What's next for the semiconductor industry?

Bell Labs established the information age in 1947 when they demonstrated the first transistor ever made using germanium semiconductors. The arrival of the silicon transistor seven years later marked the beginning of Moore's Law, as it quickly emerged as the dominant material for integrated circuit (IC) technology.

According to Moore’s Law, the number of transistors that can be accommodated on an integrated circuit would double every year, extending this later to every 18 to 24 months. This translates to a doubling of microprocessor performance roughly every two years while simultaneously reducing chip costs. For over six decades, semiconductor development has followed Moore’s Law, contributing significantly to global economic growth, driving technological innovation, social transformation, productivity enhancements and economic expansion.

However, like most phenomena in the world, even the most remarkable eras eventually come to an end. While the future of Moore’s Law remains uncertain, with ongoing discussions about its slowing and eventual failure, the industry persists in developing methods to prolong its relevance.

From silicon-based to non-silicon-based, does chip material have to change again?

For over half a century, silicon-based semiconductor technology, particularly complementary metal–oxide–semiconductor (CMOS) IC technology, has driven profound changes in human society. Nevertheless, it is now approaching both its physical and engineering limits. Since the early 21st century, the pace of silicon-based chip development has been gradually slowing, signalling the advent of the post-Moore era.

However, at the same time, the demand for data processing and storage capacity continues to grow, prompting researchers and industry leaders to explore new materials and methods to extend and expand the principles behind Moore’s Law. This is driving scientists globally to search for viable alternatives to silicon.

Examining the materials that have the potential to continue Moore’s Law

Among the materials with the potential to sustain Moore’s Law, carbon-based semiconductors have received considerable attention due to their unique physical properties and extraordinary conductivity. Carbon nanotubes (CNTs) and graphene fall into this category and are considered as potential for future electronic, optoelectronic, and quantum devices.

Another notable alternative is compound semiconductors, which differ from elemental semiconductors such as silicon. These materials, composed of two or more elements, exhibit wide band gaps, high electron mobility and superior optoelectronic characteristics, making them well-suited for high-temperature, high-frequency, and optoelectronic applications. Notable examples include silicon carbide (SiC) (Figure 1), which have gained significant attention in recent years, as well as boron arsenide (BAs).

An onsemi M3S EliteSiC MOSFET, one of the SiC devices available from Mouser Electronics. Image credit: Mouser Electronics.

Additionally, organic semiconductor materials, which include conductive polymers and small-molecule organic semiconductors, are being used extensively in flexible electronics, optoelectronic devices and display technologies.

Other emerging materials are advancing notably in specific high-tech industries. Zinc oxide (ZnO) is increasingly valued for its role in creating transparent conductive oxides essential for Liquid crystal displays (LCDs) and photovoltaic solar cells, as well as serving as a photocatalyst for hydrogen production. Molybdenum disulfide (MoS₂) is another material gaining attention for its potential in semiconductor applications, where its exceptional thin-layer properties and high electron mobility are advantageous. Meanwhile, due to its high charge capacity and favourable electrical properties, black phosphorus is being explored for light-emitting diodes (LEDs) and photodetectors.

Based on the current development, carbon nanotubes may be the direction with the greatest potential to continue Moore’s Law. Carbon-based semiconductors could offer reduced power consumption and higher efficiency, making them a strong candidate for the next generation of transistor-integrated circuits. However, it is difficult to produce materials that meet the requirements of carbon nanotubes. This is a problem that scientists have been working hard to solve. As early as 2009, carbon-based nanomaterials were included in the International Technology Roadmap for Semiconductors (ITRS) as a future technology option, but at that time, the technology was too complicated to fabricate on a mass scale.

In recent years, key material challenges in semiconductor carbon nanotubes have been overcome, and the fabricated devices and circuits have demonstrated superior real-world electronic performance compared to silicon-based products. These breakthroughs bring the theoretical potential of carbon-based Integrated Circuits closer to reality and at a research level, demonstrate the potential competitive performance compared to conventional semiconductor technologies.

At the application level, carbon-based electronics show considerable potential in areas such as digital computing, radio-frequency electronics, sensing, 3D integrated circuits, specialised chips, display drivers and optoelectronic devices. Amongst these, carbon nanotube-based digital ICs (also known as carbon-based digital circuits) represent the most commercially and technically valuable avenue for carbon electronics. Recent advancements in this area mainly focuses on four aspects, including high-performance circuit exploration, low-power device innovation, complete digital logic function demonstration and large-scale integrated system research.

Quantum and light, a different approach?

Quantum and photonics technologies represent additional promising directions for the post-Moore era. Following decades of laboratory research and academic exploration, these fields are now approaching widespread commercial viability. In recent years, interest in using quantum and photonic chips to accelerate Artificial intelligence (AI) has surged, driven by the fundamental need to enhance computational power while reducing energy consumption.

Photonic devices

Recent research highlights the potential of optical technologies to overcome the inherent limitations of electronic chip design. Photons travel faster than electrons and exhibit low power consumption, minimal latency and resistance to fluctuations in temperature, electromagnetic fields and noise. Consequently, photonic chips are widely regarded as a key technology that can surpass Moore’s Law.

Historically, fibre-optic technologies have been predominantly utilised in communication systems, taking advantage of light’s superior speed and bandwidth for long-distance data transmission. However, advancements in optical computing, including the development of embedded co-packaged silicon-photonic waveguides, are broadening the scope of applications to encompass fields such as AI, biosensing and light detection and ranging (lidar). By harnessing light instead of electrons, photonic devices can meet the increasing demand for speed across numerous applications while significantly reducing energy consumption. This paradigm shift could result in more efficient and potent computing systems, overcoming the limitations of traditional electronic architectures and unlocking new speed and energy efficiency.

Quantum computing

Quantum computing, on the other hand, utilises counterintuitive quantum mechanical properties to accelerate certain types of calculations. In principle, quantum computers possess exceptional parallel processing capabilities, enabling exponential speedups for tasks such as machine learning (ML), cryptographic decryption, big data optimisation, materials discovery and drug analysis.

Despite this potential, the commercialisation of quantum computing faces engineering and material challenges. Quantum bits (qubits), the fundamental units of quantum computation, are highly susceptible to environmental interference. To achieve practical quantum computing, researchers must develop robust error correction techniques.

Recent progress in quantum computing is driven by improvements in qubit quality, error rates, scalability and substantial investments from major technology firms and start-ups. Governments worldwide, including those of the United States, the United Kingdom, China and Germany, have collectively invested billions of dollars in quantum research, and leading technology companies such as Google, IBM, Microsoft and Intel are all actively advancing quantum computing from fundamental research to engineering realisation.

Quantum computing research follows three main technological paths, including superconducting, ion-trap and optical quantum computing. The superconducting approach is the most competitive market; however, some researchers argue that optical quantum computing offers more potential.

Optical quantum chips utilise photons’ quantum properties for information processing and transmission. Light signals travel at 300,000 km/s through small circuits, and quantum properties such as superposition and entanglement enable computational capabilities that surpass those of classical computers. Achieving this requires integrating large numbers of photonic devices via lithographic techniques and precisely controlling them, but more fundamentally, manufacturing optical quantum computers entails a complete reinvention of chip mechanisms, fabrication processes and system integration.

Conclusion

The quest to surpass Moore’s Law is advancing on multiple fronts. Scientists are looking beyond silicon-based electronics for new possibilities. Whether it is through carbon-based electronics, quantum computing or photonic chips, these new technologies offer hope, but they also require extended periods of technological improvement.

In the era of traditional electronic chips, the world’s leading technology companies achieved prominence through decades of iterative development, large-scale industrialisation and application expansion, and the same will undoubtedly hold true for the next generation of semiconductor technologies.

Top image credit: iStock.com/PonyWang