A blog from Intermolecular
Using High-throughput Experimentation (HTE) for Chalcogenide Selectors
At Intermolecular, we believe that advanced material innovation is the core of new technology in the 21stcentury. The progress in materials for digital storage is an excellent illustration of this phenomenon. A typical 32 GB micro SD card in a smart phone can store 3,000 high-resolution photos. When the cards were first introduced in 2005, they would only have been able to hold 3 of the same photos!
Until the 1990s, silicon and a handful of other elements were sufficient to manufacture semiconductor devices. Ever since that time, the need for smaller, higher performance devices coupled with advances in materials manufacturing have led to an explosion of over 45 elements in modern processing technology. As the number of materials increases, the complexity – and cost – of discovering and developing these materials multiplies.
To address these challenges, Intermolecular uses a methodology called high-throughput experimentation (HTE). HTE is a different way of conducting materials discovery. Unlike a conventional fab, where the focus is on controlling a single process and repeating it with high fidelity, HTE allows for experimentation in short periods of time under a wide range of conditions.
Screening and testing of chalcogenide selectors
Considering the above scenario of digital storage devices, here’s an example of where HTE can be useful. Part of the reason that you can fit so much data onto storage media today is because of improvements in solid-state non-volatile memory. For some types of non-volatile memory, the memory elements (or individual bits of storage) are organized into dense, cross-bar arrays. To address a memory element, you energize a row and a column in the array and access the device at the intersection. However, because these arrays are so large, the energy you apply can sometimes “sneak” through a different path, causing you to access (and potentially change) an unintended memory element. Cross-bar arrays therefore use “selectors,” which are essentially switches that block access to memory devices when they are not being used, thereby eliminating sneak paths.
Good candidates for selectors need to meet extensive electrical performance criteria for switching voltage, on/off ratio, endurance, and more. A promising class of materials that could meet these requirements is the chalcogenides, in other words, materials that include the elements of the periodic table under oxygen (notably, selenium and tellurium). However, scientists have found that the electrical characteristics of chalcogenide selectors are highly dependent on the composition of the material and the conditions under which it was processed. The best selectors often include 3, 4, or even 5 elements in exact proportions and created under specific conditions! How do engineers quickly come up with the right material for a particular application and set of requirements? The answer at Intermolecular is HTE.
In a typical chalcogenide screening program, Intermolecular uses its P-30 sputter deposition system to deposit materials with a wide range of compositions. The P-30 has four independently controlled sputtering targets, each of which can contain a different elemental makeup. By changing the power, position, and deposition time of each target, we can control the composition and thickness of the film we deposit. And by using shutters and apertures, each deposition can be isolated to a small region on a 300 mm wafer, allowing for dozens of (usually 12-36) different experiments on a single 300 mm substrate.
Next, we examine physical characteristics of the films we’ve deposited using our extensive suite of metrology tools: what is the composition, crystallinity, thickness, and roughness of the material in each spot? We can tie these physical characteristics to the deposition conditions to establish process trends and allow us to find the materials with particularly interesting characteristics.
Subsequently, we deposit the materials with the most interesting physical characteristics onto electrical test structures and use our E-60 system to determine the electrical characteristics of the material. We start with simple DC measurements, for example to establish the voltage at which the switch is “formed.” The exact compositions that have an acceptable forming voltage are identified (for example, two of the six candidates in the example below). Again, we select only the most interesting candidates and conduct pulsed or AC testing. Based on these more stringent tests, which more closely simulate real device operating conditions, we can see whether the composition of interest is a good candidate material. Finally, we conduct reliability and endurance tests to further demonstrate that the material meets the requirements for device operation.
This approach of down-selecting materials through deposition, physical characterization, and electrical characterization is not specific to chalcogenides: it could in fact be used with any of the selector technologies in the table below. Intermolecular has a proven track record across multiple technologies, conducting thousands of experiments on chalcogenides, MIEC, and transition metal oxide films for selectors.
Over more than a decade, Intermolecular has successfully deployed our approach across a variety of films, including PVD and ALD deposited metals, oxides, nitrides, and chalcogenides. We have conducted tens of thousands of experiments on materials for a wide range of applications and industries like semiconductors, metal alloys for automotive and aerospace, and glass coatings for next-generation architectural applications. Watch the full webinar here and share your thoughts with us on LinkedIn or Twitter!