Int J Life Cycle Assess (2015) 20:1127–1142

DOI 10.1007/s11367-015-0903-x

MODERN INDIVIDUAL MOBILITY

A comparative LCA of an electric vehicle and an internal

combustion engine vehicle using the appropriate power mix:
the Italian case study
Pierpaolo Girardi 1 & Alessia Gargiulo 1 & Paola Cristina Brambilla 1

Received: 4 December 2014 / Accepted: 12 May 2015 / Published online: 23 May 2015
# Springer-Verlag Berlin Heidelberg 2015

Abstract more than the 60 % of this energy is produced in efficient

Purpose The purpose of the study is to compare the perfor- combined cycle gas turbine power plants, EV performs better
mances of two passenger cars: an electric vehicle (EV) and an than ICEV in almost all the impact categories considered ex-
internal combustion engine vehicle (ICEV) paying particular cept for human toxicity and eutrophication, the only two im-
attention to the production of electricity that will charge the pact categories in which the EV car, mainly due to battery
EV. Even if many similar comparative life cycle assessments manufactory, presents more relevant potential impacts. ICEV
(LCAs) exist (Nordelöf et al. J Life Cycle Assess impacts are always dominated by well to wheel phases (use
19(11):1866–18990, 2014), only few have focused their atten- phase and fuel production). EV car and battery manufacturing
tion on evaluating which is the kind of electricity that will have higher impacts for all categories than ICEV car
recharge EV batteries (Hawkins et al. Int J Life Cycle Assess manufacturing.
17(8):997–1014, 2012). Conclusions The study demonstrates that electricity supplied
Methods Despite its relevance, many EV LCA studies have in Italy to EV today is, and will probably be in 2030, mainly
used a ready-to-use dataset to evaluate the power mix that produced by fossil fuel power plants. Nevertheless, the EV
supplies electricity to EV. The present paper tries to better proves to be able to reduce, with respect to ICEV, those im-
define the power mix that recharges EV batteries in Italy ac- pacts it is supposed to reduce: air acidification, photochemical
cording to the national power system and the national electric oxidant formation, and also greenhouse gases. Trade-offs are,
market rules. A 2013 and a 2030 scenario have been devel- as foreseeable, eutrophication and human toxicity due to EV
oped in order to understand effects in short and middle term. car and battery manufacturing.
Life cycle inventory of electricity for EV has been estimated
modifying available datasets according to official Italian data Keywords EV . ICEV . Italian power system . LCA
on power plants’ efficiency and emission rates. Finally, also
for the ICEV use phase, existing dataset have been modified
for fuel consumption and regulated emission according to the 1 Introduction
National Inventory Report results (ISPRA 2014)
Results and discussion In both 2013 and 2030 scenarios, the The penetration of electric vehicles (EVs) into the passenger
power mix that in Italy supplies energy to EV is dominated by cars’ market is seen as a great opportunity to reduce green-
fossil fuel power plants. Nevertheless, due to the fact that house gas emissions (Nemry et al. 2010) and population ex-
posure to dangerous exhaust gas emissions (Hawkins et al.
Responsible editor: Hans-Joerg Althaus 2013). The main advantages of electric vehicles rely on higher
powertrain efficiency compared to internal combustion engine
* Pierpaolo Girardi vehicles (ICEV), no matter if gasoline or diesel, and zero
[email protected]; [email protected] pollutant emissions during use phase. In order to understand
if those benefits exceed the potential environmental impacts
1
Sustainable Development and Energy Sources Department, RSE due to electricity supply, battery and vehicle production, and
Ricerca sul Sistema Energetico, Via Rubattino 54, 20134 Milan, Italy disposal, comparative LCAs should be carried out (Nordelöf
1128 Int J Life Cycle Assess (2015) 20:1127–1142

et al. 2014). In the last decade, many studies have been carried functional unit, and system boundaries of the study. Then, a
out on this topic. Hawkins has reviewed more than 50 papers description of the impact categories adopted for the LCIA
(Hawkins et al. 2012), and more recently, Nordelöf has phase and of data quality is given.
reviewed 79 papers (Nordelöf et al. 2014). The great majority
of these studies shared the conclusion that electricity produc- 2.1 Purpose and goals
tion is the main cause of environmental impact for externally
chargeable vehicles. If, and only if, the charging electricity has The goal of this LCA is to highlight by comparison the ad-
very low emissions of fossil carbon, electric vehicles can reach vantages and disadvantages linked to the diffusion of electric
their full potential in mitigating global warming (Nordelöf vehicles for private passenger transport in Italy. To this aim,
et al. 2014). Despite this conclusion, few studies focused we chose to compare the performances of a single EV with the
enough attention on defining the mix of energy sources and ones of a gasoline-fueled ICEV. Particular attention has been
generation technologies used to supply electricity to EVs or paid to the production of electricity that will charge the EV. In
used marginal electricity (Elgowainy et al. 2009; Amarakoon fact, if many studies exist on LCA of EV compared to ICEV,
et al. 2013; Hawkins et al. 2013) instead of average electricity only few of them have focused their attention on evaluating
mix. Hawkins, who found out that none of the 51 studies which is the kind of electricity that will recharge EV batteries,
analyzed has performed a complete LCI of EV, stated that while most of them use an average country electricity produc-
the most important steps toward understanding the impacts tion mix. To this end, we chose to apply our LCA to two
of EVs relative to ICEVs are improving estimates of the im- different scenarios: a 2013 scenario with few EVs and a
pacts associated with the electricity used to charge a growing 2030 scenario in which the demand of electricity due to EV
EV fleet (Hawkins et al. 2012). Moreover, Nordelöf has no- diffusion will be able to influence the structure of the electric
ticed that most assessments focus on the today’s electricity system. The purpose is to give useful information to policy
production system, although both vehicle technology and makers and stakeholders within the area of vehicle electrifica-
electricity production are supposed to change when electric tion (possible buyers too) but also to give information to re-
vehicle market share will become comparable to those of searchers who want to further investigate on this topic, on how
ICEVs (Nordelöf et al. 2014). Moreover, when and if this will to choose and analyze the energy supplied to EV.
happen, the national power system could be influenced by the
EV fleet additional electricity demand. Of course, giving that 2.2 Product system
the environmental performance of electricity supply do de-
pend on national electricity production system. In this frame- The scope of this LCA is to compare the environmental im-
work, the purpose of this paper is to compare the life cycle pacts of an EV with the ones of a gasoline-fueled ICEV. Both
environmental performances of an EVand an ICEV passenger vehicles are generic vehicles, supposed to be representative of
car in the Italian situation in the nowadays scenario and in a the lower medium market segment for passenger car, compa-
2030 scenario. We focused our attention on trying to under- rable to a VW Golf (VW a4 platform). The EVand ICEV to be
stand the real electricity used by EVs and on its LCI. compared are characterized by different propulsion systems
Life cycle assessment (LCA) is a method that can be used but are similar for glider, comfort, aesthetics, equipment,
in order to evaluate the potential environmental impacts (i.e., wheels, tires, and for all the aspects not strictly correlated to
environmental pressures) of a product, process, or activity. An the propulsion system. In this way, we ensure to compare only
LCA allows to assess impacts across the full life cycle of a those aspects linked to the propulsion system (like suggested
product system, from Bcradle to grave^ that is from materials by Amarakoon et al. (2013)). The EV is equipped with a
acquisition to manufacturing, use, and final disposal. As lithium-ion battery. Among different possibilities (as those
outlined in the International Standards Organization 14040 analyzed for example by Amarakoon et al. (2013)), we chose
series (ISO 2006), an LCA study has four major phases or a LiMn2O4 battery because it is the most used battery for
components: goal and scope definition, life cycle inventory electric propulsion systems and it is used for example in the
(LCI), life cycle impact assessment (LCIA), and interpreta- Nissan Leaf EV.1 The ICEV is an EURO5 gasoline car. Of
tion. This paper is organized in four chapters according to course, the service offered by an EV and an ICEV are not
ISO 14040. exactly the same. ICEVs have much longer range than EVs,
lower fueling time, and can rely on well-spread fuel station
grid. EVs are almost silent, can be refueled at home, offer
2 Goal and scope higher performance in terms of acceleration, and in general a
different quality of driven experience. In this sense EVs and
According to ISO 14040, in the following paragraphs, we ICEVs are, to certain extent, not really comparable. On the
describe the system we are going to analyze, in terms of pur-
1
pose and goals of the study, and definition of product system, http://www.eco-aesc-lb.com/en/product/liion_ev/.
Int J Life Cycle Assess (2015) 20:1127–1142 1129

other hand, giving that the European average daily travel dis- As regard ICEVs, we considered the lifecycle of gasoline,
tance, and in particular the Italian average daily travel dis- including crude oil extraction and transportation and gasoline
tance, is far lower than commercial EVs range (Pasaoglu refining and distribution.
et al. 2013), we can state that a comparison is somehow For EVs, we assessed the life cycle of the electricity vector,
possible. from energy sources supply, to electricity production and
transmission losses.
As regard the charging stations, we did not consider their
2.3 Functional unit
impacts because they are negligible; at the same time, the
construction of infrastructure for gasoline distribution (gas
The service provided by these vehicles is the distance driven.
station) was not taken into account.
Accordingly, the functional unit of the study is based on the
kilometers driven. We assumed the lifetime of the vehicle is
150,000 km, and we supposed that the battery will last as long
2.5 Impact categories
as the vehicle (Zackrisson et al. 2010). The functional unit
used in this LCA is 150,000 km, similar to many other studies
For the quantification of the environmental impacts (LCIA),
(Duvall et al. 2007; Sullivan and Gaines 2010; Hearron et al.
we chose to assess different impact categories, ensuring to
2011; Gao and Winfield 2012; MacPherson et al. 2012;
include almost all the main relevant effects of the diffusion
Szczechowicz et al. 2012; Hawkins et al. 2013; Sharma
of EV in the vehicles fleet. Table 1 shows the list of impact
et al. 2013).
categories and relative indicators.
The categories chosen include air pollution and global
2.4 System boundaries warming, as those are the main reasons that led to the intro-
duction of EVs in the vehicle fleet. In particular, we assessed
The approach chosen to describe and analyze the systems is air acidification (CML, Guinée et al. 2002), photochemical
the cradle to grave approach, and it considers the following oxidant formation (ReCiPe), and greenhouse effects (IPCC).
(see Figs. 1 and 2): Besides these, we included eutrophication (CML, Guinée
et al. 2002), human toxicity (CML, Guinée et al. 2002), re-
& The vehicles construction and disposal source depletion (CML, Guinée et al. 2002), cumulative ener-
& The battery construction and disposal gy demand nonrenewable (CED nonrenewable), and particu-
& The energy vectors (electricity for EV and gasoline for late matter formation potential (ReCiPe).
ICEV) supply chains In the selection of the methods to quantify the impacts, we
& Use phase of the vehicles (including vehicle maintenance) followed as close as possible the guidelines provided by JRC
& Road construction, disposal, and maintenance considering (2011) in the framework of the European Platform on Life
100 years of road life Cycle Assessment (EPLCA, http://eplca.jrc.ec.europa.eu/).

Fig. 1 Life cycle phases

considered for EV
1130 Int J Life Cycle Assess (2015) 20:1127–1142

Fig. 2 Life cycle phases

considered for ICEV

An exception was made for human toxicity. JRC (2011) sug- electricity consumption came from both literature (Hacker
gests the use of the USEtox method (Rosenbaum et al. 2008). et al. 2009; Anair and Mahmassani 2012) and RSE experi-
This methodology is under development and currently does mental data (Celaschi et al. 2013). Data for electricity 2013 are
not include the effects of fine particulate matter (PM) emis- referred to Italian efficiency for single technology (combined
sions on human toxicity. Given the fact that PM emissions are cycle, gas turbine, etc.) as declared by the national Transmis-
a big issue in the transport sector, we preferred to use CML sion System Operator (TERNA 2013) and emission rates for
method that, on the contrary, includes the contribution of regulated emission (CO2, NOx, SOx, particulate matter) for
PM10 to the total value of the indicator. CML method for thermal power plants derive from annual environmental dec-
human toxicity is anyway considered a good method in JRC laration of Italian Emas registered Power Plants (Girardi
recommendations. 2012). For 2030, data come from the EU project Needs LCI
database (http://www.needs-project.org/needswebdb/search.
php) and are referred to year 2025 (Scenario: 2025, realistic-
2.6 Data quality optimistic, 440 ppm.).

Most of the data used in this LCA are secondary data, that is,
they came from literature or from databases. The main data
source used is the database Ecoinvent v2.2 (Ecoinvent 2010). 3 Life cycle inventory
An important exception is the use phases for ICEV and the
electricity supply system for EV. For the ICEV fuel consump- In the following paragraphs, the main assumptions in the LCI
tion and the main emission factors came from the application for each life cycle phase are explained. For each paragraph
of the COPERT model (Ntziachristos et al. 2009) for the Ital- (i.e., for each life cycle phase), detailed information for ICEV
ian National Emission Inventory (ISPRA 2014). For EV, and EV are presented.

Table 1 Impact categories and indicators

Impact category Indicator U.M.

Climate change IPCC greenhouse effects (100)–2007 kg CO2 eq.

Photochemical oxidant formation ReCiPe photochemical oxidant formation potential kg NMVOC eq.
Acidification CML 2001 acidification (acidification potential, European average) kg SO2 eq.
Eutrophication CML 2001 eutrophication (eutrophication potential, generic) kg PO43 eq.
Resource depletion CML 2001 depletion of abiotic resources kg Sb eq.
Human toxicity CML 2001 human toxicity g 1,4-DCB eq.
Primary energy Cumulative energy demand nonrenewable MJ eq.
Particulate matter ReCiPe particulate matter formation potential g. eq. PM10
Int J Life Cycle Assess (2015) 20:1127–1142 1131

Fig. 3 Charge hourly profile 7%

6%

5%

4%

3%

2%

1%

0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

3.1 Well to tank (upstream) provides these increments must be considered, as pointed out
also in other similar studies (Amarakoon et al. 2013).
The Bwell to tank^ phase of ICEV system was modeled using In the Italian electricity market, the electricity price corre-
Ecoinvent Database (Frischknecht and Rebitzer 2005), ver- sponds to the clearing price resulting from the intersection be-
sion 2.2. We used the life cycle inventory Bpetrol, low-sulfur, tween the volumes of electricity demanded and offered by its
at regional storage^ relating to the distribution of petroleum participants. The schedules of injection and withdrawal of elec-
product to the final consumer (household, car, power plant, tricity into and from the grid are defined under the economic
etc.), in Europe. Operation of storage tanks and petrol stations merit-order criterion (supply offers are ranked in increasing price
and all necessary transports are included. order and demand bids are ranked in decreasing price order).
The well to tank phase of EV system consists in the elec- The GME BGestore dei Mercati Energetici S.p.A.^ (the compa-
tricity supply. The electricity mix used in the battery is one of ny in charge of organizing and managing the electricity market)
the main important parameters for the LCI of EVs, and it is publishes annual figures (http://www.mercatoelettrico.org/it/
crucial to understand which kind of electricity mix will charge download/DatiStorici.aspx) for the index of marginal
the EV battery (Duvall et al. 2007). To model this phase, technology (ITM), defined as Bthe ratio of the sum of the
hourly battery charge profile and hourly electricity mix have volumes sold in the geographical zones (included in the
to be considered. Macro-Zone) where the technology has set the price, to the
In relation to battery charge profile, we referred to Lanati sum of the overall volumes sold in the Macro-Zone.^
et al. (2011). In that study, the charging profile derives from a So, according to Italian electricity market rules, the hourly
survey on mobility carried out by Milan municipality. Figure 3 electricity mix was calculated taking into account the marginal
shows the charging profile assumed. The charging profile technologies, i.e., the technologies that during the year have
considers that almost 36 % of the charging occurs during the set the price in each hour of the day.
day, while the percentage of daytime charging considered in Using GME annual report (GME 2012), for each hour of the
other studies (e.g., Duvall et al. 2007) is usually around 25 %. year, it is possible to know which technology has been margin-
As regard electricity, we supposed that in 2013, the use of al, i.e., which is the technology that has set the price in each
few EVs determines the addition of marginal increments of market zone (Italy national grid is divided into several market
electricity demand, and consequently, the electricity mix which zone). The ratio between the volume of electricity over which a

100% 100%
90%
80% 80%
70%
60%
60%
40% 50%
40%
20% 30%
20%
0% 10%
Coal Oil Natural Gas 0%
CCGT Gas turbine Hydro- pumped
Hydro-run off river Hydro reservoir Import CCGT Coal Coal with CCs
Fig. 4 Electricity mix for EV2013 Fig. 5 Electricity mix for EV2030
1132 Int J Life Cycle Assess (2015) 20:1127–1142

Table 2 EV consumption rate according to Anair and Mahmassani acceptable also in consideration of the negligible contribute
(2012)
of RES (less than 0.1 %) to the charge electricity mix EV2013.
Brand Mitsubishi Ford Nissan Chevy In order to calculate the LCI of the described electricity mix
supplied to EV, we use the Ecoinvent database V.2.2 but, as
Car Bi^-ev Focus Leaf Volt regards fossil-fueled power plants, we made some modifica-
kWh/km 0.19 0.20 0.21 0.22 tions. First of all, we changed the efficiency of the power
plants according to the Italian actual situation as reported by
TERNA (the Italian Transmission System Operator) statistical
specific technology (e.g., CCNG) has set the price and the data (TERNA 2013). Moreover, we recalculated the emission
entire volume of electricity sold in the same period is the so rates of the regulated emissions (CO2, CO, SOx, NOx, PM10,
called marginal technology index (GME 2012). Hence, if we PM2.5) on the basis of more than 50 annual environmental
sum the volume of electricity over which a specific technology declarations of Italian power plants EU Environmental Man-
(e.g., CCNG) has set the price in the first hour of each day of agement and Audit Scheme (EMAS) registered (Girardi
every day of a year and we divide this sum by the entire volume 2012). This should lead to a more precise electricity mix
of electricity sold price in the first hour of each day of every day LCI. As regards import we used EU mix from Ecoinvent 2.2
of the same year, we have marginal technology index of the As regards the 2030 scenario, the national electricity sup-
specific technology during the first hour of the day (0–1 am). Of ply system could be influenced by the additional electricity
course, we can make the same for each hour of the day and for demand due to the EV fleet. Of course the influence on the
each technology. The so estimated marginal technology indexes national electricity supply system depends on the volume of
are the weights used to build the marginal power mix for each the national EV fleet. Moreover, even in the 2030 scenario, we
hour of the day. Using the EV battery charging profile, we should understand which part of the national electricity supply
calculate the power mix used by EV. Hence, the electricity mix will charge the battery of the EVs. To this end, we assume
mix considered is not stable during the year, but we use aggre- that the charging profile is the same of the 2013 scenario
gated average data for the entire year. (Lanati et al. 2011) as for example in Duvall et al. (2007).
The technologies taken into account in GME report and Of course, it is difficult to predict future average charging
here utilized for electricity mix calculation are as follows: profile as far as recharging behavior of EV drivers are not well
known or understood in advance and can be influenced by
1. Coal power plant several factors such as charging infrastructure, energy costs,
2. Oil power plant and of course battery technology (Azadfar et al. 2015). In
3. Combined cycle gas turbine (CCGT) order to understand which supply mix will recharge EV bat-
4. Natural gas (except CCGT and gas turbine) teries in 2030, we referred to the results of a study concerning
5. Gas turbine the impact on the development and on the operation of the
6. Hydropower (runoff river plant) Italian power system of a significant penetration of plug-in
7. Hydropower (reservoir power plant) electric vehicles (Lanati et al. 2011) for the year 2030. In their
8. Import study, Lanati et al. (2011) assumed that EVs reach about 25 %
9. RES-renewable (solar and wind) of the whole car fleet in 2030. Starting from the Bcentral^
forecast of the evolution of population by the Italian Institute
Known the charge profile and the hourly electricity mix, of Statistics ISTAT, that assumes about 62 million inhabitants
computed as above reported, the daily average electricity mix in 2030 (ISTAT 2008) and extrapolating the current trend of
for battery charge was determined and is shown in Fig. 4. the number of inhabitants per car, in 2030, we would reach a
As GME report provides aggregate data for wind and pho- value around 1.55, the total Italian car fleet in 2030 will be of
tovoltaic (renewable), we estimated the share of wind and 40 million cars of which 10 million electric. Lanati et al. then
photovoltaic electricity in proportion to corresponding annual estimate the overall electricity demand of the considered EVs
electricity productions (TERNA 2013). This assumption is that in 2030 amounts to about 17.5 TWh (including losses for

Table 3 Main characteristics of the considered ICEV and EV

Total weight Fuel consumption Range Cubic capacity

EV 2013 1479 (of which 218 Li-ion battery and 104 electric motor) 19 kWh/100 km 80 –
EV 2030 1479 (of which 218 Li-ion battery and 104 electric motor) 19 kWh/100 km 100
ICEV 2013 1059 (of which 276 drive train) 6.83 kg/100 km – 1.4–2.0 l
ICEV 2030 1059 (of which 276 drive train) 5.39 kg/100 km – 1.4–2.0 l
Int J Life Cycle Assess (2015) 20:1127–1142 1133

Table 4 Main ICEV emission data than 60 % by CCGT, but with a not negligible contribution of
g/km CO2 NOx SOx PM25 PM10 NMVOC coal-fired power plants, both with and without carbon dioxide
capture and sequestration (CCS). Figure 5 shows the electric-
ICEV 2013 212.7467 0.0435 0.0007 0.0085 0.0144 0.1383 ity mix for EV2030 scenario. Although we refer to Lanati
ICEV 2030 165.7297 0.0339 0.0005 0.0083 0.0142 0.1077 et al. (2011) for details, it is important to remember here their
main hypothesis:
Emission includes evaporation tire and brake abrasion. 2013 data came
from ISPRA (2014)
& The EV diffusion is not a key driver for renewable energy
sources (RES) development. Renewable development will
battery charging and discharging), corresponding to about be guided by European target, and EVs will not influence it.
18.7 TWh including network losses (TERNA 2013). & EV electricity demand is seen as an additional demand.
In order to evaluate how electricity will be supplied to EVs, Hence, given that RES will be already used, additional
Lanati et al. (2011) used two models. demand will be covered by fossil fuel power plants.
The MATISSE model (Cavicchioli et al. 2006; Borgarello
et al. 2008) is based on the Markal-TIMES model generator by To characterize the power plants in 2030 scenario, we used
ETSAP-IEA (Loulou et al. 2004, 2005) and has been developed the Needs project LCA dataset and more in detail:
by RSE to analyze scenarios over a long-term time horizon
(typically up to 2050). The model can combine the energetic, & For CCGT: electricity, natural gas, at combined cycle
socioeconomic, and environmental constraints of scenarios set plant, 500 MWe (RER 2025)
up by the user to determine the optimal development (in terms & For coal with CCS: electricity, hard coal power plant
of least overall cost) of the power system, making compete the 500 MW class post CCS, 200 km; 800 m aquifer (RER,
different generation technologies available to meet demand. 2015–2035) considered the most likely solutions for CCS
Medium Term SIMulator (MTSIM), developed by RSE in Italy for year 2030 (Girardi 2013)
(Zani and Migliavacca 2009), is a zonal electricity market sim- & For coal: electricity, hard coal, at power plant 800 MW
ulator able to determine the hourly dispatching of the generation (RER, 2015–2035)
set and the clearing of the market over an annual time horizon.
Lanati et al. (2011) in their study used MTSIM in a syner- From a methodological point of view, it may appear that we
getic way with MATISSE: the results of the latter, in terms of are somehow mixing attributional and consequential LCA ap-
long-term evolution of the generation set, have been used as proaches. According to Weidema (2003), the main distinctive
inputs for the former that, focusing on year 2030, has simu- characteristics of consequential LCA (CLCA) compared to at-
lated the operation of the power system in a more accurate and tributional LCA (ALCA) are that Bprocesses are included to the
detailed way. The results of the simulations on the develop- extent of their expected change caused by a demand (affected
ment of the Italian power system carried out with the MATI processes)^ and that “co-products are handled by system ex-
SSE model show that there is no significant difference, in pansion.” This is not the case of the present work. On the
terms of installed generation capacity, between the scenarios contrary, we agree with Zamagni et al. (2012), who consider
with and without the foreseen penetration of EVs. This is due CLCA not strictly a methodology but an approach to Bdeepen
to the limited impact of the electricity demand of EVs, corre- LCA,^ taking market mechanisms into account in order to in-
sponding to less than 5 % of the total 2030 end-use demand. crease the robustness of the study rather than a modeling prin-
As for the main results of the MTSIM simulations concerning ciple with defined rules. A similar approach has already been
the operation of the 2030 power system, all the electricity applied by Amarakoon et al. (2013), where marginal electricity
supplied to EVs will be produced using fossil fuels and more generation is considered so that impacts from the standpoint of

ICEV ICEV

EV2013 EV2013

0% 20% 40% 60% 80% 100% 120% 140% 0% 20% 40% 60% 80% 100% 120% 140%

Well to Tank Manufactoring- car Manufactoring- battery Well to Tank Manufactoring- car Manufactoring- battery
Use End of life-car End of life-battery Use End of life-car End of life-battery
Road Road
Fig. 6 IPCC greenhouse effect by life cycle stage for ICEV and EV Fig. 7 CML air acidification by life cycle stage for ICEV and EV (2013
(2013 scenario) scenario)
1134 Int J Life Cycle Assess (2015) 20:1127–1142

ICEV ICEV

EV2013 EV2013

0% 20% 40% 60% 80% 100% 120% 140%

0% 20% 40% 60% 80% 100% 120% 140%
Well to Tank Manufactoring- car Manufactoring- battery
Well to Tank Manufactoring- car Manufactoring- battery Use End of life-car End of life-battery
Use End of life-car End of life-battery
Road
Road
Fig. 8 CML depletion of abiotic resources by life cycle stage for ICEV Fig. 9 Cumulated energy demand (fossil) by life cycle stage for ICEV
and EV (2013 scenario) and EV (2013 scenario)

the addition of marginal increments of demand are taken into emission inventory guidebook (EEA, EMEP 2009), has de-
account. Moreover, we also make sensitivity analysis for the veloped a national methodology to estimate emissions. In par-
2030, as suggested by Mathiesen et al. (2009). ticular, the model COPERT-4 (Ntziachristos et al. 2009) is
used to estimate emissions of transport sector.
3.2 Manufacturing phase We modified the Ecoinvent dataset replacing fuel
consumption and atmosphere emissions with those of the
The manufacture of the ICEV was taken from Ecoinvent v2.2 NIR, published by ISPRA (2014) for an average Euro 5 Gas-
(2010)2, and the inventory includes processes of material, en- oline car 1400–2000 cm3, representative of the lower medium
ergy, and water use in manufacturing of a vehicle that respects European car market segment
the Euro5 normative. The process is representative of the man- Consumption and emission are not New European Driving
ufacture of a VW Golf (A4 VW platform), assumed as repre- Cycle (NEDC) based, but estimated according to national in-
sentative of the lower medium market segment (ICCT 2013). ventory methodology, and are supposed to be more realistic
Data are declared to be representative of the European situa- than the standard NEDC estimates.
tion, for year 2007 by the Ecoinvent database. The Ecoinvent flows modified are as follows: CO, NOx,
As regard the EV, data came from Ecoinvent v2.2 (2010)3 PM10, PM2,5, NMVOC, benzene, CH4, N2O, NH3, CO2,
and are representative of an electric vehicle similar to a VW SO2, Pb, cadmium, copper, chromium, nickel, selenium, and
Golf, like in Notter et al. (2010). The battery pack is consti- zinc. Also, nonexhaust emissions have been taken into account.
tuted by LiMn2O4/graphite as in many other studies (Notter EV use phase includes maintenance according to Ecoinvent
et al. 2010; Amarakoon et al. 2013; Hawkins et al. 2013). The data, as far as emissions due to abrasion (tyres, brakes…). EV
battery pack has a weight of 218 kg and a charge capacity of energy consumption considered is 0.19 kWh/km (including
24 kWh, similar to a Nissan Leaf. Data, taken from Ecoinvent charge and discharge battery efficiency), in accordance with
V2.2. (2010), are those used by Notter et al. (2010). They are literature values and in particular with Hacker et al. (2009) and
declared to be representative of the European situation, for also with RSE experimental data (Celaschi et al. 2013). This
year 2009–2010, and considered one of the most accurate value is slightly higher than the one declared for example to
available study (Dunn et al. 2012). Data for ICEV and EV the Nissan Leaf NEDC (0.17 kWh/km), but it is more similar
are mainly derived from Ecoinvent 2.2 except for EV battery to Breal use^ consumptions (Anair and Mahmassani 2012;
weight and fuel consumption. The main characteristics of Girardi 2013).
ICEV and EV for 2013 and 2030 are listed in Tables 2 and 3. Of course, for both EVs and ICEVs, efficiency improve-
ment is foreseeable until year 2030. As regard the EV, the
3.3 Use phase
ICEV
The use phase for ICEV is based on Ecoinvent data (in partic-
ular the life cycle inventory Boperation, passenger car, petrol,
EURO5^) and Italian National Inventory Report (NIR) data. EV2013
Institute for Environmental Protection and Research
(ISPRA), according to the IPCC Guidelines and Good Prac-
0% 20% 40% 60% 80% 100% 120% 140%
tice Guidance (UNEP 1995) and the EMEP/EEA air pollutant Well to Tank Manufactoring- car Manufactoring- battery
Use End of life-car End of life-battery
2
Ecoinvent v2.2, passenger car, RER [unit] (#1936) Road
3
Ecoinvent v2.2, passenger car, electric, LiMn2O4, at plant, Fig. 10 ReCiPe photochemical oxidant formation potential by life cycle
RER [unit] (#11,774) stage for ICEV and EV (2013 scenario)
Int J Life Cycle Assess (2015) 20:1127–1142 1135

ICEV ICEV

EV2013
EV2013

0% 20% 40% 60% 80% 100% 120% 140% 160%

0% 20% 40% 60% 80% 100% 120% 140%
Well to Tank Manufactoring- car Manufactoring- battery
Well to Tank Manufactoring- car Manufactoring- battery
Use End of life-car End of life-battery Use End of life-car End of life-battery
Road Road
Fig. 11 CML eutrophication by life cycle stage for ICEV and EV (2013 Fig. 13 ReCiPe particulate matter formation potential by life cycle stage
scenario) for ICEV and EV (2013 scenario)

efficiency considered (0.19 kWh/km) is already optimistic

and further improvements will be probably due to solutions Ecoinvent module considers transports and municipal and
not involving powertrain such as weight or rolling resistance hazardous waste incineration.
reduction (Offer et al. 2010). Of course, those kinds of im- BEnd of life^ of EV2013 and EV 2030 were calculated on
provements are cross-technological and can be applied also the basis of Ecoinvent data. We extrapolated Li-ions battery
to ICEV. For sake of simplicity, and because we are inter- disposal from the module relating to electric vehicle (“dispos-
ested in comparison, we consider only not cross- al, electric vehicle, LiMn2O4”) in order to have two separated
technological improvements, i.e., only improvements regard- modules and highlight the relative contributions to life cycle
ing engine and transmission. Hence, for EV in 2030, we assessment.
consider again 0.19 kWh/km. For ICEV, it has been estimat- The dataset employed for the battery disposal considers a
ed that between 2008 and 2020, thanks only to engine and mix of two technologies (pyrometallurgical and hydrometal-
transmission solutions, gasoline ICEVs can reach efficiency lurgical process) for recycling of Li-ions battery.
improvement in range of 15.6–31 % with an average of
22.1 % (Nemry et al. 2008) We consider the same efficiency
improvement (22.1 %) for our 2030 gasoline ICEV. So,
starting from 2013 data, ICEV fuel consumption and emis- 4 Results and discussion: life cycle impact assessment
sion rates related to fuel combustion are reduced of 22.1 %
in 2030 scenario. In the following paragraphs, we compare the performances
Main emission data for considered ICEV in 2013 and 2030 of ICEV and EV, separately for 2013 and 2030 scenarios.
are reported in Table 4. Data are presented as a percentage of the ICEV impacts. The
influence of the well to wheel phases (well to tank and use
phase for ICEV and electricity supply for EV) is higher than
3.4 End-of-life phase in other comparable studies (e.g., Volkswagen Group 2014).
This is probably due to the fact that we use consumption and
The disposal of passenger car was modeled using Ecoinvent emission data more similar to real use than the data from
life cycle inventory Bdisposal, passenger car^ which includes homologation cycle. If we consider the overall performance
disposal processes for bulk materials. In particular, the of EV compared to ICEV, this study highlights that EV
performs better than ICEV except for human toxicity and
eutrophication potential in line with other similar studies
(e.g., Hawkins et al. 2012; Amarakoon et al. 2013). Results
are also in line with Notter et al. (2010) as regards acidifi-
ICEV cation potential, global warming potential, and cumulative
nonrenewable energy demand, even if in their study it is
not possible to separate well to wheel from operation con-
EV2013 tributes. The trade-offs are due to the EV battery
manufacturing phase which accounts for almost 45–47 %
0% 50% 100% 150% 200% of the overall human toxicity (2013 and 2030 scenario)
Well to Tank Manufactoring- car Manufactoring- battery and 31–38 % of the overall eutrophication potential. Similar
Use End of life-car End of life-battery
Road results are declared also by Hawkins et al. (2012) and
Fig. 12 CML human toxicity by life cycle stage for ICEV and EV (2013 Amarakoon et al. (2013), even considering other typology
scenario) of Li-ion batteries.
1136 Int J Life Cycle Assess (2015) 20:1127–1142

Fig. 14 Comparison among 200%

baseline EV (EV), EV with 180%
shorter battery life and higher 160%
consumption (EV+) and gasoline 140%
ICEV in the 2013 scenario 120%
100%
80%
60%
40%
20%
0%

EV EV+ ICEV

4.1 Scenario 2013 systems. Only for air acidification, we see a nonnegligible
contribution from the use phase of ICEV. For particulate mat-
For 2013 scenario, EVs bring a reduction in greenhouse gas ter formation, Fig. 13, the use phase of both vehicles is not
emissions of about 50 % compared to similar ICEVs (Fig. 6). negligible (13 % for EV and 19 % for ICEV).
Despite the bigger effect of electricity production (well to The overall reduction, with reference to the ICEV perfor-
tank) for EV compared to gasoline production, during the mance is quantifiable in about 40, 40, 47, and 24 %, respec-
ICEV use phase much more greenhouse gases are emitted. tively, for air acidification, depletion of abiotic resources,
This result is due both to the higher overall efficiency of the CED nonrenewable, and particulate matter formation.
EV and to the fact that in our scenario, electricity is partially The value of photochemical oxidant formation potential
produced by renewable resources. (Fig. 10) over the life cycle comes mainly from the well to
As regards air acidification (Fig. 7), depletion of abiotic tank phase for both vehicles (about 40 % of the overall impact)
resources (Fig. 8), and cumulated energy demand nonrenew- with a significant contribution of the ICEV use phase. The EV
able (Fig. 9), the well to tank phase is responsible for the leads to an improvement in the indicator of about 40 %.
majority of the impacts for both kind of vehicle. ICEV impacts Eutrophication (Fig. 11) and human toxicity (Fig. 12) are
are in all the three cases higher than EV impacts. The vehicles the only environmental impact categories in which ICEV per-
manufacturing and road life cycle are comparable in the two forms better than EV. In both of these impact categories, EV

Table 5 LCIA results for 2013 scenario (complete vehicles lifetime - 150,000 km)

Impact categories Unit EV EV+ ICE_gasoline

CED nonrenewable MJ 368,480.7 415,490.5 697,203.24

CML2001 air acidification g eq. SO2 80,678.38 94,064.87 132,262.95
CML2001 depletion abiotic resources kg eq. Sb 189.6437 214.0148 318.7499
CML2001 eutrophication g eq. PO4 35,989.7 43,126.02 28,359.595
CML2001 human toxicity g eq.1,4-DCB 20,538,605 25,608,144 14,219,070
IPCC2007 greenhouse effect g eq. CO2 23,264,509 26,461,833 46,208,980
ReCiPe photochemical oxidant formation potential g eq. NMVOC 63,239.25 70,233.52 107,666.08
ReCiPe particulate matter formation potential g eq. PM10 31,653.61 35,665.71 41,429.264

EV+ is the EV with a higher energy consumption (22 kWh/100 km rather than 19 kWh) and with shorter battery life (100,000 km rather than 150,000)
Int J Life Cycle Assess (2015) 20:1127–1142 1137

ICEV
perform better anyway. For particulate matter formation
(Fig. 13) EV performs better than ICEV.
Hence, it is in these two impact categories that we can see
EV2030 an environmental impacts trade-off between EV and ICEV.
Of course, those results are strictly related to the hypothesis
0% 20% 40% 60% 80% 100% 120% 140% 160%
made. Among others, the most important hypotheses are the
Well to Tank Manufactoring- car Manufactoring- battery battery life and the EV energy consumption. In order to verify
Use End of life-car End of life-battery the robustness of the results we make the hypothesis that the
Road
EV battery could last 100,000 rather than 150,000. Moreover,
Fig. 22 ReCiPe particulate matter formation potential by life cycle stage as far as the assumed EV consumption of 0.19 kWh/km can be
for ICEV and EV (2013 scenario)
considered optimistic, we consider the higher consumption ob-
performs worst due to the battery manufacturing. It is possible served by Anair and Mahmassani (2012), that is, 0.22 kWh/km
to state that even if the EV battery manufactory contribution to (see Table 2). Results are shown in Fig. 14 where EV+ is the
these impact categories were the half, the ICEV would electric vehicle with a 10,000 km battery life and 0.22 kWh/km

Fig. 23 Comparison among 180%

baseline scenario for EV in 2030 160%
(EV_1), and EV recharged with
36 % (EV_PV_36) and 50 % 140%
(EV_PV_50) of electricity from
120%
photovoltaic power plants
100%

80%

60%

40%

20%

0%

ICEV EV EV_PV_36 EV_PV_50 EV_LB

Table 7 LCIA results for 2030 scenario (complete vehicles lifetime - 150,000 km)

Scenario 2030 Unit EV 2030 EVPV36 EV_PV_50 EVLB ICEV

CED nonrenewable MJ 354,880.74 278,385.58 248,637.47 349,828.22 568,829.85

CML2001 air acidification g eq. SO2 61,353.16 56,847.57 55,095.40 57,426.53 111,444.21
CML2001 depletion abiotic resources kg eq. Sb 193.98 150.87 134.10 191.24 261.63
CML2001 eutrophication g eq. PO4 29,320.57 29,638.84 29,762.62 26,430.31 25,441.66
CML2001 human toxicity g eq.1.4-DCB 19,691,367 19,550,393 19,495,570 17,357,927 12,863,494
IPCC2007 greenhouse effect g eq. CO2 22,332,799 17,271,631 15,303,400 21,964,550 37,426,347
ReCiPe photochemical oxidant formation potential g eq. NMVOC 54,542.36 50,028.73 48,273.43 52,931.92 91,097.32
ReCiPe particulate matter formation potential g eq. PM10 27,284.58 25,860.25 25,306.35 26,026.33 36,335.97

EVPV36 is the scenario with 36 % of energy of EV recharging from PV, EVPV50 with 50 % from PV and EVLB is the baseline EV but with a 25 %
lighter battery (164 kg instead of 218 kg)
1138 Int J Life Cycle Assess (2015) 20:1127–1142

ICEV ICEV

EV2030
EV2030

0% 20% 40% 60% 80% 100% 120% 140%

0% 20% 40% 60% 80% 100% 120% 140%
Well to Tank Manufactoring- car Manufactoring- battery
Well to Tank Manufactoring- car Manufactoring- battery
Use End of life-car End of life-battery
Use End of life-car End of life-battery
Road
Road
Fig. 15 IPCC greenhouse effect by life cycle stage for ICEV and EV
Fig. 17 CML depletion of abiotic resources by life cycle stage for ICEV
(2030 scenario)
and EV (2030 scenario)

consumption. Although impacts of EV+ are on average around eutrophication and human toxicity. For all the other
15 % higher than EV (except for eutrophication, 20 %, and indicators, EV shows lower impacts. We can observe that for
human toxicity, 25 %) from a general point of view the strength air acidification, the percentage contribution of the well to
and the weakness of EV when compared to gasoline ICEV tank phase (life cycle of electricity production) for EV
remain the same. LCIA results are summarized in Table 5. decreases from about 43 to 26 % and this makes the EV
more competitive. A higher reduction is appreciable in the
context of eutrophication, where the EV performance is
4.2 Scenario 2030 worse that ICEV, with 2030 scenario, only for about 15 %
(instead of more than 25 % of the 2013 scenario). In general
From Figs. 15, 16, 17, 18, 19, 20, 21 and 22, we show the terms, the EV weakness and strength points remain the same
results for the comparison between ICEV and EV for 2030 in the two considered scenarios. Of course, those results
scenario. It is important to remember that the differences in strongly depend on the electricity that will be used for
the two scenario are only related to the LCI of the electricity recharging the EV battery. The electricity scenario used for
used to charge EV battery and to the efficiency of the gasoline the year 2030, as mentioned, come from the Lanati et al.
ICEV that increases of 22.1 % (Nemry et al. 2008), while the (2011) study. No matter how reliable are the models they have
efficiency of the EV (0.19 kWh/km) is already high and no used, their results depend on the hypotheses they made, such
relevant efficiency improvement other than cross- as EV market penetration, renewable energy development and
technological ones are foreseeable (Offer et al. 2010). As re- so on. The main hypothesis on renewable potential of Lanati
gard the electricity, the differences rely on different power et al. (2011) is shown in Table 6. Of course, reality can evolve
supply mix but also on different power plant emission rate. in a different way. In particular, the penetration of the photo-
As already stated above, in the 2013 electricity LCI we con- voltaic (PV) in the Italian electricity production mix showed a
sidered emission rate referred to the Italian 2013 situation development far beyond many forecast made in previous
while for 2030, we refer to power generation technologies years (e.g., Contaldi et al. 2007). For this reason, for year
available in year 2025 and to their LCI as described in the 2030, we imagine two further scenarios: a scenario named
NEEDS LCA database (www.needs-progect.eu). As for the EV_PV_36, in which all the electricity recharged during the
2013 scenario, in general terms, we can assert that ICEV is day (36 % of the total energy used in average by the EV) is
able to give better environmental results only in terms of produced by photovoltaic power plants and another scenario,

ICEV ICEV

EV2030 EV2030

0% 20% 40% 60% 80% 100% 120% 140% 0% 20% 40% 60% 80% 100% 120% 140%
Well to Tank Manufactoring- car Manufactoring- battery Well to Tank Manufactoring- car Manufactoring- battery
Use End of life-car End of life-battery Use End of life-car End of life-battery
Road Road
Fig. 16 CML air acidification by life cycle stage for ICEVand EV (2030 Fig. 18 Cumulated energy demand (fossil) by life cycle stage for ICEV
scenario) and EV (2030 scenario)
Int J Life Cycle Assess (2015) 20:1127–1142 1139

ICEV ICEV

EV2030 EV2030

0% 20% 40% 60% 80% 100% 120% 140%

0% 20% 40% 60% 80% 100% 120% 140% 160% 180% 200%
Well to Tank Manufactoring- car Manufactoring- battery
Well to Tank Manufactoring- car Manufactoring- battery
Use End of life-car End of life-battery Use End of life-car End of life-battery
Road Road
Fig. 19 ReCiPe photochemical oxidant formation potential by life cycle Fig. 21 CML human toxicity by life cycle stage for ICE and EV (2030
stage for ICEV and EV (2030 scenario) scenario)

named EV_PV_50, in which the production of electricity covered by an EV (Amarakoon et al. 2013) while increasing
from photovoltaic power plant will be so competitive that energy density of 25 % would change the value of impact
50 % of the energy used by EV will come from PV (for indicators such as GWP in a range within 3–10 %
example, thanks to parking with integrated PV at working (Majeau-Bettez et al. 2011). Moreover, we think that foresee-
places, molls, and other public spaces). In both this further able increase in energy density will result in longer EV driving
scenarios, the production from PV come from tilted integrated range rather than in smaller batteries. For this reasons, even in
modules (data came from the NEEDS LCI database and are 2030 baseline scenario, we consider the same battery weight
also referred to year 2025) and the rest of the electricity is for EV, considering a smaller battery (25 % lighter) with the
produced as in Fig. 5. Results (see Fig. 23) show that even if same driving range only for a sensitivity analysis (called sce-
some impacts are strongly reduced by the use of PV for battery nario EV_LB). Results in Fig. 23 show that environmental
charging, for example, greenhouse effect or depletion of non- impact will be of course reduced by using smaller batteries,
renewable resources, the influence on eutrophication and on especially eutrophication and human toxicity. In particular, in
human toxicity is very little, and hence, again, the positive and the EV_LB scenario, the EV eutrophication potential will
negative effects of EV in comparison with ICEV remain the only be 4 % higher than the ICEV eutrophication potential,
same. Another important point of uncertainties is related to confirming the relevance of the evolution of batteries technol-
battery technology evolution. Even if it is difficult to foresee ogy as a key issue for this impact indicator. On the other hand,
the EV battery composition in 2030, they will be probable still we can state that the strength and weakness points of EV in
based on Li-ion, due to their unmatchable combination of high comparison with ICEV remain the same. LCIA results are
energy and power density (Nitta et al. 2014). Among others summarized in Table 7.
chemistry lithium cobalt phosphate (LCP), lithium iron phos-
phate (LFP), lithium manganese oxide (LMO), and lithium
cobalt manganese oxide (LCM) seem, for different reasons,
the more promising ones (Nitta et al. 2014). On the other hand, Table 6 Main
some battery comparative LCA studies have showed that the hypothesis for the 2030 RES potentials [TWh] 2030
battery chemistry influence on LCIA is little (within 10 % for electric system scenario
boundaries Existing hydro 35
LCM, LFP LMO) when the functional unit is the distance
New mini hydro 10.6
Onshore wind 22.2
Offshore wind 5.8
Photovoltaic 10
ICEV Concentrated solar power 3.4
Geothermal 9.8
Waste 5
EV2030 Biomass 8
Biogas 3.2
0% 20% 40% 60% 80% 100% 120% 140% Total RES 113
Well to Tank Manufactoring- car Manufactoring- battery Total electricity demand without 362
Use End of life-car End of life-battery EVs
Road Total electricity demand with EVs 380
Fig. 20 CML eutrophication by life cycle stage for ICEV and EV (2030
scenario) Source: Lanati et al. (2011)
1140 Int J Life Cycle Assess (2015) 20:1127–1142

5 Conclusions iron phosphate (LiFePO4) type, the analysis does not reveal
an inversion of the tendency and the use of EVs does not result
This paper compares the environmental performance during in environmental benefits if compared to traditional ICEVs for
the life cycle of two passenger cars: an electric passenger car those two impact categories. Of course, as regard 2030, it is
and a gasoline passenger car. It is known that electricity pro- possible that a technological evolution in battery production
duction is the main cause of environmental impacts for exter- can reduce those impacts beyond our expectations. On the
nally chargeable vehicles (Nordelöf et al. 2014) and improv- other hand, it is important to underline that human toxicity is
ing estimates of the impacts associated with the electricity not the only impact category considered that affects human
used to charge a growing EV fleet is the most important step health. Human health is influenced also by air acidification,
towards understanding the impacts of EVs relative to ICEVs oxidant formation, particulate matters emissions, and green-
(Hawkins et al. 2012). Instead of using an hypothetical or house effect as well. Moreover, most of the emissions leading
average power mix, the study tries to identify which is the to human toxicity occur during the mining process for the car
Breal^ power mix that is going to supply energy to EVs con- and battery production. Usually, those operations are carried
sidering the EVs electricity demand as an additional demand. out far from high density urban areas where car are used.
Of course, the geographical variable has a huge influence on
the power mix, but also on driving behavior, and we decided
Acknowledgments This work has been financed by the Research Fund
to focus our attention on Italy. Moreover, since we think that for the Italian Electrical System decree of Italian Economic Development
information on future scenarios can be more useful than on Ministry November 9th 2012 and following.
today scenario, where EVs have a limited market penetration,
we performed our analysis on two different time horizons: Funding This work has been financed by the Research Fund for the
Italian Electrical System decree of Italian Economic Development Min-
2013 and 2030. Despite of the growing penetration of renew- istry November 9th 2012 and following.
able energy sources in the national power system, we demon-
strated that the electricity used to charge the EVs batteries is, Conflict of interest The authors declare that they have no conflict of
and will probably be, mainly produced from fossil fuels. Nev- interest.
ertheless, in both scenarios (2013 and 2030), EVs present
better performances for those kind of impacts that they are
supposed to lower, that is, for impacts related to air pollutant References
emissions such as air acidification, photochemical oxidant
formation, and particulate matter formation potential. More- Amarakoon S, Smith J, Segal B (2013) Application of life-cycle assess-
over, it should be underlined that the EV major contribution to ment to nanoscale technology: lithium-ion batteries for electric ve-
these impact categories is due to electricity and car production hicles. No. EPA 744-R-12-001
Anair D, Mahmassani A (2012) State of charge: electric vehicles’ global
life cycle. It means that these pollutants are emitted far from warming emissions and fuel-cost savings across the United States.
high density populated areas (where emissions from ICEV are UCS Publications, 2 Brattle Square Cambridge, MA 02138–3780
instead relevant). Finally, pollutants are emitted by high Azadfar E, Sreeram V, Harries D (2015) The investigation of the major
stacks, leading to much bigger dispersion than those emitted factors influencing plug-in electric vehicle driving patterns and
by ICEV at low level. So, it can be argued that the emission charging behaviour. Renew Sust Energ Rev 42:1065–1076
Borgarello M, Benini M, Gelmini A, Cavicchioli C (2008) The use of the
due to ICEV use phase are likely to have higher impact on MATISSE modelling system to assess the economic implications of
human health than the same amount of emissions coming GHG emissions reduction for the Italian electricity sector. CIGRÉ
from electricity life cycle. Even if the electricity used in General Session, 24–29 August 2008, Paris
EVs, in both our scenarios, is produced mainly from fossil Cavicchioli C, Gargiulo M, Lavagno E, Vitale S (2006) The Italian elec-
tricity sector: a regional and multi grid TIMES model. International
fuels, EVs have lower impacts also for categories like green-
Energy Workshop 2006, 27–29 June 2006, Cape Town
house effect and primary energy consumption. On the other Celaschi S, Colzi F, Gianinoni I, Savaresi S, Corti A (2013) Sostenibilità
hand, we can observe a trade-off between EVs and ICEVs economica e problematiche tecniche aperte dalla mobilità elettrica
powered gasoline if we analyze the potential impacts in terms ed ibrida alla luce di studi e sperimentazioni su veicoli reali. RSE
of eutrophication and human toxicity, due mainly to the toxic report 12000993 http://www.rse-web.it/documenti/documento/
314710
emissions in the manufacturing stage of batteries. Scenarios
Contaldi M, Gracceva F, Tosato G (2007) Evaluation of green-certificates
with high percentage of electricity produced by renewables policies using the MARKAL-MACRO-Italy model. Energ Policy
(photovoltaic) show increased EV performances for impact 35(2):797–808
categories like greenhouse effect or depletion of non- Dunn JB, Gaines L, Sullivan J, Wang MQ (2012) Impact of recycling on
renewable resources but not for eutrophication and human cradle-to-gate energy consumption and greenhouse gas emissions of
automotive lithium-ion batteries. Environ Sci Technol 46(22):
toxicity. As shown in other studies (Hawkins et al. 2013), even 12704–12710
considering other types of Li-ion batteries, such as lithium- Duvall M, Knipping E, Alexander M, Tonachel L, Clark C (2007)
nickel-cobalt-manganese oxide (Li-NCM) type or lithium- Environmental assessment of plug-in hybrid electric Vehicle’s.
Int J Life Cycle Assess (2015) 20:1127–1142 1141

Volume 1: Nationwide Greenhouse Gas Emissions. EPRI, Palo MacPherson ND, Keoleian GA, Kelly JC (2012) Fuel economy and
Alto, CA: 2007. 1015325 greenhouse gas emissions labelling for plug-in hybrid vehicles from
Ecoinvent Database v 2.2 (2010) Swiss Centre for Life Cycle Inventories. a life cycle perspective. J Ind Ecol 16(5):761–773
http://www.ecoinvent.org Majeau-Bettez G, Hawkins TR, Singh B, Strømman AH (2011) Life
EEA, EMEP (2009) EEA air pollutant emission inventory guidebook— cycle environmental assessment of lithium-ion and nickel metal hy-
2009.European Environment Agency (EEA), Copenhagen dride batteries for plug-in hybrid and battery electric vehicles.
Elgowainy A, Burnham A, Wang M, Molburg J, Rousseau A (2009) Environ Sci Technol 45:4548–4554
Well-to-wheels energy use and greenhouse gas emissions analysis Mathiesen BV, Münster M, Fruegaard T (2009) Uncertainties re-
of plug-in hybrid electric vehicles. Argonne National Laboratory. lated to the identification of the marginal energy technology
ANL/ESD/09-2 in consequential life cycle assessment. J Clean Prod 17(5):
Frischknecht R, Rebitzer G (2005) The Ecoinvent database system: a 1331–1338
comprehensive web-based LCA database. J Clean Prod 13(13): Nemry F, Leduc G, Alumdena M (2010) State of the research and devel-
1337–1343 opment and comparative analysis of energy and cost efficiency. JRC
Gao L, Winfield ZC (2012) Life cycle assessment of environmental and Technical Note JRC 54699. European Commission, Luxembourg
economic impacts of advanced vehicles. Energies 5(3):605–620 Nemry F, Leduc G, Mongelli I, Uihlein A (2008) Environmental
Girardi P (2012) Ibridi plug-in e veicoli elettrici: vantaggi e svantaggi Improvement of Passenger cars (IMPRO-car). Institute for
secondo un approccio LCA. Rapporto RdS RSE prot. 12001047 Prospective Technological Studies
Girardi P (2013) Auto elettrica: un opportunità o un rischio per Nitta N, Wu F, Lee JT, Yushin G (2014). Li-Ion battery materials: present
l’ambiente?, L’Energia Elettrica, aprile 2013 and future. Mater today
GME (2012) Annual Report www.gme.it Nordelöf A, Messagie M, Tillman A, Ljunggren Söderman M, Van
Guinée JB, Gorrée M, Heijungs R, Huppes G, Kleijn R, Koning Ad, Oers Mierlo J (2014) Environmental impacts of hybrid, plug-in hybrid,
LV, Wegener Sleeswijk A, Suh S, Udo de Haes HA, Bruijn Hd, Duin and battery electric vehicles—what can we learn from life cycle
Rv, Huijbregts MAJ (2002) Handbook on life cycle assessment. assessment? Int J Life Cycle Assess 19(11):1866–1890
Operational guide to the ISO standards. I: LCA in perspective. IIa: Notter DA, Gauch M, Widmer R, Wager P, Stamp A, Zah R, Althaus H
Guide. IIb: Operational annex. III: Scientific background. Guinée (2010) Contribution of Li-ion btteries to environmental impact of
JB. Centrum Milieukunde Leiden (CML), Leiden University. electric vehicles. Environ Sci Technol 44:6550–6556
Kluwer, Dordrecht Ntziachristos L, Gkatzoflias D, Kouridis C, Samaras Z (2009) COPERT:
Hacker F, Harthan R, Matthes F, Zimmer W (2009) Environmental im- a European road transport emission inventory model. In Inf. tech-
pacts and impact on the electricity market of a large scale introduc- nologies in environmental engineering (pp. 491–504). Springer
tion of electric cars in Europe Critical Review of Literature. ETC./ Berlin Heidelberg
ACC Technical Paper 2009/4 Offer GJ, Howey D, Contestabile M, Clague R, Brandon NP (2010)
Hawkins T, Gausen O, Strømman A (2012) Environmental impacts of Comparative analysis of battery electric, hydrogen fuel cell and
hybrid and electric vehicles—a review. Int J Life Cycle Assess hybrid vehicles in a future sustainable road transport system.
17(8):997–1014 Energ Policy 38(1):24–29
Hawkins TR, Singh B, Majeau‐Bettez G, Strømman AH (2013) Pasaoglu G, Fiorello D, Martino A, Zanic L, Zubaryevaa A, Thiel C
Comparative environmental life cycle assessment of conventional (2013) Travel patterns and the potential use of electric cars—results
and electric vehicles. J Ind Ecol 17(1):53–64 from a direct survey in six European countries. Technol Forecast Soc
Hearron JD, McDonough M, Ranjbar A, Wei W, Chenjie L, Shamsi P, 87:51
Manohar S, Fahimi B (2011) The sustainability of new technologies Rosenbaum RK, Bachmann TM, Gold LS, Huijbregts MA, Jolliet O,
in vehicular transportation. In: 2011 I.E. Vehicle Power and Juraske R, Hauschild MZ (2008) USEtox—the UNEP-SETAC tox-
Propulsion Conference, Piscataway, NJ, USA, 6–9 Sept. 2011. icity model: recommended characterisation factors for human toxic-
2011 I.E. Vehicle Power and Propulsion Conference. IEEE. doi: ity and freshwater ecotoxicity in life cycle impact assessment. Int J
10.1109/vppc.2011.6043194 Life Cycle Assess 13(7):532–546
ICCT, International Council on Clean Transportation (2013) European Sharma R, Manzie C, Bessede M, Crawford RH, Brear MJ (2013)
Vehicle Market Statistics Pocketbook 2013 Conventional, hybrid and electric vehicles for Australian driving
ISO (2006) Environmental management—life cycle assessment—princi- conditions. Part 2: Life cycle CO2-e emissions. Transport Res C-
ples and framework, ISO 14040:2006. International Organization Emer 28:63–73
for Standardization, Geneva Sullivan JL, Gaines L (2010) A review of battery life-cycle analysis: state
ISPRA (2014) Italian Emission Inventory—Informative Report 2014 of knowledge and critical needs (trans: Center for Transportation
http://www.sinanet.isprambiente.it/it/sia-ispra/serie-storiche- Research ESD). ANL report. Argonne National Laboratory, U.S.
emissioni/informative-inventory-report-2012/view Department of Energy, Argonne, Illinois. (Review report ANL/
ISTAT (2008) Demographic forecasts. demo.istat.it/uniprev/ ESD/10-7)
index.html?lingua=ita Szczechowicz E, Dederichs T, Schnettler A (2012) Regional assessment
JRC (2011) ILCD Handbook: Recommendations for Life Cycle Impact of local emissions of electric vehicles using traffic simulations for a
Assessment in the European context. First Edition, EUR 24571 EN use case in Germany. Int J Life Cycle Assess 17(9):1131–1141
– 2011 TERNA (2013) Statistical Data, General Data, http://www.terna.it
Lanati F, Benini M, Gelmini A (2011) Impact of the penetration of electric UNEP, O. E. C. D. (1995) IPCC Guidelines for National Greenhouse Gas
vehicles on the Italian power system: a 2030 scenario. IEEE Power Inventories. IPCC, Bracknell, 3
Energy Soc Gen Meet 1(8):24–29 Volkswagen Group (2014) The e-Golf Environmental Commendation –
Loulou R, Goldstein G, Noble K (2004) Documentation for the MARK Background Report http://www.volkswagenag.com/content/
AL Family of Models. ETSAP. http://www.iea-etsap.org/web/ vwcorp/info_center/en/publications/publications.acq.html/icp-2/
Times.asp index.html
Loulou R, Remne U, Kanudia A, Lehtila A, Goldstein G (2005) Weidema BP (2003) Market information in life cycle assessment.
Documentation for the TIMES Model - PART I 1–78. http://www. Environmental Project no. 863. Danish Environmental Protection
iea-etsap.org/web/Times.asp Agency, Copenhagen
1142 Int J Life Cycle Assess (2015) 20:1127–1142

Zackrisson M, Avellán L, Orlenius J (2010) Life cycle assessment of Zani A, Migliavacca G (2009) A scenario analysis of the Italian
lithium-ion batteries for plug-in hybrid electric vehicles—critical Electricity market in 2020: emissions and compliance with EU tar-
issues. J Clean Prod 18(15):1517–1527 gets, International Energy Workshop 2009, Venice, Italy
Zamagni A, Guinée J, Heijungs R, Masoni P, Raggi A (2012) Lights and
shadows in consequential LCA. Int J Life Cycle Assess 17:904–918