute to an improvement of the size information for nanosized particles. In the absence of any reliable method of average particle size determination at nanometric length scale this approach of using the Scherrer equation to determine the average particles size of nanoparticles is often used. Further, it should be mentioned that we are currently working on a simulation of powder diffraction patterns which is specially designed for nanoparticles and clusters. The first results achieved by this code, which does not use the Scherrer equation, but summation of scattered wave functions from the individual atoms of a nanoparticle model, are relatively close to the results obtained here by means of the Scherrer equation. More detailed results for an extended series of anomalous XRD measurements on samples with different SiO2 reaction times are on the way and will be published in subsequent papers. In summary, both the ZnS:Mn particle size determination directly from the width of the x-ray diffraction peaks and the effective mass approximation using the position of absorption peak at 280 nm give an average size close to ⬃1.7 nm. The Raman spectroscopy results throw additional light on this situation. In Fig. 7 Raman spectra are shown for SiO2, TG, and some TG-capped ZnS and ZnS:Mn@SiO2 samples. A detailed analysis of various vibrational features and photon modes will be published elsewhere. Briefly, we observe the following: The TG-capped ZnS sample 共second spectrum from bottom兲 exhibits a large peak at about 255 cm⫺1, which is due to ZnS LO phonons, and several peaks at higher wave numbers, which are due to the capping agent

FIG. 7. Micro-Raman spectra recorded with 632 nm laser excitation for thioglycerol 共TG兲 molecules in solution, for TG-capped ZnS particles of 1.7 nm size, for TG-capped ZnS particles embedded in a disordered silica matrix 共5 min reacted兲, and in a silica shell 共180 min reacted兲, respectively, and for pure SiO2 particles of 250 nm size.

thioglycerol 共TG兲. The comparison with liquid TG 共bottom spectrum兲 indeed shows many similarities, but also some differences. The most pronounced differences are the new large peak at 740 cm⫺1 appearing upon TG capping and the complete disappearance of the largest peak at 2550 cm⫺1 共representing the S–H bond of TG兲. The latter peak is observed in the TG solution 共Fig. 7, inset兲 and disappears upon the formation of the thiol bond of TG to the ZnS nanoparticle. Of course, these two spectra indicate that the ZnS particles are TG capped and that the bonding occurs via the thiol band, as expected. It is interesting to note that in the ZnS sample all TG molecules are reacted and that significant changes occur around 740 cm⫺1, indicating the specific bonding situation of TG on the ZnS particle. What is most important for the present paper, however, is the similarity of the ZnS and all ZnS:Mn@SiO2 samples in the Raman spectra. In fact, not only the 共slightly attenuated兲 mode at 255 cm⫺1 remains, indicating the presence of ZnS nanoparticles even after 180 min of SiO2 treatment, but also the other vibrational TG modes survive 共see, e.g., the modes at 735, 1021, 1067 cm⫺1, etc.兲. This implies that the ZnS:Mn particles retain their TG capping even when completely embedded in a silica shell after 180 min reaction 共see below兲. Of course, the Raman spectrum from the ZnS:Mn@SiO2 180 min reaction sample also has similarities with that of the pure SiO2 sample 共top spectrum兲, especially in the region around 400 cm⫺1, although the sharp peaks of the latter

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

8952

Ethiraj et al.

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003

could be detected by XPS, which is not surprising in view of the very small amount of this dopant, in particular if Mn is located inside the particle. The most interesting observation in the present context is the presence of Zn and S peaks after 5 or 15 min silica treatment and their complete disappearance after 180 min treatment 共see three spectra in the middle of Fig. 8兲. While their presence clearly indicates that for the disordered silica matrix achieved after only several minutes of silica treatment, the embedding of the ZnS particles by a SiO2 shell is incomplete, their disappearance after 180 min treatment is sufficient proof that the ZnS particles are completely surrounded by a protective SiO2 layer. It is also noted that the Zn/Si and S/Si ratios do not develop precisely as expected from the reaction time as can be derived from a detailed comparison of the two lower ZnS:Mn@SiO2 spectra. The origin of this behavior is presently unclear. IV. CONCLUSIONS

FIG. 8. XPS survey scans from various samples as indicated. The spectra were recorded with 50 eV pass energy using Mg K␣ radiation 共1253.6 eV兲 for excitation. The sample material was pressed into an In foil to reduce charging effects.

sample appear to be significantly broadened for the coreshell particles for reasons that are presently unknown. The appearance of all vibrational peaks related to ZnS and TG also means that the silica shells are transparent to the 632 nm wavelength used for exciting the Raman spectra. In Fig. 8 survey XPS spectra of all samples discussed in this paper are displayed. Since XPS is a rather surfacesensitive technique 共the sampling depth here is below ⬃4 nm, depending on electron kinetic energy兲, these spectra reveal mostly the stoichiometry and chemical state of the outermost layers of the various particles 共and of course of the In foil as well兲. Numerous photoemission and Auger peaks appear in the spectra, the most prominent of which are labeled, indicating the element contributions to the various surfaces. The spectra of pure ZnS:Mn and SiO2 samples display the peak positions and relative intensities expected for these materials. It is surprising, however, that in contrast to CdS,36 the cation Zn appears much more intense 共ratio ZnS⬃1:1, in spite of the additional thiol-sulfur兲 than in the CdS case. This might be due to the fact that the surface of the nanoparticles is predominantly terminated by Zn. It is also noted that some intensity from O- and C-containing impurities is observed 共in addition to the C and O peaks arising from the TG capping visible in some of the samples兲, which is quite normal for samples that have been transferred through atmosphere and have not been outgassed by heating before measurement, as in the present case 共for avoiding any temperature induced changes兲. These impurities are probably adsorbed on top of the particles and the In foil. It is further noted that no Mn

We have synthesized silica-protected, thioglycerol共TG-兲 capped, highly luminescent, ZnS:Mn nanoparticles 共⬃1.7 nm size兲. They show a clear enhancement of photoluminescence intensity as compared to TG-capped ZnS:Mn nanoparticles without silica coating. The PL intensity peaks around 600 nm—when excited with any wavelength in the UV range—indicate that PL stems from embedded Mn ions after the excitation is transferred from the excited ZnS to the Mn atoms. The silica shells did not modify the photoluminescence excitation or emission spectra except for changing the intensities. The silica shell was found to be semitransparent to UV and visible radiation. Thus Raman spectroscopy could still detect the ZnS LO phonons and the TG-capping molecules even for ZnS:Mn particles embedded in a ⬃900 nm diameter SiO2 particle. Surface-sensitive XPS proved that the ZnS:Mn nanoparticles are indeed completely covered by SiO2. TEM was utilized to give information about the sample morphology and structure, showing a disordered silica matrix after short, and spherical core-shell particles after long reaction times. X-ray diffraction 共normal and anomalous兲 experiments finally yielded structural information 共⬃1.7 nm size, zinc-blende structure兲 which is consistent with all other findings.38 – 40 ACKNOWLEDGMENTS

S.K.K. thanks UGC, India, for continuous support, A.S.E. thanks IUC-DAEF, India, and N.H. thanks R&D Engrs. Pune, India, for financial support. Also, financial support by the Volkswagenstiftung through project I/74922 and by the Deutsche Forschungsgemeinschaft through Sonderforschungsbereich 410 共projects C3 and C5兲 is gratefully acknowledged. Two of us 共E.U. and W.K.兲 acknowledge financial support by the Fond der Chemischen Industrie. F. Caruso, Adv. Mater. 13, 11 共2001兲. Y. Xia, B. Gates, Y. Yin, and Y. Lu, Adv. Mater. 12, 693 共2000兲. 3 P. Mulvaney, L. M. Liz-Marzan, M. Giersig, and T. Ung, J. Mater. Chem. 10, 1259 共2000兲. 4 N. Gaponik, I. L. Radtchenko, G. B. Sukhorukov, H. Weller, and A. L. Rogash, Adv. Mater. 14, 879 共2002兲. 1 2

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp

J. Chem. Phys., Vol. 118, No. 19, 15 May 2003 5

B. Schreder, T. Schmidt, V. Ptatschek, U. Winkler, A. Materny, E. Umbach, M. Lerch, G. Mu¨ller, W. Kiefer, and L. Spanhel, J. Phys. Chem. B 104, 1677 共2000兲. 6 S. Peng, M. C. Schlamp, A. V. Kadavanich, and A. P. Alivisatos, J. Am. Chem. Soc. 119, 7019 共1997兲. 7 J. B. Jackson and N. J. Halas, J. Phys. Chem. B 105, 2743 共2001兲. 8 M. A. Correa-Durate, M. Giersig, and M. Liz-Marzan, Chem. Phys. Lett. 286, 497 共1998兲. 9 N. A. Dhas, A. Zaban, and A. Gedanken, Chem. Mater. 11, 806 共1999兲. 10 D. Gerion, F. Pinaud, S. C. Williams, W. J. Parak, D. Zanchet, S. Weiss, and A. P. Alivisatos, J. Phys. Chem. B 105, 8861 共2001兲. 11 L. E. Brus, Appl. Phys. A: Solids Surf. 53, 465 共1991兲. 12 Henglein, Top. Curr. Chem. 143, 113 共1988兲. 13 H. Weller, Angew. Chem. Int. Ed. Engl. 32, 41 共1993兲. 14 U. Woggon, Optical Properties of Semiconductor Quantum Dots 共Springer-Verlag, Berlin, 1997兲. 15 S. V. Gaponenko, Optical Properties of Semiconductor Nanocrystals 共Cambridge University Press, Cambridge, England, 1998兲. 16 C. B. Murray, D. B. Norris, and M. G. Bawendi, J. Am. Chem. Soc. 115, 8706 共1993兲. 17 M. Bruchez, M. Moronne, P. Gin, S. Weiss, and A. P. Alivisatos, Science 281, 2013 共1998兲. 18 W. C. W. Chan and S. Nie, Science 281, 2016 共1998兲. 19 B. J. Battersby, D. Bryant, W. Meutermans, D. Matthews, M. L. Smythe, and M. Trau, J. Am. Chem. Soc. 122, 2138 共2000兲. 20 R. N. Bhargava, D. Gallagher, and T. Welker, J. Lumin. 60-61, 275 共1994兲. 21 A. A. Khosravi, M. Kundu, B. A. Kuruvilla, G. S. Shekhawat, R. P. Gupta, A. K. Sharma, P. D. Vyas, and S. K. Kulkarni, Appl. Phys. Lett. 67, 2506 共1995兲. 22 A. A. Khosravi, M. Kundu, S. K. Deshpande, U. A. Bhagwat, M. Sastri, and S. K. Kulkarni, Appl. Phys. Lett. 67, 2702 共1995兲. 23 P. H. Borse, N. Deshmukh, R. F. Shinde, S. K. Date, and S. K. Kulkarni, J. Mater. Sci. 34, 6087 共1999兲.

Enhancement of photoluminescence in manganese-doped ZnS

8953

24

R. N. Bhargava, D. Gallagher, X. Hong, and A. Nurmikko, Phys. Rev. Lett. 72, 416 共1994兲. 25 P. H. Borse, D. Srinivas, R. F. Shinde, S. K. Date, W. Vogel, and S. K. Kulkarni, Phys. Rev. B 60, 8659 共1999兲. 26 L. Levy, D. Ingert, N. Feltin, and M. P. Pileni, J. Cryst. Growth 184-185, 377 共1998兲. 27 A. D. Dinsmore, D. S. Hsu, H. F. Gray, S. B. Quadr, Y. Tian, and B. R. Ratna, Appl. Phys. Lett. 75, 802 共1999兲. 28 A. A. Bol and A. Meijernik, J. Lumin. 87-89, 315 共2000兲. 29 X. Peng, J. Wickham, and A. P. Alivisatos, J. Am. Chem. Soc. 120, 5343 共1998兲. 30 D. V. Talapin, A. L. Rogash, M. Haase, and H. Weller, J. Phys. Chem. B 105, 12 278 共2001兲. 31 W. Sto¨ber, A. Fink, and E. Bohn, J. Colloid Interface Sci. 26, 62 共1968兲. 32 S. K. Kulkarni, U. Winkler, N. Deshmukh, P. H. Borse, R. Fink, and E. Umbach, Appl. Surf. Sci. 169-170, 438 共2001兲. 33 J. Z. Zhang, J. Phys. Chem. B 104, 7239 共2000兲. 34 D. M. Hoffman, B. K. Meyer, A. I. Akimov, Al. L. Efros, M. Rosen, G. Couino, T. Garcoin, and J. P. Boilot, Solid State Commun. 114, 547 共2000兲. 35 N. Feltin, L. Levy, D. Ingert, and M. P. Pileni, Adv. Mater. 11, 398 共1999兲. 36 W. Vogel, P. H. Borse, N. V. Deshmukh, and S. K. Kulkarni, Langmuir 16, 2023 共2000兲. 37 U. Winkler, D. Eich, Z. H. Chen, R. Fink, S. K. Kulkarni, and E. Umbach, Chem. Phys. Lett. 306, 95 共1999兲. 38 W. Chen, Jan-Olle Malm, V. Zweiller, Y. Huang, S. Liu, R. Wallenberg, Jan-Olov Bovin, and L. Samuelson, Phys. Rev. B 61, 11 021 共2000兲. 39 P. Yang, M. Lu, D. Xu, D. Yuan, and G. Zhou, J. Lumin. 93, 101 共2001兲. 40 U. Winkler, D. Eich, Z. H. Chen, R. Fink, S. K. Kulkarni, and E. Umbach, Chem. Phys. Lett. 306, 95 共1999兲.

Downloaded 07 Oct 2004 to 132.187.77.77. Redistribution subject to AIP license or copyright, see http://jcp.aip.org/jcp/copyright.jsp