PLASMA HYDROGENATION OF HYDROTHERMALLY GROWN ZNO MICROPODS

The hydrothermally grown ZnO micropods have been placed on grounded stainless-steel holder and exposed to an inductively coupled plasma (ICP) monitored in-situ by optical emission spectroscopy (OES). OES shows the immediate release of oxygen during Ar ion bombardment. The prolonged exposure to hydrogen plasma leads to deterioration of the optical properties as well. The exposure, rf power and hydrogen pressure have been optimized to enhance UV-photoluminescence peak at the wavelength 384 nm related to surface bounded excitons and reduce the defect-related photoluminescence in red spectral range. The strong UV photoluminescence appears just after 1 minute of plasma hydrogenation in a radio frequency plasma discharge with power density 40 W/dm 3 and hydrogen pressure 17 Pa.


INTRODUCTION
The zinc oxide (ZnO) crystallite powder is a low cost, environmentally friendly, wide band gap semiconducting material with unique optical properties and variety of nanoparticle (NPs) morphologies [1]. Due to the high surface-to-volume ratio and related size effects, ZnO NPs are used in many applications for energy conversion, photocatalytic wastewater treatment, electrochemical energy storage or sensing. Recently, we have applied a hydrothermal chemical synthesis as a low-cost technology for mass-production of ZnO hedgehog-like crystallite powder [2], studied their crystal quality [3] and shown that the age of precursors plays a significant role in defect formation [4]. The photoluminescence spectra (PLS) show a strong near UV emission peak related to bound exciton complexes assigned to surface states [5]. The strong correlation between surface defects and the exciton emission has been also observed in hydrothermally synthetized ZnO micropods [6].
Particularly important are electrically conductive ZnO NPs with the interstitial hydrogen acting as a donor [7]. Two-step annealing in air at around 1000 C followed by annealing in hydrogen atmosphere at 800 C has been demonstrated to significantly enhance UV-photoluminescence of ZnO NPs for highly efficient subnanosecond fast scintillators [8,9]. The inductively coupled plasma (ICP) offers an alternative way of cold hydrogenation [10]. We have already shown that the cold hydrogenation in low pressure hydrogen plasma is an effective way to significantly enhance the electrical conductivity and exciton-related emission in ZnO thin films prepared by reactive magnetron sputtering on fused silica glass substrates [11]. In this paper we optimize the plasma hydrogenation of ZnO micropods placed on grounded stainless-steel sample holder.

Hydrothermal growth
ZnO micropods were the hydrothermally synthesized in aqueous solution containing equimolar (25 mM) amount of zinc nitrate hexahydrate p.a. (Zn(NO3)2·6H2O) and hexamethylenetetramine (HMTA) (C6H12N4) p.a. Both chemicals were purchased from Slavus and used without further purification. Deionized water was purified with a system So~Safe Water Technologies, having a conductivity 0.20 µS·cm -1 (25 °C). The reaction mixture was heated at 90 °C for 3 h. The precursor salt residues were removed from the sample by washing 5-times with deionized water followed by centrifugation at 10 000 rpm (RCF: 11 510g) for 15 min. Finally, the ZnO powder was dried by lyophilization. Scanning electron microscopy (SEM) images show ZnO micropods with hexagonal shape, the length up to 10 µm and the diameter about 0.5 µm, see Figure 1.

Plasma treatment
The plasma hydrogenation has been done in a prototype inductively coupled plasma (ICP) reactor currently being developed in a cooperation with the Czech company SVCS Process Innovation, s. r. o. (Valašské Meziříčí, Czech Republic). The quartz reactor of the active volume 5 liters operates at radio frequency (rf) 13.56 MHz with maximum discharge power up to 300 W corresponding to the power density 60 W/dm 3 . The setup is evacuated with the Agilent Triscroll 300 dry vacuum pump protected with the ISO Flange Vacuum Filter (CSL-843-NW25). A 5 mg powder sample mechanically pressed into a round pellet with the diameter 3 mm was placed in the middle of the chamber on a grounded stainless-steel holder. The holder grounding is a recent improvement over the previously used ungrounded quartz holder. Prior the plasma treatment, the chamber, valves, flowmeters, gauges and all gas inlet and outlet tubes were evacuated to residual pressure below 0.1 Pa and flushed 5 min by 20 sccm flow of process gases to reduce residual gas contamination. The pellets were exposed for 30 s to Ar plasma (Ar purity 99.998%, rf power 100 W, Ar flow 5 sccm, pressure

Characterization
Optical emission spectra (OES) were measured in 400-1000 nm spectral range with 1 nm spectral resolution by spectrally calibrated fiber coupled CCD spectrometer (B&W Tek BTC112E) focused in the middle of the quartz chamber. Dark spectra were subtracted. The integration time varied from 5 ms for Ar plasma to 200 ms for H2 plasma. Each spectrum was averaged 10 times.
Raman spectra were recorded in backscatter geometry using a micro-spectroscopic Renishaw InVia Reflex Raman setup with integrated optical microscope and a charge-coupled device (CCD) camera. For excitation, a 442-nm-line of HeCd excitation laser. Other parameters were 100× Olympus objective, 65 µm slits, and a grating of 2400 grooves/mm.
The attenuated total reflectance (ATR) infrared absorption spectra were detected in 590-5000 cm -1 spectral range and 4 cm -1 resolution using Nicolet™ iS50 FTIR Spectrometer (Thermo Fisher Scientific) with build-in diamond ATR prism, KBr beam splitter and cooled DTGS detector. The ATR spectra were first normalized to zero at 5000 cm -1 and then recalculated with the build-in advanced ATR correction for diamond prism.
PLS were measured at room temperature in the 360-780 nm spectral range with 2 nm spectral resolution using the 100 mW/cm 2 UV photo-excitation provided by a focused 340 nm LED equipped with a narrow band-pass filter and operating in square pulse regime at frequency 307 Hz. The emitted and scattered light was collected and focused onto the 1 mm wide input slit of double gratings SPEX 1672 monochromator equipped with Peltier cooled multi-dynode multi-alkali red sensitive photomultiplier (PMT) to be measured at each wavelength independently by a lock-in amplifier synchronized with excitation frequency. The spectra above 600 nm were measured with a long pass filter placed in front of the monochromator. The whole setup was spectrally calibrated with Oriel #63358 Quartz Tungsten halogen lamp and converted from wavelength to energy scale for quantitative analysis [12].

RESULTS AND DISCUSSION
The OES spectra clearly show the presence of atomic oxygen in Ar plasma as demonstrated by the presence of oxygen-related emission peaks at 777 and 845 nm [13], see Figure 3. Since no nitrogen related peaks are observed, oxygen cannot originate from ambient atmosphere leak, but from ZnO powder. In pure Ar plasma the intensity of O-peaks is low and may be explained by contamination from previously processed ZnO. The intensity of O-peaks increases by an order of magnitude when a pellet is placed into the chamber while the plasma visually change color from brownish to white. Thus, OES proves the heavy ion bombardment of the sample placed on grounded holder. Since only the Balmer series lines were observed in hydrogen plasma at 656, 486 and 434 nm and no O-peaks, we conclude that ZnO was scattered by heavy Ar ions but not by light H ions.

Figure 4 (A)
shows Raman spectra of plasma hydrogenated ZnO micropods,where origin of peaks is also depicted. Nonpolar phonon mode E2(Low) expected at 100 cm -1 is associated with Zn sublattice, but it was out of the measured spectral range. Nonpolar phonon mode E2(High) associated with oxygen sublattice is observed at 440 cm -1 [14]. The sharp, strong and dominant Raman peak is significantly amplified after hydrogen plasma, but for 10-minutes hydrogenation and longer its intensity decreases again. Same as increasing intensity of the dominant peak, broad sideband at low-frequency side also increases. This broad band with maximum about 420 cm -1 is attributed to E1(TO) mode commonly seen in nano-and microstructures. The peaks at 336 and 540 cm -1 correspond to the sum and differential wavenumbers E2(High) ±E2(Low) and A1(LO), respectively [15]. The peak at 581 cm -1 has been assigned to the ZnO mode E1(LO) [16]. This mode should be inactive but it may appear as a local mode due to the disturbance of the crystal symmetry resulting from the presence of oxygen vacancies and zinc interstitials. Thus, the presence of this mode suggests that the plasma treatment increases the concentration of point defects. Moreover, from spectra normalized on the value of E2(High) peak, see Figure 4(B), could be seen that for hydrogenation time up to 5 minutes, the spectra are very similar. After longer hydrogenation, the height of the E1(LO) peak decreases. In spectra, small band near 660 cm -1 and 1150 cm -1 correspond to multiphonon processes and peak near 1100 cm -1 to combination of acoustic modes with symmetry A1 and E2.
ATR-FTIR is a sensitive probe of free carriers in ZnO. It has been shown that the room-temperature exposure of polycrystalline ZnO films to H2 plasma rapidly increases carrier density [17]. Figure 5 shows the infrared absorbance spectra of plasma hydrogenated ZnO micropods depending on the expose to hydrogen plasma. Free carrier absorption dominates infrared absorbance spectra already in the unhydrogenated sample with further increase after plasma hydrogenation. The room temperature plasma hydrogenation had been sufficient to induce significant surface conductivity changes after 1 minute expose. It should be noted that the absorbance has been somewhat arbitrary normalized to zero at 5000 cm -1 and that the signal-to-noise ratio is poor in the spectral range 2000-2300 cm -1 where diamond ATR prism heavily absorbs.

Figure 4
Raman spectra (A) and normalized Raman spectra (B) of plasma hydrogenated ZnO micropods depending on the expose to hydrogen plasma. The excitation wavelength was 442 nm (blue laser)

Figure 5
Infrared absorbance spectra of plasma hydrogenated ZnO micropods depending on the expose to hydrogen plasma measured by ATR FTIR and normalized to zero at 5000 cm -1 .
The PLS in Figure 6 shows that the strong exciton photoluminescence at 3.23 eV (384 nm) appears already after 1 min plasma hydrogenation at 200 W. This suggest that the bound exciton complexes are assigned to surface states. Furthermore, the hydrogen plasma suppresses the defect related broad PLS band centered in red spectra range around 2 eV. It takes about 10 min to suppress this band by an order of magnitude. Therefore, we suggest 5 min plasma hydrogenation as an optimal process time. This is significantly shorter time then 30 min previously used in plasma hydrogenation with ungrounded quartz holder.

Figure 6
The room temperature PLS of plasma hydrogenated ZnO micropods in semilogarithmic intensity scale depending on the expose to hydrogen plasma. The excitation energy was 3.65 eV.

CONCLUSION
Ar plasma sputters ZnO placed on a grounded stainless-steel holder in an ICP reactor. Therefore, the Ar cleaning process needs to be taken only for a short time with a relatively low rf power. The strong UV photoluminescence appears just after a minute of plasma hydrogenation of ZnO micropods which is a strong evidence that the bound exciton complexes are assigned to surface states. On the other hand, the reduction of red photoluminescence needs prolonged time suggesting that the red photoluminescence is related to bulk defect states. The optimal time for plasma hydrogenation at the rf power density 40 W/dm 3 and hydrogen pressure 17 Pa seems to be just a few minutes.