3.1. Impact Differences Due to Variations in Emission Profiles
For better clarity, we removed models 3a and 13 from the following impact overview (
Figure 3). Model #3a is just an expansion of model #3 based on CNG consumption (
Table 1). Model #13 considers wind power instead of PV electricity for battery production, which results in slight impact variations only in comparison to model #12 (
see Table S11 in Supplement #2 in Supplementary Materials). Both cases are separately discussed below.
First, the results reveal that in a minority of five of 18 impact categories the electric VW Caddy is having clear advantages over the combustion engine models (climate change, photochemical oxidant formation, fossil resource depletion, natural land transformation, and ozone depletion), while in 10 of 18 impact categories, a vice versa picture appears. The disadvantages for electric vehicles are not reflected in the single score endpoints, which almost mirror the climate change impacts (
Figure 3). This goes back to the fact that the ReCiPe endpoint evaluation scheme is very much dominated by the climate change effect [
86].
In the impact categories of terrestrial acidification, particulate matter formation, and marine eutrophication, the lowest impacts of both technologies are comparable (
Figure 3). This resembles the picture found during life cycle modelling the electric and the combustion engine SMART with the exception that the LCIs in the category of marine eutrophication (ME) of both technologies are balanced at the VW Caddy, while the electric SMART exhibited larger impacts in this category [
17].
For the first time, the results depicted in
Figure 3 enable a comparison of the different use phase emission inventories. Comparing the impacts of vehicle models #1 and #2 (petrol “Euro 5 original” and “Euro 5 scaled” inventories), there are only small differences (
Figure 3): midpoint impacts of climate change (CC), terrestrial acidification (TA), terrestrial ecotoxicity (TET), particulate matter formation (PMF), photochemical oxidant formation (POF), ME and the single score of model 2 (“Euro 5 scaled”) are slightly higher (
Figure 3). This is also the case when comparing midpoint impacts of the models #4 and #5 (diesel “Euro 5 original” and “Euro 5 scaled” inventories, see
Table 1). Model #5 (Euro 5 scaled) shows slightly larger impacts in the midpoint categories of terrestrial acidification (TA), terrestrial ecotoxicity (TET), particulate matter formation (PMF), photochemical oxidant formation (POF), and marine eutrophication (ME). Concluding, adapting LCA inventory modelling to individual mileages did not result in significantly different impacts, as shown when comparing the respective modelling results (“Euro 5 original” vs. “Euro 5 scaled”).
Due to Ei3 emission species amendments [
50] to the petrol VW Caddy a few midpoint categories scored slightly higher impacts of model #3 (particulate matter formation, PMF; marine ecotoxicity, MET; marine eutrophication, ME), compared to vehicle models 1 and 2 (
Figure 3). Adding the new group of emissions (Ei3) to the petrol car model, however, has a significant impact on the terrestrial ecotoxicity (TET,
Figure 3): Vehicle model 3 scores 57% higher in terrestrial ecotoxicity (TET)-LCI (Life cycle impact), when compared with petrol models 1 and 2 (averaged). The process module composition of the impact category terrestrial ecotoxicity (TET) of vehicle #3 exhibits that 53% of the LCI of terrestrial ecotoxicity (TET) can be traced back to the impact of the updated Ei modul “operation, passenger car, petrol EURO 5,” while another 33% of the LCI of terrestrial ecotoxicity (TET) is due to impacts caused by the Ei module “petrol, low sulphur, at regional storage [CH].” Consequently, the newly added emission species (Ei3) may have resulted this additional impact. This is also the case for the LCI in human toxicity (
Figure 3): The LCI in the impact category human toxicity (HT) of vehicle #3 increased by 17%, compared to the average of vehicles models #1 and #2 (both petrol).
Correcting the emissions of four species (
Table 5) for diesel vehicle models (#6,
Figure 3) led to more complex changes than in the impacts of the petrol vehicles. First, and analogous to petrol model #3, diesel model #6 exhibits an increase of impact in terrestrial ecotoxicity (TET) and human toxicity (HT,
Figure 3), compared to the averaged LCIs from diesel vehicle models #4 and #5 (
Figure 3). Despite this terrestrial ecotoxicity (TET) impact variations it can be concluded that in case of an unchanged adoption of the Ei3 emission numbers there would be no significant impact alterations, neither between the two attempts of mileage implementation (original vs. scaled), nor between the database advancement from Ei2.2 to Ei3.
Second, updating the emissions of four species in the inventory of diesel model #6 leads to noticeable deviations in the four midpoint categories of marine eutrophication (ME), terrestrial acidification (TA), particulate matter formation (PMF), and photochemical oxidant formation (POF) (
Figure 3, see model #6 each), by 143%, 67%, 65%, and 70%, respectively, in comparison to the averaged LCIs of vehicle models #4 and #5.
This is based on the assumption that without adapting the respective emission numbers relative to real-world emissions (
Table 5), the impact increases of diesel vehicle #6, compared to vehicle #4 and #5, would have been small or insignificant as it was observed comparing the petrol vehicle #3 with petrol vehicles #1 and #2 (see above). The corresponding original Ei3 emissions are all lower than those of Ei2 (
Table 5).
The elevated impacts observed here are due to excess use phase emissions, clearly caused by the corrections applied (
Table 5): As it can be calculated from the process modules composition of all four impact categories of marine eutrophication (ME), terrestrial acidification (TA), particulate matter formation (PMF), and photochemical oxidant formation (POF), respectively, the LCI in each is dominated by “operation, passenger car, diesel EURO 5” (to 78%, 53%, 59%, and 67%, respectively). In all these four impact categories, the second largest factor in the respective process modules compositions is “diesel, low sulphur, at regional storage [CH]” (to 12%, 28%, 21%, and 20%, respectively). This results in predominant use phase impacts over lifetime in the respective four impact categories (
Figure 3).
To the best of our knowledge, such increased impacts as a result of an updated diesel vehicle emissions inventory was first presented by Bauer (2017) [
87] but has not been documented further on in detail. Bauer (2017) [
87] pointed to a 35% increase of lifecycle particulate matter formation (PMF) impact as well as to a 64% increase of lifecycle photochemical oxidant formation (POF) impact, when updating the diesel car emission inventory to be closer to real-world results, but did not go into further detail.
We found bigger LCI increases than Bauer (2017) [
87] when applying real-world emissions, and detected increased impacts in two more impact categories. However, Bauer (2017) [
87] used NO
x emissions elevated by a factor of 6, while our updated NO
x emissions increased by a factor of 9.2, relative to Ei3 (
Table 5).
These excess LC impacts of diesel vehicle model #6 are hardly visible in the single score result that is just 4% higher than the single score indicator of models #4 and #5 (averaged). When averaging the elevated LCIs of diesel vehicle model #6 identified in the impact categories of marine eutrophication (ME), terrestrial acidification (TA), particulate matter formation (PMF), terrestrial ecotoxicity (TET), and photochemical oxidant formation (POF) for the 18 impact categories, it should result in an elevated single score LC impact of +18% of model #6. The dominance of the CC impact in the ReCiPe evaluation scheme [
86] is blurring this effect (
Figure 3).
In conclusion, switching from Ei2.2 to Ei3 emissions does not significantly change the environmental impacts of petrol and diesel cars, despite of the slightly increased impacts in terrestrial ecotoxicity and human toxicity, respectively. The high number of added chemical emission species in Ei3 does not cause significant increases in LCIs despite of the impact categories of terrestrial ecotoxicity (TET) and human toxicity (HT) (
Figure 3). However, the real-world close update of a few species, emitted by diesel cars, changes the picture.
Interestingly the switch from Ei2.2 to Ei3 also has very small and negligible effects on the use phase impact in any impact category of the electric BEV (comparing models #7 and #9,
Figure 3), although the number of emissions increased from 22 to 98 (
Table 4).
3.3. Effects of Electricity Supply Choices on Battery Production Impacts
Results depicted in
Figure 4 highlight the advantages of battery cells made under provision of renewable electricity. Particularly, human health and ecosystem related impacts decrease when substituting coal-based electricity (
Figure 4). Human health–related endpoints for the Chinese electricity provision scenario are 3.4 times higher than those of the scenario assuming wind electricity only (
Figure 4). The analogue difference in ecosystem related endpoints is a factor of 3.5 (
Figure 4). This is reflected in individual midpoint deviations: Coal dominated electricity has particularly high impacts in photochemical oxidant formation (POF), particulate matter formation (PMF), agricultural and land occupation (ALO), terrestrial acidification (TA), fossil resource depletion (FD), and marine eutrophication (ME). UCTE electricity, on the other hand, exhibits larger impacts than the other scenarios regarding human toxicity (HT), ozone depletion (OD), water depletion (WD), and ionizing radiation (IR, the latter due to the nuclear power plants still available in Europe). Individual midpoints are reported as per 1 kWh electricity in
Table S11 in Supplement #2 in Supplementary Materials. When switching from Chinese electricity toward 100% PV, the impacts in 14 categories decrease, on average by 43% per impact category (
Table S11). On the other hand, there are slight increases in the impacts of the midpoint categories of ionizing radiation, ozone depletion, terrestrial ecotoxicity, and mineral resource depletion, respectively (
Table S11). When using 100% wind electricity, on the other hand, even these impacts are lower than those under provision of Chinese electricity, despite of the impact in mineral resource depletion (
Table S11).
Battery cells produced with coal-dominated electricity have a 3.8 times higher carbon footprint (156 kg CO
2-eq/kWh) than cells made with 100% wind electricity (41 kg CO
2-eq/kWh) (
Figure 4,
Table S11). Compared to 100% PV electricity, 100% wind electricity’s carbon footprint is still an additional 16% lower. These findings coincide in principle with the 61–106 kg CO
2-eq/kWh reported in a recent review [
91], decreasing from the 150–200 kg CO
2-eq/kWh reported earlier [
92]. As other reports, also Romare and Dahllöf [
92] conclude that the magnitude of battery carbon footprint is “nearly independent of the cell chemistry.”
Based on UCTE (2004) electricity, the CC impact due to direct electricity consumption during battery production accounts for 57% of the CO
2 equivalent emissions relative to the entire CC impact of Li-Ion cell production, as derived from the process modules composition (the remaining 43% are caused by the provision of the cell components). Ellingsen et al. (2013) [
64] reported an average of 78% CO
2-eq emissions, due to direct electricity use during cell production under a comparable electricity mix. Under provision of Chinese electricity as shown in our modelling, direct electricity consumption accounts for 75% of CO
2-eq emissions during cell production, under 100% PV electricity this goes down to 18%, respectively. Under provision of wind electricity, the contribution of direct electricity consumption to the climate impacts of battery cell production is 0.4%, which is negligible. The carbon footprint of battery production under wind electricity consumption is thus almost completely dominated by the provision of the chemical/mineral battery components.
These findings suggest an enormous potential to mitigate greenhouse gas emissions by producing Li-ion battery cells with renewable electricity (
Table 3). A battery cell production in a European country with a high proportion of coal-derived energy like Poland (88% coal, 92% fossil in total, according to Frischknecht et al. [
60]) would even worsen the production backpack of the battery, while a European production with renewable electricity can deliver the savings quantified here.
Apart from the climate change impacts, the following impact categories particularly benefit from switching to renewable electricity during battery production—terrestrial acidification, particulate matter formation, photochemical oxidant formation, agricultural land occupation, urban land occupation, and marine eutrophication (
Figure 3). The high impacts from electric powertrain production (
Figure 3) are caused during the production of printed circuits. Details on this analysis can be found in Helmers et al. (2017) [
17].
3.4. The Natural Gas Alternative and Effects of Electricity Supply Choices During Battery Production on the Lifetime Impacts
CNG vehicles tend to have lower pollutant emissions (e.g., Khan et al. [
93]) than petrol vehicles. They are based on almost the same engine technology. Our approach (modelling the same chemical emissions of the petrol version) thus results in a slight impact overestimation of vehicle 3a, propelled with CNG, in impact categories like photochemical oxidant and particulate matter formation (compare
Figure 3), which affects single score impact (
Figure 5). The climate change impact of natural gas combustion was as at first glance quantified with an independent LCI model based on the Ei module “natural gas, burned in gas motor, for storage [DE].” This module revealed 214 g CO
2/kWh CNG, which was converted to 2.92 kg CO
2/kg CNG. We regard this as being unrealistically low because this would include only a 6% additional impact along the fuel supply chain. Well-to-tank efficiency, however, is 80.25% on average in CNG provision (reviewed in [
39] Helmers & Marx 2012). We accordingly add 19.75% due to fuel chain expenses on top of the 5.99 kg CNG/100 km as measured for the VW Caddy, arriving at 198 g CO
2/km (well-to-wheel) for modelling the use phase.
CNG vehicle #3a (
Table 1) illustrates the impacts of combustion engine vehicles propelled with natural gas, which in the use phase produce much lower CO
2 emissions than diesel and petrol cars (for European wide data, see Helmers et al. [
8]). The electric VW Caddy cannot compete with the NG version (
Figure 5), as long as electricity is provided by a mix close to the European average (
Figure 5). This only changes when the BEV is charged with renewable electricity (
Figure 5). The electric vehicle, however, comes with an additional production impact both due to the battery and the powertrain. Its powertrain production impact with 3.4 t CO
2-eq is 2.2 times higher compared to that of the conventional combustion engine car, according to the models run here (
Figure 5).
Lifetime CC impacts of vehicles #11-13 represent a BEV similar to model #9, with the only difference of a reduced battery production impact, when switching from a coal-dominated electricity mix during battery production (vehicle #9) to average European electricity mix (#11), 100% photovoltaic electricity (#12), and 100% wind electricity (#13) during battery production (and at the same time keeping DE 2013 electricity supply in the use phase,
Figure 5). The CC impact from battery production decreases this way from 4.0 t CO
2-eq (vehicle #9), to 2.4 t CO
2-eq (vehicle #11), 1.3 t CO
2-eq (vehicle #12), and 1.1 t CO
2-eq for vehicle #13, respectively, roughly a 50% reduction for every step excluding the last step from PV to wind electricity supply (
Figure 5, all battery sizes: 25.9 kWh). This illustrates the relevant impacts during battery production, on the one hand, and the optimization potentials, on the other hand, affecting the whole life cycle of the vehicles. With a 25.9 kWh battery made under provision of PV or wind electricity, the electric Caddy is already advantageous compared to the ICEV alternatives, even when charging DE 2013 electricity mix.
The analysis of single score endpoints reveals a pattern similar to the CC impact, which has been also observed in the earlier modelling of an electrified SMART [
16] (Helmers et al. 2017). However, the advantage of natural gas fueling appears even more pronounced in the single score endpoints than in the CC analysis (
Figure 5, model #3a). Also, the relative share of impact caused by electric powertrain production, is higher in the single score endpoints, compared with its CC impact (
Figure 5).
Battery sizes have grown continuously over the past few years—while the average BEV battery pack in the year 2015 was rated as 30 kWh, there was already a tendency toward an average of 50 kWh in the year 2017 [
94]. Accordingly we add modelling version 10a of a BEV, with a 51.8 kWh battery, twice the size of the battery assumed for the other BEV models. In case this BEV is charged from a grid with renewable electricity (130.6 g CO
2-eq/kWh, DE 2013), but the battery is still produced with coal-based electricity (1180 g CO
2-eq/kWh, e.g., China), 36% of the lifetime CC impact will be caused by the battery production only (vehicle 10a,
Figure 5). The single endpoints reveal that battery production is responsible for 31% of the lifecycle impact of vehicle 10a (
Figure 5). This finding highlights again the necessity to produce vehicle batteries with renewable electricity.
When averaging the life-cycle CC impact of the two Caddy vehicles operating an Otto engine (#3, 3a, propelled with petrol and natural gas, respectively), there is only an insignificant difference of 0.5% between this average and the CC impact of the diesel engine vehicle (vehicle model 6). Accordingly an averaged climate change impact of all the three combustion engine vehicles (petrol, diesel, natural gas) will be used in the following section to quantify the competition with the electric vehicle in better detail.
3.5. Climate Change Impact: Combining Different Mileages, Battery Sizes, and the Battery Second Use Case
In
Figure 6, seven variables of the LCA modelling process are evaluated in parallel for their influence on CC impacts, with emphasis on the choices of electricity provision during battery production. Two different battery sizes are considered (25.9 kWh, as originally installed, plus an extrapolated 52.8 kWh of battery capacity). Both battery sizes are as well modelled with and without a subsequent second-life stationary use (reducing these batteries CC impact relative to the vehicle by 50%, see above).
The choice of battery size and post-vehicle battery management (with/without second use) influences the BEV’s overall CC lifecycle impact by a factor of 1.2–2.2. When charged with renewable electricity, the lifecycle CC impact of the VW Caddy ranges from 75–166 g CO
2-eq/km (
Figure 6, bottom left and right), this increasing to 194–280 g CO
2-eq/km when charging DE 2013 electricity (
Figure 6, top left and right). Adding the battery second use case to the vehicle with 51.8 kWh battery reduces its lifecycle impact by 5–27 g CO
2-eq/km (
Figure 6). At the utmost, application of a battery second use can remove up to 15% of the lifecycle CC impact from a VW Caddy.
If battery production and vehicle use rely on electricity with a considerable carbon footprint, e.g., electricity resembling the European mix (DE 2013), then only electric vehicles with a small battery (25.9 kWh or below) can match the life-cycle carbon footprint of conventional cars (
Figure 6). Under the provision of the German electricity mix, the three vehicle combinations with a 51.8 kWh battery emit along their lifetime 281–308 g CO
2-eq/km and thus showing higher climate impacts than the conventional VW Caddy (use phase 150,000 km
Figure 6, top left). The BEV also benefits, much more than the ICEV, from a prolonged use phase of 200,000 km (
Figure 6).
3.6. The Size Effect: Comparing the Break-Even Mileages of the Electrified SMART and the VW Caddy
The electric conversion projects presented here allowed us to compare the impacts of two electric vehicles which differ distinctly in size—a mini car (by the company SMART, see Helmers et al. [
17]) and a midsize car, the VW Caddy, as modelled and presented here.
We find that, except from one case, the electric SMART and the VW Caddy drive to be advantageous in a foreseeable (realistic) life cycle, when it comes to the comparison with the combustion engine vehicles (
Figure 7). The exception is the Caddy with 51.8 kWh battery made in China with no battery second use, which needs an unrealistic mileage of 310,063 km to reach the LCI of the ICE Caddy. The break-even mileage, however, is reduced to 207,000 km or 137,000 km, when the battery is produced with electricity of the average European carbon intensity (531 g CO
2-eq/kWh) or with renewable electricity, e.g., generated from PV (92.5 g CO
2-eq/kWh), respectively (
Figure 7, use phase DE 2013). The corresponding VW Caddy with smaller battery (25.9 kWh, a size usually modelled in early investigations) provides break-even mileages of less than 75,000 km under all conditions (
Figure 7). Early LCA data published, on the contrary, did not report any advantages in CC impact of BEV compared to ICEV when charging average German electricity [
95]. While it is known that smaller batteries move EVs faster toward advantages over ICEs, the degree of variation in reaching break-even mileages depends on the use phase electricity supply. Switching from “medium carbon impact” electricity (DE 2013) to a renewable mix (DE 2050,
Table 3) reduces the number of km needed for break-even with the ICEV by a factor of 3.1–5.2 in case of the VW Caddy or by 3.8 in case of the SMART (
Figure 7).
Assuming a secondary use for battery reduces the number of km necessary to reach break-even point further by factor of 1.2–1.7 for the VW Caddy or 1.2–1.4 for the SMART, respectively (
Figure 7). The switch from batteries made in China to batteries made under provision of wind electricity reduces the number of km needed for break even by a factor of 1.6–2.5 (Caddy) or 1.4–1.8 (SMART), respectively (
Figure 7).
Under the provision of DE 2013 electricity, both cars’ overall predicted CC lifecycle impacts behave much more sensitive to a switch of battery production from higher to lower carbon footprint (
Figure 7). The reason is that the lifecycle CC footprint of BEV provided with DE 2013 charging electricity is much closer to that of the ICEV, the difference can even be very small (
Figure 5 and
Figure 6). Accordingly, under DE 2013 electricity provision the BEV needs to drive many more km to undercut the ICEV’s LC CC impact. Under provision of green electricity (DE 2050), the BEV reaches break-even mileages much faster and is thus not so sensitive to increased battery impacts: The electrified VW Caddy then passes the ICE Caddy after 13,900–59,000 km, depending on battery size and application of battery second use (
Figure 7). Accordingly, the electrified SMART needs 17,000–35,000 km to reach this target under DE 2050 electricity provision (
Figure 7).
In contrast to findings from Ellingsen et al. [
22], who reported that mileages necessary to reach break-even increased with vehicle size, this has not been confirmed here. Although the SMART with 14 kWh of battery capacity and the Caddy with 25.9 kWh of battery size provided a comparable driving range (104 vs 128 km), or in other words, possess a comparable battery in relation to vehicle size, both vehicles reveal pretty much comparable mileages necessary for the electric models to reach lower life cycle carbon impacts (
Figure 7). The smaller electric SMART even requires to drive 1.3–1.4 times more km than the electric Caddy (25.9 kWh) to reach the “green zone” based on supply of DE 2013 electricity (
Table 3). Charged with renewable electricity (DE 2050,
Table 3), both cars reveal relatively similar break-even mileages (
Figure 7). The bigger VW Caddy enables the lowest break-even mileage in this comparison: 13,900 km (25.9 kWh battery made with wind electricity, plus battery second use,
Figure 7).
While the size of the vehicle is apparently not significantly influencing the break-even mileages as shown here, it matters when it comes to quantifying the lifecycle climate impacts in terms of CO2-eq/PKT (see below).
3.7. Lifecycle Climate Change Impacts of Electric Cars in Comparison with Competing Transportation Modes
Finally, the impacts relative to passenger kilometers travelled need to be quantified—a step usually missing in LCA of electric cars. This is essential in evaluating whether the electric car can play its role in the future decarbonizing transport. There is a statistically documented occupancy rate for European cars (1.57 persons/car, Castellani et al. [
96]); however, this is an average covering all sizes/classes of cars. Only 6% of all European passenger cars registrations count among the segment of mini class cars (ICCT 2018) [
97], and the SMART, as a two-seater, is among the smallest. We assume that the occupancy rate of a two-seater will be lower than that of a five-seater, as is the vast majority among the cars on the streets. Reducing the occupancy relative to the number of seats would result in an occupancy of even < 1, so we suggest to calculate with 1 person statistically driving in a SMART.
When dividing the LCI of the VW Caddy by a factor of 1.57 and assuming one person/car for the SMART then both vehicles deliver fairly similar life cycle impacts in terms of g CO
2-eq/PKT (passenger km travelled,
Figure 8). Without accounting for occupancy, the electric VW Caddy would have with 85–248 g CO
2-eq/PKT 1.6-1.7times the impact of the SMART. It also turns out that while the electric Smart provides small advantages in g CO
2-eq/PKT over lifetime even when charged with DE 2013 electricity in comparison with the ICE SMART, this is not the case for the electrified VW Caddy. The latter needs a renewable electricity mix to deliver advantages in this comparison (
Figure 8).
The LC CC impact of the SMART was quantified as 72 g CO
2-eq/km per vehicle in the previous project [
17], if recalculated with a 150,000 km lifetime mileage and based on Chinese battery production. Additionally assuming battery production under supply of 100% wind electricity plus a battery second use scenario reduces the lifetime CC impact by 25% on 54 g CO
2-eq/km per vehicle for the SMART (
Figure 8, “el. SMART lowest”). This illustrates again the potentials of reducing the lifetime impact of an electric vehicle by adjusting production and post-use treatment of the battery.
To put the impacts into a wider transportation context, we estimate climate impacts for electric buses from literature sources and added data for alternative passenger traffic modes such as diesel buses and coaches, and trains (
Figure 8).
It turns out that both the electrified SMART and the VW Caddy can compete well with the carbon intensity of passenger km travelled with bus, coach and train, assuming the cars are charged with renewable electricity (
Figure 8). Electric buses seem to deliver the lowest CC impacts during vehicle use in terms of passenger-kilometers travelled (27–52 g CO
2-eq/PKT,
Figure 8). These results are not expected to change significantly when incorporating infrastructure construction and operation costs as these are quite similar when comparing railroads and streets [
101,
103].
This modelling reveals high savings due to electrification, when it comes to the CC impact in comparison to the cars with combustion engine. Under optimized conditions (battery produced with wind electricity, BEV charged with renewable electricity, battery second use), the BEV delivers 64% (SMART) or 65% (VW Caddy) savings in CC LCI, respectively (
Figure 8; the VW Caddy modelled with 1.57 persons travelling on average) in terms of g CO
2-eq/PKT. This would be in accordance to the EU target to reduce the transport related CO
2 emissions by 60% from the 1990 levels to 2050 (EU 2019b) [
104]. In other words, the EU CO
2 reduction target might not be achieved by electric vehicles if not reducing the life cycle impact of the batteries in the way described here.