Epitaxial Growth of Perovskite Strontium Titanate on Germanium via Atomic Layer Deposition

The overall goal of this procedure is to grow high dielectric constant crystalline perovskite oxides in a scalable process to function as gadoxides directly on geranium, a future microelectronics platform. This method can answer key questions in the micro electronics field which is can crystalline oxides which bind across the interface produce films with extremely low trap densities and interface states? The main advantage of this technique is it's easy to implement in a manufacturing setting and should be readily extendable to three dimensional device architectures.

Generally individuals new to this method will struggle because of lack of minorities with handling the precursors and that having timing the sequence to produce steps for their particular hardware in the reactor. We first had the idea for this method when collaborators grouped perovskites directly on geranium using molecular beam epitaxy and we thought to adapt their process to atomic layer deposition. Visual demonstration of this method is critical because the sample cleaning, transfer step and timing need to be modified for other systems while ensuring a clean and reconstructed growth surface.

For this protocol use a heavily doped intact geranium if electrical measurements of the film are needed otherwise all doping levels and doping types are acceptable. Place the geranium substrate polished side up into a small beaker. Add about 1 cm of acetone and place the beaker in a bath sonicator, sonicate it for ten minutes.

Decant the majority of the acetone but do not pour out or flip the geranium substrate. Then rinse the walls of the beaker with isopropyl alcohol until about 1 cm full. Then pour off most of it.

Refill the beaker with the same volume of isopropanol and repeat the sonication. Repeat the rinse procedure, this time using water. Be careful to not disturb the substrate.

Then remove the substrate with tweezers and dry under an inert gas such as nitrogen. Now clean the substrate in a UV ozone cleaner for 30 minutes. When 8 minutes remain start venting the load lock so the substrate can be immediately loaded into the vacuum system when the cleaning is completed.

Place the substrate polished side down into a 20 by 20 mm holder. Ensure the substrate is flush with the bottom of the holder. After the vacuum system has completely vented open the load lock.

Place the sample holder into an open carrier cart position by aligning the tabs of the holder with the channels of the open cart. Then close the load lock and turn on the load lock turbo molecular pump. When the pressure in the load lock has dropped sufficiently low open the load lock gate valve and move the cart through the transfer line.

Next transfer the germanium substrate into the MBE chamber. Adjust the sample to an appropriate height relative to the MBE sample heating stage. Then ramp the substrate temperature to 550 degrees Celsius at 20 degrees Celsius per minute.

Once at 500 degrees Celsius ramp the temperature up to 700 degrees Celsius at 10 degrees Celsius per minute. After holding the sample at 700 degrees Celsius for an hour cool the sample to 200 degrees Celsius, dropping the temperature by 30 degrees Celsius per minute. Next analyze the substrate by reflection high energy electron diffraction to confirm the two by one reconstructed surface.

In preparation set the ALD reactor temperature to 225 degrees Celsius and let it warm up. Heat the bis-triisopropyl cyclopentadianyl strontium to 130 degrees Celsius in a saturator wrapped with heating tape. At the same time heat the titanium tetraisopropoxide to 35 to 40 degrees Celsius in the heating tape wrapped saturator.

Regulate the room temperature water vapor flow into the ALD system via the needle valve attached to the saturator. Keep the water pressure around 0.1 torr. Also maintain constant precursor temperatures throughout the deposition process.

As soon as the ALD reactor temperature is stable quickly and carefully transfer the sample in situ to the ALD reactor. The time between obtaining a two by one reconstructed surface and loading the sample into the ALD is critical. Minimize this time to avoid any background contamination of the surface.

Crystalline films will not grow on a contaminated surface. Isolate the ALD reactor from the transfer line and switch the exhaust port of the ALD reactor from the turbo molecular pump to the mechanical pump. Turn on the flow controller to use an inert gas such as argon and maintain an operating pressure of about 1 torr during the entire growth process.

After keeping the system stable for 15 minutes program the ALD. First set the unit cycle ratio of strontium to titanium to be 2:1. Then set the strontium and titanium unit cycles to a two second dose for the strontium or titanium precursor each followed by a 15 second argon purge, a one second dose of water and another 15 second argon purge.

Next set the number of cycles for the required deposition thickness. The pressure increases during dosing in our system. Dosing times and pressures needed to reach saturation are likely ALD systems specific and will need to be determined for different systems.

As soon as the ALD deposition is completed carefully transfer the sample to the annealing chamber. There heat the sample to 650 degrees Celsius at 20 degrees Celsius per minute under UHV conditions. Once at 650 degrees Celsius hold the temperature for five minutes and then cool the sample back to 200 degrees Celsius at the same rate.

Ultimately use RHEED to evaluate the annealing result and if needed additional layers can be deposited onto the substrate by repeating the procedure. RHEED imaging of a cleaned and deoxidized germanium substrate showed that a successful deoxidation could be characterized by its smiley faced two by one reconstructed pattern. Kikuchi lines were observed in the RHEED images indicating the substrate had good cleanliness and long range order.

The sharpness and intensity of the RHEED pattern after the deposition and annealing of STO film demonstrated good epitaxial growth. The germanium 3D X-ray photo electron spectrum was free of oxidized germanium peaks with a zero valent germanium peak observed at 30 electron volts. Pre deposition, 36 cycle and 155 cycle depositions are shown respectively in red, brown and black.

The X-ray diffraction pattern of epitaxial STO on germanium had the expected characteristic peaks at 22.8 degrees, 46.5 degrees and 66 degrees. The epitaxial nature of the film was directly confirmed by cross sectional high resolution transmission electron microscopy. This showed high quality epitaxial registry between the STO and germanium and abrupt transitions between layers.

After watching this video you should have a good understanding of how to grow mono crystalline perovskites directly on germanium using atomic layer deposition. This approach is not restricted to strontium titinate. While attempting this procedure it is important to ensure that the germanium surface has the two by one reconstructions.

It's also free of carbon and oxygen contamination that can occur in the transfer lines if the sample is held there for days. Following this procedure are the measures like electrical measurements with MOS capacitor structures can be performed to investigate that interface trap density, film dilation constant and each carrier. After its development this technique paved the way for researchers in the micro electronics field to explore germanium structures that require extremely low equivalent oxide thicknesses.