Monitoring the process of formation of ZnO from ZnO2 using in situ combined XRD/XAS technique

Use of in situ combined x-ray diffraction and x-ray absorption spectroscopy for the study of the thermal decomposition of zinc peroxide to zinc oxide is reported here. Comparison of data extracted from both x-ray diffraction (XRD) and x-ray absorption spectroscopy (XAS) with thermo gravimetric analysis (TGA) enabled us to follow the nature of the conversion of ZnO2 to ZnO. A temperature range between 230 °C and 350 °C appears to show that a very poorly crystalline ZnO is formed prior to the formation of an ordered ZnO material. Both the decrease in white line intensity in the Zn K-edge XANES and resulting lower coordination numbers estimated from analysis of the Zn K-edge data of ZnO heated at 500 °C, in comparison to bulk ZnO, suggest that the ZnO produced by this method has significant defects in the system.

A c c e p t e d M a n u s c r i p t Introduction Zinc oxide is a multi-functional material that has found a plethora of applications, due to its electronic and structural properties. As an n-type semiconductor, similar to other metal oxides the structure, morphology and size of the particles have a large impact on their properties and uses [1,2] . ZnO is becoming the material of choice for a range of applications which includes transparent conducting oxides [3][4][5] , solar cells [ [6][7][8]] , photocatalytic applications, as a sensing material [9] and for antimicrobial applications [1,10] . The defects present in zinc oxide have been proposed to be a key factor for its performance [3,11,12] .
Nano sized ZnO materials have been synthesised in a variety of ways including hydrothermal [13] , mechanochemical [14] , spray pyrolysis [15] , chemical bath techniques etc [11,[16][17][18][19] . In addition to the above, ZnO can be conveniently prepared using a two-stage approach by first making zinc peroxide (ZnO2) and subsequently decomposing this in a controlled way to produce zinc oxide, the method of our choice reported in this work. It has been reported that ZnO prepared via decomposition of ZnO2 has defects compared to other methods and it is reported that it is possible to control the types of defects in ZnO which may enhance functionality [12,[20][21][22] . ZnO2 on its own found prominent use in the rubber industry for promoting cross-linking in carboxylated nitrile rubber and other elastomers and also as an antiseptic additive [23,24] . The most common method of synthesis of ZnO2 is by adding a soluble zinc salt to hydrogen peroxide, while it has also been prepared by hydrothermal or organometallic routes [9,10]. Therefore, it is of considerable interest to follow the decomposition pathway of ZnO2 producing ZnO at various temperatures. To follow the decomposition process, it is necessary to use techniques that enable the determination of both long and short-range order. This is required as whilst both the starting ZnO2 and final ZnO product are crystalline solids, the decomposition process may not have long-range order and needs to be studied via the use of methods that are suited for the determination of shortrange order present in the system. To this end, we have used combined X-ray diffraction (XRD) and X-ray absorption spectroscopy (XAS) at the Zn K-edge to monitor the process of converting crystalline ZnO2 to ZnO during a thermal decomposition process. XRD provides information regarding long-range order (crystallinity) whilst XAS (an element specific  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 3 technique that does not depend on long-range order in a system) yields the required shortrange order information around a given metal ion of interest. In addition, it is well-known that XRD can be used to determine crystallite size and monitor the formation of amorphous/poorly crystalline phases. Similarly, XANES has been noted to be sensitive to local coordination environment [18] as well as defects present in systems [25] . An added advantage of using the combined XRD/XAS technique is that it is carried out on the same sample under identical environment and utilises time-resolved methods in monitoring the changes that take place rapidly during thermal treatment process [26][27][28][29][30][31][32][33][34] . This combined approach, first established in early 1990's [26,27,29,30] , has become a "work horse" in many Synchrotron Radiation facilities; a good time-resolution in the measurement of both XRD and XAS facilitates the process of monitoring the structural changes through temperature or any reaction coordinate change. Here we report the structural changes and growth process of ZnO during the thermal decomposition of ZnO2, in air employing the in situ, time-resolved, combined XRD/XAS technique.

Experimental
Zinc peroxide was prepared via a wet chemical method. In a typical experiment, 2.0g of zinc acetate dihydrate was dissolved in 25mL water to which 25mL of 30% H2O2 aqueous solution was added and stirred. The pH of this solution was adjusted to 10 using dilute NH4OH and stirred for 12 hours at room temperature. The resultant precipitate was isolated by centrifuging, washed several times with water and dried. X-ray diffraction studies conducted using a laboratory based, Bruker AXS D8 Advance, X-ray diffractometer operated in parallel beam mode showed phase pure ZnO2 material was produced using this method (see Supplementary information (SI) Figure S1).
ICP analysis was carried out on as-synthesised ZnO2, commercial ZnO2 and ZnO2 samples heated to 300 o C. Assay results confirmed the amount of zinc to be 64.4%, 72.0% and 83.2%, respectively. Calculated Zn values are 67.1% for pure ZnO2 and 80.4% for pure ZnO. As commercial ZnO2 contains ~50% ZnO, the zinc concentration is expected to be between that of pure ZnO2 and ZnO. The obtained value for as-synthesised ZnO2 is lower than expected suggesting the presence of impurities. TG/DTA curves for ZnO2 obtained in air ambient at a heating rate of 2 o C/min are given in Figure 2. The initial weight drop of 2.8% in the temperature range of 35 -150 o C is possibly due to the loss of adsorbed moisture. A further This pellet was loaded into a high temperature custom built in situ cell (see photograph given in Supplementary Information, Figure S2) for measuring combined XRD and XAS data.
Measurements were conducted during the heat treatment of the sample in a flow of air at 10ml/min, to 500 o C at a rate of 10 o C/min and after subsequent cooling to ca 50 o C.
XAS data were processed using suite of ATHENA and ARTEMIS software [35] . For Linear Combination Fitting (LCF) analysis, an option available in the ATHENA software was utilised [35] .
Both starting material ZnO2 and final phase, ZnO (characterised by XRD) were used as standards for LCF analysis, and the fitting was performed over an energy range of -20 to 80 eV with respect to the Zn K-edge absorption edge of 9661 eV. For determining the local structural parameters from EXAFS data, a k-range of 2.4 to 10.44 Å -1 was used. Fitting was performed in R-space (R-range of 0.8 to 4 Å) using ARTEMIS software [35] . The crystal structures of ZnO2 and ZnO were used as the starting models and all the relevant paths below 4Å were considered for the fitting methods. The amplitude reduction factor, So2 was determined to be 0.87 by fixing the Coordination number (CN) to that of crystal structure data. Inter-atomic distance (R) and Debye-Waller factor ( 2 ) were determined by refining these parameters along with the Eo value. Subsequently the So2 value was fixed to determine change in coordination number. The values of first Zn-O and second Zn-Zn shells were extracted to determine the local structural change during the thermal decomposition process.  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 5

Results and Discussion
The decomposition of ZnO2 was investigated through monitoring the long-range and local structural changes during the heat treatment of ZnO2, employing in situ combined XRD/XAS techniques in combination with thermo gravimetric analysis (TGA) methods. In Figure 1 we show a stacked XRD plot recorded through the decomposition of ZnO2. It is seen that the reflections related to ZnO2 decrease beyond 200 o C and above ca 300 o C reflections belonging to ZnO begin to appear. A TGA plot is shown in Figure 2, wherein it is clear that the weight loss starts to occur gradually just below 200 o C. A sharp decrease of 15% loss is seen above 200 o C and subsequently a slow weight loss takes place before it stabilises. To cross correlate these weight losses with changes in the XRD data, we analysed the intense reflections of ZnO2 that did not have overlap with ZnO reflections to obtain the variation in the intensity of the XRD reflections and plotted these along with the TGA in Figure 2(a) (ZnO2 (022) reflection of the XRD data). Although a slight delay in the decrease of the ZnO2 reflection intensity is noted here, the exact temperature ramp used for the TGA was different to that used for the combined XAS/XRD experiment. Also, it should also be noted that the XRD measurements were conducted in sequence with respect to the XAS. Furthermore, we anticipate a small temperature difference of ca 2 o C between the point at which the sample  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 6 is irradiated with the X-ray beam and the position of the thermocouple placed in the cell (thermocouple is placed ca 7 mm away from the beam in order to avoid any interference of the thermocouple with the X-ray beam). However, the trend appears to be closely similar to the weight loss seen in TGA, which we can be attributed to the decomposition of ZnO2 and formation of ZnO taking place over the temperature range of 175 to 250 o C. In Figure 2(b) we compare the intensity variation of the ZnO2 (022) reflection and the ZnO (002) and (110) reflections, (occurring at the 2θ values of 28.99 and 47.03 degrees, respectively), that did not overlap with ZnO2 reflections. It is apparent that the intensities in this 2 range start to appear in the XRD slightly above 200 o C whilst the ZnO2 reflection intensity begins to decrease  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 8 increases to a value close to 30 to 50 nm. The value of crystallite size for the (002) reflection seems higher than the (110) reflection, suggesting a rod type morphology growth taking place in the system which is consistent with hexagonal ZnO.

Zn K-edge XAS
In order to obtain the changes in the short-range order of the system, we analysed the Zn Kedge EXAFS data. To perform XAS data analysis, it is necessary to use related materials as standard and also a starting model for extracting local structural information. Here we used crystalline ZnO2 and ZnO as our reference ZnO2 is a crystalline material that crystallises in cubic form (space group Pa-3) with cell parameter of a = 4.871Å [36] . ZnO crystallises in hexagonal form (P63mc) with lattice parameters a = 3.2427 c = 5.1948Å [37] The local structure of these systems based on the respective crystal structures are given in Table 1. It  Figure 4 shows a stacked plot of the Zn K-  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 9 edge XANES data. Figure 4 (b) shows that the normalised white line intensity (main absorption peak in XANES) is higher for ZnO2 compared to ZnO, which is related to the local coordination environment. Comparison of the Zn K-edge XANES of ZnO obtained from heating ZnO2, clearly shows a reduction in the white line intensity compared to bulk ZnO (see Figure 4(b)). It has been reported that the loss in white line intensity can be attributed to vacancies present in the system [25] . The overall features in the XANES of ZnO2 are different compared to ZnO and it is therefore possible to use these two spectra as references to estimate the fractions of ZnO2 and ZnO phases in the material from the in situ data obtained during the thermal treatment process. In Figure 5(a) we show the combination of fractions of spectral features of ZnO2 and ZnO that matches with the experimental data, recorded during the thermal treatment process. The values obtained, using linear combination fitting analysis of the two reference spectra, plotted against temperature and along with TGA is shown in Figure 5 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 10 combination of these two phases appears to represent these states very well at every stage of the thermal process.
Further comparison of the fractions of ZnO2 and ZnO estimated from LCF analysis of Zn Kedge XANES data were made with the intensities of the XRD reflections representing ZnO2 and ZnO phases (see Figure 5  This demonstrates that as XANES is sensitive to short-range order we start to see changes in the decomposition process at an earlier stage, whereas XRD is unable to follow these changes as it requires the presence of a crystalline phase with a certain degree of long-range order. We also analysed the EXAFS part of the Zn K-edge XAS data in detail from all the recorded spectrum during the thermal treatment process. A stacked plot of the Fourier Transform (FT) of the Zn K-edge EXAFS data are shown Figure 6(a). There are some clear differences between the data from the beginning of the experiment and above 200°˚C. In Figure 6(b) we highlight  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 11 the changes in the FT after 230 o C where it is clear that the second neighbour peak around 3Å starts to grow. We analysed these data sets in detail and representative best fits between experimental and calculated EXAFS after refining parameters in particular, coordination number (CN), interatomic distances and Debye-Waller factors (Eo was also refined), are given Figure 7. It is clear from the fits that all the short-range order distances are similar to the phases of ZnO2 and ZnO (given in Table 1) and consistent with the XANES observation for octahedral and tetrahedral coordination environment within the first neighbour distance. In Figure 8 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 12 Figure 7: Selected best fit between experimental and computed EXAFS using parameters listed in Table 1 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 13 Table 2: Selected structural parameters obtained from best fit between experimental and calculated EXAFS data. N is the average coordination number for the given atom-pair at a distance of R. σ 2 is the Debye-Waller factor associated with the atom-pair. The estimated S2 based on ZnO reference compound is ca 0.87

System
Atom-Pair N R/Å σ 2 / Å 2 # R-Factor This approach indeed assumes that the static disorder contribution in the actual converted samples is similar to bulk ZnO, although the static disorder component may be slightly higher.   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60 A c c e p t e d M a n u s c r i p t 15 coordination whereas the error for 2 nd neighbour appears to be larger, in the range of 20 to 30%. and second neighbours were found to be below the expected values compared to a bulk highly crystalline solid. Whilst the CN values of Zn-O is around 3.5, the values for the second neighbour in the high temperature data above 250 o C appears to be much smaller compared to the bulk ZnO and slowly increased to ca 8 for the second neighbour Zn-Zn coordination.
This could be due to defects present in the system which becomes less when heated at temperatures above 400 o C. In order to confirm whether this reduction is due temperature, the thermally treated sample measured after cooling to ca 50 o C was also analysed and the results are shown in Table 2 and the best fit for this data are shown in Figure 7(d). Indeed, the coordination number for both Zn-O and Zn-Zn remain at a lower value compared to the bulk while the DW factor decreased, as one would expect for the sample cooled close to 30 o C.
However, nano crystalline oxides also show lack of higher neighbour contributions due to size effects [11,38] It is difficult to determine or suggest whether such a decrease in CN values of higher neighbours is related to the presence of defects or nano crystalline form or both, as it was shown that additional methods are required to differentiate this aspect. [38] Combining the observations that the decrease in the white-line intensity seen in the XANES (Figure 4(b)) and the decreased average CN values of both first and second neighbour coordination environments suggests that the ZnO formed by thermal decomposition produces a defective material and possibly in addition to being a nano system consistent with earlier studies. [13,25,39] In summary, we used a rapid EXAFS and diffraction measurement technique which provided high quality EXAFS data to determine the thermal decomposition of ZnO2 material to ZnO.