17�Թ�

The technology entitled “Memristor Structures with Analog Switching Characteristics and Method for Fabricating the Same” has recently been granted United States Patent US 12,670,378 B2 by the United States Patent and Trademark Office (USPTO).

The inventors comprise Assoc. Prof. Phạm Kim Ngọc and MSc Phạm Phú Quân from VNUHCM–University of Science (HCMUS); Prof. Phan Bách Thắng from VNUHCM-University of Health Sciences; and Assoc. Prof. Nguyễn Trần Thuật from the Institute of Semiconductors and Advanced Materials, 17�Թ� National University, Ha Noi (VNU).

An idea rooted in materials science

For more than a decade, global research into memristors has concentrated predominantly on metal oxide systems such as , , and . This patent establishes a novel paradigm by constructing and protecting a materials platform based on chromium oxide () for memristor technology. Rather than merely proposing a specific device architecture, the invention positions  as the central functional material, which can be flexibly integrated with various oxide layers and metal electrodes to design new generations of memristors. This platform paves the way for devices capable of multi-state storage and non-linear operation—core characteristics required to realise artificial synapses within next-generation computing hardware.

The nucleus of a new computing architecture

The prominent feature of the patent lies not only in the memristor structure itself, but also in the deployment of this material system into synapse arrays. The invention describes a [Row  Column] array architecture integrated with passive selectors (diodes), wherein each intersection functions as a memristor acting as an electronic synapse, capable of simultaneous local data storage and processing. This architecture opens up avenues for developing In-Memory Computing and Neuromorphic Computing chips, where millions of memristors can be integrated into hardware neural networks. Such networks possess the capacity to execute parallel computations, thereby significantly reducing data transfer between memory and processors.

A foundation for the future

The journey from a novel material design in the laboratory to an international patent represents the inception of a grander vision: the development of integrated AI chips powered by memristor technology. The invention emerged from research activities undertaken at the Faculty of Materials Science and Technology, HCMUS. This milestone further strengthens the foundation for collaborative programmes between scientists at VNUHCM and VNU across the entire research chain, spanning from materials and devices to semiconductor chip technology.

The research process received funding and support from the Vingroup Innovation Foundation (VinIF) and HCMUS. This financial backing was instrumental in connecting specialized research groups, sustaining investigative activities, and enabling laboratory findings to mature into internationally protected intellectual property.