Conference PaperPDF Available

Life Cycle Assessment of a Smartphone

Authors:

Figures

Content may be subject to copyright.
Life Cycle Assessment of a Smartphone
Mine Ercan, Jens Malmodin, Pernilla Bergmark
Ericsson Research, Ericsson AB
Stockholm, Sweden
mine.ercan@ericsson.com
Emma Kimfalk (former employee), Ellinor Nilsson
Sony Mobile Communications
Corporate Sustainability Office
Lund, Sweden
ellinor.nilsson@sonymobile.com
Abstract It is of interest to understand the life cycle
contribution from the use of smartphones including their
network usage, as well as to gain knowledge regarding the
impact of the smartphone as a device to provide input for
network studies. This cradle-to- grave study is based on life
cycle assessment (LCA) methodology as outlined by the
ISO14040 series and the supplementing ICT specific LCA
standard from ETSI/ITU. The paper provides details
regarding data collection, assumptions, methods and results.
Furthermore, sensitivity analysis results for selected
parameters are presented, including variations due to different
secondary data sets. This study calculates the Global Warming
Potential (GWP) for the assessed smartphone (a Sony Mobile
Z5) including accessories) to 57 kg CO2e for an assumed
operating life time of 3 years, excluding the network usage.
Results are also presented for other impact categories and as
yearly figures. In addition, the distribution of impacts between
life cycle stages is provided for the assessed impact categories.
Integrated circuit (IC) production is identified as a major
contributor to the overall impacts followed by the production
of the display. For GWP specifically, overall results are also
provided including the network usage
Index TermsICT, GHG emissions, LCA, life cycle
assessment, smartphone, telecommunication
I. INTRODUCTION
It is commonly known that the use of smartphones in
mobile communication networks is rapidly growing
worldwide and thereby their contribution to the
environmental and economic impacts of telecommunication
networks. Although our research forecasts that the
Information and Communication Technology (ICT) sector to
remain within 2% of the total global greenhouse gas (GHG)
emissions until 2020 [1], such development indicates an
increased need for information regarding the environmental
effects of smartphones. Recently, initiatives to reduce the
energy consumption and the GHG emissions of
communication networks have gained momentum. There is
also an increasing interest in resource efficiency and the
circular economy [2], and in eco-rating of mobile phones [3-
4]. Consequently life cycle impact assessments are needed
for smartphones to build a comprehensive understanding of
their potential environmental impacts.
In their literature review of mobile phone LCAs [5],
Suckling and Lee note that most studies that assess the life
cycle impacts of mobile phones (including smartphones but
also feature phones), report impact only in terms of GWP or
energy [6]. As noted by Moberg et al. [7], the global
warming potential (GWP) category alone cannot be used to
represent all environmental impacts and thus broader studies
are needed to gain a comprehensive understanding of the
environmental impacts of mobile phones, especially toxicity
impacts (which on the other hand show large uncertainties as
this study shows). Suckling and Lee [5] also note that the
majority of studies are published by manufacturers and note
several circumstances that make results hard to compare. The
authors of this study agree with this view and make every
effort to present results for a broad range of impact
categories and a sufficient level of detail, in order for the
results to be comprehensive and interpretable. However, due
to its importance, and considered as the most important
impact category by the ETSI/ITU standard, GWP is given a
greater focus than other impact categories.
The paper outline is as follows: firstly the methodology is
presented in section II, then the goal and scope in section III
and the life cycle inventory (LCI) in section IV. In section V
the results of the life cycle impact assessment (LCIA) are
given followed by interpretation and normalization combined
with a sensitivity analysis in section VI. The paper concludes
with a discussion in section VII and conclusions are finally
drawn in section VIII.
II. METHODOLOGY
This study is based on LCA methodology as outlined by
ISO and covers a multiple number of impact categories of
the smartphone and its associated network usage, from
cradle-to-grave. Furthermore it considers the joint standard
for LCA of ICT goods, networks and services developed by
European Telecommunications Standard Institute (ETSI) [8]
and International Telecommunication Union (ITU) [9].
In line with Ercan [6] and Moberg et al. [7], who suggest
to prioritize primary data collection efforts on key
components (primarily integrated circuits) and energy use
during production and use stage, this study has used primary
data for production processes to the extent possible and
calculated results for three different usage scenarios as actual
usage varies between users. Further data details are given in
section IV.
GaBi software [10] has been used as the modeling tool
for this study and as a source for secondary data, including
the data sets from Eco-invent as well as GaBs own data.
III. SCOPE
A. Product system
The study is targeting two new high-end smartphones by
Sony (models Z3 and Z5) with accessories (see Fig. 1). It
4th International Conference on ICT for Sustainability (ICT4S 2016)
© 2016. The authors - Published by Atlantis Press
124
includes the smartphone device itself as well as, for GWP,
the associated life cycle impact from the network usage. The
main difference between Z3 and Z5 models is the total
integrated (IC) chip area, where Z5 has a larger area (9.5
cm2) compared to Z3 (7.5 cm2). The touch-screens have the
same size, and most other parts and components are the
same; so the difference in weight is only a few grams.
A. The smartphone (Z3)
152 g (see details below)
B. Headset
16 g
C. USB-cable
21 g
D. Charger
50 g
E. Documentation
48 g
X. Delivery packaging
74 g (not included in picture)
Y. Transport packaging
66 g (not included in picture)
1. Frame/backside
27 g (mainly plastics)
2. Metal sheets
15 g
3. Display
21 g (facing down, not visible)
4. Battery
48 g
5. PBAs/ICs
13 g
6. Flex-films
6.5 g
7. Cameras
1.5 g
8. Other components
11 g
Fig. 1. Smartphone composition and accessories
Embodied impacts from software developed outside
Sony (apps in general) are not included in the scope,
however software impacts are considered for the use stage as
data center services are included in network usage.
B. Functional unit
The functional unit is set to life time usage (3 years) of
the smartphone device and its accessories for a
representative usage scenario.
See section IV for specifications of the usage scenario.
For GWP, results are also presented per year and
including the life cycle impact from the associated network
usage.
C. System boundaries
All life cycle stages and processes are included in
accordance with the joint ETSI/ITU LCA standard ([8]-[9])
except reconditioning of mobile phones for reuse. For a
detailed overview of processes refer to Figure 7 of the
standard. Two cut-offs were made: Impact from third party
overhead activities (e.g. marketing services) were not
included in the supporting activities. Impact from materials
beyond the around 30 most contributing materials according
to the previous experience of the authors were not accounted
for.
D. Impact indicators
Based on recommendations from the International
Reference Life Cycle Data System (ILCD) [11], the
environmental life cycle impact assessment (ELCIA)
indicators are chosen as presented in Table I together with
the adopted impact assessment methods.
TABLE I. IMPACT INDICATORS
ELCIA indicators as
recommended by ILCD
Unit Reference
Global Warming Potential (GWP)
CO
2
-eq.
IPCC, 100 years
Ozone Depletion Potential (ODP) CFC-11-eq.
WMO model,
ReCiPe
Human Toxicity Cancer potential
effects (HumToxCan)
CTUh USEtox
Human Toxicity non-Cancer
potential effects (HumTox)
CTUh USEtox
Particulate Matter (2.5 µm) (PM)
G
RiskPoll
Photo-Oxidant Creation Potential
(POCP) NMVOC-eq.
LOTOS-EUROS
model, ReCiPe
Acidification Potential (AP) Mole of H
+
-
eq.
Accumulated
exceedance model
Eutrophication Potential (fresh
water) (EP fresh)
Mole of N-eq.
EUTREND model,
ReCiPe
Eutrophication Potential (terrestrial)
(EP terr)
g P-eq.
Accumulated
exceedance model
Eco-system Toxicity potential effects
(EcoTox) CTUe USEtox
Freshwater consumption (Water) m
3
Swiss Ecoscarcity
Abiotic Depletion Potential (ADP) Sb-eq.
CML (reserve
based)
IV. LIFE CYCLE INVENTORY
A. Emission factors
Generic GaBi models have been used for the energy and
fuel models. The emission factors include the supply chain
for the energy and fuel production, as it may have significant
environmental impacts on the total results.
For primary data, the production related electricity mixes
are based on the locations of the suppliers. For secondary
GaBi data, electricity mixes are based on locations embodied
125
in the data. For ICs and ASICs specifically, supplier
information was used corresponding to an emission factor
close to world average (around 0,6 kg/kWh).
For use stage, a world average emission factor was
applied as the assessed products are intended for a global
market.
As reflected in Section V and VI energy models were
first modelled based on Ecoinvent data, then also with GaBi
data. The Ecoinvent energy models include building
construction and materials (including metals) within their
system boundaries.
B. Raw Materials Acquisition
1) Smartphone raw materials
Primary materials data were provided by Sony per part
presented in Fig. 1, and for selected components, such as
printed board assemblies (PBAs). The component level data
were used to model other similar components. For example,
materials for all Application Specific Integrated Circuits
(ASICs) were based on materials for one ASICs scaled up to
reflect the overall ASIC weight. The acquisition processes
for these raw materials were modeled in GaBi based on
secondary data. Transportations related to the raw material
acquisition stage were included in these data to an unknown
extend and it was not possible to extract transportation
details from the available data set. Electricity mixes are
based on location and embodied in the GaBi data.
2) Packaging materials
Amounts of packaging materials for parts and final
delivery have been estimated based on Ericsson conditions
and factors have been applied that represent the packaging
material weight in relation to the part or device weight. Some
parts such as the IC and PBA require more packaging as they
are more fragile and hence have a higher packaging factor.
Packaging materials include steel, polyurethane foam,
polyurethane wood, plywood and cardboard and their
acquisition processes are based on secondary GaBi data.
3) Virgin and recycled materials
The raw material stage takes into consideration the virgin
and recycled inputs for some selected materials; copper,
gold, silver and aluminium based on world market
conditions. The virgin material input varies between 30 to 80
percent based on global recycling rates [12].
ETSI/ITU recommends a 50/50 approach to be used. This
approach seeks to distribute the impacts from primary
material production and waste treatment to the first and last
life cycle in equal amounts, however without considering
material loss at design or end-of-life treatment. Due to lack
of details in the used data base, the 50/50 approach could not
be applied consistently. For this reason this study developed
an own approach by applying the 50/50 approach for two
different scenarios for smartphone recycling (19% or 83%
depending on modelling of informal recycling) based on [13]
and a world average rate of recycled gold (28% based on
industry data from CPM group) For other metals, the GaBi
models did not distinguish between virgin and recycled
materials. For these metals the GaBi models were applied
directly representing an unknown mix of virgin and recycled
materials.
Li-battery recycling (e.g. cobalt) has not been included
but the impact will be minor and mainly add to ADP.
C. Production
1) Parts production
Production process data are based on primary data collected
from Sony’s suppliers through a questionnaire where input
was provided as annual figures for 2014 representing energy
consumption, generated waste, ancillary products, emissions
to soil, air and water and production related transportations.
With the exception of ICs and ASICs, supplier support
activities are excluded. Data were obtained from direct
suppliers to Sony but not from the sub-suppliers. Table II
below shows primary and secondary data sources for parts
production. Where primary input data were insufficient or
unavailable, data from previous studies or secondary data
were used.
TABLE II: PARTS PRODUCTION DATA TYPES FOR THE SMARTPHONE
Primary supplier data Secondary data
Headset
USB
Packaging box
Sony assembly
Key Panel
Touch and Display
Microphone
RF Switch
Vibrator
Camera
Battery
FCB
Speaker
Antenna
PCB
RF Switch
ASICS and IC (partly)
ASICS and IC (ICT
specific, see section C 2)
Connectors (GaBi)
Inductor Chip (GaBi)
Resistors (GaBi)
Capacitors (GaBi)
The collected facility data were allocated based on
Sony´s share of overall production and surface area (ASICS,
IC and PBA) or weight (all others).
Due to confidentiality, production related input data
including location of facilities have restricted availability but
the resulting potential environmental impacts are presented
in section V.
2) Integrated circuit (IC) production
The production of ICs is known as a resource intensive
production process with substantial energy and resource use
with among the highest environmental impacts per mass unit
that exist today for mass produced products. The model used
in this study considers yield and covers all main production
processes including production of silicon wafer, chip on
wafer (“wafer-fab”) and the IC packaging (encapsulation).
Also of main importance and included for the wafer fab, and
to some extent also for other processes, are production of all
special gases, chemicals and materials; emissions of gases
with high GWP; supporting activities; and; the building of
the factory itself and the production of process equipment.
For IC and ASICs the quality of the primary data
collected from suppliers were found insufficient. Instead
secondary GaBi data were used. However, the data for CO2
and some other critical GWP gases in the GaBi dataset were
considered too low when compared to supplier data from
126
earlier Ericsson LCAs including information from major
suppliers like Intel, Texas Instruments, LSI, TSMC (used
also by Texas and LSI) and former ST-Ericsson. Based on
their input the total GHG emissions for production of ICs
including life cycle impact of ancillary materials for
smartphones are estimated to be in the range of 2.7 - 4.3 kg
CO2e per cm2 of good die (total GHG emissions are only
split on fully functional chip area, i.e. yield is considered).
The data show that the electricity consumption and
emission of gases with high GWP in the wafer factory are
two of the most important individual contributors,
corresponding to an average electricity consumption for
processors and ASIC’s of about 3 kWh/cm2 and about 2
kWh/cm2 for memories (yields included, i.e. allocated to
fully functional chip area).
The average GWP for overall production impacts of all
ICs for the assessed smartphone is about 3.5 kg CO2e per
cm2 - somewhat higher for the processors and ASICs (about
4 kg) and somewhat lower for the memories (about 3 kg) as
they consume less electricity per cm2 and have higher
production yields than processors and ASICs.
To put these figures into perspective, early, unpublished
Ericsson studies in the mid-90s showed wafer factory energy
consumption for different IC types 3-10 kWh/cm2 with yield
included (or 2-4 kWh/cm2 chip area without considering
yield). Mobile phone ASICs were found in the upper part of
this range. The electricity consumption of high end IC
components today is comparable to that of low end
components in 1995, and emissions of high-emitting GWP
gases have been reduced from around 1-1.5 kg CO2e/cm2
down to about 0.5 kg CO2e/cm2.
3) Display production
Another process that gained special attention during the
study was the display production for which the authors had
limited data regarding impact levels. Further, the display
production could be expected to be an important contributor
to overall impacts as it is known to require a very clean
environment and substantial inputs of water, gases and
chemicals.
According to the supplier, the electricity consumption in
the display manufacturing is about 0.1 kWh/cm2 (including
the touch layer). As for ICs the data shows that the impact
from input of water, gases and chemicals and from
supporting activities is large compared to other components.
The supplier data included information regarding the
production process, as well as LCA results for another
display type. Of these, the LCA results could not be used for
materials to avoid double counting with the materials data
acquired through materials declarations. Furthermore, the
material content represented a display intended for a TV with
different characteristics. However the LCA data was used to
modify the factory energy data for GWP to include support
activities. This could not be done for other impact categories.
Combining the factory data and the LCA leads to higher
uncertainty for the display than for other production
processes, especially since no prior data was available for
validation.
4) Part Transportation
Transportation types, weights and distances are obtained
from the suppliers through the questionnaire. This includes
inbound (both production and ancillary materials) and
outbound (parts and waste) transports. The primary data is
combined with models for the different means of
transportation that are based on secondary data.
Packaging weights have been included in the
transportation models.
5) Assembly
Final assembly of the smartphone is performed by Sony.
Data for the assembly processes are primary data collected
from a Sony assembly site and cover energy consumption,
generated waste, ancillary products, emissions to soil, air and
water and production related transportations.
6) ICT manufacturer support activities
ICT manufacturer support activities were estimated based
on primary data regarding energy consumption for main
offices and business travelling (not hotels) which were
allocated to the device based on sales volumes. It was
assumed that the need for supporting activities was the same
for all products.
7) Distribution
Transportation types, weights and distances for
distribution to retailers were built on internal Sony data.
Based on the main transportation routes to all continents, a
global average scenario was developed for air and road
transportation of the final product. The primary distribution
data were combined with secondary data for the models for
the different means of transportation.
Packaging weight has been included in the transportation
models but user travelling was not included.
D. Use
The use stage consists of usage of the device and
associated usage of the mobile network infrastructure. A
global energy mix has been adopted for the usage stage.
1) Smartphone use
The smartphone energy consumption is based on a
reasonable usage scenario (described below) which is
assumed to be the representative case, based on Sony data
for charging time and energy consumption during charging
and for chargers in stand-by.
Two other scenarios, assumed to represent heavy and
light users, were developed for the sensitivity analysis. The
user scenarios differ with respect to the battery charging
cycles and lifetime. A representative user is considered to
charge the device once every two days whereas the heavy
and light users charge every day and every 3rd day
respectively, as shown in Table III. The phone energy
consumption includes the daily power consumption of the
device based on chargers and power consumption of battery
while charging and in standby-mode power consumption. A
charging time applicable to Z5 and Z3 conditions is
embodied in the figures. It was also assumed that the
representative user would change phone every 3 years, while
127
the corresponding operating life time for the light and heavy
users were set to 4 and 2 years respectively.
TABLE III. SMARTPHONE DEVICE ENERGY CONSUMPTION
Heavy
User
Representative
User
Light
User
Maximum charge cycle (days)
1
2
3
Phone Energy Consumption
(kWh/year)*
7.74
3.87
2.58
Usage (years)*
2
3
4
Source: *Sony
2) Associated use of networks
For GWP, the user´s use of the mobile networks was
considered. For this impact category, the studied system
hence included the wireless network which is needed in
order to have a working mobile service. For the network
related impact, data was adapted from [14] which give more
detailed information on network data, assumptions and
assessment method. Network related impacts included
impacts from mobile networks (access and core) and data
networks (data centers, data transmissions and IP core
networks).
Table IV presents the mobile and wi-fi network usage for
the different usage scenarios. Impacts considered include life
cycle impact of the listed network products allocated to the
smartphone based on [14]. More specifically, wi-fi and the
traffic dependent network energy consumption were
allocated based on traffic, while the static energy
consumption of mobile networks was allocated based on
number of subscribers. Also operator’s supporting activities
were included and allocated per subscriber.
TABLE IV. NETWORK USAGE FOR THE THREE USAGE SCENARIOS
Heavy User
Representative
User Light User
Mobile + wi-fi data
traffic
30GB** +
30GB*
11GB ** +
11GB*
5.5GB** +
5.5GB*
Use of mobile networks per year*
RBS 3G Operation
(kWh/subyear)
25 22 21
RBS Embodied
(kg CO2e/subyear)
3.5 3.5 3.5
Operator,
transmission, data
centers
(kWh/subyear)
15 8 6
Operation services
(travel)
(kg CO2e/subyear) 2.6 2.6 2.6
Other embodied
(kg CO2e/subyear)
0.7 0.7 0.7
Use of wi-fi per year**
Wi-fi Operation
(kWh/subyear) 9 3.3 1.65
Wi-fi embodied
(kg CO2e/subyear) 0.9 0.3 0.15
Source:* [14], **([15], [16])
The network data usage is representative for high end
smartphones for Swedish mobile networks and data traffic
in 2015. As Sweden, together with Finland, [17] has the
highest data usage intensity in the world the data volumes
may be too high in a global scenario, however, it provides a
reference for the current state-of-the-art networks.
The Swedish GWP figures were adjusted to reflect a global
electricity mix.
E. End-of-Life Treatment
End-of-life-treatment (EoLT) is modelled based on [13]
which investigates regional and global waste flows in order
to set up EoLT scenarios for ICT equipment. While noting
the non-consistent data types and the substantial differences
in quality between waste flow data for different countries,
Liebmann [13] attempts to model regional waste flows. For
the regional grouping, mobile phone subscription counts are
used to weight each country´s estimated percentage of
generated waste. The model considers the percentage of
waste generated within the region as well as import and
export of waste. However, in some mainly importing regions
no data were found for export flows.
Regional findings are aggregated to estimate a global end-of-
life treatment scenario for ICT equipment, also based on the
number of mobile phone subscriptions.
Informal recycling is found to be a major EoLT activity. Due
to unavailable data and obvious modelling difficulties for
informal recycling, this fraction was instead modelled as
formal recycling resulting in an overall recycling rate of 83%
vs 17% landfill based on weight. This approach should be
seen as a best case, or a non-conservative scenario. A more
conservative, yet not the worst-case, approach would be to
model the informal recycling as landfill, resulting in a
recycling rate of 19% vs 81% landfill.
The recycling process data is based on Boliden and
Kuusakoski recycling sites and landfill processes are based
on generic data from GaBi. Due to lack of specific data,
transportations related to recycling are modelled based on
Ericsson internal conditions for recycling.
V. RESULTS
A. Total Global Warming Potential of the device
Figure 2 shows the GWP of a Z5 smartphone device
including accessories but excluding network usage, over its
life time (3 years) for the representative scenario specified in
Table III. The GWP results for the life cycle impact of
smartphone model Z5 is 57 kg CO2e. For Z3 the GWP
corresponds to 50 kg CO2e. The corresponding annual
impact is 19 and 17 kg respectively. Including also the
network usage according to the representative scenario, the
total impact figure increases by 43 kg CO2e per year, or by a
total of 129 kg CO2e per year.
128
-10 010 20 30 40 50 60
EoLT
Use
Production
Raw materials
ICs Others
kg CO
2
e/3 years
See Fig. 3
1.9
-0.3
7.2
47.9
Fig.2 GWP for smartphone Z5 during its life time (3 years), including
accessories but excluding network usage
Figure 2 shows that production, about 48 kg CO2e in total for
Z5 and dominated by IC production, contributes most to
GWP for the Z5 device. The use stage gives about 7 kg
CO2e in total for the Z5 and Z3 models (corresponding to
about 4 kWh/year) over three years, assuming a global
average electricity mix (0.6 kg CO2/kWh).
B. GWP distribution within the production stage
The total GHG emissions from IC production are about
33 kg CO2e for the Z5 model and about 26 kg for Z3
reflecting the difference in IC chip area (9,5 vs7,5 cm2).
For other production activities the GWP contributions are
distributed according to figure 4 which shows that the second
most important contributor is the display with a contribution
of about 3.5 kg CO2e, as shown in Fig. 3.
00.5 11.5 22.5 33.5 4
Final transport
V. activities*
Other parts
Display
PCB
Battery
kg CO
2
e/3 years
*vendor activities include final assembly and vender supporting activities
1.4
2.1
3.5
2.7
2.4
3.0
Fig. 3 Total GWP results for all production processes but IC for Z5
C. Transport related GWP impacts
Final transportation contributes to the total results (3
years) with 3 kg of CO2e. Additionally, embedded in the
production stage, production related transportations are
included and account for 2.6 kg CO2e. The raw material
acquisition stage and waste and EoLT related transportations
are accounted for in the GaBi models but are not possible to
separate from the models. These transportation impacts are
assumed to be very small, below 0.1 kg CO2e. Thus total
transportation is estimated to be 5.6 kg CO2e, i.e. around
10% of total impacts.
D. Other impact categories
The results for other impact categories are presented in
Figure 4 and 5 for the Z5 smartphone including accessories
but excluding network usage, over its life time (3 years) for
the representative scenario specified in Table III.
Figure 4 shows the contribution from the raw material
and EoLT, production and use stage for all impact
categories based on eco-invent data for gold and energy
production, while figure 5 is based on GaBi data for gold
and energy production.
0% 20% 40% 60% 80% 100%
GWP
ODP
HumToxCan
HumTox
PM
POCP
AP
EP fresh
EP terr
EcoTox
Water
ADP
Gold
Gold
Gold
Gold Copper
Cu
Cu
Gold Cobalt Silver Li
Other
Copper
Raw materials Production Use (3 years)
57
kg
CO2-e
0.2
mg
CFC-11-e
1*10
-
06
CTUh
3*10
-
05
CTUh
0.06
kg
0.3
kg
NMVOC-e
0.5
Mole of H+
-e
1
Mole of
N-e
0.01
g
P-e
500
CTUe
50
m
3
0.002
kg
Sb-e
Fig. 4 Total life cycle result for all impact categories for smartphone Z5
with accessories using Ecoinvent database and adopting a 50/50 recycling
approach with 19% recycling of gold assumed.
0% 20% 40% 60% 80% 100%
GWP
ODP
HumToxCan
HumTox
PM
POCP
AP
EP fresh
EP terr
EcoTox
Water
ADP
Raw materials Production Use (3 years)
Other Gold Cobalt Silver Li
55
kg
CO
2
-e
0.1
mg
CFC-11-e
1*10
-07
CTUh
5*10
-06
CTUh
0.06
kg
0.3
kg
NMVOC-e
0.3
Mole of H+
-e
1.2
Mole of
N-e
2*10
-03
g
P-e
62
CTUe
3
m
3
2*10
-03
kg
Sb-e
Fig. 5 Total life cycle result for all impact categories for smartphone Z5
with accessories using GaBi database for gold and energy production and a
50/50 recycling approach with 83% recycling of gold assumed. Note that
the figure shows relative results compared to figure 4
In figure 5, Ecoinvent gold and copper data and models
are replaced by GaBi´s own data models, and the results are
expressed in percentage of the Ecoinvent based results
indicating a large difference in results due to the two data
sets. Neither of the two scenarios presented in Fig. 4 and 5
could be described as the true one, rather they represent a
129
range of possible outcomes. See section V for further details.
Also the recycling potential differs between Figure 4 and 5.
However this has only a minor impact on the result.
Both figure 4 and 5 indicate that the use stage is
relatively small for all impact categories, except for Ozone
depletion in the Ecoinvent case. For both data sets the
production stage dominates the impacts for GWP, Particulate
Matter, Photo-Oxidant Creation Potential, Acidification
Potential and Eutrophication of fresh water. For the
Ecoinvent data set also water is dominated by the production
stage. The remaining categories are dominated by the raw
materials acquisition.
V. INTERPRETATION AND SENSITIVITY ANALYSIS
The detailed results show that production and use impact
comes to a high degree from electricity consumption, and
raw material toxicity impacts are dominated by gold and
copper mining. To understand the impact of parameter
settings sensitivity analysis was made for life time, data
traffic and gold production data. Due to time restrictions the
sensitivity analysis has so far only been performed for these
parameters.
A. Impacts from gold modelling
As identified by Moberg et al. [7], gold belongs to the
prioritized data collection processes. In the present study,
when applying the Ecoinvent data set, gold is contributing to
nearly half of the abiotic resource depletion potential (also
cobalt, silver and lithium give significant contributions).
These results depend on the amount of gold that is needed to
produce a smartphone, the rate of recycled gold that enters
the smartphone life cycle, how the EoLT stage is modeled,
and the gold recycling rate at EoL.
As emphasized in Figure 4 the toxic impact potential is
dominated by the acquisition of gold, followed by the copper
processes. This is due to the time boundaries applied for the
Ecoinvent data set which covers an extremely long time
frame compared to other data sets for mining tailings
emissions (in the order of 10 000 years), resulting in high
toxicity impacts, especially related to gold and copper
mining. Furthermore, the Ecoinvent models assumes leakage
of metals based on the conditions of one mine in South
America for which mining tailings and dams are assumed to
constantly leak or even break. The Ecoinvent data model can
thus be seen as a worst case scenario. The extensive time
frame may be reasonable, but as other data sets take another
perspective the time frame should be remembered when
comparing studies.
In figure 5, Ecoinvent gold and copper data and models
are replaced by GaBi´s own data models, representing a
modern Northern Europe mine and smelter.
Besides the contribution from Gold, Abiotic Resource
Depletion Potential also gets significant contributions from
Silver, Cobalt and Lithium. However, if the Ecoinvent model
is replaced by the GaBi model representing a modern
Northern Europe mine and smelter, the potential toxic
impacts data reduces down do just a few percentage of the
impacts found with the Ecoinvent - 1% for HumToxCan, 4%
for HumTox and 4% for EcoTox.
B. Impact from the usage scenario
The main usage scenario of this study is the
representative scenario described in Table III and Table IV.
To check the importance of this scenario, the light and heavy
usage scenarios were established.The three scenarios differ
with respect to lifetime (2-4 years) and data traffic (from 5,5
GB mobile+5,5 GB wi-fi to 30 GB mobile+30 GB wi-fi).
The smartphone network usage, based on the three cases,
varies between 36, 43 and 67 kg CO2e per year - to be
compared to the yearly device impacts of 19 kg CO2e
making it clear that inclusion of the network significantly
affect the results and thus require caution when comparing
and communicating studies.
020 40 60 80
High use
Representative
use
Low use
Allocated Network GWP
kg CO2-e/ subyear
Mobile network embodied*
WiFi network embodied*
Mobile network use
WiFi network use
* Operator activities are included in network production
020 40 60 80
High use
Representative
use
Low us e
Smartphone device GWP
kg CO2-e/ subyear
Smartphone embodied
Smartphone use
2 years
life time
3 years
life time
4 years
life time
Figure 6 Sensitivity analysis of life time and data traffic
Figure 6 varies lifetime and data traffic simultanously
which makes it difficult to tell their influence apart. When
varying these parameters separately it turns out that the
decvice usage impact is sensitive to life time, while the
network usage varies with data usage.
C. Normalization
To put the results in perspective normalization was
performed. The normalization represents the yearly Z5
smartphone life cycle impact (excluding the network usage)
130
compared to the overall impact per person and year globally
according to reference values from GaBi, generally based on
the LCIA source data The normalization was performed for
the two data sets (Ecoinvent and GaBi) and recycling
conditions according to Figure 4 and 5, and Figure 7
demonstrates the substantial difference between the two.
Long-term toxicity effects (like those considered in the GaBi
data) are not included in the reference values making the
relative Ecoinvent percentages too high for toxicity.
0% 10%
0.27%
0.00038%
0.92%
1.6%
0.44%
0.33%
0.33%
0.24%
0.30%
1.7%
0.02%
0.66%
EcoInvent Normalized
results
GWP
ODP
HumToxCan
HumTox
PM
POCP
AP
EP fresh
EP
terr
EcoTox
Water
ADP
0% 10%
0.26%
0.0002%
0.12%
0.37%
0.39%
0.31%
0.23%
0.23%
0.05%
0.26%
0.001%
0.53%
GaBi Normalized
results
Figure 8 Yearly Z5 smartphone device life cycle impacts normalized based
on overall yearly impact per person per category based on the representative
scenario
VI. DISCUSSION
A. General
As reflected in the ETSI/ITU standard, the life cycle
assessment of an ICT product, in this case the smartphone is
a very complex task. As the device has multiple parts, each
with several suppliers dynamically shifting over time, the
data collection is demanding and time intense and it is not
possible to collect data from all suppliers. The geographical
dispersion of the production sites, the differences in usage of
the devices and networks operation as well as the production
and maintenance of the network infrastructure and
equipment is complex in terms of both scope and allocation.
The continuous development of technology, resulting in new
products, limits the possibility to perform complete, up to
date LCAs on all ICT equipment but makes it important to
build understanding based on representative products to
follow the development as technology advances.
B. Standard compliance
This study has had the ambition to comply with the joint
ITU/ETSI standard ([8]-[9]). Compliance was achieved with
the following exceptions:
The materials model was limited to the about 30 most
impacting materials.
The 50/50 approach for allocation between life cycles
could only be applied to gold.
Transport data for raw materials, waste and EoLT could
not be presented separately.
Reporting formats were not followed due to limitations
in number of pages.
For the same reason, a full data quality and uncertainty
description is missing, and the detailed calculations are
not described.
Results are not presented for electricity, primary energy
and fuel usage, only for impacts.
Embodied impacts due to software (including apps)
developed by other parties than Sony was not
considered. However, use of software is included in the
overall usage scenarios.
The sensitivity analysis was limited to few parameters.
C. Comparison of GWP between studiesthen and now
Already in 1995, Ericsson (through the second author of
this paper) performed an (unpublished) LCA of a mobile
phone. Also that study identified ICs production and gold
acquisition as main contributors. In that study the gold
content was about 50 mg, compared to 20 mg for the Sony
Z5 targeted by this study. In contrast the chiparea was only
40% of the Z5 chip area. However, the energy consumption
was higher and production yields lower in 1995, making ICs
the most contributing components. The production of the
Nickel-Cadmium batteries used in 1995 also had a higher
environmental impact, including GWP, than today’s Lithium
batteries.
The most important LCA results from the early LCA
studies of mobile phones were probably the insight that
chargers, usually not unplugged after use, contributed
substantially to the impact. This contribution increased as
users could have several chargers plugged-in in parallell.
This insight made the leading brands (Ericsson, NOKIA and
Motorola) develop chargers that switched themselves off,
reducing stand-by power from several Watts to below 0.5
Watt. An internal Ericsson LCA study in 2008 confirmed
this trend and showed a stand-by power below 0.1 Watt and
a charger left in stand-by all of the time resulted in less than
1 kWh/year.It is estimated that the total emission savings for
stand-by power reductions for mobile chargers was in the
range of 100 million tonnes CO2e.
The total GWP results for the production and use stages
from this study are compared to the outcome of other recent
and older LCA studies with similar scope and system
boundaries, see Fig 8.
Of recent studies, the generic smartphone study by
Fairphone/Fraunhofer [18] and Apple’s iPhone studies [19]
reviewed and approved by Faunhofer were addressed in Fig.
8. For GWP, Fraunhofer shows about the same relative
importance of the IC production and also get a similar result
for the embodied emissions, about 50 kg CO2e. For the use
stage, however, they get a higher footprint, corresponding to
131
about 15 kg CO2e over 3 years compared to 7 kg (for Z5) in
our study for comparable emisssion factors.
The Fraunhofer and Apple study assumes much higher
memory capacity which is likely to cause most of the higher
footprints in those studies.
050 100 150
Apple iPhone 6+
Apple iPhone 5
Generic smartphone
SONY Z5
SONY Z3
SONY Xperia T
Sony Ericsson 2008
Ericsson 2001
Ericsson 1998
Ericsson 1995
kg CO
2
e
128 GB
Charger stand-by 50% / 100% of time
64 GB
32 GB
Production
Use (3 years)
16 GB
16 GB
? GB
Fig 8 GWP results for different mobile LCA studies. Note that the figure
cannot be used to make conclusion regarding the relative impacts of
different smartphone models, just to compare study results.
D. Use of LCA results for eco-rating
Sustainability and environmental concerns are
increasingly gaining a global significance. To move in a
sustainable direction, human behavior needs to change as
well [20]. For this reason, many industry sectors have
developed ways to provide product related sustainability
information to customers, with the aim to positively
influence their habits and purchasing decisions. This
development is also visible in the mobile industry where
multiple eco-rating schemes are established to communicate
product related information to customers, sometimes
considering also embodied impacts.
Direct comparison of different LCA results can be
misleading if differences in methodology, assumptions,
system boundaries etc. are not considered. Additionally,
mobile devices vary in performance and features and it may
be challenging to find a good way to balance their footprints
towards their indirect effects. On a more positive note there
seem to be an opportunity in using LCA results to identify
hotspots to address and could also be used as the foundation
for improving supply chain performance and company
performance in order to make them more sustainable. As an
example LCAs could identify materials and processes with
high impacts. LCAs could also be used to show how
customer behavior affects the impacts. See figure6.
E. The smartphone footprint vs its indirect effect
Beyond the basic communication service for speech and
data, smartphones also work as music and media players,
radios, TVs, GPSs, cameras, video cameras, game consoles,
alarm clocks, etc. Consequently, it seems reasonable when
considering smartphone footprint to also consider how these
features changes usage patterns for other consumer
electronics. Sales figures for 2010, 2014 and 2015 for some
of these devices are declining globally. Although the growth
can still be high in developing countries, see Table V.
TABLE V GLOBAL SALES OF SOME CONSUMER DEVICES
Global shipments [millions]
2010
2014
2015
Game consoles
78
47
na
Portable media players
120
50
na
Digital cameras
120
40
na
TVs
250
220
na
PCs
350
310
280
Tablets
20
230
330
Smartphones
305
1300
1400
Sources: CEA, IDC, Gartner (na: not available)
Another interesting point is the total energy consumption
of all consumer electronics in use in the US, which is now
decreasing with a reduction of about 12% between 2010 and
2013. The main reason for the reduction is less PCs and TVs
in operation, and lower usage of these as media is accessed
via smartphones and tablets. This is a trend break according
to the US study which was conducted also in 2001 and 2006,
see [21] for totals for all years.
The above statistics imply that it is reasonable to put the
smartphone footprint into the perspective of the savings
induced by their additional features. Additionally, not dealt
with here, smartphones enable the use of ICT throughout
society which could have the potential to substantially
reduce overall societal GHG emissions as discussed in [22].
VII. CONCLUSIONS
The cradle-to-grave Global Warming Potential of the Z5
smartphone was estimated to 57 kg CO2e, including
accessories but excluding network usage, representing its
life time (3 years) impact for a Representative usage
scenario. The corresponding annual impact was 19 kg
CO2e without network, and 62 kg CO2e with network
usage included
The major part of the GHG emissions, 48 kg CO2e, was
related to the production processes where the IC
production dominated. The use stage resulted in 7 kg
CO2e for the smartphone itself (battery charging with a
global average electricity mix).
The use stage was relatively small for all impact
categories. The production stage dominates the impacts
for GWP, Particulate Matter, Photo-Oxidant Creation
Potential, Acidification Potential and Fresh water
Eutrophication. The remaining categories are dominated
by the raw materials acquisition
Most GWP but also most particulate matter, photo-
oxidant creation and acidification potentials are related to
the fossil fuel share of the electricity consumption.
Gold and copper are the main contributors to the toxic
impact categories and resource depletion (together with
132
battery metals) but their contribution is highly dependent
on the data source. The study uses two scenarios to show
the large variation in data for mining of rare metals.
For the use stage, the device usage impact is sensitive to
life time, while the network usage varies with data usage.
EoLT models have high uncertainty due to a profound
lack of data.
VIII. ACKNOWLEDGMENTS
The authors would like to thank the Center for
Sustainable Communications (CESC) at KTH Royal Institute
of Technology in Sweden, for continuous discussions and
knowledge sharing on ICT for sustainability and LCA.
IX. REFERENCES
[1] The future carbon footprint of ICT and E&M sectors, 2013,
ICT4S Proceedings DOI Available online: http:-
//dx.doi.org/10.3929/ethz-a-007337628 (accessed on 15
April 2016)
[2] Europe 2020, 20?? A resource-efficient Europe Flagship
initiative of the Europe 2020 Strategy Available online:
http://ec.europa.eu/resource-efficient-europe/ (accessed on
15 April 2016)
[3] ITU, 2012. Review of mobile handset eco-rating schemes.
Available online: https://www.itu.int/dms_pub/itu-
t/oth/4B/01/T4B010000030001PDFE.pdf (accessed on 15
April 2016)
[4] ITU 2016. Ongoing work at ITU: Question 16/5
Leveraging and enhancing the ICT environmental
sustainability. Available online: http://www.itu.int/en/ITU-
T/studygroups/2013-2016/05/Pages/q16.aspx (accessed on
15 April 2016)
[5] Suckling, J., Lee,J., 2015. Redefining scope: the true
environmental impact of smartphones, Int J Life Cycle
Assessments 20:1181-1196 DOI 10.1007/s11367-015-
0909-
[6] Ercan, M. 2013 Global warming potential of a Smartphone;
using life cycle assessment methodology. Available online:
http://kth.diva-
portal.org/smash/get/diva2:677729/FULLTEXT01.pdf
(accessed on 5 February 2016)
[7] Moberg, Å et.al. 2014. Simplifying a life cycle assessment
of a mobile phone. International Journal of Life Cycle
Assessment. volume 19, issue 5, pp. 979 993
[8] European Telecommunications Standards Institute. ETSI
ES 203 199 V1.3.1 (2015-02): Environmental Engineering
(EE); Methodology for environmental Life Cycle
Assessment (LCA) of Information and Communication
Technology (ICT) goods, networks and services Available
online:
http://www.etsi.org/deliver/etsi_es/203100_203199/203199
/01.03.01_60/es_203199v010301p.pdf (accessed on 2
February 2016).
[9] International Telecommunication Union. ITU L.1410
(2014-12): Methodology for environmental Life Cycle
Assessment (LCA) of Information and Communication
Technology (ICT) goods, networks and services Available
online: https://www.itu.int/rec/T-REC-L.1410-201412-I/en
(accessed on 2 February 2016).
[10] GaBi LCA software. [computer program] Available online:
http://www.gabi-software.com/sweden/index/ (accessed on
2 February 2016).
[11] European Platform on Life Cycle Assessment
http://eplca.jrc.ec.europa.eu (accessed on 1 april 2016)
[12] UNEP 2011, Recycling rates of metals- A status report.
Available online:
http://www.unep.org/resourcepanel/Portals/24102/PDFs/M
etals_Recycling_Rates_110412-1.pdf (accessed on 2
February 2016).
[13] Liebmann, A., 2015 ICT Waste Handling: Regional and
Global End-of-Life Treatment Scenarios for ICT
Equipment Royal Institute of Technology
[14] Malmodin, J., Lundén, D., Nilsson, M., Andersson. 2012.
LCA of data transmission and IP core networks. Available
online:http://www.ericsson.com/res/docs/2013/lca-of-data-
transmission-and-ip-core-networks.pdf (accessed on 15
April 2016)
[15] Cisco Visual Networking Index: Global Mobile Data
Traffic Forecast Update, 20152020 White Paper Available
online:
http://www.cisco.com/c/en/us/solutions/collateral/service-
provider/visual-networking-index-vni/mobile-white-paper-
c11-520862.html (accessed on 2 February 2016).
[16] Ericsson mobility report, 2015. On the pulse of the
networked society Available online:
http://www.ericsson.com/res/docs/2015/mobility-
report/ericsson-mobility-report-nov-2015.pdf (accessed on
15 April 2016)
[17] Tefficient AB, 2015. Mobile data usage- Price and bucket
size matter Available online:
http://media.tefficient.com/2015/06/tefficient-industry-
analysis-2-2015-mobile-data-usage-and-pricing-26-
June.pdf (accessed on 15 April 2016)
[18] Fairphone (Faunhofer), 2013. Available online:
https://www.fairphone.com/2013/08/01/whats-in-a-life-
cycle-assessment/ (accessed on 15 April 2016)
[19] Apple Inc. 2015. Environmental responsibility- Product
reports Available online:
http://www.apple.com/environment/reports/ (accessed on 4
April 2016).
[20] Urban, B. et. al. 2014. Energy consumption of consumer
electronics in U.S. homes in 2013. Final report to the
consumer electronics association (CEA). Fraunhofer USA
center for sustainable energy systems. June 2014.
[21] Electronics Association. 2016. Rateko. Consumer
Electronics Technical and Commercial Organizations.
Journal with 6 issues every year including sale statistics. In
cooperation with the Mobile Phones Association. Available
online: www.rateko.se (accessed on January 2016)
[22] The future carbon footprint of ICT and E&M sectors, 2013,
ICT4S Proceedings DOI Available online: http:-
//dx.doi.org/10.3929/ethz-a-007337628 (accessed on 15
April 2016)
133
... Since many years, researchers assess and discuss the direct environmental effects of ICT products and services [9][10][11][12][13][14][15][16] with some of them focusing on mobile networks specifically [17][18][19][20][21][22]. Scharnhorst et al. [22] compared the environmental footprint of 2G and 3G mobile networks and found that 2G networks (GSM) perform better than 3G networks (UMTS) for the impact category of climate change. ...
Chapter
5G mobile networks are intended to meet the increasing requirements placed on mobile communications. Producing and operating 5G infrastructure causes direct effects on greenhouse gas (GHG) emissions. Meanwhile, 5G is expected to support applications that contribute to GHG abatement. We investigated (i) the GHG footprint of 5G infrastructure, and (ii) the GHG abatement potential of four 5G-supported use cases (i.e., flexible work, smart grids, automated driving and precision farming) for Switzerland in 2030. Our results show that 5G infrastructure is expected to cause 0.018 Mt CO2 e/year. Per unit of data transmitted, 5G is expected to cause 85% less GHG emissions in 2030 than today’s 2G/3G/4G network mix. The four 5G-supported use cases have the potential to avoid up to 2.1 Mt CO2 e/year; clearly more than the predicted GHG footprint of 5G infrastructure. The use cases benefit especially from ultra-low latency, the possibility to connect many devices, high reliability, mobility, availability and security provided by 5G. To put 5G at the service of climate protection, measures should be taken in two fields. First, the GHG footprint of 5G should be kept small, by installing only as much 5G infrastructure as required, running 5G with electricity from renewable energy sources, and decommissioning older network technologies once 5G is widely available. Second, the GHG abatements enabled by 5G-supported use cases should be unleashed by creating conditions that target GHG reductions and mitigate rebound effects. The final outcome depends largely on the political will to steer the development into the direction of a net GHG reduction.
... Malmodin et al. (2010) estimated that ICT and entertainment & media contributed ∼10% of global energy consumption and ∼2% of global carbon emissions. Ercan et al. (2016) conducted an LCA assessment of digital products such as personal computers, smartphones, communication networks, and data centers, and estimated they together accounted for 4-10% of the world's electricity consumption. Belkhir and Elmeligi (2018) assessed the ICT global carbon footprint using an LCA framework and found that the ICT relative emissions would exceed 14% of 2016-level global emissions by 2040. ...
Article
Digital economy (DE) has become one of the most noticeable phenomena proceeding in economic, societal, and environmental systems. However, no consensus for a modeling framework of its widespread impacts has emerged. In particular, no explicit model that quantifies or analyzes the carbon impacts of DE exists. Here, we develop a phased, quantitative input-output framework that evaluates to what extent DE affects carbon emissions. Several mechanisms for the carbon effects can be analyzed using this framework, e.g., the emissions embodied in digital demand and enabled by digital supply through industrial backward and forward linkages. The China case shows that when the combined effects of digital demand and supply were considered, DE could lead to ∼6% of the country's total emissions. Driven by expanding digital demand and supply scales, digital emissions soared from 210 to 418 Mt CO2 between 2002 and 2007. However, the growth had largely slowed down in 2007-2017 thanks to improvements in sectoral efficiency and digital production and application structure. Quantitative information gained from the framework provides additional insights for policymakers to formulate economy-wide measures, e.g., modulating the expansion of digital production and sale chain, to curb the emissions from the boosting DE.
... For example, tiny amounts of gold (around 30 mg) carry the largest contribution of the CO2-equivalents of RAM bars in laptops and these in turn account for about 40% of the total GWP of the laptop [12]. Ercan [13] comes to similar results for smartphones and Bhakar et al. [14] for different types of computer monitors. High GWP`s per product weight can be found formany of the so called rare earth metals, although from a global perspective the base and bulk metals contribute significant more to the total environmental impacts of the planet because of their high production rates [15]. ...
Preprint
Improvements for the environmental footprint of gold from recycling using the method of Lifecycle Assessment.
... n the use phase(Berwald et al. 2020). However, smart TVs also contain CRMs, and extending their use phase is proven to have a positive environmental impact(Berwald et al. 2020, Prakash et al. 2016).An important aspect in the total environmental impact of both smartphones and smart TVs is energy and raw materials usage in the network infrastructure.Ercan et al. (2016) estimated that the global warming potential (GWP) is 43 kg CO2 ...
Technical Report
Full-text available
The report focuses on the mineral natural resources needed for digitalization. The use of digital applications and digital devices continues to grow and increasing amounts of information are being converted into a digital format. Prior research on digitalization in the context of sustainability has focused mainly on energy consumption and emissions. However, with the increasing demand for ICT hardware in numerous applications in modern society, the raw materials requirement of digital devices has become a crucial sustainability issue. Therefore, this report delves deeper into the topic of the raw materials consumption of digitalization. We focus on the following challenges and topics: • Sources, production, availability and sustainability of digitalization raw materials • ICT sector's raw materials consumption, with a specific focus on selected key end-user devices: smartphones and smart TVs • Key aspects of the ICT value chain • Key ICT consumer and end-user aspects • Possible solutions to support the sustainability of digital devices throughout their life cycle • Key policy aspects and recommendations
Article
Electronic devices consume energy both in the production and the use phase. Furthermore, the “hidden” impacts linked to their use are not frequently assessed and they depend on the behavior of the users, besides the servers and complex web networks. It must be underlined that many websites employ ads and trackers as part of their monetization strategy and, in order for online ads and trackers to work, they add an additional code to be executed on the users’ machines, which in turn requires more processing power. Considering that the Internet had an estimated 4.9 billion users in 2021, the global energy and carbon impacts of online ads and trackers might be significant. To investigate this phenomenon, we designed a novel automated framework for bottom‐up estimation of greenhouse gas (GHG) emissions attributable to software using exclusively free and open source software. Our process involved the building of a random sample of global news websites which we visited with and without an ad‐blocker, each time collecting power usage in identical conditions. The gathered data were put into an ordinary least squares (OLS)‐based linear regression model, which showed that ads and trackers on news websites require on average an additional 6.13 W of power on personal computers. This result was then tuned to global environmental and technological parameters to estimate that in 2019, on the client side, ads and trackers on the news websites consumed 0.61 TWh of electrical energy, emitted 0.29 MtCO2eq of GHG, and cost all Internet users approximately 140 million USD (purchasing power parity) of electrical energy. This article met the requirements for a gold‐gold JIE data openness badge described at http://jie.click/badges.
Article
Life cycle assessment (LCA) is the internationally adopted tool to assess environmental footprint. However, as highlighted by Billstein et al. [3] and Arushanyan [4] carrying out an LCA for ICT equipment is a challenging task due, to the amount of data that should be collected to achieve accurate results. This paper describes how documents such as full materials declarations can be used to solve this issue. Furthermore, the circular economy concept is introduced by analyzing alternative business models and the ITU-T L.1023 on circular economy scoring. Even if LCA was considered as a criterion in the L.1023 the link is not always straightforward between these two methods. Hence, this paper investigates how LCA results can be linked to the L.1023 criteria and proposes new criteria, for instance on recycled metals content and modularity.
Reducing residential carbon emissions is important for all countries to achieve carbon neutrality. For such task, the impact of digital development of finance on residential carbon emissions cannot be ignored. This paper estimates residential carbon emissions in 186 Chinese cities from 2011 to 2019 and examines the effects and mechanisms of digital finance on residential carbon emissions. The results show that digital finance exacerbates residential carbon emissions in China. In terms of the structural composition of carbon emissions, digital finance mainly increases residential carbon emissions by affecting electricity consumption and transportation in China. Mechanism analysis shows that digital finance can increase residential carbon emissions through the consumption effect and employment effect.
Article
Discarded consumer electronics are a significant and growing source of hazardous waste. Repurposing smartphones has the potential to reduce the rate of disposal. However, reusing smartphones for general computational tasks is difficult due to the variety and domain specificity of the mobile software stack, and the relatively low computational power of these devices. We present the design and proof-of-concept implementation of Renée, a smartphone-based cluster built from used phones that provides function-as-a-service (FaaS) capabilities. Our experience indicates that reusing decommissioned smartphones for general computational tasks is not only feasible but also that a local cluster built from these devices has the potential to provide faster response times than commercial FaaS while reducing the yearly global warming potential of the reused devices.
Conference Paper
Full-text available
Rapid development of mobile phones leads to high turnover and a lifetime shorter than the 3 years stated by manufacturers. •In the current business model, users purchase and own the phone. There is little incentive to return it at end of life. •Replaced handsets are kept by users until worthless, then often disposed to landfill, either in the EU or abroad. •This disposal of mobile phones leads to a loss of resource and valuable metals which could otherwise be recycled. •Replacement of lost resource places burden both through mining virgin ore and also intensive manufacturing processes.
Article
Purpose The possibilities for full life cycle assessment (LCA) of new Information and Communication Technology (ICT) products are often limited, so simplification approaches are needed. The aim of this paper is to investigate possible simplifications in LCA of a mobile phone and to use the results to discuss the possibilities of LCA simplifications for ICT products in a broader sense. Another aim is to identify processes and data that are sensitive to different methodological choices and assumptions related to the environmental impacts of a mobile phone. Methods Different approaches to a reference LCA of a mobile phone was tested: (1) excluding environmental impact categories, (2) excluding life cycle stages/processes, (3) using secondary process data from generic databases, (4) using input-output data and (5) using a simple linear relationship between mass and embodied emissions. Results and discussion It was not possible to identify one or a few impact categories representative of all others. If several impact categories would be excluded, information would be lost. A precautionary approach of not excluding impact categories is therefore recommended since impacts from the different life cycle stages vary between impact categories. Regarding use of secondary data for an ICT product similar to that studied here, we recommend prioritising collection of primary (specific) data on energy use during production and use, key component data (primarily integrated circuits) and process-specific data regarding raw material acquisition of specific metals (e.g. gold) and air transport. If secondary data are used for important processes, the scaling is crucial. The use of input-output data can be a considerable simplification and is probably best used to avoid data gaps when more specific data are lacking. Conclusions Further studies are needed to provide for simplified LCAs for ICT products. In particular, the end-of-life treatment stage need to be further addressed, as it could not be investigated here for all simplifications due to data gaps.
Conference Paper
The IP (Internet Protocol) core network operated by TeliaSonera in Sweden has been studied according to a described definition (system boundary) and extrapolated to a national level for Sweden. A top-down and a bottom-up data collection approach have been used to quantify network equipment and energy consumption. The electricity consumption of the IP core network was about one fourth of the total energy consumption of the connected mobile and fixed access networks. The use stage results for the Swedish IP core network is about 0.08 kWh per GB (Gigabyte) of data traffic corresponding to about 19 g CO2e/GB when applying a Swedish electricity mix. If instead a global average electricity mix is used, the resulting green house gas emission would be about 64 g CO2e/GB. A method is also proposed where the amount of data traffic is used to allocate the IP core network on different usage. The method is exemplified by the 3G mobile broadband data traffic in Sweden.
Waste Handling: Regional and Global End-of-Life Treatment Scenarios for ICT Equipment Royal Institute of Technology LCA of data transmission and IP core networks
  • A Liebmann
  • J Malmodin
  • D Lundén
  • M Nilsson
  • Andersson
Liebmann, A., 2015 ICT Waste Handling: Regional and Global End-of-Life Treatment Scenarios for ICT Equipment Royal Institute of Technology [14] Malmodin, J., Lundén, D., Nilsson, M., Andersson. 2012. LCA of data transmission and IP core networks. Available online:http://www.ericsson.com/res/docs/2013/lca-of-data- transmission-and-ip-core-networks.pdf (accessed on 15
On the pulse of the networked society Available online
  • Ericsson
  • Report
Ericsson mobility report, 2015. On the pulse of the networked society Available online: http://www.ericsson.com/res/docs/2015/mobility- report/ericsson-mobility-report-nov-2015.pdf (accessed on 15 April 2016)
Mobile data usage-Price and bucket size matter Available online
  • A B Tefficient
Tefficient AB, 2015. Mobile data usage-Price and bucket size matter Available online: http://media.tefficient.com/2015/06/tefficient-industryanalysis-2-2015-mobile-data-usage-and-pricing-26-June.pdf (accessed on 15 April 2016)
Global warming potential of a Smartphone; using life cycle assessment methodology Available online: http://kth.divaportal .org/smash/get
  • M Ercan
Ercan, M. 2013 Global warming potential of a Smartphone; using life cycle assessment methodology. Available online: http://kth.divaportal.org/smash/get/diva2:677729/FULLTEXT01.pdf (accessed on 5 February 2016)
20?? A resource-efficient Europe – Flagship initiative of the Europe 2020 Strategy Available online: http://ec.europa.eu/resource-efficient-europe
Europe 2020, 20?? A resource-efficient Europe – Flagship initiative of the Europe 2020 Strategy Available online: http://ec.europa.eu/resource-efficient-europe/ (accessed on 15 April 2016)
Review of mobile handset eco-rating schemes Available online: https
ITU, 2012. Review of mobile handset eco-rating schemes. Available online: https://www.itu.int/dms_pub/itu- t/oth/4B/01/T4B010000030001PDFE.pdf (accessed on 15
Fairphone (Faunhofer), 2013 Available online: https://www.fairphone.com
  • June
  • Pdf
June.pdf (accessed on 15 April 2016) [18] Fairphone (Faunhofer), 2013. Available online: https://www.fairphone.com/2013/08/01/whats-in-a-life- cycle-assessment/ (accessed on 15 April 2016)