chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

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

A gasification-based biorefinery for the pulp and paper

industry

Stefano Consonni a,∗ , Ryan E. Katofsky b , Eric D. Larson c

a Department of Energy, Politecnico di Milano, Milan, Italy
b Arlington, MA, USA
c Princeton Environmental Institute, Princeton University, Princeton, NJ, USA

a b s t r a c t

This paper is drawn from a 2-year study of integrated pulpmill biorefineries based on black liquor (the lignin-rich
byproduct of fiber extraction from wood) and wood residue gasification at a large kraft mill representative of those
in the Southeast United States. The study included detailed mass-energy balance simulations, financial analyses,
and energy and environmental benefits estimates for seven pulpmill biorefinery process configurations. All seven
configurations include an oxygen-blown, high-temperature black liquor gasifier, syngas cooling, clean-up by a Recti-
sol (methanol) system, and a catalytic gas-to-liquid process; six of them also include a fluidized-bed, oxygen-blown
biomass gasifier and a gas turbine combined cycle fully integrated with the gasification and syngas cooling section.
Three biofuels were examined: dimethyl ether (DME), Fischer–Tropsch liquids, and ethanol-rich mixed-alcohols.
For the integrated biorefineries analyzed here, the ratio of useful energy outputs (steam, electricity and fuels) to total
energy inputs (black liquor, wood residuals and fuel oil) ranges from 66 to 74%; these values compare with about 57%
for conventional systems based on Tomlinson boilers and 65% for gasification combined cycles that produce only
electricity. Because of the integration of the biorefinery with the pulp and paper mill, the adjusted liquid fuel yield
per unit of biomass – a measure of the effectiveness of biomass conversion to liquids – is far higher than for “stand-
alone” gasification-based biorefineries or for ethanol production via biochemical conversion (based on enzymatic
hydrolysis).
Besides better energy performance, the integration between the biorefinery and the pulp mill effectively limits the
specific capital investment associated with liquid fuels production to a surprisingly modest $60,000–150,000 per barrel
of diesel equivalent per day—specific capital costs comparable to those for much larger coal-to-liquids facilities.
Gasification-based pulp mill biorefinery technologies, once fully commercialized, offer the potential for attractive
investment returns and, if implemented widely, significant energy and environmental benefits to the United States.
© 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Gasification; Pulpmill; Black liquor; Biorefinery; Combined; Cycle; Fuel synthesis; Dimethyl ether (DME);
Fischer–Tropsch liquids; Mixed-alcohols

1. Background and scope the United States (Larson et al., 2003). The same reference mill
– a hypothetical, but representative integrated pulp and paper
This paper summarizes results from a 2-year effort assess- mill in the Southeast United States – was used as the basis for
ing prospective technical viability, commercial viability, and analysis in both studies.
national environmental and energy impacts of gasification-
based biorefineries making liquid fuels from black liquor and 1.1. Context
woody biomass residues at kraft pulp and paper mills (Larson
et al., 2006a). The study built on an earlier study examin- The U.S. pulp and paper industry is the largest producer and
ing the potential for black liquor gasification combined cycle user of biomass energy in the United States today, nearly all
(BLGCC) electricity generation at kraft pulp and paper mills in derived from sustainably grown trees. Renewable resources

∗
Corresponding author. Fax: +39 0523356837.
E-mail address: [email protected] (S. Consonni).
Received 14 November 2008; Received in revised form 23 June 2009; Accepted 9 July 2009
0263-8762/$ – see front matter © 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.cherd.2009.07.017
1294 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

ments to replace the aging fleet of Tomlinson recovery boilers

Nomenclature used today to recover energy and pulping chemicals from black
liquor. A similar situation exists in the European pulp industry
ASU air separation unit
(Ekbom et al., 2005). The large cost of new Tomlinson recov-
BL black liquor
ery boilers (for a typical mill, $100–$200 million2 ) provides an
BLG black liquor gasification
unusual window of economic opportunity for their replace-
BLGCC black liquor gasification combined cycle
ment with black liquor gasifiers. Concerted efforts are ongoing
BLS black liquor solids
in the United States and Sweden to develop commercial black
DME dimethyl ether
liquor gasification (BLG) technologies.
DMEa, DMEb, DMEc plant schemes generating DME
FSI fuel synthesis island
FT Fischer–Tropsch 1.2. Technology and fuel products
FTa, FTb, FTc plants schemes generating FT liquid
FTL Fischer–Tropsch liquid The biorefinery concept pursued in this study comprises two
GT gas turbine different types of gasifiers: one that handles the whole flow of
GTCC gas turbine combined cycle black liquor generated by the reference mill, the other for the
HP high pressure additional woody biomass needed to reach a given target for
HRSG heat recovery steam generator the size of the power island (in our case to fully fire a 70-MW
IRR internal rate of return class, heavy-duty gas turbine) and to meet the thermal energy
LHV lower heating value needs of the mill.
LPDME liquid phase DME (reactor) Gasification technology enables low-quality solid or liquid
MA or MixOH mixed alcohol fuels like biomass and black liquor to be converted with low
mt metric ton pollution into a fuel gas (synthesis gas or “syngas”) consisting
NPV net present value largely of H2 and CO. After adequate cleaning, this syngas can
REC renewable energy credit be burned cleanly and efficiently in a gas turbine to generate
REP reference energy prices electricity, or passed over appropriate catalysts to synthesize
st short ton “green” liquid transportation fuels or chemicals.
TSEP tight supply energy prices The biorefinery analysis assumed that gasification tech-
TIT turbine inlet temperature nologies have successfully moved beyond the development
TOT turbine outlet temperature stage now in progress and that the risks involved with
WGS water gas shift (reaction) installing gasification systems are comparable to those
of installing a new Tomlinson-based power/recovery sys-
tem. If the BLG experiences currently being pursued meet
used at pulp mills include bark, wood wastes, and black liquor, their promise (Lindblom and Landälv, 2007) and pressurized
the lignin-rich by-product of cellulose-fiber extraction. The biomass gasification technology (Evans et al., 1987; Lau et al.,
total of these biomass energy sources consumed at pulp mills 2003) overcomes the hurdles encountered in past demonstra-
in 2004 in the United States was an estimated 1.3 Exajoules.1 tions, this condition of commercial reliability could be reached
Additionally, there are substantial wood residues and non- within a decade. All of the equipment for downstream pro-
merchantable timber that remain behind after harvesting of cessing of the synthesis gas from the gasifiers, including for
trees for pulpwood. A major study of U.S. biomass resources sulphur capture and for catalytic synthesis of at least two
(Perlack et al., 2005) estimates there are some 2 EJ of unused of the liquid fuels of interest, is already commercially avail-
wood resources (logging residues, fire-prevention thinnings, able.
mill residues, and urban wood waste) that are recoverable on Using the potential U.S. market size and potential for
a sustainable basis at present, increasing to nearly 3 EJ in the enhancing energy security as key screening criteria, our biore-
future. finery analysis identified three liquid products for detailed
While most pulp and paper manufacturing facilities in analysis: Fischer–Tropsch liquids (FTL, a mixture of primar-
the United States today do not export electricity and none ily straight-chain hydrocarbons – olefins and paraffins – that
export transportation fuels, their established infrastructure resembles a semi-refined crude oil), dimethyl ether (DME, a
for collecting and processing biomass resources provides a propane substitute or diesel-engine fuel), and mixed-alcohols
strong foundation for future gasification-based biorefineries (MA, a mixture of ethanol and higher alcohols that is a
that could produce a variety of renewable fuels, electricity, and potential gasoline substitute). Unlike FTL or DME synthe-
chemicals in conjunction with pulp and paper products. If the sis technologies, MA technologies are not yet commercial,
biomass resources from which energy carriers are produced at although there are a number of startup companies working
such biorefineries are sustainably grown and harvested, there to try to commercialize these.
would be few net lifecycle emissions of CO2 associated with
biorefineries and their products.
Coupled with the potential to address national energy
security and global warming concerns is the looming need 2
All costs mentioned in this paper are given in year-2005 dollars.
in the U.S. pulp and paper industry for major capital invest- Capital costs in this paper do not reflect the unprecedented run-
up in costs of materials and construction labor observed during
the past several years. Thus, the absolute capital costs reported
1
One Exajoule (EJ) = 1018 Joules. Approximately 1 EJ of black in this work likely underestimate costs for systems that would be
liquor and 0.3 EJ of woody residues (hog fuel) were generated and built today. However, since the recent cost escalations affect all
consumed in the U.S. paper industry in 2004 (based on estimates technologies, relative comparisons among different systems are
from the American Forest and Paper Association). still valid.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1295

Table 1 – Reference mill characteristics (mt = metric ton; st = short ton).

Power/recovery system→ Tomlinson Gasification
Pulping chemistry→ Conventional Polysulfide

Product flow (paper) Machine-dry mt/day 1,725

Unbleached pulp rate Bone dry st/day 1,580
Hardwood (HW)/Softwood (SW) mix % HW, % SW 65 HW, 35 SW

% for softwood 45.50 48.75

Digester yield
% for hardwood 46.50 49.75

Wood to digester 3,434 3,208

Hog fuel Bone dry st/day 340 317
Total pulpwood logs 3,774 3,525

Black liquor (BL) concentration % solids 80 80

BL solids (BLS) flow rate lb BLS/day 6,000,000 5,419,646

kg BLS/day 2,721,555 2,458,311

MJ/kg of BLS, LHV (HHV) 12.46 (13.89) 12.31 (13.87)

BL energy content Btu/lb of BLS, LHV (HHV) 5,359 (5,974) 5,295 (5,966)
MW, LHV (HHV) 392.6 (437.6) 350.7 (394.7)

C 33.46 32.97
H 3.75 3.70
O 37.35 36.88
BLS composition (mass %) S 4.10 4.27
Na 19.27 20.03
K 1.86 1.93
Ash/chlorides 0.21 0.22

MJ/kg of hog fuel, LHV (HHV) 8.14 (10.0) 8.14 (10.0)

Hog fuel energy content (50% moisture content) Btu/lb of hog fuel, LHV (HHV) 3,501 (4,300) 3,501 (4,300)
MWth, LHV (HHV) 57.8 (71.3) 54.1 (66.6)

Mill O2 use (for delignification) kg/mt of pulp 23 23

Mill steam use, 55 psig steam (including kg/kg of paper 3.384 3.207
evaporators, but excl. power/recovery MWth 142.8 135.3
area) MJ/mt of paper 7,149 6,774

Mill steam use, 175 psig steam (including kg/kg of paper 1.760 1.648
evaporators, but excl. power/recovery MWth 69.3 64.8
area) MJ/mt of paper 3,469 3,247

Total mill steam use MWth 212.1 200.1

Mill electricity use (excluding kWh/mt of paper 1,407 1,407

power/recovery) MWe 100.1 100.1

1.3. Technical and economic evaluations 2. Study assumptions

Detailed process designs and simulations were conducted for 2.1. Characteristics of reference pulp/paper mill
alternative configurations for the manufacture of each fuel
product assuming projected commercial (“Nth plant”) perfor- The reference mill is a hypothetical integrated kraft pulp and
mance. Environmental impacts were estimated by quantifying paper mill producing uncoated freesheet paper from a 65/35%
emissions of carbon dioxide, sulphur dioxide, nitrogen oxides, mix of hardwood and softwood. Consistent with the forward-
carbon monoxide, volatile organic compounds, particulate looking nature of this study and the continual improvements
matter, and total reduced sulphur (TRS) along the entire sup- in process energy efficiency historically achieved by U.S. pulp
ply chain, from wood extraction to vehicle use. A detailed and paper makers, the process steam demands at the mill are
internal rate of return analysis was carried out for each pro- taken to be about 10% below current (circa 2005) U.S. “best-
cess design, and the mill-level energy and environmental practice” levels. The nominal scale of the reference mill is
performance results were used as a basis for estimating poten- 6 million lbs/day (2721 metric tonnes/day) of dry black liquor
tial national energy/environment impacts under alternative solids (BLS); for such a level of black liquor production, a mod-
assumptions about the rate at which existing Tomlinson sys- ern integrated mill would produce 1725 machine-dry metric
tems would be retired and replaced with biorefineries. The tonnes per day of paper.
value of biorefinery integration within the pulp and paper Key reference mill parameter assumptions are shown in
mill was also evaluated by comparing the results of this Table 1 for both conventional kraft pulping and polysulfide
study to other studies that looked at stand-alone biorefiner- pulping. Economic liquor recovery for the latter pulping chem-
ies. istry is enabled by the use of BLG in the recovery area, and we
1296 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

assume this chemistry is implemented in our biorefineries. 2.3. Biomass gasification

Compared to conventional kraft pulping, polysulfide pulping
raises the digester yield (Jameel and Renard, 2003), enabling a Our biorefinery designs explore the potential impacts of sup-
reduction in wood feed to the mill compared to conventional plementing to a substantial degree the relatively modest
pulping for the same level of paper production.3 The higher quantities of woody residues that are used for energy at
digester yield also reduces the amount of black liquor solids typical pulp mills today. To maximize the effectiveness of
sent to the recovery area. woody residue use, most of our biorefinery designs incor-
porate gasification (rather than combustion) of the biomass.
We have chosen to use a pressurized, oxygen/steam-blown
2.2. Black liquor gasification fluidized-bed design based on the gasifier developed at the Gas
Technology Institute (GTI).4 For synthesis of fuels, a pressur-
A number of concepts for black liquor gasification (BLG) have ized syngas undiluted by nitrogen is preferred for feeding to
been proposed in the past (Consonni et al., 1998). Our study uti- the synthesis reactor, since synthesis conversion rates gen-
lizes the pressurized, oxygen-blown, high-temperature design erally increase with pressure and decrease with increasing
being developed by Chemrec (Whitty and Nilsson, 2001; fraction of inert components in the syngas. Also, reactor size
Marklund et al., 2006). The Chemrec design is character- and cost increase with inert content.
ized by the majority of the inorganic material in the BLS Development and pilot-plant demonstration efforts with
leaving the reactor as a smelt due to the high reactor tem- pressurized oxygen/steam-blown fluidized-bed gasification
perature (950–1000 ◦ C). In the late 1990s, the Weyerhaeuser date to the early-1980s in Sweden (Engstrom et al., 1981; Strom
company installed the world’s first commercial Chemrec et al., 1984) and the mid-1980s in the USA (Kosowski et al.,
gasifier, an atmospheric-pressure, air-blown unit designed 1984; Evans et al., 1987). Most such efforts were curtailed
to process 300 metric t/d BLS at a mill in North Carolina when world oil prices fell in the late 1980s. There has been
to augment the chemical recovery capacity of the existing some recent re-assessment of pressurized oxygen-blown gasi-
Tomlinson boiler. A pilot-plant pressurized Chemrec gasi- fication in the USA (Lau et al., 2003), and a major Finnish
fier (30 bar pressure and capacity to process up to 20 t/d BLS technology development effort for production of FT fuels
with oxygen firing) has been operating under a test program from woody biomass is focusing on pressurized oxygen-blown
at a pulp mill in Piteå, Sweden, since mid-2006. This unit fluidized-bed gasification, including plans for a demonstration
will provide data for scale-up (Marklund et al., 2006) to full- of the technology before 2010 at a scale of 100–200 dry tonnes
scale Tomlinson boiler replacement applications. As of early per day of biomass feed (Mckeough and Kurkela, 2006).
2008, feasibility studies were being done for commercial-scale
installations at the Smurfit Kappa Kraftliner mill in Piteå, 3. Biorefinery designs
Sweden and the New Page pulp and paper mill in Escanaba,
Michigan. We developed seven biorefinery designs: three variants for
With gasification (unlike with a Tomlinson boiler), there DME production, three variants for FTL production, and one
is a natural partitioning of sulphur (mainly as hydrogen sul- design for MA production.
fide, H2 S) to the gas phase and sodium to the condensed The process designs all include five basic equipment
phase. This split represents an important potential bene- “islands”: (i) black liquor gasification, (ii) biomass gasification
fit to a pulp mill, since it can facilitate the implementation (or, in one case, a hog fuel boiler island), (iii) syngas heat recov-
of alternative pulping chemistries (e.g., polysulfide) that can ery and syngas clean-up, (iv) fuel synthesis, and (v) power.
lead to increased pulp yield per unit of wood consumed
(Lindstrom et al., 2002). With the high-temperature gasi-
3.1. Equipment islands and configuration of the
fier design adopted here, slightly more than half of the
integrated plant
sulphur goes to the gas phase. To take advantage of the
natural separation of sulphur (S) and sodium (Na), it is
The gasification island includes a cryogenic air separation
necessary to recover H2 S from the gas in a form suitable
unit (ASU) for the production of oxygen. The black liquor
for preparing modified pulping liquors. Another reason that
and biomass gasification islands are essentially the same for
H2 S must be removed from the syngas is to avoid poi-
all the cases considered (except for the size of the biomass
soning of downstream fuel synthesis catalysts. Capture of
gasification island, which changes very significantly from
acid gases like H2 S is routinely practiced in other industries
one case to another). The syngas clean-up island is cen-
(e.g., petroleum refining) using patented physical or chemical
tered around Rectisol® technology, except for the MA case, in
absorption processes such as Selexol® or Rectisol® . A nega-
which Selexol® technology is adopted because of less strin-
tive consequence of the S/Na split in a gasifier is a higher
gent downstream sulphur concentration requirements. Fuel
causticizing load, i.e., larger required lime kiln capacity and
synthesis is carried out in a catalytic slurry-phase reactor for
lime kiln fuel consumption per unit of BLS processed com-
FTL and DME and in a fixed bed, also catalytic, for MA. The
pared to processing in a Tomlinson boiler (Larson et al.,
source of the syngas feed of the fuel synthesis island (FSI)
2003).
gives the basic characterization to each plant scheme; after
screening a number of alternatives, we selected the following
plant schemes:
• DME from black liquor syngas, with no recycle of syngas
3
The implicit assumption here is that the mill is already unconverted after one pass through the FSI.
operating at capacity (outside of the power/recovery area) when
using conventional pulping, so that an increase in digester yield
4
can be accommodated only by decreasing wood input, not by The license for the Gas Technology Institute (GTI) technology
increasing pulp production. is currently owned by Carbona/Andritz.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1297

Table 2 – Overview of plant configurations.

• DME from black liquor syngas, with recycle of unconverted DMEc Like DMEb, but without syngas recycle in the
syngas. FSI. Similarly to DMEb, in this case some
• FTL from black liquor syngas, no syngas recycle. syngas must be burned in a duct burner
• FTL from a mix of black liquor and biomass syngas, no syn- ahead of the HRSG of the combined cycle and
gas recycle. the steam turbine is of a back-pressure
• MA from a mix of black liquor and biomass syngas, with design.
syngas recycle. FTa Production of FTL, once-through FSI, biomass
gasifier that generates enough syngas to fully
fire the medium-scale gas turbine GE 6FA.
The syngas not used in (or discharged from) the FSI feeds The steam demand of the mill is met by
the power island, which also collects the heat recovered from burning some syngas in a duct burner ahead
other islands, with a configuration tailored to the character- of the HRSG of the combined cycle and the
istics of each plant scheme to maximize the benefits of heat steam turbine is of a back-pressure design.
integration. FTb Like FTa, but with the large scale, 170 MW
The cases of DME with syngas recycle and FTL from class gas turbine GE 7FA. The much larger
black liquor syngas have been further differentiated based size of the gas turbine eliminates the need for
on the type of power island: power boilers (no biomass gasi- the duct burner and allows the production of
fier) with back-pressure steam turbine vs combined cycle for much more steam than needed by the mill;
DME; medium-size, 70 MW class gas turbine (GE 6FA) vs large, consequently, the steam turbine includes a
170 MW class gas turbine (GE 7FA) for FTL. Except for this last low pressure (condensing) section that
FTL case, all combined cycles are based on the same 70 MW generates additional power by expanding the
class gas turbine (GE 6FA). steam not needed by the mill.
Table 2 summarizes the seven designs, where the nomen-
FTc Like FTa, but with the FSI fed by the mix of all
clature is as follows:
the syngas generated by black liquor and all
DMEa Production of DME, FSI with syngas recycle, no the syngas generated by biomass. The much
gas turbine (power island comprises just a higher fuel production increases steam
steam cycle). The extra steam required to meet production in the exothermic FT reactor and
the mill demand is generated by power boilers eliminates the need for a duct burner to meet
and the FSI is fed (necessarily) with the syngas the mill steam demand. Similar to FTb, steam
generated by BLG. Given the need for the power production is larger than needed by the mill
boilers, no excess steam is available and the and the steam turbine includes a low
steam turbine is of a back-pressure design. pressure condensing section.

DMEb Like DMEa, but with a biomass gasifier that MA Production of mixed-alcohols with the same
generates enough syngas to fully fire the rationale of scheme FTc, i.e. the FSI is fed by
heavy-duty, medium-scale gas turbine GE 6FA. the mix of all the syngas generated by black
In this case the steam demand of the mill is liquor and all the syngas generated by
met by burning some syngas in a duct burner biomass. In this case however, the low
ahead of the heat recovery steam generator conversion achievable in the mixed alcohol
(HRSG) of the combined cycle. Given the need reactor makes syngas recycle imperative.
for the duct burner, no excess steam is Despite syngas recycle, fuel conversion and
available and the steam turbine is of a thus steam production in the FSI are
back-pressure design. relatively low and the steam demand
1298 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

of the mill can be met only by burning some In case DMEa, syngas that is unconverted in passing
syngas in a duct burner placed, as usual, through the synthesis reactor is separated from product DME
ahead of the HRSG of the combined cycle. and 97% is recycled to synthesis to increase DME produc-
Since no excess steam is available, the steam tion. The 3% purge stream prevents excessive buildup of inert
turbine is of a back-pressure design. gases in the recycle loop. The purge gas is burned with wood
In cases where only biomass syngas feeds the gas tur- residues in the power island boiler. Heat recovered from the
bine (DMEb, DMEc, FTa, FTb) we have considered a biomass black liquor and syngas processing areas is integrated into the
gasifier with a syngas cooler. The syngas exiting the syngas boiler to increase steam production and minimize the amount
cooler is filtered through ceramic candles and then fed to the of woody residues needed. The steam is expanded through
gas turbine. Although not yet demonstrated at commercial- a back-pressure turbine to generate electricity. After meeting
scale, pilot-scale testing has been successful (e.g., at Varnamo, the parasitic electricity demands of the biorefinery, the net
Sweden), and it is believed that this arrangement can be com- power output is nearly zero and essentially all the electricity
patible with the requirements of a gas turbine in commercial required by the mill is purchased from the grid.
service. One approach to increasing net electricity production is to
When the syngas generated from biomass is mixed with increase the use of wood residues and/or increase the effi-
the syngas generated from black liquor and the mixture feeds ciency of electricity generation. Our DMEb design adopts both
the FSI (cases FTc and MA) we have considered a biomass gasi- of these approaches. In this design (Fig. 1), the black liquor and
fier with quench. In this case the catalyst of the FSI requires synthesis gas processing areas are unchanged from DMEa, and
deep removal of all contaminants from the syngas and gas the DME production rate is identical to that in DMEa. However,
clean-up must necessarily take place at low temperature. The woody biomass is gasified and used to fuel a gas turbine–steam
penalties brought about by quenching the syngas are therefore turbine combined cycle (GTCC). Because the GTCC is charac-
limited, because the syngas has to be cooled to low tempera- terized by a higher electricity to steam production ratio than
ture anyhow. On the other hand, a quench system would be the boiler/steam turbine system in DMEa, more biomass must
lower cost than the syngas cooler. be used in the DMEb design than in the DMEa design to deliver
the same amount of process steam, but electricity production
with DMEb is considerably greater than with DMEa.
3.2. Technology status and key design parameters Another approach to increasing electricity production
involving woody biomass consumption not much greater than
Table 3 summarizes the technologies incorporated into each in DMEa, is to use a lower level of syngas recycle to the syn-
of our seven biorefinery designs. The column labeled “sta- thesis reactor, such that more unconverted gas is available for
tus” indicates the proximity to commercial offering for each power generation. In DMEc (Fig. 1), the syngas recycle loop
technology. The majority of components are already in wide is eliminated, and the syngas passes only once through the
commercial use (in non-biorefinery applications). All but one reactor, leading to much lower DME production than in DMEb
of the technologies not in the commercial category fall into but requiring relatively little purchased biomass. The design
the category of technologies that have been or are currently is otherwise very similar to DMEb.
being demonstrated at pilot-scale, such that the commercial
deployment of biorefineries may be feasible within a decade. 3.4. Production of Fischer–Tropsch liquids
Only one technology (mixed-alcohol synthesis) is relatively
under-developed commercially today. For our FT biorefinery designs, we considered only single-pass
Table 4 summarizes key characteristics for each of our synthesis due to the much higher single-pass conversion of
biorefinery designs, with comparisons to the Tomlinson and syngas that can be achieved with FT synthesis compared to
BLGCC designs we developed earlier (Larson et al., 2003). The DME synthesis. Our FTa configuration is conceptually identical
black liquor flow is the same in all cases, except for the Tom- to the DMEc design described in Fig. 1. The FTb configura-
linson case where more black liquor is available due to lower tion substitutes the 6FA gas turbine with a larger 7FA unit,
pulp yield. The amount of woody biomass input as energy but is otherwise identical to FTa. The larger fuel requirement
(including hog fuel and purchased residues) is set by the con- of the 7FA necessitates a larger input of woody biomass. The
straint that the biorefinery provide all of the pulp/paper mill’s FTc design involves passing all syngas (from both black liquor
process steam demand. Since process steam production per and biomass gasification) through the FT synthesis reactor
unit of biomass input depends on the plant configurations, the and using only unconverted syngas to fuel the (6FA) gas tur-
amount of biomass needed to satisfy the mill’s steam demand bine (Fig. 2). This design involves the largest amount of woody
is not known a priori, but is an outcome of the heat and mass biomass use among the seven designs investigated.
balance calculations. Table 4 shows the net increase in total
wood (pulpwood and purchased residues) that must be han- 3.5. Production of mixed-alcohols
dled in the wood yard at the pulp/paper mill compared to the
Tomlinson case. There is some reduction in pulpwood input For the production of mixed-alcohols (Fig. 3) we adopt a design
to the mill (compared with the Tomlinson case, due to higher similar to FTc in that syngas from black liquor and woody
yield pulping process), but this reduction is more than offset residue gasification are combined for processing through the
by the need for additional purchased wood. synthesis reactor. In the MA design, however, because the
single-pass syngas conversion is much lower than for syn-
3.3. Production of DME thesis of FTL we recycle 76% of the unconverted syngas. This
somewhat arbitrary recycle rate gives a reasonable compro-
In all our DME designs, the syngas for DME synthesis is gen- mise between syngas-to-alcohols conversion and CO2 content
erated by the black liquor gasifier and cleaned in a Rectisol in the gas fed to the reactor (which is a mixture of fresh syn-
unit. gas and recycle gas); for higher recycle rates our kinetic model
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1299

Table 3 – Summary of technologies included in our biorefinery designs indicating commercial status of each technology.
Statusa FTa FTb FTc DMEa DMEb DMEc MA

Black liquor gasification

High-temp gasifier Pilot       
Quench Pilot       
O2 feed Com       

Woody biomass conversion

Fluid-bed gasifier Pilot      
Syngas cooler Pilot    
Hot gas filter Pilot    
Quench clean-up Com  
O2 feed Com      
Boiler Com 

H2 S capture and recovery

Rectisol® Com      
Selexol® Com 
Claus/SCOT Com       

Fuel synthesis
Slurry bed reactor Com      
Fixed bed reactor Lab 
Syngas recycle Com   

Power island
Gas turbine Com      
Back-pressure ST Com     
Condensing ST Com  

a
Com, commercial; pilot, demonstrated at pilot-scale; lab, demonstrated in laboratory.

predicts higher conversion, but the CO2 content in the reactor sis of all the fuels considered in our study actively promote
increases so much that the predictions of the model become the water gas shift (WGS) reaction as part of the fuel synthe-
questionable (see further discussion on mixed alcohol syn- sis mechanism, and because including a WGS reactor would
thesis below). The unrecycled portion of the syngas fuels the further complicate the plant configuration due to the need to
GTCC island. effectively recover the heat generated by the WGS reaction.
Should one consider a fuel synthesis catalyst with no WGS
3.6. Suitability of H2 /CO ratio activity, for example cobalt for FT synthesis, then a WGS reac-
tor would need to be placed upstream of the FSI to adjust the
In all cases we did not consider a shift reactor to adjust the H2 /CO ratio to the value most suitable for fuel synthesis. The
H2 /CO ratio because the catalysts assumed for the synthe- precise location where such a WGS reactor is placed depends

Table 4 – Summary of key design parameter values for biorefinery simulations and, for comparison, BLGCC and
Tomlinson cases (n.a. = not applicable).
BLS mt/d Biomass used for Net incremental Gas Syngas from wood Synthesis
(MWLHV ) energy production biomass to mill, dry turbineb gasifier goes to loop design
mt/d (%)a
Dry mt/d MWth LHV

Biorefineries
DMEa 2,458 (351) 700 132 186 (5.4%) n.a.c n.a.d 97% recycle
DMEb 2,458 (351) 1,326 250 813 (24%) 6FA Gas turbine 97% recycle
DMEc 2,458 (351) 678 128 165 (4.8%) 6FA Gas turbine One pass
FTa 2,458 (351) 829 156 316 (9.2%) 6FA Gas turbine One pass
FTb 2,458 (351) 2,246 423 1,733 (51%) 7FA Gas turbine One pass
FTc 2,458 (351) 2,704 505 2,191 (64%) 6FA Synthesis One pass
MA 2,458 (351) 760 143 247 (7.2%) 6FA Synthesis 76% recycle

From 2003 BLGCC study (Larson et al., 2003)

Tomlinson 2,722 (393) 308 58 0 (0) n.a.c n.a.d n.a.e
BLGCC 2,458 (351) 432 81 −82 (−2.4%) 6FA n.a.d n.a.e

a
This is the net additional biomass that must be handled in the woodyard at the mill. It is the sum of pulpwood logs and purchased wood
residues less the pulpwood log flow in the reference (Tomlinson) case.
b
The rated output in simple-cycle mode when operating on natural gas is 75 MW for the 6FA and 170 MW for the 7FA.
c
These configurations include just a steam cycle, with no gas turbine.
d
In these configurations there is no wood gasifier.
e
In these configurations there is no fuel synthesis.
1300 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Fig. 1 – Schematic of DME biorefineries representing DMEb and DMEc designs—the flow represented by the dashed line is
absent in the case of DMEc. The DMEa design replaces the shaded modules of the DMEb configuration with a boiler steam
turbine system.

Fig. 2 – Schematic of biorefinery FTc. Unlike the three DME designs and the two other FT designs, in this case gasified
woody biomass is mixed with gasified black liquor flowing to the synthesis island, and only unconverted syngas is used as
fuel in the gas turbine island.

on whether the WGS catalyst is sulphur tolerant or not. With 4. Biorefinery simulation assumptions
sulphur tolerant catalysts, the WGS reactor can be placed
along the syngas cooling line, ahead of sulphur removal, The following summarizes the assumptions adopted for the
at a temperature within the optimal operating range of the most crucial components and subsystems. The full set of
catalyst. assumptions can be found in Larson et al. (2006a).

Fig. 3 – Schematic of biorefinery MA. The design is similar to FTc in that syngas from both the black liquor gasifier and the
biomass gasifier are processed through the synthesis reactor. MA differs from FTc in that a significant fraction of the
unconverted synthesis gas is recycled for further conversion.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1301

Table 5 – Basic black liquor gasifier operating Table 6 – Basic biomass gasifier operating parameters.
parameters.
Pressure 36 bar (507 psig)
Pressure of raw syngas 35 bar (483 psig) Steam mass flow 28% of bone dry wood
Temperature of raw syngas before quench 1000 ◦ C (1832 ◦ F) Steam pressure 38 bar (536 psig)
Heat loss to environment 0.5% of BL HHV Outlet temperature 950 ◦ C (1742 ◦ F)
Heat to cooling flows 2.0% of BL HHV Heat loss 0.5% of biomass HHV
Carbon conversion 99.9%
Mol fraction at cracker outlet
Methane in raw syngas 1.5% mass in raw gas
CH4 1%
Tar 0.2 g/Nm3
4.1. Black liquor gasifier

Table 7 – Composition and heating value of hog fuel and

The black liquor gasifier under development by Chemrec
wood waste.
(Fig. 4) is a high-temperature, high-pressure, entrained flow,
Mass fraction (%) Heating value, MJ/kg
oxygen-blown reactor comprising two main sections: an upper
section where black liquor is gasified at temperatures around Bone dry wood
1000 ◦ C and a bottom section where the syngas is quenched C 49.98
and the smelt is collected in a bath to produce green liquor. H 6.12
O 42.49
The basic operating conditions (Table 5) have been chosen in
N 0.55
accordance with data published in the literature and informa-
S 0.06
tion provided by the manufacturer. The composition of the Ash 0.80
syngas exiting the gasifier has been calculated assuming that: LHV 18.66
HHV 20.00
• Carbon conversion and methane in the raw syngas are as Feed wood
specified in Table 5 (unconverted carbon leaves the gasifier Moisture 50
together with the green liquor); HHV 10.01
• Ash and chlorides behave as inert material; LHV 8.12
• All the remaining material constitutes a gas/condensed- Dried wood
phase mixture at equilibrium comprising the following Moisture 20
species: Ar, CO, CO2 , COS, H2 , H2 O, H2 S, NH3 , N2 , Na2 CO3 , HHV 16.01
Na2 SO4 , Na2 S, NaOH, K2 CO3 , K2 SO4 , where the species con- LHV 14.45
taining sodium and potassium can be either liquid or solid
or both, depending on temperature and pressure. gas cooler is replaced by a quench, followed by a condensing
heat exchanger for the production of medium pressure steam.
This calculation scheme, which determines the sulphur split Biomass is pressurized in lock hoppers with a stream of pure
between the syngas and the solid/liquid phase, gives satis- N2 provided from the ASU plant and pressurized by and inter-
factory agreement with specific information provided by the cooled (IC) compressor. In addition to oxygen, the gasifier is
manufacturer of the gasifier, as well as with qualitative indica- also fed with steam to fluidize the bed. The gasifier vessel is
tions found in the literature (Fallavollita et al., 1987; Salmenoja followed by a catalytic cracker for the destruction of tar. In our
et al., 1993; Backman and Salmenoja, 1994). simulations the gasifier and the cracker are calculated alto-
gether in a single step, assuming that the syngas at the cracker
4.2. Biomass gasifier exit is at 950 ◦ C, with all species at chemical equilibrium except
for methane and tar, the latter modeled as phenol (C6 H6 O).
The biomass gasifier is a pressurized oxygen- and steam- The concentration of methane and tar at the cracker exit are
blown, fluidized-bed reactor similar to the one being specified at approximately the same values detected experi-
developed by GTI (Lau et al., 2003). In our case the use of oxy- mentally by Simell et al. (1996). Given the presence of a catalyst
gen as the oxidant does not imply major economic penalties (nickel-based catalysts appear most suited) and the relatively
because an air separation unit (ASU) must be present any- high-temperature, assuming that all species but methane and
how to supply the oxygen needed by the black liquor gasifier. tar are at chemical equilibrium at the cracker outlet appears
Moreover, the oxygen-blown design gives a syngas with supe- reasonable. Experimental data reported by Simell et al. (1996)
rior heating value, which provides benefits on equipment size show equilibrium is nearly reached even at more moderate
(and cost) and the operation of crucial components down- conditions of 5 bar and 900 ◦ C.
stream, including, higher combustion stability for the gas The operating conditions are summarized in Table 6, while
turbine combustor and higher fuel conversion for the fuel syn- Table 7 gives the composition and the heating value assumed
thesis reactor. The amount of oxygen supplied to the gasifier for woody biomass (hog fuel or purchased wood wastes). The
determines the partial oxidation of biomass and thus the gasi- operating pressure of 36 bar has been selected to allow mixing
fication temperature, which is assumed to be 950 ◦ C. This value the biomass syngas with the black liquor syngas (cases FTc
appears a reasonable compromise between the need to avoid and MA) without the need for biomass syngas compression;
ash softening, which becomes more likely when temperature although challenging, it is in the range of pressurized systems
increases, and the need to minimize tar in the syngas, which under development (Lau et al., 1993; Blackadder et al., 1994).
increases when the syngas outlet temperature decreases.
Fig. 5 shows the biomass gasifier arrangement assumed 4.3. Biomass dryer
for cases DMEb, DMEc, FTa and FTb, where the gasifier is fol-
lowed by a syngas cooler for the production of high-pressure The biomass feedstock is dried by direct contact with the flue
steam; as already mentioned, in cases FTc and MA the syn- gas exiting the HRSG to reduce its moisture content from 50%
1302 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Fig. 4 – Plant configuration for BL gasification.

Fig. 5 – Biomass gasification system with syngas cooler.

(typical value of hog fuel and wood wastes) to 20% (upper range particulates and the alkali will be removed by the quench and
of values needed to properly operate the gasifier, Lau et al., in the condensing heat exchanger placed immediately down-
2003). stream. The very stringent sulphur requirements imposed by
The temperature of the flue gas exiting the HRSG is the FSI are met by two further treatments: low-temperature
adjusted to give a temperature of the moist gas exiting the wet scrubbing and Rectisol. In the MA case, the MoS2 catalyst
dryer of 90 ◦ C. A high drier inlet temperature must be avoided considered for the mixed alcohol synthesis requires some H2 S
to minimize release of volatile organic compounds (VOCs) and in the syngas and the very deep removal achievable with Recti-
risk of ignition; Brammer and Bridgwater (1999) mention a sol is unnecessary. In this case we have therefore considered a
maximum value of 250 ◦ C. In those cases where the flue gas wet scrubber followed by a Selexol system, which is less costly
temperature needed to provide all the heat for biomass drying than Rectisol.
required a higher value than this (e.g. FTc, see Fig. 9 discussed The syngas generated from biomass is treated differently
later), we have assumed to recycle the moist gas at the drier depending on whether it is fed to the gas turbine or the FSI.
exit with the gas exiting the HRSG to limit the drier inlet tem- In the configurations where the biomass syngas goes only to
perature at 250 ◦ C. the gas turbine (DMEb, FTa and FTb) we have considered hot
Based on data reported by Consonni and Larson (1996), the clean-up by a ceramic filter operating at a temperature low
dryer electric power consumption has been set at 20 kJel per enough to allow the condensation of alkali on the particles
kg of wet biomass. captured by the filter and, at the same time, high enough to
prevent the condensation of tar all along the syngas feed line.
4.4. Syngas clean-up Alkali condensation on entrained solid particles is likely below
500 ◦ C, while tar is expected to begin condensing around 350 ◦ C
The syngas generated from BLG exhibits high concentrations (Consonni and Larson, 1996). To stay in between these two val-
of particulates, alkali, chlorine and sulphur. In the high- ues, in our calculations we have imposed that the minimum
temperature gasification systems considered here most of the temperature reached by the syngas (at the inlet of the gas
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1303

turbine combustor, after mixing with the unconverted syn- 4.6. FT synthesis
gas coming form the FSI) is 375 ◦ C; consequently, the ceramic
filter placed upstream of the syngas expander operates at a The simplest kinetic models for Fischer–Tropsch synthesis are
temperature which, depending on the ratio between biomass so-called “single-␣” models, where ␣ is the probability param-
syngas and unconverted syngas, turns out to be between 420 eter for hydrocarbon chain growth. Lox and Froment (1993)
and 500 ◦ C (for case FTb, see details in Fig. 8). present one such model, in which chain growth probability is
In the configurations where the biomass syngas goes to the the same for all olefins (Cn H2n ) and paraffins (Cn H2n+2 ). Fox and
FSI (FTc and MA), the two streams originated by the two gasi- Tam (1995) present a triple-␣ model for FT synthesis, which
fiers merge into a single flow, which are treated either in a can more accurately predict product distributions than single-
Rectisol (FTc) or Selexol (MA) system. ␣ models. We have adopted a model, based on Fox and Tam’s
approach, that was developed by Larson et al. (2005). With
Fox and Tam’s approach, the rate of olefin formation varies
4.5. DME synthesis
depending on the carbon number, and the paraffin formation
rate is related to the olefin formation rate.
The direct synthesis of DME from syngas assumed in our FSI
Our simulations assume a slurry bed reactor operating at
takes place in two simultaneously occurring steps: methanol
260 ◦ C, 30.7 bar, where an iron catalyst is dispersed in the FT
synthesis and methanol dehydration. The process comprises
wax product. Space velocity is 5800 l per kg of catalyst per
the following reactions:
hour. The advantages of the slurry bed are the same already
described for DME, with the added benefit that for FT synthesis
CO + 2H2 ⇔ CH3 OH H◦ 298 = −94.084 kJ/mol (1) the CO conversion achievable with a single pass is extremely
high: at the operating conditions considered here, the kinetic
model of the FT reactor gives a CO + H2 conversion of over 60%.
CO2 + 3H2 ⇔ CH3 OH + H2 O H◦ 298 = −52.814 kJ/mol (2)
The main disadvantage is the complexity of catalyst separa-
tion from the FT waxes. Due to its water gas shift activity, the
CO + H2 O ⇔ H2 + CO2 H◦ 298 = −41.270 kJ/mol (3) iron catalyst is particularly suited to the low H2 /CO syngas
generated by the black liquor and biomass gasifiers.
The raw liquid product of the FT synthesis reactor is
2CH3 OH ⇔ CH3 OCH3 + H2 O H◦ 298 = −19.76 kJ/mol (4)
a mixture of distillate, naphtha and wax which needs fur-
ther processing to upgrade it to gasoline and diesel fuel.
Reactions (1)–(3) are catalyzed by a methanol synthesis cat- Because of the complexity of upgrading and the small scale
alyst (e.g. CuO/ZnO/Al2 O3 ) and reaction (4) is catalyzed by of our biorefinery (compared to petroleum refineries), we have
an acidic catalyst (e.g. ␥-aluminum). Reactions (1)–(4) show assumed that the raw liquid (after stabilizing by hydrotreating)
a high degree of synergy provided that the operating condi- is exported to an existing petroleum refinery without further
tions are optimized enough to remove methanol effectively. processing. The raw FT product must be kept warm in order
Water formed in reactions (1) and (2) is removed via the WGS to maintain the waxes as liquid, thus requiring trucks with
reaction (3) to produce hydrogen which kinetically favors the heated tanks.
production of methanol.
Because of the water gas shift activity of the methanol cat- 4.7. Mixed alcohol synthesis
alyst, the relatively low H2 /CO ratios that characterize syngas
from biomass can be accommodated in a DME synthesis reac- Given the lack of any published models in the literature, a
tor, assuming that the reaction heat can be removed rapidly new kinetic model was developed to estimate CO conversion
enough to maintain an acceptable temperature for the cat- and alcohol formation over a Mo-based catalyst. The model
alyst. A liquid phase process (LPDME), where the synthesis was developed at the Department of Chemical Engineering of
reaction takes place in a slurry of inert oil and liquid reaction Politecnico di Milano (see Larson et al., 2006a, volume 2). The
products with powdered catalyst-laden particles, provides for kinetic model is based on the experimental data published by
effective mixing and heat removal. LPDME is the DME synthe- Gunturu et al. (1998) on the synthesis of methanol and higher
sis technology we simulated. alcohols from syngas by means of a C-supported, K-promoted
It was decided to base the analysis and the simulation of Co–Mo sulfide catalyst (Mo–Co–K/C). The work of Gunturu et
the process taking place in the LPDME reactor on the model for al. provides data on the whole set of reactions for higher alco-
methanol synthesis proposed by Graaf (Graaf et al., 1988; Graaf hols synthesis, as well as rate expressions and estimates of
and Beenackers, 1996), based on an extensive set of accurate kinetic parameters.
kinetic experiments, and the methanol dehydration model Since the experimental data reported by Gunturu refer to
developed by Ng et al. (1999). The model we use was originally an inlet flow with no CO2 , in the kinetic model developed here
developed by Larson and Ren (2003) and further modified by CO2 is assumed to behave as an inert and its influence on
Celik et al. (2004). the reaction rates is due solely to the decrease of the partial
The LPDME reactor is assumed to operate at 260 ◦ C, pressure of the reactants. Large concentrations of CO2 could
62.5 bar, with a space velocity of 6000 l per kg of catalyst influence the kinetic characteristics of the catalyst and reduce
per hour. The nearly isothermal conditions and the effective its activity and/or its selectivity to higher alcohols. In the MA
gas–liquid–solid heat/mass transfer allow reaching DME molar case considered here the concentration of CO2 in the reactor
fractions in the reactor output gas of the order of 5%. Such one- inlet gas is about 30%, a value for which an impact on the cat-
pass performance is inadequate for a plant aimed at producing alyst behavior cannot be excluded. Consequently, the alcohols
just DME; however, in our co-production schemes unconverted production calculated by our model may be optimistic and
syngas is effectively used to produce power, so that moderate requires verification with further experimental data under
liquid fuel productions can still be attractive. conditions with high CO2 concentration.
1304 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Given the issues yet to be solved for slurry reactors for MA

Case FTb

1 atm
Syngas
synthesis, in particular the availability of a suitable inert liquid
and the feasibility of an appropriate catalyst/liquid separa-

20 ◦ C,

7.28
tion system, in this case we have assumed that the reactor

387.9

64.9

21.5
452.8
16.0

186.5

1.0
4.0
420

1,316
626

Our calculation
is fixed bed. The operating conditions are 350 ◦ C, 100 bar, the
unconverted syngas recycle ratio is 76% and the space velocity
is 3000 l per kg of catalyst per hour. Evaporator tubes placed

7FA, 60 Hz

Conventional applications
across the bed carry away the heat generated by the synthe-
sis reaction and maintain a constant temperature throughout

1 atm)
the reaction zone; the actual arrangement needed to achieve

Natural gas
this is a crucial issue of reactor design and its feasibility needs

Table 8 – Comparison between gas turbine performance published by General Electric and predictions with simulation software GS (n.a. = not available).

(15 ◦ C,
further investigation.

48.91
432.0

9.6

16.3
441.6
15.5

171.6
36.7
0.0
0.0
Fig. 6 reports the configuration and the mass/heat balances

402

1,316
603
ISO
of the FSI and the Selexol system for H2 S removal. The fig-
ure shows the distillation column for methanol separation,

GE data
an undesirable product which is recycled back to the synthe-

432.0

171.7
15.5

36.2
n.a.
n.a.
n.a.
n.a.
n.a.

n.a.

n.a.
n.a.
602
sis reactor. The higher alcohols left after the distillation of
methanol are further treated in a molecular sieve to remove
water. The output of this sieve, which inevitably also contains

Case MixOH

Syngas
some alcohols, is sent to the gas turbine combustor.

8.80
180.0

14.9

21.1
218.4
16.5

89.7

1.0
4.0
431

1,316
633

–
4.8. Gas turbine

The calculations carried out for this study refer to two Gen-
eral Electric engines: 6FA (medium-scale) and 7FA (large scale). Case FTc

Syngas

5.95
These machines belong to the most advanced generation of

176.4

39.9

25.1
216.3
16.5

89.7

1.0
4.0
431

1,316
635

–
heavy-duty machines now in wide commercial service, also
known as “F” technology. Performance predictions have been
calibrated based on data published by General Electric; a few
1 atm
Case FTa

7.61
Syngas

189.0

29.0

22.6
218.0
16.5

83.9

1.0
4.0
relevant inputs to the gas turbine model have been fine tuned
431

1,316
624

–
to achieve the best possible agreement with the overall per-
20 ◦ C,

formance published by GE. Table 8 compares the performance

Our calculation
quoted by GE with those generated by the calculation model
for operation on natural gas fuel at ISO conditions (15 ◦ C,
Case DMEc

Syngas

1 atm). In our calculations we have assumed that:

6FA, 60 Hz

8.60
192.4

25.5

20.5
217.9
16.5

82.9

1.0
4.0
• The turbine inlet temperature (TIT) is the same as with nat-
431

1,316
620

–
ural gas firing.
• The compressor can operate with a compression ratio up to
Case DMEb

5% above its value with natural gas.

Syngas

• Further increases in fuel flow volume (as needed with syn-

6.95
177.8

32.2

20.5
210.0
16.5

89.5

1.0
4.0

gas compared to natural gas) must be accommodated by

431

1,316
635

reducing the air flow entering the compressor, i.e. by adjust-

ing the inlet guide vanes of the compressor.
Conventional applications

Due to the different flow rate and thermo-physical prop-

erties of syngas compared to natural gas, maintaining the
1 atm)
Natural gas

same TIT of the natural gas version implies higher tem-

peratures throughout the expansion and thus – everything
(15 ◦ C,

4.43
48.91

208.43

else equal – higher blade metal temperatures and shorter

204.0

16.3

15.7

75.7
34.9
0.0
0.0
409

1,316
604

life of the hot parts of the engine. This is why syngas-fired

ISO

gas turbines are typically de-rated (TIT lower by 20–50 ◦ C)

GE data

to maintain the same life and reliability of the natural gas-

204.0

fired version. Our assumption of no change in TIT implies

15.7

75.9
34.8
n.a.
n.a.
n.a.
n.a.
n.a.

n.a.

n.a.
n.a.
604

an increase in Turbine Outlet Temperature (TOT) of 10–30 ◦ C

and can be justified by considering that by the time the Nth
DP at turbine outlet, kPa
Fuel mol weight, kg/mol
Compressor outlet T, ◦ C

pulpmill biorefinery plant is realized, gas turbine technology

Ambient conditions

DP at compr. inlet, kPa

will have advanced such that TIT and TOT of state-of-the-art

Power output, MW
Exhaust flow, kg/s

LHV efficiency, %

gas turbines will be significantly higher than those adopted

Fuel LHV, MJ/kg
Fuel flow, kg/s

Pressure ratio

today.
Air flow, kg/s

As to the 5% increase in pressure ratio, it is within the range

TOT, ◦ C

typically tolerable by the compressors of heavy-duty engines.

TIT, ◦ C
Fuel

Whether such an increase is compatible with lower air flow

can only be verified by the gas turbine manufacturer (when
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1305

Fig. 6 – Energy and mass balances of Selexol system and mixed-alcohols synthesis island.

air flow is decreased by closing the Inlet Guide Vanes, the island heat, and the FSI heat is more than the amount needed
compressor stall margin decreases). by the mill, and thus the steam turbine includes a low pressure
section that expands the steam down to 0.074 bar (condensa-
4.9. HRSG and steam cycle tion temperature 40 ◦ C). In all other cases the steam turbine
is a back-pressure design and steam production is adjusted to
In case DMEa, steam is generated at moderate conditions by match the mill demand by changing the biomass input to the
recovering waste heat from the BLG island and the FSI, as biomass boilers or the biomass gasifier.
well as by the existing hog fuel boilers. In this case we have At the scale of our power islands, a combined cycle would
assumed to adapt the steam turbine already existing at the normally feature at least two evaporation pressures in the
mill to the new operating conditions, which basically involves HRSG. In our case however a single evaporation pressure at
modifying the first few stages. 130 bar is adequate for three reasons: (i) large generation of
Table 9 summarizes the design parameters assumed for the steam in the FSI, which effectively substitutes the medium
HRSG of the combined cycle. In plants FTb and FTc the steam pressure evaporator; (ii) massive steam extraction at relatively
recovered using the gas turbine exhaust heat, the gasification high pressure (13 bar) for the mill, which eliminates the ben-
efits achievable by a second evaporator pressure in the HSRG;
(iii) relatively high-temperature (above 200 ◦ C) of the HRSG flue
gases, which further downstream must provide the heat to dry
Table 9 – Basic operating parameters for the HRSG and the biomass to be fed to the gasifier.
the steam cycle.
HRSG
4.10. Heat integration
Evaporation pressure 130 bar (1870 psig)
Evaporation pressure in DME 38 bar (536 psig) Given the large amounts of waste heat made available by
and FT reactors
syngas cooling, syngas clean-up and fuel synthesis, a proper
Evaporation pressure in MixOH 130 bar (1870 psig)
integration ensuring the best use of such heat is crucial to the
reactor
Superheater pressure loss 10% achievement of superior performance.
Superheater temperature (large 565 ◦ C (1049 ◦ F) High-temperature syngas streams from the black liquor
scale GT) gasifier and the biomass gasifier are cooled to generate HP
Superheater temperature 540 ◦ C (1004 ◦ F) steam for the power island and the mill. Whenever possible,
(medium-scale GT) the heat made available at medium or low temperature is used
Heat loss 0.7% of heat released by gas
to preheat make-up water. This is particularly relevant for the
Blow-downa 1.14 kg/s
plants with a large biomass gasifier, where the biomass dryer
Deaerator pressure 4.8 bar (55 psig)
Condenser pressure (plants 0.074 bar (1.07 psia) fed with the HRSG exhausts requires a large thermal input,
FTb and FTc) thus creating a deficit of the low-temperature heat needed to
heat make-up water and condensate streams.
a
Blow-down has been kept to a fixed amount because the steam To minimize irreversibilities and thus maximize efficiency,
flow sent to the mill is also fixed (see Table 1).
we have assumed that heat is transferred across relatively
1306 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Fig. 7 – Plant configuration and mass/energy balances for DMEa case.

Fig. 8 – Plant configuration and mass/energy balances for FTb case.

 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1307

Fig. 9 – Plant configuration and mass/energy balances for FTc case.

small temperature differences: a minimum temperature dif- • Aspen Plus, a commercial software package originally devel-
ference of 10 ◦ C for gas–liquid heat transfer and ∼15 ◦ C for oped at MIT and now offered by AspenTech Inc.
gas–gas heat transfer.5 Heat exchangers operating at medium
or low temperature are assumed to incur negligible heat The final process designs were arrived at after reviewing
losses. For the syngas cooler of the biomass gasifier, which detailed simulation results from initial designs with engineers
handles gas between 950 ◦ C and 450–500 ◦ C, we have assumed at Nexant (the engineering firm that subsequently developed
a heat loss of 2% of the heat transferred. the “Nth plant” capital cost estimates) and revising the initial
designs accordingly to improve performance and/or reduce
5. Performance estimates capital costs and parasitic power consumption. As illustra-
tive examples of our calculations, Figs. 7–9 show the simulated
The calculation of mass and energy balances is the funda- mass/energy balances for cases DMEa, FTb and FTc.
mental step required to estimate the performances and costs.
Mass and energy balances determine how much electricity 6. Performance simulation results
and fuel can be generated by a given black liquor flow, as
well as how much extra fuel (biomass or possibly fossil fuel) Table 10 summarizes all of our performance results and
is needed to meet the mill steam requirements. Mass and compares them to our prior BLGCC and Tomlinson results,
energy balances also give the basic information needed to size generated using the same analysis framework (Table 11 gives
and estimate the costs of the key pieces of equipment. All higher and lower heating values for the various feedstocks
this allows estimating the capital and operating costs, and and fuels in our analysis). Fig. 10 summarizes the biorefin-
therefore the costs of the final products (electricity and/or ery and Tomlinson energy balances, showing energy inputs
fuel). and outputs. The ratio between the height of the “out” bar
The mass and energy balances and the overall performance and that of the “in” bar is the overall energy efficiency accord-
of each pulpmill biorefinery configuration have been evaluated ing to the First Law of Thermodynamics. It is an indication
by combining the use of two computer codes: of the overall energy performance. A more accurate compari-
son of performance, by assigning appropriate thermodynamic
• GS, a code developed for research purposes at Politecnico di
values to different plant outputs, is given in Fig. 11, where
Milano and Princeton University (Consonni, 1992; Macchi et
each output (fuel, electricity, heat, and chemical services) is
al., 1995; Consonni and Larson, 1996);
converted to an equivalent amount of electricity according to
the conversion coefficients in Table 12. The left bar in each
5
In an actual situation these temperature differences are 2-bar set (ETA 1st) is the efficiency that can be discerned
subject to economic optimization to determine the best trade-off from Fig. 10. The right bar (ETA eq electr) is the efficiency
between efficiency and capital costs. calculated when all outputs are converted to an electricity-
1308 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Table 10 – Biorefinery performance estimates, with comparisons to Tomlinson and BLGCC. Units are megawatts unless
otherwise indicated. Fuel values are given on a lower heating value (LHV) basis.
Power only Biorefineries

Tomlin. BLGCC DMEa DMEb DMEc FTa FTb FTc MA

Fuel inputs
Black liquor 392.6 350.7 350.7 350.7 350.7 350.7 350.7 350.7 350.7
Hog fuel 57.8 54.1 54.1 54.1 54.1 54.1 54.1 54.1 54.1
Total mill by-product fuels 450.4 404.8 404.8 404.8 404.8 404.8 404.8 404.8 404.8
Purchased wood residues 0 27.1 77.4 195.5 73.4 101.6 368.5 451.3 89.2
Natural gas to duct burner – 12.9 – – – – – – –
Lime kiln #6 fuel oil 31.1 35.9 35.9 35.9 35.9 35.9 35.9 35.9 35.9
Total purchased fuels 31.1 75.9 113.3 231.4 109.3 137.5 404.4 487.2 125.1
Total fuel inputs 481.5 480.7 518.1 636.2 514.1 542.3 809.2 892.0 529.9

Steam to mill
LP (55 psig) steam to process 142.8 135.3 135.3 135.3 135.3 135.3 135.3 135.3 135.3
MP (175 psig) steam to process 69.3 64.9 64.9 64.9 64.9 64.9 64.9 64.9 64.9
Total process steam production 212.1 200.2 200.2 200.2 200.2 200.2 200.2 200.2 200.2

Liquid fuel production – – 168.0 168.0 74.2 111.6 111.6 342.7 59.6
Barrels per day petroleum equiv.a – – 2362 2362 1043 1549 1549 4757 948

Electricity balance
Gas turbine gross output – 87.0 – 89.5 82.9 83.9 186.5 89.7 89.7
Steam turbine gross output 72.0 48.2 32.9 42.0 38.7 34.0 87.9 48.6 40.8
Syngas expander output – – 2.58 5.01 1.96 1.65 4.26 – 2.99
Total gross production 72.0 135.1 35.5 136.5 123.6 119.5 278.7 138.3 133.5
Air separation unit power use – 14.9 15.8 27.0 21.5 22.8 35.7 38.4 22.1
Syngas compressor power use – – 1.95 1.95 2.17 – – – 7.37
Gas compressors (mainly recycle) – – 7.26 7.26 – – – – 1.77
Steam cycle auxiliaries 6.7 1.2 1.38 2.30 1.52 1.31 3.76 5.40 2.75
Black liquor gasification island aux. – 2.7 2.67 2.67 2.67 2.67 2.67 2.67 2.67
Biomass gasification island aux. – – – 3.37 1.74 2.11 5.66 6.75 1.95
Biomass boiler auxiliaries 1.00 1.2 1.75 – – – – – –
Acid gas removal and S recovery aux. – 1.1 0.97 0.97 0.96 0.96 0.96 2.59 1.81
Refrigeration for AGR – – 2.15 2.15 2.07 2.01 2.01 5.11 1.21
Synthesis island auxiliaries – – 1.00 1.00 0.40 – – – –
Total recovery area use 7.7 21.1 34.9 48.7 33.0 31.8 49.8 60.9 41.6
Net electricity productionb 64.3 114.1 0.56 87.8 90.5 87.7 228.9 77.3 91.9
Power in excess of Tomlinson – 49.8 (63.7 23.6 26.3 23.4 164.6 13.0 27.6
Process use (excl. recovery area) 100.1 100.1 100.1 100.1 100.1 100.1 100.1 100.1 100.1
Mill electricity purchases 35.8 −14.0 99.6 12.3 9.58 12.4 −128.8 22.8 8.21

Energy ratios
Useful outputs/total inputsc 0.574 0.654 0.712 0.717 0.710 0.737 0.668 0.695 0.664
Effectiveness of purchased fuel used – 1.11 1.27 0.956 1.28 1.27 0.740 0.780 0.928

a
Barrels (bbl) per day of equivalent petroleum-derived fuel potentially displaced by the biorefinery fuel. The potentially displaced fuels are:
diesel (6.15 GJ/bbl, LHV) for DME, crude oil (6.22 GJ/bbl, LHV) for FTL, and gasoline (5.43 GJ/bbl, LHV) for MA.
b
For the BLG cases, this is net of 0.58 MW attributable to the generation of the oxygen used for delignification.
c
This is the sum of process steam, net electricity, and liquid fuel outputs divided by the sum of all fuel inputs.
d
This is the sum of net electricity and liquid fuel production for the gasification-based facility minus this quantity for the Tomlinson facility
divided by the difference in total purchased fuel between the gasification and Tomlinson facilities.

equivalent basis. Finally, Fig. 12 gives the variations of energy

inputs and outputs with respect to the reference Tomlinson
system.
The results reported in tables and figures suggest a number
Table 11 – Lower and higher heating values for of informative comparisons.
biorefinery feedstocks and various liquid fuels.
LHV HHV • The ratio of useful energy outputs to total energy inputs
Black liquor dry solids 12.3 MJ/kg 13.9 MJ/kg (second row from bottom in Table 10) is higher for all the
Woody residues (50% moisture) 8.1 MJ/kg 10.0 MJ/kg biorefineries than for the BLGCC case and substantially
Fischer–Tropsch synthetic crude 31.2 MJ/l 33.5 MJ/l higher than for the Tomlinson case. This reflects the more
Dimethy ether 19.2 MJ/l 21.1 MJ/l effective use of biomass with gasification-based conversion
Mixed-alcohols 28.2 MJ/l 31.0 MJ/l
and with producing a more diverse product slate.
Ethanol 21.3 MJ/l 23.6 MJ/l
• All the biorefineries (and the BLGCC) require more fuel
Petroleum gasoline 32.4 MJ/l 34.5 MJ/l
Petroleum diesel 34.5 MJ/l 36.3 MJ/l inputs than the Tomlinson case. How well are these incre-
Petroleum low-sulphur diesel 36.1 MJ/l 38.6 MJ/l mental inputs being used? The “effectiveness of purchased
fuel use” (last row of Table 10) exceeds 100%, or nearly so,
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1309

Fig. 10 – Energy balance of biorefinery simulations and Tomlinson system.

Fig. 11 – Energy efficiencies and contribution of each output (steam heat, liquid fuel and electricity) to 1st-Law efficiency
(ETA 1st) and to equivalent electric efficiency (ETA eq. electr.) ETA fuel is the ratio between fuel output and all fuel inputs
(black liquor, hog fuel, purchased wood residues, lime kiln fuel oil, see Table 10), all evaluated by their LHV; ETA heat is the
ratio between process steam production (in MWth, see Table 1) and fuel inputs. 1st-Law efficiency is simply the sum of ETA
fuel, ETA heat and ETA electricity. Equivalent electric efficiency is a weighted sum where fuel and heat output are weighed
by the conversion coefficients reported in Table 12.

Fig. 12 – Biorefinery energy inputs and outputs variations with respect to Tomlinson system.
1310 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Table 12 – Coefficients adopted to convert the energy content of liquid fuels and steam to equivalent electricity, whereby
the equivalent energy efficiencies reported in Fig. 11.
DME FT MixOH MPsteam LPsteam

K1 MWel /MWfuel,LHV 0.55 0.50 0.55 – –

K2 kJel /kgsteam – – – 610
K3 kJel /kgsteam – – – – 510

K1, electricity that could be produced by a Combined Cycle fed with the liquid fuel. The electrical efficiency of the Combined Cycle is assumed
55% for DME and MA, 50% for FT raw liquid (which implies that the efficiency of the conversion from raw FT to light fuels is 91%). K2, electricity
that could be produced by expanding Medium Pressure (MP) steam from 13 bar to 0.074 bar in a steam turbine with isoentropic efficiency 81%,
mechanical–electrical efficiency 98%. The corresponding ratio between electricity and heat is 0.31. K3, electricity that could be produced by
expanding Low Pressure (LP) steam from 4.8 bar to 0.074 bar in a steam turbine with isoentropic efficiency 84%, mechanical–electrical efficiency
98%. The corresponding ratio between electricity and heat is 0.24.

in six of the eight cases indicating the very effective use of • Scale effects also explain part of the differences between
additional energy inputs.6 In other words: (i) to satisfy the FTc (based on the medium-scale gas turbine 6FA) vs FTb.
requirements of the mill, BLG requires an energy input sig- On the other hand, the relatively low equivalent electric effi-
nificantly larger than that made available by the mill itself; ciency of FTc is also due to the irreversibilities brought about
(ii) not only does the extra energy input required with BLG by the quench biomass gasifier.
allow meeting the mill steam demand; it also allows a very • The FTc and MA designs both utilize a feed of combined
efficient production of extra electricity and/or fuels. syngas from black liquor and from biomass gasification to
• A comparison between the BLGCC and DMEa cases is of the synthesis reactor, as well as the same quench biomass
interest because the former exports electricity only and the gasifier. This translates to similar equivalent electric effi-
latter exports liquid fuels only, while both utilize a black ciencies, even though the very different fuel synthesis rates
liquor gasifier and both meet the pulp/paper mill’s steam give a very different electricity/fuel mix. Because of the
needs using hog fuel boilers. The DMEa case shows a higher low MA synthesis rates, the conversion to electricity (ETA
useful energy output ratio, although this will not necessarily electricity in Fig. 11) is relatively high, and the purchased
translate to better financial performance since consider- biomass that must be used to ensure sufficient fuel gas for
able purchases of electricity by the pulp/paper mill will be the gas turbine is much more modest than in the FTc case.
required with the DMEa system. This is true even though the FTc case uses a once-through
• A comparison between DMEa and DMEb highlights the synthesis design and the MA case uses recycle, because the
impact of the superior performances of {biomass gasifica- overall conversion of syngas to liquids (in LHV terms) is
tion + combined cycle} vs {power boilers + steam cycle} as much higher for once-through FT (49%) than for recycle MA
means to generate power. As shown in Fig. 11, biomass (17%).
gasification and combined cycle allow DMEb to achieve an • The MA and DMEb systems are similar insofar as they both
equivalent electric efficiency more than eight percentage utilize a recycle synthesis loop and the same gas turbine.
points higher than that of DMEa, even though 1st-Law effi- The higher overall conversion of syngas to liquids in the
ciencies are nearly the same. This also points the limits of DMEb case means that a larger amount of biomass must be
1st-Law efficiency as an indicator of performance. purchased since there is less unconverted syngas available
• The similar values of 1st-Law and equivalent electric to fuel the gas turbine.
efficiency of DMEb and DMEc indicate that their “thermody-
namic quality” is nearly the same. Therefore, the adoption Given all this, the selection of the most appropriate
of a once-through (like in DMEc) rather than recycle (like in technology and configuration to be adopted for a specific appli-
DMEb) configuration of the FSI “shifts” production from fuel cation will have to be assessed based on a number of local
to electricity with minor effects on efficiency. technological and market conditions, as well as on the out-
• The higher fuel conversion efficiency (ETA fuel, see again come of the pilot-scale demonstrations and developments
Fig. 11) of FTa vs DMEc, which have nearly identical now being pursued.
equipment configurations, follows from the much higher
one-pass syngas conversion efficiency of FT synthesis vs 6.1. Liquid fuel produced per ton of biomass input
DME synthesis. For the same reason, purchased biomass
use is higher in the FTa case because more of the fuel for A commonly cited metric for liquid biofuels production is
the gas turbine must be provided from gasified biomass. the liquid yield (volume) per unit of biomass input (mass).
• The higher equivalent electric efficiency of FTb vs FTa fol- Fig. 13 shows biofuel yields expressed in terms of liters of
lows from the superior performance of the much larger gasoline energy equivalent per kilogram of dry biomass feed-
scale gas turbine and steam turbine adopted in FTb, stock (lower x-axis) and also in terms of liters of ethanol
even though the adoption of a condensing (rather than energy equivalent per dry kilogram of biomass (upper x-axis).
back-pressure) steam cycle obviously decreases 1st-Law Included in this figure are results from several studies in addi-
efficiency. tion to ours, including those by engineers at the National
Renewable Energy Laboratory for corn stover conversion under
U.S. conditions (NREL, Aden et al., 2002; Aden et al., 2005;
6 Mcaloon et al., 2000; TSS Consultants, 2005); three cases for
We use “effectiveness”, rather than efficiency, since efficiency
>100% is not possible. Values of effectiveness >100% are possible European Union pulpmill conditions involving black liquor
because the denominator includes only purchased biomass (not conversion (EU, Ekbom et al., 2005); and several biorefinery
black liquor and hog fuel). designs for U.S. conditions using switchgrass (swg) analyzed
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1311

Fig. 13 – Comparison of adjusted liquid fuel yields.

by Larson et al. (2005). All of the designs included in Fig. 13 gen- of the purchased residues are charged to the added elec-
erate one or more co-products with the biofuel. Appropriate tricity production;
adjustments to the amount of biomass charged against liquid (iv) for the pulp mill biorefineriers that produce less elec-
fuel production have been made to account for co-products, tricity than the Tomlinson systems they would replace,
as discussed next (see also Larson et al., 2006a). the liquid fuels production is charged with the additional
All results above the horizontal dashed line in Fig. 13 biomass required to reach the same level of electricity
are for plant designs that include some level of electricity production as with the Tomlinson.
co-production. To obtain a measure of the effective liq-
uid fuel yield per ton of biomass for these cases, we have The up or down adjustments (iii) and (iv) in the amount of
charged a portion of the input biomass to the electric- biomass charged to liquid fuels are the difference in electricity
ity co-product. The biomass charged to electricity is the production divided by the assumed efficiency for a stand-
amount that would be required to generate the same elec- alone biomass power plant mentioned above (49.5% LHV).7
trical output at a stand-alone biomass power plant having a The adjusted liquid fuel yields are higher or substantially
lower-heating value generating efficiency of 49.5% (Larson et higher for all pulp mill biorefineries than for “stand-
al., 2006b).7 alone” biorefineries that co-produce only liquid fuels and
For the pulp mill biorefinery cases (below the dashed line in electricity.8 The high values arise primarily because the pulp-
Fig. 13), it is appropriate to make some additional adjustments mill biorefineries are providing service beyond liquid fuels and
since these biorefineries provide, in addition to liquid fuel, electricity. In effect, the biomass resource is utilized more effi-
not only some electricity but also process steam and chemical ciently by integrating the biorefinery with a pulp mill than by
recovery services to the mill. More specifically, for the process using biomass in a stand-alone biorefinery. This also has a
designs below the dashed line: positive impact on economics.

(i) black liquor and residues generated on site (hog fuel) are 7. Financial analysis
charged against the pulp mill’s steam, power, and chemi-
cal recovery needs, since this is how these inputs are used To assess prospective economics, biorefinery capital costs
with Tomlinson systems today; were estimated as a basis for calculating effective biofuel pro-
(ii) liquid fuels production is charged only with purchased duction costs and detailed cash flow/internal rates of return
woody biomass residues; on investments in biorefineries replacing existing Tomlinson
(iii) for the pulpmill biorefineries that generate more electric-
ity than the Tomlinson systems they would replace, some
8
Note that in several cases the adjusted yield corresponds to
an energy conversion of the biomass that is greater than 100%. [A
7
This is an optimistic efficiency estimate, considering a liquid yield of 126 gallons of gasoline equivalent per dry metric
large-scale system (440 MWe ) with some idealized assumptions. tonne of biomass, or 202 gallons of ethanol per dry metric tonne
More recent work by one of the authors (EDL) suggests that a of biomass, corresponds to an actual energy conversion of 100%
more realistic efficiency (but still a relatively large plant – 300 on a LHV basis]. Of course, the “adjusted yield” is not a measure
MWe ) would be between 46 and 47% on a LHV basis (42–43% on of actual energy conversion, but nevertheless is an indicator of
HHV basis). how well the biomass resource is being used.
1312 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Table 13 – Estimated overnight installed capital costs (thousand 2005$) and non-fuel operating and maintenance costs
(thousand 2005$ per year). Installed capital costs include engineering, equipment, installation, owner’s costs (including
initial catalyst), contingencies, and spare parts.
Thousand 2005$ Power/steam Biorefinery—power/steam/liquid fuel

Tomlin. BLGCC DMEa DMEb DMEc FTa FTb FTc MA

Recovery boiler 125,018 0 0 0 0 0 0 0 0

Steam system modifications 11,136 0 3,000 0 0 0 0 0 0
Air separation unit (ASU) 0 42,628 43,053 61,561 52,933 55,001 72,762 77,823 54,080
ASU increment for O2 delig. 0 1,118 1,061 879 954 933 805 776 948
BL gasifier and green liquor filter 0 63,720 63,720 63,720 63,720 63,720 63,720 63,720 63,720
Nitrogen compressor 0 0 0 1,188 810 1,071 1,757 2,013 5,181
Acid gas removal and sulphur recov. 0 19,003 37,732 37,732 27,321 27,321 27,321 42,164 24,529
Synthesis island 0 0 49,344 49,344 16,287 22,019 22,019 38,767 83,548
Combined cycle power island 0 89,243 0 105,303 100,091 90,018 171,895 104,300 90,348
Wood yard expansion 0 0 867 2,697 789 1,303 4,832 5,788 1,077
Biomass dryer, including RTO 0 0 0 50,295 32,523 37,286 72,507 45,558 31,383
Biomass gasifier and tar cracker 0 0 0 28,354 18,320 20,867 41,365 47,063 22,949
Biomass syngas cooler and filter 0 0 0 8,484 4,998 5,666 11,372 0 0
Biomass syngas cooler and wash 0 0 0 0 0 0 34,425 16,092
Biomass syngas expander 0 0 0 3,778 2,661 2,670 9,410 0 0
Hog fuel boiler 0 0 50,736 0 0 0 0 0 0
Other 0 2,359 2,359 2,359 2,359 2,359 2,359 2,359 2,359
Overnight installed capital cost 136,154 218,072 251,873 415,695 323,766 330,234 502,125 464,755 396,215
Annual non-fuel O&M cost 5,446 8,723 10,075 16,628 12,951 13,209 20,085 18,590 15,849

systems at the end of their working lives. The capital cost the investment cost for generating this electricity in a separate
considered in all cases is the incremental capital investment stand-alone facility.10
required for the biorefinery over the cost installing a new Tom- Capital charged to biofuels from our analysis and those
linson system. from some other studies are shown in Fig. 14. The full range of
scales (barrels per day) shown for biofuels is tiny compared to
7.1. Capital cost estimates typical scales for fossil fuel conversion (see reference points
on the right side of the figure). Nevertheless, the biorefinery
Capital cost estimates consistent with previous estimates for investment costs are in the same range as investment costs for
Tomlinson and BLGCC systems (Larson et al., 2003) were devel- coal-to-liquids facilities with installed capacities 20–50 times
oped for each biorefinery design by engineers at Nexant.9 as large. This highlights the benefit (capital cost credits) pro-
Based on the detailed mass/energy balances described earlier, vided by integrating the biorefinery with the pulp and paper
Nexant was asked to provide estimates with ±30% uncertainty mill.
and to assume “Nth plant” levels of technology maturity and
operational reliability. The resulting capital and non-fuel oper- 7.2. Internal rate of return analysis11
ating and maintenance costs are given in Table 13.
It is of interest to compare the capital investment charged The analysis in the previous section gives an indication of
to liquid fuels production capacity for our biorefineries with favorable economics for pulpmill biorefineries. Here we quan-
the investments needed to produce liquid fuels through other tify more precisely the economics via a detailed rate of return
routes. The capital charged to liquid fuels for our biorefineries analysis that accounts for all capital investments and all major
is the extra investment required relative to an investment for changes to mill operations that would result from investing
a new Tomlinson system. The biorefinery provides the same in an integrated biorefinery instead of a new Tomlinson sys-
process steam and chemical recovery services to the mill as tem. The analysis includes reduced pulpwood purchases due
a Tomlinson system, and provides less or more electricity as to higher digester yield with polysulfide pulping, increased use
a Tomlinson (depending on the specific biorefinery design). If of #6 fuel oil in the lime kiln, the purchase of wood residues,
the biorefinery delivers more electricity to a mill than would the sale of electricity and biofuels, and other factors. Avoided
a Tomlinson system, a credit against the capital charged to costs (e.g., avoided grid power purchases) are also included.
liquid fuels is appropriate. If the biorefinery delivers less elec- Financial parameter assumptions are in Table 14.
tricity to the mill than would a Tomlinson, then an additional Because of uncertainties about future energy prices, we
charge against liquid fuels production is appropriate. We credit consider two scenarios. We based our scenarios on the
(or charge) the biorefinery for the difference in electricity deliv- 2006 U.S. Department of Energy’s Reference and High price
ered to the mill at a rate of $968/kWe , an optimistic estimate of
10
Assuming higher than $968/kWe would result in lower values
of capital charged to biofuels in our biorefineries.
9 11
Capital costs in this paper do not reflect the unprecedented The financial analysis was prepared with cost factors and
run-up in costs of materials and construction labor observed energy price assumptions from 2006. These assumptions are
during the past several years. Thus, the absolute capital costs quite different from those that would be made today, for both
reported in this work underestimates costs for systems that capital and energy costs. Nevertheless, the relative performance
would be built today. However, since recent cost escalations is still expected to be valid, and the original financial analysis
affect all technologies, relative comparisons among different (Larson et al., 2006a) provided for sufficient sensitivities that
systems are still valid. include today’s cost environment.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1313

Fig. 14 – Effective capital investment intensity (2005$ per barrel diesel equivalent per day) for liquid fuels production as a
function of production capacity.

electricity sales. Existing and potential federal and state

Table 14 – Key input assumptions for the financial
analysis (MACRS, Modified Accelerated Cost Recovery incentives (biofuels excise tax credits, tax exemptions and
System, a common asset depreciation method of the production tax credits) may also be available. The impacts of
U.S. income tax code). such incentives and premiums were considered in sensitivity
Financial parameters analyses.
Inflation rate 2.14% Fig. 15 gives IRR and NPV results under the REP scenario.
Debt/equity 50%/50% With no incentives (left bar in each set of two bars), the IRR
Interest rate on debt 8% for most configurations exceeds a 15% hurdle rate. The largest
Return on equity (hurdle rate) 15% systems in terms of feedstock input (FTb and FTc) demon-
Income tax rate; property tax 40%; 2%
strate the most attractive biorefinery economics, but do not
and insurance
quite reach the level of return with a BLGCC system. The weak-
Economic life (years) 25
Depreciation method 20-year MACRS est financial performance is for the MA configuration.13
Construction time (months) 24 (Tomlinson) 30 When a bundle of plausible incentives is considered,
(gasification) including an excise tax credit (equivalent, on an energy basis,
P&P industry/mill assumptions to the one currently available for ethanol), an investment tax
O&M cost inflator (% per year, 2.67% credit for gasification investments (mandated by EPAct 2005),
current $) a 10-year production tax credit for renewable electricity of $9
Annual operating hours 8,330 per MWh, and a Renewable Energy Credit (REC) of $20/MWh,
Start-up (% of full output), year 80%; 100% the financial performance is overwhelmingly attractive for all
1; year 2–25
cases, with most IRRs between 30% and 35%.
In the TSEP scenario (not shown), the IRRs, even without
projections (EIA, 2006), referring to these as our Reference incentives, exceed the hurdle rate by a wide margin in all
Energy Prices (REP) and Tight Supplies Energy Prices (TSEP) cases except MA. They also surpass the BLGCC option by sev-
scenarios.12 We utilized year-by-year values for oil and other eral points, since the BLGCC option is largely unaffected by
energy prices over the 25 years spanned in each of these changes in petroleum prices.
two scenarios. Table 15 summarizes our price assumptions,
giving levelized values over the projection period. Our two
8. Industry impacts
scenarios can be characterized in short-hand by the levelized
world crude oil price calculated from our year-by-year oil price
assumptions. In the REP scenario, this levelized price is $50 per The technical potential of the U.S. kraft pulp and paper indus-
barrel (2005 $). In the TSEP scenario, it is $78 per barrel. The lat- try for liquid biofuels production, if the entire industry were to
ter represents an extrapolation of mid-2006 oil prices, but it is adopt a biorefining configuration like the FTc case (for which
worth noting that since constructing these scenarios in 2006, liquids production is highest among all cases examined), is in
energy prices have been extremely volatile, and have covered excess of 9 billion gallons per year of FT synthetic crude oil.
a range even greater than that covered by the two scenarios. On an energy equivalent basis, this corresponds to about 14
In addition to direct product sales, biorefineries may also
be able to capture monetary value from environmental and
renewable energy attributes, e.g., a premium for renewable 13
Separating the mixed alcohols and selling the components as
chemicals (rather than selling the mixture as fuel) would provide
higher returns despite the low product yield because of the high
12
Our price scenarios are not identical to the DOE forecasts, but unit values of chemical alcohols. Chemicals markets are much
our plant gate prices are based on the DOE projections. Also, we smaller than fuel markets, however, so production from pulpmill
have extrapolated a few years beyond the DOE horizon of 2030 in biorefineries would quickly saturate the markets and lower the
order to do a 25-year cash flow. prices paid.
1314 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

Fig. 15 – Internal rates of return and net present values for all cases under the Reference Energy Price Scenario, with and
without incentives included.

Fig. 16 – Net annual CO2 emissions reductions for the United States with biorefining relative to continued use of Tomlinson
systems under the aggressive biorefinery market penetration scenario. CO2 emissions from carbon that originated in
biomass and black liquor are not included in this CO2 accounting, since that CO2 is of recent photosynthetic origin, and
replacement biomass growth will absorb an equivalent amount of CO2 from the atmosphere. (Units here are short tons: 1
short ton = 0.907 metric tonne).

Fig. 17 – Cumulative (25-year) net fossil fuel savings for the United States with biorefining relative to continued use of
Tomlinson systems under different biorefinery market penetration scenarios. (One quadrillion
Btu = 1015 Btu = 1.055 EJ = 1.055 × 1018 J).
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1315

Table 15 – Levelized costs (2005$) for energy commodities (plant gate, no incentives).
Energy commodity Reference energy prices Tight supplies energy
(REP) scenario price (TSEP) scenario

World crude oil price $50/bbl $78/bbl

Electricity – avoided purchases $53.3/MWh $56.2/MWh
Electricity – sales to grid $48.2/MWh $51.8/MWh
Natural gas purchases $5.82/MCF $7.00/MCF
Residual fuel oil purchases $1.00/gallon $1.62/gallon
DME sales as motor diesel substitute $0.72/gallon $1.10/gallon
DME sales as LPG substitute $0.66/gallon $0.99/gallon
FT crude sales as petroleum crude substitute $0.96/gallon $1.54/gallon
FT crude sales as petroleum crude substitute $40.3/barrel $64.7/barrel
Mixed alcohol sales $1.43/gallon $1.77/gallon
Ethanol sales $1.09/gallon $1.35/gallon
Methanol sales $0.54/gallon $0.67/gallon
Propanol sales (representing higher alcohols) $3.64/gallon $3.90/gallon
Purchased wood (higher heating value basis) $1.53/MMBtu $1.53/MMBtu

billion gallons per year of ethanol. For comparison, the corn- as the national payoff on government support for more aggres-
ethanol industry in the U.S. produced about 9.1 billion gallons sive commercial deployment.
in 2008.
Of course, 100% of U.S. kraft pulp mills are unlikely to adopt 9. Conclusions
biorefining, even in the long term. To estimate more realistic
potential industry-wide environmental and energy impacts, Gasification-based pulp mill biorefinery technologies, once
we developed three biorefinery market penetration scenarios. fully commercialized, offer the potential for attractive invest-
The Low scenario was characterized by a 30-year period to ment returns and a revitalization of the pulp and paper
reach market saturation (defined as 90% of technical poten- industry. They also offer the potential for important con-
tial). In the Base scenario, market saturation was reached in tributions toward petroleum savings, emissions reductions,
20 years, and in the Aggressive scenario, 10 years. The large improved energy security, and rural economic development
investments that supported the rapid expansion in building – contributions that could be two to three times larger than
of corn-ethanol plants in the 2006–2007 timeframe in the U.S. those from the existing corn-ethanol industry.
suggests that capital availability alone is not a constraint on These potential private and public benefits arise, funda-
building pulp mill biorefineries, even at the Aggressive rate mentally, because of the integration of biorefining with pulp
(Larson et al., 2006a). and paper production, such that the biorefinery is provid-
Fig. 16 illustrates the potential CO2 emissions benefits for ing chemical recovery services, process steam, and process
the United States under the Aggressive biorefinery market electricity in addition to exporting liquid fuels and perhaps
penetration scenario. CO2 benefits arise primarily because some electricity. Integration can effectively enable more effi-
products of the biorefineries have low net lifecycle CO2 emis- cient use of biomass resources for liquid biofuel production
sions, and they displace fossil fuel derived electricity and compared to non-integrated biofuel production. Integration
petroleum transportation fuels with higher lifecycle CO2 emis- also can effectively reduce the capital investment required
sions. Relative to continued use of Tomlinson technology, per unit of biofuel production to levels comparable to invest-
reductions in net CO2 emissions could reach as much as ments needed for coal-to-liquids facilities that are more than
100 million metric tons per year nationally. This is equivalent an order of magnitude larger than prospective pulp mill biore-
to about 5% of the transportation sector CO2 emissions in the fineries. Finally, integration leads to favorable IRRs under a
U.S. in 2004. The main driver affecting the amount of the CO2 variety of plausible scenarios.
benefit is the amount of purchased woody biomass utilized. The variety of system configurations presented here shows
By comparing results in Fig. 16 for FTb and FTc, which use the impact of fundamental technological features like the type
similar quantities of purchased biomass but have quite dif- of power system (steam cycle vs combined cycle), arrange-
ferent outputs of electricity relative to fuel, one may note that ment of fuel synthesis (once-through vs recycle), synthesis
displacement of grid electricity (generated with a fuel mix rep- process (DME vs FT vs MA), scale of gas turbine, type of
resenting the average generating mix in the U.S.) results in biomass gasifier (syngas cooler vs quench).
greater CO2 reductions per unit of biomass consumed than All of the component technologies needed for gasification-
petroleum fuel displacement. based biofuels production at a pulp/paper mill biorefinery are
Fig. 17 shows cumulative fossil fuel savings for each of either already commercially used (in non-biorefinery appli-
the three market penetration scenarios over the first 25 years cations) or are undergoing pilot-scale demonstration, such
following initial deployment of biorefineries. Net fossil fuel that with some additional targeted research and development
savings relative to continued use of Tomlinson systems range work and effective multi-industry/government partnerships
from 1 to 5.5 quads (1.1–5.8 EJ) in the Low scenario and 3 to to manage overall risk, commercial-scale facilities could pos-
17 quads (3.2–17.9 EJ) in the Aggressive scenario. The difference sibly begin to be built within a decade. The one caveat to
between the Low and Aggressive scenarios can be viewed as this conclusion is that development and demonstration at
a proxy for the “lost” fossil fuel savings opportunity if imple- significant scale of synthesis catalysts for mixed-alcohols pro-
mentation of pulp mill biorefineries is delayed or, alternatively, duction that perform better than the one we modeled are
1316 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317

needed to gain confidence that mixed-alcohols could be pro- Energy Information Administration (EIA), 2006, Annual Energy
duced competitively as a fuel. Outlook 2006, with Projections to 2030, DOE/EIA-0383(2006),
U.S. Department of Energy, February.
Ekbom, T., Berglin, N. and Logdberg, S., 2005, Black liquor
Acknowledgments gasification with motor fuel production—BLGMF-II: A
techno-economic feasibility study on catalytic
The authors thank Jim Frederick, Kristina Iisa, and Charles Fischer–Tropsch synthesis for synthetic diesel production in
E. Courchene (Institute of Paper Science and Technology, comparison with methanol and DME as transport fuels, In
Georgia Institute of Technology), Matthew Campbell and Eti- P21384-1, Nykomb Synergetics AB Stockholm, December,
enne Parent (Navigant Consulting), Farminder Anand and Engstrom, S., Lindman, N., Rensfelt, E. and Waldheim, L., (1981).
Matthew Realff (School of Chemical and Biomolecular Engi- A new synthesis gas process for biomass and peat. In Energy
from Biomass and Wastes V. (Institute of Gas Technology,
neering, Georgia Institute of Technology), Silvia Napoletano
Chicago).
and Wang Xun (Department of Energy Engineering, Politec-
Evans, R.J., Knight, R.A., Onischak, M. and Babu, S.P., 1987, Process
nico di Milano), and Sheldon Kramer (Nexant Engineering) for performance and environmental assessment of the renugas
their contributions to the analysis reported in Larson et al., process, in Energy from Biomass and Wastes X, Klass, D.L. (ed)
2006a, on which this paper is based. We also thank the Steer- (Elsevier Applied Science (London) and Institute of Gas
ing Group that advised the work described in Larson et al., Technology (Chicago)), pp. 677–694.
2006a: Craig Brown and Del Raymond (Weyerhaeuser), Theo Fallavollita, J.A., Avedesian, M. and Mujumdar, A., 1987, Kraft
Black Liquor recovery in a fludized bed. Part I: a review.
Fleisch and Mike Gradassi (BP), Paul Grabowski (USDOE), Jen-
Canadian Journal of Chemical Engineering, 65(Oct.): 812–817.
nifer Holmgren (UOP), Gord Homer (Air Liquide), Tom Johnson Fox, J.M. and Tam, S.S., 1995, Correlation of slurry reactor
(Southern Co.), Steve Kelley (North Caroline State University), Fischer–Tropsch yield data. Topics in Catalysis, 2: 285–300.
Mike Pacheco (NREL), Lori Perrine (AF&PA), Ron Reinsfelder Graaf, G.H. and Beenackers, A.A.C.M., 1996, Comparison of
(Shell), and David Turpin (Mead Westvaco). two-phase and three-phase methanol synthesis processes.
For primary financial support, we thank the U.S. Depart- Chemical Engineering and Processing, 35: 413–427.
Graaf, G.H., Winkelman, J.G.M. and Stamhuis, E.J., 1988, Kinetics
ment of Energy Biomass Program and the American Forest
of three-phase methanol synthesis. Chemical Engineering
and Paper Association. Additionally, support is gratefully
Science, 43(8): 2161–2168.
acknowledged from the Princeton University Carbon Mitiga- Gunturu, A.K., Kugler, E.L., Cropley, J.B. and Dadyburjor, D.B., 1998,
tion Initiative and the William and Flora Hewlett Foundation. A kinetic model for the synthesis of high-molecular-weight
alcohols over a sulfided Co–K–Mo/C catalyst. Industrial &
Engineering Chemistry Research, 37: 2107–2115.
References
Jameel, H. and Renard, J., 2003, Pulping options with black liquor
gasification, In Proceedings of the Colloquium on Black Liquor
A. Aden, M. Ruth, K. Ibsen, J. Jechura, K. Neeves, J. Sheehan, B.
Combustion and Gasification (L. Baxter and K. Whitty, chairmen)
Wallace (NREL), L. Montague, A. Slayton and J. Lukas (Harris
Park City, UT, 13–16 May,
Group, Seattle, WA), 2002, Lignocellulosic biomass to ethanol
Kosowski, G.M., Onischak, M. and Babu, S.P., 1984, Development
process design and economics utilizing co-current dilute acid
of biomass gasification to produce substitute fuels, In
hydrolysis and enzymatic hydrolysis for corn stover,
Proceedings of the 16th Biomass Thermochemical Conversion
NREL/TP-510-32438, National Renewable Energy Lab, June.
Contractors’ Meeting, Pacific Northwest Laboratory Richland, WA, ,
A. Aden, P. Spath and B. Atherton, 2005, The potential of
pp. 39–59.
thermochemical ethanol via mixed alcohols production,
Larson, E.D. and Ren, T., 2003, Synthetic fuel production by
Milestone Completion Report, National Renewable Energy
indirect coal liquefaction. Energy for Sustainable
Laboratory, Golden, CO, 31 October.
Development, VII(4): 79–102.
Backman, R. and Salmenoja, K., 1994, Equilibrium behaviour of
E.D. Larson, S. Consonni and R.E. Katofsky, 2003, A cost–benefit
sodium, sulfur and chlorine in pressurized black liquor
assessment of biomass gasification power generation in the
gasification with addition of titanium oxide. Paper and
pulp and paper industry, final report, Princeton
Timber, 76(5): 320–325.
Environmental Institute, Princeton, NJ, 8 October
Blackadder, W.H., Lundberg, H., Rensfelt, E. and Waldheim, L.,
(downloadable from http://www.princeton.edu/∼energy).
1994, Heat and power production via gasification in the range
Larson, E.D., Jin, H. and Celik, F.E., 2005, Gasification-based fuels
5–50 MWe, in Advances in Biomass Thermochemical
and electricity production from biomass, without and with
Conversion, Bridgwater, A.V. (ed). (Balckie Academic &
carbon capture and storage, In Princeton Environmental Institute
Professional Press, London)
Princeton University, October,
Brammer, J.G. and Bridgwater, A.V., 1999, Drying technologies for
Larson, E.D., Consonni, S., Katofsky, R.E., Iisa, K. and Frederick,
an integrated gasification bio-energy plant. Renewable and
J.W., 2006a, A cost–benefit assessment of gasification-based
Sustainable Energy Reviews, 3: 243–289.
biorefining in the kraft pulp and paper industry, final report
Celik, F.E., Larson, E.D. and Williams, R.H., 2004, Transportation
(in four volumes), 21 December (downloadable from
fuels from coal with low CO2 emissions, In Paper presented at
http://www.princeton.edu/∼energy).
the 7th International Conference on Greenhouse Gas Control
Larson, E.D., Williams, R.H. and Jin, H., 2006, Fuels and electricity
Technologies Vancouver, BC, 5–9 September,
from biomass with CO2 capture and storage, In 8th
S. Consonni, 1992, Performance prediction of gas/steam cycles for
International Conference on Greenhouse Gas Control Technologies,
power generation, Ph.D. Thesis, 1983-T, Mechanical and
Trondheim Norway, June,
Aerospace Engineering Dept., Princeton University, Princeton,
Lau, F.S., Carty, R.H., Onischak, M. and Bain, R.L., 1993,
NJ, USA.
Development of the IGT Renugas Process, Conference on
Consonni, S. and Larson, E.D., 1996,
Strategic Benefits of Biomass and Waste Fuels, In Electric Power
Biomass-gasifier/aeroderivative gas turbine combined cycle:
Research Institute Washington, DC, 30 March–1 April,
Part A—Techonologies and performance modeling, and Part
F.S. Lau, D.A. Bowen, R. Dihu, S. Doong, E.E. Hughes, R. Remick, R.
B—Performance calculations and economic assessment.
Slimane, S.Q. Turn and R. Zabransky, 2003. Techno-economic
Journal of Engineering for Gas Turbines and Power, 118:
analysis of hydrogen production by gasification of biomass,
507–525.
final technical report for the period 15 September 2001–14
Consonni, S., Larson, E.D., Kreutz, T.G. and Berglin, N., 1998, Black
September 2002, contract DE-FC36-01GO11089 for US Dept. of
liquor gasifier/gas turbine cogeneration. Journal of
Energy, Gas Technology Inst., Des Plaines, IL, June,145 pp.
Engineering for Gas Turbines and Power, 120: 442–449.
 chemical engineering research and design 8 7 ( 2 0 0 9 ) 1293–1317 1317

Lindblom, M. and Landälv, I., 2007, Chemrec’s atmospheric & Mckeough, P. and Kurkela, E., 2006, Production of transportation
pressurized BLG (Black Liquor Gasification) fuels from biomass at pulp and paper mills—the Finnish
technology—Status and future plans, In Proceedings of the approach, In Presented at the 7th International Colloquium on Black
International Chemical Recovery Conference 2007 Quebec City Liquor Combustion and Gasification Jyvaskyla, Finland, 2 August,
(Canada), 29 May–June 1, 2007, (Part 1. Printed by Pulp and K.L. Ng, D. Chadwick and B.A. Toseland, 1999, Kinetics and
Paper Technical Association of Canada (PAPTAC), Montreal, modeling of dimethyl ether synthesis from synthesis gas,
Canada), pp. 51–57. Chemical Engineering Science, 54, pp. 3587–3592. Contract
Lindstrom, M., Kirkman, A., Jameel, H., Cheng, J., Huggins, C. and DE-AC26-99FT40675 for US DOE by Research Triangle Institute,
Bray, B., 2002, Economics of integrating black liquor Research Triangle.
gasification with pulping: Part I, Effect of sulfur profiling, In R.D. Perlack, L.L. Wright, A. Turhollow, R.L. Graham, B. Stokes and
Proceedings of the Pulping and Engineering Conference. (Tappi D.C. Erback, 2005, Biomass as feedstocks for a bioenergy and
Press, Atlanta) bioproducts industry: The technical feasibility of a billion-ton
Lox, E.S. and Froment, G.F., 1993, Kinetics of the Fischer–Tropsch annual supply, ORNL/TM-2005/66 for USDA and USDOE,
reaction on a precipitated promoted iron catalyst. 1. April.
Experimental procedure and results and 2. Kinetic modeling. Salmenoja, K., Janka, K., Kymalainen, M. and Hyoty, P., 1993,
Industrial & Engineering Chemistry Research, 32(1): Development of black liquor gasification, In Proc. Tappi 1993
61–82. Engineering Conference (Atlanta) , pp. 969–975.
Macchi, E., Consonni, S., Lozza, G. and Chiesa, P., 1995, An Simell, P., Kurkela, E., Stahlberg, P. and Hepola, J., 1996, Catalytic
assessment of the thermodynamic performance of mixed hot gas cleaning of gasification gas. Catalysis Today, 27(1–2):
gas–steam cycles. Part A: Intercooled and steam-injected 55–62.
cycles. Journal of Engineering for Gas Turbines and Power, 117: Strom, E., Liinanki, L., Sjostrom, K., Rensfelt, E., Waldheim, L. and
489–498. Blackadder, W., 1984, in Egneus, H. and Ellegard, A., Ellegard,
M. Marklund, R. Gebart and R. Tegman, 2006, Scale-up method A. (eds) (Elsevier Applied Science Publishers, London), pp.
for high temperature black liquor gasification in pressurized 57–64.
entrained flow gasifiers, Presented at the 7th International TSS Consultants, 2005, Gridley ethanol demonstration project
Colloquium on Black Liquor Combustion and Gasification, utilizing biomass gasification technology: Pilot plant gasifier
Jyvaskyla, Finland, 2 August. and syngas conversion testing, August 2002–June 2004,
A. Mcaloon, F. Taylor, W. Yee (Agricultural Research Service, NREL/SR-510-37581, National Renewable Energy Laboratory,
Eastern Regional Research Centre, US Dept Agriculture), K. Golden, CO, February.
Ibsen and R. Wooley (Biotechnology Center for Fuels and Whitty, K. and Nilsson, A., 2001, Experience from a high
Chemicals, NREL), 2000, Determining the cost of producing temperature, pressurized black liquor gasification pilot plant,
ethanol from corn starch and lignocellulosic feedstocks, In Proceedings of the Tappi Chemical International Recovery
NREL/TP-580-28893, National Renewable Energy Lab, October. Conference (pp. 655–662). (Tappi Press, Atlanta)