Pulsed laser deposition of large-sized superlattice films with high uniformity Open Access

Strong correlated oxides are known to have exciting properties, including high-T_C superconductivity, colossal magnetoresistance, and multiferroicity.^1–6 Emergent phenomena are often observed when these materials are fabricated into superlattices or heterostructures due to the interplay of factors including interfacial proximity effects, interlayer coupling, and the individual characteristics of the formation components.^7–12 Thus, superlattices of strongly correlated oxides offer unique playground for observing exotic phenomena that are important for both basic science research and future device applications. While the most common growth technique for fabrication of oxide superlattices is pulsed laser deposition (PLD),^13–15 the sample size prepared by PLD is rather small confined by the plasmon plume.^16,17 Moreover, the uniformity of the film thickness sensitively depends on the shape of the laser spot, the plume structure, and other growth conditions (the distance between the substrate and the target, oxygen pressure, etc.).^18–20 These characteristics of conventional PLD limit its potential for device applications.²¹ It is thus important to develop a method to grow large-sized oxide superlattices with high uniformity. The large-sized samples are also beneficial for neutron scattering experiments, which are crucial for determining the structural, magnetic, and electronic properties of the superlattices.²² 

Previous efforts to grow large-sized films with uniform thickness by PLD often follow two strategies: (1) moving the substrate while keeping the laser spot fixed,^22–28 which is always compatible with off-axis PLD,^27,29–31 and (2) scanning the laser beam across the target.^41–45 In the first approach, by carefully tuning the plume characteristics, adjusting the relative positions between the substrate and the plume, and synchronizing the pulsing laser with the substrate motion, the sample size is only limited by the moving range of the substrate, which can be very large.^27,32,33 Various tricks are also considered in this method to provide thickness uniformity of the deposited films in the laboratory, such as using mask slits^23,34 or tilting targets.^27,35 Uniform thin films larger than 1 inch can be realized by industrially scaled large-area PLD, which use special cylindrical targets and a homogenized laser beam.^36–39 However, the constant motion of the substrate makes it impossible to monitor the layer-by-layer growth by in situ reflection high-energy electron diffraction (RHEED) intensity oscillations, which is crucial for fabricating high quality superlattices.^38,40 In the second approach, the scanning area must be in the same focal plane to ensure that the plume energy remains unchanged. Since the laser focal plane and the target surface are not coplanar, one needs to scan the laser along the intersection line between the focal plane and the target plane.⁴¹ While these approaches have shown success in growing large-sized single films,^37,46 multilayer stacks,^32,33 and even 2D materials,^47,48 they have not been employed to grow large-sized and uniform oxide superlattices.

In this paper, we present a new PLD setup to grow large-sized oxide superlattices with uniform thickness and similar physical properties across the whole sample. The most important part of our setup is laser scanning. Three motors are used to control the rotation of the high reflective mirror and the motion of the focusing lens to keep the laser spots in the same focal plane when scanning the laser beam on the whole target surface. The high reflective mirror can be rotated along two orthogonal axes, while the focusing lens is moved along the optical rail to compensate the optical path difference caused by the small-angle rotation of the high reflective mirror. In addition, we improve the efficiency and uniformity of substrate heating by inserting a SiC thin plate between the sample holder and the substrate as an infrared absorption susceptor. Using our PLD setup, we fabricate La_0.7Sr_0.3MnO₃ (LSMO) and Pr_0.625Ca_0.375MnO₃ (PCMO) superlattices with the format of [(LSMO)₁₂/(PCMO)₆]₇. Nine different regions across the superlattices are selected to examine the quality and uniformity of the superlattices. The XRD measurements of all nine regions show distinct and identical superlattice peaks and fringe oscillations, and AFM data show that the superlattice film is droplet-free and atomically flat. Along with the ellipsometry-based thickness mapping measurements and physical property measurements, we conclude that our PLD setup is capable of growing high quality large-sized oxide superlattices with uniform thickness.

Figure 1 shows the schematic view of the PLD system employed in this study, which includes the sample stage, target stage, radiant heating setup, scanning ultraviolet laser beams, and other accessories.

The six-target stage is located at the bottom of the main chamber, with the target switching and spinning motions all controlled by using step motors. The distance between the target surface and the substrate can be varied from 30 to 110 mm.

The layer-by-layer superlattice growth process is monitored by in situ RHEED intensity oscillations. The atomically flat growth front and the uniform thickness control with atomic procession are critical for fabricating high quality superlattices.

The sample stage has 5 degrees of freedom, including three translational motions (XYZ manipulator) and two rotational motions (pitch and yaw rotary), which ensures that the substrate is placed at the hemispherical center and aligned at a grazing incident angle with the RHEED incident beam. It should be noted that once the substrate is aligned into a good RHEED geometry, the substrate position is fixed during the growth process to allow persistent RHEED oscillations for growth control.

The sample is heated by an infrared laser with the wavelength of 808 nm. The infrared laser fiber (NA = 0.22) is fixed on a bracket with its front placed at the focal point of a focusing lens (1-in. dia. and focal length: 35 mm). When the guiding laser is turned on, a red circular spot can be seen on the back of the sample holder, covering the entire substrate area. The size of the heating spot can cover the entire substrate area by adjusting the distance between the infrared laser fiber and the focusing lens, which ensures uniform heating of the entire substrate. The heating temperature can reach more than 1000 °C, and the lifetime can exceed 20 000 h under one bar oxygen environment. The entire heating assembly is fixed on a flange outside the main chamber.

The pulsed laser beam from the KrF excimer laser passes through two lenses and an ultraviolet fused quartz glass window, focusing at the center of the target. The high reflective mirror is controlled by using two step motors (actuator 1 and actuator 2), allowing the laser beam to scan the target surface by a small pitch and yaw angle rotation of the high reflective mirror, as illustrated in Fig. 2(a). It should be noted that the target surface and the laser focal plane are not coplanar [see Fig. 2(b)]. When the laser beam scans along the X axis (parallel to the intersection line of the focal plane and the target plane), the optical path remains unchanged. In contrast, when the laser beam scans along the Y axis (perpendicular to the intersection), there will be an optical path difference, i.e., A is under focus, O is on focus, and C is over focus [see Fig. 2(d)], which will influence the plume intensity and shape. In order to address this issue, in our design, the focus lens is movable along the optical rail with a servo motor to compensate the optical path difference [see Fig. 2(a)]. Since the motions of the high reflection mirrors and focus lens are adjustable, including angles and speed, they can be synchronized with the pulsed laser beam. Figure 2(c) shows a scanning path of the laser beam on the target. The laser beam scans along the X or Y axis when the high reflection mirror has a pitch and yaw angle rotation. The selection of the scanning area depends on the size of the substrate, which usually covers the projection of the substrate on the target so that the plasma plume can cover the whole substrate. It should be noted that the center of the scanning area must correspond to the center of the substrate. To keep the atomically flat growth front and the chemical uniformity of each layer in superlattices, in situ annealing is necessary after the deposition of each layer.

To ensure the uniformity of the large-sized superlattices, the heating temperature must be uniform for the whole substrate.⁴⁹ The sample holder is hollowed (15 × 15 mm²), and a SiC thin plate is inserted between the sample holder and the substrate as an infrared absorption susceptor (0.5 mm thickness) at the same area, as shown in Fig. 3. The substrate is clamped on the SiC thin plate. The beam diameter of the infrared laser is around 16 mm. This design improves the temperature uniformity on a large-sized substrate because the SiC plate has a high infrared absorption coefficient and a high thermal conductivity.⁵⁰ 

The [(LSMO)₁₂/(PCMO)₆]₇ superlattice is epitaxially grown on a 10 × 10 mm² SrTiO₃(001) substrate by using the aforementioned PLD setup (KrF excimer laser, 248 nm, 10 Hz). The substrate is heated at 760 °C under an oxygen pressure of 8 × 10⁻² Torr with 8% ozone during growth. The layer-by-layer epitaxial growth is monitored by in situ RHEED intensity oscillations, as shown in Fig. 4. The beam size of RHEED is small, and the center of the substrate is chosen. The LSMO and the PCMO targets are switched after every 12 and 6 RHEED oscillations completed for LSMO and PCMO growth, respectively. The time to complete one layer growth (∼30 s) is much longer than the time for the laser to complete one full scanning cycle on the target (∼6 s), which ensures the uniformity of the film. In order to check the uniformity of the superlattice sample, we first map the whole sample by ellipsometer imaging (ACCURION Spectroscopic Imaging Ellipsometer Nanofilm_EP4SE) to determine the thickness distribution, as shown in Fig. 5. The measurement spot size is 500 × 180 µm² for every 1 × 1 mm² area. It should be noted that the thickness of the sample in the four corners is relatively small because these corners are covered by clamps, as shown in Fig. 3(a). The standard deviation of the thickness is 0.308 nm, indicating that the thickness is highly uniform across the whole sample.

To further examine the uniformity of the superlattice, we select nine representative regions to characterize their morphological, structural, and physical properties. Figure 6 shows the morphology of the nine regions obtained by atomic force microscopy. All nine regions are droplet-free and featured by atomically flat terraces separated by parallel step edges, which are inherited from the substrate. Figure 7 shows the superlattice period determined by XRD and XRR measurements. Distinct and identical superlattice peaks and fringe oscillations are seen in all nine regions. Note that the XRD and XRR curves of the nine regions are artificially shifted in the y axis for clear visibility. The rocking curves around the 002 reflection are almost overlapped, and the average FWHM of the 002 reflection is 0.018° (see the inset of Fig. 7). The XRD and XRR measurements clearly show that the high quality superlattice is uniform across the whole sample.

The physical properties are also nearly identical for all nine regions. Figure 8(a) shows the temperature-dependent resistivity (ρ–T) curves acquired from the nine regions of the superlattice. The insulator to metal transition temperature (T_P) of the nine regions is all ∼314 and ∼322 K from cooling and worming curves, respectively. Figure 8(b) shows the temperature dependence of the magnetic properties of the nine regions of the superlattice sample. Curie temperature (T_C) and saturated magnetic moment (M_S) are nearly identical for all nine regions, which are determined to be 316 K and 3.56 µB/Mn, respectively. Considering the fact that the physical properties of the strongly correlated systems are ultrasensitive to any small variations of structural and chemical properties, our PLD setup is capable of fabricating a large-sized, high quality superlattice sample with extreme uniformity that is suitable for device fabrication.⁵¹ 

We have successfully grown large-sized, high quality manganite superlattices with high uniformity by laser beam scanning on the target surface. By combining the rotation of the high reflection mirror and the liner motion of the focus lens, the laser beam is kept focused on the target surface so that the intensity and the shape of plasma plume will not be changed during serpentine path scanning. In addition, we inserted the SiC substrate as an infrared absorption susceptor to improve the uniformity for large sample heating. Based on our setup, we fabricated the 10 × 10 mm² [(LSMO)₁₂/(PCMO)₆]₇ superlattice films. The superlattices are cross-checked by AFM, XRD, and ellipsometer measurements, which consistently indicate that our superlattices are atomically flat and droplet-free. The thickness, the superlattice period, and the physical properties are all uniform across the whole sample. The T_P of the superlattices reaches ∼314 K, which is potentially useful for room temperature device applications. Although the sample that we demonstrated in this work is only 10 × 10 mm², the sample size can be easily expanded to the same as the size of targets (∼2 in. dia). The size of the sample is mostly limited by two factors: the target-to-substrate distance and the scanning angle between the laser beam and the target surface. (1) The target-to-substrate distance mainly depends on the angular spread of the ejected flux.⁵² For most applications, this distance will be in the range of 3–10 cm.¹⁵ In this work, a distance of about 5 cm is recommended. (2) The scanning angle cannot be very small for it is difficult to control the spot size (angles larger than 45° are recommended) and cannot be too large as the sample holder may be in the beam path.¹⁵ In the single laser beam scheme, the target-to-substrate distance is ∼5 cm (∼2 in.), the laser-target angle is 45°, and the maximum sample size will be 2 in. In principle, two laser beams can be used simultaneously, and the sample size can be enlarged to ∼4 in. Our large-sized superlattice growth technique can be used to fabricate oxide superlattices on Si substrates, making oxide spintronics compatible with CMOS technology.

We acknowledge beamlines BL14B1 and BL02U2 (Shanghai Synchrotron Radiation Facility) for providing the beam time and help during experiments. This work was supported by the National Key Research and Development Program of China (Grant Nos. 2020YFA0309100), the National Natural Science Foundation of China (Grant Nos. 11904052, 11991060, 12074075, 12074073, and 12074071), the Shanghai Municipal Natural Science Foundation (Grant Nos. 19ZR1402800 and 20501130600), and the Young Scientist Project of MOE Innovation Platform, and the State Key Laboratory of ASIC & System (Grant No. 2021KF004).

The authors have no conflicts to disclose.

The data that support the findings of this study are available from the corresponding authors upon reasonable request.