Regional economic potential for recycling consumer waste electronics in the United States

Waste electronics are a growing environmental concern but also contain materials of great economic value. If properly recycled, waste electronics could enhance the sustainability of vital metal supply chains by offsetting the increasing demand for virgin mining. However, rapid changes in the size and composition of electronics complicate their end-of-life management. Here we couple material flow and geospatial analyses on over 90 critical consumer electronic products and find that over 1 billion devices, representing up to 1.5 million tonnes of mass, could be discarded annually in the United States by 2033. Emerging electronics such as connected home, health and augmented/virtual reality devices have become the fastest-growing types in the waste stream. We highlight policy opportunities to develop various sustainable circularity strategies around metal supply chains by showing the potential to integrate waste electronics and virgin mining pathways in western US regions, while new infrastructure designed specifically for waste electronics treatment is favourable in the central and eastern United States. Furthermore, we show the importance of building national-level refining and tear-down databases to improve electronics end-of-life management in the next decade. Electronic waste can harm the environment when not properly disposed of, but as a source of valuable materials, it could benefit the economy with adequate end-of-life management. This study shows the scale of the problem and assesses the potential of diverse management solutions across the United States.

Due to the growing consumption of electronic equipment and the relatively short lifetimes of many of these devices, the amount of discarded electronics in the waste streams (waste electronics) is rapidly increasing. Globally, the annual waste electronics generated has grown from less than 25 Mt per year in 2009 to around 53.6 Mt in 2019 [1][2][3] . The amount of waste electronics generated in the United States in 2019 was 6.9 Mt (approximately 12.9% of the world generation), corresponding to a generation per capita of around 21 kg, which is approximately three times the world's average value (7.3 kg) 3 .
Only 17% of the waste electronics generated globally is properly recycled, and the recycling rate in the United States is around 14% to 40% depending on the reporting agency and the selected types of electronics 3,4 . Domestic waste electronics that are exported for recycling are hard to regulate, owing to the different levels of maturity in terms of recycling technology and policy incentives between countries 5,6 .
Due to the high heterogeneity between types and brands, waste electronics are a special stream of municipal solid waste that requires complex considerations during end-of-life (EOL) management 7,8 . Without safe and effective EOL processes and government oversight, the receiving regions face various environmental pollution and public health concerns such as heavy metal poisoning and food chain contamination, which is the case for several developing nations 5,[9][10][11] . However, with proper EOL management, waste electronics can have positive socio-economic impacts, as they contain materials of important value that could allow proper recycling to have economic benefits. In terms of overall material composition, waste electronics are a combination of plastics, metals and other substances. The current waste electronics recycling process includes preprocessing and metal recycling. A general waste electronics device includes plastic covers, printed circuit boards (PCBs) that hold the various electronic Analysis https://doi.org/10.1038/s41893-022-00983-9 and the embedded value within the United States. This study identifies the underlying potential connections between gold recovery and the current metal refining industries in the United States and offers recommendations on creating an electronics-centred circular economy for different future scenarios. Other secondary resources, such as base metals (that is, copper and aluminium) and plastics, can also add value to the recycling of waste electronics; these are recognized and discussed as well (Fig. 1). Figure 2a shows the modelled number of waste electronics in the next decade, based on the historical sales data and the products' predicted life spans. In the fast-growth scenario, the amount of small-to-midsize waste electronics produced in 2033 could exceed 1 billion units; the slow-growth scenario would lead to approximately 700 million units. In terms of mass, waste electronics generated in the United States are estimated to reach 1.2 to 1.5 million tonnes (Fig. 2b), with the shift towards smaller electronics partially offsetting the growth in the number of units. Note that in this study, heavy consumer appliance electronics (such as fridges or ovens) are not included; therefore, both the number and mass are conservative estimates. Figure 2c,d shows the mean and general growth trend of representative resources in waste electronics over the next decade, within which PCBs contain the most valuable materials recoverable, including precious metals such as gold (Fig. 2c) 13,14 . The steady trend of gold availability from waste electronics in recent years (that is, 2017-2021) is consistent with previous studies. For example, Althaf et al. 21 and Golev et al. 24 estimated a steady and slow decline in gold availability from 2015 to 2018 among conventional electronics in the United States and Australia, respectively, owning to a trend towards smaller electronic devices and more efficient use of gold within those electronics.

Temporal and spatial distribution of waste electronics
However, existing tear-down studies on emerging smart devices, which often contain more complex electronics (such as smartphones and tablets), have shown higher PCB and gold content than that of larger, stationary and wired electronics (for example, DVD players and VCRs) 25 . Even though the total amount of gold used in electronics has declined in recent years, which influences the economics of recovery, the number of more complex devices with higher PCB and gold content has been increasing. These trade-offs indicate that as more smart electronics start to emerge, the amounts of both PCBs and gold that are recoverable from waste electronics could start to increase within the next decade.
Due to limited publicly available tear-down analyses, the modelled PCB and gold content assumed in different types of electronic devices has a relatively large range (Fig. 2c,d). The impact of composition is thus parts (such as resistors and capacitors), batteries and so on 12 . Each of these components offers unique value upon recycling. Existing techno-economic analysis studies have demonstrated the potential of making profits via recovering metal resources, particularly gold, from the PCBs [13][14][15] . Therefore, while improvements in international regulations on waste electronics recycling could be beneficial, it is also critical to predict and assess potential waste management strategies within the United States around metal recovery from waste electronics to identify favourable economic and environmental pathways, obviate future resource scarcities and create a more circular economy to increase the resiliency of the domestic material supply.
Within waste electronics, small-to-midsize consumer electronics (for example smartphones, fitness devices, connected home devices, augmented/virtual reality (AR/VR) equipment, drones and computers) represent an emerging stream in recent years that is heavily affected by the evolution of technology development and consumer behaviours. Material flow analysis (MFA) is an effective and strategic way of predicting the quantity and composition of waste materials on the basis of information such as historical sales and possession data, which has been applied to model waste electronic generation in different parts of the world [16][17][18] . However, existing MFAs on waste electronics in the United States have focused mostly on the temporal changes of common and relatively conventional small-to-midsize waste electronics (such as TVs, mobile phones, computers and monitors) within the past decade 19-21 . Due to potential challenges such as data availability at the time of these studies and the rapid change in consumer behaviours, limited research has been conducted regarding the impacts of the emerging small-to-midsize electronics that have been introduced in recent years. Also, the variance of the metal compositions for different types of electronics is rarely taken into account. Lastly, existing research on the geospatial modelling of waste electronics recycling reports is limited to select waste electronics in select states 22 .
On the basis of these knowledge gaps, here we couple geospatial analysis and state-of-the-art MFA to comprehensively capture the temporal variations and predict future trends regarding the amount, composition and potential value within these consumer waste electronics generated in the United States. We also predict how the newly developed, emerging electronic products can potentially reshape the resource availability from waste electronics and how these changes may affect the future metal refining infrastructure in the United States.
Since previous analyses showed that gold holds most of the metallic value (up to or higher than 80%) in waste electronics 13,14,23 , in this analysis we use gold as the primary indicator to examine the economic potential (Fig. 1). The MFA results are used in subsequent geospatial models to characterize the spatial distribution of waste electronics Analysis https://doi.org/10.1038/s41893-022-00983-9 much higher than that of the growth scenario, which further affects the economic potentials analysed in the Discussion.
In terms of zip-code-level distributions in the United States, as the model is developed on the basis of population and household possession data (Methods and Supplementary Note 4) 26-28 , the densities of waste electronics are heavier in the coastal and metropolitan areas (Fig. 3a). In the densest regions (such as certain places in New York City and Los Angeles), over 160,000 kg of waste electronics could be generated annually within one square kilometre.
For metal extraction, the mining and refining sector is an established industry in the United States that can potentially be utilized to treat waste electronics or their embedded PCBs, as both involve hydroor pyrometallurgical processes 7 . After the waste PCBs are dismantled, shredded and physically separated into metal scraps (upstream recycling), the metallurgical processes (downstream recycling) are fairly similar whether refining metal from waste or virgin mines 29,30 . In fact, a majority of the metal recovery from waste electronics is still based on pyrometallurgical pathways, which is also the primary virgin mining process with a minor difference between feedstocks (that is, precious metal versus base metal refining) [31][32][33] .
Virgin gold mines and refining plants mostly exist in the western states (such as Nevada and Arizona) of the United States, as shown in Fig. 3b. The rest of the virgin gold mines are scattered around the United States but are more concentrated in the mountain areas (for example, Utah and Colorado). Again, before waste electronics could be processed in metal refinery plants, they would need to be collected and preprocessed (that is, sorted, shredded and physically separated) by different recyclers.
The collection of waste electronics from consumers is out of the scope of this study, but Fig. 3b shows an estimate of the locations of certified waste electronics recyclers as recognized by the US Environmental Protection Agency 34 (green dots) and their proximity to the current virgin gold mines and/or refineries (triangles). The locations of certified waste electronics recyclers generally show good agreement with the distribution of waste electronics generation in Fig. 3a and indicate that there are unique challenges and opportunities for handling waste electronics among the different regions in the United States. For example, there could be an overlap between virgin mining production and the potential metal recovery from waste electronics mostly in the western United States, whereas there could be more potential for building new waste electronic mining facilities in the central and eastern regions.

Growth of different types of waste electronics
This study shows that currently, the United States is experiencing a dramatic change in waste electronics composition. The total amount of small-to-midsize consumer waste electronics and the potential precious metal (gold) recoverable from that waste have been steady in recent years; however, our model suggests that the value of waste electronics will probably increase in the next decade owing to changes in metal composition and shipments from emerging electronics. Specifically, Fig. 4a shows the predicted growth of representative waste electronics in 2021-2022 and 2032-2033 (see the whole list in the dataset at https://github.com/ppeng-cloud/ Consumer-Electronics-Recycling-Potential-in-United-States). Most of the waste electronics growing in 2033 also have positive growth rates in 2022. The overall scale of the difference in growth rates in 2022 is much larger (up to over 400%) than that in 2033 (up to 30%) owing to one or two spikes between 2018 and 2021, whose effect is diminished when projected to 2033. Notably, the waste from several emerging types of electronics (such as 5G smartphones and gateways, electric scooters and wireless earbuds) has a large range of predicted growth rates due to high sales fluctuations in recent years. For these electronics, further market analysis is recommended to provide more accurate growth predictions. Also, note that for the future scenario (2032-2033), the MFA model prediction in this study is a conservative estimate because this study does not account for completely new electronics that might enter the market with potentially high growth rates. Furthermore, emerging electronics such as AR/VR, connected home devices and internet-of-things products (defined as electronics requiring constant wireless communications) are leading the growth rate in 2032-2033. Other waste smart electronics (such as wireless headphones and drones) did not make it to the top ten growing list but also have close to or higher than 10% growth rates from 2032 to 2033. In contrast, most of the conventional waste electronics (such as desktop PCs, printers and DVD players) are generated at a steady rate or declining. These dramatic differences in the growth patterns between different waste electronics show that the United States is experiencing a shift in the composition of waste electronics, towards smaller, portable and more complex electronics, which potentially contain more resource value in terms of weight percentage 25 .
In terms of mass generated, Fig. 4b,c shows that certain heavy electronics such as LCD TVs greatly contribute to the overall mass of waste electronics, similar to how CRT TVs dominated the mass of waste electronics in the early twenty-first century 21 . Other heavy electronics, such as printers and desktop PCs, are still among the top ten waste electronics by mass. Lastly, smaller computers such as laptops and tablets will probably exceed conventional desktops and monitors in the waste stream by 2033.

Economic potential
Our geospatial modelling shows that there are potential underlying connections and opportunities between virgin mining and recycling in the United States, which can be used to help create a circular economy around metal recovery from waste electronics. By evaluating the capacity and profitability of the virgin mining refineries, we find that it is worthwhile to consider the integrated pathway of recycled waste electronics and virgin metal recovery routes in the United States.
Since gold can represent over 85% of the embedded value in consumer electronics [13][14][15] , it is used as the primary indicator to study the direct economic potential of treating electronics. One important criterion is to compare the potential gold productivity from waste electronics with the current gold productivity throughout the United States, which is approximately 220 tonnes annually 35 . Figure 5a shows that the gold recoverable from waste electronics can potentially reach the same magnitude as the national productivity from virgin resources when assuming that gold compositions of the electronics are on the higher end of values found in the literature 21,24,25,36 .
Theoretically, the virgin mining refining plants have the capacity to handle the total quantity of waste electronics for precious metal (gold) recovery. However, considering their geological distribution and current gold mining production, virgin mining with large handling capacities for waste electronics is concentrated in the west and mountain areas of the United States. The central and east regions may have more need to create new recycling infrastructure targeting waste electronics.
On the state level (estimated in Supplementary Note 5), Nevada and Alaska are the leading states for virgin gold production (Fig. 5b), with capabilities of approximately 173 metric tonnes and 21 metric tonnes per year (2018 values published in 2021), respectively 35 . Due to its high virgin gold productivity, refining plants at mines in Nevada should theoretically have the capacity to handle all of the waste electronics in the United States but would face extra burdens (that is, time, cost and emissions) when transporting waste electronics that are generated far from the region.
If the embedded gold from waste electronics in certain regions reached the maximum allocated productivity (light blue in Fig. 5c-h), the additional amount would need to be transported to the next nearest facility with excess productivity (dark blue). In this case, the transportation burden increases with the distance, which is qualitatively represented by the darkness of the yellow in Fig. 5c-  Analysis https://doi.org/10.1038/s41893-022-00983-9 efforts will be made to commercialize these technologies specifically to recover resources from waste electronics treatment. These new facilities may need to compete with virgin mining refineries that integrate electronic waste recycling into their production capacity. A higher transportation burden to integrated virgin refineries thus indicates greater potential to build such new facilities (represented by dark yellow). In Fig. 5c-h, the darker blue represents a higher opportunity to integrate waste electronics with existing virgin mining refineries, whereas the darker yellow represents the higher economic potential for new facilities.
The MFA results (Fig. 3) and Fig. 5a show that the uncertainties in the embedded gold content have a much higher influence than the growth scenario; these uncertainties are therefore chosen as one of the main parameters for sensitivity to the recycling economic potential. Another key parameter is the level of involvement, which denotes how much productivity the virgin refineries can allocate to use waste electronics as feedstock. The influence of exportation is also studied, as it directly relates to the level of domestic recycling of waste electronics. We found that the reported degree of waste electronics exportation in the United States varies widely depending on the time of study and reporting agencies 20,39 . Figure 5 shows that within the uncertainties studied, the economic potential is most sensitive to the metal content, as shown when comparing Fig. 5c-e with Fig. 5f-h. The levels of exportation and involvement of the current virgin mining industry will affect the east and central regions if the embedded values are low, as shown in Fig. 5f-h. Figure 5c-e shows that if the embedded metal content is high, there is high economic potential to develop new infrastructure around waste electronics recycling even if there is high exportation, or if the virgin mines shift a large portion (up to around 70%) of their productivity to be generated from waste electronics.
It is important to note that metal compositions differ between waste electronics and virgin mines 36,40 , which would require more separation stages. The results from this section imply that quantifying the    Analysis https://doi.org/10.1038/s41893-022-00983-9 economic trade-off between the exact procedures that need to be added for the virgin mines to handle electronics can make a notable difference in examining the nationwide profitability. Future research is thus recommended on the techno-economic comparison between adapting virgin metal refineries (particularly for gold) to include separated electronics as part of the feed stream versus building completely new plants.

Future outlook
The above discussions highlight potential opportunities to enhance the circular economic potential of metal recovery from waste electronics recycling and help address several major challenges identified in this study. Future research should focus on process development for both upstream (recyclers and collectors) and downstream refineries for this integrated waste and virgin mining pathway to enhance the economic potential. Additionally, policy efforts should focus on creating a national-level database that includes the composition-level tear-down data and the locations of small-scale refineries and collection plants for various electronics to help narrow the range for future MFAs and offer a more complete geospatial analysis of metal recovery from waste electronics. The first and most important challenge is the need for a better understanding of metal compositions within different types of electronics, especially among emerging smart electronics. The availability range plotted in Figs. 2 and 5 is relatively large because only limited tear-down studies are available for waste electronics. Between the two main uncertainties included in this analysis, the composition has a substantially higher impact on the available gold in waste electronics than the growth scenarios. The large range of resource availability indicates that different management strategies might be needed when aiming to create a circular economy around waste electronics. A database that includes the compositions of metals for different types of electronics would help anticipate future recycling infrastructure needs. Such a database could be achieved via high-quality tear-down analysis, as well as help from the electronics manufacturers without exposing company intellectual properties.
Second, there has been limited transparency at the national level on both the upstream recyclers (the green dots in Fig. 3b) and metal refineries (the triangles in Fig. 3b). There are many waste electronics recyclers and refineries (that is, small-scale and/or regional certified facilities) with minimal publicly available information and transparency that may require a national-level survey or reporting database.
Furthermore, although gold is chosen as the primary driving factor for economics due to its high embedded cost value, other resources in waste electronics can also add value to the circular economic potential of waste electronics recycling. For example, plastics in waste electronics, which can occupy up to 30 wt% 41 , could be recycled for re-manufacture or energy conversion purposes 42 . However, it is important to note that certain detoxification procedures might be required to eliminate the effects of brominated flame retardants during the high-temperature treatments 31,43 . Other metals, such as copper and aluminium, are not as valuable as gold but are also important to other manufacturing industries. More importantly, rare earth elements in magnetics and PCBs, and cobalt and lithium in batteries, can help with increasing the supply chain resiliency of critical materials 44,45 and should be considered in future studies.
Lastly, since this study does not include all of the consumer and non-consumer waste electronics, the economic potential and profitability are relatively conservative estimates. If other sources of waste electronics (such as the growing electronics in vehicles and industrial plants) are included, the yellow portion of the profitability maps can potentially be expanded.

MFA
Compared with the previous MFA models for waste electronics 19-21 , we expanded the types of waste electronics from the approximately 20 common electronics (such as waste cell phones and TVs) to over 90 different types of electronic devices. The waste electronics covered in this study include emerging waste electronics such as wearable fitness and health products, portable and wireless devices, smart home improvement devices, AR/VR sets and consumer drones. The sales data from the Consumer Technology Association were used 46 . The types of electronics modelled in this study are included in the dataset at https://github.com/ppeng-cloud/ Consumer-Electronics-Recycling-Potential-in-United-States.
The sales data on the target consumer electronics after 2021 were predicted via a logistic-based model, which is known to be capable of capturing the market behaviour of consumer products [47][48][49] . The logistic model can be described by equations (1) and (2): where t denotes the number of years for market sales, and ∂ (saturation), β (steepness) and γ (midpoint) are the logistic parameters. Note that equation (1) was used for the declining electronics, and equation (2) was used for the growing electronics.
To account for the robust change in consumer behaviours, the maximum and last-reported three years of sales data for different electronics were first used to categorize the electronics into 'decreasing' or 'increasing' patterns. For the electronics that had decreasing sales patterns, we compared the logistic fitting since the maximum reported value and the five most recent values, and we selected the one with the larger r 2 value to predict the future sales data. The same method was used to predict the sales data beyond the reported years if the reporting stopped before 2021. For the increasing electronics, we set the defined fast-growth and slow-growth scenarios using logistic fitting based on different assumptions of maximum sales penetration per household value. More details on categorizing the growth scenarios and sales data prediction are included in Supplementary Note 1.
A Weibull probability function was used to predict the temporal evolution of waste electronics in the MFA model, which has been used in previous studies to predict the flow of waste electronics on the basis of their life spans 21,50 . In this model, we assumed that the probability of a certain type of electronic device (j) reaching its EOL within its maximum usage year (n) followed a Weibull distribution, which can be described by equations (3) and (4) [19][20][21] : In the probability density function (equation (3)), t represents time, and δ and η are the parameters used to describe the Weibull probability function. The probability of reaching EOL (P) can be then used, along with the historical sales data (S) to calculate the flow (N) of waste electronics j generated after a life span of i years within its maximum life span (n) 21 : The above-mentioned MFA model was applied to the US shipment data on the electronics evaluated in this study. The model was conducted using the open-source MATLAB code as described in Althaf et al. 21 with adjustments to the numbers and types of electronics, Weibull parameters, average mass, compositions and other parameters.
For the common electronics that were analysed in previous studies (that is, waste cell phones and TVs), we combined the Analysis https://doi.org/10.1038/s41893-022-00983-9 previously published Weibull parameters and life spans from different sources [19][20][21]51 . The sum of the probability functions derived from the previous Weibull parameters is slightly less than 1 if added up to the same reported maximum life span in the previous literature, in which the deviation probability was assumed to be the devices that can be further reused when predicting the amount of waste generated. But the reused or refurbished electronics were not included for further analysis (they are included later in 'Key assumptions and uncertainties').
For the relatively sparsely studied electronics (that is, smart electronics, wireless electronics and AR/VR sets), their MFA parameters were determined from those of the 20 common electronics on the basis of UNU classification codes, Harmonized System trade codes and functionalities 51,52 . The details of the decision tree and sources for determining the MFA parameters are included in Supplementary Notes 2 and 3. The material compositions of different waste electronics were determined from previous literature. The upper and lower bounds of the reported values were used as the high-content and low-content scenarios, respectively, with details shown in Supplementary Note 3.

Geospatial modelling
To model the spatial distribution of the waste electronics generated across the United States, the total MFA results were distributed to each zip code area on the basis of the population density and the amount of electronics in residential households and commercial office buildings within different geological regions of the United States (that is, New England, the Pacific and so on) 27,28 . The regional results were normalized to waste electronics generated per capita for different zip code areas in the region.
The average percentages of ownership across different regions of the United States were calculated on the basis of the possession and ownership data for 16 types of electronics (such as smartphones, TVs, cell phones and desktop and laptop computers) in residential households and 5 types of electronics in commercial office buildings provided by the US Energy Information Administration 27,28 . The total amount of waste electronics generated from the MFA was adjusted and distributed to different regions according to their average percentage of ownership and divided by their total population to obtain the average generation per capita for different regions. Note that due to this assumption, the broader distribution of ownership per household of various electronics might be different from the representative types of electronics used.
Waste electronics generated per capita for each geological region were multiplied by the population density data for each zip code in the United States to estimate the waste electronics generated for each zip code. The waste electronics generated per zip code were combined with the corresponding geospatial data (shape, boundary, longitude, latitude and so on) to plot the distribution of the United States. The nationwide geological shape data used in this study were obtained from the US Census Bureau 53 .
To assess the potential connections between waste electronics recycling and virgin refineries, the geospatial coordinates for the mines, mining plants and their state-level productivities in the United States were obtained from various US Geological Survey and other sources 35,[54][55][56] . The nationwide certified recycler data were provided by Sustainable Electronics Recycling International (SERI) 57 on the basis of their certified recycler lists across the United States.
Note that besides virgin mining, there are several existing major refineries that list waste electronics as part of their feedstocks (Supplementary Table 4) to produce high-quality metals for the technology industry 58,59 . As there is limited information on the gold productivity and feedstock composition for these refineries, we estimated their influence on the economic potential in Supplementary Note 6 by comparing with virgin plants' productivities, previous techno-economic analysis and tear-down studies 13,14,21,25,60 .
Furthermore, we applied a distance matrix to first determine the potential capability for treating waste electronics at the nearest existing facility, through integrating with virgin mining refineries. We assumed that these facilities can allocate or expand a certain portion of their current gold productivity to waste electronics. We further characterized the United States into five greater regions shown in the legend of Fig. 5 and analysed their handling capabilities in Fig. 5c-h. More details on the geospatial modelling and productivity and profitability analyses are included in Supplementary Notes 4-6.

Key assumptions and uncertainties
All of the uncertainties and assumptions summarized in this section were also discussed in the previous text or Supplementary Notes 1-6 when they were applied to the corresponding analysis. First, the scope of this study includes small-to-midsize consumer electronics but not large electronics. Due to this assumption, we expect that the overall amount of waste electronics will be higher than what is predicted in this study. Second, refurbished or reused electronics are not modelled in this study. As described in the MFA section, reuse was recognized by assuming that not all of the electronics sales reach EOL in the probability distribution based on the reported Weibull parameters and maximum life spans, but their further re-introduction to the waste stream was not included in this analysis.
Due to limited data availability, key uncertainties in this study included the composition of resources within waste electronics, the degree of domestic recycling (as studied by the percent export), the approximation of MFA results and spatial analysis by using representative values. We also recognize that there is a high potential that new types of small-to-midsize consumer electronics will be introduced in the near future, similar to how AR/VR and 5G devices have emerged in recent years. We accounted for this by qualitatively showing the possible markup of results in Fig. 2. Lastly, as discussed in the 'Future outlook' section, although the effects of the key sources of uncertainty were studied (namely, the growth scenario, content of resources and level of export) in the Results and Discussion, future research would greatly benefit from narrowing these ranges via more comprehensive tear-down data and improved policy incentives, as stated in the Discussion.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
The data that support the findings of this study are available within the paper and its Supplementary Information. The supplementary dataset is available at https://github.com/ppeng-cloud/ Consumer-Electronics-Recycling-Potential-in-United-States.

Corresponding author(s): Arman Shehabi
Last updated by author(s): Aug 29, 2022 Reporting Summary Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy The data that support the findings of this study are available within the paper and its Supplementary Information. Supplementary dataset is available at https:// github.com/ppeng-cloud/Consumer-Electronics-Recycling-Potential-in-United-States.

nature portfolio | reporting summary
March 2021

Human research participants
Policy information about studies involving human research participants and Sex and Gender in Research.

Reporting on sex and gender
Use the terms sex (biological attribute) and gender (shaped by social and cultural circumstances) carefully in order to avoid confusing both terms. Indicate if findings apply to only one sex or gender; describe whether sex and gender were considered in study design whether sex and/or gender was determined based on self-reporting or assigned and methods used. Provide in the source data disaggregated sex and gender data where this information has been collected, and consent has been obtained for sharing of individual-level data; provide overall numbers in this Reporting Summary. Please state if this information has not been collected. Report sex-and gender-based analyses where performed, justify reasons for lack of sex-and gender-based analysis.

Population characteristics
Describe the covariate-relevant population characteristics of the human research participants (e.g. age, genotypic information, past and current diagnosis and treatment categories). If you filled out the behavioural & social sciences study design questions and have nothing to add here, write "See above."

Recruitment
Describe how participants were recruited. Outline any potential self-selection bias or other biases that may be present and how these are likely to impact results.

Ethics oversight
Identify the organization(s) that approved the study protocol.
Note that full information on the approval of the study protocol must also be provided in the manuscript.

Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.

Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Ecological, evolutionary & environmental sciences study design
All studies must disclose on these points even when the disclosure is negative.

Study description
This study uses electronics sales data to conduct material flow and spatial analyses to investigate the availability of mid-to-small consumer waste electronic devices, and their embedded materials (gold and printed circuit board) in different years and different parts of the United States.

Research sample
The research sample includes the sales data obtained from Consumer Technology Association, and geospatial data obtained from U.S. Census Bureau, United States Geological Survey (USGS), and Sustainable Electronics Recycling International (SERI), and refinery supply information from electronics sectors. All of the sources are cited in the article.

Sampling strategy
All data obtained for the selected mid-to-small size consumer electronics and metals are studied.

Data collection
The data were downloaded directly from the corresponding sources cited in the article.

Data exclusions
No data are excluded in the analysis for the selected electronics.

Reproducibility
This study does not involve experiments, and can be reproduced with the data downloaded from the cited references, steps described in the Methods and Supplementary Information, as well as the supplementary dataset and code provided.

Randomization
Randomization was not applicable because no measurement was performed in this study. All the sales data obtained for the electronics was studied, and all the scenarios for future prediction were described in the article and Supplementary Information.

Blinding
Blinding was not applicable to this study because no measurement was performed and the analysis was based on previously recorded data.
Did the study involve field work?

Yes No
Reporting for specific materials, systems and methods