Over the past half-century, integrated circuit technology has ushered in the Moore's Law era, with transistor density on chips now reaching the level of hundreds of millions per square millimeter. Under the drive of six major technological engines—materials, processes, devices, integration, architecture, and ecosystem—we have harnessed the wave of Moore's Law.

In fact, more than a decade ago, people had already recognized the device bottleneck problem that would lead to the failure of Moore's Law. With the rapid development of AI technology in recent years, higher demands have been placed on computing power.

It is important to understand that chip computing power is closely related to the computing power of individual devices, transistor density, the area of a single chip, and the integration level of a single chip.

Advertisement

Therefore, in the post-Moore era, enhancing the computing power of individual devices and chip integration through new principles and new architectures of novel components is expected to be one of the effective solutions for improving computing power in the short term.

Wang Chen, an associate professor at the School of Materials, Tsinghua University, and a researcher at the Beijing High-end Innovation Center for Integrated Circuits, is dedicated to the research of chip hard technology. He promotes the development of the most critical underlying technology of the AI era from the two aspects of the basic physical properties of new chip materials and post-Moore chips.He is committed to conducting multidimensional systematic basic research and integrated application research on post-Moore chip systems, covering research directions such as the strong field transport behavior and ultrafast dynamics of semiconductor heterojunctions, new types of semiconductor heterojunctions, chip interconnection materials, next-generation semiconductor processes, new principle high-performance devices, multi-source heterogeneous integrated microsystems, and new generation chips.

With the exploration of chip hard technology basic issues and the mapping of technical roadmaps from the five dimensions of "principle-material-device-integration-chip," and based on the development of new materials for the full adaptation of devices, he has efficiently promoted breakthroughs in post-Moore chips. Wang Chen has become one of the Chinese inductees of the "35 Innovators Under 35" by MIT Technology Review in 2023.

Efficiently Promoting Breakthroughs in Post-Moore Chips

The development of the chip industry is a dynamic process. Initially, the issues within the field were that the device sizes were not small enough, and the chip speeds were not fast enough, with the speed of electrons being considered constant. As Moore's Law developed, transistors continued to shrink.Soon, researchers discovered that when the speed increased, the performance of the interconnection materials and architectures that carry the driving and output electrical signals could not keep up with the pace of Moore's Law.

Wang Chen pointed out that, at this stage, the speed of the entire chip is no longer determined solely by the speed of the transistors, but has largely shifted to being determined by the latency of the chip's interconnection system.

In the 1970s and 1980s, the interconnection material initially used in the field was aluminum, but it had a series of problems under high current density conditions, such as high electromigration and crystalline penetration.

At the end of the 20th century, IBM proposed an innovative solution: to use copper as the interconnection material for chips. After Motorola, TSMC, and others introduced copper and industrialized it at the 0.18um node, it was confirmed that copper could reduce the power consumption and increase the speed of chips, and has strong technical potential.

To address the strong quantum effects of chip interconnection materials at the nanoscale in the post-Moore era, Wang Chen and his research team explored the specific properties of alloys such as rhodium and molybdenum, and developed new material systems (such as optical interconnections, superconducting material interconnections, etc.) and process solutions suitable for post-Moore chip nanoscale interconnections, and carried out relevant cutting-edge technical verifications.Recently, Wang Chen and his team have theoretically explored the dynamic behavior of new materials at the quantum scale using terahertz spectroscopy and strong field transport methods. This exploration has been fed back into the customization of the fundamental properties of new materials, and then through optimizing device architecture and integration technology, the technical advantages of special chips have been demonstrated.

The related technologies are expected to stand out in research directions such as new quantum interface phase modulation devices, in-plane transport exciton devices, and ultra-sensitive DNA sensing devices.

By developing highly integrable through-silicon via materials and processes, Wang Chen and his team have achieved a breakthrough in the technology of wafer-level multi-mode three-dimensional multi-source heterogeneous integrated microsystem chips.

Moreover, they have developed highly integrated multi-mode environmental sensing microsystem chips for a variety of chemical and environmental scenarios, providing a new solution to the ultra-high integration challenge of the new generation of high-performance chips.

In the field of three-dimensional high-density integration of chip devices, Wang Chen has improved the bandwidth by developing new intermediate junction layer designs and process plans, breaking through the hierarchical limitations of three-dimensional integration of storage devices and the bottleneck of commercialization in the field.Based on this technology and related research, the Intel Special Contribution Award (Group Recognition Award) was received, with the award reason being "for leading and developing the device erasure technology based on the third-order 3D flash memory chip for the first time in history."

It is important to understand that chips used in autonomous driving or AI require strong configurability and customization, which includes many different modules, different applications, and different users.

Wang Chen stated that the multi-level chip's 3D chiplet technology can flexibly combine different modules and achieve rapid delivery of the entire chip.

Whether it is logic, storage, or sensing, communication, intelligent AI chips, they can be combined through vertical stacking and horizontal expansion, without the need to develop new chips, which is very beneficial for accelerating applications and multi-scenario customization landing.

From dreaming of becoming a scientist to embarking on the path of scientific research.Wang Chen was born in Qingcheng, Inner Mongolia Hohhot, and grew up in the grassland steel city of Baotou, Inner Mongolia, where both of his parents are senior technical personnel.

Fascinated by the majestic power of heavy industry and the grand yet exquisite industrial design, he has developed an intuitive understanding of various fields such as machinery, automation, materials, chemical engineering, and electronics since childhood. As a result, he gradually nurtured the dream of becoming a scientist.

Wang Chen's scientific research interest began to be inspired during his undergraduate study of physics. He studied in the Physics Base Class at Wuhan University, under the guidance of Professor Liao Lei and Professor Xiao Wei, accumulating rich skills in experimental condensed matter physics and theoretical computational physics.

During his undergraduate studies, he conducted research on new types of transistors based on graphene and non-destructive gate dielectric materials, achieving breakthroughs in mobility and current density performance. Moreover, he graduated with the top academic ranking in his major, receiving the National Scholarship and the title of Outstanding Graduate.

Subsequently, Wang Chen pursued his Ph.D. at the University of California, Los Angeles, focusing on micro-nano electronic devices, with his doctoral advisors being Professor Duan Xiangfeng and Professor Huang Yu. Benefiting from the solid foundation in physics he gained during his undergraduate years, he published more than 20 high-impact papers during his doctoral studies.During his second year as a Ph.D. student, Wang Chen co-authored a paper published in Nature Nanotechnology[3], where he invented a pioneering single-atom-layer semiconductor lateral heterojunction, representing a cutting-edge breakthrough in the field of devices at that time.

The research was rated by the industry as a "milestone work for new structural devices," and to date, the related paper has been cited more than 1200 times. Wang Chen said: "The high-level scientific research training and cutting-edge field cognition during my Ph.D. period have greatly benefited me, and this is the foundation of my path to high-level scientific research innovation."

Professor Zhang Hua from City University of Hong Kong commented in a review: "Lateral heterojunction devices have ideal diode heterojunction characteristics because each component has a defined p-type or n-type characteristic, and they are an ideal platform for studying optoelectronic devices[4]."

In addition, Wang Chen also reported a brand-new material system of atomic layer semiconductor molecular superlattices, proposing a brand-new solution to the material bottleneck in the development of new high-performance semiconductor devices[5].

Because this is a very interdisciplinary and cutting-edge topic, almost all the testing methods and testing platforms needed to be built from scratch during the entire experimental process of the project.The seemingly ordinary issues of precisely controlling electrochemical dynamics and the leakage control of ionic gate transistors require experimental optimization that can take 3 months or even longer. Wang Chen resolved these issues by gradually improving the equipment.

He recalls, "Even the common high-resolution transmission electron microscopy images were difficult to capture due to the complexity of the material system. In the dark electron microscopy room, my colleagues and I spent nearly a hundred hours adjusting the equipment."

After several setbacks, the paper was finally published in Nature, becoming his "closing work" during his doctoral studies.

In a review, American Academy member and Stanford University professor Cui Yi, along with others, emphasized the significance of developing an "in-situ photo-electrochemical dynamics platform" in the research of atomic layer semiconductor-molecular superlattice devices [6].

To quickly promote innovation in the chip industry, after completing his doctorate, Wang Chen worked in Silicon Valley at well-known chip companies such as Intel and Lam Research, serving as a senior researcher and project leader. He was responsible for the core development of multiple generations of high-performance chips, possessing research and development capabilities in high-end chip device/architecture design, materials, process integration, wafer verification, and yield improvement.He stated: "These work experiences have endowed me with innovative thinking from basic research to the chip industry, and a macro understanding of the entire process of industrial chip development from 0 to 1. A successful product is an extremely complex system engineering, not just a single process or technology."

Overcoming numerous difficulties during the pandemic, driven by a passion for innovation and the pursuit of greater academic freedom, with the ambition to solve the national "bottleneck" problems, he joined Tsinghua University as a scientist in the interdisciplinary field of new materials and solid-state electronics.

"The childhood dream of becoming a scientist has actually come true under a fortuitous opportunity," Wang Chen exclaimed.

Using seed technology to create a "polygonal warrior" for enterprise use.

Not only did he establish an independent research group NEXT Lab, but to break down the barriers between chip research laboratories (Lab) and semiconductor manufacturing engineering (Fab), Wang Chen also created the distinctive NEXT Mini-Fab.By developing multi-level progressive systemic basic research and integrated applied research, post-Moore chip research based on new materials, new principle devices, and new processes.

"At Tsinghua, compared with follow-up research, I want and need to do more frontier exploration research with original innovation. If technology is a big tree, I believe that only when the roots are deep can its branches and leaves flourish in the future," said Wang Chen.

At present, the field's understanding of many of the most basic physical properties and mechanisms of chips is still unclear. Therefore, he and his team explore the essence of high-performance materials and devices by investigating basic physical parameters and dynamic behaviors, to determine the technical potential of post-Moore devices.

Different from the single-point breakthrough of basic research, Wang Chen's industrial work experience has made him very aware of the actual needs of enterprises. Therefore, from the beginning of the research design, he has promoted it with a systematic approach, in order to create a "polygonal warrior" suitable for the connection of enterprise technology.

It is understood that Wang Chen's laboratory has built a special chip pilot production line. Researchers will prepare the prototype chips in the laboratory and directly compare their performance with similar chips, so as to better meet and get closer to the needs of technology originality, technology development, and technology landing."To this end, we are also constructing a roadmap for key chip component technology, innovating from the perspective of theoretical working mechanisms. We hope to develop a series of seed technologies that can provide the best device technology solutions for specific post-Moore chip fields," said Wang Chen.

It is reported that Wang Chen's team has cooperated technically with several leading companies in the country and has obtained multiple independent and joint patents. In the future, they expect to promote the development and application of high-performance customized chips through the original new generation of device technology.

The main service is the exploratory research of post-Moore components, and the application scenarios include: high-performance intelligent computing chips, bio-detection chips in specific fields, comprehensive environmental perception chips, customized storage chips, optoelectronic integrated heterogeneous chips, etc.

He believes that technological breakthroughs often have non-linearity. In the post-Moore era, the new generation of device technology will inevitably give birth to brand new high-performance human-computer chips and high-performance intelligent chips. "In the future, fully intelligent chips will greatly expand human cognition, change human lifestyles, and subvert the interaction between people and between people and nature."

Next, Wang Chen plans to lead the team to expand the trial proportion of new materials in current chips and carry out larger-scale wafer testing, striving to iterate the development of post-Moore chip prototype models with broad technical potential and product potential on a three-year cycle.