In this blog post, we will examine the paradigm shift in the semiconductor industry and the direction South Korean engineering students should take, based on Thomas Kuhn’s ‘The Structure of Scientific Revolutions’.
Semiconductor research, which began with the invention of the transistor, has advanced exponentially over the past few decades. The concept that best represents this progress is “Moore’s Law.” Proposed by Gordon Moore, co-founder of Intel, this empirical law states that the number of transistors integrated onto a semiconductor chip doubles approximately every two years. While this law has long been accepted as the standard for explaining the direction of the semiconductor industry’s development, there is a growing consensus that it can no longer be applied at the same pace as in the past, as the physical and economic limits of advanced manufacturing processes have become increasingly apparent. Consequently, there is a growing expectation that next-generation semiconductors—utilizing new materials and structures beyond traditional silicon (Si)-based semiconductors—will lead the industry. Let’s examine how this paradigm shift in the semiconductor industry is taking place through Thomas Kuhn’s ‘The Structure of Scientific Revolutions’.
‘The Structure of Scientific Revolutions’ is a seminal work that explains the structure of scientific progress. According to Kuhn, scientific progress proceeds sequentially through four stages. When a community of researchers in the same field accepts the most important theory underpinning their research as a single paradigm, science enters the stage of “normal science.” In the stage of normal science, the paradigm provides the rules and direction for research. These rules are not complete from the outset but become more refined through the process of solving various problems, offering researchers conceptual, theoretical, and methodological goals. Through this process, normal science continues to grow.
However, just when the existing paradigm is believed to perfectly explain all phenomena, anomalous phenomena emerge that are difficult to explain using existing theories. At first, these anomalies do not receive much attention, but as they gradually accumulate, they shake confidence in the existing paradigm and lead to a crisis. Kuhn describes this as the “crisis” stage. This does not mean, however, that the existing paradigm is immediately discarded. Researchers strive to explain these anomalous phenomena within the existing paradigm. However, as these anomalies recur and accumulate, a “scientific revolution” occurs in which a new paradigm replaces the old one. A paradigm shift brings changes not only to research methods but also to the values and worldview pursued by science, and ultimately, a new era of “normal science” begins, centered on the new paradigm.
According to Kuhn’s theory described above, the current semiconductor industry can be viewed as being in a transitional phase where the existing paradigm has reached its limits and is shifting toward a new paradigm. Let us then examine what the existing paradigm was and how we should overcome this crisis moving forward.
Before today’s semiconductors became established as a new paradigm, vacuum tubes were the core technology used in electronic circuits. A vacuum tube is an active device that controls current using electrons emitted by a heated cathode inside a glass tube under vacuum conditions, performing functions such as rectification, detection, and amplification. However, it had limitations such as large size, high power consumption, excessive heat generation, and low reliability. It was the transistor that solved these problems.
In 1947, John Bardeen, Walter Brattain, and William Shockley at Bell Labs in the United States developed the first transistor using germanium (Ge), ushering in the semiconductor era. Two significant scientific events subsequently drove the rapid advancement of semiconductor technology.
First, in 1954, Gordon Teal of Texas Instruments demonstrated that silicon (Si) was far more suitable as a transistor material than germanium (Ge). Silicon can form a stable natural oxide layer (SiO₂), and its excellent Si-SiO₂ interface properties made it highly suitable for planar technology and the fabrication of MOS (Metal-Oxide-Semiconductor) devices. These characteristics became the cornerstone of the development of modern integrated circuit (IC) technology.
Second, in 1958, Jack Kilby of Texas Instruments developed the modern concept of the integrated circuit (IC) by integrating resistors, capacitors, and transistors onto a single semiconductor substrate. Since then, integrated circuit technology has advanced rapidly, and electronic circuits using semiconductors have established themselves as a new paradigm that replaced vacuum tubes, becoming the center of mainstream science.
South Korea is one of the countries that has achieved the greatest success through this new paradigm of applying semiconductors to electrical circuits. In particular, South Korea has secured world-class competitiveness in the memory semiconductor sector and is leading the global semiconductor industry. However, despite its global competitiveness in memory semiconductors, the country still faces the challenge of competing with global leaders in areas such as system semiconductors, intellectual property (IP), semiconductor equipment, and software. Therefore, maintaining its strengths in the memory sector while securing competitiveness in system semiconductors and next-generation semiconductor fields has become a critical challenge.
Moore’s Law, long taken for granted in the semiconductor industry, has now reached its physical and economic limits. Currently, the line width of semiconductor integrated circuits has been scaled down to the level of a few nanometers (nm), and the costs of designing and manufacturing them have also increased significantly. Consequently, there is a shared understanding within the industry and academia that it is difficult to sustain the same level of performance improvements as in the past by simply continuing to reduce the size of transistors. Recently, alongside advancements in fine-scale processes, the focus of technological development has shifted toward advanced packaging, chiplets, 3D integration, and the utilization of new semiconductor structures and materials. In other words, “normal science”—centered on silicon-based CMOS semiconductors, which has prevailed for decades—is entering a new turning point.
Drawing on ‘The Structure of Scientific Revolutions’, we need to consider what direction engineers leading South Korea’s semiconductor industry should take to resolve this crisis phase. According to Kuhn, even when a crisis begins, researchers do not immediately abandon the existing paradigm of normal science. Applying this theory to the semiconductor industry, research aimed at further advancing existing silicon-based semiconductor technology, refining processes, and improving performance and productivity will likely continue for quite some time. In fact, research is currently underway to further advance existing CMOS technologies, such as gate-all-around (GAA) transistors, advanced packaging, high-bandwidth memory (HBM), and chiplet technology. However, since even these technologies will struggle to fully overcome physical and economic limitations in the long run, research toward a new paradigm is also being conducted simultaneously.
If these limitations persist, it is highly likely that a new semiconductor paradigm will emerge, just as vacuum tubes were replaced by transistors in the past. For South Korea, which currently leads the global semiconductor industry, this change may not be entirely welcome. This is because the country must maintain the competitive edge it has secured in the existing industry while simultaneously investing boldly in new technologies. However, if South Korea does not actively pursue the development of next-generation semiconductor technologies during this period, it is highly likely to remain limited to merely improving existing technologies and may miss the opportunity to lead new markets.
In fact, South Korean universities, research institutions, and companies have long been conducting various research projects aimed at overcoming the limitations of existing CMOS technology. “Beyond CMOS” is a general term for next-generation semiconductor technologies that go beyond the concept of conventional silicon CMOS devices and utilize new device structures and operating principles. This encompasses a wide range of technologies, including carbon nanotubes (CNTs), graphene, two-dimensional semiconductor materials, spintronics, neuromorphic semiconductors, and quantum devices, and is gaining attention as a new paradigm that will drive the future semiconductor industry.
If South Korea takes the lead in establishing a new paradigm to drive the global semiconductor industry based on next-generation technologies such as Beyond CMOS, while simultaneously continuing to strengthen the competitiveness of existing silicon-based semiconductor technologies, it will be able to establish itself as a nation spearheading a new “scientific revolution.” Only when efforts to steadily advance the existing paradigm are combined with the challenge of pioneering new paradigms will South Korea’s semiconductor industry be able to maintain its global competitiveness and continue to grow in the future.
