Carrier recombination in sonochemically synthesized ZnO powders

ZnO is a wide band gap semiconductor with a high exciton binding energy which attracts considerable scientific and applied interest with respect to exploiting its optical properties [1–5]. For example, high thermal and chemical stability, simple tunability of the optical and electrical properties are widely applicable in optoelectronic devices [6, 7]. Moreover, while exhibiting an excitonic emission peak in the ultraviolet (UV) region, ZnO typically shows strong visible luminescence bands which are due to numerous point defects present in the lattice [8–11]. Due to these facts, various methods of producing zinc oxide with different morphology and point defect content have been developed so far [12].

Quite recently, the synthesis of nanoscale materials under microwave irradiation and during sonochemical reactions became particularly attractive [13–17]. The main effect of sonochemical reaction is cavitation which is accompanied by the formation, growth and collapse of cavitation bubbles. Inside the collapsing bubble, extreme physical and chemical conditions (temperatures ⩾ 5000 K and pressures ⩾ 1800 MPa) are realized [18–20]. With the chemical synthesis, metal oxides are generally grown from their acetate precursors [21–23]. Employing sonochemical reactions, a low-vapor-pressure metal acetate is usually dissolved in solvents to form a homogenous solution. Consequently, numerous metal oxide materials were successfully synthesized in sonochemical reactions and, particularly, ZnO, CuO, Co₃O₄ and Fe₃O₄ nanocrystals were sonochemically produced from respective acetate compounds in various solutions [23].

In this work, a simple sonochemical route of producing ZnO powder materials is presented and it is also shown that in those materials, concentration of defects can be changed by varying the reaction conditions.

For sonochemical synthesis of ZnO, zinc acetate and sodium hydroxide were used as starting materials. 2-propanol was used to dissolve these materials and obtain a liquid solution. At first, 0.10 g of zinc acetate was dissolved in 25 mL of 2-propanol and the resulting solution was added to 125 mL of cooled 2-propanol. Next, 15 mL of 0.050 M sodium hydroxide in isopropanol was prepared. Finally, the sodium hydroxide solution was slowly added into 2-propanol and zinc acetate solution. The final solution was transferred into the ultrasound reactor, allowing it to achieve cavitation conditions. Details of the ultrasonic setup are given in the literature [24]. When the synthesis was finished, the resulting product was continuously washed in methanol and water and dried in ambient air. The white powder, obtained in this process, was then dissolved in 2-propanol (IPA). The grain size in the powder was found to vary in the range of a few dozens of nm to ∼130 µm. In this work, the samples of ZnO powders are marked as SC1, SC2 and SC3 respectively. Subsequently, these samples are compared according to their stirring time of 1 min, 60 min and 120 min, respectively.

Photoluminescence (PL) spectra and decays were recorded at room temperature using a monochromator MDR-4 and a photomultiplier tube FEU-79 with a computer data acquisition system. PL decays were analyzed with use of a digital oscilloscope GDS-806S with a time resolution of 1 µs. An N₂ laser with the wavelength of λ = 337.2 nm (3.67 eV) was used as the photogeneration source [25]. Surface photovoltage (SPV) decays were measured in the contactless capacitor arrangement, and details of the setup are given in the literature [26]. Fourier transform infrared (FT-IR) spectra were recorded using FT-IR Spectrometer SPECTRUM BX II (Perkin Elmer).

In order to detect the surface species and the nature of surface chemical bonds, consider FT-IR spectra of the samples which are shown in Fig. 2. They exhibit a number of strong absorption peaks relevant to IPA (dotted spectrum). A number of subsidiary bands appears in the SC1, SC2 and SC3 samples in the spectral regions marked by ellipse 1 to ellipse 3 in Fig. 2. In general, electrostatic passivation of surface cations by negatively charged ligands occurs in II-VI semiconductor nanoparticles [41] which is capable of explaining consistently the observed changes in the FT-IR spectra of ZnO particles.

As can be seen in Fig. 2, the characteristic absorption band at about 510 cm⁻¹ is detected in all samples and it is attributed to the Zn–O bonds [42]. Furthermore, the transmission gradually decreases in the region from 3000 cm⁻¹ to 3500 cm⁻¹ (region 3 in Fig. 2) which indicates the more pronounced role of the v(O–H) vibration mode in the spectra of the samples produced with longer stirring time. Indeed, the broad absorption peak centered at 3342 cm⁻¹ corresponds to the – OH group of IPA. For our SC samples, this band shifts to 3350 cm⁻¹. Moreover, a small shoulder at ∼1650 cm⁻¹ (region 2) is resolved in all SC samples, and it is assigned to the O–H stretching vibrations of water molecules or moisture adsorbed on the ZnO surface [42]. Additionally, the FT-IR spectrum in the C–H vibration range (region 3 in Fig. 2) is enlarged in the inset of Fig. 2. One can see that the peak at 2830 cm⁻¹ observed in IPA completely vanishes in our SC synthesized samples of ZnO powders. The peak at 1600 cm⁻¹ is usually attributed to either O–H stretching vibrations of hydroxyl functional groups or to the bending vibrations of the surface H–OH, which shows that hydroxyl groups are adsorbed on the surface of ZnO [43]. Finally, the most pronounced changes in FT-IR spectra are monitored in region 1 of Fig. 2. The peak at 1035 cm⁻¹ in alcohols is related to the C–O stretching vibrations. This peak is clearly detectable in pure IPA but quenches remarkably in the SC samples.

The spectral changes observed in Fig. 2 can be explained by assuming that a molecule of IPA is absorbed on the surface of ZnO and then interacts with the surface of the grain via O atoms. It is interesting that the similar conclusion can be drawn from the data describing the synthesis of gold and silver nanoparticles passivated by PVP [44–47].

Therefore, the data in Fig. 2 show that the main changes in FT-IR spectra of IPA are associated with the change in transmittance of the (O–X) bond in IPA molecules. Passivation of ZnO particles in IPA is realized by the interaction of the oxygen atom in IPA and zinc in ZnO. Taking this scenario into account, it can be expected that the concentration of oxygen vacancies in SC samples will vary with varying stirring time which, in turn, can be taken into account, when analyzing the PL spectral changes given in Fig. 1.

Based on the above analysis and previously demonstrated strong correlation between the surface defects and optoelectronic properties of ZnO materials [28, 48–50], one can expect that the SPV decays would be remarkably different in the SC samples obtained at different stirring times. This is exactly what is seen in Fig. 3. In this figure, normalized SPV decays in SC1, SC2 and SC3 samples are shown. The main observation is that the increase in the stirring time (from sample SC1 to SC3) is accompanied by accelerated initial time decays in the SPV transients shortly after excitation (for times shorter than ≈5 ms in curve 1 to curve 3, respectively). These are followed by a longer nearly exponential decay component for times longer than 5 ms. This component remains practically unchanged with increasing the stirring time.

It is well known that SPV decays depend on several factors, such as (i) lifetime of photogenerated charges with respect to their recombination pathways which, in turn, determines the PL intensity, (ii) concentration of surface defects, (iii) efficiency of the inter-particle charge transfer. It seems to be very likely that the data shown in Fig. 3 indicate that the free carries trapping dynamics is mainly determined by decreasing of the surface defect concentration, provided that the distribution of the grain sizes shifts to smaller values in the sequence of SC1, SC2 and SC3 samples, respectively. This conclusion is supported by the results presented in Section 3.2., where we showed that the concentration of surface defects decreases with increasing stirring time. It is, therefore, believed that ZnO particles fabricated at short stirring times have rather high concentration of surface defects. These defects act as traps for photogenerated charges, slowing down the SPV decay. Prolonged stirring times improve the quality of the particle surface, so that the trap concentration decreases which leads to marked acceleration of the initial SPV decay.

The latter conclusion is also supported by analyzing the PL decays shown in Fig. 4. Indeed, no difference is found in the decay curve shapes reproduced in samples SC1 and SC3 which implies that the surface states do not seem to be involved into the PL emission process.

In conclusion, the carrier recombination processes in sonochemically synthesized ZnO powder particles have been investigated using PL, FT-IR and SPV techniques. It has been shown that the PL spectra were rather sensitive function of the stirring time during the ultrasonic bath. A number of defect-related PL bands, typical of ZnO lattice, has been detected, and each of the PL band was decomposed into a sum of two partially overlapping subbands. The performed measurements indicate that the sonochemical fabrication route of producing ZnO gives convenient means of the modification of the ratio surface-to-volume defect concentrations.