Developing an overall CO2 footprint for semiconductor products

Understanding the total life cycle CO
₂
impacts of different types of products and operations has been a subject of significant interest recently. Total life cycle impacts include not just the CO
₂
generated from the manufacture of a product, but also from the product's use and from the manufacture of all the inputs that go into creating the product. In theory, understanding this total impact can help manufacturers identify means of reducing the impact and could gives consumers information that can be used as a consideration in purchase decisions. Unfortunately, there are to date no broadly accepted and consistent methods for calculating this impact. Determining the boundaries that define which inputs should be included in the assessment is a significant challenge. In addition, supply chains for manufactured products can be highly complicated, involving hundreds or thousands of individual suppliers, all of which have a large number of end customers. Few if any of these suppliers have the ability to collect energy use and CO
₂
impact data in a way that breaks the impact down by product or customer. This is especially true with complex products like semiconductors, making it very difficult to allocate the CO
₂
from the supply chain in a way that a customer such as Intel can estimate its portion of the impact from the manufacture of a material that is used by many other customers as well. A number of life cycle analysis studies have been completed on semiconductors and electronics in recent years. However, the term "life cycle analysis" is broad enough that a wide range of potential focus areas and specific environmental impacts can be evaluated in any given study. Some studies reviewed attempted to make comparisons between different products while others tried to identify the portion of a manufacturing process with the largest impact or the portion of a product's life cycle with the largest impact (e.g. manufacture, use, disposal, etc.). In addition, each study had a different environmental impact metric, or range of metrics that they focused on. Some focused on energy, water or raw material use, others on environmental impacts associated with pollutant discharge (e.g. acidification, eutrophication). Intel was specifically interested in quantifying the total life cycle CO
₂
impact of our operations and products, including as many of the "indirect" impacts that could be quantified. We also wished to define a methodology and set of boundaries that could be useful in establishing consistent calculation of these impacts in the future. Intel had long been reporting its scope 1 and 2 greenhouse gas emissions publicly, but like many companies had an incomplete understanding of scope 3 emissions. Scope 1 and 2 emissions refer to direct emissions at the facility and purchased electricity. Scope 3 emissions, according to internationally accepted protocols, are indirect sources that are not directly attributed to a company's manufacturing operations, but that are nonetheless created as a result of demands that the company's operations are creating, or the use of its products. Supply chain impacts were a specific focus due to interest from customers and others, and also because preliminary assessments suggested that this was likely to be a much larger scope 3 source than some less significant ones such as employee commute and travel. Given the large number of inputs in semiconductor manufacturing, the supply chain impact appeared to have potential to be significant. To assess this impact, data was collected quantifying the amount of input chemical and materials Intel's manufacturing operations use in a year. A project was then initiated with a third party consultant to compile information on the total CO
₂
equivalent impact (including the impacts from energy required and emissions of any non-CO
₂
global warming gases) to produce a fixed quantity of each substance. The data was taken from previous research on this subject from the consultant and other sources. Information was collected on more than 90% by mass of the raw material input to the manufacturing process. This was then combined with the information on material quantities to produce an estimate of the total CO
₂
impact of Intel's supply chain. This paper will show the relative CO
₂
impacts of the various aspects of semiconductors including manufacturing operations, product use, and various scope 3 impacts such as supply chain, logistics and employee travel. It will also present what Intel believes to be a realistic definition of boundaries for defining scope 3 impacts and procedures for performing this type of assessment. The paper will also include discussion on potential focus areas for reducing impact based on the study, and will identify remaining data gaps and areas for future study.