EV Emissions: A Life-Cycle Approach

1 Introduction

1.1 Battery Electric Vehicles

With zero tailpipe emissions, the positive impact of electric vehicles (EVs) on air quality in urban centres is widely accepted. Regarding their global environmental impact and life cycle carbon emissions, they are regularly subject to debate.

EVsInstead of an internal-combustion engine,  operate on an electric motor that generates power by burning a mix of fuel and gases. There are two basic classes: all-electric vehicles and plug-in hybrid electric vehicles. The former include Battery Electric Vehicles (BEVs) and Fuel Cell Electric Vehicles (FCEVs). Besides charging from the power grid, both types are charged in part by regenerative braking, which recovers some of the energy typically lost when braking [1]. We will, from here on, refer to BEVs when mentioning the concept of electric vehicles.

A BEV utilises chemical energy stored in rechargeable battery packs its prominent feature is that drivers plug them into an off-board electric power source. Otherwise, hybrids supplement an internal combustion engine (ICE) with battery power but cannot be plugged in. Also, BEVs have traditionally been developed as city cars, while small ICE supplement electrical systems in extended-range vehicles. BEVs just run on electricity, with most having 80 to 100 miles (130 to 160 km) of all-electric range, while a few luxury models have capacities of up to 250 miles (400 km). Depending on the type of charging device and battery, when its power source is depleted, it can take 30 minutes (with quick charging) to several hours to recharge it.

With a GHG intensity of electricity generation equal to the global average (518 grammes of carbon dioxide per kilowatt-hour in 2018), BEVs, hybrid electric vehicles and fuel cell electric vehicles have similar lifetime emissions, lower than those of an average internal ICE vehicle (Figure 1).

Though the concept of electric vehicles has been around for a long time, it has gained a considerable amount of interest recently amid rising carbon footprint and air quality concerns. Electric vehicles are part of a scheme around the world focused on advancing environmentally sustainable transportation. Let’s discuss why life cycle assessments might be necessary when reporting greenhouse gas emissions, especially as we look out into the future, and we start adopting more electric vehicles.

1.2 Why Life Cycle Assessment?

To find out if an EV is better for climate, we must reason not in combustion, but life cycle: we will count the emissions due to the manufacture of the car, batteries, extraction and transport of fuel if necessary. Due to the lack of adequate evaluation tools, the indirect environmental effects of transportation disruptions on urban mobility have often been ignored. However, some trends are making life cycle assessment progressively critical [2]. Research institutes and regulatory bodies are developing methods to capture the environmental consequences of these disruptions’ entire cause and effect chain while adapting them to the city, regional, and national scales.

The term life cycle assessment or LCA refers to a structured environmental assessment method to quantify ecological flows and their impacts for a product or service from a cradle-to-grave perspective. We are therefore tracking material and energy inputs, waste and pollution outputs at each life cycle stage; but also analysing the supply chains that deliver those material and energy inputs, and pondering transportation at each stage. Life cycle assessments traditionally consider a whole range of environmental impact categories.

However, many, EV lifecycle-based studies and automotive evaluations have focused on greenhouse gas (GHG) LCA, often referred to as carbon footprints. This is acceptable when the main concern is about climate change impacts, but if we want to understand the full burden of impacts associated with the technology we prefer to think about multiple environmental aspects not to trade one impact for another. Nonetheless, this article will focus only on carbon footprints or LCA of GHG emissions.

Trends in vehicle design, fuels, and users needs are shifting environmental impacts away from the operational life cycle stage [3]. The pie charts in Figure 3 depict gasoline, a conventional fuel that modern societies have depended on for a long time for passenger vehicles, with combustion dominating impacts such as greenhouse gas emissions and energy consumption. Likewise, impacts from cars have been dominated by the operation.

Today, the use phase is the largest contributor to life-cycle GHG emissions of all powertrains. But as we look at alternative fuels and trends towards electric vehicles that run on electricity and hydrogen, the upstream or life cycle-related impacts are increasingly important. As vehicles move towards electrification, production-related emissions exhibit an increasing dominance. These vehicles are more efficient, but as grids tend to decarbonise, GHG emissions and many other air pollutants are ipso facto reduced.

1.3 Background: LCA of BEVs

Existing studies assessing BEV emissions have different life cycle scopes, which may include or exclude: vehicle and battery manufacturing emissions; power plant emissions from electricity generation; power plant fuel feedstock extraction, fuel extraction, processing, transportation, and combustion emissions; production and transportation emissions; and end of life emissions. Several studies have only included vehicle use or a subset of the emissions related to vehicle use (e.g., vehicle tailpipe emissions and power plant smokestack emissions), leading to incomplete assessments. Life cycle studies suggest that emissions implications from sources other than those two examples can comprise one fifth to one-third of vehicle life cycle GHG emissions, so addressing the full life cycle can be necessary for comprehensive comparisons.

Auke Hoekstra, Senior Advisor on Electric Mobility at the Eindhoven Technical University, is probably the world’s leading expert on cars’ life-cycle emissions. In a recent paper [4], he states that GHG emission reductions possible with BEVs are underestimated in the scientific literature. His thorough review work points out the following causes behind this issue: an overestimation of battery manufacturing, an underestimation in battery lifetime, a static electricity mix over the lifetime of the EV, unrealistic tests for energy use, exclusion of fuel production emissions, and lack of system thinking.

Moreover, previous LCA studies of BEVs have almost uniformly considered small, efficiency-oriented electric vehicles modelled on first-generation cars with approximately 25-kilowatt-hour batteries. Most of these studies have found that the electricity grid composition that charges the EVs and the climate they operate is crucial to understanding their performance [5] [6].

For instance, Specialists at Carnegie Mellon University compared life cycle GHG emissions from several light-duty passenger gasoline- and plug-in electric vehicles across US counties by accounting for driving conditions (city versus highway), ambient temperature, regional differences due to marginal grid mix, and patterns of vehicle miles travelled [7]. They found that EVs can have larger or smaller carbon footprints than gasoline vehicles, depending on these regional factors and the specific vehicle models being compared. Thus, one should look at the spatial and temporal dynamics of where these vehicles are operated and how the grids that they operate on might change over time (Figure 4). In addition to thinking about future electricity grids, it becomes crucial to discuss vehicle design trends.

2 Modelling Approaches

2.1 Trends in battery cost and performance

Some of the trends in vehicle change described in the following section can be explained, in part, by dramatic decreases in battery cost and input increasing performance of electric vehicle batteries. Figure 5 is a projection of the changing price. It stands out that the trends towards, for example, high performance and large vehicles might be something continuing to happen because of battery costs coming down. What happens to our understanding of battery electric vehicles’ life cycle impacts when considering some of these critical trends?

Regarding vehicle design and market trends, with these decreasing battery pack costs, we might see increasing battery pack size even in efficiency-oriented vehicles and greater dominance of luxury and high-performance cars in the BEV sector. We might also see the adoption of shared and autonomous vehicles, especially high-mileage applications, enabled by these larger batteries. Besides, as electricity grids decarbonise, mostly where BEVs are being adopted fastest like in California, the operation-related impacts of electric vehicles dramatically reduce but, potentially, the production-related implications of these vehicles increase. There are implications of these trends.

On the one hand, large and high-performance electric vehicles mean larger batteries and designs that may not focus on efficiency are likely to be bad for GHG emissions. Besides, increasing investment and interest in lightweight materials to deliver on performance and efficiency could also occur.

On the other hand, and likely to be suitable for emission reduction, an increase in range for these vehicles — going from a hundred miles to three hundred miles (160 km to 480 km), for instance — would mean that batteries might last a lot longer because they are not being cycled as deeply in terms of a depth of discharge which is a crucial deterioration process for batteries. Then, barriers to adoption could be reduced by including a higher mileage operation. We will provide some results where our goal is to give readers a preliminary LCA focusing on production and use of new-model and future-model BEVs considering only the most important Spatio-temporal dynamics like ageing electricity fuel mixes.

2.2 Vehicle scenarios

The majority of previous studies reflect outmoded assumptions about BEV designs and have not reflected trends in the car market. A sales review between 2012 and 2018 in the US shows a remarkable shift towards longer vehicle ranges, considerably higher capacity batteries, and a growing preference for high performance and luxury BEVs. The combined effect of these two trends results in an increase in BEV battery capacity of 6.5 kWh per year, from 20 kWh in the first quarter of 2012 to more than 70 kWh in the second quarter of 2018. As the market has grown, so too have the number of models available.

We notice significant changes in the types of vehicle that were being sold. The Nissan Leaf, shown in navy blue, dominated those small sales in the beginning, to the Tesla Model S (yellow) coming online and taking a reasonably large share of the market. For example, looking out to 2018, larger vehicles like the Tesla Model 3 and the Tesla Model X become best sellers. These figures should be highlighted because various electric cars can affect what we can say about their life cycle impact.

The recommended approach when accounting for vehicle diversity in LCA-based studies is to take a minimum of three archetypal BEVs, i.e. an efficiency-oriented compact vehicle (EOV), a high-performance luxury sedan (PLS), and a high-performance SUV (PSUV). Researchers and policymakers also need to explore some alternative use models such as shared and autonomous applications. To do this, they are sought to combine existing, and new, LCA models for vehicle components and systems, such as the battery LCA models and GREET from Argonne National Lab [8]. Some methods to estimate use phase impacts under different scenarios including annual mileage estimates from the National Highway Transportation survey or similar databases, and FASTSim, an estimator of BEV energy efficiency that assumes combined city-highway energy economy.

Studies should consider how future vehicle design changes, trends in battery performance, cleaner electricity grids, and annual mileage will affect each vehicle archetype’s total life cycle emissions. EV emissions coming from driving are supposed to be focused on the electricity mix over the vehicle’s whole endurance period. This leads to 55 gGHG per km for Europe. To calculate this accurately for a specific EV, assumptions must be made about the lifetime, e.g. 17 years, yearly mileage (let’s assume 26k km in the first year and 1k less per year), and energy use per km (0.161 kWh/km based on EPA measurements). Results are then usually presented in a functional unit of vehicle mile (or kilometres) travelled, where total emissions are divided by the vehicle’s lifetime miles. This facilitates comparisons with ICE vehicles.

2.3 Trends in electricity grids

As grids continue to get cleaner, i.e. they move towards lower emissions, both new and used electric vehicles will get cleaner as well. This is an asset EVs have over fossil-fueled cars: their emissions can get better over time. The fuel economy of gasoline vehicles is fixed and so are their emissions, as long as they depend primarily on petroleum for fuel. A future in which cars themselves, but the entire supply chain of cars operate on renewable energy could be imagined.

Batteries could power mining equipment that retrieves the ore from which windmills, solar panels, and batteries are made. Green hydrogen produced by solar and wind makes steel and aluminium zero-emission, which in turn makes the manufacturing of cars, batteries, and other clean technologies almost zero-emission. Car batteries also absorb excess solar and wind energy, stabilise the grid, and decrease the available stationary batteries. It is not an overstatement to say that the GHG emissions from batteries could be further reduced by a factor of ten in such a scenario.

Global electricity demand from electric vehicles hits 550 TWh in 2030 in the Stated Policies Scenario (SPS), which is about a six-fold increase from 2019 levels. According to the International Energy Agency projections, the share of demand due to EVs in total electricity consumption grows to as high as 4% in Europe. In the Sustainable Development Scenario (SDS), with demand rising nearly eleven-fold relative to 2019, to almost 1,000 TWh, the total market share ranges from 2% in Japan to 6% in Europe [9].

Electricity grid trends are usually modelled through a so-called business as usual scenario (BAU), which is essentially the fuel mix for ongoing conditions in a specific electric grid. In the US, for instance, increasing proportions of renewable sources are observed, yet, one would not find the same trend towards a profound reduction in fossil fuels. Consequently, emissions are evaluated under a reference case or BAU, and a carbon tax scenario assumes an allowance fee on CO2 emissions from utility-scale electricity generators.

Carbon tax scenarios are included to represent the potential impact of further changes to the grid mix, particularly for in-use vehicles, and the magnitude of possible change for the average vehicle fleet. Carbon pricing is a policy instrument aimed at minimising the amount of carbon dioxide released into the atmosphere by placing either a tax or a limit on the tons of carbon dioxide emitted. When applied to the electricity sector, carbon pricing discourages the use of carbon-intensive generators in favour of lower- or zero-emitting alternatives [10].

As a result, the average emissions rate (EF_t) can be estimated as the mass of GHG equivalent emissions per unit of delivered energy with the following general equation [11]:

where the weighted generation by year (t) and fuel source (x) is multiplied by the life cycle inventories (LCI) of emissions species (e) by fuel type (x), and the impact characterisation factors (m). Therefore, the resource mix can be broken into different fuel source categories: renewables, coal, natural gas, nuclear, fuel oil, etc.

3 Key Findings

3.1 Overall modelling results

Are EVs better for the climate? Yes. This section explores the relative importance of GHG emissions along various stages of the vehicle life cycle. It identifies drivers of emissions reduction for different vehicle technologies relative to each other. In their use phase, EVs have net GHG emissions reductions compared with fossil-fuelled vehicles, but this does not account for the emissions that occur during manufacturing and other moments of the vehicle’s full life of the vehicle.

In the studies compared below, the GHG intensity (expressed as kg CO2-eq/kWh) varies based on the assumed energy use for battery manufacturing and battery energy density. One important outcome here is that the GHG emission intensity of battery manufacturing per kWh tends to decline as assumptions on battery energy density, plant size and plant capacity utilisation increase.

Taking an integral system perspective further highlights the potential of batteries and BEVs.

The answer to this article’s prime question is not always apparent given the result of an extensive assessment whose report is available on the European Commission’s website — often better than thermal, but not always [12]. On the other hand, the difference is noticeable for a country like France or Sweden, where electricity is very low-carbon. In other words, the main GHG emissions reduction potential over the vehicle life cycle seems to be in the decarbonisation of the power system. In an example calculation, BEVs reduce emissions from 244 to 98 g/km. In a fully renewable system, BEV emission could decrease to 10 g/km [4].

The chart in Figure 9 shows on the left several vehicle scenarios and bars that consist of a yellow part — operation emissions assuming US average grid; a grey part — vehicle materials and glider assembly, i.e. vehicle production; and two blue parts — battery materials production and battery manufacturing itself. The first vital takeaway to acknowledge is that, compared to ICE and HEV, electric vehicles perform better, meaning they have fewer emissions per mile (or kilometres). On the bottom, we can observe the shared and autonomous vehicles. The first thing to watch is that there are great benefits of adopting electric vehicles in these applications relative to an ICE SUV.

With that said, the vehicles’ production-related impacts and the batteries are much smaller relative to the operation. That is happening because these types of vehicles are being used a lot every day. It is also noticeable the difference in relative impacts of vehicle production compared to battery changes, that is because in these high-mileage applications we have assumed battery replacement. Hence, the relative contribution of vehicle and battery production is slightly different.

The main takeaway from these results is that it does matter what case we consider in terms of the proportion of impacts of commute or operation emission in regards to production; but also that as EVs are operated in places like California, the production-related emissions are significant for understanding the life cycle impacts of the vehicle.

Driving emissions are where the BEV realises significant gains on the diesel car. Care must be taken to calculate the electricity mix over the lifetime of the BEV. For the diesel, GHG emissions during fuel production (predominantly from refineries) should be included. All energy use should be based on realistic road tests or numbers of the US Environmental Protection Agency (EPA). This leads to 55 g GHG per km for the BEV and 217 g for the diesel vehicle.

3.2 The California case

The Golden State scenario provides a useful comparison: California represents nearly 50% of the American BEV fleet, over 8% of new vehicles sold in the state are electric (compared with 2% nationally), a large share — around 50% — of its electricity is generated by renewable sources, and finally, the state has enacted progressive policies pushing further deployments of renewables and EVs [13]. There is a clear trend towards the dominance of renewables in California and the only fossil fuel remaining in the mix being some portion of natural gas.

There are significant differences between the values we see for electric vehicles operated on the average US grid and electric vehicles used in California: Life cycle emissions from BEVs under the California scenarios were ~45% lower than under the US average scenario according to a recent study [11]. In Figure 9 we have a national average for the US — the business-as-usual case with yellow bars — then different scenarios: with a carbon tax — the red x mark — and then for California (CAMX) — the black diamond. It should be highlighted that when operating an electric vehicle in California, vehicle production is often more than half of the impact.

Carbon dioxide-equivalent emitted through a life cycle approach for current market BEVs range from 136 gCO2e/mile for an efficiency-oriented compact vehicle in California up to 324 gCO2e/mile for the larger SUV in the US average scenario. In 2025, emissions decrease to 105 gCO2e/mile for an EOV in California, while potentially increasing to 374 gCO2e/mile for the PSUV in the US.

This leads to the net GHG emissions factor per kilowatt-hour that shows deep California reduction and only mild reductions for the US average under the business-as-usual scenario. However, when a carbon tax scenario was tested, it did not significantly affect California, maybe because the state is already increasing their renewables share quite rapidly.

3.3 Battery end-of-life

The term end-of-life (EOL) is the catchall for what is done with the product after being used. The EOL stage includes the disposal and recycling of the glider. It should be outsourced that although recycling, remanufacturing, and reuse are all options, in theory, the fate and reuse or recycling potential of EOL batteries is still uncertain and has not been addressed explicitly in most studies.

While EOL battery-related emissions and the potential benefits of reuse and recycling are not inherently high from a carbon perspective, other environmental impacts are concerned. The current falling costs of batteries and relatively low value of constituent elements post-recycling present challenges for robust recycling systems. Regarding battery lifetime, it is essential to remember that great strides have been made in recent years. Presently, when motor maintenance becomes too costly, cars are scrapped. But this would no longer be the bottleneck for a BEV because the electric motor outlasts most other car components without maintenance. It is estimated that current batteries last at least 1,500 to 3,000 cycles until they lose 20 per cent of their capacity, giving a battery life of 450k to 1,350k km to an electric car with a range of 450 km. Moreover, advances such as solid-state electrolytes could further increase cycles to between 5,000 and more than 10,000 in 2030 while also making batteries nonflammable.

The Battery Directive has set industrial and automotive battery waste management standards in the European Union, including banning landfilling and incineration,  mandating companies to collect and recycle them instead. It also fixes a minimum required recycling rate for each category. Nonetheless, this strategy was decided based on battery technologies that are no longer dominant and is subject to revision [14].

Volumes of Li-ion batteries are expected to soar with the deployment of EVs. Therefore, the legal structure will benefit from being strengthened and modified to consider the new complexities of the transition to EVs and related batteries. This could be enforced, for example, via recovery mandates for each critical battery material separately, instead of a fixed rate for the battery as a whole.

3.4 LCA-based policies

The global policy context is set to regulate vehicle efficiency, fuel economy, and tailpipe emissions. The US and California already think about the life cycle of fuels in some contexts (e.g. the Low Carbon Fuel Standard (LCFS) and RFS2 in a more limited way. But we do not find a general context for thinking about the whole vehicle life cycle (Figure 11).

By omitting life cycle emissions, policymakers could develop a paradoxical policy outcome meaning that even when intending to favour vehicles with lower GHG emissions, cars with higher emissions could result favoured instead. Another question is whether there are benefits to life cycle based policies, such as flexibility for meeting GHG targets for vehicles. Preferences for vehicle light-weighting actions, for example, can also require a life cycle assessment to understand benefits. Indeed, one of the significant challenges to deal with is that a lifecycle-based vehicle policy could be admittedly complex and challenging to achieve if not properly developed and managed.

When looking for actual lifecycle-based policies, we would find only one class of policies that have implemented life-cycle carbon accounting or greenhouse gas assessment, the low carbon fuel standards. In California, policymakers implement it through a standard government-provided model.

Providing a tool and default carbon intensity estimates, the risk of high variability and a lack of transparency is reduced. Without that kind of common modelling framework, there would be a considerable amount of flexibility for vehicle producers to play around with and, eventually, generate bad examples such as ‘Astongate’. Earlier this year, green experts hit out at data underpinning an EV life-cycle CO₂ assessment report backed by Aston Martin and Bosch [16]. The two firms were accused of using controversial figures to downplay EVs’ environmental benefits, with the former forced to delay the launch of its EVs until 2025.

Figure 12 has been extracted directly from the low carbon fuel standard website. What we can see is that these fuel pathways that the initiative has come up with. One can request a fuel pathway to receive a carbon intensity score; the tool then provides default carbon intensity estimates. However, there is also the option for vehicle producers to request their carbon intensity scores.

4 Conclusion

This article demonstrates a qualitative framework for understanding GHG emissions’ direct and indirect impacts in the electric transportation sector. By an LCA approach, we find essential feedback loops that allow for identifying and discussing trends in vehicle model scenarios, electric grids, unintended consequences, and policy synergies. The role of qualitative analysis is vital in the policymaking process, especially in electromobility analysis, where many key decision factors are uncertain or unknowable to date.

Assessing whether or not EVs bring about overall net reductions in greenhouse gas emissions to other powertrain options requires a life-cycle analysis. The extent to which a contemporary average BEV emits less GHG than a regular ICE vehicle depends on the electricity generation mix’s carbon intensity in the use phase and the car’s size. The main GHG emissions reduction potential over the vehicle life cycle is in the power system’s decarbonisation.

Electric cars are now contributing to much lower air emissions than their gasoline-powered counterparts, and are becoming even cleaner. Optimising the manufacturing of EVs and the recycling or reuse of batteries may increase their environmental benefits further. And as electricity becomes cleaner (which it is already happening), the difference between them will only grow — cementing electric vehicles’ role in cutting global warming emissions.

Life cycle-based policies are relevant for internal-combustion-engine and electric vehicles as well. The effect of light-weighting on life cycle emissions demonstrates that there are cases where achieving lower emissions during the operation of a vehicle through light-weighting materials could lead to significant increases in life cycle emission. Hence, it is crucial to think about the actual reactions and trade-offs when trying to achieve total life cycle emissions reductions, even for conventional cars.

The LCA approach may also be used as follows: i) to provide a platform by which vehicle manufacturers can engage in carbon footprint discussions (since the formulation of an LCA forces them to systematically engage in a discussion of essential assumptions, variables, and relationships); ii) to emphasise the most important relationships that may require a more thorough investigation before implementing battery end-of-life policies and recycling techniques; and, iii) to illustrate and communicate direct and indirect global warming impacts to policymakers, users, and other stakeholders.

5 References

[1] https://www.energy.gov/eere/electricvehicles/electric-vehicle-basics

[2] de Bortoli, A., & Christoforou, Z. (2020). Consequential LCA for territorial and multimodal transportation policies: method and application to the free-floating e-scooter disruption in Paris. Journal of Cleaner Production, 273, 122898.

[3] https://its.ucdavis.edu/wp-content/uploads/The-Importance-of-Life-Cycle-Assessment-in-Evaluating-the-Environmental-Performance-of-Future-El.pdf

[4] Hoekstra, A. (2019). The underestimated potential of battery electric vehicles to reduce emissions. Joule, 3(6), 1412-1414.

[5] Strømman, A. H. (2013). Comparative environmental life cycle assessment of conventional and electric vehicles.

[6] Archsmith, J., Kendall, A., & Rapson, D. (2015). From cradle to junkyard: assessing the life cycle greenhouse gas benefits of electric vehicles. Research in Transportation Economics, 52, 72-90.

[7] Yuksel, T., Tamayao, M. A. M., Hendrickson, C., Azevedo, I. M., & Michalek, J. J. (2016). effect of regional grid mix, driving patterns and climate on the comparative carbon footprint of gasoline and plug-in electric vehicles in the United States. Environmental Research Letters, 11(4), 044007.

[8] https://greet.es.anl.gov/

[9] IEA (2020), Global EV Outlook 2020, IEA, Paris https://www.iea.org/reports/global-ev-outlook-2020

[10] https://www.rff.org/publications/explainers/carbon-pricing-201-pricing-carbon-electricity-sector/

[11] Ambrose, H., Kendall, A., Lozano, M., Wachche, S., & Fulton, L. (2020). Trends in life cycle greenhouse gas emissions of future light-duty electric vehicles. Transportation Research Part D: Transport and Environment, 81, 102287.

[12] https://ec.europa.eu/clima/policies/transport/vehicles_en#tab-0-1

[13] D. Gohlke and Y. Zhou. (2020) ​“Impacts of Electrification of Light-Duty Vehicles in the United States, 2010 – 2019,” Argonne National Laboratory, Lemont, IL USA, ANL/ESD-20/4.

[14] EC (2014), Frequently Asked Questions on Directive 2006/66/EU on Batteries and Accumulators and Waste Batteries, http://ec.europa.eu/environment/waste/batteries/pdf/faq.pdf

[15] https://ww2.arb.ca.gov/resources/documents/lcfs-pathway-certified-carbon-intensities

[16] https://www.linkedin.com/pulse/astongate-fake-emission-figures-embattled-carmaker-sock-liebreich