Sie sind auf Seite 1von 11

Global environmental impact assessment

of the Pb-free shift


Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba
National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan

Abstract
Purpose – Using two different conceptual approaches to environmental life-cycle assessment, attributional and consequential, the purpose was to test
the hypothesis that a typical lead free solder paste Sn95.5Ag3.8Cu0.7 is worse than Sn63Pb37 as far as global environmental impacts are concerned.
Design/methodology/approach – Single index weighting indices within the impact methodology Life cycle Impact Assessment Method based on
Endpoint Modelling (LIME) impact methodology, were applied to the flows of three life cycle inventory models and their globally related flows. The LIME
results based on three environmental impact categories, i.e. resource consumption, global warming and ozonelayer depletion are presented and
discussed.
Findings – The attributional LCA (ALCA) results point towards a larger impact for Sn95.5Ag3.8Cu0.7 than Sn63Pb37 mostly due to the higher
Sn and Ag content. This study confirms earlier similar ALCAs. The system expansion for the Consequential LCA (CLCA) did not change this
conclusion.
Research limitations/implications – The present study has not included the affected microelectronics packaging parts of electronic products, nor has
it included toxic effects as they are local. ALCA was considered to be equal to CLCA for Sn95.5Ag3.8Cu0.7, where no Ag nor Sn recycling was included.
Originality/value – For the first time a global environmental impact assessment of the shift to Pb-free solder paste using the LIME weighting method
applied to a CLCA is reported. Environmental life-cycle investigations intended to support decisions of an ecological nature in the microelectronics
packaging industry should benefit from the consequential approach.

Keywords Solders, Modelling, Life cycle costs, Environmental management

Paper type Research paper

Introduction other and to Sn63Pb37 (TL), e.g. using environmental life-


cycle assessment (LCA). Schoenung et al. (2005) presented
Since, the beginning of electronics production, solders have an excellent review of the current scientific status of more
been used as interconnection materials and Pb has been or less the whole complex problem related to the industrial
widely used in solder alloys for its proven reliability in ecology of Pb-free electronics. For LCA, they conclude that
eutectic combination with tin (Plumbridge, 1996). previously performed case studies lack transparency and
However, the heavy metal Pb is under certain conditions
therefore could not be easily validated (Schoenung et al.,
harmful to humans, and it has therefore been banned from
2005). Nevertheless, several authors indicate that the Pb-
use in petrol, water pipes and paint. In Europe, an EU
free shift will be disadvantageous from a global warming
directive on restrictions on the use of hazardous substances
point of view, but advantageous from a toxic emissions
required the elimination of Pb in electronics, in all but a
number of special applications, by July 1, 2006 (European point of view (Verhoef et al., 2004; Itsubo et al., 2004a;
Union, 2003). Solder is the main Pb use within electronic Ekvall and Andræ 2006; Geibig and Socolof, 2005; Andrae
products, therefore Pb-free solders, anticipated to reduce and Liu, 2006). However, the resource issues, for example
toxic environmental impact, have and will come into the possible scarcity of metals replacing Pb, could also be
extensive use (Itsubo et al., 2004a). Although the Pb-free important for certain metals, as in 2003 around 120 Gg of
solders are not entirely new materials, they have not yet electronic solders were produced of which approximately
been extensively used. An analysis of the environmental 12 Gg were used to produce solder paste.
impacts of Pb-free solders as compared to Sn-Pb solders The global environmental impacts, and a tradeoff between
would help in avoiding worse alternatives. The most likely them, can be quantified using methodologies such as the Life
high-volume Pb-free alloys have similar physical properties Cycle Impact Assessment Method based on Endpoint
to Sn95.5Ag3.8Cu0.7 (LF), however, several Pb-free Modelling (LIME) (Itsubo et al., 2004b). Further, to the
material combinations using Sn in combination with Bi, author’s knowledge, no entire electronic product has so far
Cu, Co, In, Sb, Zn, have previously been compared to each been evaluated, using LCA, when comparing Pb and Pb-free
solders, but that can be explained because most of the Pb that
The current issue and full text archive of this journal is available at will be replaced in electronics originates from solders. The
www.emeraldinsight.com/0954-0911.htm hypothesis for the present research is that there will be
negative environmental impact assessment results globally, as
a consequence of introducing LF paste to replace TL paste.
Soldering & Surface Mount Technology
19/2 (2007) 18– 28
q Emerald Group Publishing Limited [ISSN 0954-0911] The staff of the LCA methodology group at AIST is acknowledged for
[DOI 10.1108/09540910710836502] their support leading to the completion of this paper.

18
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

The problems addressed in order to attempt to falsify this defined as the smallest system for which data is collected
hypothesis are: (Andræ et al., 2005). One variant of LCA is the so-called
H1. What are the attributional LIME scores, expressed attributional LCA (ALCA), sometimes also referred to as
in Japanese Yen, for an LCA comparison between LF retrospective or accounting LCA. The ALCA aims to specify
and TL? how much of the global environmental load within the system
H2. What is the consequential change in LIME score when boundaries belongs to a certain human activity. ALCA
a global change between TL and LF is made? commonly uses average data in contrast to specific data, and
is used to compare two or more alternatives and also to find
the most environmentally relevant unit processes. Another
Environmental life-cycle assessment of solders in variant of LCA is the so-called Consequential LCA (CLCA)
where the consequences of decisions, such as phasing out Pb
electronics
from electronics, are evaluated. These consequences can also
The total global environmental load and impact on the relate to activities outside of the system boundaries, should
biosphere, troposphere, and stratosphere is mainly the result these activities be affected (Ekvall and Andræ 2006). In
of the industrial metabolism connected to product systems in CLCA, the change in the global environmental load as a
which resources continuously are converted into useful result of adding or removing a specific human activity is
products and services demanded by human societies. studied. The CLCA make use of marginal data as it is the
Environmental LCA is most commonly used by universities marginal producers and consumers that are affected by a
and companies as a method to evaluate the mass balance of small change. One of the practical problems with the CLCA is
inputs and outputs of specific product systems and to organise how to identify who these marginal actors will be. An attempt
and convert those inputs and outputs into environmental was made to model the Pb and Pb scrap markets (Ekvall and
themes or categories relative to resource use, human health, Andræ 2006). The inventory flows (obtained in the data
and ecotoxicity explained by Rebitzer et al. (2004) and collection step as emissions, resource consumptions, and
Pennington et al. (2004). In Figure 1 the phases of LCA and waste amounts) from either an ALCA or CLCA, are classified
their inter-relationships are schematically shown. according to which possible environmental impacts, e.g.
In this context, an example of a LCA goal could be to global warming, they could cause. Anthropogenic and
compare the life cycle impact of a mobile phone using Pb and potential environmental impacts, which may be global,
one without Pb. The scope, which means what is intended to regional, local or a combination of these, include global
be included with the system boundary, could be just the warming, ozonelayer depletion, photo-chemical oxidant
solder life-cycle or also include the life cycles of all those parts creation, acidification, local air pollution, human toxicity,
of the phone using Pb, e.g. printed wiring board finishes, ecotoxicity, eutrophication, and resource consumption. For
termination finishes, ball grid array interconnections, and example, the CFC’s are classified as being able to contribute
internal chip-to-substrate interconnects for controlled both to global warming and to ozone layer depletion. After the
collapse chip technology (Garner et al., 2000). The classification the flows are characterised according to their
functional unit, the basis of the calculation, must be chosen relative importance for each environmental impact indicator.
and must reflect the function of the life-cycle. In this case, it The present research will focus on the integration of the
could be “the average use of one mobile phone during three environmental impact indicators for resource depletion,
years”. The scope decides which unit processes to quantify in global warming, and ozonelayer depletion as these
the inventory analysis within the system boundaries, of which environmental effects are global. In Figure 2 the scope of
examples are shown in Figure 2. In LCA a unit process is the present attributional solder paste life-cycle model, as
currently used by most LCA practitioners, is shown. The use
of the electronic product is outside the system boundary.
Figure 3 shows the scope of the present consequential
Figure 1 The phases in an LCA
model from the perspective of the global shift to LF from TL.

1. Goal and Scope Methodology for impact assessment – LIME


definition
Based on a finished inventory analysis where all data sources
are given (Ekvall and Andræ 2006), the method “Life-cycle
Impact assessment Method based on Endpoint modelling,
LIME” was applied (Itsubo et al., 2004b). The origin of this
methodology is a study conducted by the LCA National
2. Inventory analysis Interpretation Project of Japan aiming at the development of a Japanese
version of a damage-oriented impact assessment method. In
LIME, the potential damage is measured for four safeguard
objects: human health, the utilisation of non-renewable
resources (social assets), the increase of extinction risk
(biodiversity), and the loss of primary production caused by
3. Impact mining of resources (primary productivity) are individually
assessment measured. Modelling socio-economic impact was based on the
concept of user-cost, which accounts for the equity of future
generations. Interviews were performed in Japan where a
Source: International Organisation for Standardisation (1997) statistically representative population answered questions on

19
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

Figure 2 The scope of the attributional model for solder pastes

Primary Sn Primary Ag Primary Cu


production production production

Electricity
production
Bulk alloy
production

Fuel Flux
production Solder powder production
production

Preparation for
Paste solder recycling
production
Solder
incineration

Paste Use of
application Electronic Landfill of
(Reflow soldering) product solder

Figure 3 The scope of the consequential model of the global shift

Production and
Primary Sn Primary Pb Use of complimentary products
production production

Alternative Function of
World Pb use alternative Pb use
Market for
Pb
Function and
Electricity Pb recycling
use of competing products
production
Bulk alloy
production
Scrap collection
World
from
Market for
Fuel other Pb products
Pb scrap
production Solder powder
production
Waste management
of other Pb products
Preparation for
Paste
solder recycling
Flux production
production Solder
incineration

Paste Use of
application Electronic Landfill of
(Reflow soldering) product solder

how they valued different types of damage to the environment. expressed in Yen/kg enabling a comparison and integration of
This approach made it possible to make a monetary weighting the damage derived from different impact categories such as
between, e.g. human health and biodiversity. The Yen scores in global warming and ozonelayer depletion. In Figure 4 the
this context are what the Japanese society is willing to pay to LIME concept is shown with a focus on global impacts. Pb
avoid a unit of damage, caused by the environmental loadings, emissions are considered a local impact, as opposed to
to the safeguard objects shown in Figure 4. The compounds belonging to global impact categories, and are
present research made use of weighted LIME factors therefore not included within the scope.

20
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

Figure 4 Conceptual figure of LIME


Impact Category Safeguard
Inventory
category endpoint object
Human-toxic Other cancer
Pb chemical
Crop
production
Ozonelayer Human health
CFC Skin cancer
depletion

CO2 Forestry Social assets


production
Single index,
Yen, ¥
CH4 Global Thermal
warming stress Biodiversity

Infectious
Oil
diseases
Resource Net primary
consumption Terrestial productivity
Sn ecosystem

User
cost
Ag
Vascular
Damage Weighting
Characterisation plant species
assessment
Notes: CFC = ChloroFluoroCarbons. Local toxic emissions like Pb are not part of the scope

For LIME, the resource consumption indices are originally Owing to lack of comparable cradle-to-gate resource data for
based on the resource characterisation as developed by the Ag, Pb, Cu and Sn production, it was assumed that the metal
French company Conception Development Durable resource consumption was only the produced metal. For
Environment CODDE (2007) for a Raw Material Depletion example, for 1 kg of Ag produced, 1 kg of Ag metal resources
(RMD) Indicator. Equation (1) explains how the indicator is was assumed as input, whereas the extraction and processing
calculated: energy inputs were taken from the literature. For Ag a mixed
model assuming Ag to be a by-product mainly of Canadian
X  
1 1 Zn/Pb and Australian Au production was used (Teck
RMD ¼   £ Ii. . . £ kg
Ri kg £ Years Cominco Metal Ltd, 2004; Stewart and Petrie, 2006; New
Ri £ Y i ¼ M i Boliden, 2005), for Pb a model assuming Pb to be a product
of primary Pb production (Althaus and Classen, 2005), for
RMD is the total characterisation factor; Ri the available Cu a model assuming Cu to be a by-product of Ni production
reserve base of a resource (i) such as a high-grade metal ore, (Althaus and Classen, 2005), and for Sn a model assuming Sn
coal, etc. which realistically can be extracted; Yi the number to be a product of primary Sn production (Althaus and
of years left of resource (i) considering the rate at which it is Classen, 2005). However, the most relevant LCI data would
currently depleted; Mi the production of resource (i); and Ii have been a global average for ALCA and the marginal for
the inventory flow of resource (i). The inventory flow is the CLCA. Further, for the ALCAs, the outflow of printed board
amount of, e.g. silver needed to produce the functional unit. assemblies from the preparation process has a positive
Table I shows the figures used to calculate the RMD values. economic value. Therefore, it was not regarded as waste but
At this stage the recycling of metals is not included, but would a raw material inflow to the life-cycles where the recycled
presumably delay the decrease of Yi. metals are used. Thus, for ALCA, the metal recycling

Table I The basis for RMD indices


Substance Ri (kg) Mi (kg/year) Yi (year) Reference
Ag 5.7 £ 108 1.97 £ 107 28.9 USGS (2004a)
Sn 1.1 £ 1010 2.64 £ 108 41.7 USGS (2004b)
Cu 9.4 £ 1011 1.46 £ 1010 63.1 USGS (2004c)
Ni 1.4 £ 1011 1.4 £ 109 100 USGS (2004d)
Zn 4.6 £ 1011 9.6 £ 109 47.9 USGS (2004e)
Pb 1.4 £ 1011 3.15 £ 109 44.4 USGS (2004f)
Oil 1.76 £ 1014 3.87 £ 1012 45.4 GeoHive Global Statistics (2004a, b), Tomkiewicz (2006) and Asif and Muneer (2007)
Coal 9.09 £ 1014 3.79 £ 1012 240 GeoHive Global Statistics (2004c, d) and Ekawan and Duchêne (2006)
Natural gas 1.49 £ 1014 2.24 £ 1012 66.7 GeoHive Global Statistics (2004e, f) and Afgan et al. (2007)

21
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

processes belong to subsequent life-cycles. Owing to lack of Sensitivity check


economic and marginal data, e.g. for the Sn market, the
A life-cycle impact assessment method which can be
inventory result for the consequential LF study was assumed
compared to LIME is the Eco-indicator ‘99 (Eco-i. 99)
to be identical to the one for the attributional LF study. All
method as it is based on the disability adjusted life year
LIME results and others presented below are expressed per
(DALY), concept (Abou-Zahr, 1999). The DALY concept
0.53 cm3 solder paste (corresponds to 2.5 g TL having a
combines in one score the time lived with disability and the
density of 4.7 g/cm3, 90 wt% metal alloy and 10 wt% flux)
time lost due to premature mortality. As shown by Table III
applied to a printed board assembly before the reflow
Eco-i. 99 has not reported a weighted index for damage to
soldering process. The quantity 2.5 g was chosen based on the resources caused by extraction of Ag and some other
USEPA study of a typical printed circuit board assembly resources. The results shown in Figure 6 were obtained by
(Geibig and Socolof, 2005). multiplying the obtained inventory data by the corresponding
weighted damage factors. The major similarity is that
consumption of Sn resources is one of the dominating flows.
Inventory analysis results
Three LCI’s covering 416 substance flows were investigated. Completeness check
Totally 97 resources and seven gases were relevant for global
impacts. In Table II the results for nine selected flows are The robustness of the results in Figure 5 are hampered by at
displayed and they were selected based on an initial LIME least five factors:
screening of the product systems. Pb is included though it is 1 not all emissions to air and the resources, which
theoretically could have globally related LIME indices,
presently not quantified for global impact categories. The
were reported by Itsubo et al. (2004b), e.g. carbon
column farthest to the right represents the consequential
monoxide (Holloway et al., 2000);
inventory result of the shift.
2 too low resolution of the inventory result, which means
more flows of for example CFC’s could possibly be
“hidden” inside the system boundaries;
Impact assessment results 3 poor precision in the numerical values of emissions and
The results were obtained by multiplying the obtained resources, which means different kinds of temporal and
inventory data by the corresponding weighting factors measurement uncertainties;
(Itsubo et al., 2004b). The characterisation indices and 4 poor precision of the LIME indices, which means
LIME factors used in the present study are shown in Table III. different kinds of temporal, spatial, and geographical
Overall, 98 mass% of the resource input flows, 100 mass% uncertainties; and
of the greenhouse gases, and 100 mass% of the ozone 5 poor representation of unit processes, which means that a
depleting gases, had a corresponding LIME index. The total greater number of factories in reality represent the average
LIME scores for the CLCA comparison between TL and LF presently used.
were around 1.8 and 3.4¥, respectively. The total However, considering what turned out to be the dominating
LIME results for the ALCA were also near 1.8 and 3.4¥, issues of the study, Ag and Sn resources, the global perspective
respectively. In Table IV it is shown that Sn and Ag resources of the study, as well as the relatively small product systems, it is
from Sn and Ag production, respectively, were the hot spots unlikely the results would change drastically should the effects
in this study. of these five factors be greatly improved.
Figure 5 shows the LIME result where resource
consumption, global warming and ozonelayer depletion are
Discussion and interpretation
weighted into a single index.
The consequence is that the total LIME score will increase The CLCA and the ALCA results both indicate that the
by 1.6¥ or by 90 per cent. overall LIME scores in Yen will increase considerably per

Table II Selection of LCI results expressed per functional unit of TL and LF


Substance Category Unit ALCI TL ALCI LF ALCI LF – ALCI TL CLCI TL ALCI LF – CLCI TL
Coal Resource G 83 103 20 82 21
Oil Resource G 12 18 6 15 3
Natural gas Resource G 12 15 3 12 3
Sn Resource G 2.9 3.9 1 2.9 1
Ag Resource G 0 0.16 0.16 0 0.16
Pb Resource G 1.7 0 2 1.7 0.73 20.73
Cu Resource G 0 0.03 0.03 0 0.03
CH4 Emission to air G 0.77 0.94 0.17 0.78 0.16
CO2 Emission to air G 228 290 62 234 56
Pb Emission to air G 0.032 ,0 2 0.032 0.03 20.03
Pb Emission to water G 0.15 ,0 2 0.15 0.13 20.13
Pb Emission to soil G ,0 ,0 0 2 0.001 0.001

22
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

Table III Indices used


Substance Impact on LIME (Yen/kg) Eco-i. 99 (year/kg) RMD (102 15 3 kg2 1 3 year2 1) GWP20 (kg/kg) ODP (kg/kg)
Ag Resource 6,900 61,000
U Resource 1,030
Sn Resource 390 14 2,200
Wood Resource 105
Cu Resource 53 0.87 16
Ni Resource 47 0.39 71
Zn Resource 45 0.045 45
Mn Resource 39 0.0074
Pb Resource 29 0.18 160
Oil Resource 1.5 0.14 0.13
Coal Resource 1.27 0.006 0.0046
Al Resource 1.18 0.057
Natural gas Resource 0.69 0.11 0.10
Fe Resource 0.44 0.0012
Cr Resource 0.37 0.022
Bauxite Resource 0.19 0.012
Gravel and sand Resource 0.062
Mass of rock Resource 0.062
Rock salt Resource 0.062
C2F6, R-116 Air 22,900 52 7,700
Ozone-depleting gases Air 19,000 27 1
CFC/HCFC Air 19,000 27 1
CF4 Air 11,000 36 3,900
N2O Air 570 1.8 330
CH4 Air 44.3 0.11 64
CO2 Air 1.74 0.0055 1
Notes: Eco-i. 99 ¼ Eco-indicator ‘99 (hierarchist perspective, H, A) weighted damage factors (Dreyer et al., 2003), GWP20 ¼ global warming potential during
20 years, ODP ¼ ozone depletion potential

Table IV Top contributors to LIME single index given per solder paste type
Solder paste Life-cycle stage Unit process Flow Percentage of contribution
TL Manufacturing Sn production Sn resource consumption 62
Use Solder paste application CO2 emissions to air from electricity production 15
Use Solder paste application Coal resource consumption from electricity production 5
LF Manufacturing Sn production Sn resource consumption 45
Manufacturing Ag production Ag resource consumption 31
Use Solder paste application CO2 emissions to air from electricity production 9

functional unit. The most important differences between TL All impacts attributable to Pb production will not disappear as,
and LF are two-fold: Sn and Ag resource consumption. The Ag on the margin, Pb will be used in Pb-acid batteries (instead of
resources from Ag production are significant due to the solder pastes) which in combination with photovoltaic cells will
relatively high LIME factor for Ag, about 6,900¥/kg. CO2 replace diesel combustion for electricity generation. In fact,
emissions from electricity production are also noticeable. CLCA helped identify this offset in the impact related to the Pb
However, considering global warming results alone, it was production. Figure 6 shows the relative importance for
earlier reported for a similar inventory to the present study that different processes as evaluated by LIME compared to Eco-i.
the solder application process and the Sn production were the 99 and the three environmental impact categories on which
processes mainly affected (Ekvall and Andræ 2006). The they are based. Considering RMD results alone, they show a
marginal Pb usage, as a result of the ban on the use of Pb in big resemblance to the overall LIME results. The present
solder, will be where the competition is the largest and where CLCA study predicts that the LIME score will rise as the
the Pb consumers are most sensitive to a Pb price change. On avoidance of the environmental impacts connected to diesel
the margin, in remote areas, diesel combustion to generate combustion does not outweigh the increased resource
electricity is expected to be replaced by Pb back-up batteries consumptions and emissions from Ag, Sn, and energy
(the marginal Pb usage) when Pb is banned in solders. production.

23
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

Figure 5 The consequential LIME results expressed per functional unit obtained when subtracting the CLCA TL from the ALCA ( ¼ CLCA) LF

Yen / F.u.
Global LIME (toxicity not included)
1.2
1
0.8
0.6
0.4
0.2
0
–0.2
Ot So Sn Ag Di Pb
her lde pro pro ese pro
s rp du du l co du
ast cti cti mb cti
ea on on ust on
pp ion
lic
ati
on

Consequense of shift

Figure 6 The consequential LIME results compared to Eco-i. 99 and environmental impact categories GWP20
100%

80%

60%
Pb production
Diesel combustion
Ag production
40%
Sn production
Solder paste application
Others
20%

0%
Global Eco-i. 99 GWP20 ODP RMD
LIME
–20%
Notes: Global warming potential during 20 years(GWP20), ODP = ozone depletion potential,
RMD = raw material depletion

Further, it is likely that the Eco-i. 99 H, A weighted factor to the lack of inventory transparency. Nevertheless,
for Ag resources would be relatively high, and obviously two similar solder pastes to the present ones were
would indicate more similarities between overall and evaluated by the USEPA. These were Sn63Pb37 (SnPb)
individual Eco-i. 99 and LIME scores. and Sn95.5Ag3.9Cu0.6 (SAC). The results from the
Moreover, electricity production could be of higher environmental impact categories “Nonrenewable resource
importance, although not likely to change any conclusions. The use (NRR)” “global warming (GW)” and ”Ozone depletion
data quality is poor for the model of the scrap Pb market and (OD)” used by USEPA were used for a comparison with the
waste management of competing sources of scrap Pb. However, present LIME analysis. The USEPA functional unit was
this part of the model does not have a big influence on the results. 1,000 cm3 of solder metal alloy applied before the reflow
oven, compared to the present 0.53 cm3 of solder paste.
Other analyses of the shift to Pb-free solders Figure 7 shows the comparative results for the two studies
USEPA performed an attributional cradle-to-grave LCA expressed per USEPA functional unit, where 9,300 g of TL
comparison between Pb solders and Pb-free solders (Geibig and 8,170 g of LF correspond to that measure. The units
and Socolof, 2005). The present study cannot be easily are kg resources for NRR, kg CO2-equivalents for GW, and
compared to the one by USEPA, among other things due kg £ 106 CFC-11-equivalents for ODP.

24
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

Figure 7 Comparative results for the USEPA solder LCA study and the present regarding three different environmental impacts

kg

2000

1500
USEPA SnPb
USEPA SAC
1000 Present study TL
Present study LF

500

0
NRR GW ODP
Environmental impact category

The NRR difference is mainly due to the “inert rock” concluding remark is that there exists a limited production
resource consumption used in the USEPA model of electricity infrastructure for co-products in Pb ore processing. In case of
generation. The relatively low ODP value for TL is due to an extended ban on lead, both the availability and recovery of
effects identified by the use of CLCA, where the ODP from a range of metals will be affected (Verhoef et al., 2004; Reuter
alternate Pb usage (battery production) offsets ODP and Verhoef, 2004).
connected to electricity and Sn production. When the top
contributing inventory flows for NRR, GW and OD for SnPb Effect of recycling
and SAC were multiplied by the respective LIME single It is uncertain how much of the Sn, Ag, Pb, and Cu solder
index, the Zn-Pb-Cu resource consumption and CO2 metals that will be recycled will specifically be used for
emissions related to electricity generation for solder production of new solders. Quantification of this is especially
application were the dominating flows for SAC and SnPb, important for Ag and Sn. Based on this research increased
respectively. This comparison shows that USEPA did not recycling of Ag and Sn could be significant in decreasing the
characterise the resource consumption, but just reported the global impact. In the present research it was not possible to
magnitude and top contributors. For example, “inert rock” apply the LIME factors for other than global effects, as the
having a moderate LIME index, dominated the NRR which local LIME is adjusted for Japanese conditions. This
merely reports the amount of resources used. Should another screening of the global situation was however useful as it
type of characterisation by for example RMD have been strengthens earlier results showing that the social and
performed, other flows would have appeared as important. economic impacts, due to the consumption of resources for
Furthermore, the analysis of the USEPA study confirmed that Sn and Ag production, could rise as a result of the shift to Pb-
stratospheric ozone depletion is of minor importance for the free solder paste (Itsubo et al., 2004c, p. 441 Figure 2(b)-(c)).
LIME score.
Verhoef et al. (2004) and Reuter and Verhoef (2004) Consequential LCA including Ag and Sn
showed that dynamic modelling, as a life-cycle inventory of An important discussion is whether a consequential model for
the total upstream system, could be useful for assessing the LF would lead to different conclusions. How much will global
environmental aspects of the manufacturing of solders. The Sn and Ag usage rise as far as electronics solder is concerned?
system boundaries included the ore processing, metal Using Deubzer’s replacement scenario from the year 2003 to
production and solder production activities for solder alloys 2006 (Deubzer, 2007) the rise could be from about 68 Gg
including Sn60Pb40 and several Pb-free combinations. No before the shift (29 mass% of the global Sn consumption) to
inventory data were reported, but it was evaluated using the about 113 Gg (35 mass%) for Sn after the shift, and from
Eco-i. 99 method. However, the question was raised, whether 0.075 Gg (0.2 mass%) to 3.6 Gg (12 mass%) for Ag (it is
a ban on use of Pb will lead to an environmental benefit, as uncertain exactly how much electronic solder is globally used,
the manufacturing of solders could be considered a global but it was about 120,000 tonnes in 2003 and probably more
open loop recycling and production system. Verhoef argues than 10 per cent more in 2006). It mainly depends on the
that the governmental decisions to introduce Pb-free will push primary production, recycling, pricing and electronics market
the system out of the existing steady state globally, and while consumption of these metals, where economies such as that of
production of Pb, Cu and Sn would not be affected China Govern more and more the world market trends for Sn.
significantly, local changes could be important. The shift to Pb-free solder could lead to a decreased use of Ag
Nevertheless, the resource depletion dominated Eco-i. 99 and Sn in a mix of other products. Which marginal Sn
scores were higher for SnAgCu than for Sn-Pb, but became consumers, having the possibility to substitute Sn, are most
smaller as the authors changed the weighting of resource sensitive to a change in Sn price? Which marginal Ag
depletion to 5 per cent as compared to the initial settings of consumers, having the possibility to substitute Ag, are most
Eco-i. 99. The findings indicated that preventing human sensitive to a change in Ag price? A change in the Sn and Ag
toxicity will instead lead to resource damage. An important prices will affect the uses of these metals. For some products

25
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

the Sn or Ag cost is a small part of the total production cost, .


the design of quantitative partial economic equilibrium
while for others Sn or Ag have important functional models for the electricity, Sn and Ag markets and the
advantages, making the demand for Sn and Ag less sensitive corresponding scrap markets for the metals;
to changes in the Sn or Ag price. If the marginal products .
identification of marginal consumers for Sn, Ag and other
using Sn cannot replace Sn in the long run, the total Sn solder metals;
production will go up. This is under the condition that all .
marginal data for primary metal production need to be
other major users of Sn, for example producers of Sn-coated collected; and
cans, also use the same amount of Sn after the shift to Pb-free
.
system expansion including Au and Ag containing printed
solder. However, if the Sn price increases too much, wiring board surface finishes could be worthwhile.
materials such as Al, glass, paper, plastic and Sn-free steel
can substitute for Sn in, for example, pet food cans. On the References
other hand, Sn can also find new markets for example as
Abou-Zahr, C. (1999), “Disability adjusted life years
alloys in automotive balance weights. (DALYS) and reproductive health: a critical analysis”,
As for annual Ag consumption, industrial and decorative Reproductive Health Matters, Vol. 7 No. 14, pp. 118-29.
uses, photography and jewelry and silverware represent more Afgan, N.H., Pilavachi, P.A. and Carvalho, M.G. (2007),
than 95 percent, and the electronics and photography “Multi-criteria evaluation of natural gas resources”, Energy
industries are the main consumers (Lanzano et al., 2006). Policy, Vol. 35 No. 1, pp. 704-13.
The unique properties of Ag restrict its substitution in most Althaus, H.J. and Classen, M. (2005), “Life cycle inventories
applications. Further, long-term analyses of the Sn and Ag of metals and methodological aspects of inventorying
markets are required to forecast what can happen. material resources in ecoinvent”, International Journal of
Comparing the ALCA results for TL and LF is not the Life Cycle Assessment, Vol. 10 No. 1, pp. 43-9.
same as predicting the consequences of shifting from TL to Andræ, A.S.G. and Liu, J. (2006), “Consequential toxicity
LF. Occasionally, as seems to be the case regarding the solders assessment of the global shift to Pb-free solder paste”,
in the present study, the two techniques, ALCA and CLCA, Proceedings of the International Conference on Electronics
more or less provide the same conclusion. It is obvious that Packaging, 19-21 April, Tokyo, Japan, pp. 298-303.
ALCA is too rudimentary for estimating future environmental Andræ, A.S.G., Andersson, D.R. and Liu, J. (2005),
impacts, but on the other hand CLCA needs to be more “Significance of intermediate production processes in life
developed in co-operation with econometricians to be more cycle assessment of electronic products assessed using a
accurate and comprehensive. generic compact model”, Journal of Cleaner Production,
Vol. 13 Nos 13/14, pp. 1269-79.
Asif, M. and Muneer, T. (2007), “Energy supply, its demand
Conclusions and security issues for developed and emerging economies”,
Renewable and Sustainable Energy Reviews, Vol. 11 No. 7,
The following conclusions can be drawn based on the pp. 1388-413.
research carried out in the present study: CODDE (2007), Environmental Information & Management
.
as far as globally related environmental impacts are Explorer (EIME). Indicators Manual Version 1.9, Conception
concerned, the shift from TL to LF solder paste is likely to Development Durable Environment, available at: www.
increase them; codde.fr/fr/plaquette/eime%20manuel%20indicateurs.pdf
.
a significant increase was detected due to increased Ag (accessed 24 May)
and Sn production; Deubzer, O. (2007), “Explorative study into the sustainable
.
no significant increase in the LIME score could be related use and substitution of soldering metals in electronics”,
to the increased generation of electricity for the LF solder PhD thesis, Delft University, Delft, Figure 3-11.
application processes as compared to TL; Dreyer, L.C., Niemann, A.L. and Hauschild, M.Z. (2003),
.
the LIME score is highly dependent on the Ag and Sn “Comparison of three different LCIA methods: EDIP97,
weighting factors; CML2001 and Eco-indicator 99 – does it matter which one
.
ozone depleting substances had an insignificant influence you choose?”, International Journal of Life Cycle Assessment,
on the LIME scores; and Vol. 8 No. 4, pp. 191-200.
.
this study confirms earlier work reporting that the Ekawan, R. and Duchêne, M. (2006), “The evolution of hard
resource consumption will be higher for SnAgCu solder coal trade in the Atlantic market”, Energy Policy, Vol. 34
pastes than conventional SnPb. No. 13, pp. 1487-98.
Ekvall, T. and Andræ, A.S.G. (2006), “Attributional and
For LIME it has earlier been shown, in a scenario for Japan, consequential environmental assessment of the shift to lead-
that the toxicity of Pb was most important leading to the free solders”, International Journal of Life Cycle Assessment,
overall environmental superiority of Pb-free solders. The next Vol. 11 No. 5, pp. 344-53.
step would be to make a trade-off between the global impact European Union (2003), “Directive 2002/95/EC of the
of the locally related impacts, such as those originating from European Parliament and of the council of 27 January
the use of human-toxic chemicals, and the known global 2003 on the restriction of the use of certain hazardous
impacts. For example, the existence of industrial Pb aerosols substances in electrical and electronic equipment”,
could make way for updated LCA indices (Rankin et al., Official Journal of The European Union, Vol. 13.2.2003,
2005). pp. L37/19-23.
Globally focused CLCA’s of solders also need to be Garner, C.M., Gupta, V., Bissessur, V., Kumar, A. and
improved in a number of areas: Aspandiar, R. (2000), “Challenges in converting to

26
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

lead-free electronics”, Proceedings of the 3rd IEEE Electronics contamination”, Environmental Health Perspectives, Vol. 113
Packaging Technology Conference, 5-7 December, Singapore, No. 10, pp. 1344-8.
pp. 6-9. Rebitzer, G., Ekvall, T., Frischknecht, R., Hunkeler, D.,
Geibig, J.R. and Socolof, M.L. (2005), “Solders in Norris, G., Rydberg, T., Schmidt, W.P., Suh, S., Weidema,
electronics: a life cycle assessment”, United States B.P. and Pennington, D.W. (2004), “Life cycle assessment
Environmental Protection Agency Report No. EPA 744- part 1: framework, goal and scope definition, inventory
R-05-001, USEPA, Washington, DC. analysis, and applications”, Environment International,
GeoHive Global Statistics (2004a), “Proved oil reserves”, Vol. 30 No. 5, pp. 701-20.
available at: www.xist.org/charts/en_oilres.aspx (accessed Reuter, M.A. and Verhoef, E.V. (2004), “A dynamic model
22 May 2007) for the assessment of the replacement of lead in solders”,
GeoHive Global Statistics (2004b), “Oil production”, Journal of Electronic Materials, Vol. 33 No. 12, pp. 1567-80.
available at: www.xist.org/charts/en_oilprod.aspx (accessed Schoenung, J.M., Ogunseitan, O.A., Saphores, J.D.M. and
22 May 2007) Shapiro, A.A. (2005), “Adopting lead-free electronics:
GeoHive Global Statistics (2004c), “Proved coal reserves”, policy differences and knowledge gaps”, Journal of Industrial
available at: www.xist.org/charts/en_coalres.aspx (accessed Ecology, Vol. 8 No. 4, pp. 59-86.
24 May 2007) Stewart, M. and Petrie, J. (2006), “A process systems
GeoHive Global Statistics (2004d), “Coal production”, approach to life cycle inventories for minerals:
available at: www.xist.org/charts/en_coalprod.aspx South African and Australian case studies”, Journal of
(accessed 24 May 2007) Cleaner Production, Vol. 14 Nos 12/13, pp. 1042-56.
GeoHive Global Statistics (2004e), “Proved gas reserves”, Teck Cominco Metal Ltd (2004), “Trail operations and
available at: www.xist.org/charts/en_gasres.aspx (accessed sustainability – review of 2004”, available at: www.
24 May 2007) teckcominco.com/articles/environment/trail_sustainability_
GeoHive Global Statistics (2004f), “Natural gas production”, 2004.pdf (accessed 24 May 2007)
available at: www.xist.org/charts/en_gasprod.aspx (accessed Tomkiewicz, M. (2006), “Global warming: science, money
24 May 2007) and self-preservation”, Comptes Rendus Chimie, Vol. 9 No. 2,
Holloway, T., Levy, H. II and Kasibhatla, P. (2000), “Global pp. 172-9.
distribution of carbon monoxide”, Journal of Geophysical USGS (2004a), “Mineral commodity summaries”, available
Research, Vol. 105, pp. 12,123-47. at: http://minerals.usgs.gov/minerals/pubs/commodity/
International Organisation for Standardisation (1997), silver/silvemcs06.pdf (accessed 24 May 2007)
USGS (2004b), “Mineral commodity summaries”, available
“Environmental management – life cycle assessment –
at: http://minerals.usgs.gov/minerals/pubs/commodity/tin/
principles and framework”, ISO Standard 14040:1997(E),
tin__mcs06.pdf (accessed May 24 2007)
ISO, Genèva.
USGS (2004c), “Mineral commodity summaries”, available
Itsubo, N., Noh, J. and Inaba, A. (2004a), “Life cycle impact
at: http://minerals.usgs.gov/minerals/pubs/commodity/
assessment of lead-free solder considering the trade-off
copper/coppemcs06.pdf (accessed 24 May 2007)
relationship between toxic impact and other impact
USGS (2004d), “Mineral commodity summaries”, available
categories”, Material Transactions, Vol. 45 No. 11,
at: http://minerals.usgs.gov/minerals/pubs/commodity/
pp. 3194-200. nickel/nickemcs06.pdf (accessed 24 May 2007)
Itsubo, N., Sakagami, M., Washida, T., Kokubu, K. and USGS (2004e), “Mineral commodity summaries”, available
Inaba, A. (2004b), “Weighting across safeguard subjects for at: http://minerals.usgs.gov/minerals/pubs/commodity/zinc/
LCIA through the application of conjoint analysis”, zinc_mcs06.pdf (accessed 24 May 2007)
International Journal of Life Cycle Assessment, Vol. 9 No. 3, USGS (2004f), “Mineral commodity summaries”, available
pp. 196-205. at: http://minerals.usgs.gov/minerals/pubs/commodity/lead/
Itsubo, N., Noh, J. and Inaba, A. (2004c), “Life cycle impact lead_mcs06.pdf (accessed 24 May 2007)
assessment of lead-free solder considering the differences of Verhoef, E.V., Dijkema, G.P.J. and Reuter, M.A. (2004),
waste stages”, Proceedings of the Electronics Goes Green “Process knowledge, system dynamics and metal ecology”,
Conference, 6-8 September, Berlin, Germany, pp. 439-43. Journal of Industrial Ecology, Vol. 8 Nos 1/2, pp. 23-43.
Lanzano, T., Bertram, M., De Palo, M., Wagner, C., Zyla, K.
and Graedel, T.E. (2006), “The contemporary European
silver cycle”, Resources, Conservation and Recycling, Vol. 46 About the authors
No. 1, pp. 27-43. Anders S.G. Andrae received his MSc degree
New Boliden (2005) Sustainability report 2005, available at: in Chemical Engineering from the Royal Institute
http://vp031.alertir.com/sustainability2005en/ (accessed 22 of Technology, Stockholm, Sweden, in 1997, his
May 2007) Licentiate degree and PhD degree in Electronics
Pennington, D.W., Potting, J., Finnveden, G., Lindeijer, E., Production from Chalmers University of
Jolliet, O., Rydberg, T. and Rebitzer, G. (2004), “Life cycle Technology, Gothenburg, Sweden, in 2002 and
assessment Part 2: Current impact assessment practice”, 2005, respectively. Between 1997 and 2001 he
Environment International, Vol. 30 No. 5, pp. 721-39. was at Ericsson working as an Environmental Engineer with
Plumbridge, W.J. (1996), “Solders in electronics”, Journal of Life Cycle Assessment. He has published 17 papers in refereed
Material Science, Vol. 31 No. 10, pp. 2501-14. journals and conferences. Since, 2006, he has been a post
Rankin, C.W., Nriagu, J.O., Aggarwal, J.K., Arowolo, T.A., doctoral research scientist at the Advanced Industrial Science
Adebayo, K. and Flegal, A.R. (2005), “Lead contamination and Technology (AIST), Research Center for Life Cycle
in cocoa and cocoa products: isotopic evidence of global Assessment, Tsukuba, Japan. He is a Member of the IEEE and

27
Global environmental impact assessment of the Pb-free shift Soldering & Surface Mount Technology
Anders S.G. Andrae, Norihiro Itsubo and Atsushi Inaba Volume 19 · Number 2 · 2007 · 18 –28

won the IEEE Young Award at the International Conference on Information Studies Department at the Musashi Institute of
Electronics Packaging in 2006 for the paper entitled Technology, Yokohama, Japan, and splits his time between
“Consequential Toxicity Assessment of the Global Shift to Musashi Institute of Technology and AIST. E-mail: itsubo-
Pb-free Solder Paste”. Anders S.G. Andrae is the corresponding n@yc.musashi-tech.ac.jp; itsubo-n@aist.go.jp
author and can be contacted at: a.andrae@aist.go.jp
Atsushi Inaba received his BSc, MSc, and
Norihiro Itsubo received his BSc degree from PhD degrees in Chemical Engineering from
Osaka University in Osaka, Japan, and received Tokyo University, Tokyo, Japan, in 1976, 1978
his MSc and PhD degrees from University of and 1981, respectively. Between 1981 and 1986
Tokyo, Tokyo, Japan, in 1993, 1995 and 1998, he was at the National Institute for Resources
respectively. Between 1998 and 2001 he was at and Environment (NIRE) and between 1984 to
the LCA Development Department, Japan
1986 at the National Bureau of Standard in the
Environmental Management Association For
USA. Between 1990 and 1992 he was at the International
Industry. In 2001 he became a Research Scientist at AIST’s
Research Center for Life Cycle Assessment (AISTLCA), Institute for Applied Systems Analysis, Vienna, Austria before
Tsukuba, Japan. In 2003 he was awarded by the Reliability being appointed Chief of the NIRE Planning Office where he
Engineering Association of Japan for his work “LCA of IC was from 1999 to 2001. In 2001 he was appointed Director of
packages”. Between 2003 and 2005 he was a researcher for AIST’s Research Center for Life Cycle Assessment, Tsukuba,
the Environmental Assessment Research Team at AISTLCA, Japan. In 2005 he was appointed Professor for Research into
and since 2005 he has been Team Leader for the LCA Artifacts at the Center for Engineering, The University of
Methodology Research Team at AISTLCA. In 2005 he was Tokyo, Tokyo, Japan, and splits his time between the Director
appointed Associate Professor at the Environmental and and Professor roles. E-mail: a-inaba@aist.go.jp

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com


Or visit our web site for further details: www.emeraldinsight.com/reprints

28

Das könnte Ihnen auch gefallen