R E S E A R C H A N D A N A LY S I S
Carbon Nanofiber Production
Life Cycle Energy Consumption
and Environmental Impact
Vikas Khanna, Bhavik R. Bakshi, and L. James Lee
Keywords:
damage indicators
energy analysis
industrial ecology
nanomaterials
nanotechnology
process life cycle assessment (LCA)
Address correspondence to:
Bhavik R. Bakshi
Department of Chemical and Biomolecular
Engineering
The Ohio State University
140 W., 19th Avenue
Columbus, OH 43210
[email protected]
www.chbmeng.ohio-state.edu/∼bakshi/
research/
Summary
Holistic understanding of nanotechnology using systems analysis tools is essential for evaluating claims about the potential
benefits of this emerging technology. This article presents one
of the first assessments of the life cycle energy requirements
and environmental impact of carbon nanofibers (CNFs) synthesis. Life cycle inventory data are compiled with data reported in the open literature. The results of the study indicate
relatively higher life cycle energy requirements and higher environmental impact of CNFs as compared to traditional materials, like primary aluminum, steel, and polypropylene, on
an equal mass basis. Life cycle energy requirements for CNFs
from a range of feedstock materials are found to be 13 to
50 times that of primary aluminum on an equal mass basis.
Similar trends are observed from the results of process life
cycle assessment (LCA), as conveyed by different midpoint
and endpoint damage indicators. Savings in life cycle energy
consumption and, hence, reductions in environmental burden
are envisaged if higher process yields of these fibers can be
achieved in continuous operations. Since the comparison of
CNFs is performed on an equal mass basis with traditional
materials, these results cannot be generalized for CNF-based
nanoproducts. Quantity of use of these engineered nanomaterials and resulting benefits will decide their energy and environmental impact. Nevertheless, the life cycle inventory and
the results of the study can be used for evaluating the environmental performance of specific CNF-based nanoproducts.
c 2008 by Yale University
DOI: 10.1111/j.1530-9290.2008.00052.x
Volume 12, Number 3
394
Journal of Industrial Ecology
www.blackwellpublishing.com/jie
R E S E A R C H A N D A N A LY S I S
Introduction
The fast-emerging field of nanotechnology has
sparked a high level of interest from the scientific
and industrial communities in recent years and
is often touted as one of the biggest scientific
breakthroughs of the 21st century. A report by
Lux Research (2004) estimates the total value of
nanotechnology products to exceed $2.5 trillion
by 2014. Engineered nanomaterials are appearing
in a wide range of consumer products, from cosmetics to spray paints to automotive body panels.
Carbon nanofibers (CNFs) belong to the new
class of superior engineered materials because of
their exceptional mechanical and electrical properties (De Jong and Geus 2000). CNFs consist
of monomolecular carbon fibers with diameters
ranging from tens of nanometers to 200 nanometers. They are characterized by high-tensile
strength (12,000 megapascal [MPa]) and a high
Young’s modulus (600 gigapascal [GPa]) that is
approximately 10 times that of steel (Mordkovich
2003). Besides mechanical strength, CNFs possess desirable electrical properties, such as high
electrical conductivity. These properties of CNFs
are being explored in a variety of ways to impart
functionalities in various intermediate and final
value-added consumer products. One application
is the use of CNFs as polymer additives, resulting
in high-strength polymer nanocomposites (Sandler et al. 2003; Hammel et al. 2004). Other applications include the use of these engineered
nanoparticles (ENPs) in carbon-lithium batteries, start capacitors for electronic devices, and
electrically conducting polymers (Mordkovich
2003). High specific surface area of these fibers is
an additional attribute that has been investigated
for use of these ENPs as catalyst support material,
especially for liquid-phase reactions (Park and
Baker 1999; De Jong and Geus 2000). CNFs have
also been reported to exhibit remarkable hydrogen storage capacities, which can have long-term
implications for a future hydrogen economy (Fan
et al. 1999; De Jong and Geus 2000).
Although altered physicochemical properties
make CNF-based nanoproducts commercially attractive, they also raise concerns and pose difficult questions about the human and ecosystem
impact1 of these materials. We should remember the unexpected consequences of some earlier
“breakthrough” technologies and products, such
as DDT, chlorofluorocarbons, and asbestos, before reaching any concrete conclusions about the
impact of nanoproducts (Khanna et al. 2007).
The point here is not to discount the benefits of
new technologies for human welfare but that the
costs of failure to see beyond the early hype into
the broader implications of any technology can
be enormous.
In the case of emerging technologies, this entails following a proactive approach to ensure
smart and sensible research and development,
which will thereby lead to a sustainable industry.2
Focusing attention on a single process or a few
emissions without sufficient attention to others in
a product’s supply chain, however, can lead to the
unintended environmental trade-off of one problem for another. Conversely, the use of a holistic,
life cycle approach encompassing a broader system boundary can help identify this problem. One
such methodology is life cycle assessment (LCA).
This approach considers the environmental impact of products or processes over their entire life
cycle.
Researchers and government agencies have
identified the need for the LCA of potential
nanoproducts (Karn 2004; Royal Society 2004;
National Nanotechnology Initiative 2006). LCA
of nanotechnology poses several formidable challenges, however.3 First, the existing life cycle inventory (LCI) databases are limited in scope and
are useful for evaluating only common products
and processes. Establishing an LCI for nanoparticles and nanoprocesses is difficult in the absence of data about the inputs and outputs from
nanomanufacturing processes. Second, there are
very few quantifiable data available on the human
health and ecosystem impacts of products and byproducts of nanomanufacturing (Colvin 2003).
Toxicological studies of a multitude of nanomaterials would be time-consuming and require
substantial investment. Besides these obstacles,
nanotechnology is still in its infancy, and nanomanufacturing processes are evolving rapidly. All
of these challenges pose problems in evaluating
the life cycle impact of nanoproducts.
In the light of these challenges, LCA of nanotechnology has received very little attention, and
only a handful of studies have addressed the life
cycle issues and complexities of nanoproducts
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
395
R E S E A R C H A N D A N A LY S I S
and nanoprocesses. Lloyd and Lave (2003) studied the life cycle implications of replacing automotive body panels made of steel with those of
polymer nanocomposites and aluminum. Their
study employed economic input–output LCA for
quantifying savings in petroleum use and production and, hence, reduction in carbon dioxide
(CO 2 ) emissions thereof by replacing conventional steel body panels with polymer nanocomposites. They also studied the use of nanotechnology to stabilize platinum-group metals in
automobile catalytic converters (Lloyd et al.
2005). Osterwalder and colleagues (2006) compared the wet and dry synthesis methods for
oxide nanoparticle production with respect to energy consumption. In these three studies, details
such as release and impact of emissions during
nanomanufacturing are not considered, possibly due to lack of information. Furthermore,
equipment-level information, such as energy, materials use, and emissions, for the synthesis of
nanoparticles has also been ignored or approximated. Although these studies represent the first
important steps toward LCA of nanomanufacturing, their accuracy is limited for guiding selection
among alternatives.
More recently, Isaacs and colleagues (2006)4
evaluated the economic and environmental
trade-offs for synthesis of single-walled carbon
nanotubes (SWNTs). A comparison of different synthesis routes revealed that the life cycle of SWNTs is hugely energy intensive, with
numbers ranging from 1,440,000 to 2,800,000
megajoules per kilogram (MJ/kg)5 of carbon nanotubes. With the exception of this SWNT study,
no nanoparticle-specific LCAs seem to exist in
the literature that use detailed process-level information and combine that with details from
other steps in the life cycle to quantify the energy
and environmental burden associated with the
production of nanomaterials. Such studies will be
especially useful, as they will lay the foundation
for compiling the LCI of nanoparticle synthesis
and, hence, pave the way for LCA of specific
nanoproducts. This is especially important at an
early stage of research to evaluate the economic–
environmental trade-offs among manufacturing
processes and among alternative products to aid
in sensible engineering decision-making.
396
Journal of Industrial Ecology
LCA of nanomanufacturing6 will be most useful for products and processes that are most likely
to be commercially viable. With this in mind,
the goals of this article are to (1) evaluate the life
cycle energy requirements for CNFs using various
hydrocarbon feedstocks and compare the results
with those of traditional materials and (2) perform a process-based life cycle impact assessment
(LCIA) of CNF synthesis and compare the results
with those of traditional materials. The result is
a first “cradle-to-gate” LCA for CNF that encompasses the major processes starting from the
extraction and manufacture of raw materials and
ending at CNF synthesis. The use phase and endof-life issues pertaining to CNF are not included
in this study.
As this is a cradle-to-gate study, the results
cannot be extended to draw conclusions about
specific nanoproducts that incorporate CNFs.
Nevertheless, the LCA results of this study can
be readily used in an evaluation of CNF-based
nanoproducts. It is important to mention that
the release and impact of nanoparticles on human and ecosystem species along the life cycle
are not modeled due to a lack of emissions and
impact data.
The rest of this article is organized in five sections. The first section briefly describes the tools
and techniques used in the analysis. The synthesis process of CNFs, along with the data sources
and assumptions, is described in the second section. The third section presents the results of life
cycle energy analysis and environmental LCA of
CNFs. A comparison of CNFs with traditional
materials is presented to gain insights into the
relative life cycle energy intensity and environmental impact of CNFs. The fourth section provides discussion and interpretation of the life cycle results. Finally, the last section summarizes the
conclusions of the study and provides directions
for future work.
LCA Approach
This study is a cradle-to-gate analysis of CNF
synthesis, as no specific nanoproducts are investigated. The use and end-of-life phase of the CNF
life cycle are not modeled. The aim is to quantify and compare the environmental burden of
R E S E A R C H A N D A N A LY S I S
CNF synthesis with traditional materials on an
equal mass basis. In the case of CNF synthesis,
given that most of the data are either missing or
proprietary, certain assumptions are made to carefully define the process boundary. For example,
the emissions from the CNF synthesis process in
the life cycle are not included due to lack of information. Although these omissions are likely
to introduce some error and uncertainties in the
final results, the CNF LCA results will still represent preliminary lower bound estimates. Emissions corresponding to the production of all other
inputs required for CNF synthesis are incorporated in the study.
Several LCIA approaches have been described
and critically assessed in detail in the literature
(Hofstetter 1998; Bare and Gloria 2006). In this
study, a hierarchical approach, as described by
Bare and Gloria (2006), is used. The LCI results are available at the disaggregate level. These
are classified and characterized on the basis of
their impact into various impact categories, often described as midpoint indicators. Common
impact categories at the midpoint level include
global warming potential (GWP), acidification
potential, eutrophication potential, human toxicity potential, ozone layer depletion potential,
and photochemical smog formation potential.
Different chemicals can have a variety of different impacts. For example, carbon dioxide contributes to global warming, and hydrochlorofluorocarbons (HCFCs) contribute to ozone layer
depletion. Classification consists of assigning different chemicals into impact categories (e.g.,
GWP and ozone depletion potential). Characterization consists of calculating the extent of
impact for each impact category using a common metric. For example, 1 kilogram (kg)7 of
methane has a global warming impact equivalent to 23 kg of carbon dioxide; similarly, 1 kg
of methyl chloride has a global warming impact
equivalent to 16 kg of carbon dioxide. These
impacts are expressed in terms of carbon dioxide equivalents to quantify the GWP. Characterization factors are available in the literature
for a large number of chemicals and their impacts in various categories. The Centrum Voor
Milieuwetenschappen Leiden (CML) approach
developed by the Leiden University Institute of
Environmental Sciences is used in this study to
obtain midpoint indicators (Guinée 2002; Leiden
University Institute of Environmental Sciences
2003).
We can further aggregate midpoint indicators
to obtain damage indicators. The Eco-Indicator
99 methodology is used to obtain damage indicators (PRé Consultants 2001). Under EcoIndicator, three kinds of environmental damages
are weighed: damage to human health, damage
to ecosystem quality, and damage to resources.
Damage to human health is expressed in terms of
disability-adjusted life years (DALYs). This measure incorporates damage to human health in the
form of years of life lost (YLL) and years of life disabled (YLD) as a result of emissions of substances.
Damage to ecosystems is expressed as the product
of potentially disappeared fraction (PDF), area,
and year (PDF × m2 × yr), where PDF includes
the percentage of species that are threatened or
disappear from a given area in a given time. Finally, damage to resource quality is expressed in
terms of surplus energy for the future mining of
resources (MJ surplus). The underlying idea is
that extraction of mineral resources and fossil fuels now will result in a lower concentration of
these resources in the future and, hence, expenditure of increasing amounts of energy for future
extraction.
Process Description and Data
Sources
Process Description
Avoiding the formation of carbon deposits is
important in refinery applications, such as steam
reforming of hydrocarbons, hydrocracking, and
hydrotreating, primarily because these carbon filaments lead to deactivation of the catalyst surface, blockage in reactor systems, and reduction in heat transfer (Rodriguez 1993; De Jong
and Geus 2000). Nonetheless, the same fibrous
carbon products that are traditionally known
to cause problems in various refinery processes
have unique material properties when reduced to
nanoscale. For example, CNFs have received a
great deal of attention from both the research
and the industrial communities because of their
novel properties, which can be realized in a variety of ways in numerous applications. Several
academic and industrial research groups have
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
397
R E S E A R C H A N D A N A LY S I S
directed their efforts toward synthesizing and optimizing the growth of these CNFs.
In its simplest form, the catalytic synthesis
of CNFs consists of formation of these fibers
on metallic catalysts in the form of powders,
foils, gauzes, or supported particles. The process consists of reducing the catalyst sample in
a hydrogen-inert gas stream at a somewhat lower
temperature, followed by heating the catalyst up
to the reaction temperature, subsequent to which
the reaction mixture, consisting of hydrocarbon,
hydrogen, and inert gas, is introduced into the
system (Endo et al. 1988; Yang and Chen 1989;
Kim et al. 1991; Hernadi et al. 2002; Lim et al.
2004; Takehira et al. 2005). The reaction proceeds for periods ranging from a few minutes
to several hours. Several different metallic and
bimetallic catalysts can be used. Most commonly
used are iron, cobalt, nickel, and copper, both in
bulk and in supported form. Lower hydrocarbons,
such as methane, ethylene, acetylene, or benzene,
and carbon monoxide are the common sources
of carbon material. The most common mechanism proposed in the literature for the synthesis
of CNFs consists of decomposition of the hydrocarbons on the metal surface, releasing carbon
atoms. These carbon atoms then form metal carbides that dissolve and diffuse through the bulk
of the metal, resulting in the deposition of CNFs
at the other end of the metal particles (De Bokx
et al. 1985; Kock et al. 1985; Alstrup 1988).
Despite great advances in synthesis methods
and efforts at understanding the mechanism of
nucleation and growth of CNFs, continuous production of these fibers has proven to be challenging, and several issues need to be addressed. In
recent years, a new method capable of synthesizing CNFs on a continuous scale, namely the
vapor-grown carbon nanofibers (VGCNF) process, has emerged (Tibbetts and Gorkiewicz 1993;
Tibbetts et al. 1994; Rodriguez et al. 1995; Chambers et al. 1996; Fan et al. 2000a,b). A schematic
of the VGCNF synthesis is shown in figure 1.
VGCNFs are produced by catalytic pyrolysis
of hydrocarbons in the presence of a transition
metal catalyst. A trace amount of sulfur is added
to the feed to promote the formation of CNFs.
The role of sulfur is complex and has been discussed in the literature (Tibbetts et al. 1994).
The hydrocarbon feed, along with the catalyst,
398
Journal of Industrial Ecology
Figure 1 Process schematic of vapor-grown carbon
nanofiber (VGCNF) synthesis. (The schematic is
based on the process description by Tibbetts and
Gorkiewicz 1993; Tibbetts et al. 1994; Fan et al.
2000a,b.)
hydrogen gas, and a sulfur source, is introduced
into the reactor. Ferrocene (C 10 H 10 Fe), dissolved
in a suitable solvent, and iron pentacarbonyl
(Fe(CO) 5 ) are commonly used catalyst sources.
The organometallic catalyst decomposes, forming
clusters of iron (Fe) particles that act as nuclei
for the formation and further growth of CNFs.
The fibers grow as they move along the reactor, and the diameter of fibers increases by the
classic chemical vapor deposition (CVD) mechanism (Hitchman and Jensen 1993). Both horizontal and vertical reactor configurations have
been employed (Tibbetts and Gorkiewicz 1993;
Tibbetts et al. 1994; Fan et al. 2000a,b). The
fibers coming out of the reactor, along with the
off-gases, are trapped and separated via a series
of cyclone separators or trap mechanisms located
downstream of the reactor. The entire reactor assembly is enclosed within an electric furnace to
supply sufficient heat to maintain the pyrolysis
temperature of 1,100 to 1,200◦ C. The characteristics of the fibers, such as their thickness and
length, and, hence, their properties can be controlled through careful manipulation of the parameters, such as the catalyst concentration, reaction time, and size of the catalyst particles produced (Tibbetts and Gorkiewicz 1993; Fan et al.
2000a,b). In addition, the continuous process offers the advantage of achieving high throughputs,
which is commercially desirable.
Process and Life Cycle Data
As discussed in the LCA Approach section,
performing LCA of nanotechnology is difficult
R E S E A R C H A N D A N A LY S I S
because of the lack of data about the inputs and
outputs of nanomanufacturing. In this study, we
address this challenge by compiling an LCI for
VGCNF synthesis on the basis of laboratory experience and data available in the open literature
(Tibbetts and Gorkiewicz 1993; Tibbetts et al.
1994; Fan et al. 2000a,b). Missing data are estimated on the basis of suitable assumptions and
with the criterion that the conservation laws of
mass and energy are satisfied. For example, complete recycling of unreacted hydrocarbons back
to the reactor is assumed for sensitivity analysis,
with no recycling assumed to be the base case.
The catalyst life cycle is not accounted for due
to the absence of LCI data about the catalyst.
LCI data for common material inputs are obtained directly from the SimaPro 6.0 Demo Version (SimaProTM 2006). In particular, a data set
for western European technologies was used for
the sake of consistency, as this data set comprises
LCI data for most resources consumed in the CNF
life cycle. Inventory data for traditional materials, such as aluminum, steel, and polypropylene,
are also obtained from SimaPro. On the basis
of LCI data for electricity generation from the
National Renewable Energy Laboratory (2007)
U.S. LCI Database, a life cycle efficiency of 40%
is assumed to estimate the life cycle energy requirements for the electricity. The electricity mix
considered is from coal, natural gas, residual oil,
and other sources in minor quantities. Several different cases are evaluated with respect to energy
consumption and environmental impact.
Life cycle energy requirements are evaluated for VGCNF synthesis from three hydrocarbon feedstocks: methane, ethylene, and benzene.
Table 1 lists the material consumption data for
each synthesis case.
The system boundary employed for the life
cycle energy analysis and process LCA of CNFs
is depicted in figure 2. The dashed lines in
figure 2 indicate that the LCI for the corresponding material and energy sources is obtained from
the literature and LCI databases and is included
in the LCA. Process energy requirements include
the initial reactor heat-up and the energy required for the catalytic pyrolysis of hydrocarbons.
This is primarily electrical energy and is estimated from CVD equipment data sheets. The
boundaries for the process energy requirements
are limited to the reactor. The reactor assembly, along with the electric furnace, is assumed
to be operating at steady state under adiabatic
conditions. Feed is introduced at room temperature, with the cracking of hydrocarbons occurring at 1,100◦ C. A yield value of 50% reported
by Tibbetts and Gorkiewicz (1993) is assumed
for methane and ethylene cases, and 30%, as reported by Fan and colleagues (2000a,b), is assumed for the benzene case. Here, yield is calculated as the ratio of the mass of CNFs produced
to total mass of carbon in the feedstock. Material inputs are compiled on the basis of data
available in the published literature (Tibbetts
and Gorkiewicz 1993; Tibbetts et al. 1994; Fan
et al. 2000a,b). A purification efficiency of 90%
is considered, meaning that 90% of the synthesized nanofibers are recovered as useful final product. The indirect life cycle energy requirements
due to the consumption of inputs such as hydrocarbon feedstock, carrier gas, sulfur source, and
electricity are obtained directly from SimaPro.
Table 1 Material inventory for 1 kilogram (kg) of carbon nanofibers
Quantity (kg)
Material inputs
Methane feedstock
Ethylene feedstock
Benzene feedstock
Methane (carbon source)
Ethylene (carbon source)
Benzene (carbon source)
Sulfur source
Hexane (solvent)
Hydrogen
Ferrocene (catalyst)
HCl
2.67
–
–
0.28 (H 2 S)
14.30
1.57
0.22
16
–
2.33
–
0.20 (H 2 S)
11.01
1.11
0.14
16
–
–
4.70
0.03 (thiophene)
–
1.72
–
16
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
399
R E S E A R C H A N D A N A LY S I S
Figure 2 Vapor-grown carbon nanofibers–life cycle assessment (VGCNF-LCA): system boundary.
Separation and purification energy requirements
for a raw product can be significant and can exceed the main process energy costs, depending
on the level of impurities and the difficulty of
separation. As-produced or raw CNFs can have
two types of impurities: amorphous carbon and
residual metal. Hydrogen peroxide and nitric acid
are commonly used to get rid of amorphous carbonaceous content, whereas mild to strong acids
have been employed to remove metallic impurities. The chemical requirements for removing
metallic impurities and, hence, their life cycle
energy requirements are estimated on the basis
of available literature data (Choi et al. 2003;
Nguyen et al. 2003). If we assume that the acid
quantity required for continuous purification can
be significantly reduced, the calculated amount of
hydrochloric acid required for batch purification
is reduced by a factor of 100. The energy required
for separation and purification is not considered.
Although the actual quantity of distilled water
required for continuous operation is not known,
the life cycle energy contribution for distilled water (approximated with demineralized water) is
much smaller (0.006 MJ fossil eq./kg) in comparison to the process electrical energy required for
CNF synthesis and hence is ignored.
Three different feedstocks—namely, methane, ethylene, and benzene—are examined for
400
Journal of Industrial Ecology
the life cycle energy analysis of VGCNFs. The
energy associated with producing the various inputs, producing VGCNFs, and producing materials for VGCNF purification are combined to
get the overall life cycle energy requirements for
the synthesis of CNFs. Here, life cycle energy consumption refers to the cumulative nonrenewable
energy requirements for the synthesis of CNFs.
Consumption of other types of energy or materials, including raw materials such as ethylene and
benzene, is ignored.
We evaluate different scenarios for unreacted
hydrocarbon and carrier gas recycling rates to
address uncertainty and evaluate the bounds on
the life cycle energy consumption and emissions.
Several aspects of VGCNF production are excluded from the life cycle energy analysis and process LCA. These include the catalyst life cycle,
separation and purification energy requirements,
and the energy required for producing distilled
water. Accordingly, the results provide a conservative lower estimate of the life cycle energy
requirements and life cycle impacts.
Results
Life Cycle Energy Analysis
Figure 3(a) presents a direct comparison of the
life cycle energy requirements of CNF synthesis
R E S E A R C H A N D A N A LY S I S
Figure 3 Life cycle energy analysis
of carbon nanofibers (CNFs):
(a) effect of cycle time (the error
bars reflect the effect of production
cycle times ranging from 1 hr to a
continuous operation for 300 days),
(b) energy distribution along life cycle
phases, and (c) effect of feedstock
and carrier gas recycle (assuming
100% recycle of unreacted feedstock
and 90% recycle of carrier gas).
for different starting feedstocks with those of producing aluminum, steel, and polypropylene. It is
important to mention that CNFs are not likely
to be used directly in place of materials such
as aluminum, steel, and polypropylene. CNFbased nanoproducts, such as lightweight polymer
nanocomposites, are, however, being considered
as possible replacements for materials such as aluminum and steel (Sandler et al. 2003; Hammel
et al. 2004). Nevertheless, a direct comparison
of life cycle energy requirements for CNFs with
materials such as polypropylene, aluminum, and
steel is still informative and can highlight the
relative life cycle energy intensity of these ENPs.
Figure 3(a) reveals that the life cycle of CNFs
is energy intensive, with the life cycle energy requirements ranging from 2,872 MJ/kg for benzene
feedstock to around 10,925 MJ/kg for methane.
In comparison, the life cycle energy requirements
for aluminum, steel, and polypropylene are 218,
30, and 119 MJ/kg, respectively.
Table 2 and figure 3(b) show the breakdown
of energy requirements along the CNF life cycle phases. The life cycle energy requirement
for CNFs consists of three major components:
process energy, indirect effects, and purification
energy. Process energy can be further split into
energy required for initial reactor heat-up and the
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
401
R E S E A R C H A N D A N A LY S I S
Table 2 Carbon nanofiber synthesis—life cycle energy requirements in megajoules per kilogram (MJ/kg)
Energy requirements (MJ/kg)
Feedstock type
Recycle streams
Process
Indirect effects
Purification
Life cycle
Methane
Methane
Ethylene
Ethylene
Benzene
Benzene
No
Feed and 90% H 2
No
Feed and 90% H 2
No
Feed and 90% H 2
9,391
9,391
6,652
6,652
2,283
2,283
1,378
1,231
1,160
1,000
433
85
156
156
156
156
156
156
10,925
10,778
7,968
7,808
2,872
2,524
pyrolysis energy. Whereas initial reactor heating
and temperature ramp-up is a one-time operation and fixed, pyrolysis energy is a function of
the residence time.
It is evident from figure 3(b) that process energy consumption constitutes a large fraction of
the overall life cycle energy requirements, ranging from 79% to 86% for benzene and methane
feedstocks, respectively. This is primarily the
resources consumed, mainly coal, to produce
the electrical energy required to maintain the
high decomposition temperature (around 1,100
to 1,200◦ C). The process electricity requirement
for synthesizing CNFs using benzene is around
900 MJ/kg, which is very close to the process electricity required for producing polysilicon from
trichlorosilane in the semiconductor industry via
CVD (Williams et al. 2002).
A sensitivity analysis is further performed to
study the effect of varying cycle times and unreacted feedstock and carrier gas recycle rates on
the life cycle energy consumption. The positive
error bars for the CNFs in figure 3(a) represent
the effect of cycle time on the process energy
and, hence, the total life cycle energy requirements. Cycle times considered in this study range
from a run time of 1 hr to a continuous operating cycle of 300 days per year. This may be
an unnecessarily wide range of operation, and in
practice the uncertainty due to cycle time may
be much smaller than depicted in this article. A
24-hr continuous operation for a period of
300 days is considered as the base case, giving the
most conservative estimates of the CNF life cycle energy requirements for all hydrocarbon feedstocks. The life cycle energy required is highest
for a run time of 1 hr, decreases progressively as
the cycle time increases, and eventually flattens
out at longer cycle times. This is primarily because
402
Journal of Industrial Ecology
the process electrical energy required for initial
reactor heat-up gets distributed over the increasing cycle times and the increased production of
CNFs. The effect of unreacted feedstock and carrier gas recycling rates on the life cycle energy
consumption is shown in figure 3(c). Considering
the methane feedstock, one can observe that even
with complete recycling of the unreacted hydrocarbons and a 90% recycling of the hydrogen
stream, the total energy consumption decreases
from 10,925 MJ/kg to 10,778 MJ/kg. This corresponds to a 1.3% decrease in the overall life cycle
energy requirements. In practice, such a high recycling rate is unlikely, and some of the unreacted
materials may be burned for their fuel value. A
similar trend is observed for ethylene feedstock,
where complete recycling of the unreacted feed
and 90% recycling of the hydrogen streams corresponds to a 2% reduction in the overall energy
requirements. Thus, the process energy requirement still outweighs the energy savings due to
material recycling. As discussed, the energy required for separating unreacted hydrocarbons and
hydrogen from the off-gases leaving the reactor is
not accounted for here. Therefore, the numbers
presented in table 2 and figure 3(c) represent an
upper bound on the energy savings from gas separation and system recycling.
Environmental LCA of CNFs
This subsection presents the results of a process LCA of CNFs. Because the scope is restricted
to a cradle-to-gate LCA of CNFs and not specific
nanoproducts, the results of the CNF LCIA are
compared with traditional materials on an equal
mass basis. Figure 4 compares various midpoint
indicators for CNFs with aluminum, steel, and
polypropylene on a per-kilogram basis. Two base
R E S E A R C H A N D A N A LY S I S
(a) Global Warming Potential
(b) Human Toxicity Potential
1400
kg 1,4-dichlorobenzene Equivalent
600
kg CO2 Equivalent
1200
1000
800
600
400
200
500
400
300
200
100
0
0
CNF (Methane)
CNF (Ethylene)
Primary
Aluminum
Steel
Polypropylene
CNF (Methane)
(c) Ozone Layer Depletion Potential
CNF (Ethylene)
Primary
Aluminum
Steel
Polypropylene
(d) Photochemical Oxidation Potential
3.5E-05
4.0E-01
3.5E-01
kg Ethylene Equivalent
kg CFC-11 Equivalent
3.0E-05
2.5E-05
2.0E-05
1.5E-05
1.0E-05
5.0E-06
3.0E-01
2.5E-01
2.0E-01
1.5E-01
1.0E-01
5.0E-02
0.0E+00
0.0E+00
CNF (Methane) CNF (Ethylene)
Primary
Aluminum
Steel
Polypropylene
CNF (Methane) CNF (Ethylene)
(e) Freshwater Aquatic Ecotoxicity Potential
Steel
Polypropylene
(f) Terrestrial Ecotoxicity Potential
5
kg 1,4-dichlorobenzene Equivalent
4.5
kg 1,4-dichlorobenzene Equivalent
Primary
Aluminum
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
CNF (Methane) CNF (Ethylene)
Primary
Aluminum
Steel
Polypropylene
CNF (Methane) CNF (Ethylene)
(g) Acidification Potential
Primary
Aluminum
Steel
Polypropylene
(h) Eutrophication Potential
0.9
0.8
12
0.7
kg PO4 Equivalent
kg SO2 Equivalent
10
8
6
4
0.6
0.5
0.4
0.3
0.2
2
0.1
0
0
CNF (Methane) CNF (Ethylene)
Primary
Aluminum
Steel
Polypropylene
CNF (Methane) CNF (Ethylene)
Primary
Aluminum
Steel
Polypropylene
Figure 4 Vapor-grown carbon nanofibers–life cycle assessment (VGCNF-LCA): midpoint indicators. (The
positive error bars in the figure represent the effect of increasing production cycle times.)
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
403
R E S E A R C H A N D A N A LY S I S
cases are evaluated for CNF synthesis, one with
methane and the other with ethylene as the feedstock. Methane and ethylene are considered as
the base cases because they are the most preferred
material feedstocks. Both cases are considered to
have hydrogen as the carrier gas, in accordance
with the current industrial schemes. Midpoint indicators generally represent indexes somewhere
along the environmental mechanism or on the
cause–effect chain between the origin of individual emissions flow and the areas of protection:
entities considered valuable to society (Leiden
University Institute of Environmental Sciences
2003; Bare and Gloria 2006).
Figure 4 shows a higher impact potential in
all impact categories for both methane-based and
ethylene-based CNFs when compared with aluminum, steel, and polypropylene on an equal
mass basis. For example, the GWP of 1kg of
methane-based CNFs is equivalent to about 65 kg
of primary aluminum, whereas 1 kg of ethylenebased CNFs has a GWP equivalent of about 47 kg
of steel. The positive error bars for the CNFs in
figure 4 represent the effect of increasing cycle
time on the life cycle impact numbers. Higher
cycle times are associated with higher electricity
consumption, which increases the potential life
cycle impact. The error bars are smaller for some
of the midpoint indicators than for others. For
example, in figure 4(b), most of the human toxicity potential (more than 90%) results from the
hydrocarbon feedstock, and hence the effect of
increasing process electricity requirements is not
huge, as reflected by smaller error bars. Similarly,
as the ozone layer depletion potential associated
with electricity is almost negligible in comparison to the potentials for other material resources,
the effect of increasing process electricity requirements is not visible in figure 4(c).
Figure 5 presents the damage indicators obtained via Eco-indicator 99. The figure indicates
a higher impact from CNF production in each
of the three damage categories (i.e., damage to
human health, ecosystems, and resource quality) as compared to traditional materials on an
equal mass basis. Damage indicators are desirable from a decision-making point of view because they aggregate data to represent the relative severity of impact using fewer metrics. They
are more uncertain, however, because life cycle
404
Journal of Industrial Ecology
data are more aggregated and models are still
not available to determine the endpoint effect
of certain emissions (Leiden University Institute of Environmental Sciences 2003; Bare and
Gloria 2006).
It is important to reiterate that release and impact of CNFs on humans and ecosystem species
during manufacturing are not considered due to a
lack of quantifiable data about possible fate, transport, and mechanism of damage of CNFs. Release
of CNFs can occur at several stages in the life
cycle. These include the CNF reactor, the separation system, and the purification stage where
CNFs are dissolved in the solvent. Although not
included in the scope of this study, CNF release
and exposure to the environment is also likely
during the processing stages, when these ENPs are
incorporated in nanoproducts. The LCIA results
presented reflect only the material and energy use
during the synthesis of CNFs. The LCIA results
for the CNFs represent lower bounds, whereas
those for the traditional materials are likely to
be more precise, given that detailed and relatively more complete inventory data are available. Also, the emissions and, hence, the impact
from the CNF synthesis step are not accounted
for. In addition, emissions from the catalyst life
cycle have been ignored due to lack of data.
Discussion and Interpretation
of Life Cycle Results
On a per mass basis, the life cycle energy requirements for producing CNFs is 13 to 50 times
what is required for producing primary aluminum.
A similar comparison can be made for the midpoint and damage indicators evaluated in the process LCA of CNFs. There are several reasons for
this difference. First, the high-temperature vapor phase process required for CNF synthesis has
low efficiency and requires significant energy investment. Second, processes required for manufacturing traditional materials are well studied
and have been optimized with respect to material
and energy consumption over the last century or
so. Although researchers have been synthesizing
CNFs on a laboratory scale since the early 1990s,
industrial synthesis has only recently begun,
and several challenges remain, including precise
control of the process parameters to tailor the
R E S E A R C H A N D A N A LY S I S
Figure 5 Vapor-grown carbon nanofibers–life cycle assessment (VGCNF-LCA): damage indicators:
(a) disability-adjusted life years (DALYs), (b) PDF × m2 × yr, (c) MJ surplus.
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
405
R E S E A R C H A N D A N A LY S I S
diameter and length of fibers for meeting the required specifications and low yields (e.g., 10% to
30% by weight of the feedstock). Research efforts
are needed for process improvement to increase
the yield of these nanoparticles to make them
competitive with alternatives for a given application. In addition, most of industrial CNF synthesis operations are small-scale, and, as a result,
economies of scale have not yet been realized. Finally, CNFs are a highly ordered form of matter,
requiring enormous energy investment to create
the huge surface area and low entropy.
Some of the questions that may arise about
the usefulness of the results presented in this article are that (1) the life cycle results may be
gross overestimates because nanomanufacturing
processes are likely to become significantly more
efficient over time, as has happened with most
other technologies, and (2) comparing the energy
use or impact on a mass basis may be misleading
because the mass of CNFs used in commercial
products may be much less than that of conventional materials for the same application. These
issues are addressed below.
Assessing the energy requirements for future
CNF production is a challenging task that will
require knowledge about the trajectory of the
production process. Nevertheless, some useful
qualitative insights can be gained by a historical
analysis of other products and processes. A time
series of exergy-based assessments for traditional
materials, such as steel, high-density polyethylene (HDPE), pulp and paper, synthetic ammonia, and synthetic soda ash, is available in the
literature (Ayres et al. 2003). The analysis was
based on industrial data of the U.S. economy
for the 20th century. The study highlights improvements by a factor of 10 in the chemical
industry with respect to the exergy consumption of HDPE synthesis between the 1940s and
1988 and for synthetic ammonia between 1905
and 2000. The steel industry, which represents
the high-temperature process (defined as hotter
than 600◦ C), has demonstrated a threefold improvement in the exergetic efficiency over the
last 100 years. In the production of primary aluminum, the energy required for the electrolytic
smelting process decreased from 180 to 47 MJ/kg
in between 1888 and 1990. Most of these process
improvements have been attributed to changes in
406
Journal of Industrial Ecology
the production technology, feedstock, and production equipment. Incremental changes were
brought about as a result of systemwide optimization of production facilities. Experience curves
showing the relationship between cost and production rate for various energy technologies are
available (International Energy Agency 2000).
If we take the most optimistic figures based on
such historical data, a tenfold decrease in the life
cycle energy requirements for the CNFs would
mean life cycle energy values of 1,100, 800, and
300 MJ/kg for methane, ethylene, and benzene
feedstock, respectively. Even with this optimistic
scenario, the resulting energy consumption numbers are three to ten times higher when compared
on an equal mass basis with aluminum and steel.
This, of course, is based on the assumption that
the decreasing energy requirements will lead to
price reductions, causing an increase in demand,
and that large throughputs can be achieved.
Achieving these reductions will require possible changes in the production technology—for
example, switching to low-temperature processes
(Minea et al. 2004); a newer, more efficient feedstock; and possible integration with other production facilities. LCA of these new processes
should be an integral component of research efforts to ensure that savings in energy and material
resources are not offset by increased consumption or emissions in other phases of the life
cycle.
One argument that is often given in favor of
the potential impact of nanomaterials is the relatively small quantities in which these materials
will be required for various end product applications. Nonetheless, the estimated production
rates of nanomaterials are 58,000 metric tons
of ENPs over the period from 2011 to 2020
(WWICS Project on Emerging Nanotechnologies 2006). Concerns have been expressed that
this can have an impact equivalent to 5 million to
50 billion metric tons of conventional materials
when their size is accounted for (WWICS Project
on Emerging Nanotechnologies 2006). Studies in
the literature present evidence of size-dependent
hazardous and toxic effects of carbon-based nanomaterials, including CNFs (Magrez et al. 2006).
It is argued that the size, aspect ratio, and the
surface chemistry of CNFs play a significant role
in determining their biotoxicity.
R E S E A R C H A N D A N A LY S I S
Increasing process efficiency alone need not
imply a reduction in the overall impact of the
nanomaterial production. Increased efficiency
can trigger lower production costs, which, in turn,
can lead to greater demand. This has been witnessed in the past for the semiconductor industry and electronic goods specifically and is often termed the “rebound effect.” Besides, whether
material and energy efficiency leads to an overall
reduction in resource use is debatable (Herring
1999). In the case of nanomaterials production,
the answer to the questions posed above will
depend on our ability to exploit the benefits
of nanotechnology using the principles of green
manufacturing.
Conclusions and Future Work
High energy requirements as well as the associated high cost of CNFs currently restrict their use
to niche applications and possibly hinder their
use for large-volume applications. Research efforts are required to increase the process yield and
optimize the growth of CNFs to realize economies
of scale and translate novel properties into products with superior properties.
Although the results of a life cycle energy analysis and cradle-to-gate LCIA of CNF synthesis indicate that, on an equal mass basis, CNFs require
13 to 50 times more energy and have a larger life
cycle environmental impact in all impact categories than traditional materials, products based
on CNFs may be greener than alternatives for a
given application. Quantity of use of engineered
nanomaterials and material and energy savings
due to the use of these products will be the
deciding factors. Hence, comparisons of specific
nanoproducts and applications are needed.
Successful implementation of nanotechnology will require that the industry develops in a
safe and sustainable way while addressing potential risks. Nanotechnology is currently in its infancy with respect to its widespread adoption.
Currently, there are no regulations as far as its
development is concerned. Addressing environmental, health, and safety concerns in a timely
manner will help identify regulatory needs and
address public skepticism about the emerging field
of nanotechnology. It is inevitable that this powerful technology will have a pronounced impact
on future science and technology. To ensure that
the impact is constructive rather than destructive, we need to balance potential benefits and
risks as the industry proceeds toward widespread
commercialization.
The need for LCAs of potential nanoproducts cannot be overemphasized. Any LCA study
of nanoproducts will most likely suffer from high
uncertainty at the early stages of nanotechnology
research. Efforts are directed toward compiling a
life cycle inventory of engineered nanomaterials that will help pave the way for LCA of specific nanoproducts. LCA can be used to quantify
how much of the savings in energy and material
consumption is offset by increases in the manufacturing phases of the life cycle. As toxicological data on engineered nanomaterials become
available, the same will be incorporated into the
LCA framework. Current work is in progress to
use these results to evaluate claims about the potential benefits of specific CNF-based nanoproducts. Conventional LCA studies ignore the possible emissions and impact of nanoparticles, and
thermodynamic LCA methods are also being explored as proxy indicators in the absence of emissions and impact information. In the long term,
integration of this work with the detailed fate and
transport studies of these nanomaterials can provide useful insights about the possible life cycle
impacts of nanomaterials.
Acknowledgements
The research presented in this article was supported by the U.S. National Science Foundation (Grant No. EEC-0425626) and the U.S.
Environmental Protection Agency (Grant No.
R832532).
Notes
1. Editor’s note: For a discussion of nanomaterials’
biological risks to aquatic systems, see the article
by MacCormack and Goss (2008).
2. Editor’s note: For two discussions of the process
of making nanotechnology green, see the column
by Karn (2008) and the article by Eckelman and
colleagues (2008).
3. Editor’s note: This special issue contains several
articles that describe efforts to address some of these
challenges by combining LCA with other tools. See
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
407
R E S E A R C H A N D A N A LY S I S
4.
5.
6.
7.
the column by Seager and Linkov (2008) for a discussion of combining multicriteria decision analysis
with LCA and the column by Shatkin (2008) for a
discussion of using risk analysis with LCA.
Editor’s note: A more recent article on this topic
appears in this special issue on nanotechnology. See
the article by Healy and colleagues (2008).
One megajoule per kilogram (MJ/kg) is equivalent
to 108 kilocalories per pound.
Editor’s note: For an overview of nanomanufacturing methods, see the article by Sengul and colleagues (2008) in this issue.
One kilogram (kg) is equivalent to 2.2 pounds (lbs).
References
Alstrup, I. J. 1988. New model explaining carbon filament growth on nickel, iron, and Ni-Cu alloy catalysts. Journal of Catalysis 109(2): 241–
251.
Ayres, R. U., L. W. Ayres, and B. Warr. 2003. Exergy,
power and work in the US economy, 1900–1998.
Energy 28(3): 219–273.
Bare, J. and T. P. Gloria. 2006. Critical analysis
of the mathematical relationships and comprehensiveness of life cycle impact assessment approaches. Environmental Science and Technology
40(4): 1104–1113.
Chambers, A., N. M. Rodriguez, and R. T. K. Baker.
1996. Influence of copper on the structural characteristics of carbon nanofibers produced from the
cobalt-catalyzed decomposition of ethylene. Journal of Materials Research 11(2): 430–438.
Choi, W., S. Park, H. Takahashi, and T. Cho. 2003.
Purification of carbon nanofibers with hydrogen
peroxide. Synthetic Metals 139(1): 39–42.
Colvin, V. L. 2003. The potential environmental impact of engineered nanomaterials. Nature Biotechnology 21(10): 1166–1170.
De Bokx, P. K., A. J. H. M. Kock, E. Boellaard,
W. Klop, and J. W. Geus. 1985. The formation
of filamentous carbon on iron and nickel catalysts:
I. Thermodynamics. Journal of Catalysis 96(2):
454–467.
De Jong, K. P. and J. W. Geus. 2000. Carbon
nanofibers: Catalytic synthesis & applications.
Catalysis Reviews—Science and Engineering 42(4):
481–510.
Eckelman, M. J., J. B. Zimmerman, and P. J. Anastas. 2008. Toward green nano: E-factor analysis of
several nanomaterial syntheses. Journal of Industrial Ecology 12(3) doi: 10.1111/j.1530-9290.2008.
00043.x.
Endo, M., H. Yamanashi, and G. L. Doll. 1988. Prepara-
408
Journal of Industrial Ecology
tion and electrical properties of bromine intercalated vapor-grown carbon fibers. Journal of Applied
Physics 64(5): 2995–3004.
Fan, Y., B. Liao, M. Liu, Y. Wei, M. Lu, and H. Cheng.
1999. Hydrogen uptake in vapour-grown carbon
nanofibers. Carbon 37(10): 1649–1652.
Fan, Y., H. Cheng, Y. Wei, G. Su, and Z. Shen.
2000a. The influence of preparation parameters
on the mass production of vapor-grown carbon
nanofibers. Carbon 38(6): 789–795.
Fan, Y., H. Cheng, Y. Wei, G. Su, and Z. Shen. 2000b.
Tailoring the diameters of vapor-grown carbon
nanofibers. Carbon 38(6): 921–927.
Guinée, J. B., ed. 2002. Handbook on life cycle assessment: Operation guide to the ISO standard. Boston:
Kluwer Academic.
Hammel, E., X. Tang, M. Trampert, T. Schmitt, K.
Mauthner, A. Eder, and P. Pötschke. 2004. Carbon nanofibers for composite applications. Carbon
42(5–6): 1153–1158.
Healy, M. L., L. Dahlben, and J. A. Isaacs. 2008. Environmental assessment of single wall carbon nanotube processes. Journal of Industrial Ecology 12(3)
doi: 10.1111/j.1530-9290.2008.00058.x.
Hernadi, K., A. Fonseca, J. B. Nagy, A. Fudala, D.
Bernaerts, and I. Kiricsi. 2002. Catalytic production of carbon nanofibers over iron carbide doped
with Sn2+ . Applied Catalysis A: General 228(1–2):
103–113.
Herring, H. 1999. Does energy efficiency save energy?
The debate and its consequences. Applied Energy
63(3): 209–226.
Hitchman, M. L. and K. F. Jensen. 1993. Chemical
vapour deposition: Principles and applications. San
Diego, CA: Academic Press.
Hofstetter, P. 1998. Perspectives in life cycle impact assessment: A structured approach to combine models
of the technosphere, ecosphere and valuesphere. Norwell, MA: Kluwer Academic.
International Energy Agency. 2000. Experience curves
for energy technology policy. http://iea.org/
textbase/nppdf/free/2000/curve2000.pdf. Accessed October 2007.
Isaacs, J. A., A. Tanwani, and M. L. Healy. 2006. Environmental assessment of SWNT production. Proceedings of the 2006 IEEE International Symposium
on Electronics and the Environment. Piscataway,
NJ: IEEE p. 38–41.
Karn, B. P. 2004. Small stuff, big opportunity for industrial ecologists. International Society for Industrial
Ecology Newsletter 4(4): 1, 3–4.
Karn, B. 2008. The road to green nanotechnology. Journal of Industrial Ecology 12(3) doi: 10.1111/j.15309290.2008.00045.x.
R E S E A R C H A N D A N A LY S I S
Khanna V., B. R. Bakshi, and L. J. Lee. 2007. Life cycle energy analysis and environmental life cycle
impact assessment of carbon nanofiber production. Proceedings of the 2007 IEEE International
Symposium on Electronics and the Environment. Piscataway, NJ: IEEE p. 128–133.
Kim, M. S., N. M. Rodriguez, and R. T. K. Baker. 1991.
The interaction of hydrocarbons with coppernickel and nickel in the formation of carbon filaments. Journal of Catalysis 131(1): 60–73.
Kock, A. J. H. M., P. K. De Bokx, E. Boellaard, W.
Klop, and J. W. Geus. 1985. The formation of
filamentous carbon on iron and nickel catalysts:
II. Mechanism. Journal of Catalysis 96(2): 468–
480.
Leiden University Institute of Environmental Sciences (CML). 2003. Towards a life cycle impact
assessment method which comprises category indicators at the midpoint and the endpoint level—
Report of the first project phase: Design of the new
method. www.leidenuniv.nl/cml/ssp/publications/
recipe phasel.pdf. Accessed October 2007.
Lim, S., S. Yoon, Y. Korai, and I. Mochida. 2004. Selective synthesis of thin carbon nanofibers: I. Over
nickel–iron alloys supported on carbon black.
Carbon 42(7–8): 1765–1781.
Lloyd, S. M. and L. B. Lave. 2003. Life cycle economic and environmental implications of using
nanocomposites in automobiles. Environmental
Science and Technology 37(15): 3458–3466.
Lloyd, S. M., L. B. Lave, and H. S. Matthews.
2005. Life cycle benefits of using nanotechnology to stabilize platinum-group metal particles in
automotive catalysts. Environmental Science and
Technology 39(17): 1384–1392.
Lux Research, Inc. 2004. Sizing nanotechnology’s
value chain. http://luxresearchinc.com/press/
RELEASE_SizingReport.pdf. Accessed October
2007.
MacCormack, T. J. and G. G. Goss. 2008. Identifying and predicting biological risks associated with
manufactured nanoparticles in aquatic ecosystems. Journal of Industrial Ecology 12(3) doi:
10.1111/j.1530-9290.2008.00041.x.
Magrez, A., S. Kasas, V. Salicio, N. Pasquier, J. W. Seo,
M. Celio, S. Catsicas, B. Schwaller, and L. Forró.
2006. Cellular toxicity of carbon-based nanomaterials. Nano Letters 6(6): 1121–1125.
Minea, T. M., S. Point, A. Granier, and M.
Touzeau. 2004. Room temperature synthesis of
carbon nanofibers containing nitrogen by plasmaenhanced chemical vapor deposition. Applied
Physics Letters 85(7): 1244–1246.
Mordkovich, V. Z. 2003. Carbon nanofibers: A new
ultrahigh-strength material for chemical technology. Theoretical Foundations of Chemical Engineering 37(5): 429–438.
National Nanotechnology Initiative. 2006. Environmental, health, and safety research needs for
engineered nanoscale materials. http://www.
nano.gov/NNI_EHS_research_needs.pdf.
Accessed October 2007.
National Renewable Energy Laboratory. 2007. US
life-cycle inventory database. http://www.nrel.
gov/lci/. Accessed April 2007.
Nguyen, C. V., L. Delziet, K. Matthews, B. Chen,
and M. Meyyappan. 2003. Purification process
for vertically aligned carbon nanofibers. Journal
of Nanoscience and Nanotechnology 3(1–2): 121–
125.
Osterwalder, N., C. Capello, K. Hungerbuhler, and
W. J. Stark. 2006. Energy consumption during
nanoparticle production: How economic is dry
synthesis? Journal of Nanoparticle Research 8(1):
1–9.
Park, C. and R. T. K. Baker. 1999. Catalytic behavior
of graphite nanofiber supported nickel particles:
III. The effect of chemical blocking on the performance of the system. Journal of Physical Chemistry
B 103(13): 2453–2459.
PRé Consultants. 2001. The Eco-Indicator 99: A
damage oriented method for life cycle impact
assessment. http://www.pre.nl/download/EI99_
methodology_v3.pdf. Accessed February 2008.
Rodriguez, N. M. 1993. A review of catalytically grown
carbon nanofibers. Journal of Materials Research
8(12): 3233–3250.
Rodriguez, N. M., A. Chambers, and R. T. K. Baker.
1995. Catalytic engineering of carbon nanostructures. Langmuir 11(10): 3862–3866.
Royal Society. 2004. Nanoscience and nanotechnologies: Opportunities and uncertainties. www.
nanotec.org.uk/finalReport.htm. Accessed October 2007.
Sandler, J., A. H. Windle, P. Werner, V. Altstädt,
M. V. Es, and M. S. P. Shaffer. 2003. Carbonnanofibre-reinforced poly(ether ether ketone) fibres. Journal of Materials Science 38(10): 2135–
2141.
Seager, T. P. and I. Linkov. 2008. Coupling multicriteria decision analysis and life cycle assessment for nanomaterials. Journal of Industrial
Ecology 12(3) doi: 10.1111/j.1530-9290.2008.
00048.x.
Şengül, H., T. L. Theis, and S. Ghosh. 2008. Toward sustainable nanoproducts: An overview
Khanna et al., Energy Consumption and Environmental Impact From Carbon Nanofibers
409
R E S E A R C H A N D A N A LY S I S
of nanomanufacturing methods. Journal of
Industrial Ecology 12(3) doi: 10.1111/j.15309290.2008.00046.x.
Shatkin, J. A. 2008. Informing environmental decision making by combining life cycle and
risk analysis. Journal of Industrial Ecology 12(3)
doi: 10.1111/j.1530-9290.2008.00031.x.
SimaProTM . 2006. http://www.pre.nl/simapro. Accessed October 2006.
Takehira, K., T. Ohi, T. Shishido, T. Kawabata, and
K. Takaki. 2005. Catalytic growth of carbon fibers
from methane and ethylene on carbon-supported
Ni catalysts. Applied Catalysis A: General 283(1–
2): 137–145.
Tibbetts, G. G. and D. W. Gorkiewicz. 1993. A new
reactor for growing carbon fibers from liquid- and
vapour-phase hydrocarbons. Carbon 31(5): 809–
814.
Tibbetts, G. G., C. A. Bernardo, D. W. Gorkiewicz, and
R.A. Alig. 1994. Role of sulphur in the production
of carbon fibers in the vapor phase. Carbon 32(4):
569–576.
Williams, E. D., R. U. Ayres, and M. Heller. 2002. The
1.7 kilogram microchip: Energy and material use
in the production of semiconductor devices. Environmental Science and Technology 36(24): 5504–
5510.
410
Journal of Industrial Ecology
WWICS (Woodrow Wilson International Center for
Scholars) Project on Emerging Nanotechnologies. 2006. Nanotechnology: A research strategy
for addressing risk. http://www.nanotechproject.
org/process/files/2707/77_pen3_risk.pdf. Accessed January 2008.
Yang, R. T., and J. P. Chen. 1989. Mechanism of carbon filament growth on metal catalysts. Journal of
Catalysis 115(1): 52–64.
About the Authors
Vikas Khanna is a Ph.D. student in the
Department of Chemical and Biomolecular
Engineering at The Ohio State University in
Columbus, Ohio. Bhavik R. Bakshi is a professor
in the Department of Chemical and Biomolecular Engineering and codirector of the Center
for Resilience at The Ohio State University. L.
James Lee is the Helen C. Kurtz Professor in
the Department of Chemical and Biomolecular
Engineering and director of the National Science Foundation Center for Advanced Polymer
and Composite Engineering at The Ohio State
University.