surements. Finally, the grain sizes of the Bi 4 Ti 3 O 12 crystals were estimated from the AFM micrographs in Figure 4c, d to be about 0.4 and 1.0 mm for films deposited at 600°C and 700°C, respectively.Iodine-free and phase pure Bi 4 Ti 3 O 12 films with a pronounced c-axis orientation were deposited at 600°C and above. An increment of the deposition temperature increased the c-axis orientation and gave films with much smoother surfaces. Films deposited at 700°C were found to grow epitaxially on the MgO (001) substrate with the main in-plane orientational relationship [100] MgO // [110] Bi 4 Ti 3 O 12 .
ExperimentalThin films of Bi 4 Ti 3 O 12 were deposited on single-crystalline, MgO(001) substrates in a horizontal low pressure hot-wall CVD reactor. The CVD reactor is made of fused silica and consists of an outer reactor tube for both the evaporation of the metal iodides and the depositions. In order to obtain the same experimental conditions in each experiment the evaporation zones and the deposition zone are protected by exchangeable inner tubes which were cleaned in aqua regia between each experiment. Oxygen was introduced into the reactor in a separate tube to avoid oxidation of the metal iodides prior to the deposition zone. The CVD reactor is equipped with an eight-zone furnace, where each zone can be independently monitored to obtain a suitable temperature profile. The leak rate of air into the system (5 × 10 -6 Pa m 3 s -1 ) corresponds to an air contamination level of less than 5 ppm.The source materials used were titanium iodide, TiI 4 (ALFA 99.99 %), bismuth iodide, BiI 3 (ESPI, Agoura Hills, California, purity 99.999 %), oxygen (99.998 %) and argon (99.9999 %) which was used as carrier gas for the metal iodides. All gas flows were monitored by mass flow controllers. Single-crystalline magnesium oxide, MgO(001) were used as substrates. Prior to the deposition experiments, the substrates were degreased with ethanol. The amounts of TiI 4 and BiI 3 admitted into the reactor were controlled by the evaporation temperatures and checked by measuring the weight loss. The experimental parameters were the following. The total pressure was 10 Torr and the evaporation temperatures of titanium iodide and bismuth iodide were 95-120°C and 175-225°C, respectively. The deposition temperature was varied between 600 and 700°C. The oxygen and argon gas flows were kept constant at 400 sccm for Ar and 100 sccm for O 2 , which gives a linear gas flow velocity of 50 cm s -1 at the deposition zone. These flow rates were previously found to be suitable for deposition of bismuth oxides from bismuth iodide [18,19] and was retained in the present study. The deposition time was kept to 2 hours.For phase identification and determination of epitaxial relationships a Siemens D 5000 diffractometer with Cu Ka radiation was used. The geometries used were y-2y scans for phase identification, rocking curve measurements to obtain full width at half maximum (FWHM) values and j-scan measurements to evaluate the in-plane orientational relati...