Life cycle assessment of emerging technologies on value recovery from hard disk drives

Abstract

Value recovery from end-of-life hard disk drives (HDDs) is a promising strategy to promote a circular economy due to the valuable material content, large availability, and regulated disposal for data security. However, current business practices are limited to reuse or shredding of HDDs for base metal recovery, which are not optimal for environmental sustainability. In particular, the neodymium-iron-boron (NdFeB) magnets, which contain rare earth elements (REEs) that have significant environmental impacts and supply chain risks, are not recovered. To address this challenge, technologies are emerging to enable REE recovery from HDDs. This work focuses on comparing novel technologies, including direct reuse of magnet assemblies, magnet-to-magnet recycling, and recovery of REEs. Life cycle assessment (LCA) was performed on each recovery pathway to quantify and compare the environmental impacts. Primary data were collected from Seagate and other key stakeholders in the HDD value chain. LCA results showed that reusing HDDs is the most environmentally friendly option in terms of global warming potential, reducing CO₂ eq. emissions by 5∼18 kg per drive life cycle, when compared to the virgin production and shredding for aluminum recovery. Reuse of magnet assembly is the next best option (reducing ∼1.9 kg of CO₂ eq. emissions), followed by magnet-to-magnet recycling (∼1.2 kg of CO₂ eq. emissions), and metal recycling (∼0.02 kg of CO₂ eq. emissions). Environmental hotspots were also identified, revealing the significant contribution of intercontinental transportation mode for the reverse logistics. Future research is suggested on optimizing the reverse supply chain to reduce the environmental footprint of HDD value recovery.

Introduction

Hard disk drives (HDDs) contain precious metals, such as gold and silver, as well as rare earth elements (REEs) that are critical to clean energy and defense products (Chu, 2011; Eggert et al., 2016; Jin et al., 2016). HDD value recovery is proposed to be the most feasible pathway for recycling REEs including neodymium (Sprecher et al., 2014a). HDDs are a good candidate for value recovery due to high availability (e.g., a billion units of end-of-life (EOL) HDDs generated globally in a year (Sabbaghi et al., 2018)), regulated disposal in the U.S. for data security reasons (BakerHostetler, 2018), and fairly constant form factors of 2.5” and 3.5” that favor reuse and recycling (iNEMI, 2017). Currently, the United States relies on 100% importation of high-purity rare earth metals (USGS, 2018) and generates the largest share of global used HDDs (i.e., 16.9%) (Sabbaghi et al., 2018). The collection rate from data centers is high (90–95%) (Nguyen et al., 2017; Sprecher et al., 2014a) compared to residential sources (26%) (iNEMI, 2017). Therefore, this work focuses on EOL collection of enterprise HDDs in the United States.

HDD value recovery options are constrained by the drive condition (e.g., product or technology obsolescence, quality degradation, and hardware or software failure), thus a variety of feasible recovery pathways were examined, as shown in Fig. 1. New technologies have been emerging that expand the current business practices of limited reuse and massive shredding of entire HDDs (Habib et al., 2015; Rana and Brandt, 2016). These pathways include direct reuse of magnet assemblies (MAs) which was demonstrated for six HDDs (iNEMI, 2018), magnet-to-magnet recycling being commercialized at Urban Mining Company (UMC, 2019), and hydrometallurgical, pyrometallurgical, and electrochemical recovery of precious metals, base metals, and REEs that are being developed at lab scales for inclusion of REEs in the metal recovery process (Diaz and Lister, 2018; Lister et al., 2016). In addition to the development of these novel technologies, the existing HDD reuse ratio could be improved by relaxing stringent company policies of shredding fully functional HDDs for data security reasons (Neil Peters-Michaud, 2017).

Adverse environmental impacts from mining and processing of bastnäsite, monazite, or ion adsorption clays to produce REEs are well documented (Navarro and Zhao, 2014; Sprecher et al., 2014b; Vahidi et al., 2016). These impacts are especially concerning because it is estimated that 30% of global REEs are mined illegally (Packey and Kingsnorth, 2016), resulting in disastrous consequences for the environment and public health. HDD value recovery options targeting circular strategies to close material loops are a promising alternative, but may also have associated negative environmental impacts. Value recovery requires transportation, mechanical and/or chemical treatments that create environmental burden. For example, rare earth oxides recovered from EOL HDDs in the U.S. need to be sent to Asia, as there is no downstream processing facilities in the U.S., increasing the transportation related impacts. Such additional impacts are more apparent if we shift from HDD reuse to shredding through which the embedded energy and functions are lost, and more resource inputs are required to recover less value (Fig. 1). The expectation of environmental impact reduction from circular strategies underlying these value recovery options must be examined at a systems level to avoid burden shifts, or at least identify and acknowledge trade-offs. For this purpose, life cycle assessment (LCA) was conducted to quantify and compare the environmental impacts from the value recovery options with those created by current HDD production and EOL practices, namely shredding for base metal recovery. The LCA results may also help remove two major barriers to implementation of a circular economy: lack of awareness of the environmental benefits and hesitant company culture to implement new recovery pathways (Kirchherr et al., 2017).

The existing LCA literature focuses on various value recovery options for WEEE (Waste Electrical and Electronic Equipment) such as cellphones, computers, and monitors, but does not examine pathways for enterprise HDD recovery, considering the supply chain feasibility. For example, Jin et al. (2018) demonstrated the environmental benefits of NdFeB magnet-to-magnet recycling over virgin production for electric vehicle motors. However, the comparison was based on 1kg of magnets than per feedstock availability (e.g., magnets from one HDD may not be recycled to satisfy 100% of magnet demand for another HDD due to material losses), and the system boundary ended at the gate of the recycling plant without considering the next steps for closing the material loop. Lu et al. (2014) showed the environmental benefits of reusing mobile phone components (e.g., integrated circuits and cameras) over material recovery. However, the specific LCA data (e.g., unit processes), data sources, and allocation methods (e.g., EOL credit) were not disclosed, which makes it challenging for others to repeat the analysis and validate the results. Zink et al. (2014) compared the environmental impacts of refurbishing smartphones with those of repurposing into parking meters. As the LCA results relied on an LCA report from 2005 (Singhal, 2005) to estimate the production impacts of an old Nokia cell phone, the impacts might have been underestimated for new smartphones, and the impact categories were limited to five that were considered in the LCA report. (Sahni et al., 2010) evaluated the energy savings from reusing personal computer devices over purchasing new ones and found that the energy savings could be negative in certain scenarios due to improved energy efficiencies of new devices (e.g., reusing a 2005 desktop with CRT monitor was worse than purchasing a new Energy Star certified laptop of 2009). Again, the unit processes were not disclosed, and data were collected from various secondary sources. (Yung et al., 2011) conducted LCA for eco-design of electronics using primary data sourced from a company (e.g., bill of materials); however, no life cycle data was disclosed, and only disassemblability and recyclability were considered at the end-of-life. Barba-Gutiérrez et al. (2008) showed that under certain circumstances (e.g., 500 km of transportation distance for EOL collection), the environmental impact of recycling WEEE (i.e., washing machines, refrigerators, TV sets, and personal computers) could be higher than the impact of landfill. Again, the unit process data were not disclosed as well as the allocation methods (e.g., how environmental credits, if any, were assigned to the collected or recycled WEEE) that would have significant impact on the LCA results. Hong et al. (2015) conducted LCA on e-waste recycling with waste disposal (i.e., incineration and sanitary landfill) and without disposal (i.e., open dumping) in China and found that the common environmental hotspot was electricity consumption (i.e., 145.45 kWh per ton of e-waste). However, the quality of recovered materials (i.e., plastics, glass, and metals) were not discussed, and the associated environmental benefits were excluded in LCA. Xue et al. (2015) quantified the environmental impacts of recycling printed wiring boards (PWBs) in China and found that metal leaching was the most critical process followed by power source, transportation mode and distance for waste collection. However, the study relied on LCA data from multiple secondary sources (e.g., elemental composition of PWB from Park and Fray (2009) and leaching efficiencies from Chehade et al. (2012)). As such, lack of primary data, disclosure of unit processes and allocation methods, and inclusion of environmental benefits from recovered products/components/materials are critical issues with the existing literature, and our study aims to address these limitations for LCA of HDD value recovery.