Environmental footprints show the savings potential of high reparability through modular smartphone design

doi:10.21203/rs.3.rs-2724319/v1

Download PDF

Article

Environmental footprints show the savings potential of high reparability through modular smartphone design

https://doi.org/10.21203/rs.3.rs-2724319/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Two thirds of all people own a mobile phone or smartphone, which are typically not very durable and often are replaced. As a result, mobile phones already outnumber people on earth and represent the fastest growing waste stream. This entails a whole range of problems. On the environmental impact side, issues range from high input of metal ores to large amounts of electronic waste. Here, we quantify the environmental benefit of reparability on the example of a modular and easily repairable smartphone facilitating a longer lifetime. Within the scope of a life cycle assessment, we analyse the climate, energy, land, material and water footprint, focusing on the potential savings that arise from modularity and the longer lifetime. A modular use case, in which a smartphone is used for 5 years through replacement of defective modules, is compared to a reference use case with 2.5 years standard use and no replacements by means of the application-related functional unit "smartphone use for one year". The reference use case is responsible for 9 kg climate-damaging emissions, consumes 33 kWh of energy, 0.4 m² of land, 16 kg of raw material and 32 kg of primary material as well as 3 m³ of water and would require 8,000 m³ of dilution water to eliminate water pollution by dilution. The modular use case can save an average of 40% of emissions and natural resources per functional unit. In the area of gold production alone, 3 kg of raw materials or 9 kg of primary materials can be saved. Scaled to 2 billion smartphones sold worldwide yearly, raw material savings are in the order of 13,000 multi-family houses, while CO₂ emissions can be saved in the order of 12 million medium-haul flights per year. Spatial hotspots of environmental impacts can be reduced and mitigated if easy reparability is ensured through a modular design and if customers use their smartphones longer.

Earth and environmental sciences/Environmental sciences/Environmental impact

Scientific community and society/Business and industry/Technology

Earth and environmental sciences/Environmental social sciences/Sustainability

This article is the first to quantify the potential of a modular smartphone to save emissions and natural resources and mitigate hotspots of environmental burden through a high degree of reparability and a consequently longer lifetime. Background is that mobile phones and smartphones have become an integral part of the 21st century: most people in the world own one. Combined with the inherent short-lived nature of these devices, this has already led to a situation in 2011 where there were more mobile phones and smartphones on earth than people^1,2.

Primary resources have to be continuously extracted from nature for the 2 billion smartphones produced and sold annually (as of 2014, meanwhile more³). In particular, rare and precious metals are needed⁴. For example, according to the OECD (2010)⁴, at the time of their report, 84 tons of antimony, 7.1 tons of beryllium, 12.1 tons of palladium and 0.3 tons of platinum were needed to manufacture the global demand for smartphones. The extraction of minerals and metal ores entails a whole range of environmental impacts, e. g. emissions of pollutants to air and surface water, leaching from landfilled waste or impacts related to energy, land and water use⁵. Special metals that are required in smartphone production, such as gold, palladium and cobalt, in particular are associated with higher life-cycle wide environmental impacts per unit than bulk metals⁵. This is compounded by socio-economic impacts, including health hazards⁶, human rights abuses and conflicts as demonstrated particularly infamous in the DR Congo. Here, unregulated artisanal mining exacts high human health costs, accompanied by a rapid degradation of the local environment⁷.

And environmental damage also occurs further along the supply chain in the production of smartphone components, for example through the use of toxic chemicals⁸. With the increasing spread of electronic products and a reduction in usage times, the ratio of environmental costs from the production phase to the use phase has been reversed: the production phase - and the associated impacts - are becoming increasingly important, in contrast to what used to be the case with long-lasting but power-hungry large electrical devices in the past⁹.

Of the large number of smartphones on earth, only a portion is in use, while the rest represents a steadily growing electronic waste stream, the fastest growing of all waste streams¹⁰. Less than 20% of the electronic waste is collected and processed in such a way that some of the materials it contains are made available again¹¹. The rest contributes to the build-up of a currently unused anthropogenic stock, which is expected to reach the magnitude of 75 million metric tons by 2030¹¹. The flow of e-waste often escapes official controls, which means that European e-waste ends up in Ghana's capital Accra, for example, at the world's largest landfill Agbogbloshie. Residents there have resorted to informal, rudimentary e-waste recycling for lack of better options, which has led to the accumulation of copper, lead and zinc in the surrounding soils and poses a health risk to people¹².

The SHIFTPHONE from the German-based family business SHIFT GmbH attempts to counteract the described implications with a modular design following design principles seeking to retain value within the economic system¹³: The innovative smartphone is made up of individual modules that can be easily taken apart without the need for specialist knowledge or special tools. It is the most modular on the market with over 13 modules. In the event of damage, spare parts are easily obtainable and enable cost-effective replacement of specific modules by the end user through self-assembly without voiding the warranty. Theoretically, upgrades are also possible. This combination of design focusing on reparability and changed usage behaviour leads to a significant extension of the lifetime of SHIFTPHONES: Surveys have shown that SHIFTPHONE users typically can use their device for five years (including the possibly higher intrinsic motivation of SHIFTPHONE users to use longer), while the optimistically estimated average standard lifetime of a smartphone is 2.5 years¹⁴.

A life cycle assessment (LCA) of the SHIFTPHONE is conducted and the climate, energy, land, material and water footprints are calculated using selected life cycle impact assessment (LCIA) indicators (Table 1).

Table 1

LCIA methods and terminology throughout this study. FP: Footprint; *if named differently.
Footprint	LCIA method	Indicators	Impact categories	Unit (this study)	this study*
climate FP	IPCC 2013¹⁵, updated according to IPCC 2021¹⁶	Global Warming Potential (GWP)	GWP_100a	kg CO₂-eqv.
energy FP	Cumulative Energy Demand (CED)¹⁵	CED, fossil	non-renewable energy resources, fossil	MJ eqv. (kWh)
			non-renewable energy resources, nuclear	MJ eqv. (kWh)
			non-renewable energy resources, primary forest	MJ eqv. (kWh)
land FP	Land occupation¹⁷	Land occupation	Land occupation	m²
material FP	Product Material Footprint (PMF)¹⁸	Raw Material Input (RMI)	Raw material input (RMI)	kg	raw material
material FP	Product Material Footprint (PMF)¹⁸	Total Material Requirement (TMR)	Total material requirement (TMR)	kg	primary material
water FP	Water Scarcity Footprint (WSF)¹⁹	Quantitative Water Scarcity Footprint (WSF_quan)	evapotranspiration	m³	quantitative water FP
			product-incorporated water	m³
			water transfers	m³
		Qualitative Water Scarcity Footprint (WSF_qual)		m³	qualitative water FP

In order to quantify the effect of its high reparability, two use cases are distinguished: Use case M2x2.5 assumes a typical smartphone lifetime of 2.5 years and serves as reference model for a standard smartphone, whereas use case M5 assumes a lifetime of 5 years through replacement of defective modules and is specific to the SHIFTPHONE (Fig. 1).

To ensure comparability, M2x2.5 is also referenced to 5 years by the input of two SHIFTPHONES. As application-oriented functional unit, a smartphone use for one year considering the production and the use phase is chosen. The climate, energy, land, material and water footprint are calculated for the production phase of one SHIFTPHONE and per functional unit. Some of these footprints consist of several sub-indicators covering different categories of environmental impact (Table 1). Spatial and activity hotspots are determined by normalising footprint results of selected individual activities along the LCA supply chain, which contribute at least 1% to a total footprint result, with the median of all results. For spatial hotspots, the normalised results from all footprints are summarized per location if known, for activity hotspots, they are summarized per activity category.

1.1 Footprints of the production phase of a SHIFTPHONE

The production phase of the SHIFTPHONE is responsible for 20 kg emissions of CO₂ equivalents, consumes 72 kWh equivalents of energy, 1 m² a land, 38 kg raw material, 78 kg primary material as well as 7 m³ water and requires 20,000 m³ of water to virtually dilute water pollution below specific thresholds (Fig. 2, Supplementary Data 1).

It is difficult to compare the results with other smartphones because comparable studies have used other indicators for evaluation and other LCA databases. However, differences are hardly to be expected here, as the SHIFTPHONE does not use special production methods or alternative materials. A comparison with the Fairphone 3 (36 kg CO₂ equivalents²¹) shows that the climate footprint of the SHIFTPHONE is smaller, but still in the same order of magnitude. Differences in this case may result from the use of a different database. The analysis and results of the production phase of the SHIFTPHONE form the basis for the comparison of the use cases.

1.2 Footprints of the use cases

The footprints of the standard use case M2x2.5 are in the range of 35 to 42% larger than those of the modular use case M5 (Table 2, Supplementary Data 2 and 3) per functional unit. The functional unit here is the smartphone use for one year, considering production and use phase.

Table 2

Footprint results of the reference use case M2x2.5 and the modular use case M5 per smartphone use for one year. FP: footprint, RMI: raw material input, TMR: total material requirement, quant.: quantitative, qual.: qualitative, sd: standard deviation.
Footprint	unit, per smartphone use for one year	M2.5	sd	M5	sd	savings [%]
climate FP	kg CO₂-Eqv.	8.6	± 0.3	5	± 1	37
energy FP	kWh-Eqv.	33	± 6	21.4	± 0.1	38
land FP	m²*a land occupation	0.40	± 0.04	0.26	± 0.04	36
material FP (RMI)	kg raw material	16	± 1	9.8	± 0.2	42
material FP (RMI, metal)	kg raw material	9	± 1	5.1	± 0.3	44
material FP (TMR)	kg primary material	32	± 2	19.0	± 0.5	43
material FP (TMR, metal)	kg primary material	24	± 2	14	± 1	44
water FP (quant.)	m³ water	3	± 2	2	± 1	40
water FP (qual.)	m³ virtual dilution water	8119	± 5185	4914	± 3616	42

In comparison with the standard use case, the modular case shows the largest savings in the material footprint (RMI 42 and TMR 43%) of which metallic raw materials have a share of at least 60%. Smallest savings are possible in the land footprint (36%). The special role of the material footprint in savings potential can be explained by a closer look at the inputs of the two use cases: for M5, in order to extend the lifetime to 5 years, one SHIFTPHONE and various replacement modules are needed, e.g. 0.06 displays per functional unit (Supplementary Table 1), while for M2x2.5 two SHIFTPHONES are needed. The inventory analysis shows that the display and the mainboard contain comparatively the most gold of all modules directly or in the supply chain. M5 saves gold compared to M2x2.5 by consuming only 0.27 displays per functional unit (0.4 in M2x2.5) and 0.2 mainboards per functional unit (0.4 in M2x2.5). This saving in particular has an above-average effect on the material footprint, as the characterisation factor of gold in the material footprint is exceptionally high. The analysis of the activity hotspots provides more detailed information on this.

1.3 Identification of activity hotspots

Activity hotspots of the standard use case M2x2.5 are mining, mainly gold mining, energy supply as well as waste treatment. The modular use case M5 relieves in particular the mining hotspots by saving a total of more than 5 kg of raw material (RMI) and more than 14 kg of primary material (TMR) per functional unit.

Activity hotspots are determined within the framework of a hotspot analysis, which examines the LCA results in more detail according to the type of activities along the supply chain. To this end, the results of the single footprints must first be made comparable: for each footprint, the median value of all activities of the two use cases M2x2.5 and M5 is determined and used to normalise each single activity result through division (Supplementary Data 4). A normalised value represents the ratio of the raw value to the median. Normalised single footprints are combined into the dimensionless index environmental burden by summing up results of the same activity category, namely mining (of gold, other metals and minerals), energy supply (through hard coal mining, petroleum and gas production, wood chips from forestry, electricity and heat production as well as diesel use), waste treatment, display production, transport, on-site activities (referring to the location of the SHIFT GmbH) and other activities. The environmental burden per activity category is considered as hotspot if it exceeds a threshold value of 5, which means that the environmental burden is 5 times as large as a median footprint result.

Activity hotspots of the standard use case M2x2.5 are mining with a share of 31%, in particular gold mining with 22%, energy supply, mainly based on hard coal, petroleum and gas and electricity and heat, with 38% as well as waste treatment with a share of 18% of the total environmental burden (Fig. 3, Supplementary Table 2). The category “other” is also a hotspot according to its numerical value, but only because it combines numerous different activities that are not hotspots when considered individually.

1.4 Reduction of activity hotspots through the modular use case

The modular use case M5 shows a smaller absolute environmental burden than M2x2.5 of 190 compared to 295. This reduction applies to all categories, so that not only the total environmental burden decreases, but also the environmental burden of all categories with the highest absolute reduction of 28 in the category gold mining. The shares of the individual categories vary between M2x2.5 and M5 and the biggest difference is that in M5 metal mining has a 24% lower relative share (differences in the generic category “other” not considered).

In absolute numbers, the highest reductions of the dimensionless index environmental burden originate from the activity categories gold mining (27.8, Supplementary Fig. 1, Supplementary Table 2), waste treatment (21.1), hard coal mining (13.1) as well as electricity and heat production (10.9). All mining activities taken together (except for energy carriers), the modular use case M5 leads to a reduction of the environmental burden from 91 to 38 by saving 2.9 kg, 0.7 kg and 0.3 kg of raw material (RMI) in the activity hotspots gold mining, other metal mining and mineral mining compared to M2x2.5 per functional unit (Fig. 4). In terms of primary material (TMR), the savings are 8.9 kg, 1.4 kg and 0.7 kg (Fig. 4). The environmental burden of the hotspot energy supply is reduced from 114 to 74 through a reduced use of hard coal, petroleum and gas, electricity and heat and wood chips.

As regards the composition of the activity categories, compared to the reference use case M2x2.5, an 11% higher share of non-metal mining and an 18% higher share of diesel use are prominent in the modular use case M5. This contrasts with an 12% and 24% lower share of gold and other metal mining, a 23% lower share of on-site activities and a 9 and 10% lower share in use of energy wood and electricity and heat production. Thus, there is a shift in the activity categories with the modular use case M5. Since M5 has a lower environmental burden in absolute terms in all categories, this does not lead to a shift of burden. Actually, the proportionate decrease in gold and hard coal production in particular represents a real relief for the mining and energy supply hotspots.

1.5 Relief of spatial hotspots by the modular variant

Most spatial hotspots of the reference use case M2x2.5 are related to the production phase and originate from mining, in particular of gold and copper, and associated treatment of tailings. Second most are spatial hotspots due to energy supply along the supply chain. M5 reduces numerous hotspots and even dissolves some completely.

Spatial hotspots are determined in the same way as activity hotspots, i. e. spatial hotspots are places in the supply chain of the SHIFTPHONE where the sum of the normalised footprint values of all activities taking place there is greater than 5. According to the spatial resolution of the LCI, hotspots can be point coordinates, countries but also regions of two or more countries. This circumstance, which is related to the quality of the regionalisation, is also evaluated in the following.

A hotspot is usually formed by the environmental impacts of several activities (Supplementary Table 3), but most hotspots related to M2x2.5 originate from mining activities: in Australia, Canada, China, Mexico and the United States gold mining and treatment of tailings from gold mining contribute to the hotspots (Fig. 5).

Russia is a hotspot due to copper mining and treatment of associated tailings. In South Africa, there is a mixture of gold and hard coal mining responsible for the hotspot.

Second most important is energy production: in Europe, China and Chinese mines (circles located within China in Fig. 5) hotspots are related to the production of hard coal, while Algeria, the Middle East, Russia and the United States appear due to natural gas and petroleum production and Sweden because of forestry for energy wood production. Colombia appears as emerging hotspot because of electricity production from reservoir hydropower plants, which occupy relatively large areas of land with artificial lakes, and North America because of uranium production. That mining and energy production play the most important role is also confirmed, if hotspots are analysed separately for each footprint (Supplementary Figs. 3 and 4): the energy and qualitative water footprint (associated with treatment of tailings from mining) are responsible for contributions with a value greater than 5, while all others remain below 5. In M5, analysed separately by footprints, almost all locations have a value below 5. This is because M5 needs less input of primary resources and energy and relieves all hotspots.

An emerging hotspot in India (Fig. 5, Supplementary Table 3), which is due to tap and decarbonised water production, has less to do with smartphone production at first glance, but tap and decarbonised water is needed in the manufacturing of wafers for mainboards. Their production is modelled based on a global database activity, which uses a global market activity for the input of tap water. India is appearing prominently here, because no other activity is contributing to the hotspot there, but tap and decarbonised water production is also contributing to the hotspot in China.

In general, M2x2.5 shows more and more severe hotspots than M5 (Fig. 5). M5 could defuse hotspots in North America, Mexico, South Africa, Middle East, Russia, Australia and a Chinese mining region (circles in Fig. 5), while Colombia, Algeria and India disappear completely. Apart from one single contribution at the location of the SHIFT GmbH in Germany all hotspots are part of the upstream supply chain, which makes clear that direct activities are less relevant.

In comparison with the analysis of the types of activities, it shows that non-metal mining, diesel production, transport and display production are missing in the spatial analysis due to a lack of regionalised inventory data, although they together account for 15% (M2x2.5) and 17% (M5) of the total environmental burden. In addition, gold mining, natural gas and petroleum production and forestry are underrepresented in the spatial analysis: if the spatial resolution of these supply chains were better, we would probably see more hotspots.

1.6 Impact coverage of the hotspot analysis

For the hotspot analysis according to Schomberg et al. 2022²², a cut-off criterion was set so that only activities that contribute at least 1% to the overall result of a footprint are included in the analysis. For almost all footprints, the selected activities cover in total at least 61% and on average 80% of the total footprint result. However, this does not apply to the climate footprint, where a coverage of only 44% is achieved. This is an indication that there is a large number of activities which climate footprint results have a similar numerical value, so that they do not stand out as hotspots. In such cases, there are fewer hotspots, as they are intended to represent a comparatively high contribution with a clear distance to the other contributions. The coverage of on average 80% also applies to the analysis of the type of activities (Fig. 3): 80% of the total environmental burden is covered by the selected activities on average and the term “total environmental burden” in this study refers to the 80% analysed. However, random examinations of the remaining 20% of activities, each of which contributes less than 1% to the overall footprint results, show that these have no significant influence on the type and composition of the activity categories.

1.7 Spatial coverage of the hotspot analysis

Spatial coverage is determined by evaluating the quality of regionalisation: Regionalisation means that each activity in the supply chain of the SHIFTPHONE is assigned a location, which can range from point coordinates to global. Point coordinates correspond to a regionalisation of quality 1, quality 2 refers to countries, quality 3 to regions (corresponding to groupings from more than on country), quality 4 to allocations of quality 1 to 3 but which are doubtful, and quality 5 to global or rest-of-world. The last two are to be regarded unknown locations. 48% of the activities in the supply chain of the SHIFTPHONE are regionalised at least on sup-country level (Fig. 6), while the majority of activities (52%) could not be regionalised.

By using a regionalised data set for eight mining products and breaking down the SHIFTPHONE supply chain up to supplier level 3, this spatial coverage is already above the average of a standard LCA. Nevertheless, the spatial data is insufficient, why we supplement the spatial analysis of activities with an analysis of the type of activities²². Particularly affected by the lack of spatial data are the activity categories non-metal mining, diesel use, display production and transport. Here, it is unknown where in the world the associated activities take place. Conversely, hard coal and platinum mining are not affected by missing spatial data at all. In all other activity categories, the locations are partly known, mostly on a supra-country level, and partly unknown.

1.8 Variance analysis of the LCA

Monte-Carlo-Simulations of the footprints of the SHIFTPHONE as well as of both use cases are carried out with the software openLCA using 5,000 iterations and a cut-off value (0.00053 for the SHIFTPHONE model and 0.00005 for the use cases). Results show that most standard deviations are in the range of 2 to 18% of the footprint results. However, standard deviations of the water footprint are noticeably higher and can reach 74% for the qualitative water footprint (Fig. 2, Table 1). This is because the water footprint is calculated as regionalised LCIA whereby there are significantly more uncertainties to consider.

2.1 Limitations of the study and the study design

The functional unit “smartphone use for one year” is presented in this study as application-oriented unit intended to make the production and use phase of different devices easily comparable, in particular for users. It can also be suitable for other electrical devices, however, it needs to be supplemented by an identification of the technological level: different smartphone models show a very different performance, e. g. with respect to the quality of the camera. Within this study, such a distinction is not necessary, as reference use case and modular use case are based on the same dataset and have identical technical specifications. However, in order to put the savings potential in the context of technical maturity, which plays an important role for consumers when choosing a smartphone model, a corresponding modification of the functional unit together with an extension of the comparison to other smartphone models is necessary.

The quality of the input data is largely good, but there are also some gaps from which uncertainties result. In particular, the spatial and temporal resolution of data from the ecoinvent database is still too low overall, as more than half of the supply chain activities stored in the database are not spatially resolved at all. Likewise, a validation of the model is still pending. This could for example be done by determining the element contents of individual components and comparing these with the inventory results to be able to precisely determine the content of critical metals in particular. Both together, the low quality of the spatial mapping and the lack of validation of the inventory, can lead to both an underestimation and an overestimation of the number and severity of hotspots, although an underestimation is more likely and we assume to present the minimum environmental burden here.

Higher life-cycle costs arising from the sustainable design process were not taken into account in this study. Schischke et al. 2019²³ assume that the life-cycle wide impacts are about 10% higher than compared to a conventional, non-modular smartphone. Even if these results are not directly comparable with those of this study due to different framework conditions of the LCA analysis, the savings potential quantified here is already considerably higher at 40% per smartphone use per year.

2.2 Further and other savings opportunities

High reparability of smartphones alone can save large amounts of natural resources and emissions, as the results of this study show. Nevertheless, further savings and savings in other areas are also possible:

1) According to the repair statistics of SHIFT GmbH, the display breaks down most frequently, which alone is associated with high resource and energy costs due to its high weight in relation to the entire smartphone. Even if customers are already encouraged to use armoured glass and bumpers to protect the display, it is advisable to make further efforts to reduce the failure rate per functional unit. Possibilities include raising customer awareness, increasing the modularity of the display so that it can also be repaired, and evaluating and testing better display protection techniques. Since a few components typically dominate the material resource requirements of entire electrical and electronic devices, it may be worthwhile to extend such strategies to other components²⁴. However, the repair statistics of the SHIFTPHONE show that after the display and the battery, it is mainly less material-intensive, small components that are replaced.

2) Gold mining is responsible for the most and most severe hotspots. These could be relieved if it were possible to integrate recycled gold into the supply chain of the production phase of the SHIFTPHONE or at least increase the share. This would require a more in-depth analysis of the supply chain and cooperation with manufacturers. As we have also noticed in the context of this study, this is generally difficult, but for example within the framework of research projects, appropriate incentives can be created among manufacturers to participate. A look at the growing extraction of primary raw materials worldwide and the associated problems makes it clear that there is an urgent need for action here to limit the consumption of natural resources that does not stop at national borders²⁵. Such measures can also be successful indirectly, for example, when SHIFT GmbH brings recycled gold from its old equipment back into the cycle as compensation for the gold extracted for the production of new SHIFTPHONES. Such an initiative can also contribute to reducing the environmental burden of other raw materials, but it is advisable to focus on the largest hotspot, gold, first. Quantifying the potential of these measures is the subject of future work.

3) What has not yet been taken into account in this study is the possibility of integrating still functioning components from old devices into new devices. In the case of the LCA cut-off approach, their environmental impact would not be credited to the new device, as it is already taken into account in the old device. Refurbed devices, where only defective components need to be replaced in order to make them functional again beyond 5 years, can also lead to a further reduction in the environmental burden. Such end-of-life measures are already being implemented or planned as part of the circular economy strategy of the SHIFT GmbH seeking to implement value-retention processes, but the quantification of their effect is the subject of future research. As repairing devices while using new components is not automatically the most resource-efficient way, other options for extending the lifetime and best possible utilisation of the existing components and devices should also be modelled and compared²⁴.

2.3 Magnitude of savings

In the period between 2018 and 2022, an average of 22 million smartphones were sold in Germany per year. The climate and material footprint (RMI) of the use cases M2x2.5 and M5 scaled up to 22 million are very small compared to the total climate and material footprint (RMI) of Germany. The possible savings between M2x2.5 and M5 only account for a maximum of 0.02% of the total German footprints. However, savings can be achieved here purely through a change in design in combination with adapted usage behaviour without any further restrictions. The potential total CO₂ savings are in the order of 150,000 medium haul flights (at 500 kg CO₂-eqv. per flight according to German Federal Environment Agency), while resources (RMI) can be saved in the order of 220 multi-family houses (at 900 t RMI²⁶ per house) in total.

The savings potential in relation to the globally produced quantity of 2 billion smartphones in 2022 amounts to almost 13 million medium distance flights and the construction of approximately 20,000 multi-family houses. Increasing the reparability of short-lived small electrical devices to extend their lifetime can make an important contribution to saving resources and emissions.

This study has quantified for the first time the effect of high reparability of a smartphone on the environmental impact of the production and use phase per smartphone use of one year within a comprehensive LCA hotspot analysis: the use of natural resources and the emission of climate-damaging gases, and thus the total environmental burden, can be reduced by an average of 40%. Hotspots of resource use and environmental impacts along the supply chain can be considerably defused. The greatest possible savings are in the area of gold mining, which takes place worldwide at great environmental and social cost²⁷. Extrapolated to the number of smartphones produced annually in Germany or worldwide, large amounts of resources and emissions can be saved. The prerequisite is that already the design of a smartphone is geared towards the greatest possible modularity²⁸ and that spare parts are available for the replacement of defective components in order to considerably extend the lifetime. The degree of reparability decisively determines the savings potential as demonstrated within this study. The smartphone examined has the highest modularity of all smartphones on the market, but is only produced in very small numbers. Here, the major smartphone manufacturers are called upon to make reparability a priority as early as the design process by adopting modular product designs as part of their circular economy strategies. On 22 March 2023, the European Commission launched the legally binding framework conditions for this with the “proposal on common rules promoting the repair of goods”, subject to a pending approval by the European Parliament and Council²⁹. The “right to repair” is intended to enable customers to demand repairs within and beyond the legal guarantee.

In the design process it must be ensured that design tensions that require trade-off solutions are avoided, e.g. that modular smartphones are more expensive although they have a simpler design or that modularity prevents upgrades due to rapid development of software and hardware³⁰. If the design process does incur higher lifecycle-wide costs compared to conventional smartphones, referred to as “environmental activation energy” of modularity³¹, large companies that produce and sell smartphones in large quantities are probably able to absorb these. At the same time, modular design can also offer economic chances through cost reduction and employment opportunities as part of a circular economy strategy based on value retention¹³. And reparability can save natural resources not only in smartphones: in view of the Sustainable Development Goals of the United Nations and the Paris Agreement, savings through reparability should also be analysed and strived for in the design and use of other consumer electronics to contribute to the urgently needed worldwide reduction in the use of natural resources²⁵.

3.1 Phases of the life cycle assessment

An LCA consists of four steps specified in DIN ISO EN 14040³². During the determination of the goal and scope of the analyses (phase 1) the object of study is identified as the smartphone SHIFT6m manufactured by the company SHIFT GmbH, referred to as SHIFTPHONE. We consider two different uses cases with different lifetimes of the SHIFTPHONE: Use case M2x2.5 assumes a typical mobile phone lifetime of 2.5 years, which is optimistically estimated for a smartphone. It serves as reference model for a standard smartphone or mobile phone¹⁴. Use case M5 assumes a lifetime of 5 years through replacement of defective modules (internal information on usage behaviour of SHIFTPHONE customers and repair statistics according to SHIFT GmbH). The need for replacement modules over the use of five years, or to extend the lifetime to five years, is determined with the help of the repair statistics of SHIFT GmbH, which were compiled for the SHIFTPHONE from 2018 to 2022 (Supplementary Table 1). To make the two models comparable, a usage period of five years is considered in each case, which means that two SHIFTPHONES with a lifetime of 2.5 years each have to be considered for M2x2.5. The functional unit provided by the SHIFTPHONE is defined as use of the smartphone for one year considering the production and the use phase (“cradle-to-use”).

During the inventory analysis (phase 2), the models are set up in the LCA database ecoinvent 3.8³³ with the software openLCA 1.11.0 (www.openlca.org). Based on the supply chain analysis of SHIFT GmbH, separate processes are created for the replaceable modules and linked with a corresponding transport process corresponding to the real transport routes.

Product systems are being developed that connect the models to the LCA data set and represent the supply chain. In relation to eight specific mineral resources, the upstream value chains in the data basis were previously modified.

During the impact assessment (phase 3), the models are assessed with a set of selected mid-point footprint indicators that refer directly to the environmental impact categories. The calculation of the models contribution to various impact categories is performed and results are prepared for a spatially explicit hotspot analysis.

During evaluation (phase 4), the models are analysed in comparison and the supply chains are analysed spatially explicit. Uncertainties are determined using Monte-Carlo simulations.

3.2 Life cycle inventory analysis

Both models are mainly based on the input of the SHIFTPHONE. It is modelled according to the regionalised supply chain analysis provided by the SHIFT GmbH down to tier 3. All available information on the type and location of the manufacturers is taken into account in the model. Exchanged camera lenses and audioports could not be considered because suitable LCI data sets are missing in consequence of data shortages (Supplementary Fig. 4).

In general, the inventory was fed with data from the following sources in descending order: ecoinvent 3.8, literature data, estimates from own weighing of components, assumptions (Supplementary Fig. 4).

Parts of the SHIFTPHONE model were used to model the exchange modules for M5 and equipped with their own transport process under the assumption that they are purchased directly from suppliers. Accordingly, delivery processes and also transport are taken into account. This differs from the reality at SHIFT GmbH, where replacement modules can also come from old equipment returned by customers and would then be classified as burden-free according to the LCA cut-off approach. However, this practice is subject to dynamic conditions that cannot be covered by an LCA. It is the subject of further research.

For the usage phase, the electricity consumption per year is determined based on the charging capacity of the SHIFTPHONE battery and an estimated average number of charging cycles of 114.

3.3 Regionalisation of supply chains of selected mineral commodities

Up to the third upstream processing step (tier 3), the supply chain of the SHIFTPHONE is known (Supplementary Fig. 4, according to SHIFT GmbH), but not the origin of the raw materials. By updating a dataset of Schomberg et al. 2022²² to ecoinvent 3.8, we are able to regionalise the supply chains of the mineral commodities aluminium, copper, coal, cement, iron and steel, lithium and phosphorus at mine site level. To make the number of mine sites globally manageable, single mine sites are clustered with respect to geology and regional water stress, resulting in five mine sites per country that are added to ecoinvent 3.8 and linked to the existing datasets. The linkage is performed under consideration of country level and mine site production figures, which makes the dataset a model of country and world markets of the listed commodities. 80% of world production are covered, respectively. The procedure is described in detail by Schomberg et al. 2022²². Supply chains modeled using these datasets include the most likely origin of a given mineral commodity in the global supply system unless more specific information becomes available.

3.4 Life cycle impact assessment

For impact assessment, the resource footprints for energy, land, material and water are chosen, which have been found to cover more than 80% of the variance of all environmental impacts³⁴, and the climate³⁵ footprint (Table 3). Footprints are values weighted with respect to specified criteria³⁶ that are determined with selected LCIA methods to quantify and assess the inputs of elementary flows along the SHIFTPHONE supply chain.

The climate footprint³⁵ is determined following an update of the LCA implementation IPCC 2013¹⁵ with latest IPCC data from 2021¹⁶. We use the impact category climate change GWP100a (GWP₁₀₀)¹⁵, referred to as global warming, to quantify the elementary flows of carbon dioxide, carbon monoxide, chloroform, dinitrogen monoxide, different ethane and methane compounds, nitric oxide, nitrogen fluoride, sulphur hexafluoride as well as volatile organic compounds. Assessment is performed with respect to the global warming impact in kg CO₂-equivalents kg^− 1, respectively.

The energy footprint is based on the Cumulative Energy Requirements Analysis^37,38 from the LCA implementation by Hischier et al. 2010¹⁵. We limit ourselves here to the fossil cumulative energy demand, which consists of the energy supply of hard coal, lignite, crude oil, natural gas, mine gas, peat, uranium as well as wood and biomass from primary forests. Assessment is performed with respect to the energy content in MJ equivalents m^− 3, kg^− 1 or MJ^− 1 by multiplication with corresponding characterisation factors.

The land footprint is described by the total area of land occupied through the supply chain of the SHIFTPHONE in m²\(\times\)a. Here, no assessment is performed in the absence of a suitable method recommended by the Life Cycle Initiative²².

The material footprint³⁹ comprises the sub-indicators Raw Material Input (RMI) and Total Material Requirement (TMR). For their calculation, the use of abiotic and biotic materials is determined. The RMI is calculated using the ratio of extracted raw material to the corresponding material in the extracted raw material measured in kg raw material per kg material. The TMR uses extracted primary material (used and unused extraction) measured in kg primary material per kg material.¹⁸

The water footprint¹⁹ consists of the sub-indicators Quantitative Water Scarcity Footprint (WSF_quan) and Qualitative Water Scarcity Footprint (WSF_qual). WSF_quan is the quantitative water consumption composed of evapotranspiration, water incorporated into a product and transboundary water transport beyond the catchment areas in m³. WSF_qual refers to the virtual water volume (in m³) required to ensure compliance with the threshold for the concentration of aluminum emissions that occur as result of the process. Assessment is performed with respect to water stress on country level according to the LCIA method AWARE⁴⁰, respectively.

Results of indicator sub-categories are added up. In case of the WSF_quan, negative values from rounding errors in the ecoinvent 3.8 database are removed.

3.5 Hotspot analysis of LCIA results

With an LCIA, the overall footprint results are shown and also the individual contributions of activities in the upstream supply. The following is a systematic analysis of the LCIA data to make the essential information comparable:

1) The supply chain of the SHIFTPHONE is composed of approximately 14,000 individual activities, which is why only those with a share greater than 1% of the total footprint results are used for further analysis. This means that at least 44%, but on average 80% (without outlier), of the total footprint results are included (Supplementary Table 4). An analysis of the remaining activities with very little contribution would be too extensive and mislead the key messages. The term “total environmental burden” in this study therefore refers to the proportion under investigation, even if this is not 100%.

2) For the purpose of comparability of activities (with various units), the further procedure is to normalise results per footprint: individual activity results are normalised by the median of all activities of both use cases⁴⁸. Normalised values are obtained from the quotient of the individual activity results and the median.

3) The representation of the normalised value can be understood as the activity result being x times as large as the median, where x stands for the normalised value. If the value is below 1, the activity event is less than the median and therefore not a hotspot. For clarification and differentiation of the hotspots with different characteristics according to their normalised value, a classification is made using a color scale. Here, we adapt the colour scale in the lower range of the scale, because the SHIFTPHONE shows less distinct hotspots like the case studies used in Schomberg et al. 2022²². Colours divide the hotspots according to severity into emerging (1–5, dark blue), low (5–10, rose), medium (10–15, yellow), high (15–20, orange), very high (20–50, dark red) and extreme severity (> 50, dark brown).

4) Since locations can occur multiple times because varying activities proceed at the same location or one activity causes multiple exposures, all individual normalised activity results are summed for each location.

5) The quality of the regionalisation is tied to the quality of the input data, which range from point coordinates to global coordinates, where global is synonymous with unknown. Analysing international supply chains results in large amounts of data that may contain inaccuracies in regionalisation. To remedy this problem, the degree of regionalisation is presented with the help of a quality index for locations: A quality index of 1 represents locations with point coordinates. A quality index of 2 reflects countries or sub-countries and quality index 3 regions of two or more countries. Quality 4 depicts unknown sites that have the designation global or rest-of-world which are not considered in the spatial hotspot analysis.

6) In order that the significance of individual activities for the overall environmental burden of the use cases can be assessed independently of the occurrence of spatial information, these are grouped into activity categories. Activity hotspots are identified, displayed and evaluated in the same way as spatial hotspots.

3.6 Data quality evaluation, assumptions and limitations

Apart from the data the SHIFT GmbH could provide from a supply chain analysis up to tier 3 no specific manufacturer data on the components and manufacturing processes are available. The structure of the LCA model is based on weighing the individual components and parts in order to reflect the composition of a SHIFTPHONE as accurately as possible. The model was largely filled with adapted ecoinvent 3.8 data, which in most cases are not specific to smartphone components, e.g. the data set "mainboard, surface mounted". The database contains a few data sets for smartphones that have been adapted for the SHIFTPHONE. In rare cases, literature data or estimates are used. Camera lenses and audioports are the only components that could not be considered at all due to a lack of suitable LCI data.

The use of non-specific data paired with the data estimates or gaps results in a certain uncertainty of the data, which we have assessed within an LCA data quality matrix according to the criteria reliability, completeness, temporal correlation, geographical correlation and further technological correlation (Supplementary Tables 5 and 6). It shows that most data are of medium to high quality while the temporal correlation is often only poor. A comprehensive validation of the model, for example via analyses of the element composition, goes beyond the scope of this work and will be the subject of further studies.

Data availability statement

The LCI of the SHIFT6m and updated LCIA methods will be available from Mendeley Data when published.

Author Contributions

A.C.S. and C.M. designed the LCA models and the conceptual approach. A.C.S. performed the analyses, wrote the paper, and created the figures and the SI. C.M. and S.B. comprehensively reviewed and edited the manuscript. T.K and L.v.Z. provided comprehensive insights into SHIFT GmbH, provided data and proofread all passages on the SHIFTPHONE. All authors discussed on the LCA models, the analysis and the manuscript at all stages.

Acknowledgements

This research work was performed as part of the project “Closing the loop: SHIFTPHONE is ready for circular economy”, loopPHONE (033RK087B), carried out with the support of the Federal Ministry of Education and Research (BMBF). We thank María Motiño and Tabea Schneider, who have assisted with creating the LCA inventory and literature research.

Competing Interests

The authors declare no competing interest.

Bekaroo, G. & Seeam, A. Improving wireless charging energy efficiency of mobile phones: Analysis of key practices. 2016 IEEE Int. Conf. Emerg. Technol. Innov. Bus. Pract. Transform. Soc. EmergiTech 2016 357–360 (2016). doi:10.1109/EmergiTech.2016.7737366
Bekaroo, G., Bokhoree, C. & Pattinson, C. Impacts of ICT on the natural ecosystem: A grassroot analysis for promoting socio-environmental sustainability. Renew. Sustain. Energy Rev. 57, 1580–1595 (2016).
Gurita, N., Fröhling, M. & Bongaerts, J. Assessing potentials for mobile/smartphone reuse/remanufacture and recycling in Germany for a closed loop of secondary precious and critical metals. J. Remanufacturing 8, (2018).
OECD Enviroment Directorate. Materials Case Study 1: Critical Metals and Mobile Devices. OECD Glob. Forum Focus. Sustain. Mater. 1–84 (2010).
UNEP International Resource Panel. Global Metal Flows Working Group Report 3: Environmental Risks and Challenges of Anthropogenis Metals Flows and Cycles. (2013).
Elenge, M. M. & De Brouwer, C. Identification of hazards in the workplaces of artisanal mining in Katanga. Int. J. Occup. Med. Environ. Health 24, 57–66 (2011).
Nkulu, L., Casas, L., Haufroid, V. & Putter, T. De. Sustainability of artisanal mining of cobalt in DR Congo. Nat Sustain. 1, 495–504 (2018).
Park, S. H. et al. Exposure to volatile organic compounds and possibility of exposure to by-product volatile organic compounds in photolithography processes in semiconductor manufacturing factories. Saf. Health Work 2, 210–217 (2011).
Parajuly, K. et al. Future e-waste scenarios. (2019).
Perkins, D. N., Brune Drisse, M. N., Nxele, T. & Sly, P. D. E-waste: A global hazard. Ann. Glob. Heal. 80, 286–295 (2014).
Forti, V., Baldé, C. P., Kuehr, R. & Bel, G. The Global E-waste Monitor 2020. (2020).
Kyere, V. N. et al. Contamination and health risk assessment of exposure to heavy metals in soils from informal e-waste recycling site in Ghana. Emerg. Sci. J. 2, 428–436 (2018).
Nasr, N. et al. Redefining Value the Manufacturing Revolution. International Resource Panel (2018).
Abbondanza, M. N. M. & Souza, R. G. Estimating the generation of household e-waste in municipalities using primary data from surveys: A case study of Sao Jose dos Campos, Brazil. Waste Manag. 85, 374–384 (2019).
Hischier, R. et al. Implementation of Life Cycle Impact Assessment Methods. (2010).
Forster, P. et al. The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA (2021). doi:10.1017/9781009157896.009.923
Kaiser, S., Prontnicki, K. & Bringezu, S. Environmental and economic assessment of global and German production locations for CO2-based methanol and naphtha. Green Chem. 23, 7659–7673 (2021).
Mostert, C. & Bringezu, S. Measuring product material footprint as new life cycle impact assessment method: Indicators and abiotic characterization factors. Resources 8, (2019).
Schomberg, A. C., Bringezu, S. & Flörke, M. Extended life cycle assessment reveals the spatially-explicit water scarcity footprint of a lithium-ion battery storage. Commun. Earth Environ. 2, 1–10 (2021).
DIN-Normausschuss Bauwesen. Nachhaltigkeit von Bauwerken - Umweltproduktdeklarationen - Grundregeln for die Produktkategorie Bauprodukte; Deutsche Fassung EN 15804:2012+A2:2019 + AC:2021. 76 (2023).
Proske, M., Sánchez, D., Clemm, C. & Baur, S.-J. Life cycle assessment of the Fairphone 3. (2020). doi:10.1109/EVER.2014.6844034
Schomberg, A. C., Bringezu, S., Flörke, M. & Biederbick, H. Spatially explicit life cycle assessments reveal hotspots of environmental impacts from renewable electricity generation. Commun. Earth Environ. 3, 1–14 (2022).
Schischke, K., Proske, M., Nissen, N. F. & Schneider-Ramelow, M. Impact of modularity as a circular design strategy on materials use for smart mobile devices. MRS Energy Sustain. 6, 1–16 (2019).
von Gries, N. & Bringezu, S. Using New Spare Parts for Repair of Waste Electrical and Electronic Equipment? The Material Footprint of Individual Components. Resources 11, (2022).
Bringezu, S. Das Weltbudget. Das Weltbudget (2022). doi:10.1007/978-3-658-37774-8
Sameer, H. & Bringezu, S. Building information modelling application of material, water, and climate footprint analysis. (2021).
Mestanza-Ramón, C. et al. Gold Mining in the Amazon Region of Ecuador: History and a Review of Its Socio-Environmental Impacts. Land 11, 1–22 (2022).
Nissen, N. F., Schischke, K., Proske, M., Ballester, M. & Lang, K.-D. How modularity electronic functions can lead to longer product lifetimes. Prod. Lifetimes Environ. 303–308 (2017). doi:10.3233/978-1-61499-820-4-303
European Commission. Proposal for a Directive of the European Parliament and of the Council on common rules promoting the repair of goods and amending regulation. (2023).
Revellio, F., Shi, L., Hansen, E. G. & Chertow, M. Sustainability paradoxes for product modularity : the case of smartphones. Electron. Goes Green 2020 121–130 (2020).
Schischke, K. et al. The ‘Environmental Activation Energy’ of Modularity and Conditions for an Environmental Payback. in Towards a Sustainable Future - Life Cycle Management (eds. Klos, Z. S., Kalkowska, J. & Kasprzak, J.) 15–25 (2022). doi:10.1007/978-3-030-77127-0_20
Deutsches Insitut für Normung e. V. DIN EN ISO 14040:2006: Umweltmanagement – Ökobilanz – Grundsätze und Rahmenbedingungen. (2016). doi:10.1007/s00738-009-0685-2
Wernet, G. et al. The ecoinvent database version 3 (part I): overview and methodology. Int. J. Life Cycle Assess. 21, 1218–1230 (2016).
Steinmann, Z. J. N., Schipper, A. M., Hauck, M. & Huijbregts, M. A. J. How Many Environmental Impact Indicators Are Needed in the Evaluation of Product Life Cycles? Environ. Sci. Technol. 50, 3913–3919 (2016).
Egenolf, V. & Bringezu, S. Conceptualization of an indicator system for assessing the sustainability of the bioeconomy. Sustain. 11, (2019).
Deutsches Insitut für Normung e. V. DIN EN ISO 14046:2016-07: Umweltmanagement – Wasser-Fußabdruck – Grundsätze, Anforderungen und Leitlinien (ISO 14046:2014). (2016).
Pimentel, D. et al. Food production and the energy crisis: A comment. Science (80-. ). 187, 560–567 (1975).
Boustead, I. & Hancock, G. F. Handbook of industrial energy analysis. (1979).
Sameer, H. et al. Environmental assessment of ultra-high-performance concrete using carbon, material, and water footprint. Materials (Basel). 16, (2019).
Boulay, A. M. et al. The WULCA consensus characterization model for water scarcity footprints: assessing impacts of water consumption based on available water remaining (AWARE). International Journal of Life Cycle Assessment 1–11 (2017). doi:10.1007/s11367-017-1333-8

Table 3 is not available with this version.

There is NO Competing Interest.

SIfootprintsshiftphoneshowpotentialofmodularity22032023.docx
Supplementary Information
SupplementaryData1.txt
Dataset 1
SupplementaryData2.txt
Dataset 2
SupplementaryData3.txt
Dataset 3
SupplementaryData4.txt
Dataset 4

Download PDF

Version 1

posted

You are reading this latest preprint version

Environmental footprints show the savings potential of high reparability through modular smartphone design

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Conclusion

Materials And Methods

Declarations

References

table

Additional Declarations

Supplementary Files

Status:

Version 1