Information and communication technologies (ICT) are the fuel of the digital society. This article reviews the big challenges to make ICT (more) circular.

Sustainable materials and production

by Thomas Ernst, Lars-Åke Ragnarsson and Jean-Pierre Raskin

Europe is facing one of the largest challenges in its history: that of preserving a viable environment for the decades to come. On human consciousness of its environmental impact, Albert Allen Bartlett [1] said: “The greatest shortcoming of the human race is our inability to understand the exponential function”. ICT was, and still is, driven by exponential functions (Moore’s law, Cooper's law, Koomey’s law, data exchange volume worldwide, …). However, each exponential has its limits which are often reached sooner than expected as discussed further below.

Sustainability was defined by United Nations with a very broad vision, to promote prosperity while protecting the planet [2].

We will focus here on environmental aspects taken into account in the products Life Cycle Assessment (LCA) standards: resource consumption, emissions to the air, water, impacts on ecosystem, human health, and natural resources. The LCA goal is to guide product design and production to limit its impact. In ICT history to date, when some limits are reached, they are overcome by a technology breakthrough. By way of example, there was the CMOS (complementary metal-oxide-semiconductor) technology, a product of laboratory curiosity in the 1980’s when bipolar transistor power consumption was too high. New integrated systems (multi-core processors) and device (SOI, FinFET, …) architectures entered into the market when energy consumption limits were reached by central processing units (CPUs). Today, CMOS is at the core of ICT and is the cheapest and most reliably power-efficient technology for the development and growth of the internet of things (IoT).

Awareness of the rapid proliferation of wirelessly connected objects around us gives rise to questions about their environmental impact. There is a need to establish an ambitious roadmap, at research and industrial level, for moving into a virtuous cycle of eco-innovation based on precise environmental evaluations.

In this article, we propose pathways to maintain or even reduce global use of energy and critical materials for ICT.

Key insights

• ICT is energy and resource demanding and contributed to approximately 1,200 – 2,200 MtCO₂ eq in 2020, or equivalently 2.1%–3.9% of the global world greenhouse gas (GHG) emissions [3].

• Manufacturing of integrated circuits is greatly energy and material demanding and generates approximately 30% of the emissions of mobile ICT devices [3].

• The manufacturing of ICT uses many critical materials and Europe imports nearly all of them [4].

• Recycling of ICT is technically difficult, consumes energy, generates pollution, and is not currently profitable because of the low price of primary raw materials. Only 17.4% of electronics waste is recycled [5].

• With an economic model more centred around service (functionality economy) than the object itself, reparability will be more encouraged.

• The “rebound effect” could often cancel out the efficiency gains of ICT systems.

• It is time to design differently, to design within limits.

Key recommendations

• Expand the lifetime of devices through more reliable design at the system level, both by enhancing the intrinsic durability of components (batteries, screen, memories, …) and by adopting a modular approach in which the replacement of faulty or obsolete components is made easy.

• Rethink software and applications to make them less resource consuming (data, energy, hardware) and compatible with still usable hardware. Make software reconfigurable and evolutive.

• Integrate circular economy concepts, eco-design and full life cycle assessment (including fabrication, use and end-of-first-life phases) at the early stage of research and development of new ICT technologies.

• Promote research into alternative low global warming potential gases and chemical solutions. Abatement is a proven solution to reduce the impact of greenhouse gases used by the semiconductor manufacturers. Support its further adaptation, optimized maintenance and improvement.

• Avoid toxic materials where viable.

• As required by law, employ/assure that chemical processes are protective of human health and the environment.

• Develop manufacturing processes that facilitate materials recycling and low energy use.

• Develop products that are amenable to recycling at the end of life.

• Use secondary (recycled) materials, more renewable energy and develop bio-based and greener chemistry in the ICT industry. Develop and promote fabrication solutions which use energy and materials more efficiently. This requires enhanced basic and applied research into materials and fabrication processes.

• Develop new economic models in which externalities of any product or service are properly accounted for and encouraging the manufacturer to make its best effort to decrease the environmental impact of its technologies. ICT solutions must be developed and implemented in such a way that the economic profitability is (at least partially) decoupled from the depletion of natural resources.

An increasing impact of ICT growth on natural resources

The digital society we live in has never been so energy- and material-intensive and this is leading to increasing pressure on natural resources, ecosystems and the climate. Today, ICT is a major economic activity and its impact is growing, similarly to transport, energy production, manufacturing, and agriculture. Information and communication technologies including TV’s and other consumer electronics consume around 5% of the world's electricity production and were responsible for 2.1%–3.9% of the world greenhouse gas (GHG) emissions in 2020 [3], a level equivalent to air transport. Our smartphone contains electronic circuits that require more than sixty different materials. We are talking about virtually every element on Mendeleev's periodic table except radioactive materials. At the end of their life, the recycling rate of electronic equipment is very low (17.4%) [5].

It is extremely difficult to separate the sixty materials that make up electronic circuits [6]. Umicore, one of the most advanced companies in the field of electronic materials recycling, manages to extract 17 elements out of the 60 [6]. Material recovery from obsolete equipment is not profitable given the low cost of raw materials imported from the so-called global South. The environmental and social costs are not fully accounted for in the price of the final product. This leads to a double penalty for the global south countries: they suffer from environmental pollution (loss of biodiversity, pollution of the air and groundwater, etc.) during the extraction of raw materials, as well as receiving 75% of our electronic waste [7].

Figure Estimation of distribution of the energy consumption of digital technologies for production (45%) and use (55%) Source: The Shift Project, 2019 [8]

Toward efficient integrated circuits production

The manufacturing of electronic devices used in the ICT industry is very complicated and requires large amounts of energy and materials (Figure 1). Especially the manufacturing of integrated circuits is an extremely complex and energy intensive endeavour done in very large fabs using hundreds of processing tools installed in clean room environment. IC manufacturing involves many types of complex process steps which use very high purity (also referred to as semiconductor grade) gases, chemicals, precursors, metals, and water. Some of the process steps use fluorinated gases which have very high global warming potential (orders of magnitude higher than CO₂) and therefore need to be abated to minimize the CO₂ equivalent emissions of a fab. The water used in wet cleaning processes needs to be ultrapure to minimize added particles and contaminants which could reduce yield and device performance. Finally, the production of the semiconductor grade materials required by the fab use complex and high energy demanding manufacturing processes themselves.

As a result, for mobile applications and devices, it is estimated that the manufacturing of IC’s contributes significantly to the embodied emissions of a device. It is estimated that around 75% of the total Carbon footprint of mobile devices are from manufacturing, and about 50% of that (or ~40% of the total) is from the IC manufacturing of the many IC’s [8].

The relentless efforts to develop denser memories and logic circuits continue to provide more functionality per chip area. However, this comes at a cost in process complexity and an increase in fabrication steps. Furthermore, the new applications are becoming more sophisticated and require higher complexity at the system architecture level [9]. This results in an increase in the number of interconnect (metal) levels. As a result, more advanced semiconductor technologies have significantly higher embodied energy and associated emissions than former technologies for a chip with the same area. Generally speaking, newer technologies use more electrical energy, more raw materials, more chemicals and water. Figure 2 shows the evolution of the emissions related to process gases (Scope 1) and electricity (Scope 2) from the 28 nm to 5 nm estimated using a bottom-up virtual fab model [10]. The technology nodes are mapped approximately onto their year of introduction. For these technology nodes, the Scope 1 and 2 emissions, normalized per wafer, have increased by approximately 2x in the past 10 years.

Interesting to note is that when normalizing the emissions to the size of a function (e.g. a GB of memory, or a logic gate) the emissions are reducing. In other words, for a fixed function, a more advanced technology has lower emissions. However, due to the demand at the system level, this benefit rarely makes it to the final device. Instead, the improved scaling is used to create ever higher complex circuits and systems which end up increasing the total fabrication energy (rebound effect).

To enable these advances in technology scaling, inside the fab there are process steps that produce extremely high-quality materials and well-defined features with high aspect ratios and sub-nanometer precision. These processes require complex equipment using high purity process chemicals and gases. To keep the cost of manufacturing as low as possible, the efficiency of these processes is optimized as part of the technology development and maintenance. However, some processes are quite inefficient due to technical and process limitations. For example, the deposition efficiency using chemical vapor deposition (CVD) techniques of a material on the targeted wafer ranges from 1 to 20% [11]. This means that most (>80%) of the input high purity process gases is passing through the chamber to the exhaust without being actually deposited.

Furthermore, to enable the high density of advanced technologies there are many process steps which rely on the use of sacrificial layers. A first classical example is the photoresist used to transfer patterns from a mask (reticle) to the Si wafer surface. After development, the pattern is transferred to the surface by etching and then the resist is removed. Secondly, the deposition of “hard-masks” is introduced in many process flows nowadays to reach denser pattern densities when resist alone is not adequate. These hard-masks may include metals. Third, the definition of many three-dimensional structures (for example 3D NAND) relies on deposition of stacks of materials which are later etched and replaced by other materials through complex integration schemes. Finally, to enable high density lithography it is essential that the Si wafer remains flat throughout the manufacturing. This requires that material deposition filling small features will need to overflow and cover the full wafer and will be polished at the end using chemical mechanical polishing (CMP) resulting in further materials losses.

Due to these practices it is important to understand that the amount of material that ends up in the final chip is small compared with the amount of silicon, while the manufacturing processes require and waste significant amount of materials other than silicon. This puts some significant constraints on the materials circularity of IC manufacturing: It needs to happen inside the fab.

Efforts to reduce power consumption and industrial waste during the manufacturing of electronic devices and their components have been made since the beginning of high-volume IC manufacturing. This work continues with for example, low temperature processes, reduction of heat dissipation in and from the oven, and reduction and recycling of chemicals and water. To take care of the direct emissions impact from fluorinated process gases, efficient abatement solutions have been introduced and are progressively used in the industry. There are also significant efforts towards finding replacement gases to lower the impact even further [12].

There are also good examples of circularity in the fab. Many manufacturers are today reusing their ultra-pure water [13] thus reducing their impact in regions where water scarcity is a concern. Other examples include recovery of copper from CMP slurry waste [14], recovery of H₂ used by extreme ultra-violet (EUV) scanners [15] and recirculation of hot water in cleaning processes [16].

Fab circularity and low-waste production

Beyond resource efficiency and recycling waste, the transition towards a cost-effective circular economy needs to be implemented as well as the general design methodologies of materials for sustainable development proposed by Ashby [17] adapted to specific ICT domains such microelectronics.

Some companies began to adopt the lifecycle analysis years ago [18], for example, in the production and recycling of the ultra-pure water needed [13] for the microelectronics industry. There are active research programs, both on the part of equipment makers and in research labs, seeking to reuse exhausted gases or fluids within the fab [19], and also to develop much more efficient material deposition techniques [11].

The reduction of toxic chemicals in the semiconductor manufacturing fabs is investigated through different methods. As an example, biowaste-based [20] chemicals and materials may be used to reduce significantly the use of solvents and chemicals in lithography, as suggested by preliminary results [21]; more research efforts are required. Bio-based materials are also being investigated for use in packaging in the ICT domain.

More circularity between companies means developing supply chains in which the by-products or wastes from one industry could be the supply material for another. For instance, the hafnium required in CMOS production is a by-product of ultrapure zirconium used by the nuclear industry and produced in several countries including France and the United States [22].

To further reduce the waste of energy and materials, the ICT industry must adopt a holistic approach to developing sustainable products. Several initiatives already exist in the private sector. As an example, we can point out the significant and long-term efforts of several companies such as ST-Microelectronics which evaluates (by life cycle assessment) the carbon footprint of their microcontrollers [18] [23] and established a clear material declaration available online.

To envision a more sustainable future, Europe must:

• Take actions to make the ICT supply chain more transparent in collaboration with European Semiconductor Industry Association (ESIA);

• Systematically make life cycle assessment and declarations of materials (including for imported products) with shared methodology worldwide;

• Implement a clear and ambitious plan to maximize ICT product lifetime and anticipate its end of life.

This will encourage both research and industry sectors to innovate for the good of everyone.

Looking at minerals

The electronic industry needs a wide variety of minerals. For example, since the 1970’s the silicon-based complementary metal-oxide-semiconductor (CMOS) field-effect transistor (FET) has been the mainstream technology for most transistor applications, thus making today’s digital economy possible. Over the years, the number of elements used in their manufacture has increased greatly (Figure 3), especially since 2000 with the implementation of high-k dielectrics and metal gate stacks which are essential to minimize short-channel effects and gate-leakage current of short transistors (today gate length shorter than 20 nm).

A growing awareness of the limited nature of the supplies of some elements that have specialized and important uses is reflected in the proliferation of terms to describe them and the ores from which they are derived, including “gateway minerals” and “critical” or “endangered” elements. Some countries have adopted policies recognizing the high strategic importance of some of these for their physical and economic security.

In 2010, 14 elements were considered as critical by the European Commission (EC) according to both their strategic importance for future technology and their scarcity, while in 2023 the number rose to 34 [18].

Modern devices and systems rely heavily on a high degree of control of material properties and a mastery of manufacturing techniques and, to date, the ICT industry has been remarkably successful in fulfilling these needs.

The semiconductor manufacturing process is a “top-down” or “subtractive” one based on UV photolithography, etching and many sequential, highly organized and efficient steps of chemical and physical treatment of the chip, layer after layer.

Work is being undertaken on substituting or decreasing the use of hazardous and critical raw materials. For currently crucial elements such as indium, ruthenium, platinum, gallium, arsenic and gold, new technologies and materials are being investigated with a view to replacing them or drastically limiting their use in some critical devices (e.g. in sensors, memories, optoelectronics and spintronics) [25]. For instance, gold wire bonding has been replaced in IC packaging when possible and replaced by other copper-based techniques like copper through silicon vias (TSV).

Other examples of the move towards sustainable electronics include the avoidance of lead in micro-components like actuators included in cell phones, use of 2D mono-atomic or ultra-thin atomic-deposition layers to reduce the use of some active materials by a factor of up to 10⁶, and use of silicon-based substrates such as silicon-on-insulator (SOI), instead of materials made from combinations of group III and group V elements, for radiofrequency (RF) technologies [26].

The scaling down of dimensions of high-tech devices in recent decades and the multiplication of materials in the components – some of them in extremely small quantities of a few micrograms – are leading to new challenges in recycling. The need for large amounts of power and the use of aggressive acids and solvents can make recycling of such electronics impractical [6].

Approaches to increasing the sustainability of microelectronic devices must include extending their lifetime through better design, by both enhancing the intrinsic durability and reliability of components and adopting a modular approach in which replacement of faulty or obsolete components is made easy.

These approaches can draw on the experience of, for example, some European microelectronics manufacturers and R&D laboratories (e.g. On-Semi, X-FAB, Infineon, STMicroelectronics, NXP, GlobalFoundries) that are designing or fabricating highly reliable components for automotive, energy management and security applications.

Steps towards the ambitious goal of achieving the sustainability of the physical layer of the digital society will require concerted actions covering a range of interlocked approaches. These will address the entire lifecycle of not only the digital devices themselves but also the services that support them, paying attention to their energy and environmental footprints as well as economy and efficiency in the utilization of resources.

Conclusion

The development of a more sustainable semiconductor industry is quite challenging and not only at the technical level. A mindset change has to be induced not only in the industry but also at the research centre organizations and in the academia. Sustainability criteria must be defined at the early stage of development of any research project, product or service and their have to guide the choices of each stakeholder only the entire supply chain. There are encouraging signs such as the multiplication of policies at the European level to guide and support concrete actions for developing greener electronics, and strong engagement of several semiconductors companies to minimize their environmental footprint. In order to drastically reduce the ICT e-waste growth we have to multiply and disseminate the good practices in term of eco-design and circular economy.

Acknowledgement

The present article summarizes the main discussion outcomes from a panel of experts in the fields of geology, materials science, micro and nanoelectronics fabrication process, electronics circuits design, electronics packaging, supply chain management, and IT systems: Mathilde Billaud, Fraunhofer IZM, David Bol, UCLouvain, Thierry Baron, CNRS, Patrice Christmann, BRGM, Marie Garcia-Bardon, imec, Léa Di Cioccio, CEA-Leti, Laurent Pain, CEA-Leti, Bertrand Parvais, imec, Cédric Rolin, imec, Karine Samuel, UGA, Lutz Stobbe, Fraunhofer IZM.

AUTHORS

Thomas Ernst is a chief scientist at CEA-Leti, France

Lars-Åke Ragnarsson is a program director at imec, Belgium

Jean-Pierre Raskin is a professor at the Université Catholique de Louvain, Belgium

REFERENCES

[1]: A. Bartlett. [Online]. Available: https://www.albartlett.org/.
[2]: "www.un.org," [Online]. Available: https://www.un.org/sustainabledevelopment/.
[3]: C. Freitag, M. Berners-Lee, K. Widdicks, B. Knowles, G. S. Blair and A. Friday, "The real climate and transformative impact of ICT: A critique of estimates, trends, and regulations," Patterns, vol. 2, no. 9, 2021.
[4]: "Critical raw materials - European Commission," [Online]. Available: https://single-market-economy.ec.europa.eu/sectors/raw-materials/areas-specific-interest/critical-raw-materials_en.
[5]: World Health Organization, "Electronic waste," 18 10 2023. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/electronic-waste-(e-waste). [Accessed 20 12 2023].
[6]: C. Hagelüken, "Recycling of Electronic Scrap at Umicore. Precious Metals Refining," Acta Metallurgica Slovaca, vol. 12, pp. 111-120, 2006.
[7]: A. Shamim, A. Mursheda and I. Rafiq, "E-Waste Trading Impact on Public Health and Ecosystem Services in Developing Countries," International Journal of Waste Resources, vol. 5:4, 2015.
[8]: U. Gupta, Y. G. Kim, S. Lee, J. Tse, H.-H. S. Lee, G.-Y. Wei, D. Brooks and C.-J. Wu, "Chasing Carbon: The Elusive Environmental Footprint of Computing," 2020. [Online]. Available: https://doi.org/10.48550/arXiv.2011.02839.
[9]: M. Koyanagi, "System Integration Technology Based on 3D Integration," in IEEE International Meeting for Future of Electron Devices, Kansai (IMFEDK), Kyoto, Japan, 2022.
[10]: "netzero.imec-int.com," imec, [Online]. Available: https://netzero.imec-int.com.
[11]: H. Pedersen, N. T. Barry and J. Sundqvist, "Green CVD—Toward a sustainable philosophy for thin film deposition by chemical vapor deposition," J. Vac. Sci. Technol. A, vol. 39, no. 5, p. 051001, 2021.
[12]: C. Jennings, P. Nguyen, S. Biltek and N. Stafford, "Advances in low GWP etch gasses," in Proc. SPIE 12499, Advanced Etch Technology and Process Integration for Nanopatterning XII, 124990G, 2023.
[13]: W. Den, C.-H. Chen and Y.-C. Luo, "Revisiting the water-use efficiency performance for microelectronics manufacturing facilities: Using Taiwan’s Science Parks as a case study," Water-Energy Nexus 1, pp. 116-133, 2018.
[14]: X. Wang, G. Buer, W. Fan, L. Gao and M. Huo, "Copper removal from semiconductor CMP wastewater in the presence of nano-SiO2 through biosorption," Water Reuse, vol. 11, no. 2, pp. 289-300, June 2021.
[15]: C. Bailey, "Hydrogen Recovery in EUV Lithography," Semiconductor Digest, Vols. 23-26, April/May 2023.
[16]: J. Wang, C.-T. . Ni, S. Ku and A. Tzeng, " TSMC Develops Hot DI Water Circulation System 2.0 for Wafer Cleaning Tool, Saving 380,000 kWh/yr of Electricity," TSMC, [Online]. Available: https://esg.tsmc.com/en/update/greenManufacturing/caseStudy/67.
[17]: M. F. Ashby, Materials and Sustainable Development 2nd edition, Butterworth-Heinemann, 2022.
[18]: "STMicroelectronics, “Footprint of a ultralowpower MCU," [Online]. Available: https://www.st.com/content/st_com/en/about/st_approach_to_sustainability/sustainability-priorities/sustainable-technology/eco-design/footprint-of-an-ultra-low-power-mcu.html.
[19]: V. Revankar and H. S. Lahoti, "CVD-semens monosilane reactor process with complete utilization of feed gases and total recycle”, patent US 8,657,958 B2, 2014.". Patent US 8,657,958 B2, 2014.
[20]: M. Caillau, P. Crémillieu, E. Laurenceau, Y. Chevolot and J.-L. Leclercq, "Fifty nanometer lines patterned into silica using water developable chitosan bioresist and electron beam lithography," J. Vac. Sci. Technol. B, vol. 35, p. 06GE01 , 2017.
[21]: I. Servin, A. Teolis, A. Bazin, P. Durin, O. Sysova, C. Gablin, B. Saudet, D. Leonard, O. Soppera, J.-L. Leclercq, Y. Chevolot, R. Tiron and T. Delair, "Water-soluble bio-sourced resists for DUV lithography in a 200/300 mm pilot line environment," Micro and Nano Engineering, vol. 19, 2023.
[22]: C. T. Matos et al., "Material System Analysis of Nine Raw Materials: Barytes, Bismuth, Hafnium, Helium, Natural Rubber, Phosphorus, Scandium, Tantalum and Vanadium," 2021. [Online]. Available: https://doi.org/10.2760/677981.
[23]: T. T. Pirson, T. P. Delhaye, A. G. Pip, G. Le Brun, J.-P. Raskin and D. Bol, "The environmental footprint of IC production: review, analysis, and lessons from historical trends," IEEE Transactions on Semiconductor Manufacturing, vol. 36, no. 1, pp. 56-67, Fe, vol. 36, no. 1, pp. 56-67, 2023.
[24]: S. Dhong and J. P. Colinge, "FinFET to Nanowire Transistor," in DAC 2014 Tutorial, 2014.
[25]: J.-P. Raskin, "No Digital Society without Sustainable Communication Technologies," in Workshop Toward sustainable IoT from rare materials to big data at the 50th European Solid-State Device Conference – ESSDERC 2020, Grenoble, France, 2020.
[26]: J. Raskin, "FinFET and UTBB for RF SOI communication systems," Solid-State Electronics, vol. 125, p. 73–81, 2016.
[27]: H. Ferrboeuf, "'Lean ICT towards digital sobriety' and 'Lean ICT Materials'," 2019. [Online]. Available: https://theshiftproject.org/en/article/lean-ict-our-new-report/.