You are on page 1of 11

Resources, Conservation and Recycling 54 (2010) 256–266

Contents lists available at ScienceDirect

Resources, Conservation and Recycling


journal homepage: www.elsevier.com/locate/resconrec

Facility-level energy and greenhouse gas life-cycle assessment of the global


nickel industry
Matthew J. Eckelman ∗
Center for Industrial Ecology, Program in Environmental Engineering, Yale University, 195 Prospect Street, New Haven, CT 06511, United States

a r t i c l e i n f o a b s t r a c t

Article history: Nickel is an integral material to our modern, high-performance technological society. With increasing
Received 4 August 2008 emphasis being put on energy efficiency and global climate change, it is important for companies to
Received in revised form 14 August 2009 understand in detail the energy use and greenhouse gas implications of their business. The present
Accepted 18 August 2009
analysis is a facility-level life-cycle assessment of these twin impacts covering the entire global nickel
industry. Cradle-to-gate results (including extraction, production, and fabrication) are presented here
Keywords:
for selected nickel and nickel alloy products, including upstream energy required for fuel production.
Nickel production
Stainless steel is one of the most highly recycled metals in the world. In order to assess the energy and
Ferronickel
Nickel pig iron
carbon implications of secondary material use, recycling scenarios for three grades of stainless steel
Nickel oxide (AISI 304, 409, and 430) were considered. Using the current scenario as a baseline, maximum use of
Stainless steel scrap (within technical limits) and all-virgin production results varied widely. Smelting/Class II refining
Life-cycle assessment was the most energy intensive step of production, accounting for 50–90% of total primary energy use.
Primary energy Transport contributed 2–11% of the total, depending on the nickel product considered. A sensitivity
Greenhouse gas emissions analysis revealed that the results are highly dependent on the energy requirements for upstream fuel
production, which apply to all steps of the assessment. These results will help the nickel industry
navigate energy and climate change concerns in the coming years.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction lion metric tons), more than 100 times greater than production
levels of a century before (Harper et al., 2006; Kuck, 2006).
Nickel is in many ways a metal of affluent societies. It is not nec- Nickel is also quite energy intensive to produce, as the results
essary for the construction of large structures, as iron and steel are, here will show. As international discourse and policy on energy
or for electrical conduction, as copper and aluminium are. Its major consumption, greenhouse gas emissions, and climate change devel-
uses are for high-performance technological products, which take ops, large industrial facilities such as nickel smelters and refineries
advantage of its valuable characteristics such as high heat resis- will feel financial, regulatory, and public pressure to act respon-
tance, strength, and durability. That said, nickel and its primary sibly. Most large companies now produce annual reports that
alloy, stainless steel, provide such utility that their presence in mod- attempt to quantify the energy and carbon impacts of their opera-
ern technological society is pervasive, from cutlery to industrial tions, and detail plans and programs that further sustainability in
reactor tanks to jet engines. the metals sector. The question of which metrics to use to evalu-
The life cycle of nickel, as for most metals, is relatively com- ate sustainability is under debate; for example, the International
plicated: even just for nickel fabricated semi-products (prior to Council for Mining and Minerals has developed a set of Principles
manufacturing into final products), there are numerous grades of that members must implement (ICMM, 2008).
metal with varying compositions and properties. Nearly 70% of Because of its large environmental footprint, metal mining and
nickel is used for stainless steel fabrication. Nickel is also used production has been a focus of energy analysis and life-cycle assess-
in non-stainless alloy steels (7%), nickel- and copper-based alloys ment for several decades. Chapman and Roberts (1983) provided
(11%), nickel plating (6%), foundry (3%), and other specialty uses an extensive treatment of energy models for metal production and
(4%) (Reck et al., 2008). In 2005, world production was 1.3 Tg (mil- recycling, though many of their results are now outdated. Kusik
and Kenahan (1978), working for the then U.S. Bureau of Mines
(now part of the U.S. Geological Survey), gave energy use results
for primary and secondary production for many metals that showed
∗ Tel.: +1 203 432 4985; fax: +1 203 432 5556. explicitly the energy benefits of recycling. Yoshiki-Gravelsins et al.
E-mail address: matthew.eckelman@yale.edu. (1993) updated this work for five base metals, followed by Norgate

0921-3449/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.resconrec.2009.08.008
M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266 257

such as transport energy intensity. As often happens, significant dif-


ferences emerged between these studies, as additional data were
considered and incorporated.
The uncertainty of LCA results can be reduced by incorporat-
ing more detailed inventory data. The purpose is to provide an
updated and extremely detailed accounting of the energy use and
greenhouse gas impacts of nickel production worldwide, covering
all major facilities and representing nearly 100% of world produc-
tion. This study is based on publicly available facility data, which
can have their own limitations, including out-of-date information,
aggregation of process steps, ambiguous treatment of co-products,
and exclusion of low-volume inputs. Therefore, while the present
study is more comprehensive in its coverage of facilities than those
Fig. 1. Final energy intensities of several metals. previous, the lack of up-to-date survey information introduces
Adapted with permission from Barkas (2009). some uncertainty into the results.
Notes: Average production intensities are indicated by vertical line segments; Nickel
pig iron production is not included.
2. Methods
(2004) among many others. Fig. 1 shows ranges of final energy
Current processing of nickel ores utilizes an assortment of tech-
intensities for primary nickel and several base metals, which clearly
nologies, both pyro- and hydrometallurgical, with much more
shows the wide variations in global metals production.
variety than for base metals such as copper or aluminium (Boldt,
Other researchers have attempted to look at the entire metals
1967). Nearly every facility is unique in its process flows. Even in
sector as a system, with links among metals at all stages of the life
facilities that utilize similar technologies, the primary fuels used are
cycle (Reuter, 1998; Verhoef et al., 2004). This enabled them to look
often different, particularly in the on-site or grid mix of electricity.
for optimization opportunities through the use of system dynamics
Though there are a small number of large ore producers, such as
tools. Energy use specifically in nickel production has been studied
Russia, Canada, and Australia, ore is mined over a wide range of
previously, although not as extensively as base metals such as cop-
grades, from an average of 6.5% at the new Cosmos mine in Aus-
per and usually as part of a large compilation, such as Boustead
tralia to 0.1% at the nickel-containing talc mines of Finland or the
and Hancock (1979). Academic and industry reports have been
platinum group metal (PGM) mines of southern Africa. Some nickel
produced, usually in countries that have significant stainless steel
complexes are integrated to include mining, smelting, and refining,
industries, such as Germany (Adelhardt et al., 1998). Environmen-
while others perform only one step and sell their products on for
tal impacts other than energy and greenhouse gas emissions have
additional processing. Joint ownership of many facilities among the
generally been considered individually, such as water (Norgate and
large nickel companies has resulted in a complicated web of mate-
Lovel, 2004), land use, and toxicity (Gustavson et al., 2007). Social
rial flows as refined nickel producers struggle to ensure consistent
impacts of mining in particular are another focus of serious concern
supply. All of this variety makes it difficult to describe an “average”
and research, but these are not considered here. With the develop-
process for the production of nickel from the perspective of energy
ment and standardization of life-cycle assessment (LCA), data on
use and greenhouse gas emissions.
the environmental impacts of nickel and other metals over their
In 2005, there were approximately 66 active nickel-producing
life cycles are available from several life cycle inventory databases,
mines, over 50 smelters and Class II processing facilities, and 25
such as Ecoinvent (2007) and JEMAI (2007). Many LCA studies have
Class I refineries in operation. (Since that time, several additional
been conducted for metals and metal-containing products, though
production facilities have come on-line.) The defining character-
only a very few have dealt specifically with nickel (Althaus and
istic of the present study is that it accounts for energy use and
Classen, 2005; Forbes et al., 2000; Narita et al., 2001; Rydh and
greenhouse gas emissions on a facility level instead of using average
Karlstroem, 2002; Sagisaka et al., 2007).
values. As described in detail below, this assessment incorporates
Concerns of the energy and carbon implications of the nickel
data on the ore grade and type, processing technology, fuel types
industry have increased in recent years with the re-introduction of
(including for electricity generation), geographical location, prod-
nickel pig iron from China for use as a low-grade ferronickel substi-
uct quality, and co-products of every nickel production facility in
tute. Nickel pig iron was produced in Europe several decades ago,
the world. The methodology used here resembles that described
before the introduction of new pyrometallurgical processes that
by Schuller et al. (2008) in their investigation of copper production
were more energy efficient and produced higher grade ferronickel.
in Chile and Germany, though the present study has a much larger
In China, nickel pig iron is typically produced in small-scale blast
scope.
furnaces that had been taken out of carbon steel production for eco-
The functional unit for the assessment is one ton of nickel or
nomic and environmental reasons, though the recent ramp up in
nickel alloys. Results are given for Class I nickel (refined to greater
production has brought several electric arc furnaces on line as well.
than 99% purity), Class II nickel (refined to less than 99% purity, such
Chinese demand for nickel in stainless production has exploded
as ferronickel and nickel oxide sinter), as well as several grades of
in recent years, as with most metals, as production levels have
stainless steel, which is the most important first use of nickel. The
increased nearly 10 times since the start of the decade. Chinese
products considered here are listed in Table 1.
domestic demand for nickel has also increased dramatically dur-
ing that period, from approximately 50 Gg (kt, or thousand metric
tons) Ni in 2000 to over 350 Gg Ni currently. 2.1. System boundary
In an effort to quantify the environmental impacts of nickel
specifically, the international Nickel Institute commissioned a A complete energy life-cycle assessment should track the direct
series of life-cycle assessment studies (Ecobalance, 2000, 2003). and indirect energy inputs into all processes and stages of the
These studies were based on survey data and as such were restricted product life cycle. Indirect inputs are those that are upstream of
to those facilities that opted to participate. There was also a high a process, such as the energy needed to construct the smelting
sensitivity to certain parameters that were used across facilities, facility itself. In practice, including all indirect inputs is extremely
258 M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266

Table 1 products of nickel smelting and refining, namely copper matte and
Nickel and nickel products considered in the present study.
blister copper, cobalt compounds, and sulfuric acid, are included in
Product name % Nickel Notes the analysis. Following previous studies (Ecobalance, 2003), alloca-
Class I (refined) nickel >99% tion of energy use and greenhouse gas emissions is done by mass
Ferronickel 15–45% for copper and cobalt, and by system boundary expansion for sulfu-
Nickel oxide sinter 75–78% ric acid. As all nickel production facilities were included, no cut-off
Nickel pig iron 1.5–8% From blast furnaces criteria were used when considering global production. All results
8–17% From electric arc furnaces are in terms of primary energy and mass of CO2 equivalent, and
Copper–nickel matte 40–80%
are specific to the on-site or local grid sources of electricity. Lower
Stainless steel 304 8% heating values for fuels are used throughout. Data sources for each
Stainless steel 409 0.1–0.5% process step are documented below.
Stainless steel 430 0.1–0.5%

Source: International Nickel Study Group (2006). 2.2. Mining, milling, and beneficiation

Mine production statistics are generally widely published; how-


difficult without using gross average data from life-cycle inven- ever, there are often inconsistencies across sources. A number of
tory databases or environmental input–output (EIO) tables. As the sources were consulted in order to create a table of nickel mines
present study focuses on energy and greenhouse gas emissions, that included nickel and/or ore production tonnages, ore grade and
the upstream energy required for fuel production is included in the type, geographical location, and mining technique. These included
analysis. This is the only category of upstream (or indirect) inputs Brook Hunt (2007); United States Geological Survey (Kuck, 2006);
that is included explicitly. A schematic diagram of the nickel pro- Natural Resources Canada (McCutcheon, 2005); the International
duction system and system boundaries used here is shown in Fig. 2. Nickel Study Group (2006); and Raw Materials Group (2007).
The system includes the production of selected nickel alloys, such Electricity requirements for hard rock mining are generally
as stainless steel. much higher for underground mining than for open-pit mining
Production, trade, and energy data are all for 2005, or the clos- because of the need for drilling, ventilation, water removal, and
est available year. Losses are considered during mining, milling, hoisting of rock to the surface. Open-pit mines generally consume
and beneficiation, smelting, refining, and scrap preparation. Trans- more liquid fuel because of the use of mining trucks to transport
portation is considered between each step and is assumed to be rock up from the pit. For the present study, consumption of diesel
100% material efficient (that is, no nickel is lost). The major co- fuel and electricity was derived from a previous study that grouped

Fig. 2. Nickel production system boundaries.


Notes: FeCr = ferrochromium; DRI = direct reduced iron; FeNi = ferronickel; NiO = nickel oxide sinter; EW = electrowinning.
M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266 259

Table 2 • Krupp-Renn direct reduction of lateritic ores into ferronickel


Average final energy consumption values for nickel mining processes.
(only at the Oheyama smelter, Japan).
Process Liquid fuels Electricity • HPAL hydrometallurgy, where an acid solvent is applied to lat-
(MJ/ton ore) (MJ/ton ore) erite ores that dissolves much of the nickel.
Open pit mining 141 60 • The Caron process, where ammonia is applied to roasted cal-
Underground mining 66 149 cine and nickel and cobalt are extracted electrolytically from the
Milling and beneficiation 0 200 resulting liquor.
• Pyrometallurgy of sulfidic ores using Direct Outokumpu Nickel
(DON), conventional, and Inco Oxygen flash furnaces to produce
different mining regions according to ore type and grade (Adelhardt
nickel matte.
et al., 1998) and a pair of benchmarking studies conducted in • Nickel matte production using a reverberatory furnace (only at
Canada for gold, copper, iron, lead, and zinc (Mining Association
the Chengdu smelter, China).
of Canada, 2005a,b). Older data that cover a wide range of mine • Blast furnace (BF) and EAF production of nickel pig iron (only in
types and metals can be found in Chapman and Roberts (1983).
China) and EAF production of secondary nickel pig iron (7.9% Ni,
Mining also uses large quantities of explosives, which themselves
13.8% Cr, 70.3% Fe, only at the International Nickel Reclamation
contribute to primary energy use and more significantly to green-
Company, United States).
house gas emissions. Here, the treatment of explosives follows that
of the U.S. Department of Energy (ITP, 2002) with the assumption This list corresponds for the most part with the processes con-
that 1.1 kg of explosives are required to mobilize one ton of ore and sidered in the nickel LCA conducted by Ecobalance (2003). Newer
concentrate (Johnson et al., 2008). processes such as atmospheric heap leaching and bio-leaching were
Milling and beneficiation also consumes large quantities of elec- not significant in 2005 and so are not included here.
tric power, as large boulders must be crushed down to the size As with mine production, a production table of intermediate
of the mineral grains. Electricity consumption therefore depends products from nickel smelters was created using the same sources
crucially on this grain size, as well as the rock hardness, number listed above, with additional data derived from company and trade
of screenings, and factors around the particular process of bene- websites. This table includes production statistics (in contained
ficiation. In general, milling and beneficiation are applied only to nickel), nickel content in the final product, feed ore/concentrate
sulfide ores, though some laterite ores do get milled and most go type, and process used.
through a screening process to remove large boulders. The use of Energy is consumed for different purposes depending on the
flotation agents in beneficiation leads to direct emissions of CO2 as process. Pyrometallurgical processing of laterite ores has high fos-
many agents contain carbon; accounting of these direct emissions sil and electrical energy requirements, as laterite ores have a high
assumes an emissions rate of 18 kg CO2 per ton nickel concentrate water content and the presence of magnesia leads to high melt-
(Ecobalance, 2003). ing temperature. For most pyrometallurgical processes, fossil fuel
Average final energy consumption values are shown in Table 2. is consumed both as a source of heat and as a reductant. HPAL pro-
Final energy refers to that which is utilized at the facility level and is cessing requires energy to heat the acid solvent and to pressurize
generally much lower than the total upstream (or primary) energy. the chamber. Electric arc furnaces consume significant amounts
of electricity, and also use carbonaceous electrodes and electrode
2.3. Smelting/processing of Class II nickel paste, which produces CO2 directly as it is consumed (these emis-
sions were assumed to be 0.3 tons CO2 per ton contained nickel,
Nickel smelting is the most complicated stage of the global Sjardin (2003)).
nickel life cycle, and the most important in terms of energy con- Energy use data were derived on a facility level using sev-
sumption and greenhouse gas emissions. Its complexity stems from eral global nickel smelter surveys (Bergman, 2003; Warner et al.,
the range of technologies used as well as the variety of intermediate 2006, 2007), research reports (Adelhardt et al., 1998; Kerfoot,
nickel products that are outputs of the smelting process, including 2002), energy analysis research (Boustead and Hancock, 1979),
ferronickel, nickel matte, nickel oxide sinter, and nickel pig iron. life-cycle inventory databases (Argonne National Laboratory,
In past energy analysis research on nickel, two processes have 2008; Ecoinvent, 2007; Japanese Environmental Management
generally been considered: (1) pyrometallurgy of sulfide ores Association for Industry 2007), and company reports and websites.
using flash furnaces and (2) hydrometallurgy of lateritic ores Information regarding energy consumption for Chinese nickel pro-
using high-pressure acid leaching (HPAL) and solvent extrac- duction facilities was particularly difficult to gather (especially for
tion/electrowinning (SX/EW) (Norgate, 2004). There are, however, nickel pig iron) and various domestic and international sources
many other important processes that are used in the smelting and were consulted (Lennon, 2007; Non-ferrous metals industry of
processing of nickel. The processes included in the present analysis China 2005; Xu, 2007; Xu et al., 2004; Yu and Chen, 2007). Where
are as follows: specific data were not available for the smaller nickel pig iron pro-
ducers, slightly dated blast furnace energy consumption was used
• Drying, calcining, and electric arc furnace (EAF) smelting of lat- as a proxy for the older Chinese blast furnaces that have been con-
erite ores for the production of ferronickel, and a small amount verted to nickel production (Brown, 1996). Process information for
of matte. INMETCO’s nickel recovery operation was gathered from Lankey

Table 3
Representative final energy consumption data for nickel smelting and Class II production.

Process (product) Fossil fuels (GJ/t product) Electricity (GJ/t product) Facility

EAF (FeNi) 26.6 25.1 Hachinohe


Flash (Matte) 21.0 16.0 Fortaleza
EAF (Matte) 9.6 14.1 INCO Sudbury
BF (Ni pig iron) 117.3 2.1 Huaguang
Caron (NiO) 405.6 21.6 Nicaro

Note: These data differ from energy consumption per ton of contained Ni, due to the differing Ni contents of each product.
260 M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266

Table 4 consumption during electric arc furnace fabrication of stainless


Average final energy consumption values for Class I nickel refining processes.
steels followed Johnson et al. (2008). EAF was the only process
Process Fossil fuels Electricity technology considered.
(GJ/t Ni) (GJ/t Ni) There are many non-nickel inputs into stainless steels, as listed
Electrorefining 0 15.3 in Table 5. Energy consumption for many of these intermediate
Ammonia pressure electrowinning 5.7 7.2 products again followed Johnson et al. (2008), updated for 2005
Sulfate/chloride electrowinning 0.7 9.2 production statistics. Updates are important because the relative
Caron 0 9.0
output of producing countries changes the average electric effi-
Carbonyl 13.2 3.2
ciency used in metal production, which in turn affects primary
energy use in fuel type and quantity.
and McMichael (1999) and company websites. Table 3 shows a rep-
resentative list of final energy requirements at different types of 2.6. Transportation
processing facilities.
The transport of intermediate products represents a signifi-
2.4. Class I refining cant and often over-simplified portion of overall primary energy
consumption for metals processing and production. Traded prod-
Additional refining to remove the impurities remaining after ucts included in this analysis were chromite, nickel, and iron ores
smelting or extract nickel from solution is necessary to produce and concentrates, ferrochromium, ferronickel, nickel oxide sin-
Class I (>99% purity) nickel. As with smelting, there are several ter, nickel pig iron, nickel matte, refined nickel, direct reduced
processes that are used, including electrorefining, electrowinning, iron (DRI), and scrap of all kinds. A trade matrix was created
hydrogen sulfide reduction (Sherritt), Caron, and carbonyl (Mond) for each of these products, derived from global trade statistics
processing. The majority of Class I nickel in 2005 was produced for 2005 (International Nickel Study Group, 2006; Kuck, 2006;
from matte, though the future dominance of laterites as a source of UNCOMTRADE, 2008). Distances between each pair of shipping and
nickel ore means that HPAL and related technologies will be much freight points were derived from global shipping tables and maps.
more important in the future (Dalvi et al., 2004). While Ni–Cu matte Energy intensities (or energy use per ton-km) for transport by ship
and other intermediate products are usually traded and trans- and train were taken from the Energy Information Administration
ported among different countries before refining – nearly 300 Gg of (U.S. Department of Energy, 2006) and an average value of energy
traded matte were recorded in 2005 (United Nations Department of consumption per ton of traded product was calculated. These aver-
Economic and Social Affairs/Statistics Division 2008) – the Murrin age values are shown in Table 6. It was assumed that only liquid
Murrin complex in Australia engages in hydrometallurgical Class I fuels (diesel and bunker fuel oil) were consumed. Adjustments were
refining of laterite in a vertically integrated nickel processing com- made to account for intermediate products that are not traded
plex. internationally but are transported internally among production
Again using global production data, a table of refineries was facilities in a single country.
created that shows production statistics, geographical location, Chromite and nickel ore have relatively low transport-related
and process technologies for every Class I nickel refinery in the energy consumption figures because smelters are often located in
world. Sources included Brook Hunt (2007) and the Raw Materials mining regions. Nickel pig iron is assumed to be used entirely in
Group (2007). Facility-level energy consumption data were gath- the Chinese stainless steel industry.
ered mainly from research reports (Adelhardt et al., 1998; European
Integrated Pollution Prevention and Control Bureau 2006; Johnson 2.7. Electricity and upstream energy
et al., 2008; Norgate and Rankin, 2002). Soda ash (Na2 CO3 ) is also
used in electrorefining for purification of the electrolyte, which In order to convert from local electricity consumption to primary
results in direct CO2 emissions. These were assumed to be approx- energy use, the effective electrical efficiency was derived for each
imately 1.0 ton CO2 per ton refined nickel product (Ecobalance, country (and individual facilities, where possible) using data from
2003; Falconbridge Limited et al., 1999). Average final energy con- the International Energy Agency (IEA) and company reports and
sumption values are shown in Table 4. websites (IEA, 2007a,b). Electric efficiency is the ratio of electrical
energy produced by power plants to the fuel energy that enters.
2.5. Stainless steel fabrication Electricity production by autoproducers (facilities that generate
their own electric power) and combined heat and power (CHP)
Country-level production data for austenitic and ferritic stain- plants were included in this calculation (hence the term “effec-
less steels for 2005 were gathered from Brook Hunt (2007). Energy tive”). The hybrid substitution method was used, meaning that

Table 5
Energy consumption of non-nickel input processes to stainless steel fabrication (per ton product).

Process input to stainless steel Direct fossil Direct Electric Upstream Upstream fuel Total primary
fuel (MJ/ton) electricity efficiency electricity prod (MJ/ton) energy
(MJ/ton) (MJ/ton) (MJ/ton)

Chromite mininga 41 180 44% 411 20 471


Ferrochromium productiona 14,600 13,300 44% 30,600 2300 47,500
Iron mininga 90 12 53% 24 19 133
Direct reduced iron productiona 11,300 360 49% 700 3000 15,000
Scrap processinga 0 25 40% 61 6 67
Molybdenum productionb – – – – – 148,000
Manganese productionb – – – – – 58,200
Titanium productionb – – – – – 644,000

Note: Total primary energy is the sum of direct fossil fuel use and upstream energy for electricity and fuel production.
a
Source: Johnson et al. (2008).
b
Source: Ecoinvent (2007) (total primary energy only).
M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266 261

Table 6
Average energy consumption for transport of intermediate products.

Intermediate product Liquid fuels (MJ/ton) Upstream (MJ/ton) Total (MJ/ton)

Chromite ore 593 119 712


Ferrochromium 2735 547 3282
Nickel ore 245 49 294
Ferronickel 2115 423 2538
Nickel oxide sinter 959 192 1150
Nickel pig iron 70 14 84
Nickel matte 732 146 878
Refined nickel 1724 345 2069
Iron ore 1256 251 1508
DRI 318 64 382
Scrap metal 1807 361 2169

hydropower is assumed to be 100% efficient. In addition to the them into iron oxide and SO2 . In many metal production facili-
electric efficiency, the local grid mix of fuel sources for electric- ties, this concentrated stream of SO2 is sent to an acid plant for
ity was derived, and these were combined to create a matrix that additional conversion to sulfuric acid. The amount of sulfuric acid
returned consumption of primary energy (coal, oil, gas, etc.) per generated at major nickel smelters was gathered from an industry
unit of electricity consumption. These facility level primary energy survey (Warner et al., 2007) and company reports and websites.
values were weighted by the share of global production at each The global weighted average co-production of sulfuric acid is 2.8
facility to find the weighted average for primary energy consump- t/t Ni in matte. The allocation method used was system bound-
tion for each product. The electricity conversion matrix for nickel ary expansion, to include conventional production of sulfuric acid.
oxide sinter production is shown in Table 7. The upstream energy This is significantly higher than the value of 1.5 t/t Ni given in the
used for the extraction, preparation and refining, and transporta- Ecoinvent LCI database (Ecoinvent, 2007). The primary energy con-
tion of primary fuels was derived from life-cycle assessment studies sumption for conventional sulfuric acid production was taken to
for coal (Spath et al., 1999), oil (Sheehan et al., 1998), natural gas be 2.6 MJ/kg acid (Boustead and Hancock, 1979); the energy credit
(Spath and Mann, 2000), and nuclear fuel (ERDA, 1976). Upstream was applied to the smelting process.
energy comprises 15–30% of total primary energy consumption for Cobalt is an important co-product of nickel production; refined
most processes considered here. cobalt is a product of nearly all Class I nickel refineries. Cobalt
production statistics from nickel refineries were gathered from
2.8. Co-products and allocation the United States Geological Survey (Shedd, 2006) and com-
pany reports and websites. Following previous nickel LCA studies
Three major co-products of nickel production were considered (Ecobalance, 2003), allocation for cobalt was made by mass. The
here: copper, cobalt, and sulfuric acid. These are produced at dif- global weighted average percent of cobalt of the total useful metal
ferent stages of the nickel production cycle and slightly different output of nickel refineries was 4%; this energy credit was applied
allocation procedures were used for each. to the nickel refining process.
Copper is often co-mined with nickel in large sulfide deposits
and smelted into a mixed-metal matte, which also usually con- 2.9. Scrap
tains a small amount of cobalt. These mattes have highly variable
compositions from facility to facility; the ratio of nickel to cop- Stainless steel is one of the more highly recycled metals; the sec-
per ranges from 0.8 at Canada’s Inco Sudbury smelter to nearly 40 ondary nickel content of stainless steel production was more than
at Australia’s Kalgoorlie smelter (Warner et al., 2007). The global 40% in 2000 (Reck et al., 2008). Non-stainless nickel scrap, how-
weighted average percent of nickel in the metal contained in matte ever, is much less widely used and constitutes appreciable inputs
is 70%. Following previous nickel LCA studies (Ecobalance, 2003), to production in only a handful of countries. Some LCA research
allocation for copper was made by mass. Therefore, a credit of 30% of has been conducted on the recovery of non-stainless nickel from
the matte smelting energy is applied to copper production. For the certain waste streams, such as plating sludge or nickel-hydride bat-
mining stage, energy use at nickel-producing PGM mines has been teries, but the results are restricted to Japan (Sagisaka et al., 2007).
allocated according to value, as these mines are primarily driven The International Nickel Reclamation Company (INMETCO) recov-
by the economics of gold and PGM extraction. ers nickel in relatively pure form from Ni–Cd batteries, but these
Sulfur dioxide (SO2 ) is produced in prodigious quantities at plates are added to its nickel pig iron melt for use in stainless steel
metal production facilities that process sulfide ores. Oxygen is manufacturing.
blown through nickel-containing mattes in converter plants in Nickel scrap inputs to Class I refining were assumed to be 0.7% of
order to strip the iron and sulfur contained there and convert the total, with a maximum content of 5% (International Nickel Study

Table 7
Electricity conversion matrix for nickel oxide sinter production (units per unit electricity).

Smelting facility Prod. kt NiO Coal Oil Gas Nuclear Hydro Geo Bio Total

Philippines—Coral Bay 5.2 0.93 0.27 0.55 0.00 0.15 1.75 0.00 3.65
Australia—Yabulu QNI 4.4 2.28 0.03 0.37 0.00 0.06 0.00 0.06 2.79
Canada—Inco Sudbury 41 0.44 0.08 0.13 0.45 0.58 0.00 0.04 1.71
Cuba—Nicaro 12 0.00 1.86 0.00 0.00 0.01 0.00 0.50 2.37
Cuba—Punta Gorda 31 0.00 1.86 0.00 0.00 0.01 0.00 0.50 2.37
Japan–Tokyo Ni 56 0.70 0.21 0.49 0.88 0.07 0.03 0.04 2.42

Weighted average 0.48 0.89 0.28 0.43 0.05 0.10 0.23 2.46
% of Total 19% 36% 11% 18% 2% 4% 9% 100%
262 M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266

Table 8 oxygen through nickel matte can also produce nickel oxide by the
Average energy consumption for transport of intermediate products.
oxidation of Ni3 S2 , which takes place in a roasting furnace. Nickel
Process Losses oxide is again reduced directly with carbon to form CO2 , though
Nickel mining and separation 20% of nickel mined reduction with hydrogen is also common (Boldt, 1967). Carbon
Ferronickel production 5% of nickel dioxide emissions from direct reduction of nickel oxides during
Nickel oxide sinter production 5% of nickel smelting and refining have been included in the analysis, based
Nickel pig iron production 8% of nickel on stoichiometric calculations.
Nickel matte production 5% of nickel
Refined nickel production 0.5% of nickel
Chromite mining and separation 17% of chromite mined 2.12. Scenarios
Ferrochromium production 20% of chromium
Iron mining and separation 19% of iron mined For each of the nickel products included in the analysis (Table 1),
DRI production 0.4% of iron
three scenarios were considered, following Johnson et al. (2008):
Scrap metal preparation 1% of scrap collected
‘current operations’, reflecting real production material flows for
Stainless steel production 0.7% of nickel 2005; ‘maximum recycling’, reflecting production using the maxi-
1.4% of chromium
mum technologically feasible quantities of secondary material; and
1.8% of iron
0.5% of molybdenum, ‘all virgin’, which includes no secondary material. Recycling statis-
manganese, titanium tics for 2005 were gathered from industry reports (Brook Hunt,
2007; International Nickel Study Group, 2006) and extrapolated
from Reck et al. (2008).
Group, 2008). Secondary inputs of various scrap grades to stainless
steel fabrication for austenitic and ferritic steels were ascertained 2.13. Sensitivity/error analysis
from the International Stainless Steel Forum (2004).
Energy used for scrap collection and preparation is generally dif- In order to assess the sensitivity of the present analysis to certain
ficult to ascertain with any precision, due to the quickly changing input variables, each major variable was increased in isolation by
nature of the scrap market, lack of robust trade statistics, and highly 10% to examine the effect on total primary energy use and green-
disaggregated nature of local scrap industries. Energy consumption house gas emissions. Ferronickel and Class I refined nickel were
was assumed to be 67 MJ/ton scrap for processing and 2170 MJ/ton considered. Comparable results for austenitic (304 grade) stainless
scrap for collection and transportation, following Johnson et al. steel can be found in Johnson et al. (2008). It is very difficult to
(2008). assign accurate quantitative uncertainties to the input variables, so
this analysis does not include any stochastic models of uncertainty
2.10. Losses (such as Monte Carlo simulation).

Production losses of nickel and alloying elements increase the 3. Results and discussion
amount of primary material and energy required to produce a unit
of final product. Elemental losses from the various stages of pro- The energy intensities of nickel and nickel products are highly
duction are shown in Table 8. Losses were derived from Johnson et variable, depending on raw material sources, process routes, and
al. (2008) and Reck et al. (2008). product specifications. Fig. 3 shows average results by final product.
For all nickel products, the smelting and refining (production)
2.11. Greenhouse gas emissions steps consume the most primary energy and are responsible for the
highest greenhouse gas emissions of any category. Transportation
Fossil-based energy use is considered as an environmental accounts for 2–11% of energy use while mining and concentrating
impact category as it results in the depletion of non-renewable account for 7–35%, depending on the nickel product. Considering
resources, as well as emissions to the atmosphere. Here, we con- fuel type, coal/coke contributed the most primary energy of any
sider only those emissions that contribute to global climate change fossil fuel type for each nickel product. Coal/coke is used for heat-
(greenhouse gases). Direct fuel use, upstream energy, and fuel use ing and directly in smelting as a reductant, and it is the source
for electricity generation were combined for each process to deter- of a large portion of the electricity grid in many countries with
mine the amount of each fuel used per ton of product. This result large nickel industries, such as Australia. While some smelters are
was coupled with carbon intensity figures (mass of carbon dioxide located near hydroelectric installations that provide cheap, clean
emissions per ton product) from the Energy Information Adminis- power, such as the Xstrata refinery in Norway, the overall contribu-
tration (U.S. Department of Energy, 2006) to give carbon dioxide tion of renewables (hydro, biomass, and geothermal energy) to the
emission factors for each process. global production of nickel is at most 7%. The majority of primary
There are many material inputs to nickel production that cause, energy is used directly, that is, as fossil fuel during the production of
either directly or indirectly, additional greenhouse gas emissions. nickel. For nickel matte, however, the prevalent processes of elec-
Carbon-containing materials can evolve CO2 gas through their use. trorefining and electrowinning require much more electricity than
Several such materials have been included in the analysis, includ- they do direct fossil fuels. Ferronickel also has a large portion of its
ing explosives, flotation agents, electric arc furnace electrodes primary energy requirements due to electricity use in electric arc
and paste, and soda ash; other minor input materials have been furnaces.
excluded, as noted. 2005 production levels were 1100 Gg of ferronickel, 110 Gg of
There are also greenhouse gas emissions that result directly nickel oxide, and 880 Gg of Class I nickel. Overall, the produc-
from the processing of metal ores. These occur primarily from tion of these materials consumed 337 PJ (1015 joules) of primary
the pyrometallurgy of lateritic ores. Much of the nickel in lat- energy and contributed 27 Tg (million metric tons) of carbon diox-
erite ores occurs as nickeliferous limonite of the chemical form ide equivalents to the atmosphere.
(Fe,Ni)O(OH). During the ore preparation and smelting stages, the None of the intermediate nickel products contain much sec-
hydroxyl groups in the ore decompose and reform as water, in a ondary nickel and so scrap collection and processing represents
process known as dehydration. The remaining oxygen is reduced an insignificant contribution to total energy consumption. Because
directly during refining to form carbon dioxide. The blowing of relatively little Class I nickel scrap is available for recycling and few
M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266 263

Fig. 4. (a) Primary energy consumption for nickel mining and beneficiation in 2005.
(b) Primary energy consumption for Class II nickel smelting and refining.
Fig. 3. Primary energy consumption for the production of one ton of nickel product, (c) Primary energy consumption for Class I nickel refining (excluding smelting).
broken down by (a) process, (b) fuel type, and (c) energy stage.

underground while laterite ores are amenable to less-energy inten-


facilities are designed to accommodate secondary material, there is sive surface mining operations. Sulfide ores undergo concentration,
not much opportunity to realize significant energy and greenhouse while laterite ores are only screened. Laterite ores, however, are
gas savings through substitution for virgin material. also generally dried from 30% moisture to 11–17% moisture, which
These average results mask a wide variation in energy con- requires heavy use of fossil fuels.
sumption at the facility level. On a contained nickel basis, energy Nickel occurs in natural deposits at much lower concentrations
consumption for nickel production spans a wide range of values, as than many other base metals, such as iron and aluminium, and so
seen in Fig. 4a–c, covering (a) mining and beneficiation, (b) Class II nickel production requires correspondingly high volumes of rock
smelting and processing, and (c) Class I refining. All figures reflect that must be mined and processed. As clearly shown in Fig. 4a,
cumulative production, ordered on the x-axis by the energy con- ore grade is the most significant determinant of energy use per
sumption per ton contained nickel at each facility. unit of contained nickel. Barring discovery of spectacularly rich new
The different mining and beneficiation practices for under- nickel deposits, it is certain that the energy requirements for nickel
ground and open-pit mines, as well as for sulfide and laterite ore mining, milling, and beneficiation will continue to increase as
ores, result in a wide range of energy requirements, as shown in average ore grades decrease and more rock needs to be processed
Fig. 3a. The primary energy requirements for sulfide ores are par- for an equivalent amount of metal.
ticularly variable, ranging from 60 GJ/t contained Ni to 1300 GJ/t It is clear from Fig. 4b that there is a marked difference in the
contained Ni. There are no clear patterns in the energy require- energy requirements for the five Class II and intermediate nickel
ments for sulfide versus laterite ores, as energy use in the different products considered in this report. Please note that both final
processing steps for the two ore types roughly balance each other Class II products and intermediate products that will undergo fur-
(again, on a contained Ni basis). Sulfide ores are generally mined ther refining to Class I nickel.are included. Nearly all nickel matte
264 M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266

Table 9a
Effect of a 10% increase in model parameters on total results: Ferronickel.

Process Primary Greenhouse


energy use gas emissions

Upstream energy of coal 1.9% 1.9%


Open pit mining liquid fuels use 0.7% 0.8%
Cerro Matoso smelter fossil fuel use 0.6% 0.6%
Donaimbo smelter electricity use 0.5% 0.5%
Nickel content of Japanese FeNi 0.5% 0.7%
% of ore traded intl 0.3% 0.3%
Upstream energy of natural gas 0.2% 0.2%
Domestic transport distances 0.1% 0.1%
New Caledonia elec eff 0.0% 0.0%
PT Aneka Tambang ore grade −0.1% −0.1%

imum recycling would require 24 GJ/t, and all-virgin production


would require 52 GJ/t of primary energy. There are many grades
Fig. 5. Primary energy requirements for various AISI grades of stainless steel over
three scenarios (global averages). of ferritic stainless steels, but assuming an equal share of 409 and
430 grade of the year 2005 production levels of 6.2 Tg (Brook Hunt,
2007), the current scenario conserves 52 PJ, or 16% compared with
derived from sulfide ores and concentrate requires the least pri-
all-virgin production. The maximum recycling scenario uses fer-
mary energy. This is largely due to the fact that the smelting of
ritic stainless scrap for all chromium units, conserving half of all
sulfide ores utilizes the exothermic heat of oxidation of iron and
primary energy use compared with all-virgin production.
sulfur, and thus requires relatively little fossil fuel. Pyrometallurgy
These energy savings have significant impacts on greenhouse
of laterite ores requires more energy than for sulfide ores, primar-
gas emissions. The use of secondary material in austenitic and fer-
ily because lateritic ores contain no sulfur and so can not make
ritic stainless steel production in 2005 prevented approximately
use of exothermic (heat-producing) reactions, instead relying on
30 million tons of CO2 from entering the atmosphere. If related
fossil fuel and electricity. Hydrometallurgy of lateritic ores (leach
metals industries were given credits for these reductions as part
and nickel oxide) consumes even more primary energy per unit
of a global climate policy, assuming a market price of US$10–$20/t
nickel. This is largely due to the copious use of fossil fuels at the
CO2 eq, these reductions would be worth $300–$600m.
calcining and leaching steps. Inco Sudbury produces nickel oxide
Nickel pig iron production has constituted much of the expan-
sinter using pyrometallurgical techniques, largely for refining at
sion of the nickel industry in China over the past several years. In
its Clydach, Wales facility; this material consumes relatively little
2005, Chinese production of nickel pig iron was 4.5 Gg contained
primary energy, as seen in Fig. 4b. At the high end of the spec-
nickel, though that number surged to between 70 and 90 Gg in
trum, Chinese nickel pig iron requires huge amounts of coking
2008. The production of nickel pig iron per unit contained nickel
coal and is by far the most energy intensive nickel product on a
requires large amounts of energy compared to ferronickel, particu-
contained nickel basis. INMETCO’s scrap-derived nickel pig iron
larly for coke. As such, the introduction of nickel pig iron in Chinese
requires roughly 207 GJ/t contained nickel, much less than Chinese
production in 2005 (in very small quantities compared to current
virgin nickel pig iron.
2008 levels) increased the global average energy requirement of
The refining step for Class I nickel is much less energy inten-
austenitic stainless by 0.15 GJ/t or nearly 0.3%. If nickel pig iron use
sive than either the mining or smelting steps above. Most refining
rises to contribute roughly 20% of nickel in global stainless produc-
processes rely primarily on electricity as an energy source and
tion, the energy requirement of austenitic stainless steel production
does not require extensive heating. The energy credit from the
would increase by approximately 3.5%.
co-production of cobalt during refining also lowers the net energy
Increased use of nickel pig iron has other environmental con-
requirements slightly.
sequences as well, stemming from the high level of coke use and
The analysis for stainless steel is more complicated, as these
the vintage of most small Chinese blast furnaces used for smelting.
alloys incorporate many more materials and process steps than for
Elevated local levels of particulate matter and phosphorus contam-
primary nickel production. For 304 grade austenitic stainless steel,
ination of stainless steel are two trends that should be expected,
current production requires 57 GJ/t, maximum recycling would
in the absence of dramatic improvement in industry practices in
require 25 GJ/t, and all virgin would require 79 GJ/t of primary
China.
energy, as shown in Fig. 5.
Energy analysis and life-cycle assessment results can be quite
Given year 2005 production levels of 18.5 Tg of austenitic stain-
sensitive to certain model parameters. Sensitivity analysis revealed
less steel (Brook Hunt, 2007), the recycling of stainless steel in
current operations conserves 29% of primary energy, or 420 PJ com-
pared with all-virgin production. Table 9b
Unlike austenitic stainless steel, 409 grade ferritic stainless con- Effect of a 10% increase in model parameters on total results: Class I refined nickel.
tains 0–0.5% titanium, which is extremely energy intensive to Process Primary Greenhouse
produce, requiring 644 GJ/t (Ecoinvent, 2007). Scrap use in fer- energy use gas emissions
ritic stainless also differs from austenitic: there is more use of Allocation for Ni in matte prod. 4.5% 4.5%
carbon steel scrap but much less use of stainless scrap. How- Upstream energy of coal 1.9% 1.9%
ever, there is only 10–12% chromium in 409 grade stainless as Beneficiation electricity use 1.6% 1.6%
opposed to 18% in 304 grade, so that 409 grade production % Ni lost to tailings 0.9% 0.9%
Nickel content of Russian matte 0.5% 0.8%
requires a relatively smaller amount of energy-intensive primary % Ni lost during pyrometallurgy 0.5% 0.5%
ferrochromium. Russian electrorefining electricity use 0.4% 0.4%
430 grade ferritic stainless steel contains no titanium but Upstream energy of sulfuric acid −0.3% −0.3%
16% chromium. This increases its energy requirements slightly Norlisk average ore grade −0.8% −0.8%
Canada elec eff −0.9% −0.9%
compared to 409 grade: current production requires 43.3 GJ/t, max-
M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266 265

that changes in all major parameters had a much smaller effect than Dalvi A, Bacon WG, Osborne RC. The past and the future of nickel laterites. Prospec-
in previous work (Johnson et al., 2008). This is mostly due to the tors & Developers Association of Canada; 2004.
Ecobalance. Life cycle assessment of nickel products, Final report for the nickel
independence of facility-level energy consumption data. Parame- industry LCA group. Bethesda, MD; 2000.
ters that apply to multiple facilities, such as the energy used in the Ecobalance. Life cycle assessment of nickel products, Final report for the nickel
production of coal, have the most influence on the total result. Sum- industry LCA group. Bethesda, MD; 2003.
Ecoinvent, Ecoinvent database for life cycle analysis of emissions and materials, Data
maries of the sensitivity analyses for ferronickel and Class I refined Vers. 1.3. Zurich, Switzerland: Swiss Center for Life Cycle Inventories; 2007.
nickel are shown in Table 9a and Table 9b. ERDA. A national plan for energy research, development and demonstration: cre-
Although commercial ore grades are decreasing and nickel pig ating energy choices for the future. Washington, D.C.: Energy Research and
Development Administration; 1976.
iron is growing in importance as a source of supply, new processing European Integrated Pollution Prevention and Control Bureau. BREF: Non-
technologies may lead to an overall decrease in the energy require- ferrous metal processes; 2006. Available from: http://eippcb.jrc.es/pages/
ments for ferronickel and refined nickel production. Atmospheric FActivities.htm.
Falconbridge Limited, Shane D, Mouland J, Uceda D. Recovery of nickel from copper
acid leaching of ore heaps does not require the intensive pressuriza-
refinery tankhouse electrolyte. Canada: World Intellectual Property Organiza-
tion necessary for current HPAL technology. Combined biological tion; 1999.
and chemical hydrometallurgy of sulfide ores also has the poten- Forbes P, Von Blottnitz H, Gaylard P, Petrie JG. Environmental assessment of base
tial to reduce dramatically the energy used in beneficiation and metal processing: a nickel refining case study. Journal of the South African Insti-
tute of Mining and Metallurgy 2000;100:347–53.
leaching. Bio-leaching does not require fine milling of sulfide ores Gustavson KE, Barnthouse LW, Brierley CL, Clark EH, Ward CH. Superfund and mining
prior to processing and also takes place at atmospheric pressure. megasites. Environmental Science & Technology 2007;41:2667–72.
Detailed energy consumption data are not available for these pro- Harper EM, Johnson J, Graedel TE. Making metals count: applications of material
flow analysis. Environmental Engineering Science 2006;23:493–506.
cesses, but it will be important in coming years to examine whether ICMM. Sustainable development framework. International Council on Mining and
these technological advances can balance the increases in energy Metals; 2008. Available from: http://www.icmm.com/page/200.
requirements due to other factors. IEA. Energy balances of non-OECD countries: 2005. Paris, France: International
Energy Agency; 2007.
Detailed accounting of energy use and greenhouse gas emis- IEA. Energy balances of OECD countries: 2005. Paris, France: International Energy
sions is now an integral part of corporate environmental strategy Agency; 2007.
and responsibility. Accurate accounting over the product life cycle International Nickel Study Group. World nickel statistics. Bulletin of the
INSG2006;15: 124.
requires a large number of specific data, especially so for products International Nickel Study Group. Private communication; 2008.
with long and shifting supply chains. The use of average values International Stainless Steel Forum. ISSF scrap survey; 2004.
can in many cases distort the picture by grossly under- or over- ITP. Mining, industry of the future: energy and environmental profile of the U.S.
mining industry. Washington, D.C.: U.S. Department of Energy; Office of Energy
estimating life-cycle impacts. In a culture where consumers of all
Efficiency and Renewable Energy; Industrial Technologies Program; 2002.
kinds demand information about the energy and greenhouse gas Japanese Environmental Management Association for Industry. JEMAI-LCA Pro v4;
impacts of a specific product in a specific place, and money is trans- 2007.
acted based on that information as part of a carbon market, it is Johnson J, Reck BK, Wang T, Graedel TE. The energy benefit of stainless steel recycling.
Energy Policy 2008;36:181–92.
essential to perform life-cycle assessments with as much detail as Kerfoot DGE. Nickel. In Ullmann’s Encyclopedia of Industrial Chemistry, Vol. A17.
possible. The present study provides results for nickel and nickel New York: VCH Verlagsgesellschaft mbH; 2002.
products and alloys based on facility-level data, accounting for ore Kuck P. Nickel, 2005 minerals yearbook. Reston, VA: United States Geological Survey;
2006.
type and grade, processing technology, fuel type, geographical loca- Kusik CL, Kenahan CB, 1978. Energy use patterns for metal recycling. Washington,
tion, country-to-country trade, and product quality. D.C.: U.S. Dept. of Interior.
Lankey R, McMichael F. Rechargeable battery management and recycling: a green
design educational module. Pittsburgh: Carnegie Mellon University, Green
Acknowledgements Design Initiative; 1999.
Lennon J. The Chinese nickel outlook and the role of nickel pig iron. Presentation to
International Nickel Study Group; 2007.
Thanks are due to my colleagues Barbara Reck, Jeremiah John- McCutcheon B. Nickel. Canadian Minerals Yearbook; 2005.
son, and Thomas Graedel at the Center for Industrial Ecology at Yale Mining Association of Canada. Benchmarking the energy consumption of Canadian
University; Sophia Wong and Bruce McKeon at the Nickel Institute open-pit mines. Ottowa: Natural Resources Canada; 2005.
Mining Association of Canada. Benchmarking the energy consumption of Canadian
for guidance and suggestions; Jim Middleton for careful reading of
underground bulk mines. Ottowa: Natural Resources Canada; 2005.
the work; and John Barkas of Metalytics Pty Limited for valuable Narita N, Sagisaka M, Inaba A. Life cycle inventory analysis of metals. Materia Japan
comments and data. 2001;40:705–9.
Non-Ferrous Metals Industry of China. Yearbook of the Non-Ferrous Metals Industry
of China; 2005 [in Chinese].
References Norgate TE. Metal recycling: an assessment using life cycle energy consumption as
sustainability indicator. CSIRO Minerals, Rep. DMR-2616; 2004. 39 pp.
Adelhardt W, Antrekowitsch H, Atmaca T, Neumann W, Thormann A. Stoffmen- Norgate TE, Lovel RR. Water use in metal production: a life cycle perspective. DMR-
genflusse und Energiebedarf bei der Gewinnung ausgewahlter mineralischer 2505. CSIRO; 2004.
Rohstoffe, Geologisches Jahrbuch Sonderhefte. Hannover: Wuppertal Institute Norgate TE, Rankin WJ. The role of metals in sustainable development. Green pro-
for Climate, Environment, and Energy; 1998 [in German]. cessing conference, Cairns, QLD, Australia; 2002.
Althaus HJ, Classen M. Life cycle inventories of metals and methodological aspects of Raw Materials Group, Raw Materials Database. Solna, Sweden; 2007.
inventorying material resources in Ecoinvent. The International Journal of Life Reck BK, Müller DB, Rostkowski K, Graedel TE. Anthropogenic nickel cycle:
Cycle Assessment 2005;10:43–9. insights into use, trade, and recycling. Environmental Science & Technology
Argonne National Laboratory, Greenhouse Gases. Regulated emissions, and energy 2008;42:3394–400.
use in transportation (GREET) v. 1.8. U.S. Department of Energy; 2008. Reuter MA. The simulation of industrial ecosystems. Minerals Engineering
Barkas JP. Patterns of energy consumption and greenhouse gas emission across pri- 1998;11:891–918.
mary base metals production. In: Joint ILZSG-ICSG-INSG energy and climate Rydh CJ, Karlstroem M. Life cycle inventory of recycling portable nickel-cadmium
change policy seminar; 2009. batteries. Resources, Conservation and Recycling 2002;34:289–309.
Bergman RA. Nickel production from low-iron laterite ores: process descriptions. Sagisaka M, Onoye T, Tashiro K, Kitamura O. Nickel recycling system and its life-
CIM Bulletin 2003;96:127–38. cycle environmental assessment. Resources Processing 2007;54:195–200 [in
Boldt JR. The winning of nickel. Toronto: Longmans Canada; 1967. Japanese].
Boustead I, Hancock GF. Handbook of industrial energy analysis. Chichester, UK: Ellis Schuller M, Estrada A, Bringezu S. Mapping environmental performance of inter-
Horwood; 1979. national raw material production flows: a comparative case study for the
Brook Hunt, 2007. The long term outlook for nickel, metal service. 3rd Quarter Data copper industry of Chile and Germany. Minerals & Energy—Raw Materials Report
Volume. 2008;23:29–45.
Brown HL. Energy analysis of 108 industrial processes. Lilburn, GA, USA: The Fair- Shedd K. Cobalt, 2005 Minerals Yearbook. United States Geological Survey; 2006.
mont Press; 1996. Sheehan J, Camobreco V, Duffield J, Graboski M, Shapouri H. Life cycle inventory of
Chapman PF, Roberts F. Metal resources and energy. London: Butterworth and Co; biodiesel and petroleum diesel for use in an urban bus. Final report. NREL/SR–
1983. 580-24089. Golden, CO (US): National Renewable Energy Lab.; 1998.
266 M.J. Eckelman / Resources, Conservation and Recycling 54 (2010) 256–266

Sjardin M. CO2 emission factors for non-energy use in the non-ferrous metal, ferroal- Warner AEM, Díaz CM, Dalvi AD, Mackey PJ, Tarasov AV. JOM world nonferrous
loys and inorganics industry. Utrecht, The Netherlands: University of Utrecht, smelter survey, part III: nickel: laterite. JOM Journal of the Minerals, Metals and
Department of Science, Technology, and Society; 2003. Materials Society 2006;58:11–20.
Spath PL, Mann MK. Life cycle assessment of a natural gas combined cycle power Warner AEM, Díaz CM, Dalvi AD, Mackey PJ, Tarasov AV, Jones RT. JOM world nonfer-
generation system. NREL/TP-570-27715. Golden, CO (US): National Renewable rous smelter survey Part IV: nickel: sulfide. JOM Journal of the Minerals, Metals
Energy Lab.; 2000. and Materials Society 2007;59:58–72.
Spath PL, Mann MK, Kerr DR. Life cycle assessment of coal-fired power produc- Xu A. Thoughts on nickel pig iron manufacturing. China Metal Bulletin 2007:11–2
tion. NREL/TP-570-25119. Golden, CO (US): National Renewable Energy Lab.; [in Chinese].
1999. Xu J, Jiang J, Chen H, Wang T, Hao W. LCA study for metallic nickel produced by
U.S. Department of Energy. Energy Information Administration; 2006. pyrometallurgy processes. Xiyou Jinshu Cailiao yu Gongcheng (Rare Metal Mate-
UN COMTRADE. Trade Statistics; 2008. rials and Engineering) 2004;33:75–8 [in Chinese].
United Nations Department of Economic and Social Affairs/Statistics Division. Yoshiki-Gravelsins KS, Toguri JM, Choo RTC. Metals production energy and the envi-
United Nations Commodity Trade Statistics Database; 2008. ronment, part I. Energy consumption. Journal of Metals 1993;45:15–20.
Verhoef E, Dijkema G, Reuter MA. Process knowledge, system dynamics and metal Yu HQ, Chen JC. Study on life cycle assessment for the process of nickel mining and
ecology. Journal of Industrial Ecology 2004;8:23–43. metallurgy. Yunnan Metallurgy 2007;36:38–41 [in Chinese].

You might also like