Advancing Materials and Integration Technologies for Next-Generation Ferroelectric Devices
As the semiconductor industry continues to push the physical limits of scaling, materials science is key to unlocking performance improvements. Novel material systems and integrations are critical for scaling of the semiconductor devices and for enabling new types of devices. New material system becomes more complex in terms of number of elements and the control of interfaces, making development longer and more costly.
We use a method of quickly down selecting and fine-tuning multi-element materials which meets both physical and electrical key performance indicators. Starting with combinatorial physical vapor deposition and unit cell electrical test vehicles a wide elemental space is screened, and compositions are down selected. With a narrower materials composition space defined, the process is transferred to atomic layer deposition to check the physical and electrical performance with a 3D compatible process. At each step, different types of test devices (HAR structure for conformality, two-terminal devices and three-terminal devices with critical dimension below 100nm scale) are used to evaluate for improved thermal stability, better step coverage, reduced impurity levels, and the right electrical performance. Machine learning algorithms are also used to fine tune the materials for accelerated learnings.
Several ferroelectric material compositions and stacks have been optimized for Ferroelectric Field Emission Transistor (FeFET) devices using this method. These ferroelectric materials are being investigated for logic device scaling, enhancing existing memories such as DRAM and NAND flash and new types of memory.