The perovskite manganese oxides with the general formula of R_1−xA_xMnO₃ where R is typically a trivalent rare earth (R = La, Pr, Sm, ...) and A is a divalent alkaline (A = Ca, Sr, Ba) exhibit rich phase diagrams as a result of interactions among the charge, spin, orbital, and lattice degrees of freedom of Mn 3d electrons. La_1−xCa_xMnO₃ is famous because for a broad range of doping, these materials have a paramagnetic to ferromagnetic transition upon cooling, which is accompanied by a sharp drop in the resistivity. The behavior is usually explained with double exchange (DE) theory based on exchange of electrons between Mn3+ and Mn4+ ions. When DE interaction wins over super exchange and thus, ferromagnetic metallic phase prevails, the colossal magnetoresistance (CMR) is observed [1–5].

Among the phases observed in La_1−xCa_xMnO₃ manganites phase diagram, La_0.7Ca_0.3MnO₃ (LCMO) is a ferromagnetic metal with mixed valence state of t2g3eg1 and t2g3eg0 configurations and CaMnO₃ (CMO) is an antiferromagnetic insulator with t2g3eg0 configurations. An interesting situation is when a ferromagnetic metallic LCMO interfaces with CMO, which is an effective vacuum to the e_g electrons in LCMO. The e_g electrons in LCMO are expected to be delocalized across the interface within a certain length scale [3–7].

In this work, we characterize a heterostructure composed of a thin layer CMO with thickness t sandwiched by top and bottom LCMO layers. In particular, we focus on the optimization of sample synthesis and the characterization of crystallinity, surface morphology and electrical transport of the LCMO/CMO/LCMO heterostructures.

The La_0.7Ca_0.3MnO₃ and CaMnO₃ targets were prepared by standard solid state reaction. First, the pure La₂O₃, CaCO₃ and MnO₂ powder were calcined at 1000 °C for 6 hours in air. Next, they were ground in an agate mortar for thorough mixing before pressing this mixture into a round pellet with diameter of 2 cm by using a pressing die. Then, the pellets were calcined in air inside furnace. The process of mixing and calcinations were repeated several times. Finally, the pellets were sintered in air at 1350 °C for 24 hours. High density of pellets is necessary to minimize formation of particulates during deposition. Prior to deposition, the SrTiO₃ (STO) (001) substrate (0.25 × 0.25 cm from Crystec) were cleaned twice in acetone and twice in ethanol to remove dust or any other greasy contaminants. Then, substrates were glued on substrate holders with silver paste to ensure good thermal contact.

A KrF excimer laser with wavelength of λ = 248 nm was used to irradiate the targets with repetition rate of 2 Hz. After stabilization of pressure (background pressure – 10⁻⁶ Torr), temperature and laser power for about 15 minutes, the target was ablated for several minutes to remove all the contaminants on the target surface and then, deposition starts.

XRR and XRD results of heterostructures was carried out using synchrotron radiation in Pohang Light Source. The morphology of our samples was measured by the AFM technique Here, we used the contact mode. RT curves were measured using Van der Pauw method. In order to use this method, the size of sample with 2.5 × 2.5 mm was contacted by four copper wires at the edges of sample by silver paste. Here, current flow along one edge of the sample and the voltage across opposite edge is measured.

1. The surface morphology of heterostructures

Fig.1 shows the AFM results of CMO, LCMO and LCMO10/CMO20/LCMO10 thin films, respectively, where the numbers in the heterostructure denote the deposition time. A uniform and homogeneous surface morphology with small RMS roughness is seen. The RMS roughness Rq was calculated to be 1.96 nm, 0.26 nm and 0.24 nm for CMO, LCMO, and LCMO10/CMO20/LCMO10, respectively.

Figure 1. (Color online) The AFM images of thin films: a) STO/CMO, b) STO/LCMO, c) STO/LCMO10/CMO20/LCMO10 deposited at T_s = 700 °C, P = 0.2 torr, F =1.5 J/cm².

2. The structural characterization of thin films

The XRR results of heterostructures is shown in Fig.2. The film thickness of the CMO30 and LCMO30 samples are determined to be 30.1 nm and 15.4 nm by extracting from fitting data. The XRD pattern is shown in Fig.3, the LCMO peak in single layer is located at lower angles than the CMO peak, which indicate the out of plane lattice constant of LCMO layer higher than that of CMO layer. Compared to the single LCMO and CMO layers, the position of the film peaks shifts to the higher angles when the thickness of the middle CMO layer is increased. This confirms that the CMO layer in the middle is tensilely strained by the top and bottom LCMO layers and its unit cell elongates along the a- direction by elastic deformation. The Kiessig fringes in the heterostructures are more complex than in the single layer because the heterostructure films have two more interfaces. Judging from AFM, XRR, and XRD data, all the samples are comparable quality in terms of surface smoothness and crystallinity.

Table 1 Deposition conditions of La_0.7Ca_0.3MnO₃/CaMnO₃/ La_0.7Ca_0.3MnO₃ heterostructures on SrTiO₃ (001) substrates.

	SrTiO3/ La0.7Ca0.3MnO3/CaMnO3/ La0.7Ca0.3MnO3 (STO/LCMO/CMO/LCMO)
	STO/LCMO10/CMO(t1)/LCMO10	STO/LCMO(t2)/CMO05/LCMO10
Substrate temperature (Ts)	700 °C	700 °C
Energy density (F)	1.5 J/cm2	1.5 J/cm2
Oxygen Pressure (P)	0.2 – 0.3 torr	0.2 – 0.3 torr
Deposition time (t)	t1 = 10, 15, 20, 30 mins	t2 = 10, 20, 30, 60 mins
Note: LCMO10 is a LCMO layer which was deposited for t = 10 mins. CMO05 is a CMO layer which was depsoited for t = 5 mins.

Figure 2. (Color online) The XRR pattern of CaMnO₃ and La_0.7Ca_0.3MnO₃ thin film on the SrTiO₃ (001) substrates deposited at T_s = 700 °C, P = 0.2 torr, F =1.5 J/cm², t = 30 min.

Figure 3. (Color online) The XRD pattern of thin films on the SrTiO₃ (001) substrates deposited at T_s = 700 °C, P = 0.2 torr, F =1.5 J/cm²

3. The resistivity of thin films

We measured the sheet resistance as a function of temperature (RT) of single CMO, LCMO, and heterostructures by using Van der Pauw method. Temperaturedependent resistivity of the CMO thin film is shown in Fig. 4a, which shows the expected insulating behavior. On the other hand, the LCMO thin films in Fig. 4b exhibit the metallic and insulating behaviors at low and high temperatures, respectively.

Figure 4. (Color online) The RT curves of a) STO/CMO, b) STO/LCMO, c) STO/LCMO10/CMO20/LCMO10 deposited at T_s = 700 °C, P = 0.2 torr, F =1.5 J/cm²

In case of heterostructures, we investigated the resistivity of LCMO/CMO/LCMO heterostrucutures while changing the thickness of the middle CMO layer or the bottom LCMO layer. First, we investigated LCMO/CMO/LCMO heterostructures with the bottom LCMO layer varied from 10 to 60 mins and with the middle CMO and top LCMO layers whose deposition times are 5 mins and 10 mins, respectively. As the deposition time of the bottom LCMO increases, the heterostructures become more metallic.

Then, we prepared LCMO/CMO/LCMO heterostructures while deposition time of the middle CMO layer are changed from 5 mins to 30 mins while keeping the deposition time of bottom and top LCMO to be 10 mins. The RT result is shown in Fig.4c for the film with middle CMO20 layer deposited for 20 mins. This heterostructure shows metallic behavior from 2 K to 25 K and insulating behavior at higher temperatures. This heterostructure shows an insulator-metal transition around 30 K – 70 K. This reduced insulator-metal transition temperature is related to the increased localization tendency of the e_g electrons due to the influence of the middle CMO layer. As the quality of CMO and LCMO in single-layer and heterostructure samples is similar, the change of the RT curve is attributed to heterostructuring.

We have reported the synthesis of LCMO/CMO/LCMO heterostructures. After we found an optimal growth condition common for LCMO and CMO, we successfully grew LCMO/CMO/LCMO heterostructures with sharp interfaces and good crystallinity. We observed a very pronounced peak at low temperature in the RT curve of LCMO/CMO/LCMO heterostructures. At present we do not fully understand its origin and so, further experiments are necessary.

This research was supported in part by the Daegu University Research Funds.