Quantifying spatiotemporal impacts of the interaction of water scarcity and water use by the global semiconductor manufacturing industry

The high-tech semiconductor manufacturing sector is integral to the international electronics industry and was valued at over $400 billion USD in 2017. Intensive water use by this industry is well-documented and this work provides a spatially explicitly assessment of water use impacts by nearly 100% of global semiconductor manufacturing capacity. Both direct manufacturing water use and water use from electricity were considered as part of a facility's total withdrawal. Manufacturing water withdrawals were estimated by using technology-specific water and electricity use data, reported at the semiconductor chip or wafer level from the life cycle literature and industry estimates. Electricity water use intensity (WUI) factors were gleaned from the literature and regional electricity WUI factors were derived for China and the U.S. Geolocation of semiconductor manufacturing facilities allowed for summation of water withdrawals at various spatial extents (i.e. watershed, country, and globally). This data was combined with calculated regional or country-level electricity water use factors to estimate total water withdrawals by a facility. Geolocation of data also allowed for calculation of watershed specific scarcity-weighted withdrawals. Scarcity-weighted withdrawals were ascertained by multiplying facility water withdrawal data by the AWaRE water scarcity characterization factor available for each of 202 watersheds associated with semiconductor manufacturing facilities. These data were used to identify and map hotspots of industry water use, which is especially important for areas of industry growth such as China. This analysis is useful as a benchmark for global semiconductor industry water withdrawals and may assist OEMs in decisions about supply chain sourcing. This could also guide semiconductor manufacturers in prioritizing locations and time periods to implement water-saving technologies or employ less water intensive electricity sources. Additionally, the spatially explicit water use data for the semiconductor sector can be used to improve existing databases of national and regional sector-specific water use coefficients that are often applied in LCA input-output studies.