Tailoring the Dopant Distribution in ZnO:Mn Nanocrystals

The synthesis of semiconductor nanocrystals with controlled doping is highly challenging, as often a significant part of the doping ions are found segregated at nanocrystals surface, even forming secondary phases, rather than incorporated in the core. We have investigated the dopant distribution dynamics under slight changes in the preparation procedure of nanocrystalline ZnO doped with manganese in low concentration by electron paramagnetic resonance spectroscopy, paying attention to the formation of transient secondary phases and their transformation into doped ZnO. The acidification of the starting solution in the co-precipitation synthesis from nitrate precursors lead to the decrease of the Mn2+ ions concentration in the core of the ZnO nanocrystals and their accumulation in minority phases, until ~79% of the Mn2+ ions were localized in a thin disordered shell of zinc hydroxynitrate (ZHN). A lower synthesis temperature resulted in polycrystalline Mn-doped ZHN. Under isochronal annealing up to 250 °C the bulk ZHN and the minority phases from the ZnO samples decomposed into ZnO. The Mn2+ ions distribution in the annealed nanocrystals was significantly altered, varying from a uniform volume distribution to a preferential localization in the outer layers of the nanocrystals. Our results provide a synthesis strategy for tailoring the dopant distribution in ZnO nanocrystals for applications ranging from surface based to ones involving core properties.

. XRD pattern of the ZHNMn sample annealed at 200 o C, showing the transformation into ZnO as main crystallographic phase as well as very low intensity peaks from the ZnOHNO 3 ·H 2 O minority phase (stars). The very broad XRD line in the 20 -30 degrees range corresponds to the glass support. Table S1. Samples selected for in-depth spectroscopic and microstructural investigations: synthesis temperature and amount of HNO3 1M added to the starting solution; composition and crystallite size resulted from XRD at the end of the co-precipitation synthesis; maximum annealing temperature T max ; composition and crystallite size after annealing at T max .

Sample
Synthesis temperature The amplitudes of the calculated spectra of the different centres were adjusted for an optimum fit of the experimental spectrum. The simulated spectra of each sample were obtained by summation of the calculated spectra of all centres present in the sample. The intensity of the EPR spectrum of a centre, calculated by double integration of the spectrum, is proportional to the amount of that centre in the measured sample. The relative concentration of a Mn 2+ centre in a sample was determined as the ratio between the intensity of the corresponding calculated spectrum and the intensity of the simulated spectrum of the sample.
The EPR spectra of the Mn 2+ paramagnetic centres were analysed using the following spin Hamiltonian (SH): 1 The first term represents the main electron Zeeman interaction of the S = 5/2 electron spin with the external magnetic field B. The second term represents the hyperfine interaction of the electron spin with the I = 5/2 nuclear spin of the 55 Mn (100% abundance) isotope, responsible for the characteristic six-lines hyperfine structure. The next two zero-field-splitting (ZFS) terms describe the interaction of the electron spin with the local crystal field, which gives rise to a fine structure of 2S = 5 component lines, while the last term describes the nuclear Zeeman interaction. The SH parameters of the Mn 2+ centres, determined by simulation and lineshape fitting of the Q-band spectra of the investigated samples, are given in Table S2, together with reference SH parameters for the Mn 2+ ions in other host lattices of interest. For the ZFS parameters the absolute values are given, although a prediction of the parameter signs is possible based on the reference data. For example, in the case of the Mn 2+ ions in ZnO single crystals, the sign of the D parameter value was determined from low temperature measurements to be negative 2 . It is thus likely that the sign of D is negative in the case of the Mn 2+ ions in nanocrystalline and disordered ZnO as well.
As-prepared ZOM1 sample Figure S2 displays the deconvolution of the EPR spectrum of the as-prepared ZOM1 sample. The two centres associated with the Mn 2+ ions embedded in ZnO nanocrystals, i.e. Mn 2+ (c) and in the disordered phase, Mn 2+ (d), have very close EPR parameters, the main difference being the value of the broadening parameter. As initial parameters in the simulation of both centres we have used the reported EPR parameters of the Mn 2+ ions in the ZnO single crystal, and then adjusted them until the best fit of the experimental spectra was achieved.
The effect of the broadening parameter on the aspect of the EPR spectrum is quite dramatic. The spectrum of the Mn 2+ (c) centres, where the broadening effect is moderate, can be described as consisting of six lines with a complex shape, each one extending over a magnetic field range B, measured as the distance between the maximum and the minimum of the S5 complex line (see Figure S2), which is further called "powder linewidth". The powder line shape and B strongly depend on the D parameter value, allowing thus the estimation of this parameter from the lineshape fitting of the experimental spectrum. The disorder induced broadening effect leads to the decrease of B, until only six symmetric lines associated with the central allowed M S :-1/2 ↔ +1/2, m = 0 transitions can be seen, as is the case for the spectrum of the Mn 2+ (d) centres. In this case the D value was assumed to be the same as for the Mn 2+ (c) centres, while the g and A parameter values could be determined with a good accuracy. The very small differences (less than 2%) in the hyperfine parameters of these two centres can be explained by slightly different average coordination numbers of the Mn 2+ ions in the disordered phase associated with small variations in sample stoichiometry [3][4][5] .
A similar disorder effect can be observed in the spectrum of the Mn 2+ (x) centres. In this case a range of possible values was determined for the D-parameter for which the spectrum lineshape and width were reasonably well reproduced.  Table S2 are represented below. The arrows mark the powder linewidth ΔB for the Mn 2+ (c) and Mn 2+ (d) centres.  g, A, D and a), broadening parameter (D) and individual peakto-peak linewidth (B pp ) of the Mn 2+ centres in the investigated samples. Reference parameters of the Mn 2+ centres in other host materials discussed in the text are given for comparison.

Host
Centres