Research blunder leads Holonyak Lab group to discovery, multiple patents
8/19/2020 10:39:22 AM
From potato chips to penicillin, some of the most famous inventions in history have been accidents. Rather than getting rid of their potential blunders, the creators of these new innovations changed the world. Like these other inventors, Holonyak Lab Interim Director Xiuling Li encountered accidental wafer breakages during an experiment in her early career. Her "misfortune" ended up changing the way semiconductor materials are etched.
Etching is an important step in device fabrication processes whether it be with transistors or lasers. Li’s patented etching technique, metal-assisted chemical etching (MacEtch), started with etching silicon to make it porous, and evolved into anisotropic etching. This allows for the production of extremely high-aspect-ratio damage-free 3D structures for a wide range of semiconductors.
In 2000, Li was trying to make porous silicon by the conventional anodic etching method when the silicon wafer that was used to seal the etching solution broke during the experiment. This resulted in the electrolyte (HF + H2O2) leaking out of the Teflon beaker and causing a potential safety hazard. That piece of wafer was also overlaid with a thin gold (Au) film, which was supposed to be deposited only on the back side of the wafer, but because of a separate mishap was adhered to the front of the wafer at one corner. This gold was then covered by the leaked electrolyte, compounding the two unintended consequences.
“While I was cleaning up the mess from this disastrous experiment, I decided to check the wafer anyway by shining a hand-held UV lamp to see if there was any red light coming out. It is well-known that anodic etching produces light emitting silicon resulting from a quantum confinement induced phenomenon,” said Li, Donald Biggar Willett Professor in Engineering.
Lo-and-behold, that corner of the wafer, where two blunders converged (the metal deposition on the wrong side and the leaked electrolyte), had the brightest red color she had ever seen. Li then intentionally repeated the experiment by simply depositing metal (Au or Pt and other noble metals) and placing the wafer into the electrolyte solution, and proved that in the presence of metal, etching takes places under open circuit. From that point on, she ditched the potentiostat (an equipment that controls voltage or current in a circuit) and relied on the catalytic metal to induce local etching. The rest, as the saying goes, is history.
The paper that documented the results from this discovery, “Metal-assisted chemical etching in HF/H2O2 produces porous silicon,” Appl. Phys. Lett. 77, 2572 (2000), has been cited over 1010 times according to Google Scholar. Today, Li and her students apply MacEtch as a plasma-free anisotropic etching technique to produce device quality 3D structures, not only for silicon, but also for many other semiconductors beyond Si, including Ge, GaAs, InGaAs, InP, GaN, SiC, b-Ga2O3, etc. Eleven MacEtch related patents have been issued, including magnetic-field guided, self-anchored-catalyst, and inverse-MacEtch, all with University of Illinois.
The latest patent, issued this month, involves vapor-phase based MacEtch, which means the etching takes place in the gas phase with independent control of gaseous etchant precursors, while maintaining the plasma-free, damage-free, clean, and simple nature. This patent represents a significant step for scalability. Jeong Dong Kim, Dane J. Sievers, Lukas Janavicius are the co-inventors.
Li believes that serendipity is no accident.
“It requires observations and insights going beyond the obvious,” said Li. “This discovery, and the device research it enabled, would not have happened if we stopped at ‘crying over’ the accidents during the experiments.”
Li and her research group continue looking into improving the material and structure versatility of this technique and demonstrating device feasibility for various applications, including aggressive scaling of high power and high frequency devices, and possibly quantum technologies.