Softwood forest thinnings as a biomass s

The co-located case refers to a specific site with an existing biomass power facility near Martell, California.. A two-stage dilute acid hydrolysis process is used for the production of

Softwood Forest Thinnings as a Biomass Source for Ethanol

Production: A Feasibility Study for California

Kiran L Kadam,*,†Robert J Wooley,†Andy Aden,†Quang A Nguyen,†

Mark A Yancey,†and Francis M Ferraro‡

National Renewable Energy Laboratory, 1617 Cole Boulevard, Golden, Colorado 80401-3393, and Merrick and

Co., 2450 South Peoria Street, Aurora, Colorado 80014

A plan has been put forth to strategically thin northern California forests to reduce

fire danger and improve forest health The resulting biomass residue, instead of being

open burned, can be converted into ethanol that can be used as a fuel oxygenate or an

octane enhancer Economic potential for a biomass-to-ethanol facility using this

softwood biomass was evaluated for two cases: stand-alone and located The

co-located case refers to a specific site with an existing biomass power facility near Martell,

California A two-stage dilute acid hydrolysis process is used for the production of

ethanol from softwoods, and the residual lignin is used to generate steam and

electricity For a plant processing 800 dry tonnes per day of feedstock, the co-located

case is an economically attractive concept Total estimated capital investment is

approximately $70 million for the co-located plant, and the resulting internal rate of

return (IRR) is about 24% using 25% equity financing A sensitivity analysis showed

that ethanol selling price and fixed capital investment have a substantial effect on

the IRR It can be concluded that such a biomass-to-ethanol plant seems to be an

appealing proposition for California, if ethanol replaces methyl tert-butyl ether, which

is slated for a phaseout

Introduction

California is faced with several critical issues related

to how its forest resources are used and managed In

particular, because of suppression of forest fires, mainly

during the last century, growing quantities of dead/

diseased trees, underbrush, and small-diameter green

trees have accumulated in the forests, creating a severe

fuel loading problem that threatens human life and

property Resulting fires are so intense that they can

literally turn forests into sterile deserts Air quality

impacts on ecological systems and human health are also

enormous as these fires emit large quantities of smoke

and particulate matter, especially PM10 (particulate

matter <10 µm) An estimated 600,000 tons of air

pollutants are emitted annually from California wildfires

(1).

Furthermore, this large quantity of underlying biomass

material affects the vitality of large trees, impairing

forest health and creating ecological imbalances To

address these issues, several stakeholder groups have put

forth plans to strategically thin the forests to reduce fire

danger, improve forest health, and restore ecosystem

balance One such stakeholder is the Quincy Library

Group (QLG), a forum for local environmental

organiza-tions, county officials, and timber industry groups QLG’s

objective is to facilitate forest thinning in the vicinity of

Quincy, California, i.e., the Plumas, Lassen, and part of

the Tahoe National Forests However, one of the major

obstacles to thinning large numbers of acres each year

is cost Moreover, a key question is the fate of the biomass

once it is removed from the forests The choices are to continue to dispose of forest residue, e.g., by open burning, to reduce the fuel loading and the risks to the ecological systems, or to find economic markets for the biomass that will cover the costs of collecting, processing, and transporting this potential resource to an end user With the state government’s concern about burning as a management option, one avenue is to convert the biomass into fuel ethanol and electricity Utilizing the forest thinnings as feedstock for ethanol and electricity produc-tion may be a way to offset all or a significant porproduc-tion of the thinning costs

The U.S Clean Air Act Amendments of 1990 mandated the sale of oxygenated fuels in ambient air quality nonattainment areas (areas in which air quality does not meet federal air quality standards for carbon monoxide, ozone, and particulate matter) As a consequence, 30%

of the gasoline consumed nationally has to meet federal reformulated gasoline (RFG) requirements, so there has been a sharp increase in the demand for fuel oxygenates There are three nonattainment areas (or air basins) in California: Sacramento, South Coast (Los Angeles and surrounding areas), and San Diego These regions col-lectively account for approximately 70% and 10% of the gasoline sold in California and the entire nation,

respec-tively (2).

The two main oxygenates sold in the United States are

methyl tert-butyl ether (MTBE) and ethanol MTBE is

currently being used in California RFG (also called California clean-burning gasoline or CBG), but a recent

Gubernatorial Executive Order (3) calls for its removal

from California RFG by the end of 2002 because of the high risk of contamination of surface and groundwater

† National Renewable Energy Laboratory.

‡ Merrick and Co.

947

Biotechnol Prog 2000, 16, 947−957

10.1021/bp000127s CCC: $19.00 © 2000 American Chemical Society and American Institute of Chemical Engineers

from faulty storage systems A recent study by the

California Energy Commission states that the existence

of a viable alternative to MTBE is fundamental to the

phaseout dictated by the Executive Order and perceives

ethanol as an alternative to MTBE (4), although other

choices are possible Even if the oxygenate requirement

for California RFG is rescinded, ethanol is still needed

to replace lost volume and octane when MTBE is phased

out A least cost scenario is to use 1% oxygen or 3.8%

ethanol for Phase III RFG (4).

The Energy Information Administration predicts that

U.S ethanol consumption would average 3.2% growth

annually over the next two decades, compared to 1.3%

annual growth for petroleum consumption (5) Although

corn-derived ethanol will play a major role in supplying

the California demand, there is a rolesalbeit smallers

for biomass-derived ethanol, especially considering the

externalities, e.g., the need to reduce fuel loading in the

forests and a ban on open-field burning of rice straw,

which would render this option attractive In this study,

we show that a softwood biomass-to-ethanol facility,

co-located with an existing biomass power plant, can be an

economically feasible proposition, with ethanol replacing

MTBE as a fuel oxygenate

Technology Options

Most of the biomass-to-ethanol studies have focused

on the conversion of hardwoods and herbaceous crops to

ethanol One of the reasons for this is that the hardwood

and herbaceous biomass species are in general easier to

pretreat than softwoods Softwoods contain more lignin

than hardwoods and lack the long, continuous vessels

found in hardwoods, which permit greater penetration

of heat, chemicals, and enzymes into the hardwood

matrix This contrast in structure and the higher lignin

content of softwoods explains the differences in

suscep-tibility to pretreatment techniques that exist between

hardwood and softwood species Some of the relevant

work on softwood pretreatment and hydrolysis is

re-viewed here

Ramos et al (6) compared the enzymatic hydrolysis of

eucalyptus (Eucalyptus viminalis), aspen, and spruce

wood chips All of the chips were subjected to SO2

impregnation followed by steam explosion They found

that the cellulosic fractions (peroxide-treated,

water-insoluble, and alkali-insoluble) from eucalyptus exhibited

a susceptibility to enzymatic hydrolysis greater than that

of the equivalent fractions from aspen or spruce,

par-ticularly at high substrate concentrations (10% w/v)

Olsson et al (7, 8) investigated fermentatability of

softwood-derived spent sulfite liquor (SSL) and

hardwood-derived enzymatic hydrolyzate (EH) The main sugars

in EH were glucose and xylose, whereas in SSL the

sugars were in the form of mannose (43%), xylose (27%),

galactose (14%), and glucose (11%) Another difference

between the two substrates was the pretreatment used:

steam pretreatment combined with enzymatic hydrolysis

for EH versus sulfite pulping for SSL SSL contained 5

g/L acetic acid versus EH, which had 10 g/L acetic acid

The fermentatability of the SSL was good; however,

ethanol yield for SSL was lower than that for EH (0.3 vs

0.45 g ethanol/g sugar) For a given organism, the

fermentation rate was also lower, by a factor of about

two, than that in EH

Clark and co-workers (9, 10) studied pine as a feedstock

for ethanol production They specifically used Pinus

radiata, which is a predominant commercial forest

spe-cies in New Zealand, and plenty of pine-derived forest

residue and mill residues such as shavings, sawdust and

chips are available Clark and Mackie (9) took logs from

an 18-year old pine tree and chipped them in whole-tree chipper after debarking They pretreated the resultant chips by means of SO2impregnation (2.55 wt %) followed

by steam explosion (215 °C, 3 min) The pretreated substrate showed good enzymatic digestibility, with a

total sugar yield of 57 g/100 g oven-dry wood Pinus radiata is also Australia’s principal softwood timber species, while Eucalyptus regnans, a hardwood, is

con-fined to the southeastern states Saw-milling operations

of both wood species generate large quantities of waste

products that can be converted to ethanol Dekker (11) studied P radiata and E regnans sawdust pretreatment using steam explosion at 200 °C P radiata pulps were poorly hydrolyzed by Trichoderma cellulases compared

to E regnans pulps Impregnation of P radiata with

4.44% (w/w) SO2 followed by autohydrolysis explosion significantly improved its enzymatic digestibility

Von Sivers and Zacchi (12) evaluated three technology

options for a pine feedstock: concentrated acid process (CHAP), SO2/dilute acid process (CASH), and enzymatic hydrolysis process This is a unique study in that it evaluates these options for the conversion of pine to ethanol using a uniform platform Conversion of Swedish kronors (SEK) to American currency is as follows: 8 SEK ) $1 U.S., for the 1992/93 time frame of the study The major conclusion, which can be drawn from this study,

is that none of the three processes can be eliminated as less economical than the other two The ethanol produc-tion price varies between 4.0 and 4.3 SEK/L

($0.50-$0.54/L), and the difference is definitely within the margin of uncertainty in the various assumptions un-derlying the technical and economic calculations

In the CHAP process, the energy consumption figures are high and have a large impact on the process economy Increased sugar yield, energy integration, or larger plant capacity are improvements that reduce the ethanol production cost for all three processes The sensitivity analysis conducted by these investigators showed that reduced capital cost considerably decreased the ethanol production price They suggested that integration of the ethanol plant with existing facilities, such as a pulp and paper plant or a heating plant, could further reduce the ethanol production cost

Another approach to softwoods utilization is the acid-catalyzed Organosolv saccharification or the ACOS

proc-ess (13, 14) Aqueous acetone (acetone/water 9:1) in the

presence of a mineral acid (0.04 N HCl) at 200 °C can be used to pretreat hardwoods and softwoods In a two-stage process, selective hemicellulose removal and delignifica-tion in the first stage followed by saccharificadelignifica-tion in the second stage are achieved This process has been applied

to aspen sawdust and Douglas fir sawdust at the bench scale A total sugar recovery of 95% under continuous percolation mode is possible using optimal process

condi-tions for each stage (14).

Based on the above discussion, the following process configurations can be identified as potential options for converting softwoods to ethanol:

• SO2steam explosion, followed by enzyme hydrolysis and fermentation with yeast

• SO2steam explosion, followed by dilute acid (sulfuric

or nitric acid) hydrolysis and fermentation with yeast

• Two-stage dilute acid hydrolysis, followed by fermen-tation with yeast

• Concentrated acid hydrolysis, followed by fermenta-tion with yeast

• Organosolv pretreatment followed by fermentation

with yeast

Criteria for selecting a technology for the study were

as follows: (1) the technology must be at least at the

pilot-scale stage of development or, alternatively, extensively

studied by multiple research organizations with positive

results for softwoods at the bench-scale and (2) pilot-scale

or commercial pretreatment equipment must be

avail-able On the basis of the comparison provided by the

Swedish researchers (12), the three processes (CHAP,

CASH, and enzymatic hydrolysis) have a similar

eco-nomic potential However, the concentrated acid process

(CHAP) results in high capital and energy costs and

needs an elaborate acid recovery system The enzymatic

hydrolysis process suffers from the uncertainty associated

with cellulase production costs The authors had used

laboratory-scale data for their analysis, and although

cellulase production on a large scale (1000 m3fermenters)

is possible, achieving high enough productivities at such

scales will need more research Although high sugar

yields are possible with the ACOS process under

continu-ous percolation conditions, pilot-scale data for such

conditions are not available Hence, a CASH-type process,

i.e., two-stage dilute acid hydrolysis followed by

fermen-tation with yeast, was chosen as a suitable technology

in this study for a near-term softwood-to-ethanol plant

This approach has been described in some detail by

Harris et al (15).

Study Objective and Scope

The objective of this work was to conduct a feasibility

analysis of ethanol production from forest thinnings

(softwoods) in northern California Specifically, economic

potential for a softwood biomass-to-ethanol conceptual

facility was evaluated for two cases, stand-alone and

co-located The stand-alone plant is a “greenfield” site with

no existing infrastructure The co-located case refers to

a specific plant site near Martell, California, which

represents a typical desirable co-location site Available

infrastructure at this site includes steam and electricity

from an 18-MW biomass power plant; wood chip

receiv-ing, storage, and processing equipment; process and

potable water; sanitary sewer facilities, and office

build-ings and maintenance facilities The biomass power plant

can also utilize the ethanol plant’s lignin residue as boiler

fuel

The size of the ethanol facility is restricted by the

amount of forest thinnings and timber harvest residue

available within a reasonable haul distance from the site

On the basis of a study by QLG (16), the plant is sized

for processing 800 dry tonnes per day of feedstock

A two-stage dilute acid hydrolysis process is used for

the production of ethanol from softwoods, and the

re-sidual lignin is used to generate steam and electricity

For the stand-alone plant, lignin residue is burned in a

boiler, which is part of the plant design and produces the

necessary steam and electricity for the plant, with excess

electricity sold to the local grid The costs of the boiler

and turbine generator set are included in the cost

estimate, as is the equipment operating cost The boiler

and turbine generator set are specifically sized to

accom-modate the lignin produced by the ethanol plant Capital

cost for other utilities, as well as buildings, yard

im-provements, etc., is also included in the analysis

For the co-located case, the existing boiler and power

generation facility could be adapted to burn the lignin

produced by the ethanol process The softwood feedstock,

along with the regular boiler feed, is channeled through

the existing biomass handling system of the power plant The chip handling capability is sized to handle softwood thinnings and the regular boiler feed; the cost to upgrade the handling system is included in the estimate The capital and operating costs of the boiler and generator are, however, not included in the estimate The ethanol plant would purchase its steam and electricity from the adjoining power plant

Assessment of Co-Location Site

The co-location site refers to the plant at Sierra Pacific Industry’s lumber mill (since closed) in Martell, Califor-nia, and is owned and operated by Wheelabrator Envi-ronmental Systems Inc The generating station of the power plant of interest produces 18 MW of electricity while burning variable amounts of agricultural wastes, municipal wastes, and forest trimmings with average moisture content of 40-45% The plant processes about 130,000 dry tonnes per year (∼140,000 dry short tons per year) of biomass The station pays $15-$25 per dry tonne for biomass delivered to the site It is assumed that the generating capacity remains the same, i.e., the power plant will accept lignin residue and buy the requisite amount of other feedstock to generate 18 MW Another assumption is that lignin from the ethanol plant can be processed in the existing boiler with minor modifications The ethanol plant would be located in close proximity

to the existing 18-MW biomass power plant The co-located case assumes that a single owner or a similar business entity allowing shared operation and mainte-nance would own and operate the ethanol facility and the biomass power plant A co-located facility would allow the steam and electricity costs to be shared with the power plant because the “sale” of the utilities would be

an internal cost between the Martell facility and the wholly/majority-owned biomass-to-ethanol facility A ma-jor advantage of co-location is the use of existing facilities and services, which provides the following benefits:

• Lower capital expenditure for the ethanol plant The Martell site would allow use of the chip handling equip-ment and surge piles, boiler equipequip-ment, condensate equipment, and boiler-feed treatment system

• Decrease in operator, security, maintenance, man-agement, and administration personnel The combined sites can use some of the same work force; the Martell power plant is assumed to provide a majority of these personnel

• Shared chemical and utility costs The co-located facility would allow sharing of chemical and utility (excluding steam and electricity) costs between the two plants, thereby allowing more aggressive negotiation with suppliers

• Other shared buildings and facilities

Utilities and infrastructure available at the site that could potentially benefit an adjacent ethanol plant are

as follows:

• Steam The boiler generates 90,700 kg/h (200,000 lb/ h) of steam at 6300 kPa (900 psig) Extraction steam is available at 1300-1500 kPa (180-200 psig) and

700-800 kPa (85-100 psig) Currently the power generating station supplies 9,070 kg/h (20,000 lb/h) at 1300-1500 kPa (180-200 psig) to an adjacent particleboard plant, and more steam is available for export

• Electricity The generating station currently sells power to the grid at $0.05/kWh The turbogenerator will continue to produce 18 MW or more when ethanol plant steam is supplied from the turbine extraction ports

• Cooling Water The generating station utilizes a

cooling tower for heat rejection operating at about four

cycles of concentration This system could provide cooling

water to an adjacent ethanol plant The match would be

beneficial because there is a lower load on the cooling

towers when steam is exported The cooling tower is sized

for turbogenerator operation at as much as 18 MW of

power generation without steam export (full condensing)

• Demineralized Water Boiler makeup at the

generat-ing station is produced in parallel cation/anion

deminer-alizers There is some excess capacity in these units to

provide demineralized water to an adjacent plant

How-ever, additional storage would be required to allow

production and storage of demineralized water during

off-peak generating periods

• Miscellaneous Fire protection, site security, and

railcar access are available and can be shared

Process Description

For this study, the biomass feedstock for the ethanol

facility is whole tree chips from forest thinning and

timber harvest operations The species mix is assumed

to be 70% white fir and 30% ponderosa pine (by weight);

this is based on the prevalent species in the area The

composition of a 70/30 mix of white fir and ponderosa

pine is shown in Table 1 (based on unpublished analysis)

Figure 1 shows a simplified process flow diagram

illustrating the major processing steps and flow paths

in the plant (Detailed process flow diagrams were used

to develop the material and energy balance, equipment

list, and capital cost estimate for the process (see ref 17)

but are not included here.) A process description is

provided below based on published data (15) and work

conducted at the National Renewable Energy Laboratory

(NREL)

Forest thinnings and other biomass residue that have

been chipped in the forest without debranching or

debarking are received at the ethanol plant The wood

chips are conveyed from the chip pile past a magnetic

cleaner and through a screen to ensure a chip size of 2.5

cm or smaller Oversized chips are sent to a slicer and

returned to the screen A second conveyor moves the

chips to the first-stage acid impregnator Here the chips

are heated to approximately 50 °C and soaked in water

containing 0.7% sulfuric acid

From the impregnator the chips enter the first-stage

hydrolysis reactor, where the temperature is raised to

190 °C (167 psia, 1150 kPa) At the liquid-phase acid

concentration of 0.7%, a residence time of 3 min is

required to achieve efficient hydrolysis of hemicellulose

Approximately 80% of the hemicellulose and 20% of the

cellulose are hydrolyzed in the first reactor The stream

leaving the first hydrolysis reactor (hydrolyzate) contains about 30% total solids (suspended and dissolved solids) The hydrolyzate pressure is reduced in the flash tank, lowering the temperature of the hydrolyzate to about 130

°C Following a residence time of 2 h in the flash tank to hydrolyze most of the oligosaccharides to monosaccha-rides, the hydrolyzate enters the countercurrent slurry washer where the solids and liquids are separated, and the solids are washed to remove the sugars and other soluble compounds Efficient washing is critical to maxi-mize sugar recovery and to minimaxi-mize the dilution of the sugar stream Too much washwater will increase the size and cost of the fermenters and distillation system A washwater flow rate of 3-4 times the dry weight of solids being washed is used The combined liquid hydrolyzate and wash water are sent to the first-stage pH adjustment tank The solid stream (30% solids) from the washer is further dewatered by a screw press The pressate is returned to the slurry washer, and the solids, now at about 45% solids, are sent to the second-stage acid impregnator

The acid-impregnated material then enters the second-stage hydrolyzer In the second hydrolysis reactor, the temperature is raised to 220 °C (322 psia, 2220 kPa), and the prevailing acid concentration in the liquid phase is 1.6% A residence time of 3 min is required at these conditions to achieve the necessary reactions About 60%

of the remaining cellulose is hydrolyzed, with about 70% going to glucose and 30% degraded to hydroxymethyl furfural (HMF) and other degradation products Follow-ing the second hydrolysis, the hydrolyzate slurry is flashed butsunlike in the first hydrolysis stepsnot washed

Lime is added to the first-stage pH adjustment tank

to neutralize the sulfuric acid and raise the pH of the hydrolyzate to about 5.5 This causes most of the calcium sulfate (gypsum) to precipitate out of the solution The calcium sulfate is removed by a filter and sent to disposal The relatively solids-free sugar stream is cooled to 35 °C and sent to the fermenters Ammonium sulfate and nutrients such as corn steep liquor are added to the two fementers

The hydrolyzate slurry from the second-stage pH adjustment tank is cooled to 35 °C before it enters the fermentation section The fermentation section has a total residence time of 32 h, with an overall ethanol yield

of 90% of the six-carbon sugars and 75% of the xylose entering the fermenters A recombinant xylose-ferment-ing yeast can be used in this case

The fermentation broth, also known as “beer”, is sent

to the distillation/molecular sieve dehydration systems where the ethanol is separated from the water The ethanol distillation/dehydration technology is well de-veloped and will not be described here (see ref 18) The 99.9% ethanol is denatured with 5% gasoline and sent

to storage The stillage from the distillation column is processed using centrifuges and evaporators Lignin separation is accomplished in solid-bowl decanting cen-trifuges The lignin stream, containing about 45% solids,

is sent to the boiler/power generation section (18) The

liquid stream is sent to a wastewater treatment system, and reclaimed water is recycled to the process (see ref

19 for a description of the wastewater treatment system)

Model Development

The softwood biomass-to-ethanol process was modeled

on the basis of the above description and the performance parameters delineated in Tables 2 and 3 The ASPEN

Table 1 Feedstock Compositiona

% dry weight

aHemicellulosic glucan is assumed to be 10% of total glucan.

Acetate is calculated by difference to normalize result to 100%.

Plus process simulation software was used to generate

mass and energy balances for the entire process The

overall model is thermodynamically rigorous and uses

built-in physical properties as well as properties

devel-oped at NREL The individual unit models are

thermo-dynamically consistent and can either be rigorous, e.g.,

the simulation of distillation unit, or simple Simple unit operations are typically modeled using key performance parameters based on typical industrial practice or ex-perimental data, if available For example, solid-liquid separations are modeled using an estimated fixed solids removal and liquid retention in the solids stream Process flow rates for certain streams generated from the mass and energy balances were then used to perform detailed design for each piece of equipment The equip-ment costs were obtained from vendor quotations when possible, especially for uncommon equipment such as the hydrolysis reactors The ICARUS process evaluator software was used to cost more common types of equip-ment such as pumps, agitators, and conveyors For quotations without an installed cost and for common types of equipment, Lange-type factors were applied to the equipment costs to account for the added cost of installation These factors were based on engineering

experience and standard texts such as Garrett (20) and Peters and Timmerhaus (21).

Once the installed equipment costs were determined, various indirect construction costs and other overhead factors were applied to determine a total plant invest-ment cost The indirect construction costs were estimated

as a percentage of the capital equipment costs on the basis of experience with similar types and sizes of projects The indirect costs are defined as follows:

• Prorated These include fringe benefits, insurance, bonding, and overhead burdens The prorated costs are adjusted to include the prorated costs of skid mounting equipment in a nearby industrial center

• Process Development These costs include final development of the specialized parts of the process that may require additional study or research

• Field Expenses Normally these include consumables, equipment rental, field services, temporary facilities, and

Figure 1 A simplified process flow diagram illustrating the two-stage, dilute sulfuric acid process for ethanol production.

Table 2 Hydrolysis Yields for Softwood

Biomass-to-Ethanol Process

first stage

hydrolysis yield

% (based on

original

feedstock)

second stage hydrolysis yield

% (based on original feedstock)

overall hydrolysis yield

% (based on original feedstock)

Table 3 Other Process Parameters for Softwood

Biomass-to-Ethanol Process

Operating Conditions first stage acid hydrolysis 190 °C, 0.7% acid, 3 min

second stage acid hydrolysis 220 °C, 1.6% acid, 3 min

Countercurrent Washing

insoluble solids lost to sugar stream, % 0.0

insoluble solids recovered in stream

to second stage impregnator, %

98.0

Overall Fermentation Yield, %

supervision The co-located site will have existing

facili-ties that would be used during construction The

co-located site also has services (water, sewer, electric,

phones, roadways, etc.) that would be used during

construction The co-located field expense was reduced

because certain facilities and services are available

• Home Office Construction Fee This includes detailed

engineering of the plant, purchasing of the equipment

and bulks, and field construction support

• Contingency This is an allowance for expected but

undefined costs

• Start-up, Permits, and Fees These include plant

commissioning, construction permits, and fees

Indirect costs that are not included in the capital costs

include

• Owner supervisory personnel for engineering,

con-struction, and start-up

• Engineering/construction overtime pay

• Owner/engineering scope changes

Other assumptions used include the following:

• There are no land acquisition costs

• There are no off-site costs (e.g., public road

improve-ments, extension of power, water, telephone services)

• There is a source of qualified construction personnel

within daily driving distance of the site

• There are adequate roads, railroads, or ship docks to

allow equipment deliveries

• The costs of obtaining air and water permits are not

incurred by the project and are not included

• Soils are adequate for conventional foundation design

• The escalation factors used are as follows:

)Commodities are escalated at 1% per year These

include ethanol and electricity sales as well as raw

material costs

)Cost items are escalated at 3% per year These include

utilities purchase costs and operations and maintenance

costs

The variable operating costs, which include raw

ma-terials, waste handling charges, and byproduct credits,

were determined using the ASPEN mass and energy

balance model Fixed operating costs (salaries, etc.) were

obtained from standard labor rates for the study area as

well as from information about corn-to-ethanol plants of

similar size (18).

The total plant investment cost and the plant operating

costs were used in a discounted cash flow analysis to

determine the internal rate of return (IRR) Depreciation

was determined via the IRS MACRS (U.S Internal

Revenue Service’s Modified Accelerated Capital Recovery

System) method, as recommended by Short et al (22).

Iterative calculations on the IRR were performed until

the net present value of the project was zero for a given

ethanol selling price, e.g., $0.32/L ($1.20 per gallon),

which is a base-case selling price This cash flow analysis

was the primary economic comparison tool employed

Results and Discussion

Economic basis and general plant data given in Table

4 and the above assumptions were used in assessing the

economic performance of the two cases The IRR on an

after-tax basis was used as a measure of project

profit-ability IRRs for the stand-alone case and the co-located

case were 14% and 24%, respectively (Table 5) For this

reason, a detailed economic analysis is provided for the

co-located case only (Table 6)

The lower IRR for the stand-alone case is mainly due

to the higher capital investment needed Capital cost is

heavily influenced by the availability of existing

infra-structure at each site The undeveloped or greenfield site requires the installation of a boiler to provide steam to the ethanol process, as well as buildings and other infrastructure, which adds significantly to the total capital cost, making this site less attractive Conversely,

a biomass power plant and other infrastructure are available at the Martell site, and the capital for power generation is a “sunk” cost for the co-located case The capital investment for the co-located case ($70 million

or $3.5 per annual gallon of ethanol) is therefore 30% lower than that for the stand-alone case ($100 million or

$5.0 per annual gallon of ethanol)

On the basis of these numbers, the co-located project looks attractive from an economic viewpoint Further-more, there are environmental benefits in using Califor-nia biomass for ethanol production versus open burning

it, such as lower criteria air pollutants on a life-cycle basis

(23), which are extremely important in California because

of its strict air quality standards However, such exter-nalities are beyond the scope of current discussion

Sensitivity Analysis A sensitivity analysis was

performed for the co-located case for ethanol selling price,

Table 4 General Plant Data and Assumptions

General Plant Data plant basis: feedstock processed, dry metric t/d 800

plant capacity, million liters (gallons) of formulated product per year

76 (20) Economic Assumptions

owner equity financing, % of fixed capital investment

25

Feedstock purchase cost, $/tonne (dry basis) 22

Plant Personnel Data

nonskilled laborer’s hourly rate, $ 11.50

Product and Coproduct Data ethanol selling price, $/L ($/gal) 0.32 (1.20) electricity selling price/cost, $/kWh

(stand-alone case/co-located case)

0.05 steam 300 psig cost, $/tonne (co-located case) 3.86 lignin credit, $/dry tonne (co-located case) 23.5b

aThe loan rate is lower than normal because of a subsidy available from California state.bTakes into account higher energy and water contents of the residue compared to the base feedstock.

Table 5 Economic Performance of the Stand-Alone and Co-Located Cases

total capital investment, million $

total capital investment,

$/annual gallon

of ethanol IRR, %

feedstock cost, and fixed capital investment, which were

considered to be key parameters The effect of electricity

selling price on project profitability was also evaluated

for the stand-alone case For the co-located case, the

higher electricity selling price reduces the IRR because

electricity is a cost center This is assuming that the

credit for lignin is not linked to the electricity-selling

price Hence, evaluating this parameter for the

stand-alone case is more illuminating

Changing the ethanol selling price in either direction

has a great impact on the IRR (Figure 2) Higher ethanol

prices significantly increase the revenue and the

eco-nomic performance and vice versa A $0.01/L change in

ethanol selling price changes the IRR by about three

percentage points Hence, properly estimating the actual

selling price the regional/local market will bear, including

all tax incentives and discounts, is essential The $0.32/L

($1.20/gal) price takes into account the federal excise tax

credit to the blender ($0.14/L or $0.54/gal), as well as the

federal small producer tax credit ($0.03/L or $0.10/gal)

for the first 57 million L (15 million gal) of ethanol

The feedstock cost has a moderate impact on the IRR (Figure 3) A $2/tonne change in feedstock cost changes

the IRR by about one percentage point QLG (16) has

estimated a cost of $20/ton for collecting, processing, and transporting the forest thinnings to an end user Al-though the feedstock cost has a relatively moderate influence on overall economics, the QLG estimate needs

to be validated

The total fixed capital investment as estimated has some uncertainty attached Hence, studying the effect of under- and overestimating the capital investment is important A 10% change on either direction changes the IRR by four to five percentage points (Figure 4) Keeping the capital expenditure within the forecast budget is therefore essential, and an Engineering, Procurement and Construction firm would need to confirm capital estimates to secure a project financing performance guarantee It should be noted that the capital investment responds nonlinearly as opposed to the other parameters Electricity selling price for the stand-alone case has a positive impact because this is a coproduct of the plant This relationship is shown in Figure 5 The

electricity-Table 6 Economic Evaluation of Co-Located Case: Capital Investment and Cash Flow Analysis

Capital Costs

Cash Flow Analysis

(i) balance on borrowed capital 52.80

Figure 2 Effect of ethanol selling price on IRR (co-located

case)

Figure 3 Effect of feedstock cost on IRR (co-located case).

selling price would probably not be higher than $0.05/

kWh Hence, only that portion of the response curve with

prices e$0.05/kWh is applicable in this case

Further-more, the stand-alone case is not as practical an option

as the co-located case, and this response curve is shown

for illustrative purposes

Power Plant Capacity The existing power plant

processes about 130,000 dry tonnes per year (140,000 dry

tons per year) of biomass and produces 18 MW of power

This analysis assumed that the 18-MW capacity does not

change when an adjacent ethanol plant is built The

ethanol plant produces about 80,000 dry tonnes per year

of ligneous residue If all the lignin is accepted as fuel,

about 50,000 dry tonnes per year of other additional

biomass can be used Accounting for higher heating value

of the ligneous residue, more than 18 MW of power can

be produced for the same throughput of biomass

(assum-ing that the boiler and power generator can produce over

18 MW without major modifications) However, the

electricity selling price and other market and regulatory

conditions, such as peak versus off-peak prices and

capacity payments, greatly influence the economics and

practicability of such increased capacity, and this

discus-sion is provided for illustrative purposes

Process Status and Uncertainties The two-stage

dilute sulfuric acid process presented here, although not

mature, is not overly complicated either There are,

however, several assumptions made in the process design

and performance that need to be validated before a

commercial facility is built To improve the dependability

and consistency of the process, additional design

inves-tigations and testing need to be conducted to resolve the

following issues and to secure process performance guarantees The cost estimates already include funds for final development of the specialized parts of the process that would require additional scrutiny or research

Hydrolysis Sugar Yields The hydrolysis sugar yields

used in the process design presented here are based on

published data (15) and NREL’s ongoing work on soft-wood hydrolysis (24, 25) Sugar yields assumed for this

study are achieved by NREL in the laboratory, but work

at a larger scale is needed Hydrolysis sugar yield is an important parameter that directly affects the ethanol yield and should be confirmed by pilot-scale tests

Material of Construction for Hydrolysis Reactors.

Use of dilute sulfuric acid at elevated temperatures requires special alloys or metals such as hastelloy, titanium, or zirconium or an acid brick reactor lining

As a material of construction, 316 stainless steel is not adequate for the primary and secondary hydrolysis reactors’ wetted parts Titanium and zirconium are too expensive to consider until other options have been exhausted Acid-brick-lined reactors with hastelloy-clad nozzles and openings have been assumed for costing purposes The suitability of this design needs to be confirmed through corrosion tests Using sulfur dioxide

or an acid that is less corrosive to stainless steel, such

as nitric acid, is another option

Hydrolyzate Fermentability The hydrolyzate

pro-duced from the first- and second-stage hydrolysis reac-tions may be toxic to the fermentation yeast The design and economic analysis presented here assumes that the hydrolyzate is not toxic and does not adversely influence the assumed fermentation time (32 h) or ethanol yield

A hexose-fermenting yeast has been adapted to inhibitors present in the hydrolyzate milieu and shown to perform well in terms of ethanol yield at hydrolyzate

concentra-tions assumed in the model (26) This actually is a

significant benefit for the process because contaminants cannot easily compete with the adapted yeast In this analysis a small fraction of fermentations are assumed

to be lost owing to contamination, and the economics would be slightly improved if contamination is shown to

be nonexistent

Ethanol Yield Like the hydrolysis sugar yield, the

fermentation ethanol yield directly affects the plant’s ethanol production capacity and therefore its profit-ability The 90% overall fermentation yield from six-carbon sugars is achieved at the bench scale using

Saccharomyces cerevisiae, but needs to be demonstrated

at the pilot scale

S cerevisiae, a commonly used ethanologen, has the

ability to ferment biomass-derived six-carbon sugars (glucose, mannose, and galactose) to ethanol but lacks the biochemical machinery to ferment the five-carbon sugars (xylose and arabinose) More ethanol could be produced from the same amount of feedstock if an organism with the ability to ferment five-carbon sugars

is utilized, as assumed in this analysis A recombinant

xylose-fermenting organism, e.g., S cerevisiae, Zymomo-nas mobiliz, E coli, or Klebsiella oxytoca, can be used to

ferment both six- and five-carbon sugars None of these recombinant organisms, however, has been used com-mercially Although xylose-fermentation proficiency is assumed, the relatively low amounts of xylose in soft-woods (about 7% versus 20-25% in hardsoft-woods) makes this a less critical capability Nevertheless, the 75% overall fermentation yield from xylose also needs confir-mation

Yeast Propagation A small percentage of the glucose

entering the fermenters is converted to yeast cell mass;

Figure 4 Effect of fixed capital investment on IRR (co-located

case)

Figure 5 Effect of electricity selling price on IRR (stand-alone

case)

2% has been assumed for this design The amount of cell

mass produced affects the ethanol yield, and the 2% value

needs to be substantiated

The design also assumes that a yeast seed train is not

needed to provide fresh yeast to the fermenters This is

a fairly safe assumption for the fermentation considered

here but needs to be confirmed by continuous bench- or

pilot-scale fermentation tests Bench-scale batch

fermen-tation shows that yeast can be grown in the fermenter

by adding a very small amount of air, which does not

negatively affect ethanol yield (26).

Neutralizing Base The use of lime to adjust the

hydrolyzate-liquor pH to 5.5 will produce calcium

pre-cipitates (mainly calcium sulfate or gypsum), which may

foul the distillation column and the heat exchanger

surfaces throughout the ethanol facility This could cause

serious operating problems and reduce the capacity of

the plant Use of ammonia for pH adjustment would

eliminate the potential fouling problem but would

in-crease the raw materials cost Similarly, using nitric acid

with lime or ammonia would eliminate the potential

fouling problem but would also increase the raw

materi-als cost for the plant The potential for fouling by calcium

sulfate in the system needs to be carefully evaluated

Facility Thermal Design The overall plant thermal

design and energy use for the process presented have not

been optimized, i.e., no heat integration was attempted

except in the wastewater treatment area, which was

integrated with the distillation column It is assumed

that a chilled water system is required to maintain the

fermenter temperature at 35 °C In the study region, it

may be possible to maintain the fermenter temperature

with the cooling water system (cooling tower) only This

could result in significant capital and operating cost

savings Thermal optimization of the facility design is

recommended when conducting follow-on engineering

work

Solid/Liquid Separation Equipment Two solid/

liquid separation steps are included in the process

design: following the first stage hydrolysis and after the

pH adjustment step (Figure 1) Separating the

hydro-lyzate liquids from the remaining solid biomass can be

problematic A rotary vacuum drum filter with

counter-current washing was selected for this study However,

the performance of this equipment in these applications

has not been established and needs to be demonstrated

at the pilot scale

Ligneous Residue The lignin/cellulose residuals

re-covered after distillation are used as boiler fuel Small

residue samples were produced and analyzed for volatile

ash and ash fusibility to assess the potential for slagging

in biomass furnaces and boilers and fireside fouling of

heat exchangers In general, these residues appear to be

of low to moderate fouling type (27) To augment these

bench-scale tests, larger quantities of representative

residue need to be produced for test burns to determine

whether this fuel would cause boiler fouling or problems

with boiler emissions or ash characteristics

Water Recycle An 80% recycle rate has been assumed

for the design presented here If the amount of recycle

water is reduced, additional makeup water will be

required However, the impact on overall costs and

economics should not be significant Nevertheless, the

potential adverse effect of byproducts buildup in the

recycle water on the fermentation time and ethanol yield

needs to be investigated

Project Deployment Deployment of the preferred

option, i.e., the co-located case, depends on several

factors The following enabling activities are needed for the development and deployment of this project:

• Validate feedstock supply plan It is necessary to ascertain that the resources and specific supply systems needed to support a sustainable ethanol plant operation

do exist or can be economically established Also, because feedstock cost is an important parameter, it is imperative

to define feedstock harvesting and delivery requirements, which can be used to establish feedstock costs that support the economic assessment of the project Eventu-ally, long-term (at least spanning the loan duration) feedstock supply agreements/contracts will need to be in place For this study, the size of the ethanol facility is based on the amount of forest thinnings and timber harvest residue available within a 40-km radius plus any mill residue that may be available The Martell site was not specifically included in the previous feedstock

as-sessment (16); it was assumed that 800 dry tonnes per

day are available at this site This is an average amount

of biomass available at the four other sites studied in the

QLG area (16) Although the assumption is reasonable

and plausible, feedstock availability and cost will need

to be verified

• Perform further development work to minimize process uncertainties The above-mentioned uncertainties will need to be addressed to render the process robust enough for large-scale operation

• Establish an owner/operator organization for the ethanol facility to carry out further project development The co-located case assumes that a single owner would own and operate the ethanol facility and the biomass power plant However, a multiparty joint ownership may

be needed to bring the project to fruition

• Analyze market and pricing issues pertaining to ethanol and set up market relationships that establish

a contract basis for ethanol sales The economic analysis

is based on an ethanol selling price of $1.20/gal, after incorporating the various tax credits and discounts associated with long-term market relationships This appears to be reasonable However, the proforma is very sensitive to ethanol selling price; hence, estimating the actual selling price the regional/local market will bear, including all tax incentives and discounts, is essential

To minimize market risk, long-term contractual relation-ships regarding the sale of ethanol must be established

Project Financing Project financing is a strategically

important issue in successful deployment Projects of this magnitude are rarely financed with 100% owner equity

It is well-known that favorable financing terms can be used to leverage the owner’s equity and dramatically improve the IRR A scenario with 25% owner equity and 75% debt financing, i.e., debt:equity ratio of 3, was used

in this analysis A loan interest rate of 5% was chosen to represent perhaps the best possible financing scenario

A 5% interest rate loan may be available through Public Interest Energy Research Program (to be administered

by the California Energy Commission), through the California Pollution Control Financing Authority, or the California Alternative Energy Financing Authority Results of the IRR calculations with 25% owner equity and 5% interest rate were discussed earlier These results, showing higher IRRs at higher debt:equity ratios, demonstrate that robust IRRs are possible if the debt: equity ratio is increased beyond 3 (data not shown) Although optimistic, a debt:equity ratio higher than 3 may be feasible

Key elements in securing financing include having sales and feedstock supply agreements in place Project financing performance guarantees are proforma driven

and need guarantees in several areas, such as

unit-operations performance and product yield, to attract

private investors Moreover, financing of potential projects

may be encumbered by new-technology risks and high

capital requirements Spreading new-technology risks

among all beneficiaries of the project, including the

public, can expedite financing Public and private

part-nerships can play a significant role in accelerating the

development of forest biomass to ethanol in northeastern

California and as such should be an integral component

of the efforts to advance the opportunities Thus, a

conglomeration of entities will probably be needed to put

together the financing Project financing participants

would include owner/operator entity, state/federal

agen-cies (mostly as guarantors and facilitators), and private

investors

Legislative and Other Issues The

Herger-Fein-stein Quincy Library Group Forest Recovery Act (28)

allows thinning of 28,300 ha (70,000 acres) per year The

amount of potential thinnings available in the Lassen,

Plumas, and Tahoe National Forests is, on average, 27

dry tonne/ha This amounts to 770,000 dry tonnes/year

of biomass or about three times that required by the

plant

However, as mentioned earlier, one of the major

obstacles to thinning large areas is cost, which can be

reduced by state and federal subsidies Monetization of

the benefits of forest thinning, such as lower fire-fighting

costs and improved air quality, is necessary to justify and

secure such subsidies Also, there is a precedence for

providing state-backed incentives to promote off-field use

of waste biomass in California For example, rice straw

burning faces severe restrictions or an eventual

prohibi-tion mandated by the Rice Straw Burning Reducprohibi-tion Act

of 1991 (29) To promote its off-field use, state tax credit

of $16.6/tonne ($15 per ton) was available to offset the

cost of transporting the straw (30) As open burning of

biomass is antithetic to California’s environmental goals,

a similar tax credit or subsidy can be justified for a

gainful use of forest thinnings As already demonstrated

in Figure 3, this can favorably affect the economics

Conclusion

This study concludes that converting forest thinnings

to ethanol appears to be economically feasible at the

Martell site, which has an existing biomass power plant

and other infrastructure The undeveloped or greenfield

site requires the installation of a boiler to provide steam

to the ethanol process, as well as buildings and other

infrastructure that escalate significantly the total capital

cost, making this site less appealing The co-located

ethanol plant, with a 76 million L/year (20 million gal/

year) ethanol production capacity, can be constructed for

a total capital investment of approximately $70 million

Incorporating depreciation results in annual project net

revenues of approximately $12 million The resulting IRR

is about 24% using 25% equity financing Hence,

replac-ing MTBE, which is the current fuel oxygenate but is

scheduled to be phased out by the end of 2002, with

biomass-derived ethanol appears to be an attractive

option for California Follow-on engineering design and

experimental studies are recommended to confirm the

design and process performance assumptions made in

this study

Acknowledgment

This work was funded by the Bioconversion Element

of the Office of Fuels Development of the U.S

Depart-ment of Energy We wish to thank Eric Selya of Whee-labrator Environmental Systems Inc., Martell, California, for his help in ascertaining the utilities and infrastruc-ture available at their site and Jim Sharpe and Dick Voiles of Merrick & Co for their help in the economic analysis

References and Notes

(1) CDF 1995 Fire Plan California Department of Forestry and

Fire Protection: Sacramento, CA, 1995

(2) CEC Timetable for the Phaseout of MTBE from California’s Gasoline Supply, Docket No 99-GEO-1; Fuel Resources Office,

Energy Information and Analysis Division, California Energy Commission: Sacramento, CA, June 1999

(3) Davis, G Executive Order D-5-99; Office of the Governor of

California: Sacramento, CA, March 1999

(4) Blackburn, W.; MacDonald, T.; McCormack, M.; Perez, P.;

Brown, S.; Deller, N Evaluation of Biomass-to-Ethanol Fuel Potential in California; California Energy Commission’s

Report to the Governor of California and the Secretary of California; Environmental Protection Agency: Sacramento,

CA, August 1999

(5) EIA Annual Energy Outlook 2000 (AEO2000), Report DOE/

EIA-0383(2000) Energy Information Administration (EIA): Washington, DC, December 17, 1999

(6) Ramos, L P.; Breuil, C.; Saddler, J N Comparison of steam pretreatment of eucalyptus, aspen, and spruce wood chips and

their enzymatic hydrolysis Appl Biochem Biotechnol 1992,

34/35, 37-47.

(7) Olsson, L.; Linden, T.; Hahn-Hagerdal, B Performance of microorganisms in spent sulfite liquor and enzymatic

hy-drolysate of steam pretreated Salix Appl Biochem

Biotech-nol 1992, 34/35, 359-367.

(8) Olsson, L.; Hahn-Hagerdal, B Fermentative performance

of bacteria and yeasts in lignocellulose hydrolysates Process

Biochem 1993, 249-256.

(9) Clark, T A.; Mackie, K L Steam explosion of the softwood

Pinus radiata with sulphur dioxide addition I Process

optimisation J Wood Chem Technol 1987, 7(3), 373-403.

(10) Clark, T A.; Mackie, K L.; Dare, P H.; McDonald, A G

Steam explosion of the softwood Pinus radiata with sulphur dioxide addition II Process characterisation J Wood Chem.

Technol 1989, 9(2), 135-166.

(11) Dekker, R F H The utilization of autohydrolysis-exploded

hardwood (Eucalyptus regnans) and softwood (Pinus radiata)

sawdust for the production of enzymes and fermentable

substrates Biocatalysis 1987, 1, 63-75.

(12) Von Sivers, M.; Zacchi, G A techno-economical comparison

of three processes for production of ethanol from pine

Bioresour Technol 1995, 51, 43-52.

(13) Paszner, L.; Cho, H J High-yield Organosolv process for

conversion of cellulosic biomass to ethanol Proceedings of Energy from Biomass and wAstes XII; Klass, D L., Ed.; IGT:

Chicago, IL, 1988; pp 1300-1318

(14) Paszner, L.; Jeong, C.; Quinde, A.; Awardel-Karim, S Two-stage, continuous hydrolysis of wood by the acid-catalyzed

Organosolv saccharification (ACOS) process Proceedings of Energy from Biomass and Wastes XIII; Klass, D L., Ed.; IGT;

Chicago, IL, 1990; pp 629-664

(15) Harris, J F.; Baker, A J.; Conner, A H.; Jeffries, T W.; Minor, J L.; Pettersen, R C.; Scott, R W.; Springer, E L.;

Wegner, T H.; Zerbe, J I Two-Stage, Dilute Sulfuric Acid Hydrolysis of Wood: An Investigation of Fundamentals; Gen.

Tech Rep FPL-45; Forest Products Laboratory, U.S Depart-ment of Agriculture: Madison, WI, 1985

(16) Quincy Library Group (QLG), California Energy Commis-sion, California Institute of Food and Agriculture, Plumas Corporation, TSS Consultants, and National Renewable

Energy Laboratory Northeastern California Ethanol Manu-facturing Feasibility Study; Quincy Library Group: Quincy,

CA, November 1997

(17) Merrick & Company Softwood Biomass to Ethanol Fea-sibility Study NREL Subcontract AXE-8-18020-01, Final

Report; National Renewable Energy Laboratory: Golden, CO, 1999