Life cycle assessment (LCA) of waste management strategies: [616227]

Life cycle assessment (LCA) of waste management strategies:
Landfilling, sorting plant and incineration
Francesco Cherubinia,*, Silvia Bargiglib, Sergio Ulgiatic
aJoanneum Research, Elisabethstraße 5, 8010 Graz, Austria
bUniversita `degli Studi di Siena, Via Aldo Moro 2, 53100 Siena, Italy
cUniversita `degli Studi di Napoli ‘‘Parthenope’’, Dipartimento di Scienze per l’Ambiente, Centro Direzionale, Isola C4, 80133 Napoli, Italy
article info
Article history:
Received 16 January 2008
Available online 1 November 2008
Keywords:
LCAWaste managementLandfillMSW sorting plant
MSW incinerationabstract
This paper focuses on a Life Cycle Assessment (LCA) of four waste management strategies: landfill
without biogas utilization; landfill with biogas combustion to generate electricity; sorting plant which
splits the inorganic waste fraction (used to produce electricity via Refuse Derived Fuels, RDF) from the
organic waste fraction (used to produce biogas via anaerobic digestion); direct incineration of waste.These scenarios are applied to the waste amount and composition of the Municipality of Roma (Italy) and
are evaluated under different points of view: global and local emissions, total material demands, total
energy requirements and ecological footprints. Results, reliable for most of the European big cities, showlandfill systems as the worst waste management options and significant environmental savings at global
scale are achieved from undertaking energy recycling. Furthermore, waste treatments finalized to energy
recovery provide an energy output that, in the best case, is able to meet the 15% of Roma electricityconsumption.
/C2112008 Elsevier Ltd. All rights reserved.
1. Introduction
In most developed and developing countries with increasing
population, prosperity and urbanization, one of the major chal-
lenges for municipalities is to collect, recycle, treat and dispose of
increasing quantities of solid waste and wastewater.
A cornerstone of sustainable development is the establishment
of affordable, effective and truly sustainable waste management. It
must be emphasized that multiple public health, safety and envi-
ronmental cobenefits accrue from effective waste management
practices which concurrently reduce Greenhouse Gas (GHG)
emissions and improve the quality of life, promote public health,
prevent water and soil contamination, conserve natural resourcesand provide renewable energy benefits.
For all these reasons, there is an increasing interest in the
several options for management of resources and waste in order to
design strategies for integrated, sustainable resource and waste
management policies.
Life Cycle Assessment (LCA) methodologies can be used in this
context as an input to decision-making regarding the choice of
waste management systems, or strategic decisions concerning
resource use priority. In fact, a LCA is able to provide an overview of
the environmental aspects of different waste managementstrategies, and makes possible to compare the potential environ-mental impacts of these options.
The present paper focuses on the LCA of alternative urban solid
waste management strategies that could be applied in large
municipality, where a significant critical mass of waste is present.
The assessment is assumed to be applied to the waste stream of the
biggest Italian city, Roma, but the final results can be considered
reliable for most of the European cities, which have a similar
Municipal Solid Waste (MSW) composition [1]. The city of Roma
has an area of 1.29E ț09 m
2, 2.81E ț06 inhabitants and a production
of MSW equal to 1.46E ț06 ton per year (recycled wastes not
included). These materials are to date sent to the landfill of Mala-grotta (Roma), where a partial landfill gas recovery for electricitygeneration takes place.
The aim of this study is to compare selected waste disposal
alternatives in a Life cycle perspective, considering both landfill
systems, where no recycle takes place, and systems which are able
to minimize the amount of landfilled waste while maximizing
material and energy recovery.
2. Methods
2.1. The LCA
A LCA study is a tool able to evaluate environmental burdens
associated with a product, process, or service by identifying energy
*Corresponding author. Tel.: ț433168761327; fax: ț433168761330.
E-mail address: cherufra@yahoo.it (F. Cherubini).
Contents lists available at ScienceDirect
Energy
journal homepage: www.else vier.com/locate/energy
0360-5442/$ – see front matter /C2112008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2008.08.023Energy 34 (2009) 2116–2123

and materials used and emissions released to the environment;
moreover it allows also an identification of opportunities for
environmental improvements [2–4] .
LCA methodology is the internationally standardized method
that is considered one of the most effective management tools for
identifying and assessing the environmental impacts related with
waste management options [5–8] . In particular, the broad
perspective of LCA makes possible to take into account the signifi-cant environmental benefits that can be obtained through different
waste management processes. For instance, waste incineration with
energy recovery reduces the need for other energy sources, material
from recycling processes replaces production of virgin material and
biological treatment may reduce the need for production of artificial
fertilisers and transportation fuel [9]. Some models, which are able
to perform LCA of waste management systems also exist, with thepurpose of speed up the analysis and allow the analysts to know
how the changes in the system affect the environmental impacts
through scenario analysis [10].
The methodological framework used in this paper is the LCA as
defined by ISO standards (International Standard Organization, ISO
14040:14043), with several methods jointly applied in order toinvestigate system performances under different points of view,
such as material and energy requirements, environmental impacts
and ecological footprint. Such approach is applied because the LCA
of a product or service should be the assessment of the product with
regard to its impacts on the environment and on human health, and
should aim to be an overall ecological assessment. Therefore, a range
of specific and selected environmental impacts is assessed, but other
aspects such as economic and social factors are not considered.
A typical LCA study consists of the following stages: goal and
scope definition; life cycle inventory (LCI) analysis, with compila-
tion of data both about energy and material flows and on emissions
to the environment, throughout the life cycle of the case study (ISO
14041); assessment of the potential impacts (Life Cycle Impact
Assessment, LCIA) associated with the identified forms of resource
use and environmental emissions (ISO 14042); interpretation of the
results from the previous phases of the study in relation to the
objectives of the study (ISO 14043).
The goal of this study is to provide a transparent and compre-
hensive environmental evaluation of a range of waste management
strategies for dealing with mixed waste fractions of a city waste
stream (Roma in this work). However, even if the impacts of these
systems are very dependent on local, regional and national condi-
tions, including consumer habits, transport means, generation of
by-products and energy, or the local energy supply systems (fossil
fuels, biomass, hydropower, nuclear), the results can be interpreted
as significant in a general framework.
The definition of the functional unit, that is the focus of the
study, is a special issue for LCA studies of waste management
options, since LCA for waste management differs from product LCA.
In a product LCA the functional unit is usually defined in terms of
the system’s output, i.e. the product (for example, per MJ or km of
transportation biofuels, or per number of mobile phones
produced). In an LCA for waste management the functional unit
must be defined in terms of system’s input, i.e. the waste. The
management of the quantity of specific waste, or the waste of onehousehold, or the total waste of a defined geographical region in
a given time (e.g. 1 year) can be chosen as the functional unit. In this
study, the functional unit to which the results are referred is the
amount of waste produced in a year (2003), by the city of Roma
(geographical system boundary), which must be disposed of:
1460 kton of wastes contained in the so-called ‘‘black sacks’’ (i.e.
pre-sorted and recycled wastes not included).
Regarding the scope of the assessment, four different waste
management strategies are investigated (the collection is not
investigated because it is assumed to be the same for all scenarios):/C15Scenario 0 : wastes are delivered to landfill without any further
treatment;
/C15Scenario 1 : part of the biogas naturally released by the landfill is
collected, treated and burnt to produce electricity;
/C15Scenario 2 : a sorting plant is present at landfill site for sepa-
ration of the organic and inorganic fractions and for ferrous
metal recovery. Electricity, biogas and compost are thenproduced on site;
/C15Scenario 3 : wastes are directly incinerated to produce
electricity.
For each scenario, a detailed LCI, containing the mass and energy
flows directly and indirectly involved in urban waste system, iscompiled (see Table 1 ). Results from LCI are then used for the
upstream and downstream characterization of impacts (LCIAphase). Liquid, solid and gaseous emissions are carefully evaluated
and classified into impact categories in order to estimate indicators
such as Global Warming Potential (GWP), Acidification Potential
(AP) and Eutrophication Potential (EP); local emissions are evalu-
ated as well. The GHGs considered as potential contributors to
global warming are CO
2,C H 4,N2O; gases responsible of rain acid-
ification are SO 2,N O x, HCl, H 2S, and HF; chemical compounds that
contribute to aquatic system eutrophication are total-N and total-P
(equivalency factors for each impact category are available in [4]).
Furthermore, the assessment also includes indicators of resourcedepletion such as Material Intensities (MIs) (derived from the
Material Flow Accounting, MFA) and Gross Energy Requirements
(GERs), and of ecological footprint such as the Sustainable Process
Index (SPI).
Several assessment methods were chosen because a single
method cannot be sufficient, in any circumstances, to provide
comprehensive information on environmental impact assessment.
In fact, LCAs based on only one approach invariably lead to partial,
and sometimes even worthless, indications. A complete LCA should
carefully rely on a selection of impact assessment methods, which
account for both the upstream and the downstream categories
1.
Early efforts of the scientific community in this direction can berecognised in the scientific literature [11–13] . A short description of
these assessment methods is provided in the following section.
2.2. The assessment methods
The MFA method aims at evaluating the environmental distur-
bance associated with the withdrawal or diversion of material flows
from their natural ecosystemic pathways [14–16] . In this method,
appropriate MI factors (g/unit) are multiplied by each input,respectively, accounting for the total amount of abiotic matter,
water, air and biotic matter directly or indirectly required toprovide that very same input to the system. The resulting MIs of the
individual inputs are then summed together for each environ-
mental compartment (again: abiotic matter, water, air and biotic
matter), and assigned to the system’s output as a quantitative
measure of its cumulative environmental burden from that
compartment.
The GER is an indicator coming from the Embodied Energy
Analysis method, which deals with the GER (direct and indirect) of
the analysed system, and offers useful insight on the first-law
1All impact assessment methods can be divided in two broad categories: those
that focus on the amount of resources used per unit of product (‘‘upstream’’
methods), and those that deal with the fate of system emissions (‘‘downstream’’
methods). The former can provide invaluable insights into the hidden environ-mental costs and inherent (un)sustainability of systems (even when they seemclean). On the other hand, downstream methods are often more closely related tothe impact immediately perceived on the local ecosystem, and can unveil largedifferences between systems with similar upstream performance.F. Cherubini et al. / Energy 34 (2009) 2116–2123 2117

energy efficiency of the analysed system on the global scale, taking
into consideration all the employed commercial energy supplies
[17,18] . In this method, all the material and energy inputs to the
analysed system are multiplied by appropriate oil equivalentfactors (g
oil/unit), and the cumulative embodied energy require-
ment of system output is then computed as the sum of the indi-vidual oil equivalents of the inputs, which can be converted to
energy units by multiplying by the standard crude oil equivalency
factor of 41.860 J/g. The chosen cumulative indicator is the so-called
GER, expressing the total commercial energy requirement of 1 unit
of output in terms of equivalent Joules of petroleum oil.
The SPI, a member of the ecological footprint family, relies on
a method that converts mass and energy flows used and produced
by the process into area [19–21] . The sum of all areas results in
a highly aggregated number, an ecological footprint. As area isdefinitely a limited source, competitiveness with respect to
sustainability can be assessed and processes can be compared
easily. The concept of the SPI measures the potential impact
(pressure) of processes (or, more generally, activities) on the
ecosphere. The SPI compares mass and energy flows induced by
human activities with natural flows. As natural flows are always
linked to area (examples are the growth of biomass, precipitation
and, most importantly, solar radiation) the basic unit of the SPI isarea: it is the total surface area that is required by any activity that
exchanges material with the environment to be ‘‘sustainably
embedded into the ecosphere ( ¼environment)’’. Thus, the lower
the requirement of area for a given activity is, the lesser is theimpact of this activity on the environment; the SPI concept allows
a quick and reliable evaluation of very diverse processes according
to their environmental impact from a sustainable development
point of view [22]. The software tool able to calculate the ecological
footprint of a process from a set of input/output data (an LCI) can bedownloaded from the SPIonExcel homepage ( www.SPIonExcel.tugraz.at ) for free. Also available are additional databases contain-
ing detailed data for a wide range of processes.
3. Description of the scenarios
The MSW elementary composition (i.e. the fraction of C, H, N, S
and O) assumed for the assessment is derived from Ref. [1], while
the waste type abundance in the total waste mass (i.e. the % of
kitchen garbage, paper, plastics .) comes from the environmental
report of AMA, the company in charge of waste disposal in Roma[23]. These results, shown in Table 2 , can be considered typical for
most of European cities.
The scenarios investigated in this paper are based on the
following assumptions: since the output of a scenario is usuallycomposed of one or two energy types (electricity and/or biogas),
we consider that part of this amount is used as a feedback for
internal energy consumption within the landfilling/waste treat-
ment facility (as a consequence, the quantities of the final outputs
are ‘‘net’’); if a scenario has an energy output which provides
environmental benefits, both gross and net impact indicators (e.g.
GWP and AP) are shown; environmental benefits derived from
recycled material (ferrous metals and compostable matter) are not
accounted for, due to uncertainty of available data; if a process stephas two or more co-products (i.e. the organic and inorganic waste
fraction, as well as ferrous metals, which come out from the sorting
plant of Scenario 2), an allocation of material and energy flows
based on exergy content of the different outputs is applied.
3.1. Scenario 0: the landfill
Wastes are buried in a monitored landfill. For more information
about a landfill system see Refs. [1,24,25] . In this system, the
organic fraction of wastes undergoes decomposition in anaerobicTable 1
LCI of the main input flows to waste management options; amounts are normalized per g of waste
Input flow (per g of waste) Unit Scenario
0 Ref 1 Ref 2 Ref 3 Ref
1 Water from wells g 0.10 [32]ț[36] 0.16 [36]
2 Natural gas g 1.98E-03 [35]ț[36] 6.01E-05 [1]
3 ElectricityakW h 9.63E-07 [23] 5.31E-07 [23]ț[31] 8.16E-05 [32]ț[35] 6.68E-05 [1]
4 Diesel g 6.24E-04 [23] 6.24E-04 [23] 3.79E-04 [35] 1.57E-04 [36]
5 HDPE (pipes) g 1.25E-04 [1] 1.25E-04 [1]
6 HDPE (landfill walls) g 6.06E-05 [1] 6.06E-05 [1] 9.83E-06 [1] 1.33E-05 [1]
7 Clay (landfill walls) g 4.47E-02 [1] 4.47E-02 [1] 7.26E-03 [1] 9.80E-03 [1]
8 Concrete g 6.83E-03 [34]ț[42] 6.87E-04 [34]
9 Steel g 4.20E-07 [1] 1.77E-03 [34]ț[42] 5.62E-04 [34]
10 PVC reactors (H 2S removal) g 1.45E-08 [1]
11 Iron sponge (H 2S removal) g 4.89E-07 [31]
12 Copper cables g 1.76E-05 [36]ț[42]
13 Polyethylene film g 1.60E-04 [36]
14 Urea (NH 2CONH 2)g 6.44E-04 [36] 3.00E-03 [36]
15 Activated carbon g 1.34E-03 [36] 2.50E-03 [36]
16 Ca(OH) 2 g 1.72E-03 [36] 3.20E-03 [36]
17 CaO g 1.34E-02 [36] 2.50E-02 [36]
18 Cement g 7.25E-04 [36] 1.35E-02 [36]
19 Sodium silicate g 8.05E-04 [36] 1.50E-03 [36]
Useful output20 Electricity
a
gross kW h 8.05E-05 7.67E-04 6.61E-04
net kW h 8.00E-05 6.85E-04 5.94E-04
21 Upgraded biogas J 2.55E ț03
22 Compostable matter (d.m.) g 0.12
23 Ferrous metals g 2.80E-02
Waste to landfill24 Untreated waste g 1.00 1.00
25 Heavy wastes g 3.73E-02
26 Ashes g 0.12 0.22
aA fraction of the electricity produced in the Scenario is used to meet the landfill/plant energy needs.F. Cherubini et al. / Energy 34 (2009) 2116–2123 2118

conditions, releasing the so-called ‘‘landfill (bio-)gas’’. This is
mainly composed of CH 4(58%) and CO 2(41%), but it may also
contain traces of H 2S, HCl, HF and other chemical compounds. On
a yearly basis, a methane production rate of 140 Nm3per ton of
landfilled waste is estimated [1]. A fraction of 50% of biogas is
assumed to be collected by pipes and burnt in flares to convert CH 4
to the less impacting CO 2(mainly). Such a CO 2emission from
landfill gas flaring is quantified but it is not accounted for in the
GWP because it has not a fossil origin (it comes from organic
fraction, because plastic does not decompose); the remaining 50%
of landfill gas is assumed to be directly released to the atmosphere.
Other basic assumptions regarding landfill activities (valid both
for Scenarios 0 and 1) are the following: 3% of the total Sulphur
disposed of to landfill in one year is released to atmosphere as H 2S
[26]; the main emissions released from biogas combustion in flares
are CO, NO 2, HCl, HF (emission factors: 800,100,12 and 0.02 mg/m3,
respectively [27]) and dioxins (emission factors from [28]); the
main emissions freely released from the landfill together with
biogas (i.e. CH 4and CO 2) are CO, HCl and HF (emission factors: 13,
65 and 13 mg/m3respectively [27]); heavy metals released to
atmosphere are only Mercury (Hg) and Cadmium (Cd), the most
volatiles [1]; leachate emission (and its composition) is averaged
from several data reported in [29].
According to such assumptions, the airborne emissions due to
combustion processes directly or indirectly involved in the systemare also evaluated, together with (as well as for Scenarios 1, 2, 3):
emissions from spontaneous landfill fires (CO
2, CO, NO x, dioxins
among others [1])2, from combustion of fossil fuels such as natural
gas and diesel (CO 2,C O ,N O x, PM10, SO 2,C H 4,N2O[30] and dioxins
[28]), from electricity consumption (according to the Italian electric
supply mix, made of coal 14.2%, oil 30.8%, natural gas 34.8%, hydro
16.6%, other renewable 3.6%) and from production and deliver of
the input flows required by the system.
All dioxin and furan emissions are referred to g of 2,3,7,8-TCDD
with apposite equivalence factors [28]. In this scenario, the most
relevant input flows are clay and earth for landfill walls (Table 1 ).
3.2. Scenario 1: landfill with biogas recovery
The biogas naturally released from landfill is collected (about
50% of the total), treated and burnt (in situ) in order to produce
electricity. The other part of the biogas is burnt in flares (25%) or
released to the atmosphere (remaining 25%). As mentioned in the
previous section, the CO 2released by biogas combustion is not
accounted for in the GWP but it is included among the localemissions. The estimated energy content of the collected biogas is3.12Eț09 MJ, thanks to a biogas lower heating value of 17.73 MJ/m
3
[31]. The biogas is thus burnt in turbines at 28% efficiency to
produce 2.43E ț08 kW h of electricity per year. Airborne emissions
from combustion are calculated according to Ref. [31]. Material
inputs for biogas collection and use are to be considered negligible
in terms of mass contribution compared to the clay and earth for
landfill walls (see Table 1 ).
3.3. Scenario 2: MSW sorting plant
This scenario (see Fig. 1 ) is based on a sorting plant (for infor-
mation on sorting plant see [32–34] ) that is able to split the waste
into several components: the organic from the inorganic part (themain flows), as well as ferrous metals from lighter materials and
the heavy fraction of the waste. The organic fraction is essentially
made of kitchen garbage (about 50% of the original MSW weight)
while mainly plastics, paper and cardboard, wood, textiles and
rubber constitute the inorganic fraction
3. After the MSW sorting
phase, an allocation, based on the exergy content of the products, isapplied to the different outputs. This results in a share of the
environmental burdens equal to 79% to the inorganic fraction, 19%
to the organic fraction and 2% to ferrous metals. The landfilled
heavy waste is not a product and therefore it is not accounted for.
After this step, all the flows are independently treated and no
further allocation is needed (biogas is assumed as the main product
of the anaerobic digestion step and no burdens are assigned to the
compostable matter, as well as electricity is the main product of the
combustion step and ashes are landfilled).
The organic part is then delivered to another plant in which it
undergoes anaerobic digestion in order to produce biogas (70% CH
4
and 30% CO 2), with an average biogas yield of 3.74E ț03 MJ/ton of
organic waste [35]. This results in a total biogas production of
2.19Eț09 MJ. Afterwards, the biogas is upgraded, in order to
remove H 2S (with a dry sorption process, through its reaction with
a recyclable metal oxide such as iron sponge to form a metal sulfide
compound) and other impurities, with an increase of the
percentage of CH 4up to 97%. The energy demand of this step is
equal to the 11% of the energy content of the biogas [35]. The
digestate, i.e. the residue of the anaerobic digestion that remains
inside the reactor, can be treated and used as fertilizer.
The inorganic part of the waste is delivered to a Refuse Derived
Fuel (RDF) production plant. The RDF bricks are then burnt in an
incineration plant to generate electricity. Generally, it is preferableTable 2
MSW composition
Component wt(%) Moisture (%) Ash (%) C (%) H (%) N (%) S (%) O (%)
Kitchen garbage 49.5 70.0 5.0 48.0 6.4 2.6 0.4 37.6
Paper 12.0 10.2 6.0 43.5 6.0 0.3 0.2 44.0
Cardboard 6.8 5.2 5.0 44.0 5.9 0.3 0.2 44.6
Plastics 22.9 0.2 10.0 60.0 7.2 0.0 0.0 22.8
Textiles 2.4 10.0 2.5 55.0 6.6 4.6 0.2 31.2
Rubber 0.3 10.0 78.0 10.0 2.0 0.0 0.0 10.0
Leather 0.3 10.0 60.0 8.0 10.0 0.4 11.6 10.0Wood 1.3 1.5 2.4 49.4 5.7 0.2 0.0 42.3Glass 1.5 2.0 98.9 0.5 0.1 0.1 0.0 0.4Ferrous metals
a1.9 2.0 90.5 4.5 0.6 0.1 0.0 4.3
Aluminiuma0.9 2.0 90.5 4.5 0.6 0.1 0.0 4.3
Other 0.2 3.2 68.0 26.3 3.0 0.5 0.2 2.0
aOrganic content is from coatings, labels, and other attached materials.
2Regarding the dioxins, PAH and PCB emissions, only 35% of their amount is
treated as airborne emission, since they are mostly absorbed to the particle mattersand fall down to the ground close to the point from where they are emitted.3Cellulosic material is not an inorganic matter, but since they cannot be fer-
mented (cellulose and hemicellulose do not depolymerize in the anaerobic condi-tion required for biogas production) they are stored with the inorganic fraction ofthe waste to be burnt (via RDF).F. Cherubini et al. / Energy 34 (2009) 2116–2123 2119

to burn RDF than untreated wastes. In fact, RDF has a higher heating
value, a more homogeneous chemical composition, an easier
storage and handling ability and less emission factors. The
combustion of RDF bricks, having a lower heating value of 17 MJ/kg
[36], delivers 1.33E ț16 J of energy that is converted (with an effi-
ciency of 30%) to 1.11E ț09 kW h of electricity. The emission factors
for RDF brick combustion can be found in Ref. [34].
Finally, the heavy waste fraction and the ashes (as a residue of
combustion as well as of the flue gas treatment) from RDF
combustion are sent to landfill, while ferrous metals are recovered.
In this scenario, the most relevant flows are the equipment for flue
gas cleaning system (urea, activated carbon, CaO, Ca(OH) 2) and the
energy for biogas enrichment (see Table 1 ).
3.4. Scenario 3: incineration
Waste is directly transported to the incineration plant in order
to recover electricity through combustion (further information
about incineration plant in Refs. [37–39] ). The non-differentiated
waste fraction has an LHV of 8.85 MJ/kg [36] and can generate
1.29Eț 10 J of energy. This energy is converted to 9.67E ț08 kW h of
electricity at 27% efficiency. Emission factors for MSW incinerationcome from Ref. [1]. Bottom ashes and flue gas treatment ashes are
delivered to landfill. According to Table 1 , the most relevant inputs
are the equipment for flue gas cleaning system (urea, activated
carbon, CaO, Ca(OH)
2).
4. Results and discussion
4.1. MFA
Table 3 shows the main product outputs of each scenario (waste
disposed of, electricity, upgraded biogas or ferrous metals) in the
second column, their abiotic MI in the fourth column and their
water MI in the fifth column.
Table 3 can be read as follows: to dispose of 1 g of waste with, for
example, Scenario 2, 0.3 g of abiotic materials must be altered
(extracted, moved, used .). This means that the disposal of 1 g of
waste in the city of Roma requires the production, somewhere elsein the world, of 0.3 g of further waste. This is a sort of cycle of waste
production that can be broken only by means of waste preventionpolicies. Similarly, the water MIs represent the water demand of
each process, i.e. the amount of water that is altered (moved,
heated, deviated and so on) to feed the process.
Among the investigated strategies, Scenarios 0 and 1 show
similar results for the abiotic MIs, while Scenarios 2 and 3 have the
highest values.
Concerning the MIs for electricity production, 334 g of abiotic
matter are required to produce 1 kW h of electric energy in
Scenario 2 (and 552 g
abfor Scenario 3). These values are lower than
those for electricity from coal, 5137 g ab, and comparable, at least inScenario 2, to electricity from oil and natural gas, respectively 349
and 274 g ab[15]. Even the MI of enriched biogas (97.3% CH 4) is very
close to that of natural gas, 1.22 g ab[15]. In Scenario 1, the water MI
for electricity production is lower than other scenarios because
both Scenarios 2 and 3 are affected by an elevated water
consumption at the treatment plants and by a high demand of CaO
and urea (which have high water factors). In fact, their contribu-
tions to the total water MI are, in Scenario 3, 20%, 35% and 13%,
respectively. Scenario 1 does not have such inputs (see Table 1 )
because it just involves the combustion of collected biogas.However, water MIs of Scenarios 2 and 3 for electricity generation
are in line with those of electricity produced from conventional
fossil sources.
Finally, a comparison among the final amounts of waste sent to
landfill by each scenario can be done. The quantities of final wastelandfilled are the following: 100% for Scenarios 0 and 1, 15% for 2
and 20% for 3. Despite none of them is able to avoid landfilling, in
Scenarios 2 and 3 there is a strong mass reduction (up to 80% less
waste to final disposal) and a change in waste composition and
reactivity. In fact, they are made of inert or burnt materials, which
do not undergo decomposition (with no release of CH
4and other
landfill gas to the atmosphere), although they are still responsiblefor the presence of heavy metals in the leachate.
4.2. GER
The GERs of the analysed scenarios are compared in Table 4 ,
while Table 5 shows the possible contribution of each scenario’s
energy output to the energy demand of Roma (city energy
requirement: electricity 2.91E ț16 J/year, natural gas 4.51E ț16
J/year, energy for transports 1.38E ț17 J/year; source: [40]).
The fourth column in Table 4 indicates that about 53.5 J of
energy must be used up to dispose of 1 g of waste within Scenario 0.In Scenario 2, this value rises up to 9.7 kJ, while approximately
14 MJ of energy must be invested in order to generate 1 kW h of
electricity, also including as input energy the energy content of
waste itself (RDF, in this case). If wastes were not included as an
energy cost, since they can be assumed as a renewable source, these
figures would decrease to 66.8 J per g of waste disposed of and to0.5 MJ per kW h of electricity produced (fifth column of Table 4 ).
Thus, Scenario 1 requires less energy per g of waste than Scenario0 thanks to the feedback of electricity produced from collected
biogas. These indicators also point out that Scenario 2 requires
about 20% more non-renewable energy (mainly fossil) than
Scenario 0 for providing the same service, i.e. the waste disposal;
however, it is also able to release 5 kJ of energy (electricity and
biogas) per g of waste treated, which can replace an equivalent
amount of fossil energy. Therefore, it is important to relate the
invested energy with the energy output delivered by the system
(the parameter ‘‘energy output’’ J
out/gwaste treated ofTable 4 ), by
means of an energy recovering efficiency index (the ratio betweenOrganic
FractionMSW collectionMSW sorting
phaseRDF
production
plant
Anaerobic
digestion and
upgradingCombustion of
RDFElectricity
Landfill Ashes
Biogas
Heavy wastes
to landfillFerrous metals
to recycling
Compostable
matterInorganic
Fraction
Fig. 1. Diagram of the main steps involved in Scenario 2.F. Cherubini et al. / Energy 34 (2009) 2116–2123 2120

the energy output and the total energy input, ‘‘E out/Ein’’). Based on
above data, it is possible to state that Scenario 2 appears as the best
waste management option: it produces more than twice the energyoutput of Scenario 3 and has the highest energy recovery efficiency.
The explanation for such a striking result is that this is the only
scenario which takes into account both components of waste, the
organic one, to produce biogas, and the inorganic one, to produce
electricity via RDF combustion. Instead, Scenario 1 only exploits the
organic part (landfill gas) and Scenario 3 the inorganic part (direct
combustion).
A comparison among the energy demand for the production of
1 kW h of electricity from different sources can also be performed.
The energy requirements per kW h of electricity produced from
Scenarios 1, 2 and 3 are shown in the fourth and fifth columns of
Table 4 ; these values are comparable with the energy demand for
electricity production by conventional sources, such as oil(1.24E ț07 J/kW h), coal (1.21E ț07 J/kW h), natural gas (9.50E ț06
J/kW h) and hydro (4.72E ț06 J/kW h) [41].
Table 5 shows that Scenario 2 could meet an important fraction
of Roma energy demand (15.47% of electricity and 8.24% of naturalgas) and would contribute to reach a target of 6.22% of renewable
sources in the total city energy consumption (also including thecontribution from hydroelectricity). If only transport sector is
considered, its renewable energy fraction could be equal to 5.3%,
very close to the EU target of 5.75%, which should be achieved bythe year 2010.
4.3. Ecological footprint
The ecological footprint of each scenario has been determined
using the software tool SPIonexcel and the results are shown in
Fig. 2 . A distinction is made between the gross and the net
ecological footprint. The gross ecological footprint is the total arearequired by each scenario to absorb or neutralize its emissions into
the ecosphere and provides its inputs from the environment (the
resources exploited), without considering the benefits of co-prod-
ucts. Landfill scenarios (0 and 1) have the lowest values because
they require no significant inputs and their ecological footprint is
mainly constituted by landfill methane emission (around 85% of the
total ecological footprint). In particular, Scenario 1 requires
a smaller area than 0 thanks to the electricity feedback produced by
biogas combustion. Instead, Scenarios 2 and 3 show higher values,
since the combustion step strongly affects final results: CO
2from
fossil carbon (mainly coming from the plastic waste fraction) andNO
xemissions are responsible for about 70% and 10%, respectively,
of the total area required by Scenarios 2 and 3; an important role isalso played by the provision of lime (CaO), used in flue gas cleaning
system, which is responsible for about 7% of their total ecological
footprint.
The net ecological footprint has been calculated by including
co-products in the evaluation. In fact, Scenarios 1, 2 and 3 also
co-produce energy outputs (80 kW h
e/ton waste for Scenario 1,
685 kW h e/ton waste and 52 kg biogas /ton waste for Scenario 2,
594 kW h e/ton waste for Scenario 3) that are able to replace an
equivalent energy amount coming from conventional processes:
Italian electric mix for electricity, having a specific footprint factor
of 238.8 m2/kW h and natural gas instead of biogas, having
a specific footprint factor of 40,008 m2/ton. The resulting land-
benefits can thus be subtracted from the gross results. For example,Table 3
Performance indicators according to MFA method
Process ProductaUnit MFA
Abiotic MI (g ab/unit) Water MI (g wt/unit)
Scenario 0 Landfilled waste g 0.24 0.03
Scenario 1 Landfilled waste g 0.24 0.02
Electricity kW h 1899 0.82
Scenario 2 Landfilled waste g 0.30 2.09
Electricity kW h 334 2398
Biogas ( >97% CH 4) g 1.14 8.06
Ferrous metals g 0.13 1.00
Scenario 3 Landfilled waste g 0.36 1.04
Electricity kW h 552 1578
aTotal amount of product is not indicated, because focus is on input material flows per unit of output. In each scenario, the amount of landfilled waste as well as its physical–
chemical state can be different.
Table 4
Performance indicators according to GER method
Process ProductaUnit GER
Energyc(J/unit) Energyd(J/unit)
Scenario 0 Landfilled wastebg 53.51
Scenario 1 Landfilled wastebg 2.15E ț03 12.67
Electricity kW h 2.67E ț07 871.08
Energy output: 288 J/g waste treated
Energy efficiency ( Eout/Ein): 13%
Scenario 2 Landfilled wastebg 9.71E ț03 66.80
Electricity kW h 1.38E ț07 5.06E ț05
Biogas ( >97%CH 4) g 3.79E ț03
Ferrous metals g 2.06E ț01
Energy output: 5014 J/g waste treated
Energy efficiency ( Eout/Ein): 52%
Scenario 3 Landfilled wastebg 9.52E ț03 673.94
Electricity kW h 1.60E ț07 1.13E ț06
Energy output: 2140 J/g waste treated
Energy efficiency ( Eout/Ein): 22%
aTotal amount of product is not indicated, because focus is placed on input
energy per unit of product and output energy per unit of waste treated.
bIn each scenario, the amount of landfilled waste as well as its physical–chemical
state can be different.
cIncluding the energy content of the waste.
dWithout the energy content of the waste (for biogas the energy content of the
waste is not yet exploited).Table 5
Energy output and Rome energy demand
Scenario 1 Scenario 2 Scenario 3
Electricity produced (kW h) 1.17E ț08 9.98E ț08 8.67E ț08
Biogas produced (J) 3.72E ț15
Electricity out/electrical demand of Romaa1.81% 15.47% 13.44%
Biogas (97% CH 4) out/gas demand of Roma 8.24%
Achievable quota of renewable energy in Romab2.97% 6.22% 4.24%
Energy out/energy for transport in Roma 0.30% 5.30% 2.26%
aElectricity from hydroelectric plants and other renewable energies not included.
bIncluding electricity from hydroelectric plants and other renewable sources.F. Cherubini et al. / Energy 34 (2009) 2116–2123 2121

Scenario 1 has a gross ecological footprint of 0.65 ha/ton waste , but
thanks to its electricity output can replace 80 kW h e/ton waste of
conventional electricity, which would require 1.9 ha: the net
ecological footprint of Scenario 1 is thus – 1.3 ha/ton waste .
Fig. 2 shows that despite of Scenario 2 requires more area than
landfill scenarios, it is able to provide energy benefits that can
replace products (conventional electricity and natural gas) having
a significant land pressure: this brings to the highest negative
ecological footprint, which means that Scenario 2 ‘‘sets free’’ about
15 ha of area per ton of waste treated.
4.4. Emissions at global and local scale
The environmental performances of each scenario (expressed as
GWP, AP, EP and dioxin emission) are depicted in Table 6 . These
values refer to the larger global scale, while Table 7 shows the
estimated airborne emissions at plant local scale.
Global impacts are reported both as gross and net impacts. The
gross impact refers to the effective emissions released by each
scenario during its life cycle, without considering the environ-
mental benefits coming from energy outputs. Instead, the net
impact refers to the ‘‘Total emission for each scenario’’ minus
‘‘Avoided emissions thanks to the energy output of each scenario’’.
For Scenario 0, net and gross emissions coincide because no useful
outputs are provided. If electricity is produced, we consider that it
replaces the electricity derived from the Italian electric grid, while
biogas substitutes natural gas.
Impacts of Scenarios 0 and 1 are strongly affected by landfill gas
emissions such as CH 4(equal to 90% of GWP for the first case and
about 30% for the second), H 2S, HCl and others, due to the anaerobicdecomposition of landfilled organic waste. The GWP of landfill
scenarios is even higher than the impact of Scenario 3 (direct
incineration). Therefore, from a GW point of view, burning is better
than burying. The critical point of a landfill is that it is not an iso-
lated system and it is not inert at all: it can be considered as
a chemical reactor which remains active for thousands of years. All
the out coming chemicals will inevitably pass the landfill system
and will undergo to environmental fate. In order to avoid/minimize
the environmental impact, a landfill needs to be monitored for
centuries.
Furthermore, Table 6 highlights that Scenario 2 is the best waste
management option since it has negative values for GWP, AP anddioxin emissions. Its implementation would therefore help
reducing GHG effect, acid rain deposition and dioxin emissions. The
difference in dioxin emissions between Scenarios 2 and 3 is due to
the fact that in Scenario 3 there is a co-firing of inorganic materials
(including Cl-rich plastics) with organic matter. This condition
favours the formation of dioxins (instead, in Scenario 2, only the
inorganic fraction of the waste is burnt).
Despite these global performances, the change from global to
local scale gives rise to different results. In fact, in Table 7 , the main
local airborne emissions are listed; they were calculated consid-ering the emissions at local scale, i.e. direct combustion in plant or
landfill site, excluding the electricity required to feed the scenarios
and the emissions for auxiliary material supply.
Results show that landfill options have lowest values except for
those of CH
4and H 2S, while Scenarios 2 and 3 have more relevant
emissions because of their combustion steps. Thus, there isa conflict between the global and local results: what is positive at
global scale (Scenario 2) it is not at local scale.
5. Conclusions
A LCA of different waste disposal strategies, i.e. landfilling with
(Scenario 1) and without (Scenario 0) biogas exploitation, sorting
plant to produce electricity via RDF and biogas via anaerobic
digestion (Scenario 2), and finally waste incineration (Scenario 3),
was performed by means of a multi-method multi-scale approach.
Results show landfill systems (Scenarios 0 and 1) as the worst
waste management options and that significant environmental
savings are achieved from undertaking energy recycling. The higher
the yields of energy recovered with waste disposal, the greater the
savings. Results also show that a sorting plant coupled with elec-
tricity and biogas production (Scenario 2) is very likely to be the
best option for waste management, despite the non-negligible
-15-12-9-6-3036
Scenario 0 Scenario 1 Scenario 2 Scenario 3ha/t waste
Gross
Net
Fig. 2. Comparison among the gross and net areas (in hectares) required by each
scenario to dispose of 1 ton of waste.
Table 6
Impact categories of the investigated scenarios
Scenario GWP, kt CO 2 AP, t SO 2 EP, t NO 3 Dioxins, g TCDD
Scenario 0
Gross 1914 546 126 0.24
Net
Scenario 1
Gross 966 338 126 0.35
Net 868 186 126 0.29
Scenario 2
Gross 704 852 n.a.a0.25
Net /C0340 /C0441 n.a.a/C00.28
Scenario 3
Gross 948 1902 n.a.a1.38
Net 224 780 n.a.a0.92
aFor these scenarios landfilled wastes are without a significant organic content.Table 7
Comparison among airborne emissions at a local scale (expressed in g)
Species (g) Scenario 0 Scenario 1 Scenario 2 Scenario 3
CO2 3.15E ț11 3.95E ț11 6.51E ț11 1.44E ț12
CO 1.60E ț08 7.38E ț07 3.92E ț07 2.11E ț09
NO x 2.45E ț07 6.27E ț07 5.43E ț08 1.09E ț09
PM10 1.13E ț05 5.02E ț05 7.83E ț06 2.60E ț07
SO2 1.13E ț08 1.14E ț08 3.13E ț07 8.10E ț08
CH4 7.55E ț10 3.78E ț10 1.30E ț05 2.01E ț04
N2O n.a. 5.27E ț03 2.35E ț07 1.17E ț08
HCl 4.80E ț07 2.75E ț07 2.74E ț07 1.61E ț08
HF 4.58E ț06 8.10E ț05 2.74E ț06 n.a.
H2S 1.60E ț08 8.02E ț07 n.a. n.a.
Hg 2.81E ț01 2.81E ț01 3.92E ț04 1.46E ț06
Pb 5.54E ț02 5.54E ț02 3.88E ț05 7.81E ț04
Cd 3.51E ț00 3.51E ț00 3.92E ț04 4.81E ț05
Cu 3.53E ț02 3.53E ț02 n.a. 4.81E ț04
Zn 9.07E ț02 9.07E ț02 n.a. 1.24E ț05
Ni 3.87E ț01 3.87E ț01 n.a. 4.90E ț03
Cr 1.00E ț02 1.00E ț02 n.a. 1.31E ț04
TCDDeq 0.24 0.35 0.19 1.38
PAH 1.01E ț02 1.01E ț02 n.a. 1.38E ț04F. Cherubini et al. / Energy 34 (2009) 2116–2123 2122

problem of local emissions (NO x, PM10, heavy metals, Polyciclic
Aromatic Hydrocarbons (PAH), among others). Furthermore, within
Scenario 2 and, to a smaller extent, Scenario 3, a non-negligible
amount of energy becomes available while waste residues to be
landfilled are minimized and inerted. The energy outputs of
Scenario 2 (i.e. electricity and biogas) are able to replace conven-
tional Italian electric mix and natural gas, thus providing environ-
mental benefits at global scale (e.g. GWP, AP) and a reduction of the
ecological footprint. In particular, energy and material input flows
required for the production of electricity within Scenario 2, are
comparable and to some extent competitive with other conven-
tional sources; in other words, an efficient waste-to-energy plant
with extensive emission controls can replace electricity generated
from an environmentally less than optimal power supply system.
Finally, if we have to choose only among the investigated
options, even the incineration alternative (Scenario 3) seems to be
better than landfilling (Scenarios 0 and 1), also from an environ-
mental impact point of view.
However, none of the investigated scenarios is completely able to
avoid the landfill. This is a second issue (in addition to uncontrolled
local emissions) that urgently calls for a breakthrough improve-ment: finding a convenient way for reuse of bottom and fly ashes
(produced in Scenarios 2 and 3) which are now delivered to landfill.
References
[1] Sundqvist JO, Finnveden G, Albertsson AC, Karlsson S, Berendson J,
Ho¨glund LO. Life cycle assessment and solid waste, AFRReport 173. Stockholm,
Sweden: AFR; 1997.
[2] Consoli F, Allen D, Boustead I, Fava J, Franklin W, Jensen AA, editors. Guidelines
for life-cycle assessment: a ‘Code of practice’. Society of Environmental Toxi-
cology and Chemistry (SETAC); 1993 [SETAC Workshop, Sesimbra, Portugal, 31
March–3 April 1993].
[3] Lindfors L-G, Christiansen K, Hoffmann L, Virtanen Y, Juntilla V, Hanssen OJ,
et al. Nordic guidelines on life-cycle assessment. Nord 1995:20. Copenhagen:Nordic Council of Ministers; 1995.
[4] LCA Guide. A guide to approaches, experiences and information source.
Environmental issue series n
/C146. European Environmental Agency; 1997.
[5] Finnveden G. Methodological aspects of life cycle assessment of integrated
solid waste management systems. Resources, Conservation and Recycling1999;26:173–87.
[6] Clift R, Doig A, Finnveden G. The application of life cycle assessment to inte-
grated waste management. Part 1. Methodology, Trans. IchemE
2000;78(B):279–87.
[7] Bjarnaddottir HJ, Fridriksson GB, Johnsen T, Sletsen H. Guidelines for the use of
LCA in the waste management sector. Nordtest TR 517. Espoo, Finland: Nordtest,
<http://www.p2pays.org/ref/37/36469.pdf >; 2002 [accessed 09/07/08].
[8] Cherubini F, Bargigli S, Ulgiati S. Life cycle assessment of urban waste
management: energy performances and environmental impacts. The case ofRome, Italy. Waste Management 2008;28:2552–64.
[9] Ekvall T, Assefa G, Bjo ¨rklund A, Eriksson O, Finnveden G. What life-cycle
assessment does and does not do in assessments of waste management. WasteManagement 2007;27(8):989–96.
[10] Winkler J, Bilitewski B. Comparative evaluation of life cycle assessment models
for solid waste management. Waste Management 2007;27(8):1021–31.
[11] Ulgiati S. Energy, emergy and embodied exergy: diverging or converging
approaches? In: Proceedings of the first biennial emergy analysis research
conference, 15–32/328, UFL, Gainesville, FL, USA: 2000.
[12] Ulgiati S, Raugei M, Bargigli S. Overcoming the inadequacy of single-criterion
approaches to life cycle assessment. Ecological Modelling 2006;190:432–42.
[13] Khan FI, Sadiq R, Husain T. GreenPro-I: a risk-based life cycle assessment and
decision-making. Environmental Modelling and Software 2002;17:669–92.
[14] Schmidt-Bleek F. MIPS re-visited. Fresenius Environmental Bulletin
1993;2:407–12.
[15] Hinterberger F, Stiller H. Energy and material flows. In: Ulgiati S, editor.
Advances in energy studies, energy flows in ecology and economy. Roma,
Italy: Musis Publisher; 1998. p. 275–86.
[16] Bargigli S, Raugei M, Ulgiati S. Mass flow analysis and mass-based indicators.
In: Jorgensen Sven E, Costanza Robert, Xu Fu-Liu, editors. Handbook ofecological indicators for assessment of ecosystem health. CRC Press; 2005. p.353–78.
[17] IFIAS. International federation of institutes for advanced study. In: Slesser M,
editor. Energy analysis workshop on methodology and conventions. Report
IFIAS no.89; 1974. Stockholm.
[18] Herendeen RA. Embodied energy, embodied everything .now what? In:
Ulgiati S, editor. Advances in energy studies, energy flows in ecology andeconomy. Roma, Italy: Musis Publisher; 1998. p. 275–86.
[19] Krotschak C, Narodoslawsky M. The sustainable process index – a new
dimension in ecological evaluation. Ecological Engineering 1996;6(4):241.
[20] Narodoslawsky M, Krotschak C. Integrated ecological optimization of
processes with the sustainable process index. Waste Management
2000;20(8):599–603.
[21] Sandholzer D, Narodoslawsky M. SPIonExcel – fast and easy calculation of the
sustainable process index via computer. Resources, Conservation and Recy-cling 2007;50(2):130–42.
[22] Krotscheck C. How to measure sustainability? Comparison of flow based
(mass and/or energy) highly aggregated indicators for eco-compatibility.EnvironMetrics 1997;8:661.
[23] AMA. Environmental report, <http://www.amaroma.it/index.
php?option ¼com_content&task¼ view&id ¼294&Itemid ¼136>; 2003 [in
Italian, accessed 09/07/08].
[24] Erses AS, Onay TT, Yenigun O. Comparison of aerobic and anaerobic degra-
dation of municipal solid waste in bioreactor landfills. Bioresource Technology
2008;99(13):5418–26.
[25] Obersteiner G, Binner E, Mostbauer P, Salhofer S. Landfill modelling in LCA –
a contribution based on empirical data. Waste Management 2007;27(8):
S58–74.
[26] Nielsen PH, Hauschild M. Product specific emissions from municipal solid
waste landfills, part I. Landfill model. International Journal of Life CycleAssessment 1998;3:158–68.
[27] White PR, Franke M, Hindle P. Integrated solid waste management – a life
cycle inventory. Gaithersburg, MD, USA: Aspen Publishers Inc.; 1999. New
York: Chapman & Hall.
[28] USEPA. Locating and estimating air emissions from sources of dioxins and
furans. Research Triangle Park, NC: Office of Air Quality Planning andStandards; 1995. <http://www.epa.gov/ttn/chief/le/dioxin.pdf >[accessed
09/07/08].
[29] Kylefors K. Evaluation of leachate composition by multivariate data analysis
(MVDA). Journal of Environmental Management 2003;68:367–76.
[30] EPA document. Compilation of air pollutant emission factors. vol. I, 5th ed,
point sources AP-42. <http://www.epa.gov/ttn/chief/ap42/ >[accessed 09/
07/08].
[31] Conte I. La produzione di energia dal biogas della discarica Basse di Stura,
Energy Manager Amiat S.p.A., Technical Report; 2001 [in Italian] <http://
www .amiat.it/File/abstr3.pdf >[accessed
09/07/08].
[32] Caputo AC, Pelagagge PM. RDF production plants: I. Design and costs. Applied
Thermal Engineering 2002;22:423–37.
[33] Chang YH, Chen WC, Chang NB. Comparative evaluation of RDF and MSW
incineration. Journal of Hazardous Materials 1998;58(3):33–45.
[34] Consonni S, Giugliano M, Grosso M. Alternative strategies for energy recovery
from municipal solid waste Part B: emission and cost estimates. WasteManagement 2005;25:137–48.
[35] Berglund M, Borjesson P. Assessment of energy performance in the life-cycle
of biogas production. Biomass and Bioenergy 2006;30:254–66.
[36] Arena U, Mastellone ML, Perugini F. The environmental performance of
alternative solid waste management options: a life cycle assessment study.Chemical Engineering Journal 2003;96:207–22.
[37] Ruth LA. Energy from municipal solid waste: a comparison with coal
combustion technology. Progress in Energy and Combustion Science1998;24:545–64.
[38] Liang L, Sun R, Fei J, Wu S, Liu X, Dai K, et al. Experimental study on effects of
moisture content on combustion characteristics of simulated municipal solid
wastes in a fixed bed. Bioresource Technology 2008;99(15):7238–46.
[39] Chen JC, Huang JS, Chen CM, Guo JS. Emission characteristics of PAHs, benzene
and phenol group hydrocarbons in O2/RFG waste incineration processes. Fuel2008;87(12):2787–97.
[40] Statistical yearbook of Rome. [Annuario Statistico di Roma 2003]. Roma:
Department XVII, Statistical Office, <www.comune.roma.it/uffstat/ >; 2003 [in
Italian, accessed 09/07/08].
[41] Frischknecht R, Jungbluth N, Althaus HJ, Bauer C, Doka G, Dones R, et al.
Implementation of life cycle impact assessment methods. Final report ecoin-
vent 2007. Duebendorf, CH: Swiss Centre for LCI; 2003. <http://www.
ecoinvent.org/fileadmin/documents/en/03_LCIA-Implementation.pdf >
[accessed 09/07/08].
[42] Bjorklund J, Geber U, Rydberg T. Emergy analysis of municipal wastewater
treatment and generation of electricity by digestion of sewage sludge.Resources, Conservation and Recycling 2001;31:293–316.F. Cherubini et al. / Energy 34 (2009) 2116–2123 2123