65 eV for the BFO film ascribed to Bi3+-related emission [30]. Thus, it is reasonable to believe that the near-band-edge transition contributes to our shrunk bandgap. Figure 7 Plot of ( α▪E ) n vs photon energy E . (a) n = 2 and (b) n = 1/2. The plots suggest that the BFO has a direct bandgap of 2.68 eV. On the other hand, it deserves nothing that there is controversy about bandgap sensitivity of the epitaxial thin film to compressive strain from heteroepitaxial this website structure [5, 7]. Considering that the degree of compressive stress imposed by the epitaxial lower layer progressively decreases with increasing BFO thickness [3], our result 2.68 eV from the BFO thin film prepared

by PLD with a 99.19-nm thickness is compared to the reported ones of the BFO film on DSO or STO with comparable thickness as well as that deposited by PLD, as listed in Table 1. Table 1 Bandgap of BFO thin film (prepared by PLD) on different substrate Bandgap (eV) Substrate Film thickness (nm) 2.68 (this work) SRO-buffered STO 99.19 2.67 [8] DSO 100 2.80 [7] Nb-doped STO 106.5 The bandgap of BFO on SRO is almost the same as that on DSO and is smaller than that on Nb-doped STO. It is noted that the in-plane (IP) pseudocubic lattice parameter for SRO and DSO is 3.923 and 3.946 Å [11], respectively, AZD5363 while STO has a cubic lattice parameter of 3.905 Å [7]. Considering the IP

pseudocubic lattice parameter 3.965 Å for BFO [11], the compressive strain for the BFO thin film deposited on STO substrate is larger than that on SRO and DSO. Thus, the more compressive Terminal deoxynucleotidyl transferase strain imposed by the heteroepitaxial structure,

the larger bandgap for the BFO thin film, which agrees with the past report [7]. The obtained direct bandgap 2.68 eV of the epitaxial BFO thin film is comparable to 2.74 eV reported in BFO nanocrystals [31] but is larger than the reported 2.5 eV for BFO single crystals [32]. This can be understood because even for the epitaxial thin film, the existence of structural defect such as grain boundaries is evitable, which will result in an internal electric field and then widen the bandgap compared to single crystals. On the other hand, a bandgap of 3 eV for BFO single crystals through photoluminescence investigation is also reported [33]. The broad and asymmetric emission peak at 3 eV in the photoluminescence spectra presented in [33] is attributed to the bandgap together with the near-bandgap transitions arising from oxygen vacancies in BFO. However, the Lorentz model employed to depict BFO optical response in our work reveals the existence of a 3.08-eV transition, which is the transition from the occupied O 2p to unoccupied Fe 3d states or the d-d transition between Fe 3d valence and conduction bands rather than the bandgap [26]. Therefore, the broad and asymmetric peak is more likely to be explained as the overlap of the 3.08-eV transition and the bandgap transition with lower energy.