Just as automakers have coalesced on moving towards electric vehicles, it’s become clear the affordable EV is dead. Back in 2017, media anticipated that affordable mass-market EVs were just around the corner. The compact sedan Tesla Model 3 was announced at $35,000 as a Ford Focus competitor, while the Chevy Bolt was already available at $37,000 as a Honda Fit competitor, and after tax credits both sets of vehicles were within striking distance of cost.
Today the Ford Focus and Honda Fit are long discontinued in the US market, as are any plans for similarly economical new EVs. The Model 3 now starts at $48k, while the Bolt has been burning quietly behind ads for a new Hummer. Automakers almost universally decided to go for the higher margins and lower volumes achievable with SUVs and culled their economy cars, an approach that’s been vindicated by the chip shortage while called into question by rising gas prices. Every EV hitting the US market from the major automotive companies is a $45k and up crossover SUV or truck. This includes the Ford Mustang Mach-E, Toyota BZ4X, VW ID.4, Honda Prologue, Nissan Ariya, Cadillac Lyriq, Hyundai Ioniq 5, Volvo XC40, and Kia EV6. Each weigh in the mid 4,000lbs range, and are 30-40% less energy efficient than smaller EVs.
At the extreme end of this trend is GM’s new $110k Hummer, weighing in at 9,063lb and using a whopping 246kWh of batteries. In global contrast, GM’s top selling global model is the $4,200 Chinese-market Wuling Mini EV, using as little as 9.6kWh and weighing under 1,500lbs. With “mass-market” EVs in North America strictly following this broader trend towards bigger, heavier, and more expensive vehicles, it’s worth questioning whether SUVs are the best use of lithium ion batteries, a limited and dirty resource to extract.
Literature Review
Used for the past 30 years in nearly all consumer electronics, global lithium ion battery production dramatically increased from 95.3GWh in 2020 to 410.5GWh in 2022, and will only go up from there. Each major automaker is currently planning and constructing facilities to individually produce more batteries than the total global consumption was in 2020. That includes VW planning 240GWh, Tesla 250GWh, GM 140GWh, and Ford 140GWh, in addition to production planned by other global automakers.
In 2020, a brutal right wing dictatorship seized power in Bolivia and maintained it for a full year before allowing new elections. Responding to theories that a goal of privatizing Bolivia’s lithium supplies encouraged US support for the fascist coup, Tesla CEO Elon Musk joked(?) “we will coup whoever we want” on Twitter. Whether a genuine connection or not, this illustrates the politically contentious and dirty nature of precious metal extraction involved in battery production.
Current lithium carbonate mining is insufficient to meet the demand for lithium ion batteries, which is rapidly rising. Industry consumption is estimated at 641,000 megatons this year, compared to 342,000 in 2020. While in 2020 and years prior mining resulted in surplus of 66,000mt, last year more lithium was consumed than could be mined, dramatically increasing the price. From March 2021 to March 2022, the price of lithium carbonate increased by 465%.
This tenfold increase in need for batteries does not come without severe environmental and social consequences. Thacker Pass in northern Nevada is the site home to the US’ largest lithium deposit, and its development is imminent. Lithium extraction at Thacker Pass will require 12,204 liters of water per minute in an arid environment, which will require purchasing and diverting water from neighboring communities. The US BLM environmental impact study of the project anticipates a likelihood that lithium production will leak arsenic into this already limited group water table. The study also anticipates lithium production will threaten thousands of acres of wildlife habitat. The mine will be depleted by 2065.
Neighboring Indigenous Nations have organized against the project concerned about the prevalence of of MMIW (Missing and Murdered Indigenous Women, Girls, and Two-Spirit People) in Indigenous communities near industrial boomtowns. In a press release, The Atsa koodakuh wyh Nuwu (The People of Red Mountain) wrote:
Fighting climate change, however, cannot be used as yet another excuse to destroy native land. We cannot protect the environment by destroying it. In addition, there is the likelihood that “man camps” will form to supply labor force for the mine, and that the mine will strain community infrastructure, such as law enforcement and human services. This will lead to an increase in hard drugs, violence, rape, sexual assault, and human trafficking. The connection between man camps and missing and murdered indigenous women is well-established.
A US District court ruled in favor of a Canadian lithium mining company in continuing with Thacker Pass development despite the Reno-Sparks Indian Colony and Burns Paiute Tribe claiming ancestral sovereignty amid evidence of an 1865 US massacre of 31 Paiute men, women, and children on the site. The Judge admitted the evidence, “further highlights the shameful history of the treatment of Native Americans by federal and state governments,” but contended, “it does not persuade the Court that it should reconsider.” Will Falk, the lawyer representing the Indigenous nations, reflected on the decision, “The memory of the massacre lives on in the land. Destroying the land where the massacre happened makes it more difficult to imagine that massacre. Destroying massacre sites destroys the evidence of the massacres.”
Echoing oil pipeline boosters, executives of the mining companies often use rhetoric borrowing from both the nationalistic demand for “energy independence” and tugging at human rights and environmental heart strings by comparing their projects to less regulated mines in the Global South. “This is the most sustainable lithium in the world, made in America,” said Rod Colwell of Controlled Thermal Resources. Severe environmental and social impacts of mining certainly also exist in Australia, Chile, China, and Zimbabwe, the current top lithium producing countries, but should not justify harmful actions in the United States. These impacts are ultimately driven by demand from the Global North for large battery EVs.
Lithium is just the namesake of metals in Lithium Ion batteries sourced from often violent and environmentally destruction mines. Two other key ingredients include Cobalt and Nickel, which each have similar shortages and blood-stained mining.
While less flammable and resource intensive solid state batteries are on the horizon for the 2030s, they will not eliminate the need for lithium and nickel, just make better use of them while demand for batteries continues to skyrocket.
Methodology
To get a better understanding of an accurate comparison between new EVs and other gas, hybrid, and plug-in hybrid vehicles, we can start with Tesla’s own data. In 2020, Tesla released an impact report aiming to illustrate how the Tesla Model 3 (entry-level 54kWh, since replaced with a 62kWh model) has dramatically lower lifecycle emissions than combustion vehicles. Tesla’s methodology assumes a US average vehicle lifespan of 200,000 miles and 17 years, representing approximately 11,750 miles per year. Tesla’s impact report compares the compact Model 3 to a “premium mid-size” sedan averaging 25 MPG, mentioning in footnotes that a Mercedes C300 is one of their benchmark examples.
The most common car for Tesla buyers to switch from is the Toyota Prius, which achieves over double the efficiency of Tesla’s comparison. This 25 MPG choice of comparison is however conveniently the exact average new vehicle fuel economy, and not far off from the total US fleet average, making it easy to ask the question, “what if everyone went electric?”
Tesla calculates total lifecycle emissions at for the Model 3 at 177gCO2 per mile and the ICE car at 446gCO2, of which 45gCO2/mi on the gas car came from production compared to 52g/mi on the Tesla (p.87). Multiplying across the 200,000 miles of use, this means Tesla emitted 10.4t in production, while the gas car took 9t. Tesla’s report explains their methodology of arriving at that 9t for the gas vehicle by counting 5.5kg CO2 per kg of vehicle weight. Tesla does not specify how much of their vehicle’s embodied emissions are from the battery versus other parts, so I’m left to go to third party research.
In a literature review of estimates made between 2011 and 2017, lithium ion battery production emission estimates range from Dunn etc al’s 30-50kg/kWh to Ambrose & Kendal’s 194-494kg/kWh. It’s unlikely that the Model 3’s non-battery components are dramatically less carbon intensive to produce than gas vehicles. In Volvo’s report, even without the battery the XC40 EV is 4t more carbon intensive to build than the gas version, likely due to requiring larger suspension components to support the heavy battery. With Tesla claiming a total production carbon difference of just 15% compared to the gas car, it’s sensible to assume Tesla claims to achieve the lowest estimate of 30kg/kWh. This would put their 54kWh battery at 1.62t of CO2, and the non-battery component emissions at 5.37kgCO2/kg of vehicle weight. My analysis of other vehicles uses the 5.5kgCO2/kg figure for the standard Prius and Escape Hybrids, a halfway mark (5.43kgCO2/kg) between the two figures for the plug-in hybrid Prius Prime, and the 5.37kgCO2/kg figure for all other EVs. This means the battery in a standard Prius takes just 0.04t of CO2 to produce, while the Prius Prime takes 0.26t, and the VW ID.4 does 2.46t.
Volvo’s own impact report directly compares the gas and electric versions of their XC40 compact crossover. The gas XC40 happens to get the same 25mpg and weigh roughly the same as Tesla’s gas comparison vehicle, and is also representative of the most vehicles currently sold. While Tesla’s gas car rate of production emissions per kg of vehicle weight is 5.5, Volvo’s XC40 is 8.62. Volvo’s report is notably more transparent than Tesla’s, breaking down emissions by component, and even raw material, where Tesla forces us to guess. Volvo’s numbers point to dramatically higher production emissions for both its gasoline XC40 and electric variant than Tesla’s report. The EV’s production emissions are 25.4t with 7t of that from the 78kWh battery alone, nearly 2.5x higher than the Model 3, while the gas variant creates 15.7t of emissions, 75% higher than Tesla’s example gas car. Volvo also considers a separate end-of-life scrapping emission at 0.5t for its EVs and 0.6t for its combustion vehicles. 7t to produce the XC40 EV’s 78kWh battery would indicate 89.75kgCO2/kWh, three times the figure we used for Tesla’s estimate, but still within the lower end of third party research. Adjusting Volvo’s 200,000km (124,275mi) expected vehicle lifetime in Europe to the US market standard of 200,000 miles (322,000km), the gas XC40 will emit over 96t in its lifespan.
Tesla Production Emission Estimate Equations
Volvo Production Emission Estimate Equations
This is not to risk Elon Musk’s legal wrath by claiming Tesla’s report is a lie, it’s possible the production difference is from greater use of renewables at their facility than Volvo’s Belgian and Luqiao plants. In third party research, Qinyu Qiao et al estimated subcompact gas car production emissions in China at 9.17t and EV production at 14.7t in a 2017 study. The Burning Question authors Mike Berners-Lee and Duncan Clark estimated European production of a city car to do 6T, a Ford Fusion to hit 17T, and a Land Rover to top out at 35T. Both Volvo and Tesla’s estimates are within a range supported by other research.
Electricity Use Emissions
To calculate the carbon intensity of electricity used to power the car, instead of using the US grid as a whole, Tesla says they weighted it by the grid average that Teslas operate in, without directly revealing a figure. The Model 3 achieves 24kWh/100 miles, and Tesla counts 125g per mile from charging. After accounting for a roughly 13% efficiency loss that occurs from charging a battery, this means the grid mix Tesla is assuming emits 453g/CO2/kWh. This is slightly dirtier than the 2022 US average, which is about 395gCO2/kWh based off of the grid mix estimated by the US Energy Information Administration’s 2021 Annual Energy Outlook report. The Tesla’s European use emissions are estimated at 60g, indicating 217gCO2/kWh, while Mainland China, with a grid makeup comparable to North Dakota, carries an emissions rate of 311gCO2/mi, or 1,126gCO2/kWh.
Rather than multiplying Tesla’s static 453gCO2/kWh across the life of the vehicle, as Tesla does in their report, I recalculated use emissions annually using the US EIA AOE estimated fuel mix with figures each year through 2050. We can multiply the fuel mix by the CO2 emissions generated per kWH by energy source. The US EIA provides these figures for coal, natural gas, and petroleum, while we can go to the IPCC estimates for renewables and nuclear. This estimate indicates the US average will gradually make it down to 288gCO2/kWh in the 17 years leading up to 2039. While gas cars (including regular hybrids like the Prius) can never become more efficient, all EVs will be less carbon intensive over time, provided the grid moves away from fossil fuels as expected.
Both Tesla and Volvo illustrate very clearly that their EVs generate fewer emissions than 25MPG gas cars over the course of their lifetimes. But 25MPG cars are not the only comparison to be made.
My analysis estimates production and use emissions for the Toyota Prius and Ford Escape hybrids, a plug-in hybrid Prius Prime, and the all-electric Nissan Leaf, VW ID.4, Ford Mustang Mach-E, Hummer EV, Wuling Mini EV, and an e-bike. I produced two sets of numbers, one using Tesla’s reported production emissions by kWh and vehicle mass, and one using Volvo’s equivalents. The Mach-E and XC-40 EV are both highly representative of the auto industry’s rapidly growing lineup of crossover SUV EVs. While most drivers travel short distances each day, automakers are happily accommodating a desire for edge use-case range, adding a thousand pounds to modern EVs.
Meanwhile, using a tiny fraction of battery capacity of a full EV, the Prius has been a staple of efficient cars for 25 years, and happens to be the vehicle that new EV buyers most often flock from. The Prius Prime can run for 30 miles from the grid on a battery 1/8th the size of most EVs before it starts operating like a regular Prius. In a study prepared for the US DOE, drivers of the similar Chevy Volt did about 73% of their mileage on electricity. For the purposes of this analysis and to account for the Prius Prime having a slightly lower EV range than the Volt, we will round that down to 70%. Both Prius variants remain practical and reliable family hatchbacks available for under $30,000 and have similar equivalents made by Hyundai, but are ultimately a dying breed in North America. Nissan’s Note, a subcompact hatchback discontinued for the US market in 2019, lives on in Japan and other markets as a hybrid achieving 87MPG. The Toyota Yaris, Honda Fit, and Ford Focus similarly live on as ultra-efficient hybrids elsewhere. In the US, however, many speculate whether the Prius is the next non-SUV to die.
Findings
Given Lithium Ion batteries are a limited and destructive resource, it begs the question whether large battery EV crossovers result in greater reduction in emissions than far fewer total batteries being used to power greater numbers of more affordable hybrids, plug-in hybrids, and e-bikes.
Using Tesla’s production emissions estimates while adjusting charging emissions annually, the Model 3 will emit 27.8t of CO2 over the course of its 200,000 mile, 17 year lifespan. No other EV sold in the US currently achieves the Model 3’s fuel efficiency, at 141MPGe (0.24kWh/mile). The Nissan Leaf comes close at 111MPGe and 31.4 of lifecycle emissions, but the Prius Prime, if driven 70% on EV, does better than any of the larger EV SUVs, at 30.7. With the exception of the Hummer, the dirtiest EV in this comparison, the VW ID.4, still bests the standard Prius 38.8t, while the 9,000lb Hummer doubles that footprint at 81.5t. All numbers of electrified cars are put into perspective the sheer inefficiency of a 25 MPG gas car, hitting 89.3t, and the economy of an e-bike, burning just 1.8t even after accounting for three new battery packs (my own e-bike’s original battery lasted 4 years before needing replacement). Except the Hummer, all EVs make up for their higher manufacturing emissions and are cumulatively better than the 25MPG vehicle by just 2 years into operation, and the Leaf and Model 3 beat the Prius at just 4 years.
Cumulative lifecycle emissions over time by vehicle:
Lifecycle total emissions by vehicle:
Volvo’s methodology paints a slightly bleaker picture for all EVs due to their significantly less optimistic production emissions estimates. Unlike Tesla’s methodology, the HUMMER comes out as worse than the 25MPG vehicle with Volvo’s numbers. Besides that, the conclusions are largely the same: EVs are better than typical gas cars by wide margins, but dependent on the model, are within a close margin of efficient hybrids like the 56MPG Prius and 41MPG Ford Escape. Meanwhile, e-bikes have roughly 1/20th the impact of electric cars. The Prius Prime’s exceptional performance in this ranking is enabled by its 133MPGe electric mode efficiency, the closest match to Tesla’s 141MPGe, while vehicles like the XC40 EV gets just 85.
One consideration regarding the contextual sustainability of plug-in hybrids is their general incompatibility with apartment living. While modern EVs can fast charge at public DC stations in 15 minutes, a weekly chore only moderately less convenient than filling a tank of gas, the 30 mile or less battery of a PHEV really needs to be plugged in after each use to make any sense. Unless their building has a parking garage with available outlets, something that increases the cost of apartment construction, apartment dwellers are most likely left without the ability to plug-in. The PHEV technology will not make its way into carshare fleets for the same reason. Reducing car dependency and replacing car trips with walking, transit, and bicycling is a more systemically important approach to reducing transportation energy intensity than moderate gains in fuel economy.
In CO2 footprint, the HUMMER EV is undoubtedly an improvement on an equivalent full-size SUV like the Chevy Suburban, with 34.4t (Volvo method) to 51t (Tesla) fewer lifetime emissions. However, that improvement is made at the cost of a 246.8kWh of lithium ion battery, and weight that will wear on public roads. That enormous battery and the blood metals contained could instead power a 40 foot long city bus serving hundreds of daily passengers, four typical modern EV crossovers, 18 Wuling Mini EVs, 28 plug-in hybrids, 189 hybrids, or 368 e-bikes. Assuming a corresponding number of 25MPG vehicles can be taken off the road by their production, these alternate uses of the batteries would make a greater reduction in carbon emissions. The demand for larger and more profitable vehicles is a result of successful marketing, not customers needing the utility. Automakers currently sell every vehicle they can make, and chose to push the market exclusively towards SUVs that are more likely to kill a pedestrian on impact than the sedans of yore, causing a troubling rise in pedestrian deaths.
While the Hummer is General Motors’ flagship product domestically, their top-selling vehicle globally is an entirely different approach to electrification, the Wuling Mini EV. The Smart Car-sized vehicle launched in 2020 at under $5,000 USD and is the clear winning strategy for electric cars in the Global South. The 13.8kWh battery gives it a claimed range of 110 miles using the NEDC test, which is typically more forgiving than the US EPA test. The 75kWh 2016 Model S achieved 249 miles in the EPA test cycle and 298 miles in the NEDC. Assuming the same 0.835x conversion rate, the Wuling Mini would achieve 91.8 miles in its 13.8kWh variant, providing for a consumption rate of 15kWh/100 miles, half that of Nissan Leaf, but still 6.7x that of the e-bike. A typical entry-level 14Ah, 48V e-bike (0.672kWh) can travel 30 miles on a charge, accomplishing 2.24kWh/100 miles. This demonstrates both that vehicles can be made with less than half the energy consumption of US-desired models, and that even the smallest and most energy efficient car is a nearly magnitude less efficient of a way to get around than attaching a motor to a bicycle.
Electric vehicles are a necessary piece of decarbonization, but without caution their proliferation may only justify larger problems created by auto dependency.
If someone were to switch most of their commuting to an e-bike while driving 5,000 miles a year in a Prius (like you can rent with the AAA-operated GIG/EVO carshare service in West Coast cities), their lifecycle-adjusted share of annual emissions would be lower than a 10,000 mile-per-year Nissan Leaf driver. This is a reminder that reducing vehicle miles travelled is more important than EV adoption, and investments in public transportation and pedestrian safety infrastructure should be a priority to allow more people to avoid driving.
| Prius Hybrid driven 5,000mi/year | E-bike ridden 5,000mi/year | Prius + E-bike Combined Emissions | Nissan Leaf driven 10,000mi/year | |
| Use emissions per mile (g) | 179.0 | 8.1 | 93.6 | 108.4 |
| Annualized share of production emissions (kg) | 610.3 | 13.6 | 623.9 | 798.0 |
| Annual emissions (kg) | 1505.4 | 54.1 | 1559.5 | 1882.3 |
Urban Geometry and Efficiency Considerations
Overall per capita transportation energy use in transit-heavy and walkable cities such as Barcelona, Shanghai, Tokyo, and Helsinki are a small fraction that of car-dependent Atlanta or Houston due to commuter use of single occupancy vehicles. The halving of emissions achieved by electrification still leaves car dependent cities far from the efficiency of cities with mass transit and walkability.
Within US Metro areas, per capita carbon emissions are dramatically lower in transit-served cities than the surrounding suburban sprawl. In Downtown Seattle’s 98104 zip code, that figure is 21.6tCO2e per year, while 20 miles east in the car-oriented 98053 zip code of Redmond, the figure is 72tCO2e per year. This dramatic difference cannot be solved by electric single occupancy cars alone. The physically larger geometry requires additional concrete, utility infrastructure, and delivery route mileage for each resident.
Parking is also an unsustainable geometry. In metro Vancouver BC, each car is served by 2.2 parking spots, amounting to 3.3 million spaces, yet complaints of not enough parking are abound. Each parking spot requires 288-398 square feet of impermeable real estate, comparable to many smaller studio apartments, while driving the cost of the associated housing or commercial space up.
Conclusion
Electric vehicles undoubtedly have a smaller carbon footprint than strictly equivalent combustion vehicles, but their use of extractive mineral resources is a harder to quantify issue with both environmental and social impacts. Distributing batteries efficiently to meet global transportation needs should be a public interest priority – and can be best achieved with a mix of e-bikes, hybrids, smaller EVs, and public transportation. Instead, automakers are increasingly promoting oversized vehicles that make a downright gross misuse of materials. Mass adoption of electric SUVs does not make the best use of limited urban space or environmentally and socially costly mineral resources. Urban planners should not overturn best sustainability practices because of electrification, and regulators should reconsider the ways US CAFE standards have encouraged the proliferation of larger, less efficient, and more hazardous vehicles.