Sputter XPS measurement of the region of a 5 nm praseodymium oxide on Si(111) sample after PDA at in 1 bar oxygen. After 3 min of sputtering (top line), an intensity peak is visible at 967 eV. This peak, which is characteristic of stoichiometric , disappears during the sputter process, although there is still praseodymium on the sample, as verified by the residual and peaks. This effect is related to preferential sputtering of oxygen and thus reduction in the oxide.
Sputter XPS measurement of the region of a 5 nm praseodymium oxide on Si(111) sample after PDA at in 1 bar oxygen. The top line was measured after 3 min of sputtering to clean the surface. After 33 min of sputtering, silicate and Si are detected. After further sputtering, the silicate signal disappears and the signal of bulk Si remains.
Sputter XPS measurement of the region of a 5 nm praseodymium oxide on Si(111) sample after PDA at in 1 bar oxygen. The double peak at the beginning of the sputter process is due to . The relative intensities of the two peaks shift due to preferential sputtering of oxygen. The single peak after a sputter time of is related to silicate.
Top: experimental XRR data (dotted line) and simulated fitting curves (solid line) for three samples postdeposition annealed at different temperatures. Experimental and simulated data agree excellently. Bottom: thickness of the Pr-rich interface for different PDA temperatures. No Pr-rich layer is detected for temperatures below . At higher temperatures, a Pr–O–Si layer is detected. Its thickness depends strongly on the PDA temperature.
Two-dimensional view of the reciprocal space of different bulk praseodymium oxide species and bulk Si, all with (111) orientation. The broad yellow lines describe the scans shown in the following pictures in , , and orientations. The narrow black lines indicate Si CTRs. Kinematically forbidden peaks are ignored. The exception to this rule is Si, where the kinematically forbidden peak is indicated by a smaller symbol than the nonforbidden Si peaks because the intensity of this peak is not completely erased due to the nonsymmetric electron distribution of the Si atoms.
scans of several 5 nm praseodymia films on Si(111) samples after PDA in 1 bar oxygen for 30 min at different temperatures from 100 to . Sharp Si peaks and broad underlying praseodymia peaks are visible at , , and for . For lower annealing temperatures the typical diffraction pattern of is observed. Therefore, the phase transition from hexagonal to cubic praseodymium oxide happens between 200 and .
CTR scans of several praseodymia films on Si(111) postdeposition annealed from 100 to . All peaks and fringes are due to the praseodymia film. No -type cubic praseodymium oxide, which would fulfill Bragg conditions at , is detected on the samples. The oxides on the 100 and samples fulfill Bragg conditions at and at and are thus hexagonal. The remaining samples have -type cubic structure, as the oxides on them fulfill Bragg conditions at and at . The discussed Bragg peak positions are also presented graphically in Fig. 5.
In-plane scan of the sample annealed at . Praseodymia CTRs are detected at and , the Bragg peak with a broad underlying praseodymia Bragg peak at . Bragg peaks, which are related to are not observed (cf. Fig. 5). The inset shows the region around for three different samples. The lattice constants of these samples in -direction are extremely close to each other.
Lateral lattice constant determined for the praseodymia films from the data around in Fig. 8. The lattice constant matches almost exactly the lattice constant of bulk , although CTR measurements verified that no hexagonal structure is on the sample. The layers are therefore strained, and their horizontal lattice constants seem to be pinned to the value of bulk , which was originally grown on the sample and only later transformed to films with cubic structure.
Top: close-up of three CTRs from Fig. 6. Data (black points) was fitted by a function (black line), which consists of two subfunctions originating from the two coexisting (green line) and (red line) praseodymia species. Both functions are calculated by kinematic diffraction theory. The peaks are located at higher scattering vectors and show weaker fringes compared to the peaks. This corresponds to a smaller vertical layer distance and higher roughness. The sharp and kinematically forbidden peak was removed for clearness. Bottom: vertical layer distances determined for the praseodymia films from the data around in Fig. 6. After the phase transition between 200 and , the peak can only be fitted using two coexisting praseodymia species as shown above. The points marked with a triangle have stoichiometry, those with a square and those with a circle .
Schematic drawing of the phase transition from to . The stacking of the Pr and O crystal planes is signed by capital and low-case letters, respectively. Red and black arrows indicate the lateral movement of Pr and O ions, respectively. Blue arrows (labeled by ) show additional O crystal planes due to the oxidation of the film. For the structural phase transition only small lateral shifts are necessary to change the stacking from of to the stacking of -oriented , while the stacking of the first layer at the interface is conserved. This explains the pinned lateral lattice constant we observed in our experiments.
Article metrics loading...
Full text loading...