Renewable Energy 160 (2020) 428e449

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Techno-economic analysis and life-cycle assessment of jet fuels

production from waste cooking oil via in situ catalytic transfer
hydrogenation
Elena Barbera a, *, Rustem Naurzaliyev a, Alexander Asiedu b, Alberto Bertucco a,
Eleazer P. Resurreccion c, Sandeep Kumar b
a
Department of Industrial Engineering DII, University of Padova, Via Marzolo 9, Padova, 35131, Italy
b
Department of Civil and Environmental Engineering, Old Dominion University, Norfolk, VA, 23529, United States
c
Department of Civil Engineering Technology, Montana State University-Northern, Havre, MT, 59501, United States

a r t i c l e i n f o a b s t r a c t

Article history: This work evaluates the feasibility of renewable jet-fuel production from waste cooking oil (WCO) via
Received 16 October 2019 catalytic transfer hydrogenation (CTH) using isopropanol as hydrogen donor. Results were compared to a
Received in revised form commercial hydroprocessed renewable jet (HRJ) fuel technology, employing process simulation-based
23 April 2020
techno-economic analysis (TEA) and life-cycle assessment (LCA). The two routes were compared in
Accepted 15 June 2020
Available online 15 July 2020
terms of product yield, energy consumption, economic and environmental metrics, and allocation
methods. The total capital expenditure of CTH plant (7.3M$) was significantly lower than that of HRJ
($149.7M$). The annual revenues were comparable (~150M$/year), due to similar fuel yields. To be
Keywords:
Renewable jet fuels
profitable, the liquid fuel should be sold at $3.00/gal and $1.67/gal for CTH and HRJ, respectively. The
Waste cooking oil cumulative fossil energy demand (CED) of HRJ was 1.6 times that of CTH and the total 100-year GWP of
Process simulation CTH was 8% less than HRJ’s, with both systems not sequestering CO2 through co-product offsets. Mass-,
Catalytic transfer hydrogenation energy-, and market-value allocations were utilized. Sensitivity analysis indicated that both systems
were driven by transportation factors and not process inputs. Trend analysis on CTH’s energy-return-on-
investment (EROI) showed that wide improvements could be made in energy efficiency (EROI ¼ 10.30
e11.30). From an investment/construction perspective, CTH (95% cheaper) appears to outperform HRJ at
similar revenues.
© 2020 Elsevier Ltd. All rights reserved.

1. Introduction approximately 45% in jet fuels consumption is expected by 2040

[2]. About 278 billion liters of jet fuels were consumed by com-
The progressive depletion of fossil fuels, along with its envi- mercial operators in 2016, leading to 781 million tons of CO2
ronmental impacts, has driven the scientific community to search emissions [3].
for sustainable and renewable energy sources. The search has been While other energy sources (e.g. electricity, fuel cells) are being
compelled by the burgeoning global population at 0.9% annual rate, investigated for road transportation, the aviation sector relies on
reaching over 9 billion by 2050, resulting to 29% increase in world liquid fuels with strict quality requirements to power gas-turbine
energy consumption [1]. One of the highest energy-consuming engines [4]. These are kerosene-range hydrocarbons with C8eC16
sectors is transportation. In the U.S., 92% of transportation fuels carbon atoms. These hydrocarbons can be obtained from renewable
are petroleum-derived, while only 5% are biofuels obtained from sources such as triglycerides. Among the numerous fatty acid
renewable sources [2]. Among transportation fuels, the develop- sources (mainly animal fats and vegetable oils), waste cooking oil
ment of bio-based jet fuels is of tremendous interest. According to (WCO) is particularly promising. Unlike virgin oils, WCO is not used
the U.S. Energy Information Administration, an increase of in food applications and is also a cheaper feedstock. It is largely
available in the U.S. and around the globe, with an annual global
production of 29 million tons [5]. Moreover, WCO is already largely
used in the U.S. for biodiesel production, so that a well-developed
* Corresponding author.
E-mail address: [email protected] (E. Barbera). transportation and supply infrastructure already exists [6].

https://doi.org/10.1016/j.renene.2020.06.077
0960-1481/© 2020 Elsevier Ltd. All rights reserved.
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 429

Several technologies have been investigated for the conversion fats, and greases, can be processed using the UOP-Eni Ecofining. The
of fatty acids into jet fuels [4,7]. These involve oxygen removal from first reactor, the HDO reactor R1, is fed with the feedstock and
the fatty acids via hydrodeoxygenation (HDO) or decarboxylation/ hydrogen to produce n-alkanes [16,17]. The nature of the product is
decarbonylation reactions using hydrogen. Depending on the na- dictated by the feedstock’s fatty acid composition, catalyst type,
ture of catalyst used, its support, and reaction conditions such as and operating conditions. The by-products are separated from the
temperature and pressure, one reaction pathway can be favored n-alkanes in a flash drum, operated at the reaction pressure. The
over another [8]. Most commonly used catalysts are sulfided gaseous stream is a mixture of unreacted hydrogen, CO2, and CO.
bimetallic materials such as Ni-Mo or Co-Mo, or noble metals like Pt The liquid is an immiscible mixture of organic liquid and water. In
or Pd [9]. Veriansyah et al. [10] investigated the hydroprocessing the second reactor, the hydroisomerization/hydrocracking (HH)
conversion of soybean oil on different supported catalysts, and reactor R2, the liquid product is isomerized and cracked in the
found that the most efficient one was NiMo/geAl2O3. Clearly, the presence of hydrogen [7,15]. This step is necessary to obtain a
desired catalyst characteristics are high stability and low cost. For kerosene boiling range and jet fuel requirements. Light fuels and
this reason, due to the high cost of noble metals, extensive effort residual diesel (if any), are the valuable by-products. The latter
has been directed towards the development of transition metals, might be recycled to increase jet fuel production or directly sold to
and particularly Ni-based catalysts [11]. These catalysts have the market.
reached a mature level, and are already widely employed Despite its maturity level, HRJ has drawbacks associated with
commercially for hydrodesulphurization (HDS) processes [8]. In the use of hydrogen gas. First, H2 is poorly miscible with oil at at-
particular, Ho
cevar and co-workers reported that the sulfided form mospheric conditions, thus, 25-100 bar pressure is required to
appeared the most active compared to the oxide and reduced enhance mass transfer. This results in high capital and operating
(metal) forms in the HDO of lignocellulosic derived compounds costs. Second, storage and transportation of hydrogen increases
[12,13]. On the other hand, the catalyst support also plays a major cost and creates safety issues due to its high reactivity and flam-
role in the catalyst performances [14]. mability. Third, the production of hydrogen in refineries relies on
The hydrotreating process for biojet fuels production from fossil sources resulting into large CO2 emissions.
triglyceride-based feedstock, known as hydroprocessed renewable This study addresses the three issues associated with HRJ by
jet (HRJ) or hydroprocessed esters of fatty acids (HEFA), is in fact a employing catalytic transfer hydrogenation (CTH) as an alternative
mature technology and has already been commercialized. For process to convert WCO into jet fuels (Fig. 1B.) [18]. CTH utilizes
example, Honeywell’s UOP, in collaboration with ENI has success- hydrogen-donating compounds that, in the presence of a catalyst,
fully commercialized the Ecofining™ process that converts vege- release H2 to saturate and reduce triglyceride compounds in oils.
table oils into green jet fuels by deoxygenation [7,15] as shown in Suitable hydrogen-donor compounds typically contain hydrogen
Fig. 1A. Accordingly, the distillate products are obtained from a bonded to elements or groups with similar electronegativity to that
series of two-reaction system with an intermediate separation of of hydrogen itself. Therefore, organic acids, alcohols, amines, and
by-products. Various feedstock types, such as vegetable oils, animal hydrocarbons are suitable hydrogen donors in CTH [19]. Using

Fig. 1. Block-flow diagram of HRJ (A) and CTH (B) processes for the conversion of triglyceride into biojet fuels.
430 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

hydrocarbons as in situ hydrogen donors gives the advantage that 2.1. Waste cooking oil feedstock
the bond energy of C-H is lower than that of the H-H bond in H2 [20].
However, most of these compounds are costly, toxic and volatile. The The fatty acids composition of the WCO feedstock considered is
most available, less toxic and less expensive solvents are common reported in Table 1 [18]. The number of unsaturated bonds in each
alcohols (methanol, ethanol, n-propanol and isopropanol), which fatty acid is specified after the length of the carbon chain. Roughly
have been reported to be effective hydrogen donors [21]. Methanol 70 wt% of the oil was composed of unsaturated fatty acids (oleic and
has been used to hydrogenate styrene and nitrobenzene [22], while linoleic acids), while the remaining compounds were saturated
ethanol and 2-propanol were used to deoxygenate levulinic acid carbon chains. It should be noted that typical waste cooking oils are
over 5% Pd/C [23]. Among these common alcohols, primary alcohols composed of 80e96 wt% triglycerides (TG), 2-7 wt% diglycerides
are generally less active than secondary alcohols. The improved (DG), 0.3e2 wt% monoglycerides (MG), and free fatty acids (FFA)
hydrogenation ability of secondary alcohols has been attributed to contents from 1 to 16 wt% [26e29].
the highest reduction potential of the corresponding ketones or the
smaller electron-releasing inductive effect of one alkyl group 2.2. Hydroprocessed renewable jet (HRJ) process
compared to two alkyl groups [24]. For example, the two methyl
groups in 2-propanol donate more electrons to weaken the O-H The process flowsheet for the HRJ process was built in Aspen
bond compared to the lower electron-donating ability of one methyl Plus, based on the UOP commercial process (Section 1) as shown in
group and one ethyl group in methanol and ethanol, respectively. Fig. 2. The corresponding material and energy balances resulting
Despite the use of a variety of alcohols, isopropanol remains the most from the simulations are reported in Appendix A (Table A1). The
popular donor, because of its simplicity, cheapness, availability, and WCO feed was pumped and heated up to the operating conditions
ease of removal from reaction systems. of the HDO reactor R1 (HDOR1) together with H2, which was fed at
Therefore, based on the literature, our preliminary studies, and 2.6 wt% ratio with respect to WCO [30]. The HDO unit was modelled
the possibility of generating this solvent from renewable feedstock, as a stoichiometric reactor RStoich. For each saturated fatty acid, the
we proposed to use and compare the effectiveness of using iso- stoichiometric reactions of decarboxylation, decarbonylation and
propanol as in situ H2 source for jet fuels production from WCO. To hydrodeoxygenation were defined according to:
the best of our knowledge, isopropanol has not been used for hy-
drogenation or deoxygenation of WCO. Hence, in this study, cata- Cn H2n O2 / Cn1 H2n þ CO2 (1)
lytic transfer hydrogenation of WCO to produce jet fuels was
analyzed using isopropanol as an H-donor in a continuous-flow Cn H2n O2 þ H2 /Cn1 H2n þ CO þ H2 O (2)
reactor over a fixed bed of activated carbon. The experimental
study has been previously published by our group [18], proving that Cn H2n O2 þ 3H2 /Cn H2nþ2 þ 2H2 O (3)
the process was effective in producing jet/diesel range fuels. With
this work, we aimed at carrying out a detailed, process simulation- For the unsaturated fatty acids, additional hydrogen was
based techno-economic analysis (TEA) and life-cycle assessment consumed to hydrogenate the double bonds. The detailed stoichi-
(LCA) of the proposed CTH process at industrial-scale and ometry is reported in Table S1. The operating conditions and fatty
compared its market and environmental performance with that of acids conversions for the three reactions were taken from the
commercial HRJ process. The novelty of this work lies in the fact works of Chu et al. [30] and Veriansyah et al. [10], which analyzed
that CTH of WCO was performed at nearly atmospheric pressure the conversion of vegetable oil into fuels at 400  C, 92 bar, and 2-h
conditions, which cannot be achieved using gaseous hydrogen. reaction time over nickel-molybdenum catalyst. Accordingly, the
Moreover, detailed techno-economic and environmental perfor- extent of reaction was set equal to 0.68 for decarboxylation, 0.03 for
mance data, which is limited in scientific literature for this novel decarbonylation and 0.29 for hydrodeoxygenation, with overall oil
process, were generated and reported. conversion equal to 1. The hydrogen consumption associated with
the cleavage of triglycerides backbone (3 mol of H2 for each tri-
2. Methods glyceride producing the corresponding free fatty acids) and the
corresponding formation of 1 mol of propane was also considered
Process simulations of renewable jet fuels production from WCO (Table S4), according to the reaction pathway proposed by the same
by direct hydrogenation (HRJ via pressurized hydrogen) and CTH authors [10,30].
were performed using Aspen Plus® software v.9. The Peng- The gaseous products (CO, CO2, H2 and H2O) might give rise to
Robinson equation of state and the Non-Random Two-Liquid Water Gas Shift (WGS) and Methanation reactions. However, Ver-
(NRTL) activity coefficient model were selected as thermodynamic iansyah et al. [10] reported that no methanation was observed over
methods. The former was selected as it is suitable to describe the Ni-Mo catalyst under the experimental conditions investigated. To
behavior of hydrocarbon systems in high-pressure conditions, allow for possible WGS, a REquil reactor unit (WGS-HDO) was
while NRTL was used to model specifically the low-pressure vapor- added to calculate the gas phase equilibrium reaction (Eq. (4)),
liquid-liquid equilibrium in the products separation following CTH which was evaluated at the same operating conditions (400  C and
reaction (see section 2.3). The goal of this study was to perform a 92 bar).
comparative TEA and LCA (attributional) between HRJ and CTH
CO þ H2 O 4H2 þ CO2 (4)
process. For the TEA, a basis of 1000 ton/day WCO feedstock was
employed because this represents the throughput of a small-to
medium-scale refinery in the U.S [25]. The two processes were
Table 1
compared in terms of economic metrics (internal rate of return
Fatty acids composition of waste cooking oil feedstock.
(IRR) and net present value (NPV)) and environmental impacts
(cumulative fossil energy demand (CED) and life cycle greenhouse Compound Formula wt%
gas (GHG) emissions). All the assumptions and methods used in the C15:0 Pentadecanoic acid C15H30O2 17.9
analysis are described in the succeeding sections. The TEA consid- C19:0 Nonadecanoic acid C19H38O2 11.3
ered a WCO feed flow rate of 1000 ton/day and the LCA adapted a C18:1 Oleic Acid C18H34O2 18.6
C18:2 Linoleic Acid C18H32O2 52.3
1 MJ of useable energy produced embodied in liquid biojet fuel.
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 431

Fig. 2. Process flowsheet for the HRJ process. Black solid lines indicate material flows while gray dashed lines indicate energy/heat flows.

A high-pressure flash was used to separate the liquid and The liquid fuels mixture was then sent to an atmospheric frac-
gaseous products, which were cooled to 40  C, making the water tionation tower (FRAC) to recover the different product fractions,
and hydrocarbons completely immiscible. Hence, a 3-phase adia- i.e. naphtha and kerosene. Distillation was modelled as a RadFrac
batic flash (HIGH-P-F) was used to achieve complete separation of unit, operating with 30 ideal stages and reflux ratio of 0.8.
water, organic liquids, and gases (Table S6). The large excess of H2 recovered from the HIGH-P-F and HP-SEP
The organic liquid products were then sent to a hydrocracking flash units is valuable, so that high recovery from the gaseous
unit (HCC) that cracked and isomerized the paraffins to obtain the streams was desired to maximize recycle. The gas mixture separated
desired product mixture quality at 350  C and 90 bar [31,32]. from the first flash unit was characterized by large amounts of CO
Because of the complexity of modelling hydrocracking reactions, and CO2, while the one recovered after the hydrocracking unit
the reactor was simulated with a RGibbs model, considering a total mostly comprises light hydrocarbons, such as CH4, C2H6 and C3H8.
of 358 hydrocarbon components from the Aspen Plus database, Industrially, the recovery of hydrogen is performed by Pressure
including all isomers ranging from C1 to C18 carbon atoms. The Swing Adsorption (PSA), with operating pressures varying between
validity of this approach was verified by checking the product 10 and 40 bar [33,36]. As a higher recovery pressure results in lower
yields (LPG, naphtha, kerosene and diesel ranges) with those re- compression loads required for H2 recycling, 40 bar was chosen for
ported in the literature [30,33e35]. The amount of H2 required for the simulations. For the purposes of the simulations, the PSA units
hydrocracking was evaluated based on the properties of the feed, were modelled by simple SEP blocks assuming complete H2 recov-
namely  API gravity and Watson factor (Table S7), and was equal to ery. The H2 make-up stream was compressed up to 40 bars by means
84.44 Nm3 of H2 per m3 of oil (Table S8) [35]. Hydrogen is typically of a 3-stage compressor (C2, compression ratio ¼ 3.14, which
fed in large excess to absorb the heat of reaction by direct quench at considered the final temperature of each stage, not exceeding 225  C
different stages of the reactor. Therefore, a higher amount of inert for H2), with inter-stage cooling at 40  C. The make-up and recycled
hydrogen is required in the reactor depending on the heat pro- H2 streams were then mixed and compressed up to the final pressure
duced. The amount of hydrogen required for quench was estimated of 92 bar in a second single-stage compressor (C1) and delivered to
through an energy balance to keep the difference between inlet and the HDO and HCC reactors respectively.
outlet of reactor in the maximum range of 25  C (assuming adia- Finally, pinch analysis was performed in order to optimize heat
batic operation). This amount of H2 was set as inert fraction in the integration (Section 1.6 of SI). The optimized heat exchanger net-
RGibbs unit, so it was not accounted for in the equilibrium calcu- works are displayed in the process flowsheet (Fig. 2), where inte-
lation. HCC process input parameters in Aspen Plus are shown in grated heat exchangers are connected by energy streams (dashed
Table S9. lines). Accordingly, the external process utilities needed consisted
The products from the HCC reactor were separated by means of of a fired heater to achieve 400  C prior to the HDO reactor, and
two flash units in series. The first one operated at high pressure cooling water to cool the products at 40  C in the units CW1 and
(HP-SEP) to recover the unreacted hydrogen, while the second one CW2. The column reboiler and condenser duties were also provided
operated at low pressure (LP-SEP) to separate C1-C4 gaseous by external utilities.
products from the liquid mixture, which was to be sent to an at-
mospheric fractionation tower to recover the different product 2.3. Catalytic transfer hydrogenation (CTH) process
fractions. The higher the operating pressure of the LP-SEP unit, the
higher the recovery of liquid fuels (Table S14). However, it was The process flowsheet of the CTH process is shown in Fig. 3, and
preferred to operate the flash at 1 bar so that a higher amount of the corresponding material/energy balances are reported in
propane C3H8 (94%) was recovered in the gaseous stream, to be Appendix A (Table A2). The simulation was based on the block flow
used as additional fuel. diagram of Fig. 1B and on the experimental data obtained in the
432 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

laboratory using a fixed-bed tubular reactor filled with charcoal Table 2

catalyst [18]. The WCO feedstock was mixed with isopropanol (ISO- Mass-, energy-, and market-based equivalency (impact) factors used in the impacts
assessment phase of HRJ and CTH life cycle. Factors were calculated from Table S69.
P) at a flowrate of 341.8 ton/day (ISO-P/WCO ratio ¼ 0.3418), and
heated up to the CTH reactor operating temperature (380  C), found Outputa Mass Energy Markete
to be the optimal value [18]. The CTH operating pressure was set to HRJ Products 0.8176 0.8109 0.0254
2 bar, and the reaction was simulated using the RYield model. The
Naphtha 0.2934 0.2993 0.0098
product yields, defined as mass of product over inlet mass of re- Jet fuel 0.5242 0.5116 0.0156
agents, were calculated based on the experimental material bal-
HRJ Co-products 0.1824 0.1891 0.9746
ance results (Table S24).
The gaseous products considered were H2, H2O, CO, CO2, CH4, Propane gas 0.0045 0.0047 0.0675
C1eC3 gasb 0.1070 0.1146 0.9043
C2H6 and C3H8, which were formed either by decarboxylation, C4eC8c 0.0568 0.0569 0.0022
decarbonylation, or hydrodeoxygenation reactions as well as by C9eC15d 0.0013 0.0013 0.0000
isopropanol decomposition. Concerning the liquid products, for the C1eC3 liquid (LPG) 0.0082 0.0087 0.0004
sake of simplicity, only three reference components were taken as C16eC18 (diesel) 0.0029 0.0028 0.0001
Liquid CO2 0.0017 0.0000 0.0000
representatives for the naphtha, kerosene and diesel range hydro-
carbons, respectively. Based on the average properties (boiling CTH Products 0.7899 0.7867 0.0167
point and molecular weight) of the experimental products distri- Naphtha 0.2428 0.2495 0.0056
bution, C8H16, C12H26, and C17H36 were selected as the three Jet fuel 0.4200 0.4129 0.0086
respective liquid fuel fractions (Table S18-S23). Diesel 0.1270 0.1243 0.0026
Because the reaction leads to coke formation, catalyst regener- CTH Co-products 0.2101 0.2133 0.9833
ation was included by simulating the coke removal using a SEP unit. Coke 0.0307 0.0201 0.0006
The composition of the outlet stream from CTH reactor was char- C1eC3 gasb 0.1687 0.1821 0.9823
acterized by a mixture of hydrocarbons (C8H16, C12H26, C17H36), Octane 0.0061 0.0062 0.0003
water (H2O) and gaseous products (CO2, CO, C1-C3). The separation Dodecane 0.0001 0.0001 0.0000
C1eC3 liquid (LPG) 0.0046 0.0049 0.0002
of hydrocarbon mixture from gases and water could be achieved at
a
low temperature, since the solubility of water is inversely depen- Output flow rates obtained from mass balance (tons/day).
b
Represented by ethane gas.
dent on temperature. Accordingly, a 3-phase flash model was used c
Represented by compressed liquid hexane.
to simulate the required separation (VLL-SEP) at atmospheric d
Represented by compressed liquid dodecane.
pressure, after cooling the products to 25  C. For a correct estima- e
Obtained based on wholesale market price in 2019 US dollars excluding taxes.
tion of the vapor-liquid-liquid equilibrium at low pressure, NRTL
was used as thermodynamic model. The solubility of the incon- Heat integration was also carried out between the cold feed and
densable compounds (CO, CO2 and H2) was determined from the the hot products, as shown by the energy stream (dashed line) in
Henry constants, available in the Aspen Plus database. The validity Fig. 3. By analyzing the vaporization range of the WCO-isopropanol
of the binary interaction parameters of the H2O-hydrocarbon sys- feed, the feed outlet temperature from the integrated heat
tems was first verified against experimental data (Table S26-S27). exchanger was set to 300  C. As most of the WCO vaporizes in the
Finally, the three liquid fuels fractions were separated by means of range of 300e350  C (Fig. S8), it was preferred to avoid the phase
two atmospheric distillation columns in series: the first one change within the heat exchanger, and to supply the latent heat
(FRAC1) separated the naphtha from the heavier fractions, which duty by means of a fired heater. Accordingly, the external utilities
were then separated in the second one (FRAC2). The columns were represented by the fired heater required to reach 380  C and
specifications were, respectively: 10 ideal stages and reflux refrigerated cooling water (available at 5  C) to cool down the
ratio ¼ 0.2 for the first separation; 15 ideal stages and reflux products to 25  C, in addition to the reboiler and condenser duties
ratio ¼ 0.2 for the second one. of the two distillation columns.

Fig. 3. Process flowsheet for the CTH process. Black solid lines indicate material flows while gray dashed lines indicate energy/heat flows.
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 433

2.4. Techno-economic analysis Oil/Molten Salts for reactor cooling [44]. More specifically, the cost
was based on the price of natural gas that would be required to heat
For a 1000 ton/day WCO feed flow rate (approximately 270,000 the hot oil/molten salts. The number of operators was calculated
gal/day), the economic and profitability analysis of both HRJ and based on the number of equipment units and was equal to 29 for
CTH processes were performed according to the method proposed the HRJ and 26 for the CTH processes. Indirect manufacturing costs
by Towler et al. [37] for chemical plants. Accordingly, the Fixed and general expenses were instead estimated as percentages of the
Capital Investment (FCI) and the Cost of Manufacturing (COM) were FCI [35,37].
estimated together with the revenues. The FCI consists of the The plant was assumed to operate for 8000 h/year, i.e. a stream
capital expenses of the plant including construction and engi- factor of 0.913 was applied. The economic profitability of the pro-
neering costs (CAPEX), contingency charges, and offsite infra- cesses was evaluated by means of discounted cash flow analysis,
structure investment (OSBL) (Section 5.1.1 of SI). The CAPEX were according to four criteria: discounted payback period (DPBP), net
estimated based on the Bare Module Cost of each equipment unit, present value (NPV), internal rate of return (IRR), and present value
evaluated using the correlations proposed by Guthrie [38]. Ac- ratio (PVR). The following additional assumptions were made [45]:
cording to this method, the purchase cost of the equipment was first, the useful life of the plant was taken equal to 25 years, the first
calculated as a function of the size, as well as the construction 2 of which were used for construction and start-up, while full
material and the operating pressure, which become particularly regime production started at year 3. In particular, 70% of the capital
relevant for the HRJ process, characterized by high pressures. was invested at year 1, while the remaining 30% in year 2. Second,
Furthermore, the material chosen for HDO and HCC reactors were the discount rate was taken as an interest rate of 8%, hence, it is not
stainless steel (SS) because of hydrogen gas in compatibility with based on loan interests or debt ratio. Third, depreciation was
respect to the normal carbon steel (CS) [39]. Moreover, to account evaluated according to the Modified Accelerated Cost Recovery
for inflation, all the costs were referred to the year 2017 using the System (MACRS). The depreciation period was 7 years, starting at
Chemical Engineering Plant Cost Index (CEPCI) [40]. the beginning of the 3rd and ending at the end of the 10th year. In
A rigorous sizing procedure was applied to determine the size of particular, the double declining balance was applied for the first 5
the HDO and HCC reactors for the HRJ process, and that of the CTH years after the plant start-up, after which it was switched to the
reactor. A Trickle Bed Reactor (TBR) composed of adiabatic multi- straight-line method. No salvage value of the plant was assumed.
stage beds was selected as HDO reactor [41]. Considering the Thus, the total capital to be depreciated equaled the fixed capital
properties of the feed, catalyst hold-up of 60%, and liquid residence investment (FCI). Fourth, the income tax rate was assumed to be
time of 2 h, the resulting reactor was determined to have 4 stages, 35% of the gross profit. Fifth, the cost of the land was not consid-
each of diameter D ¼ 3.048 m and H ¼ 8.39 m, for a total volume of ered, as it was not relevant for the sake of comparison between the
245 m3 (Section 4.1.1 of SI). The HCC reactor was sized as a multi- profitability of the two processes. Finally, annual operators’ salary
stage packed bed reactor, such as those used at commercial level. was taken from U.S. Bureau of Labor Statistics [33].
Taking a typical gas residence time of 1 h, the reactor volume The revenues were evaluated based on the wholesale prices of
equaled 392.5 m3, with 5 stages of D ¼ 3.81 m and H ¼ 6.9 m [42] the refinery products, which included the production costs and
(Section 4.1.2 of SI). Finally, a packed bed reactor was selected also refiner profit, while distribution and transportation costs and taxes
for the CTH reactor. The scale-up was based on the experimental were not considered. The wholesale prices corresponding to the
value of the weight hourly space velocity (WHSV) of 6.8 h1. The different refinery products (naphtha, kerosene, and diesel) were
resulting volume, assuming 60% catalyst hold-up, was equal to retrieved from the U.S. Energy Information and Administration
27.4 m3 (Section 4.2.1 of SI). Given the high coke formation, 2 [46], and were equal to 0.426 $/L, 0.443 $/L, and 0.440 $/L,
parallel reactors were considered, in order to allow a continuous respectively, as average prices for 2017. The large amount of fuel
operation when performing catalyst regeneration. The sieved tray gases produced by both the HRJ and CTH process was assumed to be
distillation columns were sized according to the Fair Method [43]. sold at the same price as natural gas (0.13 $/m3), considering that
In particular, the distillation tower of the HRJ process resulted the lower heating value (LHV) of these gases was very similar to
having a diameter D ¼ 1.8 m and H ¼ 15.7 m, which allowed that of natural gas.
operating at 74% of the flooding condition. The diameters and the
heights of columns of the CTH process were H ¼ 17.5 m, D ¼ 1.35 m 2.5. Life cycle assessment
and H ¼ 17.5 m, D ¼ 1.7 m, respectively (approximately 78% of the
flooding). In addition, the capital cost related to building an internal 2.5.1. Goal and scope
H2 production facility by Steam Methane Reforming (SMR) was The comparative environmental impacts of HRJ and CTH processes
considered for the HRJ process [37]. The capital cost of the PSA unit using WCO as feedstock were evaluated in terms of CED and life cycle
was instead determined using a capacity factor with respect to a GHG emissions. It was assumed that one biorefinery was located in
similar reference unit [36]. According to the simplified procedure any medium-sized U.S. city (population: 100,000e300,000 persons)
employed, the OSBL were estimated as 50% of the total CAPEX (with [47]. A biorefinery serves one division for each U.S. region (e.g. one
reference to atmospheric pressure purchase costs), while contin- biorefinery for Division 2 [Mid-Atlantic] of Region 1 [Northeast],
gencies amount to 15% of the actual CAPEX [37]. which includes New Jersey, New York, and Pennsylvania). WCO was
The COM considers both direct and indirect manufacturing assumed to be collected from restaurants, hotels, casinos, malls, and
costs, as well as general expenses related to business administra- other food providers operating in these cities. The volume of WCO
tion. Direct manufacturing costs are directly correlated with the generated from household sources was insignificant relative to the
production rate and comprise raw materials and utilities or oper- food service industry; restaurants still dominate total spent oil output.
ations expenses (OPEX), as well as staff and operating labor. The These cities are ideal test locations for a WCO-based refinery for the
OPEX were evaluated based on the results of the process simula- following reasons. First, they are affordable, offering abundant op-
tions. For the estimation of natural gas utility in the fired heater portunities for business investments. Second, these cities have
unit, an efficiency of 80% of the thermal system was assumed. numerous restaurants that can provide appreciable amount of WCO.
Moreover, the utility used for cooling the CTH reactor was hot oil or They have high average number of restaurants per 1000 people
molten salts due to the high operating temperature, i.e. 380  C. The compared to the national average of 1.52 for cities whose population
price of this utility was estimated based on the reference for Hot is greater than 50,000 [48]. Third, these cities are not as busy as
434 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

megacities which cuts transportation costs yet were assumed to be 9.2 MPa and 2 h of residence time [14,15]. A four-stage HY unit was
located within 150-mile maximum radius from a Metropolitan Sta- utilized in this sub-system. Material inputs included pretreated
tistical Area (MSA) (>1 million population) to ensure reliable supply of crude oil, H2 as hydrogenating agent, and nickel-molybdenum
spent oil. Examples of these medium-sized U.S. cities are Pittsburgh (NiMo) catalyst. The output from HY was a mixture of straight-
(2019 population: 302, 908) serving Division 2, Region 1; Greensboro, chain alkanes consisting of tetradecane, pentadecane, heptade-
North Carolina (2019 population: 292,265) serving Division 5, Region cane, and nonadecane. Moreover, gases such as CO, CO2, and pro-
3; or Fremont, California (2019 population: 238, 281) serving Division pane were generated in addition to water. The amount of total H2
9, Region 4 [36]. Based on the 270,000 gal/day (98.55 Mgal/year required and propane generated were based on stoichiometry
continuous operation) TEA basis, the nine biorefineries in the U.S. will presented in Table S4. The energy consumption for HY was attrib-
require 8.87  108 gal/yr. The U.S. generates more than enough WCO uted to: (1) pressurization assuming isentropic compression of H2
at 3.8 Mton WCO annually (9.92  1011 gal/yr) [49], with 2.84 Mton at 85% efficiency and (2) WCO pumping at 95% efficiency. Heat was
yellow and other grease [50]. utilized to raise the oil temperature from 380 to 400  C, determined
The system boundary for the analysis was “well-to-tank” (WTT), as the heat duty to be supplied resulting from heat integration.
in which all associated energy and material flows were determined HRJ Sub-System 3: CO and CO2 Flash Separation (FS). Flash sepa-
and quantified for each unit operation: hydrogenation/hydro- ration was performed to separate the gaseous by-products from the
cracking/separation for HRJ and hydrogenation/separation for CTH. output liquid straight chain alkanes. Separation input parameters
Fuel combustion was excluded in the analysis. The WTT system for FS were as follows: 9.2. MPa, 10 min liquid hold-up [23], adia-
boundary starts with the acquisition of all raw materials up to batic operation, and vapor-liquid-dirt water phases, employing the
distribution of products to end-users (e.g. airports). For ease of Peng-Robinson property method in Aspen Plus modeling. A single
analysis, HRJ was modelled as the base case using the UOP-Eni vertical-type separation vessel was used. These gases include CO,
Ecofining™ process. CTH was considered as the alternative case. CO2, water vapor, propane, and H2. Along with liquid alkanes, sour
water was generated. Energy requirement for this sub-system was
2.5.2. Functional unit attributed to the lowering of product temperature from 231 to 40  C
The functional unit (FU) chosen for this analysis was 1 MJ/year (after heat integration).
embodied energy in liquid biojet fuel (main product) produced HRJ Sub-System 4: Hydroisomerization and Hydrocracking (HH).
from either HRJ or CTH process using WCO as feedstock. Such FU is The separated liquid straight chain alkanes were isomerized/
conveniently adapted to provide a quantitative reference on which cracked to get smaller chains of iso-paraffines. Mild conditions
all calculated materials and energy flows were based. Co-products were employed in this sub-system: 350  C and 9.0 MPa at 1-h HH
were light fuels and naphtha from HRJ and naphtha and diesel from reaction time. Detailed process input parameters used in Aspen
CTH. Plus modeling are tabulated in Table S9. The hydroisomerization
and hydrocracking unit comprised 5 stages. Properties of feed oil
2.5.3. System boundary and hydrogen are presented in Table S7 and Table S8, respectively.
The HRJ process was divided into eight sub-systems which Material inputs to this sub-system include the alkanes, H2 for
include WCO transport (WT), hydrodeoxygenation (HY), CO and isomerization and cracking, and H2 for quenching. Output iso-
CO2 flash separation (FS), hydroisomerization and hydrocracking paraffins include N-paraffins (C1-C3, C4-C8, C9-C15, C16-C18) and
(HH), high pressure flash (HF), low pressure flash (LF), distillation their methylated counterparts (mono-x-methyl and mono-yy-
(DT), and fuel transport (FT). Fig. 1A represents hydrotreating as HY, methyl) (naphtha, jet fuel, and diesel). The compression energy
separation as FS, hydrocracking as HH, and products recovery as HF, was already accounted for in the HH unit. All input H2 was previ-
LF, and DT. In the case of CTH, there were six sub-systems identi- ously compressed and then split into the two reactors thereby
fied: WCO transport (WT), catalytic transfer hydrogenation (CH), making pressurization energy of this unit to be 0. Direct heat use
low pressure flash (LF), distillation 1 (DT1), distillation 2 (DT2), and was also 0, due to heat integration (See Section 1.5 of SI). The heat to
fuel transport (FT). These sub-systems are represented in Fig. 1B as: reach the operating temperature was recovered from other
hydrogenation for CH, and separation for LF, and products recovery streams.
for DT1 and DT2. Both HRJ and CTH have identical WT and FT sub- HRJ Sub-System 5: High Pressure Flash (HF). High-pressure flash
systems. was employed to recover H2 and other gaseous compounds from
HRJ Sub-System 1: Waste Cooking Oil Transport (WT). This sub- the liquid products. Herein, the process conditions were: 9 MPa, 10-
system involved the systematic collection of WCO from restau- min liquid hold-up, and no heat duty (adiabatic operation, tem-
rants and food service businesses in cities where a biorefinery was perature after cooling was stable at 40  C). There was one vertical-
to be established. Additional WCO were to be collected from MSA type HF vessel used. Two material streams were produced in this
located within 150-mile maximum radius of the biorefinery to sub-system: (1) the gaseous stream consisting of the recovered H2,
meet the 1000 ton/day biorefinery requirement used in TEA CO, CO2, N-paraffins, and methylated paraffins, and (2) the liquid
(approximately 270,000 gallons/day WCO). WCO were transported stream consisting of naphtha, jet fuel, and diesel. There were re-
from sources near the biorefinery or from MSAs using diesel trucks sidual H2, CO, and CO2 present in the output liquid. Like sub-system
with likeliest truck mileage efficiency of 6.5 miles/gallon [51]. FS, the energy requirement was cooling energy use, or the latent
Diesel Class 8 heavy trucks were assumed to travel for 6 h in a day at heat expelled to lower the product temperature from 196 to 40  C
50 miles/hr. Material input to WT was transportation diesel while (after heat integration) (exothermic).
material outputs were combustion products CO2 and H2O. HRJ Sub-System 6: Low Pressure Flash (LF). Atmospheric sub-
Maximum travel distance per truck was 300 miles and the mileage system LF was employed following HF to separate light hydro-
efficiency was taken to be 5.5 miles/gal [52]. carbons to be used as fuel gas and residual H2, CO, and CO2 from
HRJ Sub-System 2: Hydrodeoxygenation (HY). WCO entered the the liquid fraction in a horizontal vessel. Process conditions for
biorefinery as pretreated spent oil suitable for processing. this sub-system were: 0.10 MPa, 10-min liquid hold-up, and no
Depending upon the nature of fatty acids, conversion can be heat duty (temperature was stable at 24  C). The output liquid
decarboxylation, decarbonylation, or hydrodeoxygenation was heavy with biojet fuel, diesel, and naphtha with very little
(Table S1). HY is adapted as general term for hydrotreatment. The CO2 left. There was no cooling energy requirement for this sub-
chemical reaction was operated at the following conditions: 400  C, system.
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 435

HRJ Sub-System 7: Distillation (DT). Final stage of product recovery 2.5.4. Allocation
was atmospheric distillation, to separate the liquid fractions naphtha Allocation is a critical step in LCA as it directly affects results of
(C6-C8) from the top and jet fuels (C9-C15) from the bottom. DT impacts assessment. Allocation involves the partitioning of the
process conditions were as follows: 0.10 MPa, 30 stages, feed input at environmental impacts from the process based on products (ma-
the 15th stage, 0.80 reflux ratio, 164.8  C reboiler temperature, and terials output) flows [53]. Allocation is not a straightforward divi-
65.30  C condenser temperature. Reflux ratio was optimized by sion of environmental impacts between products and co-products,
sensitivity analysis to maximize the recovery of N-octane (light-key thus, protocols set forth by ISO 14044:2006 guidelines [54] were
component) in the distillate. The output distillate (naphtha) was adapted. For this analysis, mass-, energy- and market-based allo-
heavy in mono-x-methyl (C5-C8) (44 wt%) while the output residue cation methods were necessary. In both HRJ and CTH, the products
(biojet fuels) was predominantly mono-x-methyl (C9-C15) (38 wt%) were identified as naphtha/gasoline (C4-C8, e.g. octane C8H18) and
and multi-yy-methyl (C9-C15) (57 wt%). The energy use for this sub- jet fuels (C9-C15, e.g. dodecane C12H26) obtained at the very last
system came from two sources: the heating duty at the reboiler and sub-system (distillation). In some intermediate sub-systems, co-
cooling energy at the condenser, both were function of WCO feed rate. products were generated. For example, the sub-system HF in HRJ
HRJ Sub-System 8: Fuel Transport (FT). Like WT, FT utilized diesel produced output liquid (gasoline and jet fuel mixture) that fed into
to transport naphtha and biojet fuel to distribution endpoints (i.e. subsequent LF sub-system. Concurrently, it also generated a co-
gate). Designated fuel distributors were assumed to be located product output gas mixture consisting of short-chain N-paraffins
within 150-mile maximum of WCO collection. The trucks were C1-C3 (e.g. ethane, propane), short-chain N-paraffins C4-C8 (e.g.
assumed to have a mileage efficiency of 6.5 miles/gallon. The dis- butane, octane), and mono-x-methyl or multi-yy-methyl C9-C15
tance hauled was calculated to be 300 miles/truck-day at 6 h per (e.g. dodecane, hexadecane or cetane), which can be pressurized
truck per day at 50 miles/hr. Material input to FT was diesel and and distilled into LPG, naphtha, and jet fuel, respectively. In the case
material outputs were CO2 and H2O. of HRJ, the co-products were as follows: FS ¼ propane; HF ¼ [C1-C3,
CTH Sub-System 1: Catalytic Transfer Hydrogenation (CH). CH was C4-C8, C9-C15] gaseous hydrocarbons; LF ¼ [C1-C3, C4-C8, C9-C15]
modelled based on the experimental work of Asiedu et al. [18]. gaseous hydrocarbons, [C1-C3 (LPG), C16-C18 (diesel)] liquid hy-
Hydrogenation and deoxygenation of WCO was accomplished by drocarbons, and liquid CO2. For CTH, the co-products were as fol-
means of a hydrogen donor solvent. Herein, gaseous H2 was lows: CH ¼ coke; LF ¼ [C1-C3, octane, dodecane, heptadecane]
replaced by isopropanol. Details on lab-scale reactor geometry and gaseous hydrocarbons, and [C1-C3] (LPG) liquid hydrocarbons.
catalyst properties are presented in Table S16 and experimental Table 2 details the energy- and market-based allocation factors for
conversion at different temperatures are shown in Table S17. From HRJ and CTH processes. Thus, the CED and life cycle GHG emissions
these results, an industrial model was built based on scale-up of for each process were apportioned between products and co-
experimental reactor and CH reactor material balance at T ¼ 380  C products using mass flows, lower heating values, and unit selling
presented in Fig. S5. The reaction conditions for CH sub-system price at prevailing market conditions.
were as follows: 380  C, 2 bar, 6.8 h1 weight hourly space veloc-
ity in a continuous fixed-bed plug-flow reactor (PFR). There were 2 2.5.5. Inventory
PFRs, each with a total capacity of 27.4 m3, to be operated alter- Life cycle inventory (LCI) is the most tedious process in LCA as it
nately while the other undergoes catalyst regeneration. Activated involves the acquisition of high-quality data essential to accurate
carbon catalyst was assumed to be replaced every 90 days. Material environmental impacts assessment. The development of HRJ and
input streams consisted of WCO, isopropanol as hydrogenating CTH models in this study entailed data from various sources and
agent. The main liquid fuel product contained C1-C3 gases, methods. The models were created and developed using Excel
naphtha, diesel (heptadecane), and biojet fuels. Gases were also spreadsheet with Crystal Ball™ add-in functionality. This allows
generated in the reaction: CO, CO2, and H2, in addition to water and certain material inputs and process conditions to contain proba-
coke. Energy was utilized to pressurize WCO and 2-propoanol to bility distributions (Table S70) that were subsequently used sto-
2 bar and heat was used to raise the feed’s temperature from 300 to chastically for the Monte Carlo simulation, resulting into
380  C after heat integration. probabilistic outputs (“endpoints”). The upstream impacts of
CTH Sub-System 2: Low Pressure Flash (LP). Like HRJ, output liquid electricity and selected materials were calculated using U.S.
from the main chemical reaction (CH) was subjected to LP to equivalency factors (or European where U.S. values were not
separate gases from main products (naphtha, diesel, biojet fuels). available) presented in Table 3.
However, unlike HRJ, CTH has three output streams: gas (mainly CO The hydrodeoxygenation LCI were largely based on the experi-
and CO2), organic liquid (main products), and aqueous liquid (water mental results of Veriansyah et al. [10], appropriately scaled-up to
and residual main products). Conditions were like HRJ’s sub-system 1000 tons/day FU. Gaseous H2 was stoichiometrically supplied to
LF, albeit condition was exothermic; heat was expelled from the crude WCO using with nickel-molybdenum (NiMo) catalyst. The
system to cool the reaction products from 50 to 25  C. The cooling electrical energy for pressurization was determined from Aspen,
down to 50  C was achieved by heat integration. assuming isentropic compression with H2 (85% efficiency) and
CTH Sub-System 3: Distillation 1 (DT1) and Distillation 2 (DT2). continuous pumping of WCO (95% efficiency). LCI data for the rest
Products recovery involved a series of successive distillations DT1 of the sub-systems were obtained from various literature sources
and DT2, each employing one column. In DT1, naphtha (C6-C8) was and detailed technical descriptions were reported in Section 1 of SI.
recovered from the top, while the heavier fuels (C9-C18) were Electrical energy for pressurization, heat, and material flows were
separated from the residue. The residue was further distilled in DT2 determined from Aspen modeling. Compressed hydrogen gas was
to extract jet fuels as distillate and diesel from the bottom. Process supplied as make-up contribution because part of the non-reacted
conditions for DT1 were: 0.10 MPa, 10 stages, feed input at the 3rd H2 is recovered and recycled. As for CTH, large-scale LCI were based
stage, 0.20 reflux ratio, 220  C reboiler temperature, and 130  C on the experimental study by Asiedu et al. [18]. Specifically, iso-
condenser temperature. Process conditions for DT2 were: 0.10 MPa, propanol was supplied in large quantity (341.8 tons/day) following
15 stages, feed input at the 9th stage, 0.20 reflux ratio, 299  C scale-up of CTH reaction of WCO at different temperatures (Section
reboiler temperature, and 212  C condenser temperature. For both 2.1 of SI). The electrical energy was applied to the pump to pres-
distillation columns, heat was applied to the reboilers and heat was surize the WCO and isopropanol to 2 bar. The heat needed to in-
removed from the condensers. crease the feed’s temperature from 300 to 380  C was calculated by
436 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

Table 3
Equivalency factors used in modeling HRJ and CTH processes.

Impact Value Unit Notes

Electricity, U.S..
Energy use 12.5747 MJ/ 1.43% from surface coal mining, 1.85% from transportation, and 96.72% from electricity generation. Normal, SD ¼ 1 [55].
kWh
Water use 0.00004 m3/ Hard coal extraction region in the U.S., low-volatile bituminous coal at LHV ¼ 33,412 kJ/kg. Normal, SD ¼ 1 [56].
kWh
CO2 to air 1.0220 kg/ 0.94% from surface coal mining, 1.72% from transportation, and 97.34% from electricity generation. Normal, SD ¼ 1 [55].
SO2 to air 0.0067 kWh 1.06% from surface coal mining, 1.42% from transportation, and 97.52% from electricity generation. Normal, SD ¼ 1 [55].
kg/
kWh

Petro Diesel
Energy use 51.0996 MJ/kg 47.75% domestic crude production, 45% foreign crude production, 0.27% domestic crude transport, 1.09% foreign crude transport, 5.38% crude
refining, 0.51% diesel transport. Normal, SD ¼ 1 [57].
Water use 0.0004 m3/kg Normal, SD ¼ 1 [57].
CO2 to air 10.0476 kg/kg 86.54% tailpipe fossil (combustion), 13.46% production. Normal, SD ¼ 1 [57].
SO2 to air 0.0027 kg/kg Normal, SD ¼ 1 [57].

H2, from fossil fuels (RER), as liquid H2

Energy use 33.6420 MJ/kg As oil, crude, in ground. Normal, SD ¼ 1 [58].
Water use 0.0739 m3/kg 0.000723 m3 water estimate þ 0.0732 m3 cooling water, unspecified origin. Normal, SD ¼ 1 [58].
CO2 to air 1.2328 kg/kg 1.23 kg fossil CO2 e 0.00278 kg biogenic CO2. Normal, SD ¼ 1 [58].
SO2
4 to 0.0005 kg/kg Normal, SD ¼ 1 [58].
water

Isopropanol
Energy use 3.2000 MJ/kg Propylene oxide production via direct hydration demands a split of 50% natural gas, 38% electricity, and 12% steam. All energy used for heat of
steam was obtained from natural gas. Lognormal, SD ¼ 1.88 [59].
Water use 0.0271 m3/kg 0.0032 m3 water estimate as 10x stoichiometric amount þ 0.024 m3 cooling water. Lognormal, SD ¼ 1.88 [48].
CO2 to air 0.0990 kg/kg From wastewater treatment. Lognormal, SD ¼ 1.88 [60].
SO2
4 to 0.1740 kg/kg Calculated from mass balance. Lognormal, SD ¼ 2.11 [60].
water

Carbon black, at plant (RER)

Energy use 9.8820 MJ/kg As heat from industrial furnace, >100 kW. Lognormal, SD ¼ 1.24 [61].
Water use 0.0000 m3/kg Uniform ¼ 0.00 (minimum), 0.01 (maximum) [60].
CO2 to air 1.9700 kg/kg Calculated from mass balance. Lognormal, SD ¼ 1.38 [60].

Cryolite (2 AlF3·3NaF)a
Energy use 3.2000 MJ/kg 2 MJ/kg natural gas and 1.2 MJ/kg electricity. Normal, SD ¼ 1 [62].

Water use 0.0300 m3/kg 24 kg/kg þ 6 kg/kg cooling water. Normal, SD ¼ 1 [62].

CO2 to air 0.0000 kg/kg Uniform ¼ 0.00 (minimum), 0.01 (maximum) [62].

Steel
Energy use 2.1826 MJ/kg 18.66% coke-making, 9.59% sintering, 67% blast furnace, 4.75% LD steel converter. Cradle-to-casting plate gate (steel plates). Normal, SD ¼ 1
[63].
Water use 0.0003 m3/kg Normal, SD ¼ 1 [63].
CO2 to air 1.1586 kg/kg Normal, SD ¼ 1 [63].
a
Due to missing production data for Ni-Mo/Al2O3 catalyst, the impacts of cryolite production was adapted as a rough estimate on energy use, water use, and CO2 emissions
according to the overall reaction: 12 HF þ Al2O3  3H2O þ 6 NaOH / 2 AlF3  3 NaF þ 12 H2O. To produce 1 kg cryolite, these reactants are needed stoichiometrically:
28.580 mol HF, 2.382 mol Al2O3  3H2O, and 14.290 mol NaOH at 100% assumed yield. Gendorf (2000) [62].

Aspen’s heat integration. Catalytic transfer hydrogenation output to PSA unit. Hence, the amount of hydrogen consumed would be
succeeding sub-systems were based on experimental yields by slightly larger than indicated. The overall hydrogen-to-WCO ratio of
Asiedu and colleagues. Detailed technical information for CTH LCI 3.5% w/w is similar to the ones calculated by Chu et al. (2.6e3% w/
data were outlined in Section 2 of SI. Section 3 of SI provides Aspen w), Han et al. (2-3% w/w), and Pearlson et al. (4% w/w) [30,33,45].
simulation. A full accounting of energy/materials inputs/outputs is Nevertheless, Chu et al. did not consider the amount of hydrogen
presented in Table 4. used for hydrocracking reactions assuming it as a factor of the
degree of cracking instead. Pearlson et al. proposed that 4% w/w
3. Results and discussion overall process ratio is appropriate for maximum jet fuel produc-
tion based on literature data. The consumption of isopropanol in
3.1. Process performance CTH is equal to its supply (341.8 ton/day), since it was completely
decomposed inside the reactor and hence, non-recyclable. CTH’s
The results of process simulation for HRJ and CTH systems are isopropanol use is equal to 45.5 ton/day H2 consumption equiva-
summarized in Table 5. Aspen simulation resulted in an overall HRJ lent. Normalized on per ton biojet fuel produced, HRJ consumes 1.4
hydrogen consumption of 35.2 ton/day, of which 22.93 ton/day ton H2 while CTH utilizes 2.4 ton H2. On a per ton WCO consumed,
were consumed in the HY reactor (at hydrogen conversion of HRJ uses 0.035 tons H2 while CTH utilizes 0.045 ton H2.
88.2%), 11.34 ton/day in the hydrocracking unit, and 0.93 ton/day The yield of liquid fuel products obtained from both systems
lost in the light gas stream separated by the low-pressure flash. were similar, comparable to previous studies [10,30,33,45] (77 wt%
Hydrogen was assumed to be completely recycled without any for HRJ and 76 wt% for CTH). The liquid mixture obtained from the
losses, although in practice these may vary between 15 and 25% for conventional HRJ system were characterized by a larger amount of
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 437

Table 4
Global LCI for HRJ and CTH detailing energy/materials inputs/outputs to and from the technosphere/environment based on Table S71. Energy and heat are in MJ/day and
materials and chemicals are in tons/day.

HRJ

Energy, chemicals (input from technosphere) Products, co-products (output to technosphere)

Electricity 683,059.20 Naphtha 267.80

Heat 512,599.92 Jet fuel 478.37
Diesel 16.53 Propane gas 4.12
WCO 1000.00 C1eC3 gas 97.64
NieMo/Al2O3 3.26 C4eC8 51.81
H2 (gross) 83.44 C9eC15 1.19
H2 (net, minus recycle)a 35.20 C1eC3 liquid (LPG) 7.52
Steel 0.0017 C16eC18 (diesel) 2.64
Emissions (output to environment) Liquid CO2 1.52
Air Heat (cooling) 1,207,360.70
CO2 150.71
CO 11.42
H2 25.75
Water
Pure H2O 18.69
Sour water (H2O þ CO2) 37.77

CTH

Energy, chemicals (input from technosphere) Products, co-products (output to technosphere)

Electricity 158,696.70 Naphtha 232.86
Heat 485,100.00 Jet fuel 402.88
Diesel 14.44 Diesel 122.13
WCO 1000.00 Coke 29.41
Isopropanol 341.80 C1eC3 gas 161.85
Activated carbon 0.0913 Octane 5.83
Emissions (output to environment) Dodecane 0.11
Air C1eC3 liquid (LPG) 4.40
CO2 210.79 Heat (cooling) 493,849.44
CO 127.90
H2 6.85
Water
Pure H2O 16.26
Sour water (H2O þ CO2) 72.06
a
Net H2 requirement is 35.20 tons/day, of which 22.93 tons/day is consumed in the hydrodeoxygenation reactor, 11.34 ton/day in the hydrocracking unit, and 0.93 ton/day
are lost in the light gas stream separated by the low-pressure flash.

Table 5 10% of the WCO feed was converted into light gas (mostly propane)
Process performances of HRJ and CTH systems. All specific energy consumption of in HRJ, comparable to what is reported in other studies, ranging
utilities is scaled to the mass of WCO feed.
between 7% and 10% [30,33,45]. This light gas had low LHV because
Performance parameter HRJ CTH Unit it contained large amount of CO and CO2. The CTH system produced
Waste cooking oil basis rate 1000 1000 ton/day around 162 ton/day of fuel gas with LHV of 21.2 MJ/kg. The thermal
H2/isopropanol consumption 35.2 341.8 ton/day and electricity comparison between the two systems reflected their
H2/isopropanol consumption 17,600 5678.2 kmol/day relative process conditions: temperature and pressure. The thermal
H2 consumption equivalent 17,600 22,748.8 kmol/day energy (heat) consumptions were comparable between the two
H2 consumption equivalent 35.2 45.5 ton/day
H2 consumption/ton WCO 0.035 0.045 ton
systems (513 GJ/day for HRJ, 485 GJ/day for CTH), due to the opti-
H2 consumption/ton biojet 1.4 2.4 ton mized heat integration network and the similar operating tem-
peratures. CTH’s cooling energy requirement was 59% lower than
Liquid fuels products
Total liquida 770.7/40,207.2 763.2/40,139.4 ton/day; L/hr HRJ’s due to the latter’s combined exothermic effect via flash sep-
Naphthab 8041.5 12,041.8 L/hr aration and high-pressure flash. In terms of direct electricity, HRJ
Kerosenec 32,165.8 24,083.6 L/hr consumed more than four times electricity than CTH because of its
Diesel 0 4013.94 L/hr high operating pressure and the handling of gaseous hydrogen via
Energy consumption multi-stage compressors. CTH utilized low electricity because it
Heating energy 0.243 0.298 kWh/kg operates at relatively low pressure and pumps were employed in
Cooling energy 0.517 0.365 kWh/kg
place of compressors to transport liquid streams. In general, CTH
Electric energy 0.092 7.4$105 kWh/kg
Maximum pressure 9.2 0.2 MPa consumed far less heat and electricity than HRJ because the number
Maximum temperature 400 380 
C of pieces of equipment necessary in an HRJ reaction was much
a
Total liquid yield: HRJ ¼ 77 w/w%, CTH ¼ 76 w/w%.
larger (two reaction systems: HY and HH), whereas only one reactor
b
Naphtha yield: HRJ ¼ 20 v/v%, CTH ¼ 30 v/v%. is used in CTH. Further, CTH does not require large recycle of
c
Kerosene yield: HRJ ¼ 80 v/v%, CTH ¼ 60 v/v%. gaseous streams, which is beneficial in reactor sizing and operation.

3.2. Capital expenditure

kerosene products, roughly 16% higher than that obtained from
CTH. However, CTH produced more naphtha and diesel than HRJ. The results of CAPEX and FCI for both HRJ and CTH systems are
Both systems produced gaseous by-products. Specifically, about summarized in Table 6 and Table 7, respectively. A summary of
438 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

direct manufacturing costs for both systems are reported in Table 8. Table 7
Moreover, the results of COM for HRJ and CTH systems are shown in Fixed capital investment in HRJ and CTH jet fuel production processes.

Table 9 and Table 10, respectively using direct, fixed, and general Fixed Capital Investment HRJ CTH
costs of manufacturing data (Table S58 and Table S59). ISBL þ Engineering $ 149,681,462 $ 7,319,681
In both CTH and HRJ systems, CAPEX was controlled by their main OSBL (50% total Cbm base) $ 7,691,843 $ 3,659,841
reaction sub-systems: 91% in HRJ and 45% in CTH, calculated in Contingency (15% total Cbm) $ 26,942,663 $ 1,097,952
reference to the bare module cost assuming a factor of 4.74 to fluid Total $ 184,315,968 $ 12,077,474
process as single Lang factor, Cbm [22]. It is worth noting that the
great difference in the capital costs of reactors between HRJ and CTH
in Table 6 is due to a few factors. First factor is reactor volume. HDO
and HCC reactors in HRJ operate with a large excess of gaseous H2, Table 8
resulting in a much larger (about 10 times) reactor volume Summary of direct manufacturing costs.
compared to CTH. Second factor is catalyst price. While conventional
Unit Value Reference
hydrotreating processes employ metal-based catalysts (Ni-Mo), CTH
uses a sensibly cheaper catalyst that is made of activated carbon. Raw Material
Waste cooking oil $ ton1 150.00
Third factor is operating conditions. HRJ operates at a pressure of 92
Hydrogen $ kg1 1.60 [45,64]
and 90 bars for HDO and HCC, respectively, compared to 2 bars for Isopropanol $ kg1 1.30 [65]
CTH. High pressures require high quality and costly stainless steel as
Utility
construction material, which increases the Cbm of HDO and HCC by Electric power $ kWh1 0.087 [46]
about 25 times relative to the base value. The heating component in Natural gas (LHV ¼ 38.42 MJ/m3) $ m3 0.130 [46]
CTH’s main reactor costs over 3 M$, about 2% the cost of the heating Cooling tower water (30  C) $ GJ1 0.354 [38]
component in HRJ’s main reactor. CTH required pumping, as the Refrigerated water (5  C) $ GJ1 4.430 [38]
Medium pressure steam (10 bar, 184  C) $ GJ1 6.870 [38]
solvent was isopropanol while HRJ required compression, as the
High pressure steam (41 bar, 254  C) $ GJ1 9.830 [38]
source was gaseous hydrogen. Gas compression is prohibitively Hot oil $ GJ1 f (kW) [44]
expensive, in the order of M$. Both systems needed infrastructure
Operators
for separation, although CTH demanded a single low-pressure Salary $ year2 60,000.00 [66]
separator while HRJ required multiple high- and low-pressure
separators including a PSA hydrogen recovery unit.

Table 6
Total bare module cost in HRJ and CTH jet fuel production processes. For HRJ the cost at atmospheric pressure (base) is reported in addition to the final one.

Equipment Type HRJ CTH

Reaction System Cbm,base Cbm Reaction System Cbm

Hydrodeoxygenationa $ 1,650,989 $ 40,814,774 Hydrodeoxygenation $ 132,396

Hydrocrackingb $ 2,972,309 $ 76,618,490 2nd hydrodeox. regen. $ 132,396
Steam methane reform. $ 3,441,561 $ 16,313,000
Heatingc $ 1,450,520 $ 1,675,585 Heatingc $ 3,010,654

Compression Cbm,base Cbm Pumps Cbm

Compression 1 $ 448,539 $ 1,704,448 Feed pumping WCO $ 12,046

Compression 2 $ 1,234,069 $ 4,689,461 Feed pumping ISOPd $ 11,158
Feed pumping WCO $ 73,966 $ 174,098

Separation Cbm,base Cbm Separation Cbm

By-P HP separation $ 74,603 $ 663,506 LP separation $ 74,603

Product HP separation $ 53,829 $ 581,562
Product LP separation $ 63,636 $ 63,636
PSA-hydrogen recov. $ 889,876 $ 2,936,590

Distillation Cbm,base Cbm Distillation Cbm

Fractionation 1e $ 487,216 $ 487,216 Fractionation 1e $ 387,179

Fractionation 2e $ 491,028

Heat Exchangers Cbm,base Cbm Heat Exchangers Cbm

Heat exchange 1 $ 104,310 $ 123,010 Heat exchange 1 $ 469,833

Heat exchange 2 $ 157,024 $ 185,174
Heat exchange 3 $ 387,390 $ 456,232
Heat exchange 4 $ 159,540 $ 187,891
Cooling 1f $ 139,372 $ 164,358 Coolingf $ 151,479
Cooling 2f $ 144,417 $ 169,854
Heatingc $ 1,450,520 $ 1,672,574 Heatingc $ 2,446,909

Total (2017) $ 15,383,686 $ 149,681,462 Total (2017) $ 7,319,681

a
4x5 bed,
b
5x5 bed,
c
Fired heater,
d
Three stages.
e
Trays þ columns.
f
Cooling water.
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 439

Table 9 system, respectively (Table 7). The direct manufacturing costs in

Total cost of manufacturing in HRJ jet fuel production process. HRJ and CTH as summarized in Table 8 demonstrated that on a per
Raw Material (RM) Value Unit $/year kg basis, isopropanol and hydrogen’s unit costs were relatively
WCO 1000 ton/day 50,000,000
similar. However, when mass flow rates and hours of operation are
Hydrogen 35.2 ton/day 18,773,333 factored in, the total annual cost of isopropanol ballooned to
148.11 M$, in comparison to hydrogen’s 18.77 M$, regardless of
Utilities (UT) Value Unit $/year
hydrogen’s compression requirement. In fact, isopropanol’s cost
Compressor 1 1196 kW 836,586
represented 67.8% of CTH’s total COM (without depreciation,
Compressor 2 (make-up) 3191 kW 2,230,902
Feed WCO pump 125 kW 87,540 Table 10). Table 9, on the other hand, shows that the driver in HRJ’s
Cooling water 1 6197 kW 63,175 total COM is WCO, not hydrogen. Hydrogen compression was
Cooling water 2 5736 kW 58,477 another significant driver, in conjunction with the cost associated
Fired heater (LHV ¼ 38.42 MJ/m3) 704 kW 85,807 with heating the single reboiler in the HRJ system.
Condenser 2042 kW 20,815
Reboiler 4189 kW 828,901
3.3. Operations expenditure
Waste Treatment (WT) - - -

DMC þ FMC þ GE (Table S58) 29,251,124 The breakdown of total OPEX (raw materials, utilities) for the
Total COM (without depreciation) 102,236,661 two systems are shown in Fig. 4 (bottom panel). It is worth noting
that OPEX for HRJ amounted to 73 M$/year, while OPEX for CTH
amounted to 203 M$/year. This huge difference was determined by
the cost of isopropanol (148 M$/year), around 68% of CTH’s total
Table 10
Total cost of manufacturing in CTH jet fuel production process. operating cost (COM without depreciation). In general, the feed-
stock supply represented the major input to both systems. The
Raw Material (RM) Value Unit $/year
overall COM, accounting for operating labor, fixed cost of
WCO 1000 ton/day 50,000,000 manufacturing, and general expenses, amounted to 102 M$/year
Isopropanol 341.8 ton/day 148,113,333
and 218 M$/year for HRJ and CTH system, respectively.
Utilities (UT) Value Unit $/year In terms of revenue, the two systems were very similar, i.e.
Hot oil CTH 9508 kW 2,138,041 150.8 M$/year for HRJ and 153.9 M$/year for CTH (Table 11). By
Feed WCO pump 2.1 kW 1458 comparing the revenues with the total COM, CTH generates a
Feed ISOP pump 1 kW 696 negative gross profit, with most of the income offset by the cost of
Cooling water 1 3153 kW 32,140
supplying isopropanol.
Fired heater (LHV ¼ 38.42 MJ/m3) 5614 kW 683,904
Reboiler 1 5034 kW 1,425,077
Reboiler 2 1747 kW 867,511 3.4. Break-even analysis
Condenser 1 1058 kW 10,788
Condenser 2 1505 kW 15,348 Break-even analysis was performed to ascertain the profitability
Waste Treatment (WT) - - - of CTH and HRJ systems using total CAPEX (CTH ¼ $7.3M,
DMC þ FMC þ GE (Table S59) 15,147,753 HRJ ¼ $149.7M), annual total COM (without depreciation)
(CTH ¼ $218.4M, HRJ ¼ $102.2M), revenues from liquid fuel
Total COM (without depreciation) 218,436,050
(CTH ¼ $140.4M, HRJ ¼ $141.4M), revenues from gaseous fuel
(CTH ¼ $13.5M, HRJ ¼ $9.4M), and depreciation (use MACRS for 7
years). With these inputs, the net present value (NPV) for both
The breakdown of CAPEX for HRJ and CTH systems are shown in processes under 10-year debt financing term was evaluated
Fig. 4 (top panel), respectively. The total cost of CTH plant (7.3 M$) (CTH ¼ $216.6M, HRJ ¼ $6.4M). To ascertain how NPV is sensitive
was significantly lower (around 98%) than the total cost of HRJ plant to cost inputs (i.e. feedstock and products prices), these were varied
($149.7 M$). This was mainly due to the high operating pressure in by ±10% with respect to the base case values. The overall results are
HRJ, which increased the bare module cost by about 10 times reported in Table 12, while Fig. 5 shows the impact of each single
relative to atmospheric conditions. Together, the cost of HY and HH factor on the NPV of HRJ and CTH processes, respectively. From the
reactors in HRJ represented more than three-fourths of the 2017 latter, it is shown how jet fuel price plays an important role in both
total CAPEX or 64% of FCI. This large contribution was attributed to processes, while the cost of isopropanol is the most impacting
the slow reactions that unavoidably resulted in large reactor vol- factor in CTH.
umes and critical operating conditions. Thereafter, the break-even point (i.e. time in years at NPV ¼ 0)
The SMR plant for hydrogen production substantially contrib- for the 10-year term was calculated for both systems by setting the
uted to the overall capital investment, about 11%. Fractionation in NPV ¼ 0 and generating a new annual cumulative NPV profile (0-25
both HRJ and CTH systems operated atmospherically. HRJ’s single years) with year 10 as the break-even point. This method was
fractionation column had a negligible contribution to total CAPEX repeated for other break-even points: 5, 15, 20, and 25 years for
(<1%). But unlike HRJ, the CTH system consisted of two fraction- both CTH and HRJ systems. The results are plotted in Fig. 6.
ation columns in series, thereby doubling the cost of distillation. The 25-year NPV profiles of CTH (left) and HRJ (right) systems
The contribution of these two columns to the total CAPEX was 12%. reveal the general trend in cumulative NPV across all financing
The CTH main reactor was comparably priced relative to other terms. The initial dip in NPV from 0 to 2 years indicates the M$
units, thanks to the low cost of the catalyst employed. In terms of initial investment. Seventy percent of the initial investment was
heat exchange, CTH has one integrated heat exchanger, one cooling spent on the first year and the remaining 30% on the second year. At
unit, and one fired-heating element as opposed to HRJ’s four inte- year 2, the profile vertically dips even further with magnitude
grated heat exchangers, two cooling units, and one fired-heating corresponding to the working capital. From 2 years onward, the
element. profile increases until it hits the break-even point. Thus, DPBP co-
The FCI of the two systems, accounting for the offsite and con- incides from year 2 to break-even point. This discounted cash flow
tingency contributions, are 184.3 M$ and 12.1 M$ for HRJ and CTH analysis revealed that the cumulative cash flow became positive
440 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

Fig. 4. Capital expenses of HRJ (above left) and CTH (above right) systems. Operating expenses for raw materials and utilities of HRJ (below left) and CTH (B) systems.

Table 11
Product incomes and gross profits in HRJ and CTH jet fuel production processes.

Fuel Product HRJ CTH

Produced Unit Income ($/yr) Produced Unit Income ($/yr)

Fuel gas 6969 Nm3/hr 9,433,018 12,978 Nm3/hr 13,497,224

Naphtha 8041 L/hr 27,405,246 12,042 L/hr 41,038,506
Kerosene 32,166 L/hr 113,969,798 24,084 L/hr 85,333,119
Diesel e L/hr e 4014 L/hr 13,997,406
Total revenue (R) 150,808,063 153,866,254
Cost of manufacturing (COMd) 102,236,661 218,436,050
Gross profit (GP) 48,571,402 64,569,796

Table 12
Sensitivity of NPV on feedstock and products prices.

10% Base case þ10% NPV CTH NPV HRJ

a
WCO ($/t) 135 150.00 165 Min ¼ 335 M$ Mina ¼ 73 M$
H2 ($/kg) 1.44 1.60 1.76 Maxb ¼ 100 M$ Maxb ¼ 21.5 M$
Isopropanol ($/kg) 1.17 1.300 1.43 Base case ¼ 216.6 M$ Base case ¼ 6.4 M$
Naphtha ($/L) 0.383 0.426 0.469
Jet fuel ($/L) 0.399 0.443 0.487
Diesel ($/L) 0.392 0.436 0.479
Fuel Gas ($/Nm3) 0.117 0.13 0.143
a
NPV value corresponding to a 10% increase in feedstock price and 10% decrease in products sale price.

passed the break-even point. Depreciation was factored in from both CTH and HRJ systems, the salvage value of the project de-
year 4 to year 11, and the tail end of the profile shows the NPV of creases as the financing term increases.
each system at the end of its life. For example, CTH’s end-of-life Negative cumulative NPV indicates that any system is deemed
NPVs were as follows: 35.29 M$, 8.01 M$, 3.05 M$, 1.04 M$, and unprofitable while positive cumulative NPV indicates otherwise.
0 M$ for financing term of 5, 10, 15, 20, 25 years, respectively. In For a system to be viable using the break-even analysis, one must
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 441

Fig. 5. Sensitivity of process NPV to feedstock and products prices for HRJ (left) and CTH (right). Dotted yellow bars indicate þ10% change in assumption from base case and full red
bars 10% change in assumption from base case. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6. Cumulative NPV profiles for CTH (left) and HRJ (right) systems at various debt financing terms showing corresponding break-even points (5, 10, 15, 20, and 25 years).

understand that the current cumulative NPV at year n is a function (12.94% difference) such that in the long run, CTH fuel would be
of revenue, total COM, and depreciation at that year, including the affordable as a result of its low total CAPEX. It appears that CTH
initial investment. The break-even point is the time when annual would recover its initial investment faster than HRJ through faster
revenue equals total business costs (total COM, depreciation) pro- payment of its CAPEX. As capital recovery within 5 years appears
gressively achieved by paying a fraction of the initial investment ambitious, maintaining DPBP (8 years after the construction) and
year after year. Typically, the break-even point in most economic an interest rate of 8% for both processes would be ideal.
analyses determines the amount or volume of product sales in a An IRR analysis was performed for both CTH and HRJ systems
year that equals total dollar sales to total dollar COM (i.e. gross covering interest rates from 1 to 50% for the entire 25-year life. The
profit ¼ $0). In this analysis, however, the break-even point was inputs to this analysis were base case inputs, that is, CTH-NPV was
calculated as the year beyond which total dollar sales exceeds total negative and HRJ-NPV was positive at the end of 25 years under a
dollar COM, coinciding with cumulative NPV>$0. There was no set 10-year financing term. The results are presented in Fig. 7. The IRR is
break-even point for which a business model ensures profitability; an index of the investment risk, a method used in capital budgeting
however, if total COM is higher than expected total sales, the to estimate the profitability of investments. It refers to the discount
business plan can be re-examined, and alternative strategies must rate that makes NPV of all cash flows equal to 0. IRR calculations
be implemented. This study determined that selling the biojet fuel followed the same method as NPV. As the interest rate increased,
at $2.99/gal ($0.79/L) and $1.32/gal ($0.35/L) for CTH and HRJ, HRJ-NPV decreased while CTH-NPV increased. Since HRJ-based
respectively, would take 25 years to break even. Since most cor- biojet fuel production is a mature technology with low risk, its
porations prefer break-even points to be less than 10 years, it IRR of 15.91% represents an attractive value for the investment. As
would, therefore, be profitable to sell biojet fuel at $3.00/gal ($0.79/ of current market without any subsidies, incentives, or any of the
L) and $1.67/gal ($0.44/L) for CTH and HRJ, respectively. At these investigated case scenarios from Table S68, CTH’s IRR profile seem
rates and for a project life of 25 years, the NPV for CTH and HRJ were to suggest that the system is untenable.
estimated to be 8 M$ and 121 M$, respectively. It was observed that At the present state, the CTH process based on experimental
the biojet fuel selling price for the CTH system is higher than that of conditions investigated has potential to be economically profitable
HRJ due the higher operation cost associated with CTH. However, as despite large input cost associated with isopropanol. Given that
the break-even point approaches 5 years, the biojet fuel selling isopropanol cost represents 68% of CTH’s total COM, total COM
price for HRJ ($2.70/gal) and that for CTH ($3.10/gal) converges must be reduced by around 30% to even start realizing net profit.
442 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

input. On the other hand, HRJ operations involves hydrogen pres-

surization energy in HY and HH units assuming isentropic
compression (85% efficient) and WCO pumping (95% efficient). It
also includes cooling energy in the flash separator, high-pressure
flash unit, distillation column (exothermic conditions), and heat
input. From a gross operations perspective, HRJ exhibits 1.6 times
the CED of CTH. This is not surprising given that it has greater
upstream fossil energy brought about by the production, pressur-
ization, and delivery of hydrogen. Gaseous hydrogen is primarily
produced from stream reforming of natural gas. Isopropanol is the
main upstream driver in CED of CTH; albeit insufficient amount to
cause higher CED of HRJ.
Direct CED of CTH and HRJ systems are negative, that is, surplus
electricity is generated by the system. The energy generated from
cooling the reactor is re-captured and utilized as in-process energy.
HRJ operates at higher temperature than CTH, enabling larger en-
Fig. 7. IRR profiles for CTH and HRJ systems under 10-year debt financing term.
ergy release. It also has more reactor units. Consequently, HRJ’s net
heat consumption is 26% higher than CTH’s. HRJ’s direct heat use
This can be done using a combination of two strategies: (1) reduce comes from heating the oil from 380  C to 400  C plus the heat duty
both isopropanol cost and total direct/fixed/general cost by 60%, (2) of the distillation column. Although CTH consists of two distillation
reduce isopropanol cost by 55%. It is highly unlikely that a 60% units in series needing heating of WCO from 300  C to 380  C, the
reduction in staff salaries, plant overhead, and distribution/selling combined heat demand is still lower than HRJ’s.
costs will be achieved because these variables are set by prevailing Total CED is equal to operations CED plus infrastructure CED. The
market conditions and for an emerging technology such as CTH- results suggest that total CED for both systems are equal to their
based biojet fuel, these costs will be slightly elevated, leaving the operations CED due to the miniscule contributions of steel. Reactor
latter option to be realistically feasible. It was calculated that to steel demand for both systems are between 33 and 177 kg/yr,
obtain the same Present Value Ratio of HRJ (a more meaningful annualized over a 30-year period.
index than the NPV when two processes with very different capital The results of operations GWP is consistent with the results of
investments are compared), the price of the isopropanol should not operations CED in both systems, i.e. operations GWP-
exceed 0.735 $ kg1 (i.e. a 43.5% reduction from current cost) CTH < operations GWP-HRJ; however, the difference is low at
(Table S68). Alternatively, at the current market price, the around 8%. This result considers that the net GHG emissions has
maximum isopropanol inlet was calculated to be 196.54 ton/day two components e direct GHG emissions and upstream GHG
(isopropanol/WCO feed ratio of 0.204 by weight) to be profitable. emissions. Both CTH and HRJ systems do not sequester CO2 through
However, performances and product yields should be verified in co-product offsets because there is no biogenic carbon embodied in
this scenario. the final fuel unlike jet fuel from algae or switchgrass. While it is
true that some WCO are plant-based (olive oil, palm oil, soybean oil,
or canola oil), others are animal-derived (butter, lard) or synthetic.
3.5. Environmental metrics In fact, even if this study assumes that all WCO are plant-derived, it
is still outside the scope to identify which location produced waste
The results of the inventory analysis described in Section 2.5.5 olive oil or canola oil. Thus, this analysis does not consider carbon
was used in life cycle impacts assessment, a step in LCA that sequestration. Direct GHG emissions from CTH is higher than direct
quantifies environmental impacts, selects priorities from these GHG emissions from HRJ, mainly due to CO2 emissions from low-
environmental impacts, and transforms these priorities into pressure flash. However, upstream GHG emissions from CTH is
meaningful environmental endpoint (damage category or forecast). 26% lower than HRJ’s because of the considerable emissions asso-
In his study, two major environmental endpoints were identified: ciated with hydrogen production. Total GWP is equal to operations
CED and life cycle greenhouse gas GHG emissions using the Eco- GWP due to the insignificant GHG emissions associated with steel
Indicator 99 method [67] and the equivalency factors (damage in the reactors. Further, both CTH and HRJ systems have no heat
factors) detailed in Table 3. CED represents the cumulative non- offset, i.e. total heat use equals net heat use. The hydro-
renewable fossil energy requirement of the entire process in MJ deoxygenation reaction in HRJ consumes more heat than the CTH
per 1000 tons WCO. Life cycle GHG emissions is reported as 100- reaction.
year global warming potential (GWP) in kg CO2 equivalents per The conversion of 1000 tons/day WCO generates approximately
1000 tons WCO. 60  106 MJ useable fuel energy in the form of jet fuel, naphtha, and
Fig. 8 and Table 13 depict the characterization results in the diesel from both CTH and HRJ systems. As evident in the right pane
production of biojet fuel from WCO via CTH and HRJ processes. The of Fig. 8, relative magnitudes of these energy are similar, except that
left pane demonstrates environmental endpoints (CED and GWP, HRJ reaction does not produce diesel product. HRJ has a slightly
positive bar) and environmental credits (negative bars). The right lower total energy than CTH despite its higher jet fuel and naphtha
pane shows the energy distribution of fuel products produced from production because HRJ does not produce diesel. The EROI of both
CTH and HRJ processes including the energy-return-on-investment systems were calculated using total energy output and total energy
(EROI), defined as the ratio of the total energy output from the input. EROI is the system’s energy efficiency measured in terms of
system to the total energy input into the system. energy productivity normalized by energy expenditure. It is used as
To facilitate analysis, it is important to note that CTH operation a decision tool in adapting any process from an energetic stand-
includes pressurization energy required to pump WCO and iso- point with EROI>1 to be desirable (i.e. energy break-even ¼ 1, total
propanol to the CTH reactor at 2-bar pressure, cooling energy energy output ¼ total energy input). However, different EROIs can
(negative energy) needed to cool reaction products in low-pressure be calculated from the same dataset as a result of dissimilar energy
flash unit, and distillation units (by heat integration), and heat accounting methods [68]. In this study, EROI is evaluated as the
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 443

Fig. 8. Results of characterization in the production of CTH- and HRJ-based biojet fuel from 1000 tons WCO. Left pane corresponds to environmental endpoints (CED and GWP) and
right pane corresponds to distribution of output energy. Energy-return-on-investment (EROI) is included in the right pane.

Table 13
Characterization results of Fig. 7. Cumulative fossil energy demand (CED) and output
fuel energy (OUT) are expressed as  106 MJ/1000 tons WCO and global warming
potential (GWP) is expressed as 106 kg CO2-equivalents/1000 tons WCO.

Input (CED/GWP) Operations Direct Heat Upstream

CED-CTH 1,625,836 447,771 583,780 1,489,773

CED-HRJ 2,539,483 804,010 793,395 2,550,098
GWP-CTH 327,414 192,557 0 134,240
GWP-HRJ 356,414 137,451 0 218,686

Output (Fuel Energy) Naphtha Diesel Jet Fuel Total

OUT-CTH 9,485,084 4,720,114 15,716,052 29,921,250

OUT-HRJ 10,908,355 0 18,660,657 29,569,011

ratio of the summation of all energy embodied in the main product

and co-products plus surplus electricity (if any), and, the summa- Fig. 9. Electricity and heat use breakdown among sub-processes in the production of
CTH- and HRJ-based biojet fuel from 1000 tons WCO.
tion of net energy (if any), net heat, upstream energy, and infra-
structure energy consumption. Although both energy platforms are
energetically favorable, CTH’s EROI is 1.8 times higher than HRJ’s,
products. The system has neither provision for recouping this large
attributable to its lower heat and upstream impacts.
emission for re-use nor method of sequestration. In contrast,
Fig. 9 shows that CTH has lower electricity and heat consump-
emissions associated with low-pressure flash in HRJ is one-fifth
tion than HRJ; in fact, CTH’s negligible electricity use is credited to
that for CTH’s, as exhibited by the stable reactor temperature of
pressurization, while the rest of the sub-processes yield in-process
HRJ at 24  C.
heat. Both systems have comparable direct heat use. Because CO2
separation, high-pressure separation, and fractionation units in HJR
operate at high temperatures, large amount of heat is expended to 3.6. Effect of allocation method
lower product temperature. HRJ’s single fractionation unit con-
sumes more than twice the electricity of CTH’s. The allocation methods were analyzed in the context of un-
The relative global warming potential (GWP) contributions re- derstanding the relative effect on CED and GWP (Fig. 11). For CTH,
ported as kg CO2 equivalents/1000 tons WCO of the different sub- CED for main products via mass- and energy-allocations are in the
processes for CTH and HRJ systems are presented in Fig. 10. The order of millions while CED for main products via market-
results indicate that both systems create similar warming within allocation is in the order of thousands. Conversely, CED for co-
100 years (CTH ¼ 376,403. HRJ ¼ 375,420). CTH’s GWP is dominated products via mass- and energy-allocations are two orders of
by emissions from low-pressure flash while HRJ’s is driven by WCO magnitude greater than CED for co-products via market-allocation.
transport emissions. This is surprising given that the CED of HRJ is This is due to mass rates (tons/day) corresponding to energy rates
1.56 times that of the CED of CTH. A closer look at the GWP (MJ/day) and the LHVs (MJ/kg) among fuel products and co-
breakdown reveals that the main reaction in CTH produces half the products are close to each other (min ¼ 42.791, max ¼ 47.206).
GHG emissions of HRJ’s main reaction. Nonetheless, the sum of Yet, market-allocation is based on price ($/gal) and density (kg/m3).
GHG emissions associated with the sub-processes in CTH, except Hence, the resulting allocation is remarkably different from mass-
for WCO/fuel transport and diesel emissions, is higher than HRJ’s by and energy-allocations. In the case of CTH, 28.20 M$ is attributed to
33%. Beyond CTH’s main reactor, the low-pressure flash generates co-products, of which 28.17 M$ is generated by the sales of gaseous
the most emissions among all sub-processes at 169,390 kg CO2-eq/ products (C1-C3 gas). Consequently, 98.33% was allocated to co-
1000 tons WCO. Such phenomenon is associated with: (1) the at- products while only 1.67% was allocated to main products. Similar
mospheric 3-phase (VLLE) flash unit following CTH that separates trend is observed in HRJ main products: 2.08 MMJ/1000 tons WCO
gaseous products from the liquid organic fraction and water, and and 2.06 MMJ/1000 tons WCO for mass- and energy-allocations,
(2) the large temperature differential in cooling the reaction respectively, while 0.06 MMJ/1000 tons WCO for market-
444 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

Fig. 10. GWP breakdown among sub-processes in the production of CTH- and HRJ-based biojet fuel from 1000 tons WCO.

Fig. 11. Analysis of CED and GWP in the production of CTH- and HRJ-based biojet fuel from 1000 tons WCO using mass rate (MRA), energy rate (ERA), and market value (MVA)
allocations.

allocation. In the case of co-products, mass- and energy-allocations The tornado plots presented in Fig. 12 is the result of indepen-
are in the order of hundred thousand while market-allocation is in dently testing each input variable one at a time against the target
the order of millions. Similar trend is observed in GWP. From an forecast, CED and GWP for CTH. The Monte Carlo simulation of the
environmental life cycle perspective, price and density have more models utilized assumptions and generated several charts that
pronounced effect on impacts assessment than LHV. Because this depicted the relationships between the assumptions and the target.
analysis pertains to comparing an emerging energy platform to a It is evident that CED and GWP of CTH are dependent on trans-
more established conventional processing, reporting CED and GWP portation e hauling of WCO/fuels and not on process parameters.
impacts based on the three allocation methods is vital. For example, a 9% increase in truck mileage efficiency from 5.50 mi/
gal base case to 6.05 mi/gal resulted in a 2.4% decrease in CED from
3.7. Sensitivity the median value of 0.0835 MJ/MJ WCO. The results reveal that as
the truck mileage efficiency increases, CED and GWP decrease;
The construction of CTH and HRJ models requires the incor- whereas, as distance hauled increases, both CED and GWP increase.
poration of materials and process inputs with distributions (“as- Of special case is the GHG emissions associated with the elec-
sumptions”) obtained from this study’s experimental results and tricity equivalency factor (U.S. grid). The base case is 1.0220 kg/kWh
from various literature sources (Table S70). The models also apply and SD ¼ 1. A 125% decrease in electricity equivalency factor
the equivalency factors in Table 3 during life cycle impacts resulted in a 68% increase in GWP (0.015e0.047 kg CO2 eq/MJ
assessment. These factors contain distributions. When analyzing a WCO). Similarly, a 51% decrease in electricity equivalency factor
model, it is useful to know which input variables have the greatest caused a negligible increase in GWP (0.015e0.016 kg CO2 eq/MJ
impact on the model’s result (endpoint or forecast). The goal is to WCO).
determine the driver of a forecast while ignoring inconsequential Results of the sensitivity analysis highlighted two key findings:
inputs. (1) fossil energy consumption and GWP are highly sensitive to
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 445

Fig. 12. Tornado plots depicting drivers in CED and GWP of CTH. Dotted yellow bars indicate þ5-10% change in assumption from base case and lined red bars indicate -5-10% change
in assumption from base case. Bar labels show the test range for each input variable. Tornado method used is percentiles of the variables and test range is between 10% and 90%.
Explained variation in forecast is cumulative. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

externalities associated with transportation of WCO and fuels, and energy as HRJ, expends half the fossil energy as HRJ, and creates
(2) the relative contributions of surface coal mining, transportation, 92% GWP of HRJ’s. Despite isopropanol’s high cost in the TEA
and electricity generation to GWP are insignificant because the (Section 3.3), it only accounts for 57% of CTH’s total CED and sur-
combined effects of these processes tend to amplify GHG emis- prisingly, around 9% of CTH’s total GWP.
sions. The first finding has an advantage and disadvantage. The The contributions between the CTH’s main reaction system
advantage relates to the minimal influence of CTH’s reaction versus HRJ’s on EROI projections are evident in their individual CED.
chemistry on the environment which allows any research and While isopropanol’s CED represents 99% of CTH’s main reaction
development to further enhance energy and emissions perfor- system (activated carbon’s CED is merely 1% of CTH’s main reaction
mance. The CTH-based biofuel plant does not require reconfigura- system and is assumed to be regenerated and replaced every 90
tions of the main reactor to cut costs, increase profits, and improve days), CTH reaction constitutes half the CED of HY and HH reactions
sustainability other than what was detailed in this analysis. Because combined. More importantly, isopropanol’s CED in producing CTH
the CED and GWP drivers are transportation-related and not biojet fuel is less than hydrogen’s CED in producing HRJ biojet fuel
process-related, the disadvantage would be the effect of diesel fuel by a factor of 0.5. Clearly, the environmental opportunities in using
used in heavy trucks on the circular bioeconomy, to the extent that an alcohol-based solvent remarkably outweighs any cost barrier.
diesel’s volatile market creates a gap between any gain or loss
incurred by the plant. Such disadvantage has a more pronounced 4. Conclusions
impact on the TEA aspect than the environmental outcome.
Given the knowledge that the environmental impacts of CTH are This study analyzed the techno-economic and environmental
driven by externalities, it is essential to examine the CTH model life-cycle implications of biojet fuel production from WCO via the
with regards to variations truck mileage efficiency and distance CTH process. Results were compared to the conventional HRJ pro-
hauled. There are two types of variations: uncertainty and vari- cess. TEA results demonstrate that the total CAPEX of CTH plant (7.3
ability. Uncertainty is the lack of pertinent information about the M$) is significantly lower (~95%) than the total CAPEX of HRJ plant
driver while variability is the characteristic spread of the driver in a ($149.7 M$). In terms of other economic metrics, CTH outperforms
set of observations or population. Truck mileage efficiency has HRJ based on revenues from gaseous fuels (CTH ¼ $13.5M,
variability [39,40] while distance hauled has uncertainty. The cur- HRJ ¼ $9.4M) but underperforms based on annual total COM
rent state of trucking industry in the U.S. has been essentially stable
and therefore emphasis must be placed on distance traveled, which
is a function of the study’s scope. This analysis determined that the
most optimal hauling distance is 300 mi/truck-d (6 h per truck per
day at 50 miles/hr), obtained from the four-region, nine-division
scheme (Section 2.5.1).
To capture CTH’s energy outlook as a function of distance hauled
and truck mileage efficiency across all regions using process con-
ditions and reaction chemistry described herein, a trend analysis on
CTH’s EROI was made on a single CTH-based biojet fuel plant. This
plant assumes to transition from HRJ to CTH in one year. The results
are presented in Fig. 13. HRJ’s median EROI of 10.5 lies within the
25% certainty band (range: 10.4e10.6). In other words, HRJ will
have a 25% probability of enhancing EROI to 10.6 from the median
value of 10.5. It also has a 90% chance of improving EROI to 11.30
from the median value of 10.5. Clearly, the latitude on improvement
(or deterioration) of energy efficiency through optimization in the
HRJ system is limited (EROI ¼ 10.30e11.30). In contrast, the range
by which CTH’s EROI could be improved is quite wide, from 18.6 to a
maximum of 25. This finding is significant considering that on a per Fig. 13. Trend chart showcasing the shift in EROI from HRJ to EROI at median EROI
volume WCO basis, CTH yields statistically equivalent biofuel values (HRJ-EROI ¼ 10.5, CTH-EROI ¼ 18.6.
446 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

(without depreciation) (CTH ¼ $218.4M, HRJ ¼ $102.2M). The allocation. In both systems, price and density have more pro-
revenues from liquid fuels in both systems are comparable nounced effect than LHV has on impacts assessment. The fossil
(CTH ¼ $140.4M, HRJ ¼ $141.4M). NPV analysis reveals that to be energy consumption and GWP are highly sensitive to externalities
profitable, and for a break-even point to be less than 10 years, biojet associated with transportation of WCO and fuels and the relative
fuel can be sold at $3.00/gal ($0.79/L) and $1.67/gal ($0.44/L) for contributions of surface coal mining, transportation, and electricity
CTH and HRJ, respectively. At these rates and for a project life of 25 generation to GWP are insignificant because the combined effects
years, the NPV for CTH and HRJ are 8 M$ and 121 M$, respectively. of these processes tend to amplify GHG emissions.
IRR for HRJ is 15.91% while CTH’s IRR profile is untenable without
subsidies, incentives, or any of the investigated case scenarios. This CRediT authorship contribution statement
TEA demonstrates that the CTH process based on experimental
conditions has potential to be economically profitable despite large Elena Barbera: Methodology, Validation, Data curation, Writing
input cost associated with isopropanol. Given that isopropanol cost - original draft. Rustem Naurzaliyev: Software, Methodology.
represents 68% of CTH’s total COM, total COM must be reduced to Alexander Asiedu: Investigation, Methodology, Writing - original
around 30% to realize net profit. This is done using a combination of draft. Alberto Bertucco: Supervision, Writing - review & editing.
two strategies: (1) reduce both isopropanol cost and total direct/ Eleazer P. Resurreccion: Methodology, Data curation, Writing -
fixed/general cost by 60%, and (2) reduce isopropanol cost by 55%. original draft. Sandeep Kumar: Conceptualization, Project admin-
The LCA results indicate that the environmental opportunities in istration, Resources, Writing - review & editing.
using an alcohol-based solvent remarkably outweighs any cost
barriers. The latitude on improvement (or deterioration) of energy Declaration of competing interest
efficiency through optimization of CTH’s process parameters is
wide. Despite isopropanol’s restrictive cost, CTH outperforms HRJ, The authors declare that they have no known competing
regardless of the allocation method employed. The CED of HRJ is 1.6 financial interests or personal relationships that could have
times the CED of CTH. Direct CED of CTH and HRJ systems are appeared to influence the work reported in this paper.
negative, meaning surplus electricity is generated. In cooling the
reactors, they also produce heat, which is re-captured and utilized Acknowledgements
as in-process energy. HRJ operates at higher temperature than CTH,
so HRJ consumes four times more electricity than that of CTH. CTH’s This research did not receive any specific grant from funding
cooling energy requirement is 59% lower than HRJ’s due to the agencies in the public, commercial, or not-for-profit sectors.
latter’s combined exothermic effect via flash separation and high-
pressure flash. Although EROIs for both CTH and HRJ are energet- Appendix A. Supplementary data
ically favorable (EROI > 1), CTH’s EROI is 1.8 times higher than HRJ’s.
The total 100-year GWP of CTH is 8% less than HRJ’s, and both Supplementary data to this article can be found online at
systems do not sequester CO2 through co-product offsets. CTH’s https://doi.org/10.1016/j.renene.2020.06.077.
GWP is dominated by emissions from low-pressure flash while
HRJ’s is driven by WCO transport emissions. Environmental end- Appendix A. e Stream Tables
points based on mass- and energy-allocations are relatively similar
to each other but remarkably different from market-value

Table A.1
Material and energy balances of HRJ process (Fig. 2)

BLOCK COMPRESSION HDO REACTOR FLASH-1

Inlet Outlet Inlet Outlet Outlet

STREAM H2-MAKEUP H2-REC H2 H2-HDO WCO PROD-HDO WATER HC GAS-HP

Temperature ( C) 25 36 155 155 25 155 59 59 59
Pressure (MPa) 0.1 4.0 9.2 9.2 0.1 9.2 9.2 9.2 9.2
Total mass flow rate (t/d) 35.2 24.8 60.0 18.7 1053.5 1072.2 37.8 1002.0 32.4
Components mass fraction
Linoleic acid C18H32O2 0 0 0 0 0.496 0 0 0 0
Oleic acid C18H34O2 0 0 0 0 0.177 0 0 0 0
Pentadecanoic acid C15H30O2 0 0 0 0 0.170 0 0 0 0
Nonadecanoic acid C19H38O2 0 0 0 0 0.106 0 0 0 0
N-tetradecane C14H30 0 0 0 0 0 0.097 0 0.104 0
N-pentadecane C15H32 0 0 0 0 0 0.042 0 0.045 0
N-Heptadecane C17H36 0 0 0 0 0 0.402 0 0.430 0
N-octadecane C18H38 0 0 0 0 0 0.237 0 0.254 0
N-Nonadecane C19H40 0 0 0 0 0 0.027 0 0.029 0
C1-C3 gases 0 0 0 0 0.051 0.050 0 0.049 0.127
Napthta (C4-C8) 0 0 0 0 0 0 0 0 0
Jet fuels (C9-C15) 0 0 0 0 0 0 0 0 0
Diesel (C16-C18) 0 0 0 0 0 0 0 0 0
H2 1 1 1 1 0 0.003 0 0.001 0.072
CO 0 0 0 0 0 0.005 0 0.002 0.104
CO2 0 0 0 0 0 0.098 0.001 0.082 0.694
H2O 0 0 0 0 0 0.039 0.999 0.004 0.003
Energy
Heating (MW) e e E2þFired Heater 5.93
Cooling (MW) Intercooling 2.58 Reactor cooling 5.16 CW1 6.2
Power (MW) Compression (C1þC2) 3.71
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 447

BLOCK HCC REACTOR HP-FLASH LP-FLASH

Inlet Outlet Outlet Inlet Outlet

STREAM H2-HCC HC-IN PROD-HCC V-HP L-HP L-LP LIGHTGAS LIQFUELS

Temperature ( C) 155 350 196 40 40 24 24 24
Pressure (MPa) 9.2 9.0 9.0 9.0 9.0 0.1 0.1 0,1
Total mass flow rate (t/d) 34.0 1002.0 1035.9 141.01 894.9 894.9 124.2 770.7
Components mass fraction
Linoleic acid C18H32O2 0 0 0 0 0 0 0 0
Oleic acid C18H34O2 0 0 0 0 0 0 0 0
Pentadecanoic acid C15H30O2 0 0 0 0 0 0 0 0
Nonadecanoic acid C19H38O2 0 0 0 0 0 0 0 0
N-tetradecane C14H30 0 0.104 0 0 0 0 0 0
N-pentadecane C15H32 0 0,045 0 0 0 0 0 0
N-Heptadecane C17H36 0 0.430 0 0 0 0 0 0
N-octadecane C18H38 0 0.254 0 0 0 0 0 0
N-Nonadecane C19H40 0 0.029 0 0 0 0 0 0
C1-C3 gases 0 0.049 0.102 0.385 0.057 0.057 0.355 0.010
Napthta (C4-C8) 0 0 0.334 0.063 0.377 0.377 0.351 0.383
Jet fuels (C9-C15) 0 0 0.448 0.002 0.518 0.518 0.007 0.602
Diesel (C16-C18) 0 0 0.003 0 0.003 0.003 0 0.003
H2 1 0.001 0.023 0.159 0.001 0.001 0.008 0
CO 0 0.002 0.008 0.052 0.001 0.001 0.006 0
CO2 0 0.082 0.080 0.337 0.039 0.039 0.273 0.002
H2O 0 0,004 0 0 0 0 0 0
Energy
Heating (MW) e e e e
Cooling (MW) e CW2 5.74 e
Power (MW) e e e e

BLOCK H2-RECOVERY FRAC

Outlet Outlet

STREAM CO2 FUELGAS H2-REC1 H2-REC2 NAPHTHA JETFUELS

Temperature ( C) 51 37 51 37 65 165
Pressure (MPa) 4.0 4.0 4.0 4.0 0.1 0.1
Total mass flow rate (t/d) 30.1 118.6 2.34 22.5 276.4 494.3
Components mass fraction
Linoleic acid C18H32O2 0 0 0 0 0 0
Oleic acid C18H34O2 0 0 0 0 0 0
Pentadecanoic acid C15H30O2 0 0 0 0 0 0
Nonadecanoic acid C19H38O2 0 0 0 0 0 0
N-tetradecane C14H30 0 0 0 0 0 0
N-pentadecane C15H32 0 0 0 0 0 0
N-Heptadecane C17H36 0 0 0 0 0 0
N-octadecane C18H38 0 0 0 0 0 0
N-Nonadecane C19H40 0 0 0 0 0 0
C1-C3 gases 0.137 0.458 0 0 0 0
Napthta (C4-C8) 0 0.080 0 0 0.988 0.016
Jet fuels (C9-C15) 0 0 0 0 0.012 0.984
Diesel (C16-C18) 0 0 0 0 0 0
H2 0 0 1 1 0 0
CO 0.112 0.062 0 0 0 0
CO2 0.748 0.400 0 0 0 0
H2O 0.003 0 0 0 0 0
Energy
Heating (MW) e Reboiler 4.19
Cooling (MW) e Condenser 2.04
Power (MW) e
448 E. Barbera et al. / Renewable Energy 160 (2020) 428e449

Table A2
Material and energy balances of CTH process (Fig. 3)

BLOCK CTH REACTOR VLL-FLASH

Inlet Outlet Outlet

STREAM WCO ISO-P COKE PROD GASES WATER HC

Temperature ( C) 25 25 380 50 15 15 15
Pressure (MPa) 0.1 0.1 0.2 0.2 0.1 0.1 0.1
Total mass flow rate (t/d) 1000 341.8 29.4 1312.4 483.3 71.2 757.9
Components mass fraction
Linoleic acid C18H32O2 0.523 0 0 0 0 0 0
Oleic acid C18H34O2 0.186 0 0 0 0 0 0
Pentadecanoic acid C15H30O2 0.179 0 0 0 0 0 0
Nonadecanoic acid C19H38O2 0.112 0 0 0 0 0 0
Isopropanol C3H8O 0 1 0 0 0 0 0
C1-C3 gases 0 0 0 0.124 0.340 0 0
Naphtha (C8H16) 0 0 0 0.178 0.012 0 0.307
Jet fuels (C12H26) 0 0 0 0.300 0 0 0.532
Diesel (C17H36) 0 0 0 0.091 0 0 0.161
CO 0 0 0 0.095 0.265 0 0
CO2 0 0 0 0.127 0.351 0.0004 0
H2O 0 0 0 0.057 0.012 0.9995 0
H2 0 0 0 0.006 0.015 0 0
Coke 0 0 1 0 0 0 0
Energy
Heating (MW) Fired heater 5.61 e
Cooling (MW) e e CW 3.15
Power (MW) Pumps (P1þP2) 0.003 e

BLOCK FRAC1 FRAC2

Outlet Outlet

STREAM NAPHTHA HEAVY JETFUEL DIESEL

Temperature ( C) 130 220 212 299
Pressure (MPa) 0.1 0.1 0.1 0.1
Total mass flow rate (t/d) 233.0 524.9 403.0 121.9
Components mass fraction
Linoleic acid C18H32O2 0 0 0 0
Oleic acid C18H34O2 0 0 0 0
Pentadecanoic acid C15H30O2 0 0 0 0
Nonadecanoic acid C19H38O2 0 0 0 0
Isopropanol C3H8O 0 0 0 0
C1-C3 gases 0 0 0 0
Naphtha (C8H16) 0.985 0.006 0.008 0
Jet fuels (C12H26) 0.015 0.761 0.989 0.008
Diesel (C17H36) 0 0.233 0.003 0.992
CO 0 0 0 0
CO2 0 0 0 0
H2O 0 0 0 0
H2 0 0 0 0
Coke 0 0 0 0
Energy
Heating (MW) Reboiler 5.03 Reboiler 1.75
Cooling (MW) Condenser 1.06 Condenser 1.51
Power (MW) e e e e

References hydrodeoxygenation of triglycerides: an approach to clean diesel fuel pro-

duction, Renew. Sustain. Energy Rev. 80 (2017) 1072e1088, https://doi.org/
10.1016/J.RSER.2017.05.268.
[1] IEA International Energy Agency, World energy outlook. https://webstore.iea.
[9] S. Popov, S. Kumar, Renewable fuels via catalytic hydrodeoxygenation of lipid-
org/world-energy-outlook-2018, 2018.
based feedstocks, Biofuels 4 (2013) 219e239, https://doi.org/10.4155/
[2] (EIA) U.S. Energy Information Administration, Annual energy outlook 2017,
bfs.12.89.
2017, p. 3. https://www.iea.org/reports/world-energy-outlook-2018.
[10] B. Veriansyah, J.Y. Han, S.K. Kim, S.A. Hong, Y.J. Kim, J.S. Lim, Y.W. Shu, S.G. Oh,
[3] Air Transprt Action Group (ATAG), Aviation Benefits beyond Borders, 2016.
J. Kim, Production of renewable diesel by hydroprocessing of soybean oil:
https://aviationbenefits.org/media/149668/abbb2016_full_a4_web.pdf.
effect of catalysts, Fuel 94 (2012) 578e585, https://doi.org/10.1016/
[4] W. Wang, L. Tao, J. Markham, Y. Zhang, E. Tan, L. Batan, M. Biddy, W. Wang,
j.fuel.2011.10.057.
L. Tao, Y. Zhang, E. Tan, E. Warner, M. Biddy, Review of Biojet Fuel Conversion
[11] C. Kordulis, K. Bourikas, M. Gousi, E. Kordouli, A. Lycourghiotis, Development
Technologies, vol. 98, 2016, https://doi.org/10.1016/j.rser.2015.09.016. Www.
of nickel based catalysts for the transformation of natural triglycerides and
Nrel.Gov/Publications.
related compounds into green diesel: a critical review, Appl. Catal. B Environ.
[5] K. Nanou, T. Roukas, Waste cooking oil: a new substrate for carotene pro-
181 (2016) 156e196, https://doi.org/10.1016/J.APCATB.2015.07.042.
duction by Blakeslea trispora in submerged fermentation, Bioresour. Technol.
[12] B. Ho
cevar, M. Grilc, M. Hu
s, B. Likozar, Mechanism, ab initio calculations and
203 (2016) 198e203, https://doi.org/10.1016/J.BIORTECH.2015.12.053.
microkinetics of hydrogenation, hydrodeoxygenation, double bond migration
[6] M.G. Kulkarni, A.K. Dalai, Waste cooking OilAn economical source for bio-
and cisetrans isomerisation during hydrotreatment of C6 secondary alcohol
diesel: a review, Ind. Eng. Chem. Res. 45 (2006) 2901e2913, https://doi.org/
species and ketones, Appl. Catal. B Environ. 218 (2017) 147e162, https://
10.1021/ie0510526.
doi.org/10.1016/J.APCATB.2017.06.046.
[7] G. Liu, B. Yan, G. Chen, Technical review on jet fuel production, Renew. Sus-
[13] B. Ho
cevar, M. Grilc, M. Hu
s, B. Likozar, Mechanism, ab initio calculations and
tain. Energy Rev. 25 (2013) 59e70, https://doi.org/10.1016/j.rser.2013.03.025.
microkinetics of straight-chain alcohol, ether, ester, aldehyde and carboxylic
[8] M. Ameen, M.T. Azizan, S. Yusup, A. Ramli, M. Yasir, Catalytic
acid hydrodeoxygenation over Ni-Mo catalyst, Chem. Eng. J. 359 (2019)
 E. Barbera et al. / Renewable Energy 160 (2020) 428e449 449

1339e1351, https://doi.org/10.1016/J.CEJ.2018.11.045. 2012.

[14] M. Ojeda, N. Osterman, G. Dra
zi

c, L. Fele Zilnik, A. Meden, W. Kwapinski, [38] R. Turton, R.C. Bailie, W.B. Whiting, J.A. Shaeiwitz, D. Bhattacharyya, Analysis,
A.M. Balu, B. Likozar, N. Novak Tu
sar, Conversion of palmitic acid over Bi- Synthesis, and Design of Chemical Processes, Prentice-Hall, 2012.
functional Ni/ZSM-5 catalyst: effect of stoichiometric Ni/Al molar ratio, Top. [39] C. San Marchi, B.P. Somerday, Technical Reference for Hydrogen Compatibility
Catal. 61 (2018) 1757e1768, https://doi.org/10.1007/s11244-018-1046-7. of Materials, Sandia Tech. Rep., 2012, p. 292. http://energy.sandia.gov/wp/wp-
[15] T.N. Kalnes, M.M. McCall, D.R. Shonnard, Renewable diesel and jet-fuel pro- content/gallery/uploads/SAND2012_7321.pdf.
duction from fats and oils, Thermochem. Convers. Biomass Liq. Fuels Chem. [40] Last access: 06 March 2018), (n.d.), http://www.chemengonline.com/.
(2010) 468e495, https://doi.org/10.1039/9781849732260-00468. [41] V.V. Ranade, R. V Chaudhari, P.R. Gunjal, Trickle Bed Reactors: Reactor Engi-
[16] M.J. McCall, J.A. Kocal, A. Bhattacharyya, T.N. Kalnes, T.A. Brandvold, Produc- neering and Applications, Elsevier, Amsterdam, 2011, https://doi.org/10.1016/
tion of aviation fuel from renewable feedstock, 2011. US 8,039,682 B2, https:// B978-0-444-52738-7.10012-9.
patentimages.storage.googleapis.com/3d/d5/a0/ad92fb84cd73b6/US8039682. [42] G. Eigenberger, Fixed-Bed Reactors, 2012, https://doi.org/10.1002/
pdf. 14356007.b04.
[17] J.A. Petri, T.L. Marker, Production of diesel fuel from biorenewable feedstocks, [43] C.C. Herron Jr., B.K. Kruelskie, J.R. Fair, Hydrodynamics and mass transfer on
U. S. Jpn. Outlook 7 (2009) 511, 181 B2. three-phase distillation trays, AIChE J. 34 (1988) 1267e1274, https://doi.org/
[18] A. Asiedu, E. Barbera, R. Naurzaliyev, A. Bertucco, S. Kumar, Waste cooking oil 10.1002/aic.690340805.
to jet-diesel fuel range using 2-propanol via catalytic transfer hydrogenation [44] G.D. Ulrich, P.T. Vasudevan, How to estimate utility costs, Chem. Eng. 113
reactions, Biofuels (2019) 1e14, https://doi.org/10.1080/ (2006) 66e69.
17597269.2018.1532754, 0. [45] M. Pearlson, C. Wollersheim, J. Hileman, A techno-economic review of
[19] R.A.W. Johnstone, A.H. Wilby, I.D. Entwistle, Heterogeneous catalytic transfer hydroprocessed renewable esters and fatty acids for jet fuels production,
hydrogenation and its relation to other methods for reduction of organic Biofuels, Bioprod. Biorefining. 7 (2013) 89e96.
compounds, Chem. Rev. 85 (1985) 129e170, https://doi.org/10.1021/ [46] (EIA) U.S. Energy Information Administration (Last access 06 March 2018),
cr00066a003. (n.d.), https://www.eia.gov/.
[20] K.M. Isa, T.A.T. Abdullah, U.F.M. Ali, Hydrogen donor solvents in liquefaction of [47] World Population Review: the 200 Largest Cities in the United States by
biomass: a review, Renew. Sustain. Energy Rev. 81 (2018) 1259e1268, https:// Population 2019, 2019. http://worldpopulationreview.com/us-cities/.
doi.org/10.1016/J.RSER.2017.04.006. accessed April 4, 2019.
[21] T. Kilpio€, P. Ma €ki-Arvela, T. Salmi, D.Y. Murzin, Modeling of the catalytic [48] J. Miller, Determining the Rate of Used Cooking Oil Output by the Restaurant
deoxygenation of fatty acids in a packed bed reactor, in: Multiph. Catal. React, Industry in the Salt Lake Valley, 2007. UT.
John Wiley & Sons, Ltd, 2016, pp. 365e376, https://doi.org/10.1002/ [49] A. Radich, Biodiesel Performance, Costs, and Use, 2006.
9781119248491.ch15. [50] N. Brett, S. Searle, Projected Availability of Fats, Oils, and Greases in the, U.S.,
[22] Y. Xiang, X. Li, C. Lu, L. Ma, Q. Zhang, Water-improved heterogeneous transfer 2016.
hydrogenation using methanol as hydrogen donor over Pd-based catalyst, [51] U.S. Department of Tranportation (DOT), National Highway Traffic Safety
Appl. Catal. A Gen. 375 (2010) 289e294, https://doi.org/10.1016/ Administration, Commercial Medium- and Heavy-Duty Truck Fuel Efficiency
J.APCATA.2010.01.004. Technology Study e Report #1, 2015.
[23] A.S. Amarasekara, M.A. Hasan, Pd/C catalyzed conversion of levulinic acid to g- [52] The state of fuel economy in trucking (n.d.), https://www.geotab.com/truck-
valerolactone using alcohol as a hydrogen donor under microwave conditions, mpg-benchmark/. accessed August 15, 2019.
Catal. Commun. 60 (2015) 5e7, https://doi.org/10.1016/ [53] M. Brander, R. Tipper, C. Hutchinson, G. Davis, Consequential and Attributional
J.CATCOM.2014.11.009. Approaches to LCA: a Guide to Policy Makers with Specific Reference to
[24] P. Panagiotopoulou, N. Martin, D.G. Vlachos, Effect of hydrogen donor on Greenhouse Gas LCA of Biofuels, 2008. http://www.globalbioenergy.org/
liquid phase catalytic transfer hydrogenation of furfural over a Ru/RuO2/C uploads/media/0804_Ecometrica_-_Consequential_and_attributional_
catalyst, J. Mol. Catal. A Chem. 392 (2014) 223e228, https://doi.org/10.1016/ approaches_to_LCA.pdf.
J.MOLCATA.2014.05.016. [54] International Organization for Standardization, ISO 14044:2006(E): Environ-
[25] O.J. Guerra, G.A.C. Le Roux, Improvements in petroleum refinery planning: 2. mental Management e Life Cycle Assessment e Requirements and Guidelines,
Case studies, Ind. Eng. Chem. Res. 50 (2011) 13419e13426, https://doi.org/ 2006.
10.1021/ie200304v. [55] M. Spielmann, C. Bauer, R. Dones, M. Tuchschmid, Transport Services, Data
[26] K.F. Haigh, G.T. Vladisavljevi c, J.C. Reynolds, Z. Nagy, B. Saha, Kinetics of the v2.0, Ecoinvent Report No. 14, Dübendorf CH, Switzerland, 2007.
pre-treatment of used cooking oil using Novozyme 435 for biodiesel pro- [56] R. Dones, C. Bauer, R. Bollinger, B. Burger, T. Heck, A. Ro€der, M.F. Emmenegger,
duction, Chem. Eng. Res. Des. 92 (2014) 713e719, https://doi.org/10.1016/ R. Frischknecht, N. Jungbluth, M. Tuchscmid, Life Cycle Inventories of Energy
J.CHERD.2014.01.006. Systems: Results for Current Systems in Switzerland and Other UCTE Coun-
[27] Y. Liu, R. Sotelo-Boya s, K. Murata, T. Minowa, K. Sakanishi, Production of Bio- tries, Ecoinvent Report No. 5, Dübendorf CH, Switzerland, 2007.
Hydrogenated diesel by hydrotreatment of High-Acid-Value waste cooking oil [57] J. Sheehan, V. Camobreco, J. Duffield, M. Graboski, H. Shapouri, An Overview of
over ruthenium catalyst supported on Al-Polyoxocation-Pillared montmoril- Biodiesel and Petroleum Diesel Life Cycles, a Joint Study by, U.S. Department
lonite, Catalysts 2 (2012) 171e190, https://doi.org/10.3390/catal2010171. of Agriculture (DOA) and U.S. Department of Energy (DOE) prepared by Na-
[28] L.F. Chuah, A.R.A. Aziz, S. Yusup, A. Bokhari, J.J. Kleme
s, M.Z. Abdullah, Per- tional Renewable Energy Laboratory (NREL), 1998.
formance and emission of diesel engine fuelled by waste cooking oil methyl [58] R. Hischier, Life Cycle Inventories of Packaging and Graphical Paper, Data v2.0,
ester derived from palm olein using hydrodynamic cavitation, Clean Technol. Ecoinvent Report No. 11, Dübendorf CH, Switzerland, 2007.
Environ. Policy 17 (2015) 2229e2241, https://doi.org/10.1007/s10098-015- [59] A.J. Papa, Propanols, Ullmann’s Encyclopedia of Industrial Chemistry, Wiley-
0957-2. VCH Verlag GmbH & Co. KGaA, 2000.
[29] M. Toba, Y. Abe, H. Kuramochi, M. Osako, T. Mochizuki, Y. Yoshimura, [60] H.-J. Althaus, R. Hischier, M. Osses, A. Primas, S. Hellweg, N. Jungbluth,
Hydrodeoxygenation of waste vegetable oil over sulfide catalysts, Catal. Today M. Chudacoff, Life Cycle Inventories of Chemicals, Data v2.0, Ecoinvent Report
164 (2011) 533e537, https://doi.org/10.1016/J.CATTOD.2010.11.049. No. 8, Dübendorf CH, Switzerland, 2007.
[30] P.L. Chu, C. Vanderghem, H.L. MacLean, B.A. Saville, Process modeling of [61] M. Voll, P. Kleinschmit, Carbon e 6: carbon black, Ullmann’s Encyclopedia of
hydrodeoxygenation to produce renewable jet fuel and other hydrocarbon Industrial Chemistry, Wiley-VCH Verlag GmbH & Co. KGaA, 2002.
fuels, Fuel 196 (2017) 298e305, https://doi.org/10.1016/j.fuel.2017.01.097. [62] W. Gendorf, Umwelterkl€ arung, 2000.
[31] M. Rigutto, Cracking and Hydrocracking, 2010, https://doi.org/10.1002/ [63] P.A. Renzulli, B. Notarnicola, G. Tassielli, G. Arcese, R. Di Capua, Life cycle
9783527630295.ch18. assessment of steel produced in an Italian integrated steel mill, Sustain. 8
[32] G. Valavarasu, M. Bhaskar, K.S. Balaraman, Mild hydrocrackingda review of (2016), https://doi.org/10.3390/su8080719.
the process, catalysts, reactions, kinetics, and advantages, Petrol. Sci. Technol. [64] F. Mueller-Langer, E. Tzimas, M. Kaltschmitt, S. Peteves, Techno-economic
21 (2003) 1185e1205, https://doi.org/10.1081/LFT-120017883. assessment of hydrogen production processes for the hydrogen economy for
[33] J. Han, A. Elgowainy, H. Cai, M.Q. Wang, Life-cycle analysis of bio-based the short and medium term, Int. J. Hydrogen Energy 32 (2007) 3797e3810,
aviation fuels, Bioresour. Technol. 150 (2013) 447e456, https://doi.org/ https://doi.org/10.1016/J.IJHYDENE.2007.05.027.
10.1016/J.BIORTECH.2013.07.153. [65] https://www.icis.com/ (Last access: 06 March 2018), (n.d.).
[34] L. Tao, A. Milbrandt, Y. Zhang, W.C. Wang, Techno-economic and resource [66] U.S. Bureau of Labor Statistics. https://www.bls.gov/home.htm (Last access 15
analysis of hydroprocessed renewable jet fuel, Biotechnol. Biofuels 10 (2017) March 2018), (n.d.).
1e16, https://doi.org/10.1186/s13068-017-0945-3. [67] H.-J. Althaus, C. Bauer, G. Doka, R. Dones, R. Frischknecht, S. Hellweg,
[35] J.H. Gary, G.E. Handwerk, M.J. Kaiser, Petroleum Refining: Technology and S. Humbert, N. Jungbluth, T. Ko €llner, Y. Loerincik, M. Margni, T. Nemecek,
Economics, CRC Press, 2007. Implementation of Impact Assessment Methods, Data v2.2, Dübendorf CH,
[36] S. Peramanu, B.G. Cox, B.B. Pruden, Economics of hydrogen recovery processes Switzerland, 2010. Ecoinvent report no. 3.
for the puri_cation of hydroprocessor purge and o}!gases, Int. J. Hydrogen [68] Y. Zhang, L.M. Colosi, Practical ambiguities during calculation of energy ratios
Energy 24 (1999) 405e424. and their impacts on life cycle assessment calculations, Energy Pol. 57 (2013)
[37] G. Towler, R.K. Sinnot, Chemical Engineering Design: Principles, Practice and 630e633, https://doi.org/10.1016/J.ENPOL.2013.02.039.
Economics of Plant and Process Design, second ed., Butterworth-Heinemann,