Self-assembled single-crystal silicon circuits on plastic

Abstract

We demonstrate the use of self-assembly for the integration of freestanding micrometer-scale components, including single-crystal, silicon field-effect transistors (FETs) and diffusion resistors, onto flexible plastic substrates. Preferential self-assembly of multiple microcomponent types onto a common platform is achieved through complementary shape recognition and aided by capillary, fluidic, and gravitational forces. We outline a microfabrication process that yields single-crystal, silicon FETs in a freestanding, powder-like collection for use with self-assembly. Demonstrations of self-assembled FETs on plastic include logic inverters and measured electron mobility of 592 cm2/V-s. Finally, we extend the self-assembly process to substrates each containing 10,000 binding sites and realize 97% self-assembly yield within 25 min for 100-microm-sized elements. High-yield self-assembly of micrometer-scale functional devices as outlined here provides a powerful approach for production of macroelectronic systems.

Fig. 1.

The heterogeneous self-assembly process. Microcomponents are introduced over a template submerged in a liquid medium and moved with the fluid flow. Self-assembly occurs as microcomponents first fall into complementarily shaped wells and then become bound by the capillary forces resultant from a molten alloy.

Fig. 2.

Details of the self-assembly process for a single microcomponent. (A) The microcomponent approaches a binding site with complementary shape. (B) The microcomponent is held by capillary force resultant from molten-alloy-bridging the metal pads positioned on the microcomponent and on the template.

Fig. 3.

Transistor microfabrication process flow. Completion of the last step yields a collection of micrometer-scale, released components.

Fig. 4.

Released silicon FETs. (A) A collection of freestanding single-crystal silicon FETs. (B and C) Optical microscope images of and triangle-shaped (B) cross-shaped (C). (D) Measured performance of a typical triangular transistor before release from SOI wafer.

Fig. 5.

Plastic template fabrication flow.

Fig. 6.

Heterogeneous self-assembly results. (A) Optical microscope image of multiple types of microcomponents self-assembled on plastic to complete an electrical network. (B) Measured current–voltage curve between the two leftmost ports of the network verifying proper electrical connection between the microcomponents and the template.

Fig. 7.

Self-assembly of FETs. (A) Optical microscope image of a plastic substrate with empty binding sites (two outlined with white lines for clarity). (B) The template after completion of the self-assembly process showing the position of FETs and diffusion resistors. (C) Measured performance of the inverter shown in B. The overshoot is likely due to the parasitic line capacitance of the interconnection on the substrate.

Fig. 8.

High yield (≈97%) self-assembly of 100-μm, circular, single-crystal silicon elements onto a flexible plastic template containing 10,000 binding sites. (A) Completed template. (B) Close-up image of ≈2,000 binding sites, self-assembled elements, and electrical interconnects.

Fig. 9.

Measured self-assembly rate. The percentage of correctly self-assembled elements after each element pass step is shown, indicating a 99% average self-assembly yield after five element passes. Error bars indicate the standard deviation.

Fig. 10.

Modeling results. (A) Calculated energy landscape of component as it joins with solder. (B) Surface Evolver model of capillary forces of molten solder. (Inset) An optical microscope image used to determine the contact angle of molten alloy on gold under ethylene glycol. (C) Capillary forces in the z direction.