Fabrication of A Barium Titanate-Strontium Titanate Strained Layer Superlattice
- PDF / 1,650,520 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 103 Downloads / 295 Views
EXPERIMENT lxl cm 2 SrTiO3 wafers were diced into quartiles and cleaned to remove carbaceous grease and heavy metal impurities. The samples were monitored by energy dispersive spectroscopy (EDS) before and after cleaning. The cleaning treatment involved immersion in boiling .IM HNO3 acid as well as ultrasonic immersion in acetone and methanol baths. It is observed through EDS experiments that the as-received samples contain iron impurities on the surface and that they are effectively removed with the acid treatment. The clean surfaces were then investigated with atomic force microscopy (AFM) in the tapping mode to determine the morphology of the surface. Pulsed laser deposition (PLD) was then used to ablate a strontium titanate target. Two separate regimes of process conditions were investigated for effect on surface roughness: 1) the standard perovskite 2 parameters of 650 'C substrate temperature, 200 mTorr partial pressure of oxygen, and 2 J/cm 2 incident laser fluence, and 2) soft effusive conditions of 550 'C, 6 mTorr, and 500 mJ/cm . Multilayer structures composed of a YBCO electrode layer, and 20 periods of alternating STO/BTO layers were then fabricated in-situ using multitarget ablation. These films were carefully analyzed for microstructural variations using HRXRD and RBS. RESULTS AND DISCUSSION Fig. 1 is the resulting micrograph of the bare STO surface showing a peak-to-valley roughness of 2.349 nm. Atomic ledge lines are observed with contrast variations equating to multiples of the lattice parameter for STO (3.905 A). These atomic steps arise due to miscut of the 211 Mat. Res. Soc. Symp. Proc. Vol. 397 01996 Materials Research Society
wafer which is approximated to be 1 ' based on the average step height and step separation. To corroborate this data, an optimization scan was performed for high-resolution x-ray diffraction
(HRXRD). In this procedure, the co-value will vary sinusoidally with sample rotation, the amplitude of which is the sample miscut. Results from the optimization indicate a 1V miscut which is in good agreement to the results found in the initial AFM scan.
2.50-
-2.50,
, 0
0.25
0.50
0.75
1.00
1.25
Rm
Fig. 1 AFM image of cleaned STO substrate showing a mean roughness of .227 nm. Due to a 10 miscut of the wafer, atomic ledge lines are visible across the surface. A line scan is provided to demonstrate that the lines represent jumps in thickness equal to multiples of the lattice parameter for SrTiO3. The analyzed clean substrate surface was then subjected to sequential low energy fluence pulses from a bulk STO target at the process conditions listed above. Fig 2a-c shows the threedimensional scans of the resulting surface after 1, 5, and 15 pulses were applied. A gradual decrease in mean roughness is observed for increasing number of pulses. After 15 pulses, the surface remains mirrorlike with no detectable asparities. This demonstrates the Stranski-Krastanov layer by layer growth mechanism for STO films. Layer growth is crucial to the fabrication of ultrathin layers, and it is
Data Loading...
