ARTICLE IN PRESS

Energy Policy 34 (2006) 32683283

www.elsevier.com/locate/enpol

Outlook for advanced biofuels

Carlo N Hamelincka,,1, Andre P.C. Faaijb
a

Azielaan 774, 3526 SZ Utrecht, the Netherlands

Utrecht University, Copernicus Institute, Heidelberglaan 2, 3584 CS Utrecht, The Netherlands

Available online 8 August 2005

Abstract
To assess which biofuels have the better potential for the short-term or the longer term (2030), and what developments are
necessary to improve the performance of biofuels, the production of four promising biofuelsmethanol, ethanol, hydrogen, and
synthetic dieselis systematically analysed. This present paper summarises, normalises and compares earlier reported work. First,
the key technologies for the production of these fuels, such as gasication, gas processing, synthesis, hydrolysis, and fermentation,
and their improvement options are studied and modelled. Then, the production facilitys technological and economic performance is
analysed, applying variations in technology and scale. Finally, likely biofuels chains (including distribution to cars, and end-use) are
compared on an equal economic basis, such as costs per kilometre driven. Production costs of these fuels range 1622 h/GJHHV now,
down to 913 h/GJHHV in future (2030). This performance assumes both certain technological developments as well as the
availability of biomass at 3 h/GJHHV. The feedstock costs strongly inuence the resulting biofuel costs by 23 h/GJfuel for each h/
GJHHV feedstock difference. In biomass producing regions such as Latin America or the former USSR, the four fuels could be
produced at 711 h/GJHHV compared to diesel and gasoline costs of 7 and 8 h/GJ (excluding distribution, excise and VAT; at crude
oil prices of
35 h/bbl or 5.7 h/GJ). The uncertainties in the biofuels production costs of the four selected biofuels are 1530%.
When applied in cars, biofuels have driving costs in ICEVs of about 0.180.24 h/km now (fuel excise duty and VAT excluded) and
may be about 0.18 in future. The cars contribution to these costs is much larger than the fuels contribution. Large-scale
gasication, thorough gas cleaning, and micro-biological processes for hydrolysis and fermentation are key major elds for RD&D
efforts, next to consistent market development and larger scale deployment of those technologies.
r 2005 Elsevier Ltd. All rights reserved.
Keywords: Biofuels; Prospects; Well-to-wheel

1. Introduction
Bioenergy is seen as one of the key options to mitigate
greenhouse gas emissions and to substitute fossil
fuels (Hall et al., 1993; Goldemberg, 2000). Largescale introduction of biomass energy could contribute
to sustainable development on several fronts, environmentally, socially, and economic (Ravindranath
and Hall, 1995; Turkenburg, 2000; van den Broek,
2000).

Abundant biomass resources are available in most

parts of the world (Rogner, 2000). The present
bioenergy use covers 914% of the global demand (of
about 400 EJ in 1998), most of which as traditional, lowtech and inefcient cooking and heating in developing
countries (Hall et al., 1993; Turkenburg, 2000). Modern
production of energy carriers from biomass (heat,
electricity and fuels for transportation or biofuels)2
contributes a lower, but signicant 7 EJ (Turkenburg,
2000). Different global energy scenario studies indicate
that in this century biomass may contribute much
more: up to 30% of the 2100 energy supply to

Corresponding author. Tel.: +31 30 2897414.

E-mail address: carlohamelinck@yahoo.com (C.N. Hamelinck).

Currently employed with Ecofys, P.O. Box 8408, 3503 RK Utrecht,
the Netherlands.
1

0301-4215/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enpol.2005.06.012

2
Biofuels means liquid or gaseous fuel for transport produced from
biomass (EU, 2003).

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

biomass (Intergovernmental Panel on Climate Change,

2000), an average 50250 EJ/yr in 2060 (2003); a
global (technical) potential of primary biomass in
2050 of 331135 EJ/yr (Hoogwijk et al., 2003a),
depending on population growth and food demand
(diet), economic development, food production
efciency, energy crop productivity on various land
types, competing biomaterial products, and land use
choices. This range may be narrowed to 300675 EJ
or 4060% of the energy demand in 2050, which could
be produced on 410% of the terrestrial surface
(Hoogwijk et al., 2003c). Large-scale bioenergy production also includes environmental, social and
economic risks. They have not been evaluated here,
but deserve attention in discussions about sustainability
of bioenergy.
Biomass is currently almost exclusively used for
the generation of heat and power. Only some 0.56 EJ
of the biofuels are produced worldwide, largely ethanol
from sugar/starch and a small amount of biodiesel
from oil crops (His, 2004), of which 62 PJ in the
European Union (EurObserER, 2004). Also, the
possibilities for electricity and heat production
from biomass (esp. combustion) are well known (Faaij,
1997; van den Broek, 2000) and widely applied
in various markets. But, while there are many

3269

other new or renewable technologies emerging

for large-scale electricity production with low or no
carbon emissions, the transportation sector, which is
almost entirely based on fossil oil use, has fewer
alternatives.
Transportation represents about 27% of the worlds
secondary energy consumption (21% of primary) and is
almost exclusively fuelled by mineral oil. The share may
increase to 2932% in 2050 (Intergovernmental Panel
on Climate Change, 2000; IMAGE-team, 2001; EIA,
2003). The rapidly increasing demand for transportation
fuels is combined with rapidly decreasing mineral oil
reserves of non-OPEC states. This increases dependency
on a limited number of oil-providing countries (Rogner,
2000), with inherent risks for energy security and sudden
price distortions.
Biofuels can play an important role in addressing
both the greenhouse gas emissions of transport and the
dependency on mineral oil.
A few main routes can be distinguished to produce
biofuels: extraction of vegetable oils, fermentation of
sugars to alcohol, gasication and chemical synthesis,
and direct liquefaction. Many eventual fuels are
conceivable: methanol, ethanol, hydrogen, synthetic
diesel, biodiesel, and bio oil (Fig. 1). All have very
different properties.
Hydrogen
(H2)
(H
2)

Water gas shift

+ separation
Gasification

DME
(CH3OCH3)
(CH3OCH3)

Syngas
Catalysed
synthesis

Lignocellulosic
biomass

Anaerobic
digestion

Purification

SNG
(CH4)
(CH4)

Bio oil

Hydro treating
and refining

Biodiesel
(CxHy)
(C
xH y )

Sugar

Fermentation

Ethanol
(CH3CH2OH)
(CH3CH2OH)

Esterification

Biodiesel
(alkyl esters)

Hydrothermal
liquefaction

Sugar/starch
crops

Oil plants

Milling and
hydrolysis
Pressing or
Extraction

FT Diesel
(CxHy)
(C
xH y)

Biogas

Flash pyrolysis

Hydrolysis

Methanol
(CH3OH)
(CH3OH)

Vegetable oil

Fig. 1. Overview of conversion routes from crops to biofuels (Elam, 1996; van den Broek, 2000).

Bio oil
(vegetable oil)

ARTICLE IN PRESS
3270

C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

1.1. Advanced fuels

With the exception of sugar cane ethanol, the shortterm (van den Broek et al., 2003) or traditional
(Turkenburg, 2000) biofuels have a number of severe
disadvantages that are related to the feedstock. The
current costs of rapeseed biodiesel and ethanol from
cereals or beets are much higher than the costs of
gasoline and diesel. Substantial subsidies are needed to
make them competitive. These high costs are a result of
the low net energy yield of most annual crops
(100200 GJ/ha yr in the long-term), the high-quality
(valuable) agricultural land required, and the intensive
management. The lower productivity per hectare and
high fertilizer requirement also limit the well-to-wheel
reduction of fossil energy use, and the environmental
benets are also limited (Ranney and Mann, 1994; Van
Zeijts et al., 1994; van den Broek, 2000; van Thuijl, 2002;
Berndes et al., 2003).
The net energy yield of perennial crops (220550 GJ/
ha yr), grasses (220260) and sugar cane (400500) is
much higher. These crops can be grown on less valuable
land (Rogner, 2000). Compared with sugar, starch, and
oil crops, the application of lignocellulosic biomass (e.g.
wood and grasses) is more favourable and gives better
economic prospects to the future of biofuels. Also, more
types of feedstock are in principle suitable to produce a
broader range of fuels than when applying traditional
biofuels feedstock.
Higher overall (production, distribution and use)
energy conversion efciencies and lower overall costs
are the key criteria for selecting biofuels for the longer
term. Various options are considered/developed that
have good potentials. Key examples are ethanol
produced from lignocellulosic biomass, synthetic diesel
via FischerTropsch (FT), methanol, and hydrogen
(Katofsky, 1993; Williams et al., 1995; Arthur, 1999;
Turkenburg, 2000). But the research, development and
demonstration status of all these fuels vary considerably.
These fuels are selected for a detailed analysis of their
long-term perspectives and RD&D needs.
1.2. Production of the selected biofuels
Methanol, hydrogen and FT diesel can be produced
from biomass via gasication. Several routes involving
conventional, commercial, or advanced technologies
under development are possible. Fig. 2 pictures a
generic conversion owsheet for this category of
processes. A train of processes to convert biomass to
required gas specications precedes the methanol or FT
reactor, or hydrogen separation. The gasier produces
syngas, a mixture of CO and H2, and few other
compounds. The syngas then undergoes a series of
chemical reactions. The equipment downstream of the
gasier for conversion to H2, methanol or FT diesel is

the same as that used to make these products from

natural gas (Williams et al., 1995), except for the gas
cleaning train. A gas turbine or boiler, and a steam
turbine optionally employ the unconverted gas for
electricity co-production.
Ethanol, instead, is produced via (largely) biochemical
processes (Fig. 3). Biomass is generally pretreated by
mechanical and physical actions (steam) to clean and
size the biomass, and destroy its cell structure to make it
more accessible to further chemical or biological
treatment. Also, the lignin part of the biomass is
removed, and the hemicellulose is hydrolysed (saccharied) to monomeric and oligomeric sugars. The cellulose
can then be hydrolysed to glucose. The sugars are
fermented to ethanol, which is to be puried and
dehydrated. Two pathways are possible towards future
processes: a continuing consolidation of hydrolysisfermentation reactions in fewer reactor vessels and with
fewer micro-organisms, or an optimisation of separate
reactions. As only the cellulose and hemicellulose can be
used in the process, the lignin is used for power
production.

1.3. Status of knowledge

Several studies exist that provide an overview of (part
of) the biofuels eld (Van Zeijts et al., 1994; Arthur,
1999; van den Broek et al., 2003). Other studies present
the techno-economic performance of individual biofuels
(Katofsky, 1993; Williams et al., 1995; De Jager et al.,
1998; Wooley et al., 1999). However, often these studies
incorporate only small biomass input scales (authors
assume that large scale is a priori not feasible) and
existing technologies. The potential for a better performance that could be obtained by applying improved or
new (non-commercial) technologies, combined fuel and
power production, and increasing scale giving higher
efciencies and lower unit capital costs, has not
exhaustively been explored.
Another problem is the comparability of the results.
The capital analysis for biofuels producing facilities has
been done in different ways. The data quality is very
variable. Also, the level of detail in analyses varies
enormously: from very supercial, to thorough plant
analysis. In either case the inuence of individual
parameters (e.g. feedstock costs) on the nal product
price is unclear.
This study is for a large part based on our earlier
reported techno-economic studies on the production of
methanol and hydrogen (Hamelinck and Faaij, 2002),
FT diesel (Tijmensen et al., 2002; Hamelinck et al.,
2004a) and ethanol (Hamelinck et al., 2005), and on the
long-distance transport of biomass (Hamelinck et al.,
2004b), which have resulted in a PhD thesis (Hamelinck,
2004).

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

3271

Recycle

Biomass

Drying and
Chipping

Gasification and
gas cleaning

Reforming, Shifting,
CO2 separation

Catalysis,
Separation

Methanol

Separation

Hydrogen

Catalysis,
Separation

Refining

Gas Turbine or
boiler

FT Diesel

Electricity

Steam
Turbine

Fig. 2. Generic owsheet for methanol, hydrogen, or FT diesel production, via gasication of biomass.

Enzyme growth

Biomass

Chipping

Hemicellulose
hydrolysis

Cellulose
Hydrolysis

Fermentation

Distillation

Gas Turbine or
boiler

Ethanol

Electricity

Steam Turbine

Fig. 3. Ethanol production by hydrolysis fermentation schematically.

1.4. Objective
The central questions are as follows: Which of the
biofuel options have the better potential for the shortterm and which have the best long-term (2030)
prospects? And what developments are necessary to
improve the performance of advanced biofuels production and use?
To answer these questions the short- and long-term
technological and economic performance of biofuels are
analysed and compared. This present study summarises
and normalises results for the four selected fuels
(methanol, ethanol, hydrogen, and synthetic diesel)
from the named earlier studies, compares their well-towheel performance, and indicates the key factors
inuencing that performance. First, the key technologies
for the production of these fuels, such as gasication,
gas processing, synthesis, hydrolysis, and fermentation, and their improvement options are studied
and modelled. Then, the production facilitys technological and economic performance is analysed,
applying variations in technology and scale. Finally,
major biofuels chains (including distribution to cars,
and end-use) are compared on an equal economic
basis, such as costs per kilometre driven. The results

are compared with the reviewed performance of classic

biofuels such as rapeseed biodiesel and sugar/starch
ethanol.
It is assessed which factors inuence the fuel production and fuel chains performance most, and which
aspects are most uncertain. This gives insights both in
the possible barriers to implementation that need to be
overcome, and in the technological improvement
options that should be stimulated by research development and demonstration.

2. Research method
2.1. Modelling mass and energy balances
For analysing the production of methanol, hydrogen
and FT diesel, Aspen Plus (Aspen Technology Inc.,
2003) owsheet models were made and used for
optimisation purposes. The gasier, reformer and gas
turbine deliver heat, whereas the dryer, gasier, reformer, and water gas shift reactor require steam. The
supply and demand of heat (taking into account steam
conditions) is added to or drawn from the steam turbine,
such that the surplus heat is turned into electricity.

ARTICLE IN PRESS
3272

C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

Ethanol production was for the greater part modelled

in Excel (except for the power isle, which was modelled
in Aspen Plus). For each process step in the hydrolysis
fermentation process, conversion extents (of hydrocarbons to sugars, of sugars to ethanol, of energy
generation, etc.) and losses were applied, so that each
step yielded intermediate amounts of sugar, ethanol and
solid residuals.

Table 1
Unieda set of input parameters
Scale
Electricity price (supply and
demand)
Economic lifetime (depreciation
time)
Technical lifetime
Interest rate
Load
Investment path

2.2. Economic evaluation

400-2000 MWHHV input (shortterm-long-term)

0.03 h/kWhe
15 yr
25 yr
10%
8000 h (91% of time)
20% in rst year, 30% in second
and 50% in last year

The resulting mass and energy balances served as a

basis for economic evaluation. Fuel production costs are
calculated by dividing the total annual costs of a system
by the annually produced amount of fuel. The total
annual costs consist of annual capital costs, operating
and maintenance (including maintenance, consumables,
labour, waste handling), biomass feedstock costs and
costs of electricity supply/demand (xed power price).
The Total Capital Investment, or TCI, is calculated by
factored estimation, based on known costs for major
equipment as found in the literature or estimated by
experts, and translated to the actual equipments size.
Scaling-up has been done by using individual scale
factors for each piece of equipment. The uncertainty
range of such estimates is up to 730% (Peters and
Timmerhaus, 1980).
For some of the process equipment costs are known
from current practical reality (e.g. compressors, heat
exchangers, reactor vessels, reformers, etc.). Other costs
have been estimated in the literature, or by consulted
experts.
While largely the same method was applied in the
analysis of the selected fuels, there are differences in the
degree of detail. The methanol and hydrogen concepts
were rst chosen for their expected technical performance, then individually modelled and optimised for
efciency, and eventually economically evaluated; they
included far reaching heat integration. The model of the
second FT study allowed for far going optimisation, by
variation of gasier oxygen level and pressure, and
applying many step-by-step process conguration
changes, but the heat integration focussed on the major
heat sources and sinks. In the FT process, pressure,
temperature, and (relative) concentration of the reactants inuence the product quality and yield, and this
was also accounted for. Ethanol production has not
been modelled in Aspen Plus (though the power isle was)
because it incorporates biological rather than chemical
processes. They were analysed at a lower level of detail
because only generic information (mostly conversion
efciencies per step) was available. The studies have
been made comparable by applying a unied set of input
parameters (Table 1).

Several input parameters were slightly different between the

original papers, which may complicate direct comparison. Therefore,
the results have been recalculated from the separate studies by using
this unied set of parameters.

2.3. Feedstock costs

The biomass feedstock costs are a major input
parameter for the calculation of the biofuel production
costs. Hoogwijk et al. (2003b) found that in the year
2050, 170290 EJHHV may annually be produced at costs
below 2 h3/GJ4HHV. In Latin America, Africa, Asia and
Eastern Europe signicant amounts of biomass are
already supplied at 0.52.0 h/GJHHV (Williams and
Larson, 1993; Marrison and Larson, 1995, 1996; Perlack
and Wright, 1995; Phillips et al. 1995; Azar and Larson,
2000; van den Broek, 2000). For the rst decade, forest
or agricultural residues and waste will play an important
role, as they are potentially available in large amounts
(3090 EJ/yr in 20202050) at mostly low costs. Largescale international transport from bioenergy producing
regions to Western Europe can be done via a broad
variety of chains, comprising different biomass production systems, pre-treatment and conversion operations,
and transport of raw (chips, logs, bales) and rened
biomass (pellets) by different means (truck, train, ship).
The projected costs for delivered biomass feedstock
has been analysed separately (Hamelinck et al., 2004b).
It has been shown that 300 MWHHV (8.6 PJHHV over
8000 h) solid biomass in the compressed form (pellets or
bales) could be delivered to a Western European
location from Latin America (11,000 km sea transport)
at about 3.1 h/GJHHV, adding 2 h/GJHHV to the local
production costs of 1 h/GJ (Hamelinck et al., 2004b), of
which only 0.5 h/GJHHV is in the actual international
shipping. The present paper further assumes that
3
All costs in this introductory essay are in h2003, ination 2.5%
annually, 1 h2003 1 US$2003.
4
Energy is preferably expressed on Higher Heating Value basis,
indicated by the subscript HHV. LHV indicates Lower Heating Value,
if no subscript is given, the denition is unknown.

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

biomass in compressed form can be delivered in Western

Europe at 3 h/GJHHV. In biomass producing regions as
Brazil and Eastern Europe, biomass can be delivered to
local plants at 2 h/GJHHV.

3273

short-term. Plausible improvement options were summarised, and their effect estimated. The concepts from
the background articles that will be compared here are
summarised in Table 2.

2.4. Selection for comparison

Based on the earlier broader evaluations, for each of
the four fuels we select one concept for the short-term,
and one concept for the longer term. For the short-term,
we choose the best performing from concepts that are
possible with currently available technology. For the
long-term, we compare the ultimately best performing
concepts, which may need further technological development. The study on methanol and hydrogen (Hamelinck and Faaij, 2002) yielded six promising concepts for
the production of methanol and ve for the production
of hydrogen. About half of these concepts made use of
technologies already available now, the other half
assumed technologies that are still under development.
Of each category, the best performing concept was
chosen. The study on ethanol (Hamelinck et al., 2005) in
contrast yielded a development pathway with one
plausible concept for the short-, one for the medium-,
and one for the long-term. The study on FT diesel, via a
tree of process choices and directly visible economic
results, yielded one best performing concept for the

3. Results and discussion

3.1. Technological insights
3.1.1. Gasification-based fuels
The ndings of the previously published papers can be
summarised as follows: gasication-based fuel production systems that apply pressurised gasiers have higher
joint fuel and electricity energy conversion efciencies
than atmospheric gasier-based systems. The total
efciency is also higher for once-through congurations,
than for recycling congurations that aim at maximising
fuel output. This effect is strongest for FT production,
where (costly) syngas recycling not only introduces
temperature and pressure leaps, but also material leaps
by reforming part of the product back to syngas. For
methanol and hydrogen, however, maximised fuel
production, with little or no electricity co-production,
generally performs economically somewhat better than
once-through concepts.

Table 2
Denition of the selected processes for biofuels production
Fuel
Methanola

Technical description
Now
Future

Ethanolb

Now
Future

Hydrogenc

Now
Future

FT dieseld

Now
Future

Atmospheric indirect gasier, wet gas cleaning, steam reforming (partly fed by off gas), shift reactor, lowpressure gas phase methanol reactor with recycle, and a steam turbine
Atmospheric indirect gasier, wet gas cleaning, steam reforming (partly fed by off gas), a liquid phase
methanol reactor with steam addition and recycle, and a steam turbine
Dilute acid pre-treatment, on-site enzyme production, enzymatic cellulose hydrolysis, SSF conguration
(cellulose hydrolysis and C6 fermentation integrated in one reactor vessel), boiler and steam turbine
Liquid hot water pre-treatment, CBP conguration (enzyme production, enzymatic cellulose hydrolysis
and co-fermentation in one reactor vessel), boiler and steam turbine
Atmospheric indirect gasier, wet gas cleaning, shift reactor, pressure swing adsorption for H2 separation,
and a combined cycle
Pressurised direct oxygen red gasier, hot gas cleaning, ceramic membrane with (internal) shift, and a
combined cycle
Direct 25 bar oxygen red gasier, tar cracker, wet gas cleaning, no reforming, and once-through FT
synthesis at 60 bar with 90% conversion
Direct 25 bar oxygen red gasier, tar cracker, wet gas cleaning, no reforming, and once-through FT
synthesis at 60 bar with 90% conversion

a
Hamelinck and Faaij (2002) assessed six concepts for the production of methanol. Methanol now is the sixth of those concepts; it is the concept
that performs best with currently available technology. Methanol future is the fourth and overall best of those concepts, but it applies a technology
that is not yet available. The quoted results are compensated for 15% cost reduction through learning, which was not incorporated in that study.
b
Hamelinck et al. (2005) assessed three concepts for the production of ethanol via hydrolysis fermentation for short-, medium-, and long-term.
Ethanol now is ethanol short-term concept, and ethanol future is the long-term concept of that study.
c
Hamelinck and Faaij (2002) assessed ve concepts for the production of hydrogen. Hydrogen now is the fth of those concepts; it is the concept
that performs best with currently available technology. Hydrogen future is the third and overall best of those concepts, but it applies technologies
that are not yet available. The quoted results are compensated for 15% cost reduction through learning, which was not incorporated in that study.
d
Hamelinck et al. iteratively assessed a broad range of concepts for the production of FT diesel (Hamelinck et al., 2004a). The FT diesel now
concept is the concept that was found to perform best; it incorporates technology that is currently available. The future concept is the same concept
with 15% and 5% cost reduction (learning+process improvement).

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

3274

Hot (dry) gas cleaning generally improves the total

efciency, but the economical effects are ambivalent,
since the investments also increase. Similarly, CO2
removal does increase the total efciency (and in the
FT reaction also the selectivity), but due to the
accompanying increase in investment costs this does
not decrease the product costs. The bulk of the capital
investment is in the gasication and oxygen production
system, syngas processing and power generation units.
These parts of the investment especially prot from
cost reductions at larger scales. Also, combinations with
enriched air gasication (eliminating the expensive
oxygen production assumed for some methanol and
hydrogen concepts) may reduce costs further.
Several technologies considered here are not yet fully
proven or commercially available. Pressurised (oxygen)
gasiers still need further development. At present, only
a few pressurised gasiers, operating at relatively small
scale, have proved to be reliable (Larson et al., 2001).
Consequently, the reliability of cost data for large-scale
gasiers is uncertain. A very critical step in all thermal
systems is gas cleaning. It still has to be proven whether
the (hot) gas cleaning section is able to meet the strict
cleaning requirements for reforming, shift and synthesis.
Liquid phase reactors (methanol and FT) are likely to
have better economies of scale. The development of
ceramic membrane technology is crucial to reach the
projected hydrogen cost level. For FT diesel production,
high CO conversion, either once through or after recycle
of unconverted gas, and high C5+ selectivity are

important for high overall energy efciencies. Several

units may be realised with higher efciencies than
considered in this paper: new catalysts and carrier
liquids could improve liquid phase methanol single pass
efciency. At larger scales, conversion and power
systems (especially the combined cycle) have higher
efciencies, but this has not been researched in depth.
3.1.2. Ethanol
The assumed conversion extent of (hemi)cellulose to
ethanol by hydrolysis fermentation is close to the
stoichiometric maximum. There is only little residual
material (mainly lignin), while the steam demand for the
chosen concepts is high. This makes the application of
BIG/CC unattractive at 400 MWHHV. Developments of
pre-treatment methods and the gradual ongoing reactor
integration are independent trends and it is plausible
that at least some of the improved performance will be
realised in the medium-term. The projected long-term
performance depends on development of technologies
that have not yet passed laboratory stage, and that may
be commercially available earlier or later than 20 yr
from now. This would mean either a more attractive
ethanol product cost in the medium-term, or a less
attractive cost in the long-term.
The investment costs for advanced hemicellulose
hydrolysis methods need to be assessed more exactly.
Continuing development of new micro-organisms is
required to ensure fermentation of xylose and arabinose,
and decrease the cellulase enzyme costs.

Table 3
Technological and economic performance of biomass to fuels facilities for now-future: efciencies to fuel and electricity, capital investment, scale
factor, annual O&M costs, and fuel production costs are summarised and recalculated from the background articles
ZHHVa (%)

Fuel

Methanol
Ethanol
Hydrogen
FT diesel

Now
Future
Now
Future
Now
Future
Now
Future

Fuel

Electricity

58.9
57.0
34.9
47.3
34.8
41.3
42.1
42.1

4.0
0.1
4.1
4.0
16.9
19.7
3.2
3.2

TCIb (Mh)

Rc

O&Md (% of TCI)

Production costse (h/GJHHV)

Now

235
188
291
218
247
207
292
235

0.79
0.84
0.84
0.82
0.81
0.86
0.85
0.85

4.0
4.0
6.4
3.6
4.0
4.0
4.4
4.4

future

Local future

12
9

11

13

11

22
16
18

Parameters hold at 400 MWHHV biomass input. The production costs for the future include a larger scale (2000 MWHHV input). Electricity buy/sell
costs 0.03 h/kWhe. Delivered feedstock costs 3 h/GJHHV (Western Europe), or 2 h/GJHHV (local in biomass producing region). The processes assume
wet (30% moisture) chipped biomass, drying to 1015% and pulverisation are included in the concepts.
a
Electricity is co-produced in most processes (Paper 2 also shows methanol concepts co-producing electricity). Some processes require extra
electricity.
b
From the Total Capital Investment (TCI) follows the Total Capital Requirement (TCR) assuming a correction for lifetime (90.4%) and
investment path (20%, 30% and 50%, in rst, second and last year: 118%). The methanol and hydrogen study (Paper 2) and the Fischer-Tropsch
study (Paper 4) did not include an investment path; the values presented here are therefore somewhat higher. The TCR is used for determining the
annual capital costs.
c
R value found for upscaling from 400 to 2000 MWHHV input, smaller R are found for downscaling.
d
O&M for the methanol and hydrogen processes is xed at 4%. In the FischerTropsch process, O&M consists of a xed part (4% of TCI) and a
part decreasing with scale (0.4% at 400 MWHHV, R 0:85). O&M in ethanol production is very dependent on cellulase required.
e
The time path also incorporates a scale increase: now: 400 MWHHV and future: 2000 MWHHV.

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

3.2. Selected processes

Table 3 gives the resulting key parameters for biomass
to fuel conversion facilities in the short- and long-term,
the latter includes process improvements and technological learning. In the short-term ethanol and FT diesel
facilities are the most expensive. These processes also
have the lowest total (fuel+electricity) efciency in the
short-term. This, in combination with high operating
costs, makes cellulose ethanol the most expensive of
these biofuels in the short-term. Through expected
process improvements, technological learning and scale
enlargement, production costs can decrease and/or
efciencies can go up. Technological learning can take
place through increasing production capacity and
experience.
A breakdown of the production costs into capital,
O&M, feedstock and power costs is shown in Fig. 4. A
30% uncertainty should be applied to the total capital
investment of methanol, hydrogen, and FT diesel
concepts. The capital costs for the ethanol concepts
are estimated to have a higher uncertainty (50%),

3275

because of the less detailed analysis. The uncertainty

bars in Fig. 4 show that the eventual inuence of these
uncertainties to the production costs could be up to
3050%.
The bare inuence of scale (for all concepts of Table
3) on the biofuel production costs is made visible in Fig.
5. Analysis of the curves yields different overall scale
factors for the production facilities capital investments
for the different technologies. Also, these scale factors
change over the whole 802000 MWHHV range. All
thermal gasication-based processes experience a stronger inuence of scale between 80 and 400 MWHHV, than
ethanol production. The hydrolysis fermentation takes
place in vessels that have a small maximum size,
decreasing scale advantages at larger scales compared
to gasication-based processes.
In the base situation (Fig. 4) feedstock (at 3 h/GJHHV)
accounts for 4558% of the total product costs. The
inuence of biomass feedstock price depends on the
conversion efciency from feedstock to fuel, e.g. a
ZHHV,fuel of 35% (ethanol-now) implies that with every h
feedstock cost reduction, the production costs reduce

25
Capital

Now

O&M

Future

Biomass
Costs/Income Power

15

Fuel production costs ( /GJHHV)

20

10

-5
Methanol

Ethanol

Hydrogen

FT diesel

Methanol

Ethanol

Hydrogen

FT diesel

Fig. 4. Breakdown of the production costs of selected biofuels (methanol, ethanol, hydrogen and FT diesel) now and in future. Feedstock costs 3 h/
GJHHV. Time path also incorporates a scale increase: now 400 MWHHV, medium-term: 1000 MWHHV, and ultimate: 2000 MWHHV. Uncertainty
ranges of 30% are applied to capital (and O&M, because this is a linear function of capital), 50% for the ethanol concepts.

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

3276

25

Fuel production costs ( /GJHHV)

20

15

10
MeOH now
MeOH future
EtOH now
EtOH future

H2 now

2000 MWHHV
and
2 /GJHHV
feedstock
costs

H2 future

FT now
FT future

0
0

400

800

1200

1600

2000

Input scale (MWHHV)

Fig. 5. Inuence of input scale on the production costs of selected biofuels. Feedstock costs 3 h/GJHHV, with a sensibility towards 2 h/GJHHV at
2000 MWHHV input.

with 1/0.35 or 2.9 h. A much more efcient process, such

as methanol now thus becomes relatively more
attractive at high feedstock costs (refer to Fig. 6). Since
cheap feedstock will be used rst (up to 3 h/GJ at gate),
process improvements may initially focus on capital,
O&M and power cost reduction.
3.3. Broader comparison
Table 4 presents the estimated production costs for a
broader range of biofuels, calculated from information
on TCI, scale, and efciencies, by using the same base
assumptions. For ethanol production from sugar cane a
3840 h operation window is used (limited harvest season
and impossibility to store cane for longer time) (Damen,
2001).
The calculated ethanol from sugar cane production
costs are higher than the current prices in South
East Brazil (79 h/GJHHV currently without subsidies
(Goldemberg et al., 2004)). The (large) difference may
be explained by the fact that many existing installations
were built with subsidies in the eighties, and are at
present completely depreciated, and by the direct link
between the sugar and ethanol production, which allows
for cost allocation over the two products. International
alcohol shipment from Latin America to Europe will not
add more than 0.5 h/GJ. If road tanker transport from

an inland production location to the harbour were

necessary, this would add half a euro extra per GJ per
100 km (Hamelinck et al., 2004b).
The calculated production costs of most fuels agree
with the literature: delivered costs are reported to
amount 2138 h/GJHHV for ethanol from wheat,
1529 h/GJHHV for RME, and 13 h/GJHHV for dimethylether (DME). Sugar beet ethanol is often projected
cheaper, than calculated here, at a delivered cost of
2640 h/GJHHV, and substitute natural gas (SNG) more
expensive at 13 h/GJHHV (Van Zeijts et al., 1994; Arthur,
1999; Reinhardt and Zemanek, 2000; van den Broek et
al., 2003).
The fuels could also be produced in the biomass
supplying countries, provided that the amount of
cheaply available biomass locally sufces, after which
the biofuel is internationally shipped. This has two
advantages: the elimination of the costly densication step (pellets) otherwise necessary to minimise
long-distance transport costs, and the higher energy
density of the transported commodity. The joint cost
reduction is about 9% for methanol (Hamelinck
et al., 2004b). Although ethanol has a higher
energy density than methanol, the cost advantage
will be practically the same. It can be concluded
that this route is very attractive for sugar cane ethanol
from Brazil.

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

3277

25

Fuel production costs ( /GJHHV)

20

15

10
MeOH now
MeOH future
EtOH now
EtOH future

H2 now
H2 future
FT now
FT future
0
0

Feedstock costs ( /GJHHV)

Fig. 6. Inuence of feedstock costs on the production costs of selected biofuels (400 MWHHV input).

3.3.1. Distribution and dispensing

The fuels physical properties have a strong inuence
on costs and energy use during distribution to the enduser (Table 5). The energy density of ethanol and
methanol is lower than that of gasoline and diesel, which
results in proportionally higher costs. Hydrogen is by
far the most challenging commodity to distribute and
store. Pipeline transport in gaseous form is the cheapest
and the most energy efcient approach at present, but
only suitable for large-scale, e.g. at national to regional
level. Supplying individual gas stations require the
transport of liquid hydrogen by road tankers, which
increases the costs (van Walwijk et al., 1996). Chemical
properties of individual fuels require special materials
and safety precautions. Alcohols and RME can best be
blended with gasoline and diesel right at the reneries,
or at intermediate depots between reneries and gas
stations (van den Broek et al., 2003). The choice depends
mainly on the biofuel production facilitys location(s),
on the blends stability (van den Broek et al., 2003), and,
in the case of alcohols, on whether or not they are also
dispensed as neat fuels.
The resulting delivered costs at the car are shown in
Fig. 7, left. In the short-term, lignocellulosic methanol
and sugar cane ethanol are the cheapest biofuels. They
may be delivered at 1314 h/GJHHV. RME from rapeseed is very expensive, but as has been reported by

others (van den Broek et al., 2003), it has a very broad

cost uncertainty range. Hydrogen has very high
distribution costs (De Jager et al., 1998; Ogden et al.,
1999), so end consumers costs are about 20 h/GJHHV,
which is comparable with FT diesel. In the long-term the
delivered costs range from 10 to 15 h/GJHHV for most
biofuels, but the traditional biofuels lag behind. For
comparison: gasoline over the last decennium cost
2.57.2 h/GJHHV at Rotterdam port and diesel
2.46.6 h/GJHHV (BP, 2003), while distribution to gas
stations adds about 1.4 h/GJHHV (van Walwijk et al.,
1996), so that the high ends of the delivered costs5 were
about 9 h/GJHHV. This corresponded with oil prices of
1234 h/bbl (1 bbl equals 6.12 GJHHV). At the current
crude oil prices of above 35 h/bbl or 5.7 h/GJ, diesel and
gasoline cost over 7 and 8 h/GJ (excluding distribution,
excise and VAT), at 50 h/bbl, diesel and gasoline costs.
As projections for future fossil fuels costs are highly

5
The sale price at the gas station further usually includes excise duty,
in the Netherlands this is currently 18 h/GJHHV (0.64 h/l) for gasoline
and 9 h/GJHHV (0.34 h/l) for diesel (van den Broek et al., 2003), and
value added tax (VAT, 20%). (Partial) duty exemption for biofuels
could make their price competitive with gasoline and diesel. VAT is
compulsory and the same percentage for all vehicle fuels. The current
sale prices in the Netherlands are about 33 h/GJHHV (1.15 h/l) for
gasoline and 21 h/GJHHV (0.80 h/l) for diesel.

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

3278

Table 4
Production costs of fuels from various crops, not analysed further in this paper, based on literature
Fuel

Ethanol

Biodiesel (RME)
DMEg
DMMh
SNGh

Feedstocka

Maizec
Wheatc
Sugar beetd
Sugar canee
Now
Sugar cane
Future
Rapeseedf
Lignocellulose
Lignocellulose
Lignocellulose

(h/GJHHV)

ZHHV (%)

TCI (Mh)

O&M (% of TCI)

Production costsb (h/GJHHV)

25
29
40
11

Fuel

Electricity

8.3
10
14
2.4

18
19
37
43

16
15
2.2
0

289
280
149
48

0.75
0.75
0.75
0.8

8.3
8.3
5.0
13

2.4

89

1.2

153

0.8

10

6.6
3
3
3

31
58
46.4
60

12
3.6
0
0

397
306
242
141

0.95
0.7
0.7
0.7

0.9
7
4
4

19
24
39

8
25
15
15
9

23
11
11
7

Technological and economic fuel production parameters for short-term (400 MWHHV input), and production costs for short-term-long-term.
Efciencies are on whole used feedstock basis (see notes 3, 4 and 6). Ethanol from cane at Brazil conditions, all other fuels European conditions.
a
Costs for maize, wheat and rapeseed are for grain+straw (straw is free at eld) (De Jager et al., 1998). All include 100 km truck transport of the
raw material (0.5 h/GJHHV). Sugar cane feedstock price at conversion installation in Latin America, leaves are left in the eld (Damen, 2001).
b
The production costs are recalculated from total capital requirement and efciency, assumptions: 8000 h load (3840 h for sugar cane), 10%
interest, 15 yr lifetime, the relation between TCR and TCI was explained in Table 3, note 2. Electricity buy/sell 0.03 h/kWhe, short-term 400 MWHHV
and long-term 2000 MWHHV input. Long-term includes cost reductions by learning (see respective notes).
c
Ethanol from maize and wheat by the same wet milling-fermentation process. Conversion grain to ethanol is 37-40% by HHV, conversion straw
to electricity is 37-45% by HHV (BIG-CC). The milling-fermentation process requires electricity 0.06-0.05 GJe/GJHHVfuel, and heat 0.240.20 GJth/GJHHVfuel. Efciencies and costs reported in this table are for the total process of grain and straw to ethanol and electricity. Investment for
the milling-fermenting part is 72-62 Mh for 104 MWHHV grain input, for the BIG/CC 2986-2398 h/kWhe. The co-produced fodder has a value,
which is settled with the O&M (Elam, 1996; De Jager et al., 1998; Faaij et al., 1998).
d
Ethanol from sugar beet conversion is 43%, but requires heat 0.35 GJth/GJHHVfuel and electricity 0.06 GJe/GJHHVfuel. The heat demand is
deducted from the feedstock, the electricity supplied by the grid. A 139 MWHHV input facility has TCR of 72 Mh now, and 10% reduction towards
future. Co-produced fodder has value, which is deducted from the O&M (Elam, 1996; De Jager et al., 1998).
e
Conversion efciency for an average Brazilian sugar/ethanol mill is 85 l/tonnewet (mc 73%) or 7.4 MJHHV/tonnedry or 42.8% by HHV (sugar cane
17.35 GJHHV/tonnedry) but could increase to 177 litre/tonnewet (mc 73%) or 89% HHV. Sugar mills assumed to produce or consume no extra energy
in the short-term, but in the long-term produce electricity 55.4 MJe/tonnewet (electric efciency 1.2% by HHV). TCR for a 523 MWHHV input facility
now is 63.2 Mh, for a future 1951 MWHHV input 581 Mh (Damen, 2001).
f
Efciencies and costs reported in this table are for the total process of seed and straw to RME and electricity. Conversion of rapeseed to RME is
49% by HHV. The extraction-esterication requires electricity 0.05 GJe/GJHHVfuel. A 1.875 MWHHV input esterication facility has a TCR of 2.11.9 Mh (10% cost reduction by learning), costs and efciencies for the BIG/CC as in note 3 (De Jager et al., 1998).
g
Dimethylether. Efciency and capital costs from Elam (Elam, 2002).
h
DMM is dimethoxymethane; SNG is substitute natural gas; Efciency and capital costs from Arthur D. Little (Arthur, 1999).

uncertain,6 they may well come in the range of biofuels

costs.
3.4. Biofuels use in cars
Different fuels perform differently in various drive
chains. Therefore, the various biofuels should be
6
The worlds oil supplies are not unlimited (Meadows et al., 1972),
and although the insights differ from increasing oil production until at
least 2025 (EIA, 2003), down to declining oil production within the
next 20 (Cavallo, 2002) or 510 yr (Campbell and Laherrere, 1998;
Deffeyes, 2001). The differences are partly explained by the denition
of supplies. This may or may not include unconventional oil that
although it cannot be exploited economically at presentcould
eventually be exploited economically depending on both technology
improvement and the markets demand (Rogner, 2000). After a
production peak, the world does not run out of oil, but the oil prices
likely rise (EIA, 2003). For comparison, over the last hundred years the
crude oil price has usually been between 1.6 and 4.9 h/GJ (BP, 2003),
and up to 13 h/GJ during the oil crisis.

Table 5
Costs assumed for distribution and dispensing different fuels from a
central production facility to gas stations (van Walwijk et al., 1996; De
Jager et al., 1998; Ogden et al., 1999)
Fuel

Costs (h/GJHHV)

Gasoline
Diesel
Methanol
Ethanol
Compressed hydrogen
LPG, DME

1.41
1.33
2.12
1.84
4.50
2.40

compared on an equal basis, such as driving costs, their

energy saving potential, or specic costs as CO2
reducing option. Moreover, the prospects for biofuels
cannot be seen apart from (the future of) vehicle
propulsion. A selection of fuels, shown in Fig. 7 (left),
is used for further comparison. Table 6 sums up the

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

10

15

20

25
Now

in FCV-PEM
in FCV-SR-PEM
in FCV-SR-PEM

Future

Future

Now

Gasoline incl . duty ex VAT

0.30

as E10 in ICEV-SI
in ICEV-SI
in ICEV-SI
in ICEV-CI
in ICEV-SI
in ICEV-CI

in ICEV-SI
in ICEV-CI
as RME5 in ICEV-CI
as E10 in ICEV-SI
in ICEV-SI
in ICEV-SI
in ICEV-CI
in ICEV-CI

lignocellulose ethanol
lignocellulose methanol
lignocellulose diesel (FT)
lignocellulose DME
lignocellulose hydrogen
fossil gasoline
fossil diesel
lignocellulose methanol
lignocellulose DME

0.25

as RME5 in ICEV-CI
as E10 in ICEV-SI
as E10 in ICEV-SI

fossil gasoline
fossil diesel
rapeseed biodiesel
sugarcane ethanol

0.20

lignocellulose hydrogen
lignocellulose methanol
lignocellulose DME

0.15

lignocellulose ethanol
lignocellulose ethanol
lignocellulose methanol
lignocellulose diesel (FT)
lignocellulose hydrogen
lignocellulose DME

Fuel

in ICEV-SI
in ICEV-CI

fossil gasoline
fossil diesel
rapeseed biodiesel
maize ethanol
sugarcane ethanol

Car O&M

Car
0.10

30

Costsof driving ( /km)

0.05

Diesel incl. duty ex VAT

Fuel delivery

0.00

Fuel production

Fuel delivered costs ( /GJ)

3279

in FCV-PEM
in FCV-POX-PEM
in FCV-POX-PEM
in FCV-SR-PEM
in FCV-SR-PEM

Fig. 7. Delivered biofuels costs (left), and marginal costs of driving (right) for different fuelcar combinations for now-future. Costs exclude excise
duty, fuel VAT and road tax. Both fuel and driving costs are on pure biofuels basis: all extra or avoided costs of blends, relative to pure gasoline or
diesel, are allocated to the biofuel. Mineral oil-derived gasoline costs 5-5.3 h/GJHHV and diesel 4.4-4.9 h/GJHHV. For all vehicles (100 kW wheel
power passenger cars), annual distance is 20,000 km, lifetime is 10 yr, annual O&M is 400 h. RME5 means 5% biodiesel in diesel, E10 means 10%
ethanol in gasoline.

Table 6
Car costs and fuel efciencya (km/GJHHV) on short-term-long-term, for passenger cars (about 100 kW wheel power), for various fuels in various
engine systems
Engine system

Car costsa (h)

Fuel

Short-term
ICEV-SI

Gasoline
Methanol
Ethanol

Efciencyb (km/GJfuel)
-Long-term

Short-term

-Long-term

430
540
440
520
510

-950
-950
-950
-950

H2

16,500
17,800
16,500
18,500
21,450

ICEV-CI

Diesel (fossil, synthetic, RME)

M100

18,000
18,000

-17,750

470
470

-760

FCV-PEM

H2

25,400

-18,750

880

-1690

FCV-SR-PEM

M100

26,050

-22,100

760

-840

FCV-POX-PEM

Gasoline/diesel/E100

26,500

-22,500

630

-960

M100
E10
E85 (FFV)

Annual O&M is assumed 400 h/car.

a
Costs for the gasoline ICEV SI and diesel ICEV CI are set; the others are derived from absolute costs (extra costs for drive train) and relative costs
in literature (van Walwijk et al., 1996; De Jager et al., 1998; Arthur, 1999; Ogden et al., 1999).
b
Besides the fuel and engine, the fuel efciency depends on other car features (weight, form, energy regeneration options) and the drive cycle
(Ogden et al., 1999; Maclean and Lave, 2003). Efciency is found expressed in many different ways: mile per gallon (mpg), 100 km per certain litres of
gasoline equivalent, %, MJ/km. To calculate from km per litre gasoline equivalent to km per GJ, multiply by 28.4 l/GJ (1 l for 15 km equals 430 km/
GJ). Efciencies estimated from average drive cycles in literature (Williams et al., 1995; Lynd, 1996; van Walwijk et al., 1996; De Jager et al., 1998;
Arthur, 1999; Weiss et al., 2003). Efciencies for blends are allocated to the biofuel only.

ARTICLE IN PRESS
3280

C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

performance of (bio)fuels in different engines and car

costs. Parameters have been derived from relative
numbers in the literature that compared the drive trains
running on biofuels with comparable drive trains on
fossil fuels. The fuel cell vehicle has been included as an
(expensive) option for the short-term, and is assumed to
be available at projected costs in the future.
The engine efciency (measured in km/GJ fuel) of
ICEVs can be improved for both spark and compression
ignition engines. Alcohols may have a higher efciency
in spark-ignition ICEVs than gasoline, if the vehicles are
adapted to prot from the slightly higher octane7
number. Efciency differences found for different fuels
in compression-ignition (diesel) ICEVs are so small that
they will not be reckoned with (van Walwijk et al.,
1996). Vehicle efciency improvements in transmission
and aerodynamics are generally independent of the fuel
type (Lynd, 1996) and are not anticipated here. The
efciencies in Table 6 hold for a drive chain of 100 kW
wheel power. Because of differences in energy content
between the fuels, equal driving ranges require different
fuel and tank loads (volume or mass), which indirectly
inuences the performance (Elam, 1996; Ogden et al.,
1999). It is not known to what extent this is dealt with in
the literature.
All biofuels require adaptations in ICEV cars, leading
to extra costs. Only relative costs from comparisons of
different fuelcar combinations in the literature are
used.
Many advantages are attributed to future fuel cell
vehicles8 over ICEVs. Potentially, they have a high
energy efciency (not limited to the Carnot efciency of
thermal energy processes), hardly produce emissions,
operate silently, require little maintenance (no moving
parts), and would be cheap in owning and operation
(Katofsky, 1993; Williams et al., 1995; van Walwijk
et al., 1996; Ogden et al., 1999; E-lab, 2000). However,
fuel cells are not commercial yet and expected expensive
in the short-term, large and heavy per kiloWatt output.
Also, exible car operation (changing load) requires the
use of very efcient batteries. Short-term efciency may
not be so high compared to hybrid vehicles. Whether
7
The octane number expresses a fuels quality rating, indicating the
ability of the fuel to resist premature detonation and to burn evenly
when exposed to heat and pressure in an (spark ignited) internal
combustion engine. Normal gasoline has an octane number of 8789.
The cetane number is a primary measure of fuel suitability for diesel
engines (compression ignited), but not related to fuel efciency. It
essentially expresses the delay before ignition. The shorter the delay the
betterand the higher the cetane number.
8
Different types of fuel cells are considered for vehicles: Alkaline
Fuel Cell (AFC) and Proton Exchange Membrane (aka Polymer
Electrolyte Membrane PEM) fuel cell that process hydrogen, and the
Direct Methanol Fuel Cell (DMFC) that processes methanol. All have
low operation temperatures (20120 1C). Most literature and demonstration programs focus on PEM FCVs. The AFC is extremely
expensive and requires very pure H2.

and when FCVs will really perform better than future

ICEVs is not clear (Weiss et al., 2003).
To use a fuel other than hydrogen in PEM Fuel cells,
it must be converted into H2 (and CO2) onboard the
vehicle. The most simple fuels methanol and DME can
be steam reformed (SR) at relatively low temperatures
(van Walwijk et al., 1996), while others require (harsher)
partial oxidation (POX) or autothermal reforming.
Onboard reforming implies an energy (direct and via
vehicle weight) and cost penalty. The total fuel efciency
of a fuel cell system must then take into account the
efciency of the reforming process as well as the
efciency of the fuel cell. The heat generated as a byproduct of the fuel cell can be used to increase the
systems energy efciency. Onboard reforming is only an
option if the reformer is exible in providing hydrogen
to the fuel cell, as fast or slow as it is being consumed by
the fuel cell. If additional hydrogen storage would be
necessary, the onboard reformer loses its advantage
(Brydges, 2000). Direct methanol fuel cells (DMFC)
circumvent reforming, thereby offering lower system
complexity. Although the methanol to electricity efciency is theoretically high, results reported so far are
not optimistic, and the catalyst is very expensive (Dillon
et al., 2004).
For a 100 kW wheel power passenger car with annual
20,000 km driving cycle, the resulting costs per kilometre
driven are shown in Fig. 7. Biomass alternatives to diesel
are all slightly more expensive than fossil diesel in the
short-term, but synthetic diesel and DME become
competitive in the longer term. Most fuel cell options
are signicantly more expensive than gasoline/diesel
ICEVs; only hydrogen in a future FCV PEM may
become competitive.
In this gure, excise duty and VAT are excluded, and
the resulting fuel contribution may become strikingly
small: only 9% of the total driving costs for hydrogen in
FCVs. This agrees well with values reported by others
(Arthur, 1999; Ogden et al., 1999). Considering these
ndings with the fact that car costs are dictated by other
factors than engine and fuel system costs, and with the
observed uncertainty in future fossil fuel-derived gasoline and diesel costs, there are no a priori economic
barriers to develop and introduce cars that run on
advanced biofuels.
For indication, estimated distances that could be
driven on fuel produced on 1 ha of feedstock are given in
Table 7. The larger distances driven in the long-term are
caused by the higher yields per hectare combined with
the higher engine efciencies projected. Compared to
these, the increase in conversion efciency plays a
marginal role. Although the ranges found are broad, it
can be concluded that much larger distances can be
driven on fuels from lignocellulosic biomass than on
fuels from sugar beet or rapeseed, and that these
differences become more pronounced in the long-term.

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283
Table 7
Distance that can be driven per hectare of feedstock for several
combinations of fuels and engines, derived from the net energy yield
and vehicle efciency as reported in Hamelinck and Faaij (2002), and
Hamelinck (2004)
Feedstock

Fuel

Engine

Distance (thousands km/ha)

Short-term

Lignocellulose

Hydrogen
Methanol
FT
Ethanol

Sugar beet

Ethanol

Rapeseed

RME

ICEV
FCV
ICEV
FCV
ICEV
FCV
ICEV
FCV
ICEV
FCV
ICEV
FCV

26 37
44a140b
34b49a
68a83b
22b38a
50a67b
29b30a
38a72b
15a37b
19a93b
5a28b
6a84b

Long-term
80b97a
189b321a
75b287a
113b252a
56b167a
97b211a
82b238a
129b240a
57a88b
58a138b
15a79b
19a137b

Derived from area net fuel yield and vehicle efciency in Hamelinck
(Hamelinck, 2004, Tables 7 and 8).
b
Derived from net energy yield, energy conversion efciency and
vehicle energy use in Faaij and Hamelinck (2002, Tables 2, 3 and 5).
From short- to long-term, the feedstock yield increases from 11 to
30 tonnedry/ha for lignocellulose crops, from 13 to 19 tonnedry/ha for
sugar beet and from 3 to 4 tonnedry/ha for rapeseed.

It seems that modern ICEVs (e.g. hybrids) combined

with advanced biofuels have a very strong position as
carbon neutral alternative for the transport sector
compared to hydrogen fuelled FCVs for a long period
to come. The distance that can be driven on hydrogen is
more than double, when FCVs are applied instead of
ICEVs.

4. Conclusion
4.1. Main conclusions
Biomass could play a large and important role in a
future sustainable energy supply as a source for modern
energy carriers as electricity and transportation fuels.
Especially the introduction of biofuels is attractive
because it is one of very few options for low CO2
emission transport systems against (eventually) reasonable costs, and because it decreases or spreads fuel
dependency. Of the many conceivable biofuels, fuels
from lignocellulose biomass are the most attractive,
because they allow for a higher fuel yield per hectare,
have better projected economics, their feedstock requires
less additional energy for growth and harvest and can be
grown under many different circumstances in contrast to
annual crops that require good-quality land.
In the present study, the production of four promising
biofuelsmethanol, ethanol, hydrogen, and synthetic

3281

dieselwas systematically analysed, based on detailed

pre-engineering analyses. Production costs of these fuels
range 1622 h/GJHHV now, down to 913 h/GJHHV in
future (2030). This performance assumes both certain
technological developments as well as the availability of
biomass at 3 h/GJHHV. The feedstock costs strongly
inuence the resulting biofuel costs by 23 h/GJfuel for
each h/GJHHV feedstock difference. In biomass producing regions such as Latin America or the former USSR,
the four fuels could be produced against 711 h/GJHHV.
The uncertainties in the biofuels production costs of the
four selected biofuels are 1530%, which is small when
considering the large uncertainty in future (2030)
gasoline and diesel prices.
The production costs of other biofuels were calculated
using the same assumptions on feedstock costs, scale,
load, and capital depreciation. This showed that also
dimethoxymethane (DMM), DME and SNG fall in the
same cost range. These fuels were however not analysed
at a similar level of detail.
When applied in cars, biofuels have driving costs in
ICEVs of about 0.180.24 h/km now and may be about
0.18 in future (fuel excise duty and VAT excluded). This
is only slightly higher than the driving costs of current
gasoline/diesel. Moreover, the cars contribution to
these costs is much larger than the fuels contribution,
and the differences in the fuels contributions are only
about 1 cent/km (which is small compared to the
uncertainties). The driving costs for fuel cell vehicles are
higher because of the projected larger vehicle costs.
From this perspective one can conclude that cars that
run on advanced fuels can be brought on the market at
competitive prices, and that the biofuels driving costs
could compete with those of fossil-derived fuels.
Furthermore, much larger distances per hectare can be
driven on fuels from lignocellulosic biomass than on
fuels from sugar beet or rapeseed, and these differences
become more pronounced in the long-term. The
distances that can be driven on hydrogen more than
double, when FCVs are applied instead of ICEVs.
The key fuel chains for the short-term seem to be
methanol and FT diesel, while ethanol from lignocellulose emerges in the medium-term. Ultimately, hydrogen
may offer the best perspective; but requires breakthroughs in hydrogen storage technology to tick the
balance. Compared with these advanced biofuels,
biodiesel from rapeseed and ethanol from sugar beet
or starch crops, already available today, are expensive
and inefcient with very little room for improvement.
4.2. RD&D issues
The gasication-derived fuels require the development
of large (about 400 MWHHV input) pressurised gasiers,
a gas cleaning section that matches the catalysts
specication, increased catalyst selectivity (for FT diesel

ARTICLE IN PRESS
3282

C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283

production), and ceramic membranes in the case of

hydrogen. Hot gas cleaning and CO2 removal positively
affect the total plant efciency, but the economic effect
is ambivalent. Co-producing electricity and biofuels
deserve further research. Synergy with fossil fuels (cofeeding) could facilitate both scale enlargement and cost
reduction. More research is desired in this eld.
The production of ethanol from lignocellulose biomass requires the development of more efcient pretreatment technology, and of micro-organisms that yield
higher conversions, as well as the integration of several
conversions into fewer reactors.

Acknowledgements
The PhD research (Hamelinck, 2004) underlying this
article was made possible by nancial help from the
National Research Programme on Global Air Pollution
and Climate Change (NOP-MLK), the Technology
foundation (STW), the Cooperation for Sustainable
Energy (SDE), Shell Global Solutions, Netherlands
Agency for Energy and the Environment (Novem),
and Essent. The author is much indebted to professor
Wim Turkenburg (promoter) for detailed commenting
on draft papers, and to many other colleagues who
contributed to discussions on biomass and biofuels and
wishes to thank them.

References
Arthur, D.L., 1999. Analysis and integral evaluation of potential CO2neutral fuel chains, GAVE reports 99089910 (Management
Summary, Sheetpresentation and Appendices). Netherlands
Agency for Energy and the Environment Novem, Utrecht, the
Netherlands, 391pp.
Aspen Technology Inc., 2003. Aspen Plus release 11.1. Cambridge,
MA, USA.
Azar, C., Larson, E.D., 2000. Bioenergy and land-use competition in
Northeast Brazil. Energy for Sustainable Development 4 (3),
5158.
Berndes, G., Hoogwijk, M., Van den Broek, R., 2003. The contribution of biomass in the future global energy supply: a review of 17
studies. Biomass and Bioenergy 25, 128.
BP, 2003. BP Statistical Review of World Energy (www.bp.com/
centres/energy/). Bournemouth, UK, 44pp.
Brydges, J.E., 2000. A hydrogen fueling station in 2005will it
happen? How do we get from here to there? M.Sc. Thesis,
Massachusetts Institute of Technology, Cambridge, MA, USA,
70pp+appendices.
Campbell, C.J., Laherrere, J.H., 1998. The end of cheap oil. Scientic
American 278 (3), 7883.
Cavallo, A.J., 2002. Predicting the peak in world oil production.
Natural Resources Research 11 (3), 187195.
Damen, K., 2001. Future Prospects for Biofuel Production in Brazil, a
Chain Analysis Comparison of Ethanol from Sugarcane and
Methanol from Eucalyptus in Sao Paulo State. Utrecht University,
Copernicus Institute, Department of Science Technology and
Society, Utrecht, the Netherlands 67pp.

De Jager, D., Faaij, A.P.C., Troelstra, W.P., 1998. Cost effectiveness

of transport biofuels. Netherlands Agency for Energy and the
Environment Novem, Utrecht, the Netherlands, 82pp+appendices
(in Dutch).
Deffeyes, K.S., 2001. Hubberts Peak: the Impending Oil Shortage.
Princeton University Press, Princeton, NJ, USA.
Dillon, R., Srinivasan, S., Arico`, A.S., Antonucci, V., 2004. International activities in DMFC R&D: status of technologies and
potential applications. Journal of Power Sources 127 (12),
112126.
EIA, 2003. International Energy Outlook 2003 (www.eia.doe.gov/oiaf/
ieo/index.html). Energy Information Administration, Ofce of
Integrated Analysis and Forecasting, Washington, DC, USA,
272pp.
E-lab, 2000. Running Buses on Hydrogen Fuel Cells: Barriers and
Opportunities, JulySeptember issue. Massachusetts Institute of
Technology, Cambridge, MA, USA.
Elam, N., 1996. Automotive Fuels Survey, Part 2: RAW Materials and
Conversion. International Energy Agency IEA/AFIS, Breda, the
Netherlands.
Elam, N., 2002. The bio-DME project, Phase 1 Non-condential
version, Report to Swedish National energy Administration STEM
(www.atrax.se/PDF/Final_report_DME.pdf). Atrax Energi, Stockholm, Sweden, 61pp.
EU, 2003. Directive 2003/30/EC of the European Parliament and of
the Council, of 8 may 2003, on the promotion of the use of biofuels
and other renewable fuels for transport. Ofcial Journal of the
European Union 2003 (L123), 4246.
EurObserER, 2004. Biofuels barometer 2003. Systemes Solaires (161),
5365.
Faaij, A.P.C., 1997. Energy from Biomass and Waste. Utrecht
University, Depertment of Science Technology and Society,
Utrecht, the Netherlands (180pp).
Faaij, A., Meuleman, B., van Ree, R., 1998. Long Term Perspectives
of Biomass Integrated Gasication with Combined Cycle Technology. Netherlands Agency for Energy and the Environment Novem,
Utrecht, the Netherlands (93pp+appendices).
Faaij, A.P.C., Hamelinck, C.N., 2002. Long term perspectives for
production of fuels from biomass; integrated assessment and
RD&D priorities. Proceedings of the Twelfth European Biomass
conference: Biomass for Energy, Industry and Climate Protection.
ETA Florence, Florence, Italy.
Goldemberg, J., 2000. World Energy Assessment, Preface. United
Nations Development Programme, New York, NY, USA.
Goldemberg, J., Coelho, S.T., Lucon, O., 2004. How adequate policies
can push renewables. Energy Policy 32 (9), 11411146.
Hall, D.O., Rosillo-Calle, F., Williams, R.H., Woods, J., 1993.
Biomass for energy: supply prospects. In: Johansson, T.B., Kelly,
H., Amulya, K.N.R., Williams, R.H. (Eds.), Renewable Energy,
Sources for Fuels and Electricity. Island Press, Washington, DC,
USA, pp. 593653.
Hamelinck, C.N., 2004. Outlook for advanced biofuels. Ph.D. Thesis,
Utrecht University, Copernicus Institute, Utrecht, the Netherlands,
p. 232.
Hamelinck, C.N., Faaij, A.P.C., 2002. Future prospects for production
of methanol and hydrogen from biomass. Journal of Power
Sources 111 (1), 122.
Hamelinck, C.N., Faaij, A.P.C., den Uil, H., Boerrigter, H., 2004a.
Production of FT transportation fuels from biomass; technical
options, process analysis and optimisation, and development
potential. Energythe International Journal 29 (11), 17431771.
Hamelinck, C.N., Suurs, R.A.A., Faaij, A.P.C., 2004b. International
Bioenergy Transport Costs and Energy Balance. Utrecht
University, Copernicus Institute, Utrecht, the Netherlands.
Hamelinck, C.N., Van Hooijdonk, G., Faaij, A.P.C., 2005. Ethanol
from lignocellulosic biomass: techno-economic performance in

ARTICLE IN PRESS
C.N. Hamelinck, A.P.C. Faaij / Energy Policy 34 (2006) 32683283
short-, middle- and long-term. Biomass and Bioenergy 28 (4),
384410.
His, S., 2004. Panorama 2005A look at y Biofuels Worldwide.
Institut Franc- ais du Petrole IFP, Rueil-Malmaison France (6pp).
Hoogwijk, M., Faaij, A., van den Broek, R., Berndes, G., Gielen, D.,
Turkenburg, W., 2003a. Exploration of the ranges of the global
potential of biomass for energy. Biomass and Bioenergy 25 (2),
119133.
Hoogwijk, M., Faaij, A.P.C., de Vries, B., Turkenburg, W.C., 2003b.
Potential of biomass energy under four land-use scenarios, Part B:
exploration of regional and global cost-supply curves. Manuscript.
Hoogwijk, M., Faaij, A.P.C., Eickhout, B., de Vries, B., Turkenburg,
W.C., 2003c. Potential of biomass energy under four land-use
scenarios, Part A: the geographical and technical potential.
Manuscript.
IMAGE-team, 2001. The IMAGE 2.2 implementation of the SRES
scenarios, CD-ROM publication 481508018. National Institute of
Public Health and the Environment RIVM, Bilthoven, the
Netherlands.
Intergovernmental Panel on Climate Change, 2000. In: Nakicenovic,
N., Swart, R. (Eds.), Special Report on Emission Scenarios, A
Special Report of IPCC Working Group III. Cambridge University
Press, Cambridge, UK, 612pp.
Katofsky, R.E., 1993. The production of uid fuels from biomass.
Report no. 279. Princeton University, Center for Energy and
Environmental Studies, Princeton, NJ, USA, 330pp.
Larson, E.D., Williams, R.H., Regis Lima Verde Leal, M., 2001. A
review of biomass integrated-gasier/gas turbine combined cycle
technology and its application in sugarcane industries, with an
analysis for Cuba. Energy for Sustainable Development V (1),
5476.
Lynd, L.R., 1996. Overview and evaluation of fuel ethanol from
cellulosic biomass: technology, economics, the environment, and
policy. Annual Reviews, Energy Environment 21, 403465.
Maclean, H.L., Lave, L.B., 2003. Evaluating automobile fuel propulsion system technologies. Progress in Energy and Combustion
Science 29 (1), 169.
Marrison, C.I., Larson, E.D., 1995. Cost versus scale for advanced
plantation-based biomass energy systems in the USA and Brazil.
In: Proceedings of the Second Biomass Conference of the
Americas. National Renewable Energy Laboratory NREL, Golden, CO, USA.
Marrison, C.I., Larson, E.D., 1996. A preliminary analysis of the
biomass energy production potential in Africa in 2025 considering
projected land needs for food production. Biomass and Bioenergy
10 (5/6), 337351.
Meadows, D.H., Meadows, D.L., Randers, J., Behrens III, W.W.,
1972. The Limits to Growth, A Report to The Club of Rome.
Universe Books, New York, NY, USA.
Ogden, J.M., Steinbugler, M.M., Kreutz, T.G., 1999. A comparison of
hydrogen, methanol and gasoline as fuels for fuel cell vehicles:
implications for vehicle design and infrastructure development.
Journal of Power Sources 79 (2), 143168.
Perlack, R.D., Wright, L.L., 1995. Technical and economic
status of wood energy feedstock production. Energy 20 (4),
279284.
Peters, M.S., Timmerhaus, K.D., 1980. Plant Design and Economics
for Chemical Engineers, third ed. McGraw-Hill Book Company,
New York, NY, USA.

3283

Phillips, V.D., Liu, W., Merriam, R.A., Bain, R.L., 1995. Shortrotation forestry as alternative land use in Hawaii. Biomass and
Bioenergy 8 (4), 235244.
Ranney, J.W., Mann, L.K., 1994. Environmental considerations in
energy crop production. Biomass and Bioenergy 6 (3), 221228.
Ravindranath, N.H., Hall, D.O., 1995. Biomass, Energy, and
Environment, A Developing Country Perspective from India.
Oxford University Press, Oxford, UK.
Reinhardt, G.A., Zemanek, G., 2000. Okobilanz bioenergietrager,
Basisdaten, ergebnisse, bewertungen. Erich Schmidt Verlag, Berlin,
Germany.
Rogner, H.H., 2000. Energy resources. In: Goldemberg, J. (Ed.),
World Energy Assessment. United Nations Development Programme, New York, NY, USA, pp. 135171.
Tijmensen, M.J.A., Faaij, A.P.C., van Hardeveld, M.R.M., Hamelinck, C.N., 2002. Exploration of the possibilities for produciton of
Fischer Tropsch liquids via biomass gasication. Biomass and
Bioenergy 23 (2002), 129152.
Turkenburg, W.C., 2000. Renewable energy technologies. In: Goldemberg, J. (Ed.), World Energy Assessment. United Nations
Development Programme, New York, NY, USA, pp. 219272.
van den Broek, R., 2000. Sustainability of biomass electricity systems.
Ph.D. Thesis, Utrecht University, Department of Science Technology and Society, Utrecht, the Netherlands, 216pp.
van den Broek, R., van Walwijk, M., Niermeijer, P., Tijmensen, M.,
2003. Biofuels in the Dutch market: a fact-nding study. Report
Novem 2GAVE03.12. Ecofys bv/Netherlands agency for energy
and the environment Novem, Utrecht, the Netherlands,
256pp+appendices.
van Thuijl, E., 2002. Large scale application of biofuels in road
vehicles, a transition to low emission transport in the Netherlands.
MSc Thesis, Energy Research Center of the Netherlands (ECN),
Petten, the Netherlands, 89pp+appendices (in Dutch).
van Walwijk, M., Buckmann, M., Troelstra, W.P., Achten, P.A.J.,
1996. Automotive Fuels Survey, Part 2: Distribution and Use.
International Energy Agency IEA/AFIS, Breda, the Netherlands.
Van Zeijts, H., Oosterveld, E.B., Timmerman, E.A., 1994. Can the
agricultural sector supply clean energy? Assessment of sustainability of energy crops. Centre for Agriculture and Environment
CLM, Utrecht the Netherlands (78pp+appendices) (in Dutch).
Weiss, M.A., Heywood, J.B., Schafer, A., Natarajan, V.K., 2003.
Comparative assessment of fuel cell cars. Report MIT LFEE 2003001 RP (lfee.mit.edu/publications/PDF/LFEE_2003-001_RP.pdf).
MIT Laboratory For Energy and the Environment LFEE,
Cambridge, MA, USA, 29pp.
Williams, R.H., Larson, E.D., 1993. Advanced gasication-based
biomass power generation. In: Johansson, T.B., Kelly, H., Amulya,
K.N.R., Williams, R.H. (Eds.), Renewable Energy, Sources
for Fuels and Electricity. Island Press, Washington, DC, USA,
pp. 729785.
Williams, R.H., Larson, E.D., Katofsky, R.E., Chen, J., 1995.
Methanol and hydrogen from biomass for transportation, with
comparisons to methanol and hydrogen from natural gas and coal.
PU/CEES Report 292. Princeton University/Center for Energy and
Environmental Studies, Princeton, NJ, USA, 47pp.
Wooley, R., Ruth, M., Glassner, D., Sheehan, J., 1999. Process design
and costing of bioethanol technology: a tool for determining the
status and direction of research and development. Biotechnology
Progress 15 (5), 794803.