Heterogeneous integration of single-crystalline complex-oxide membranes

Here we introduce and demonstrate a universal method of generating freestanding complex-oxide membranes and stacking them together to artificially create novel heterostructures with unprecedented characteristics.

Like Comment
Read the paper

Epitaxy, the art of growing thin single-crystalline materials on thick crystallographically oriented substrates, is a key enabler for state-of-the-art electronic and photonic devices used today. However, with the advancement and miniaturization of consumer gadgets, integration of many different types of devices grown on difference substrates have become increasingly important, as well as thinness and flexibility. Unfortunately, substrates are typically several hundred microns thick, preventing monolithic integration of devices fabricated on dissimilar substrates. Additionally, due to strict epitaxial constraints [1], it is impossible to create useful heterostructures which could significantly advance many industries. It has been the dream of many material scientists and device engineers to be able to make any arbitrary stacks of ultrathin single-crystalline membranes for ultimate control of physical and quantum properties to create materials with totally new functionalities.

To make this one step closer to reality, we report in Nature a universal method of exfoliating thin epitaxially grown membranes off from its host substrate, and utilize our method to create several artificial heterostructures only several hundred nanometers thick, and manually stack these ultrathin membranes together to create a device with hybrid properties. This work builds upon our previous report on remote epitaxy, which allows epitaxial lift-off by utilizing graphene as the release layer while retaining the substrate as the epitaxial seed [2]. Our material of choice was complex-oxides due to their rich physical and quantum characteristics, such as superconductivity, piezoelectricity, magnetostriction, pyroelectricity, and most ferroic properties.

Figure 1. Exfoliated freestanding complex-oxide membranes

Although we were confident about the remote epitaxial process (in which we were able to create membranes of perovskite, spinel and garnet crystalline structures such as SrTiO3,BaTiO3, CoFe2O4, and Y3Fe5O12), we were not prepared for the challenge of growing PMN-PT, which is a material with one of the highest piezoelectric coefficients today. We quickly found out that graphene could not survive the harsh plasma environment in which PMN-PT was grown, thus precluding the possibility of remote epitaxy for this material. However, we accidentally loaded a PMN-PT film grown on SRO/STO substrate for exfoliation and surprisingly discovered that, even without graphene, PMN-PT would exfoliate from the SRO/STO substrate with atomic precision at the interface. Although the exact mechanism is unclear at this moment, we believe it possibly may be due to the period misfit dislocations generated at the PMN-PT/SRO interface and in-depth investigation is underway.

Figure 2. (a) Magnetoelectric coupled device consisting of freestanding CFO/PMN-PT membranes. (b) In-situ TEM of the CFO membrane as a function of voltage across the PMN-PT membrane. (c) Voltage generated due to magnetoelectric coupling effect for a freestanding CFO/PMN-PT membrane compared to clamped CFO/PMN-PT film.

Excited by our discovery, we wasted no time in stacking freestanding CFO/PMN-PT membranes, a magnetostrictive and piezoelectric material, to create a magnetoelectric (ME) coupled device with unprecedented ME coupling coefficient reported to date. We were also able to physically visualize the strain mediation between the CFO and PMN-PT via in-situ TEM. We also verified other physical couplings such as magnetostatic coupling by stacking CFO and YIG membranes, and also electrical coupling by observing the Fermi-level shift in graphene as it is sandwiched between CFO and YIG membranes (CFO p-dopes graphene while YIG n-dopes graphene). 

These results clearly demonstrate for the first time that new and exciting single-crystalline heterostructures can be made artificially, bypassing the epitaxial limitations altogether. We are very eager to see what combinations of materials the community comes up with to pioneer the next generation of functional devices.

For more information, please refer to our recent publication in Nature, “Heterogeneous integration of single-crystalline complex-oxide membranes” (https://www.nature.com/articles/s41586-020-1939-z).    

[1] Hyun Kum, Doeon Lee, Wei Kong, Yongmo Park, Yunjo Kim, Yongmin Baek, Sang-Hoon Bae, Kyusang Lee, and Jeehwan Kim, “Epitaxial growth and layer-transfer techniques for heterogeneous integration of materials for electronic and photonic devices”, Nature Electronics 2, 439-450 (2019).

[2] Yunjo  Kim,  Samuel  S.  Cruz,  Kyusang  Lee,  Babatunde O.  Alawode,  Chanyeol  Choi,  Yi  Song,  Jared M.  Johnson,  Chris Heidelberger, Wei Kong, Shinhyun Choi, Kuan Qiao, Eugene A. Fitzgerald, Jing Kong, Alexie M. Kolpak, Jinwoo Hwang, and Jeehwan Kim*, “Remote epitaxy through graphene enables two‐dimensional material‐based layer transfer” Nature 544, 340–343 (2017).

Hyun S. Kum

Postdoctoral Associate, Massachusetts Institute of Technology

No comments yet.