Citation
The effects of molecular saturation of feedstocks on renewable diesel production in the absence of hydrogen gas

Material Information

Title:
The effects of molecular saturation of feedstocks on renewable diesel production in the absence of hydrogen gas
Creator:
Dubinsky, Jonathan
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file : ;

Subjects

Subjects / Keywords:
Feedstock ( lcsh )
Biodiesel fuels ( lcsh )
Biomass energy ( lcsh )
Biodiesel fuels ( fast )
Biomass energy ( fast )
Feedstock ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Catalytic deoxygenation of fatty acids was carried out using Palladium on Carbon (Pd/C) and Magnesium Oxide (MG70) catalysts and a Parr bench top micro reactor. The effect of feedstock saturation on the decarboxylation reaction was measured by varying the feedstock from stearic acid (fully saturated) to oleic acid (mono-unsaturated) to linoleic acid (poly-unsaturated). The prevalence of the decarboxylation reaction was measured by analyzing both the liquid products and the gaseous products. The liquid products were collected and analyzed for the major product, heptadecane (C17), as well as straight chain alkane in the diesel range content using a gas chomatograph with a flame ionization detector. Concentration of CO2 in the gaseous product was analyzed using gas chromatography. As saturation increased, the concentration of CO2 in the gas phase increased, showing that decarboxylation occurred. The analysis of the liquid products showed as feedstock saturation increased the concentration of saturated straight chain hydrocarbons in the diesel fuel range with high selectivity to C17 heptadecane increased, confirming that the decarboxylation reaction favors more saturated feedstocks.
Thesis:
Thesis (M.S.)--University of Colorado Denver. Civil engineering
Bibliography:
Includes bibliographical references.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Jonathan Dubinsky.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
862934824 ( OCLC )
ocn862934824

Downloads

This item has the following downloads:


Full Text
THE EFFECTS OF MOLECULAR SATURATION
OF FEEDSTOCKS ON RENEWABLE DIESEL
PRODUCTION IN THE ABSENCE OF HYDROGEN GAS
By
Jonathan Dubinsky
B.S., University of Kansas, Environmental Science
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Masters of Engineering
Civil Engineering
2013


This Thesis for the Master of Engineering degree by
Jonathan Dubinsky
has been approved for the
Civil Engineering Program
by
Arunprakash Karunanithi, Chair
Ron Roher
Jason Ren
April 16, 2013


Dubinsky, Jonathan (MEng, Civil Engineering)
The Effects of Molecular Saturation of Feedstocks on Renewable Diesel Production in
the Absence of Hydrogen Gas
Thesis directed by Assistant Professor Arunprakash Karunanithi
ABSTRACT
Catalytic deoxygenation of fatty acids was carried out using Palladium on Carbon
(Pd/C) and Magnesium Oxide (MG70) catalysts and a Parr bench top micro reactor. The
effect of feedstock saturation on the decarboxylation reaction was measured by varying
the feedstock from stearic acid (fully saturated) to oleic acid (mono-unsaturated) to
linoleic acid (poly-unsaturated). The prevalence of the decarboxylation reaction was
measured by analyzing both the liquid products and the gaseous products. The liquid
products were collected and analyzed for the major product, heptadecane (C17), as well as
straight chain alkane in the diesel range content using a gas chomatograph with a flame
ionization detector. Concentration of CO2 in the gaseous product was analyzed using gas
chromatography. As saturation increased, the concentration of CO2 in the gas phase
increased, showing that decarboxylation occurred. The analysis of the liquid products
showed as feedstock saturation increased the concentration of saturated straight chain
hydrocarbons in the diesel fuel range with high selectivity to C17 heptadecane increased,
confirming that the decarboxylation reaction favors more saturated feedstocks.
The form of and content of this abstract are approved. I recommend its publication.
Approved: Arunprakash Karunanithi
111


DEDICATION
I dedicate this work to my loving and supportive sweetheart, Elizabeth. She has
encouraged me to always pursue my dreams with confidence and grace.


ACKNOWLEDMENTS
It is with great honor that I am able to contribute this paper to the scientific
community. As an owner and operator of a vegetable oil powered educational bus, I have
seen first hand the power and limitations of alternative fuel sources. I would like to thank
Dr. Arun Karunanithi for going out on a limb and taking on a greasy mechanic like
myself. I would like to thank my parents for being the change that they want to see in the
world, and always telling me that I can do what ever I put my mind to. Thank you to all
the merry pranksters and joy seekers in the world...you know who you are. I plan to not
only work to advance the field of bio-fuels technology, but to educate the next generation
with the concepts of sustainability and biophilia. Long live spaceship Earth.
v


TABLE OF CONTENTS
CHAPTER
I EXPERIMENTAL............................................................1
Introduction...........................................................1
Background and Problem..............................................1
Biomass Based Diesel/Renewable Diesel...............................3
Purpose of the Project..............................................5
Experimental...........................................................7
Methods.............................................................7
Chemicals and Catalysts...........................................7
Decarboxylation Reaction..........................................8
Analysis of Products..............................................9
Results...............................................................10
Effects of Feedstock Saturation on Liquid Products.................10
Effects of Feedstock Saturation on Gaseous Products................15
Discussion............................................................16
II ENVIRONMENTAL ASSESMENT................................................18
Introduction..........................................................18
Goal and Scope........................................................19
System Boundaries..................................................19
Process flow diagram...............................................20
Experimental Lab data............................................21
Feedstock........................................................22
vi


Life Cycle Inventory....................................................22
Allocations...........................................................22
Method of Data Collection.............................................22
Decarboxylation Reaction Methods......................................23
Conversion Methods....................................................24
Environmental Impact Assessment.........................................24
Interpretation..........................................................28
Goal and Scope........................................................28
Results and Discussion..................................................29
III CONCLUSIONS..............................................................31
Significant Findings....................................................31
Next Steps..............................................................32
REFERENCES...................................................................34
APPENDIX.....................................................................38
Soybean Agriculture Unit Process Data...................................38
Soybean Transport Unit Process Data.....................................39
Soybean Crushing Unit Process Data......................................40
Soybean Oil Transport Unit Process Data.................................41
Soybean Oil Conversion to Renewable Diesel using Hydrogen Process.......42
Soybean Oil Conversion to Renewable Diesel using Non-Hydrogen Process..44
vii


LIST OF TABLES
Table
1 This table show levels of saturation in some common feedstock oils.............7
2 Concentration in mg/1 of straight chain alkanes from C7 to C17 in the liquid products. 12
3 A collection of calculations needed to compute environmental impacts of biofuels. .. 24
4 Cradle-to-grave comparison of petroleum diesel, HDO renewable diesel and non-
hydrogen renewable diesel.....................................................27
5 Soybean agriculture raw materials data.........................................38
6 Soybean agriculture emissions data.............................................38
7 Soybean transport raw materials data...........................................39
8 Soybean transport emissions data................................................39
9 Soybean crushing raw materials data............................................40
10 Soybean crushing emissions data................................................41
11 Soybean transport data raw materials..........................................41
12 Soybean transport emissions data...............................................42
13 HDO raw materials data assuming an 84% yield..................................43
14 HDO emissions data assuming an 84% yield......................................43
15 Non-hydrogen raw materials data assuming a 45% yield..........................44
16 Non-hydrogen emissions data assuming a 45% yield..............................44
viii


LIST OF FIGURES
Figure
1 Chemical formulas for three common biofuels: Biodiesel, Ethanol, and Renewable
Diesel........................................................................3
2 This diagram, which was extracted from Bambang etal, 2011, shows the stoichiometry
of the renewable diesel reaction..............................................4
3 Summary of total expenses in a current HDO renewable diesel processing plant....5
4 Stoichiometric equation for the decarboxylation reaction........................6
5 This diagram, extracted and modified by from Kubickova, 2007, is of the Parr 4590
bench top reactor setup used in the experiment................................9
6 These figures show the yield of both total fuel range alkanes and the major product
heptadecane..................................................................12
7 GC chromatograms of alkane reaction products at 673 K...........................13
8 Concentration of liquid Alkanes in the liquid product...........................14
9 This chart shows the selectivity to C17 heptadecane in the total alkane product.14
10 Real time data of produced gas from the reaction..............................15
11 This chart show the concentration of CO2 in the gaseous products from GC......16
12 Process Flow Diagram..........................................................20
13 Global Warming Potential Indicator............................................25
14 Carcinogenic Effects Indicator................................................26
15 Aquatic Eco Toxicity Indicator................................................26
16 This chart show the life cycle GHG emissions from cradle to grave of three pathways
to fuel production normalized to petroleum diesel............................27
IX


17 C02e emissions from only the processing stage of three possible pathways to diesel
fuel.
28
x


CHAPTER I
EXPERIMENTAL
Introduction
Background and Problem
If you are in a shipwreck and all the boats are gone, a piano top . that comes
along makes a fortuitous life preserver. But this is not to say that the best way to design a
life preserver is in the form of a piano top. I think that we are clinging to a great many
piano tops in accepting yesterdays fortuitous contrivings.
Buckminster Fuller [Critical Path, 1981]
Society is at a transition point and now we have an opportunity to re-design our
energy future. With the uncertainty of the worlds oil supply, economic crisis, and global
climate change all on the horizon, a suitable fuel replacement must be identified. Wind,
solar, fuel cells, geo-thermal and other renewable energy sources will play a part in a
diverse energy portfolio, but liquid fuels which work with current infrastructure must be
utilized during the transition. Bio-fuels are a potential fit for a liquid fuel replacement,
but rigorous study needs to be done to determine the true costs. An exponential increase
in the consumption of such biofuels throughout the world has taken place in the past few
years [Serrano-Ruiz etal, 2012],
Bio-fuels have existed for thousands of years. The first fuel source for humans
was wood, which we utilized for heat and to cook our food. Wood, or cellulose, contains
carbon, which is sequestered from the air using photosynthesis. As technology and
human ingenuity grew, we figured out how to extract fats and oils from plants and
1


animals. This became our primary light source, which was a great advancement for
humanity because it meant we did not have to use fire in the summer time just for a little
light. Health wise, candles were a great improvement over the soot and smoke from a
fire [Barnes etal, 2005], Human kind, however, did not abandon the use of fire once they
discovered candles. There was a need and a use for both. It was not until thousands of
years later, with the discovery of fossil fuels, that civilization resorted to a one-size fits
all fuel source.
Rudolf Diesel observed and disapproved of this back in the 1890s. His engine
did not use petroleum but locally grown peanut oil. The use of vegetable oils for engine
fuels may seem insignificant today, Diesel said in 1912, but such oils may become in
the course of time as important as the petroleum and coal tar products of the present
time [Paula etal, 2011], In the late 1800s, other engines were released on the market
that ran on gasoline alone. However, the diesel engine could operate on many different
fuels sources including coal dust, animal fats, vegetable oil, and petroleum. The problem
that we face today is a growing demand for fossil fuels with a shrinking supply. Our
economy is dependent on growth, which results in increased energy consumption. Since
the early 20th century we have been designing our systems to be reliant on one source of
fuel, oil, unlike Rudoulf Diesels engine. The transportation sector of our society heavily
relies on petroleum, which accounts for essentially all (96%) of the transportation energy
[Serrano-Ruiz etal, 2012], The natural world evolved to have biodiversity and resilience,
the measure of which is an ecosystems ability to bounce back after a catastrophic event
like a hurricane or a volcano. This ability to bounce back is now our great struggle as a
human nation. We are now faced with the challenge of coming up with replacements to a
2


cheap hydrocarbon based system or face the slow sinking of our great ship and the frantic
clinging to piano tops as we struggle to survive.
Biomass Based Diesel/Renewable Diesel
The commercial market for biodiesel is approximately 2.5 billion gallons per year
and will grow rapidly in the near future due to the Renewable Fuels Standards (RFS-2),
which establishes renewable fuel consumption and production mandates in the United
States [EPA, 2012], The global market for biofuels is expected to reach $280 billion by
the year 2022 [Solecki etal, 2011; REN21, 2011], Renewable diesel, or Biomass Based
Diesel, refers to fuel produced from biomass that has as least a 50% reduction in life
cycle C02 emissions as compared to petroleum and can be used as a direct substitute (not
blended) in internal combustion engines such as diesel, gasoline, or jet [EPA, 2012], The
chemical makeup of renewable diesel is exactly identical to petroleum, unlike other
biofuels such as biodiesel and ethanol (see figure 1).
(rtf h**
Biodiesel
"OH
Ethanol
Petroleum and
Renewable Diesel
Figure 1 Chemical formulas for three common biofuels: Biodiesel, Ethanol, and
Renewable Diesel
The most commonly known biofuel today is biodiesel or F. A.M.E (fatty acid
methyl esters). This fuel has a chemical makeup that is compatible with diesel engines,
but not identical to petroleum diesel [Mikulec etal, 2010], In recent years a new biofuel
based on pure hydrocarbons has been heavily researched. This differs from biodiesel
significantly in molecular structure and production process. From an environmental


standpoint, both renewable diesel and biodiesel outperform petroleum diesel in all
categories of emissions, including particulates, unburned hydrocarbons, NOx, SOx, and
carbon monoxide [Kalnes etal, 2009], This new class of second generation biofuels is
referred to as Biomass Based Diesel or Renewable Diesel in the EPAs Renewable Fuel
Standards (RFS2) [EPA, 2012], These fuels can come from diverse feedstocks such as
vegetable oil, fats, algae oil, and waste oil. As opposed to biodiesel, which can only be
used in a diesel engine, renewable diesel can be refined to make jet fuel, gasoline, green
napthala, and other hydrocarbon based products [Knothe etal, 2010], There are three
major chemical pathways to remove oxygen from the fatty acids founds in plant based
lipids, and hence produce renewable hydrocarbons. Those pathways are
hydrodeoxygentation, decarboxylation, and decarbonylation (see figure 2).
O
H2C-C-0-CnH2(n.x)+1 (=)
I Q
HC-COCnH2(n.y)+1 (=)
0
I II
H2C C O Cr,H2(n.z)+1 (-)
Triglycerides
H,CC-OCnH,r
HC-C-0-CnH2n+1 (-)
P
H2CCOCnH2n+| (-)
Hydrogenated triglycerides
H2
Cracking
O
II
HC OCnH2n+i + CH3CH2CH3
Free fatty acids
n : odd number
x, y, z : number of double bonds
=: double bond
single bond
Aromatics ^------------ n-CnH2n+2 + C02 n-CnH2n+2 + CO + HoO n-Cn+iH2n+4+ 2H20
Cyclization
Isomerization Cracking Isomerization
iso-CnH2n+2 Lighter hydrocrabons iso-Cn+|H2n+4
Figure 2 This diagram, which was extracted from Bambang etal, 2011, shows the
stoichiometry of the renewable diesel reaction
The typical cascade reaction for upgrading triglycerides to hydrocarbon liquid fuels with
three main routes: Decarboxylation, Decarbonylation, and Hydrodeoxygentation.
The major body of research on renewable diesel, and all current commercial
endeavors, has been focused on the hydrodeoxygentation pathway, since it is most similar
4


to current refinery technology and has provided the highest yields of hydrocarbons
[Simacek etal, 2009; Huber etal, 2007; Veriansyah etal, 2012], Other studies have looked
just at the decarboxylation route or the decarbonylation route [Na etal, 2012;
Rozmyszowicz etal, 2011], One appeal to the decarboxylation route is that it does not
consume hydrogen gas in the reaction, which is expensive and mainly produced from
non-renewable sources [Abbasi etal, 2011], Figure 3 clearly shows the high cost of
hydrogen gas in the breakdown of capex and opex for a current commercial renewable
diesel plant [Pearlson, 2011],
19% 26%
Figure 3 Summary of total expenses in a current HDO renewable diesel processing
plant.
For each tonne of hydrogen produced from natural gas, approximately 2.5 tonnes
of carbon is released as C02. When hydrogen is produced from coal, approximately 5
tonnes of carbon is emitted to the atmosphere per tonne of hydrogen. Currently over 90%
of our hydrogen production comes from fossil fuels (coal, natural gas, and petroleum)
[Abbasi etal, 2011],
Purpose of the Project
The goal of the project is to determine a safe, economic, and environmentally
friendly means to convert biomass into renewable diesel that does not require large
5


amounts of hydrogen gas and hence is less expensive and safer. The project aims to
assess the benefits of catalytic deoxygenation to convert lipids from non-food oil crops
(e.g. camelina, castor beanjatropha, and algae and waste biomass) into renewable diesel
fuel through an energy densification process using thermal decarboxylation, which does
not utilize molecular hydrogen, to deoxygenate lipids to hydrocarbon fuels. The reaction
scheme for decarboxylation is presentenced in figure 4.
O
II
R'C OH
r-h + eo2
Figure 4 Stoichiometric equation for the decarboxylation reaction
This research looked at the impacts of lipid saturation on the extent of the
decarboxylation reaction. Since feedstocks vary dramatically in fatty aid makeup, it was
important for me to understand how saturation will impact the selectivity of the reaction
towards our desired product. With this understanding it will allow operators to select
appropriate feedstocks in their area as well as look into possible pre-treatment steps that
end in desired saturation levels (see table 1). Algae oil also has the potential to be breed
for desired traits such as saturation level [Schlagermann etal, 2010],
6


Table 1 This table show levels of saturation in some common feedstock oils
Saturated Monounsaturated Polyunsaturated
g/100g g/100g g/1 OOg
Animal fats
Lard 40.8 43.8 9.6
Duck fat 33.2 49.3 12.9
Butter 54 19.8 2.6
Vegetable fats Coconut oil 85.2 6.6 1.7
Palm oil 45.3 41.6 8.3
Cottonseed oil 25.5 21.3 48.1
Wheat germ oil 18.8 15.9 60.7
Soya oil 14.5 23.2 56.5
Olive oil 14 69.7 11.2
Com oil 12.7 24.7 57.8
Sunflower oil 11.9 20.2 63
Safflower oil 10.2 12.6 72.1
Hemp oil 10 15 75
Canola/Rapeseed oil 5.3 64.3 24.8
Algae Species Lipid Content Saturated Mono-Unsaturated Poly-Unsaturated
(% Dry yyt) % % %
Nannochloropsis salina 31-68 45 35.2 20
Botrvococcus braunii 25-75 26 5 62
In addition to the laboratory research, I also compared the environmental impacts
of this process to hydrodeoxygenation and petroleum diesel using life cycle analysis and
literature review. The main difference between the two pathways to renewable diesel is
the use or non-use of hydrogen gas in the bio-fuel production process. The focus of the
comparison was on energy requirements and GHG emissions of each process as well as
other inputs and outputs deemed important by the industry and EPA.
Experimental
Methods
Chemicals and Catalysts
Oleic acid, stearic acid, and linoleic acid was purchased from Sigma Aldrich and
used without further purification. A hydrocalcite solid phase powdered catalyst with
7


magnesium to alumina ratio of 70:30 (MG70) was purchased from Sasol. Palladium on
activated carbon (Pd/C) with a 10% loading rate was purchased from Acros chemicals.
Decarboxylation Reaction
Several studies have previously explored the decarboxylation approach to
renewable diesel production. The methods for this experiment were collected through
rigorous literature review [Na etal, 2010; Na etal 2012; Rozmyszowicz etal, 2011; Snare
etal, 2007; Huber etal, 2007], The decarboxylation reaction was performed using a Parr
bench top stirred reactor (4590 micro reactor with a 100 ml reaction vessel and a 4848
controller). This batch reactor was customized for this reaction and fit underneath a
chemical fume hood to evacuate any harmful gasses during sample ejection. The reactor
was designed for a maximum pressure of 345 bar and 500 C with a flat graphite gasket.
A multi blade impeller was purchased for mixing a liquid reactant with a solid phase
catalyst. The temperature and pressure were plotted using the Parr software provided
with the reactor. In one batch reaction 28g of reactant was placed in the reactor with 1,4g
of catalyst (reactant catalyst=20:1). After loading the reactant and catalyst, the reactor
was purged with nitrogen to remove any oxygen from the vessel. The purge method was
based on standard procedures from literature [Kinsley etal, 2001], After purging, the
reactor was heated to 400 C and maintained until the reaction was completed. The
reaction was monitored in real time using the pressure curve from the Parr software and
was determined to be complete when the curve reached the upper asymptote (about 3
hours at 400 C and 4 hours total). Stirring speed was maintained at 150 rotations per
minute during the reaction.
8


Figure 5 This diagram, extracted and modified by from Kubickova, 2007, is of the
Parr 4590 bench top reactor setup used in the experiment
After completion the reactor was cooled down to 70 C using a customized cold
water jacket (below this temperature unreacted stearic acid is solid). Then the gaseous
products were collected in a tedlar bag and the liquid products were collected in a glass
vial. A 1 ml sample of the liquid product was taken and dissolved in 25 ml of solvent
(50/50 Dicholomethane and Acetone) and centrifuged at 5000 rpm for 5 minutes to
separate any remaining catalyst. A polar and non-polar solvent was chosen because the
unreacted starting material, fatty acids, are polar and the product, hydrocarbons, are non-
polar.
Analysis of Products
The liquid products were first characterized using a Gas Chromatograph- Mass
Spec (Shimadzu GC/MS-QP2010) equipped with a Restek 30m x 0.25mm x 0.5um Rtx-
Wax column. The carrier gas was helium and the flow rate was 1.0 ml/min. The column
temperature was programmed to increase from 40 C to 240 C at 5 C/min. Samples were
derivitized to protect the column from unreacted carboxylic acids using
9


Dimethylformamide dimethyl acetal and then dissolved in Dichloromethane before
injection. After characterization of the products pure hydrocarbons standards, alkanes
from Ce to C30, were obtained and calibration curves were developed for each individual
alkane using a GC-FID. Peak areas and retention times were associated with know
concentrations of pure samples. The liquid products from the reaction were then
analyzed using a Thermo trace GC-FID equipped with a Restek 30m x 0.18mm x 0.18um
Rxi-lms column for quantification of desired products. The carrier gas was helium and
the flow rate was 2.75 ml/min. The column temperature was programmed to increase
from 40 C to 330 C at a 20 C/min rate. The final temperature was held for 20 minutes
and the FID base temperature was held at 350 C. The gas products were analyzed using a
gas chromotograph (SRI 8610 C) equipped with an FID detector.
Results
Effects of Feedstock Saturation on Liquid Products
The chromatograms (GC results) of the liquid product analyses are presented in
figure 7. The major product was heptadecane, C17, and other minor products were
alkanes from C(> to C17. Other hydrocarbons may be present in the product but were not
considered in this study. Alkanes are viewed as more desirable over aromatics due to
toxicity and suitability to diesel fuel [Senate, 2013; KubicAova etal, 2005], Straight
chain alkanes also have a higher cetane number than aromatics, however isomers and
cracked hydrocarbons have physical properties desirable for fuels such as lower cloud
points [Huber etal, 2007], For the purpose of this research, I only looked at straight chain
alkanes. Since the feedstocks used have a carbon count of Ci8 the main product was C17,
heptadecane, due to the decaboxylation reaction removing one carbon along with the O2
10


molecule (see fig 4). Equation 1 shows how yield was calculated in the analysis. The
reactants were fatty acids and the products were either heptadecane or total alkanes in the
fuel range. Both results are presented in figure 6. Equation 2 shows how selectivity was
calculated. Selectivity shows what percentage of the product was alkanes in the fuel
range and then what percentage of the alkanes was the major product heptadecane. To do
this, the mass concentrations from the GC were converted to volume % based on the
mass and density of the different hydrocarbons. Both yield and selectivity were looked at
in this study and the results are presented in figures 6, 8, 9.
Yield = moles of recovered products / moles of reactants (1)
Selectivity = volume of desired products / volume of total products (2)
The yield and selectivity for C17 heptadecane was much higher when the Pd/C
catalyst was used when compared to MG70. The results of the MG70 catalyst showed an
even distribution among the fuel range hydrocarbons with a slightly higher concentration
in the lighter range hydrocarbons. When Pd/C was used as the catalyst there was a clear
increase in diesel range alkanes as feedstock saturation increased. Selectivity towards
C17 also increased when stearic acid was used as a feedstock when compared to oleic and
linoleic acids. In addition to an overall increase in alkane yield and higher selectivity
towards heptadecane, stearic acid had no peak for unreacted parent material in the
chromatorgram (fig 7). Overall Pd/C produced a higher concentration of diesel range
hydrocarbons when compared to MG70. The concentrations of alkanes in the total liquid
products (selectivity) from various experiments are presented in table 2 and figure 8.
11


Figure 8 represents various jars of product and shows what percentage of the jar is diesel
range alkanes. It can be seen that as feedstock saturation increases, the percentage of the
liquid product that is alkanes increased. Figure 9 is a similar chart, but looks at the
selectivity of heptadecane amongst the alkanes. Again it can be seen as feedstock
saturation increases, the percentage of the alkanes that is heptadecane increases. The
selectivity for heptadecane when MG70 was used as a catalyst was non-existent. The
product yield is also significantly lower then the product catalyzed by Pd/C.
Table 2 Concentration in mg/1 of straight chain alkanes from C7 to C17 in the liquid
products.
Stearic (Pd/C) 1 Oleic (Pd/C) I Linoleic (Pd/C) I 100% Oleic (MG70)
mg/l mg/l mg/l mg/l
Heptane 14029.048 15088.5805 8049.1225 14824.525
Octane 15839.248 15645.5755 8222.62675 16526.4135
Nonane 15218.82175 14777.687 6046.99825 11269.644
Decane 15031.03425 14747.7495 7193.92525 6221.37675
Undecane 13734.25925 13071.76 4612.82525 5830.232
Dodecane 13050.8775 10249.18875 3692.4235 5382.75375
Tridecane 12606.30125 9525.33875 3549.927 4092.063
Tetradecane 11817.1735 8099.932 2410.85525 3892.421
Pentadecane 14395.158 6995.4755 4898.19325 5864.8425
Hexadecane 5651.258 5241.062 3156.19 7240.094
Heptadecane 369051.304 211211.624 19603.772 5153.536
Total 500424.4835 324653.9735 71436.859 86297.9015
Total in (g/l) 500.4244835. 324.6539735 71.436859 86.2979015
Molar Yield of Alkanes
70
60 j
20
10
l
l
0
saturated (Pd/ mono-
C) unsaturated
(Pd/C)
poly-
unsaturated
(Pd/C)
(
mon-
unsatu rated
(MG70)
Molar Yield of Heptadecane
4S
40 £
35
?
20
I
* 15
10
i
5
0
-s
saturated (Pd/ mono-
C) unsaturated
(Pd/C)
poly-
unsaturated
(Pd/C)
-5-
mon-
unsaturated
(MG70)
Figure 6 These figures show the yield of both total fuel range alkanes and the major
product heptadecane.
12


6 7 8 9 10 11 12
6 7 8 9 10 11 12
- y
12
6 7 8 9 10 11
I i i i t
12
67 89 10 11 I
I I i i 1 .
13
Oleic (MG70)
13
Linolcic (Ed/C)
Oleic fPdO
Stearic CPd/Cl
Figure 7 GC chromatograms of alkane reaction products at 673 K
(l:hexane; 2:heptane; 3:octane; 4:nonane; 5: decane; 6: dodecane, 7: undecane; 8:
tridecane; 9: tetradecane; 10: pentadecane; 11: hexadecane; 12: heptadecane; 13: oleic
and stearic acid).
13


The Percentage of the Liquid Product that
is Alkanes
100
1
?
Q.
2
90
80
70
60
SO
| 40
o 30
* ~
10
0 Saturated (Pd/ mono- poly- mono-
C) unsaturated unsaturated unsaturated
(Pd/C) (Pd/C) (MG70)
Other 34.691203 57.33785 90.38017473 88.16811627
-Alkanes 65.308797 42.66215 i 9.619825267 11.83188373
Figure 8 Concentration of liquid Alkanes in the liquid product
Each of the bars represent one liter of liquid product. The blue section of the bar is
identified and quantified diesel range alkanes present in the product converted from mass
per volume to volume per volume using the density of each component. The red sections
are other compounds most likely made up of alkenes, aromatics, isomers, and unreacted
feedstock.
The percentage of the Alkane Product
that is Heptadecane
1
6.
I
*
a*
100
90
80
70
60
50 '
40
30 ;
20
10
0 *
Other Alkanes
Saturated mono poly' 100% Oleic
(Pd/C) unsaturated unsaturated (MG70)
(Pd/C) (Pd/C)
27.27327328 31.4770961 73.77202945 94.3942929
Heptadecane 72.72672672 68.5229039 26-22717055 5.605707099
Figure 9 This chart shows the selectivity to C17 heptadecane in the total alkane
product
Each of the bars represent one liter of alkane product. The blue section of the bar is
percentage of heptadecane present in the alkane product converted from mass per volume
to volume per volume using the density of heptadecane. The red sections are other
compounds made up of diesel range alkanes.
14


Effects of Feedstock Saturation on Gaseous Products
The gaseous products were collected, characterized, and quantified using GC and
the results are presented in figures 10 and 11. After the gas was collected, it was left for
12 hours to cool and acclimate to room temperature and pressure. The total quantity was
measured using the ideal gas law:
P1*V1 /T1 = P2*V2/T2
Total Gaseous Product Produced in Real Time
900
Time (min)
Figure 10 Real time data of produced gas from the reaction.
This graph shows pressure data from the reactor with varying levels of feedstock
saturation and using two different catalysts (MG 70 and Pd/C). The reactor was cleared
of all gasses before the reaction so this pressure curve relates to quantity of total gasses
produced from the reaction. [blue=stearic (Pd/C), red = oleic (Pd/C), green=linoleic
(Pd/C), purple=oleic (MG70)]
As you can see from figure 10, the pressure in the reactor increased more quickly
with increasing feedstock saturation when the Pd/C catalyst was used. When the MG70
catalyst was used the overall quantity was much lower. The concentration of CO2 in the
15


products was quantified by first injecting a known sample of CO2 into the GC to develop
a calibration curve. Then based on peak area, I was able to determine the concentration
of CO2 in the gaseous products. This value decreased as unsaturation in the feedstock
increased, showing that the decarboxylation reaction, which produces C02, was more
prevalent when a saturated feedstock was used. When MG70 was used as a catalyst, the
CO2 levels were much lower showing less decarboxylation occurred. This confirms that
unsaturation of the feedstock leads to deactivation of the catalyst as well as a suppression
of the decarboxylation reaction [Rozmyszowicz etal, 2011 pg 2818],
C02 Concentration in the Gaseous Product
Saturated (Pd/C) Moo o-Unsatu rated Poly-Unsaturated Mono-Unsaturated
(Pd/C) (Pd/C) (MG70)
Figure 11 This chart show the concentration of CO2 in the gaseous products from
GC analysis
Discussion
From the results, we see that the decarboxylation reaction was more active based
on both the liquid phase and gas phase analyses when the feedstock was more saturated.
This is because the reaction order goes from hydrogenation to isomerization to
16


decarboxylation. When double bonds are present in the feedstock, they are susceptible to
isomerization, hydrogenation, and aromatization reactions [Rozmyszowicz, 2011], These
reactions inhibit decarboxylation and therefore reduce the quantity of CO2 and suppress
the formation of desired alkanes. When oleic acid was used (one double bond) the
tendency is for the molecule to become hydrogenated from H2 produced in situ from side
reactions to stearic acid and then to saturated alkanes. This is even more apparent with
the use of linoleic acid (two double bonds). When a fully saturated feedstock such as
stearic acid is used, the decarboylation reaction happens much more quickly with fewer
side reactions. This is also why stearic acid has higher selectivity to C17 heptadecane
when Pd/C is used as a catalyst. When the feedstock is saturated, the decarboxylation
reaction happens much more quickly so cracking and isomerization are less prevalent,
leading to a more predictable product based on the stoichiometric reaction. The fact that
there is no remaining unreacted parent material in the product when stearic acid was used
with Pd/C catalyst, also shows a more complete reaction scenario. There are still many
undefined liquids in the samples that are most likely some combination of alkenes and
aromatics as well as isomers of the parent material. Overall when MG70 was used, the
reaction was much less complete and there was almost no selectivity to C17 heptadecane.
It was therefore concluded that PD/C is a much more effective catalyst for this reaction
and no further study was done using MG70.
17


CHAPTER II
ENVIRONMENTAL ASSESMENT
Introduction
This chapter takes a brief look at the environmental impacts associated with
renewable diesel production. This was not intended to be a full life cycle analysis on the
production of renewable diesel. I coupled data found in the literature on life cycle
analysis with data on yields from experimental research to compare two pathways for
renewable diesel. This study was designed to see the value proposition and
environmental potential in a hydrogen-free renewable diesel scenario. The main
difference between the two pathways is the use or non-use of hydrogen gas in the bio-fuel
production process. I focused on the non-hydrogen method because the hydrogen
process has been well documented [Pearlson etal, 2011], I have shown through laboratory
experiments and environmental assessment that a viable alternative drop in replacement
fuel can be produced with out the use of hydrogen gas, which could reduce cost and
environmental impacts.
The decarboxylation process is currently on a laboratory scale. Data was
collected in the lab and the yield of renewable hydrocarbons fuel was reported (see
chapter 1). The hydrogenation process (HDO) data was gathered from an LCA study
done by the Helsinki University of Technology on the Next BTL process at the Neste Oil
plant in Kilpilahti [Nikander, 2008],
18


Goal and Scope
This chapter is not designed to be a complete life cycle assessment, but rather a
literature review and look at the potential environmental and economic benefits of a non-
hydrogen gas process to produce renewable hydrocarbons. The purpose of this chapter is
to better understand the potential impacts of two different paths to renewable diesel
production. One is hydrodeoxygentation (HDO) process, which is well known, and the
other is a fragment of that reaction, decarboxylation, which has been isolated in this
study. Can this one reaction suffice in a commercial scale renewable diesel operation?
Will the yields be great enough to make this technology cost competitive, both
environmentally and economically, with the better-researched HDO reactor? The
stakeholders being addressed through this study are venture capitalists and the fuel
production industry. The goal is to compare the energy efficiency and environmental
impacts of both HDO and non-hydrogen processes.
System Boundaries
The system boundary for this study is from agriculture to fuel transportation (fig
11). Both processes start with agricultural production of soybean oil and end with
renewable diesel (hydrocarbon mix). One process uses only decarboxylation (no
hydrogen), and the other uses Hydrogenation/Hydrodeoxygentaion (HDO). The
functional unit for this LCA is the amount of fuel needed to travel 1 vehicle mile
travelled (VMT). This unit is based on previous LCA studies in this field and is an easy
to understand unit. For the purpose of this study I gathered data from an LCA study done
by the University of Helsinki for all data sets relating to the HDO process [Nikander,
2008], Lab scale data was used to extrapolate the yields of the decarboxylation process.
19


limitation of this work was that the carbon emissions from the conversion processes were
not considered emissions to the air. Since the emissions could be captured during the
process, it was assumed that they would be sequestered in some way.
Feedstock
The only feedstock considered was soybean oil. All data on agriculture to refined
oil came from data on soybean oil. Other feedstocks would offer different LCA results.
All upstream data on unit processes other then the conversion step was gathered from an
NREL LCA study on biodiesel [Sheehan, 1998],
Life Cycle Inventory
Allocations
In the LCA study from Helsinki, all beneficial co-products were assessed on their
energy and GHG implications by mass balance, or boundary expansion. All impacts
were assigned to the primary product. Credits were assigned based on the amount of co-
product production. The two major co-products relevant to this analysis are soy meal
from the agriculture phase, and fuel gas from the processing phase. The amount of
energy and GHG it takes to make an equivalent amount of co-product using the
traditional methods was credited to each unit process (Nikander, 2008],
Method of Data Collection
Each unit process presented in the process flow diagram is represented in its own
separate table (see appendix). The data for the unit processes for agriculture, transport,
and crushing were taken from an LCA study conducted by NREL on biodiesel production
[Sheehan, 1998], Here we assume that these processes will remain the same for
22


renewable diesel production. All data on input flows relating to the HDO process was
acquired from the LCA conducted by University of Helsinki. Output data on C02e from
the HDO process was reported from the previously mentioned LCA study conducted
Helsinki University and based on an 84% yield from pre-treated soybean oil to
hydrocarbon fuel [Nikander, 2008], Other output data for the HDO process was obtained
from the NREL LCI database [U.S. LCI Database, 2010], For the non-hydrogen reaction,
process lab scale data on potential yields of hydrocarbon fuel was coupled with the
industrial scale process data from the Helsinki LCA study to model production of 1 VMT
of fuel quantity. There are many limitations to this approach, but it is the best option
available to me since there is no commercial scale data to date on a non-hydrogen
process.
All data has been normalized using the functional unit of one vehicle mile
traveled (1 VMT). All data streams have been chosen because they are relevant to both
processes. All emissions streams were chosen based on a previous LCA study, which
examined biodiesel production. Below the data is represented in graphical form (figures
13-15). Data from the five unit processes were taken to mid point indicators using
IMPACT 2002.
Decarboxylation Reaction Methods
I used a Parr bench top reactor with a 100 ml reaction vessel to collect data on the
non-hydrogen process. During each reaction the vessel was loaded with oleic acid,
stearic acid, or linoleic acid and catalyst (20:1 ratio), and the headspace purged with
nitrogen [Kinsley, 2001], At the conclusion of the reaction, all gaseous products were
collected in a tedlar bag and analyzed using GC. The liquid products were collected and
23


analyzed using GD-FID. Based on the findings, the median conversion of a mixture of
fatty acids to fuel range hydrocarbons was 45%. Using these numbers and standardizing
to 1 VMT of fuel produced the results that are presented in table 3. All data sets on
material flows are presented in the appendix.
Conversion Methods
During the course of the study it was appropriate to use different units based on
the process being analyzed. All data was eventually converted to 1 vehicle mile traveled.
Included below is a chart referencing all relevant conversions made during the study. All
numbers were based off of the 25-mpg light truck average in the US [US DOT, 2011],
Table 3 A collection of calculations needed to compute environmental impacts of
biofuels.
Variable value source
Miles per gallon of petroleum 25MPG [10]
1 gallon of Highway Diesel 135.8 MJ [ORNL, 2008]
1 VMT Needs 5.432 MJ [ORNL, 2008]
1 kg soy beans 0.18 kg soy oil [USDA, 2010]
0.147 kg soy oil (165.7 ml) 1 VMT calculation
0.139 kg renewable Diesel 1 VMT [ORNL, 2008]
Denver electricity C02e emission 0.8 [Ramaswami,
factor kgC02e/kWh 2008]
Denver electricity emission factor 0.22 kg C02 e/MJ calculation
49 MJ 1 kg Natural gas NREL database
1 kWh 3.6 MJ [Cengal, 2011]
Environmental Impact Assessment
Emission factors for the unit processes of agriculture, transportation, and crushing
were gathered from the NREL LCA study conducted on biodiesel [Sheehan, 1998],
Emission factors for the hydrogen NexBTL process were gathered from two different
LCA studies on Neste Oils plant in Proovo, Finland [Nikander, 2008; Reinhardt, 2006],
24


All data for the non-hydrogen process was scaled up using the study from Finland and
laboratory data on yield and catalysts. Characterization factors came from IMPACT
2002. This study only analyzed mid point categories and did not address damage
categories.
Mid Point Impacts analyzed: global warming potential, carcinogenic effects, and aquatic
ecotoxicity. The data is reported in the figures below.
Global Warming Potential
(kgC02e/VMT)
4.00E-01
3.5GE-01
3.00E-01
2.50E-O1
2.00E-01
1.50E-01
1.00E-01
5.O0E-O2
O.OOE+OO
Figure 13 Global Warming Potential Indicator
GWP is reported in carbon dioxide equivalents. All life cycle transportation emissions
were combined into one bar. The processing stage has the HDO process in blue on the
bottom and the non-hydrogen process in red on top.
25


Carcinogenic Effects
(kgeq Chloroethylene into air/VMT)
8E-08
7E-08
6E-08
5E-08
4E-08
3E-08
2E-Q8
1E-Q8
0
! 73082E-01
agg
crushing
processing transport
Figure 14 Carcinogenic Effects Indicator
Carcinogenic effects are reported in chloroethylene equivalents. The unit processes are
listed separately and all life cycle transportation data is shown in one bar. The processing
stage has the HDO process in blue on the bottom and the non-hydrogen process in red on
top.
Aquatic Ecotoxicity
(kgeq Triethylene glycol into water/VMT)
0.005
0.004
0.003
0.002
0.001
0D29222511
0
agg crushing processing transport
Figure 15 Aquatic Eco Toxicity Indicator
Aquatic eco toxicity is shown as Triethylene glycol equivalents. All unit processes are
show separately and life cycle transportation is collected into one bar. The processing
stage has the HDO process in blue on the bottom and the non-hydrogen process in red on
top.
26


Table 4 Cradle-to-grave comparison of petroleum diesel, HDO renewable diesel
and non-hydrogen renewable diesel
Unit Process Petroleum diesel Commercial Renewable Diesel usana Hvdroaen (HDO) CU Hydrogen-Free renewable fuel orocesa
Energy Inputs MJ/MJ Fuel Emissions gO02e/MJ fuel Energy Inputs MJ/MJ Fu Emissions qC02e/MJ fuel Energy Input MJ/MJ Fuel Emissions ciC02e/MJ
Agrtcufture/Mirang/Procssa "fl Energy GHG Emissions 0.03 3.3 0.116 58.69 0.116 58.69
Transport of Raw Materials Energy GHG Emissions 0.01 0.8 0.01 0.3 0.01 0.3
Processing of Fuel Energy Inputs GHG Emissions 0.1 8.6 0.0935 9.33 0.03 2.6
Transport of Products Energy GHG Emissions 0.01 0.7 0.01 0.7 0.01 0.7
Fuel Use" Energy 73
GHG Emissions
Total 0.15 86.4 0.19 69.86 0.17 62.29
100
90
80
70
60
50
40
Petroleum
HDO Non-H2
- Fuel Use
Transport of Products
Pre-Treatment and Processing
* Transport of Raw materials
" Agricufture/Mining
Figure 16 This chart show the life cycle GHG emissions from cradle to grave of
three pathways to fuel production normalized to petroleum diesel.
The chart shows that the HDO process has a 20% decrease in life cycle GHG emissions
and the un-optimized non-H2 process having a 30% decrease when compared to
petroleum diesel.
27


Greenhouse Gas Emissions in the
Processing Stage
10
9
I 7

6
O
O
5
2 4
o
M
3
£
2
1
0

Petroleum
HDO Non-H2 (un- Non-H2 (optimized)
optimized)
Figure 17 C02e emissions from only the processing stage of three possible
pathways to diesel fuel.
Interpretation
Goal and Scope
The goal of the analysis was to determine which unit process has the greatest
impact on the three mid point categories selected: global warming potential (GWP),
carcinogenic effects, and aquatic eco toxicity. In addition to determining which unit
process had the greatest impact, a direct comparison of two approaches to renewable
diesel conversion was analyzed using energy inputs and GHG emissions.
28


Results and Discussion
This study found that the agricultural phase of the biofuels industry is the largest
producer of GHG emissions and therefore has the highest global warming potential. That
is comparible with other LCA studies in this field [Sheehan, 1998], In addition to
emitting GHG emissions, agriculture also has the largest impact on ecosystem toxicology
and cancer causing substances [figures 14 and 15], This is mainly due to the energy and
fossil fuel intense nature of industrial agriculture as well as runoff to rivers and streams.
This study shows that the agriculture industry has more work to be done before being
considered a viable alternative to the current petroleum industry. According to the LCI,
even with 40% lower yields, the non-hydrogen process had less of an impact on global
warming potential and aquatic eco toxicity then did the HDO process. The renewable
diesel plant in this study uses natural gas reforming to produce hydrogen, as is the case
with most hydrogen production in the world today [Abassi etal, 2011], Hydrogen
production uses large quantities of fossil energy and is responsible for the high GWP and
aquatic eco toxicity of the HDO process in addition to contributing to 30% of the
financial costs. The carcinogenic mid point indicator showed the HDO process being
slightly less harmful then the non-hydrogen process. This is attributed to the increased
usage of electricity, both from coal and natural gas, needed in the non-hydrogen process
to produce an equivalent yield in biofuel as the HDO process. This resulted in the
increased release of cancer causing gasses to the air and eventually into humans. The
results from comparing the GHG emissions of the HDO process, the non-hydrogen
process, and petroleum diesel are presented in figures 15 and table 4. The data for the
petroleum life cycle energy inputs and emission was collected and compiled from an
29


NREL study comparing petroleum to first generation biodiesel [Sheehan, 1998], The
data for the HDO process and the non-hydrogen process was taken from the Helsinki
study [Nikander, 2008], All of the upstream energy inputs and emissions for the HDO
process and the non-hydrogen process are exactly the same. The difference that can be
seen between the two renewable diesel processes starts in the processing and pre-
treatment stage highlighted in yellow in table 4. Here it can be seen that hydrogen has
been removed from the process resulting in a decrease in energy inputs as well as GHG
emissions. This decrease is seen even when a 40% drop in biofuel yield from the non-
hydrogen process. It can also be seen from figure 16 that the current commercially
available HDO renewable diesel processes has higher GHG emissions then even
petroleum diesel. This means that the fuel would not qualify as Biomass Based Diesel,
which calls for a 50% reduction in life cycle GHG emissions [EPA 2012],
30


Chapter III
CONCLUSIONS
Significant Findings
The journey towards developing a viable alternative to cheap energy dense fossil
fuels is an incremental one. Many research groups and industries are working
simultaneously at a multitude of routes for achieving that end. This study has also added
to that body of knowledge. Through laboratory experiments, literature review, and
environmental assessment, I have verified that a viable alternative drop-in replacement
fuel can be produced with out the use of hydrogen gas, which in turn could reduce cost
and environmental impacts of biofuels.
I have shown that a significant quantity of saturated hydrocarbons can be
produced from fatty acids found in soybean oil using the decarboxylation route. This
means that the potential for producing biofuels with increasingly lower green house gas
emissions and reduced fossil energy inputs is possible. In addition, this study has also
directly compared the energy inputs and GHG emissions of petroleum diesel, commercial
renewable diesel, and renewable diesel without the use of hydrogen gas. These
preliminary findings show that even with an un-optimized reaction, a non-hydrogen
process could have half of the emissions of current commercial renewable diesel in the
processing stage. As we continue to optimize the reaction and increase yields, energy
demands and GHG emission will look even more promising. This shows that it is
possible to eliminate the need for large amounts of environmentally detrimental hydrogen
gas in commercial scale renewable diesel production.
31


Next Steps
The results presented in this thesis are exciting and hopeful. But this is
just the beginning. The most promising aspect is the capacity building that took place
over the course of the research. The Green Engineering and Systems Analysis Lab at CU
Denver now has the foundation to build on this research and continue producing state of
the art research in advanced biofuels. At the beginning of the research there was a
reactor sitting in a corner of a dusty unused lab space. Now the lab is fully equipped to
use both batch and semi-batch processes to produce biofuels as well as to analyze and
report on the products and results of the reactions. Yield optimization of the reaction is
underway, looking at temperature and pressure, as well as catalyst
activation/deactivation. At 400 C cracking and isomerization is prevalent, so decreasing
the reaction temperature is highly favorable for both product selectivity and energy
reduction [Huber, 2007], More analysis and research needs to be done to further
characterize the complete liquid product to better understand the reaction mechanics. In
addition, research on a catalytic process to hydrogenate unsaturated feedstock in the
reaction vessel in underway. The use of a proton donating solvent has been reported to
be effective [KubicAova etal, 2008], In addition I plan to employ a food industry process
called catalytic transfer hydrogenation as a pre-treatment step. Differing from the
classical hydrogenation techniques using molecular hydrogen, hydrogen donors are used
as a source of hydrogen in a catalytic transfer reduction to saturate the feedstock. The
generalized equation 1 represents this process:
DH+ A catalyst D+ AH
Here A represents the acceptor and D represents the hydrogen donor [Naglic, 1998
32


pg 1; Dumesic, 2011], This process could greatly decrease the cost and environmental
issues with using molecular hydrogen while increasing the yields of fuel range
hydrocarbons from the decarboxylation reaction presented in this paper.
Based on these findings we plan to work on catalytic transfer hydrogenation to
saturate feedstocks before the decarboxylation reaction. Little work has been done in this
realm and it could produce significant increases in yields. In addition we plan to look
into wet algae processing for biofuels production. This paper showed that the most
significant contributor to life cycle environmental impacts is the agricultural phase. In
order to reach the 50% reduction in life cycle GHG emission mandated by the RFS2 fuel
standards, much work must be done in the feedstock production realm. Algae are being
considered as a possible solution to environmental problems of in-ground agriculture. I
have started cultivating strains of algae in the lab and am looking into oil extraction
methods as well as wet processing. Now that we have biofuels research capacity at UCD
there is limitless directions that we can take the research. Cheers.
33


REFERENCES
Abbasi, T., & Abbasi, S. a. (2011). Renewable hydrogen: Prospects and challenges.
Renewable and Sustainable Energy Reviews, 75(6), 3034-3040.
doi: 10.1016/j.rser.2011.02.026
Barnes, B., Mathee, A., & Moiloa, K. (2005). Assessing child time-activity patterns in
relation to indoor cooking fires in developing countries: a methodological
comparison. International journal of hygiene and environmental health, 208(3),
219-25. doi:10.1016/j.ijheh.2005.01.022
Cengel, Yunus A. Boles, Michael A. (2011) Thermodynamics: An Engineering
Approach. New York, NY, McGraw-Hill
Chia, M., & Dumesic, J. a. (2011). Liquid-phase catalytic transfer hydrogenation and
cyclization of levulinic acid and its esters to y-valerolactone over metal oxide
catalysts. Chemical communications (Cambridge, England), 47(44), 12233-5.
doi:10.1039/clccl4748j
EPA 1320 SUMMARY : Regulation of Fuels and Fule Additives: 2012 Renewable Fuel
Standards (2012).
Huber, G. W., Iborra, S., & Corma, A. (2006). Synthesis of transportation fuels from
biomass: chemistry, catalysts, and engineering. Chemical reviews, 106(9), 4044-98.
doi:10.1021/cr068360d
Huber, G. W., OConnor, P., & Corma, A. (2007a). Processing biomass in conventional
oil refineries: Production of high quality diesel by hydrotreating vegetable oils in
heavy vacuum oil mixtures. Applied Catalysis A: General, 329, 120-129.
doi:10.1016/j.apcata.2007.07.002
Huber, G. W., OConnor, P., & Corma, A. (2007b). Processing biomass in conventional
oil refineries: Production of high quality diesel by hydrotreating vegetable oils in
heavy vacuum oil mixtures. Applied Catalysis A: General, 329, 120-129.
doi:10.1016/j.apcata.2007.07.002
Kalnes, T. N., Koers, K. P., Marker, T., & Shonnard, D. R. (2009). A Technoeconomic
and Environmental Life Cycle Comparison of Green Diesel to Biodiesel and
Syndiesel, 25(1), 111-120. doi:10.1002/ep
Kinsley, G. R. J. (2001). Properly Purge and Imert Storage Vessels. Chemical
Engineering Progress, 97(2), 57-61.
Knothe, G. (2010a). Biodiesel and renewable diesel: A comparison. Progress in Energy
and Combustion Science, 36(3), 364-373. doi: 10.1016/j.pecs.2009.11.004
34


Knothe, G. (2010b). Biodiesel and renewable diesel: A comparison. Progress in Energy
and Combustion Science, 36(3), 364-373. doi:10.1016/j.pecs.2009.11.004
Koloini, T., & Andrej, S. (1998). Kinetics of Catalytic Transfer Hydrogenation. Journal
of American Oil Chemists Society, 75(5), 629-633.
Kubickova, I., Snare, M., Eranen, K., Maki-Arvela, P., & Murzin, D. Y. (2005).
Hydrocarbons for diesel fuel via decarboxylation of vegetable oils. Catalysis Today,
706(1-4), 197-200. doi:10.1016/j.cattod.2005.07.188
Kubickova, I., Snare, M., Era, K., Murzin, D. Y., & Turku, F.-. (2007). Catalytic
Deoxygenation of Fatty Acids and Their Derivatives. Energy and Fuel, 27(16), 30-
41.
Mikulec, J., Cvengros, J., Jorikova, E., Banic, M., & Kleinova, A. (2010). Second
generation diesel fuel from renewable sources. Journal of Cleaner Production,
18(9), 917-926. doi:10.1016/j.jclepro.2010.01.018
Na, J.-G., Han, J. K., Oh, Y.-K., Park, J.-H., Jung, T. S., Han, S. S., Yoon, H. C., et al.
(2012). Decarboxylation of microalgal oil without hydrogen into hydrocarbon for
the production of transportation fuel. Catalysis Today, 185(1), 313-317.
doi: 10.1016/j. cattod.2011.08.009
Na, J.-G., Yi, B. E., Kim, J. N., Yi, K. B., Park, S.-Y., Park, J.-H., Kim, J.-N., et al.
(2010). Hydrocarbon production from decarboxylation of fatty acid without
hydrogen. Catalysis Today, 756(1-2), 44-48. doi:10.1016/j.cattod.2009.11.008
Nikander, S. (2008). GREENHOUSE GAS AND ENERGY INTENSITY OF PRODUCT
CHAIN: CASE TRANSPORT BIOFUEL. Helsinki University of Technology.
Oak Ridge National Laboratory (2008), Conversion Factors for Bioenergy, retrieved from
www.ces.ncsu.edu/forestry/biomass.html
Paula, Matthew de (2011). Veggie Oil Smells Better Than Diesel, But It's No Slam Dunk.
Forbes,. Retrieved from
http://www.forbes.com/sites/matthewdepaula/2011/05/27/veggie-oil-might-smell-
better-than-diesel-but-its-no-slam-dunk/
Pearlson, M. N. (2011). A Techno-Economic and Environmental Assessment of
HydroprocessedRenewable Distillate Fuels by. MIT.
Ramaswami, A., Hillman, T., Janson, B., Reiner, M., & Thomas, G. (2008). Policy
Analysis A Demand-Centered Hybrid Life-Cycle Methodology for City-Scale
Greenhouse Gas Inventories. American Chemical Society, 42(11), 6455-646 T
35


Reinhardt, G. (2006). Final Report An Assessment of Energy and Greenhouse Gases of
NExBTL (pp. 1-30). Porvoo, Findland.
REN21. (2011). Renewables 2011 Global Satus Report (pp. 1-116). Retrieved from
www.ren21.net
Roh, H.-S., Eum, I.-H., Jeong, D.-W., Yi, B. E., Na, J.-G., & Ko, C. H. (2011). The effect
of calcination temperature on the performance of Ni/Mg0-A1203 catalysts for
decarboxylation of oleic acid. Catalysis Today, 164(1), 457-460.
doi: 10.1016/j.cattod.2010.10.048
Rozmyszowicz, B., Lestari, S., Simakova, O., Er, K., Salmi, T., & Murzin, D. Y. (2011).
Catalytic Deoxygenation of Tall Oil Fatty Acid over Palladium Supported on
Mesoporous Carbon. Energy and Fuels, (25), 2815-2825.
Schlagermann, P., Gottlicher, G., Dillschneider, R., Rosello-Sastre, R., & Posten, C.
(2012). Composition of Algal Oil and Its Potential as Biofuel. Journal of
Combustion, 2012, 1-14. doi: 10.1155/2012/285185
Senate Low Aromatic Fuel Act 2013 (2013). Retrieved from www.comlaw.gov.au/
§enol, O., Ryymin, E.-M., Viljava, T.-R., & Krause, a. O. I. (2007). Effect of hydrogen
sulphide on the hydrodeoxygenation of aromatic and aliphatic oxygenates on
sulphided catalysts. Journal of Molecular Catalysis A: Chemical, 277(1-2), 107-
112. doi:10.1016/j.molcata.2007.07.033
Serrano-Ruiz, J. C., Pineda, A., Balu, A. M., Luque, R., Campelo, J. M., Romero, A. A.,
& Ramos-Fernandez, J. M. (2012). Catalytic transformations of biomass-derived
acids into advanced biofuels. Catalysis Today, 195(1), 162-168.
doi:10.1016/j.cattod.2012.0T009
Sheehan, J. et al. (1998). Life Cycle Inventory of Biodiesel and Petroleum Diesel for Use
in an Urban Bus (pp. 1-314).
Simacek, P., Kubicka, D., Sebor, G., & Pospisil, M. (2009). Hydroprocessed rapeseed oil
as a source of hydrocarbon-based biodiesel. Fuel, 88(2), 456-460.
doi: 10.1016/j.fuel.2008.10.022
Sinclair, C., & Manager, E. (2007). Cremation of Human Remains: A Comparison of
Alkaline Hydrolysis versus Combustion (pp. 1-6). Retrieved from
www.resomation.com
Snare, M., Kubickova, I., Maki-Arvela, P., Chichova, D., Eranen, K., & Murzin, D. Y.
(2008). Catalytic deoxygenation of unsaturated renewable feedstocks for production
of diesel fuel hydrocarbons. Fuel, 87(6), 933-945. doi: 10.1016/j.fuel.2007.06.006
36


Snare, M., Kubickova, I., Maki-Arvela, P., Eranen, K., Warna, J., & Murzin, D. Y.
(2007). Production of diesel fuel from renewable feeds: Kinetics of ethyl stearate
decarboxylation. Chemical Engineering Journal, 134(1-3), 29-34.
doi:10.1016/j.cej .2007.03.064
Solecki, M., & Epstein, B. (2011). Advanced Biofuel Market Report 2011 Meeting the
California LCFS Authors Table of Contents (pp. 1-22).
Tat, M. E., Wang, P. S., Gerpen, J. EL, & Clemente, T. E. (2007). Exhaust Emissions
from an Engine Fueled with Biodiesel from High-Oleic Soybeans. Journal of the
American Oil Chemists Society, 84(9), 865-869. doi: 10.1007/sl 1746-007-1109-6
U.S. Department of Agriculture (2010), Soybeans and Oil Crops: Background. Retrieved
from: http://www.ers.usda.gov/Briefing/SovbeansOilcrops/background.htm
U.S. Department of Transportation, Federal Highway Administration, (2011) Highway
Statistics (Washington, DC: Annual Issues), table VM-1, available at
http://www.fhwa.dot.gov/policvinformation/statistics.cfm
"U.S. Life Cycle Inventory Database." (2012). National Renewable Energy Laboratory,
2012. Accessed November 19, 2012: https://www.lcacommons.gov/nrel/search
Veriansyah, B., Han, J. Y., Kim, S. K., Hong, S.-A., Kim, Y. J., Lim, J. S., Shu, Y.-W., et
al. (2012). Production of renewable diesel by hydroprocessing of soybean oil: Effect
of catalysts. Fuel, 94, 578-585. doi: 10.1016/j.fuel.2011.10.057
Wang, W.-C., Roberts, W. L., & Stikeleather, L. F. (2012). Hydrocarbon Fuels From Gas
Phase Decarboxylation of Hydrolyzed Free Fatty Acid. Journal of Energy Resources
Technology, 134(3), 032203. doi:10.1115/1.4006867
Wang, W.-C., Thapaliya, N., Campos, A., Stikeleather, L. F., & Roberts, W. L. (2012).
Hydrocarbon fuels from vegetable oils via hydrolysis and thermo-catalytic
decarboxylation. Fuel, 95, 622-629. doi: 10.1016/j.fuel.2011.12.041
37


APPENDIX
Soybean Agriculture Unit Process Data
This includes fuels, fertilizers, and inputs used on the farm and their associated emissions, as well
as the upstream emissions associated with those inputs used to grow soybeans. These upstream
environmental flows are combined with the flows associated with the actual soybean growing and
harvesting to calculate the total emissions associated with soybean agriculture [Sheehan, 1998],
All data has been normalized to 1 VMT.
Table 5 Soybean agriculture raw materials data
Raw Materials for Soybean Agriculture Units kg of raw material/1 VMT
Coal (in ground) kg 0.0474376
Oil (in ground) kg 0.2439248
Natural Gas (in ground) kg 0.0928704
Uranium (U, ore) kg 0.000002604
Phosphate Rock (in ground) kg 0.335776
Potash (K20, in ground) kg 0.1578192
Perlite (Si02, ore) kg 0.000047544
Limestone (CaC02, in ground) kg 0.0061376
Sodium Chloride (NaCI) kg 0
Water Use kg 3081.96
Elec from Coal kg 2.44048
Elec from Natural gas kg 2.44048
Table 6 Soybean agriculture emissions data
Emissions from Soybean
Agriculture Units Total emissions /1 VMT
AIR
Carbon Dioxcide (C02, fossil) kg 1.024072
Carbon Dioxcide (C02, biomass) kg 0
Methane (CH4) kg 0.001012592
Nitrous Oxcide (N20) kg 0.00004676
Carbon Monoxcide (CO) kg 0.004890312
Hydrocarbons kg 0.006069504
Benzene kg 4.6928E-08
Formaldehyde kg 6.328E-07
Particulates, unspecified kg 0.001043084
Sulfur oxides kg 0.003357558
Nitrogen oxides kg 0.007168
Hydrogen chloride kg 0.00000952
Hydrogen fluoride kg 4.3848E-07
Ammonia (air) kg 0.0026208
38


WATER Agrochemicals kg 0.00037352
BOD5 kg 0.000189896
COD kg 0.00134624
Metals kg 0.000005824
Ammonia kg 0.000022512
Nitrates kg 9.576E-08
Soybean Transport Unit Process Data
The transport of soybeans from the field to the crusher is based on current practices in the
soybean farming and soybean crushing sectors. The actual data used in this study were based on a
combination of conversations with industry representatives and modeled data. They ignored
transporting soybeans for the export market, which represents the longest transportation distances
in the industry [Sheehan, 1998].
Table 7 Soybean transport raw materials data
Raw Materials for Soybean Transport Units Kg of material /1 VMT
Coal (in ground) kg 0.00059472
Oil (in ground) kg 0.01903664
Natural Gas (in ground) kg 0.00163464
Uranium (U, ore) kg 1.316E-08
Phosphate Rock (in ground) kg 0
Potash (K20, in ground) kg 0
Perlite (Si02, ore) kg 0.000004088
Limestone (CaC02, in ground) kg 0.00011312
Sodium Chloride (NaCI) kg 0
Water Use kg 0.00264936
Elec from Coal kg 0.126017889
Elec from Natural gas kg 0.126017886
Table 8 Soybean transport emissions data
Emissions from Soybean
Transport Units Total emissions /1 VMT
AIR
39


Carbon Dioxcide (C02, fossil) kg 0.0634228
Carbon Dioxcide (C02, biomass) kg 0
Methane (CH4) kg 2.27276E-05
Nitrous Oxcide (N20) kg 6.30168E-06
Carbon Monoxcide (CO) kg 0.00021512
Hydrocarbons kg 0.0000714
Benzene kg 4.2728E-09
Formaldehyde kg 5.712E-08
Particulates, unspecified kg 8.54661 E-05
Sulfur oxides kg 0.000093408
Nitrogen oxides kg 0.00059673
Hydrogen chloride kg 3.192E-07
Hydrogen fluoride kg 3.9928E-08
Ammonia (air) kg 8.12E-11
WATER
Agrochemicals kg 0
BOD5 kg 0.000012768
COD kg 0.000108024
Metals kg 5.2808E-07
Ammonia kg 1.86704E-06
Nitrates kg 5.0568E-10
Soybean Crushing Unit Process Data
The crushing process in this study is assumed to be representative of typical soybean crushing
operations in the United States. The major distinguishing feature among the types of processes
used in the U.S. crushing plants is in the oil extraction method. Our model facility uses solvent
extraction to recover the oil. The other major method for extraction is mechanical crushing. Only
1% to 2% of the soybeans processed in the United States are recovered via mechanical extraction
methods [Sheehan, 1998]
Table 9 Soybean crushing raw materials data
Raw Materials for Soybean
Crushing
Coal (in ground) kg
Oil (in ground) kg
Natural Gas (in ground) kg
Uranium (U, ore) kg
Phosphate Rock (in ground) kg
Potash (K20, in ground) kg
Perlite (Si02, ore) kg
Limestone (CaC02, in ground) kg
Units Raw materials /1 VMT
0.019754224
0.003182424
0.053525584
4.7348E-07
0
0
0
0.00376656
40


Sodium Chloride (NaCI) kg 0
Water Use kg 0.004359208
Elec from Coal kg 0.508763304
Elec from Natural gas kg 0.508763304
Table 10 Soybean crushing emissions data
Emissions from Soybean Crushing Units Total emissions from Agg/1 VMT
AIR
Carbon Dioxcide (C02, fossil) kg 0.21234768
Carbon Dioxcide (C02, biomass) kg 0
Methane (CH4) kg 0.000395656
Nitrous Oxcide (N20) kg 1.92284E-06
Carbon Monoxcide (CO) kg 6.43876E-05
Hydrocarbons kg 0.001979569
Benzene kg 0
Formaldehyde kg 3.18E-15
Particulates, unspecified kg 0.000281171
Sulfur oxides kg 0.001516627
Nitrogen oxides kg 0.000398267
Hydrogen chloride kg 1.06201E-05
Hydrogen fluoride kg 1.32712E-06
Ammonia (air) kg 4.8972E-08
WATER
Agrochemicals kg 0
BOD5 kg 2.6818E-07
COD kg 2.19314E-06
Metals kg 6.2858E-08
Ammonia kg 3.53404E-07
Nitrates 1.6854E-08
Soybean Oil Transport Unit Process Data
The NREL LCA practitioners studied average transport distances for soybean oil in
depth. This data includes average fuel emissions from wells to pump and pump to wheels
[Sheehan, 1998],
Table 11 Soybean transport data raw materials
Raw Materials for Soybean Oil
Transport Units
Kg of raw material /1
VMT
41


Coal (in ground) kg 0.00021702
Oil (in ground) kg 0.006945164
Natural Gas (in ground) kg 0.000596416
Uranium (U, ore) kg 5.1072E-09
Phosphate Rock (in ground) kg 0
Potash (K20, in ground) kg 0
Perlite (Si02, ore) kg 1.57472E-06
Limestone (CaC02, in ground) kg 4.11768E-05
Sodium Chloride (NaCI) kg 0
Water Use kg 0.000966573
Elec from Coal kg 0.0459648
Elec from Natural gas kg 0.0459648
Table 12 Soybean transport emissions data
Emissions from Soybean Oil Transport Units Total emissions from Oil Transport /1 VMT
AIR
Carbon Dioxcide (C02, fossil) kg 0.023296044
Carbon Dioxcide (C02, biomass) kg 0
Methane (CH4) kg 8.37718E-06
Nitrous Oxcide (N20) kg 2.4839E-07
Carbon Monoxcide (CO) kg 7.77499E-05
Hydrocarbons kg 2.80794E-05
Benzene kg 1.5476E-09
Formaldehyde kg 2.0776E-08
Particulates, unspecified kg 1.42888E-05
Sulfur oxides kg 3.39094E-05
Nitrogen oxides kg 0.00038425
Ammonia (air) kg 2.9574E-11
WATER
Agrochemicals kg 0
BOD5 kg 4.64074E-06
COD kg 3.92678E-05
Metals kg 1.9193E-07
Ammonia kg 1.8338E-10
Nitrates kg 1.5052E-08
Soybean Oil Conversion to Renewable Diesel using Hydrogen Process
This data set includes the energy use from the HDO reaction at the NESTE Oil plant in
Finland plus all inputs from production of hydrogen gas [Nikander, 2008],
42


Table 13 HDO raw materials data assuming an 84% yield
Raw Materials for HDO conversion
Treated Soybean Oil
Natural Gas (for hydrogen)
Fuel Gas (for hydrogen)
Process Water
Cooling water Use
Steam
Elec from Coal
Elec from Natural gas
Kg of raw material /1
VMT
kg 0.14699
kg 0.00588
kg 0.00196
kg 0.02861
kg 0.0004887
MJ 0.003801
kWh 0.00332
kWh 0.00332
Table 14 HDO emissions data assuming an 84% yield
Emissions Units Total emissions from Agg/1 VMT
AIR C02e (from electricity) kg 0.0011
C02e (from Steam) kg 0.00396
C02e (from hydrogen) kg 0.04561
Methane (CH4) kg 2.37478E-05
Nitrous Oxcide (N20) kg 0
Carbon Monoxcide (CO) kg 3.31169E-06
Hydrocarbons kg 0
Benzene kg 1.87249E-09
Formaldehyde kg 2.33507E-09
Particulates, unspecified kg 2.95796E-07
Sulfur oxides kg 0.002156816
Nitrogen oxides kg 3.11198E-06
Ammonia (air) kg 0 0 0 1.7376E-08
WATER Agrochemicals kg
BOD5 kg 0
COD kg 2.37478E-05
Metals kg 0
Ammonia kg 3.31169E-06
Nitrates kg 0
43


Soybean Oil Conversion to Renewable Diesel using Non-Hydrogen Process
This data set looks at all inputs and relevant outputs from the lab scale yield data overlaid
on the process data from the Neste Oil plant [Nikander, 2008],
Table 15 Non-hydrogen raw materials data assuming a 45% yield
Raw Materials for Non-H2 conversion
Treated Soybean Oil
Palladium on Carbon
Process Water
Cooling water Use
Steam
Elec from Coal
Elec from Natural gas
Kg of raw material /1
kg 0.1940268
kg 0.0096888
kg 0.0377652
kg 0.000645084
MJ 0.00501732
kWh 0.0043824
kWh 0.0043824
Table 16 Non-hydrogen emissions data assuming a 45% yield
Emissions
Units
Total emissions from Agg/1
VMT
AIR
C02e (from electricity) kg 0.00151
C02e (from Steam) kg 0.00554
C02e (from Palladium) kg 0.0000015
Methane (CH4) kg 2.99515E-05
Nitrous Oxcide (N20) kg 0
Carbon Monoxcide (CO) kg 4.20173E-06
Hydrocarbons kg 0
Benzene kg 2.35595E-09
Formaldehyde kg 2.93798E-09
Particulates, unspecified kg 2.60605E-05
Sulfur oxides kg 0.002713707
Nitrogen oxides kg 4.35272E-06
44


Ammonia (air) kg 0
WATER
Agrochemicals kg 0
BOD5 kg 0
COD kg 0
Metals kg 0
Ammonia kg 2.18624E-08
Nitrates kg 7.10528E-12
45


Full Text
Assumptions and Limitations
Many assumptions were made while compiling this LCA study. These
assumptions will be looked at in two major categories: experimental lab data and
feedstock selection.
Experimental Lab Data
The single biggest limitation was moving from the laboratory data collected on
the non-hydrogen process to a viable comparison with a commercial scale process. I
assumed that all inputs for the non-hydrogen process at commercial scale would be the
same as reported for the NexBTL process minus the energy and emissions associated
with hydrogen gas. Further modeling would be needed for a more accurate
representation. Also all experimental data was gathered using fatty acids as model
compounds in lieu of using soybean oil. These compounds are similar to soybean oil in
composition, but they do not contain glycerin backbones. It was assumed that a
hydrolysis step would be added to the industrial reaction scheme to first produce fatty
acids from the soybean oil. Fat splitting, or hydrolysis, of triglycerides produces fatty
acids and glycerol as seen in reaction below.
RCOOCH2-CHOOCR-CH2OCOR + 3 H20 -> 3 RCOOH + HOCH2-CHOH-CH2OH
My data assumes high yield of fatty acids from soybean oil and therefore might
limit the accuracy of the study, but provides a good comparison. Since the conditions for
hydrolysis is very similar to decarboxylation (350 C in an aqueous environment with a
catalyst) I have considered the GHG emissions associated with hydrolysis to be
equivalent to the emissions from the reactor itself [Sinclair etal, 2007], Another
21



PAGE 1

THE EFFECTS OF MOLECULAR SATURATION OF FEEDSTOCKS ON RENEWABLE DIESEL PRODUCTION IN THE ABSENCE OF HYDROGEN GAS By Jonathan Dubinsky B.S., University of Kansas, Environmental Science A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Masters of Engineering Civil Engineering 2013

PAGE 2

ii This Thesis for the Master of Engineering degree by Jonathan Dubinsky h as been approved for the Civil Engineering P rogram by Arunprakash Karunanithi, Chair Ron Roher Jason Ren April 16, 2013

PAGE 3

iii Dubinsky, Jonathan (MEng, Civil Engineering ) The Effects of Molecular Saturation of Feedstocks on Renewable Diesel Production in the Absence of Hydrogen Gas Thesis directed by Assistant Professor Arunprakash Karunanithi ABSTRACT Catalytic deoxygenation of fatty acid s was carried out using Palladium on Carbon (Pd/C) and Magnesium Oxide (MG70) ca talyst s and a Parr bench top micro reactor. The effect of feedstock saturation on the decarboxylation reaction was measured by varying the feedstock from stearic acid (fully saturated) to oleic acid (mono unsaturated) to linoleic a cid (poly unsaturated) The prevalence of the decarboxylation reac tion was measure d by analyzing both the liquid products and the gaseous products The liquid products were collected and analyzed for the major product, heptadecane (C 17 ) as well as straight chain alkane in the diesel range content using a gas chomatograph with a flame ionization detector. Concentration of CO 2 in the gaseous product was analyzed using gas chromatography. As satura tion increased the concentration of CO 2 in the gas phase increased showing that decarboxylation occurred The ana lysis of the liquid products showed as feedstock saturation increased the concentration of saturated straight chain hydrocarbons in the diesel fuel range with high selectivity to C 17 heptadecane increased confirming that the decarboxylation reaction favors more saturated feedstocks The form of and content of this abstract are approved. I recommend its publication. Approved: Arunprakash Karunanithi

PAGE 4

iv DEDICATION I dedicate this work to my loving and supportive sweetheart Elizabeth She has encouraged me to always pursue m y dreams with confidence and grace.

PAGE 5

v ACKNOWLEDMENTS It is with great honor that I am able to contribute this paper to the scientific community. As an owner and operator of a vegetable oil power ed educational bus, I have seen first hand the power and limitations of alternative fuel sources. I would like to thank Dr. Arun Karunanithi for going out on a limb and taking on a greasy mechanic like myself. I would like to thank my parents for being the change that they want to see in the world, and always telling me that I can do what ever I put my mind to. Thank you to all the merry pranksters and joy seekers in the world you know who you are. I plan to no t only work to advance the field of bio fuels technology, but to educate the next generation with the concepts of sustainability and biophilia. Long live spaceship Earth

PAGE 6

vi TABLE OF CONTENTS CHAPTER I EXPERIMENTAL ................................ ................................ ................................ ......... 1 Introduction ................................ ................................ ................................ ................. 1 Background and Problem ................................ ................................ ........................ 1 Biomass Based Diesel/Renewable Diesel ................................ ............................... 3 Purpose of the Project ................................ ................................ ............................. 5 Experimental ................................ ................................ ................................ ............... 7 Methods ................................ ................................ ................................ ................... 7 Chemicals and Catalysts ................................ ................................ ..................... 7 Decarboxylation Reaction ................................ ................................ ................... 8 Analysis of Products ................................ ................................ ........................... 9 Results ................................ ................................ ................................ ....................... 10 Effects of Feedstock Saturation on Liquid Products ................................ ............. 10 Effects of Feedstock Saturation on Gaseous Products ................................ .......... 15 Discussion ................................ ................................ ................................ ................. 16 II ENVIRONMENTAL ASSESMENT ................................ ................................ .......... 18 Introduction ................................ ................................ ................................ ............... 18 Goal and Scope ................................ ................................ ................................ ......... 19 System Boundaries ................................ ................................ ................................ 19 Process flow diagram ................................ ................................ ............................ 20 Experimental Lab data ................................ ................................ ...................... 21 Feedstock ................................ ................................ ................................ .......... 22

PAGE 7

vii Life Cycle Inventory ................................ ................................ ................................ 22 Allocations ................................ ................................ ................................ ............ 22 Method of Data Collection ................................ ................................ .................... 22 Decarboxylation Reaction Methods ................................ ................................ ...... 23 Conversion Methods ................................ ................................ ............................. 24 Environmental Impact Assessment ................................ ................................ ........... 24 Interpretation ................................ ................................ ................................ ............. 28 Goal and Scope ................................ ................................ ................................ ..... 28 Results and Discussion ................................ ................................ ............................. 29 III CONCLUSIONS ................................ ................................ ................................ ......... 31 Significant Findings ................................ ................................ ................................ .. 31 Next Steps ................................ ................................ ................................ ................. 32 REFERENCES ................................ ................................ ................................ ................. 34 APPENDIX ................................ ................................ ................................ ....................... 38 Soybean Agriculture Unit Process Data ................................ ................................ ... 38 Soybean Transport Unit Process Data ................................ ................................ ...... 39 Soybean Crushing Unit Process Data ................................ ................................ ....... 40 Soybean Oil Transport Unit Process Data ................................ ................................ 41 Soybean Oil Conversion to Renewable Diesel using Hydrogen Process ................. 42 Soybean Oil Conversion to Renewable Diesel using Non Hydrogen Process ......... 44

PAGE 8

viii LIST OF TABLES Table 1 This table show levels of saturation in some common feedstock oils ............................ 7 2 Concentration in mg/l of straight chain alkanes from C 7 to C 17 in the liquid products. 12 3 A collection of calculations needed to compute environmental impacts of biofuels. .. 24 4 Cradle to grave comparison of petroleum diesel, HDO renewable diesel and non hydrogen renewable diesel ................................ ................................ ........................ 27 5 Soybean agriculture raw materials data ................................ ................................ ........ 38 6 Soybean agriculture emissions data ................................ ................................ .............. 38 7 Soybean transport raw materials data ................................ ................................ ........... 3 9 8 Soybean transport emissions data ................................ ................................ ................. 39 9 Soybean crushing raw materials data ................................ ................................ ............ 40 10 Soybean crushing emissions data ................................ ................................ ................ 41 11 Soybean transport data raw materials ................................ ................................ ......... 41 12 Soybean transport emissions data ................................ ................................ ............... 42 13 HDO raw materials data assuming an 84% yield ................................ ....................... 43 14 HDO emissions data assuming an 84% yield ................................ ............................. 43 15 Non hydrogen raw mate rials data assuming a 45% yield ................................ ........... 44 16 Non hydrogen emissions data assuming a 45% yield ................................ ................. 44

PAGE 9

ix LIST OF FIGURES Figure 1 Chemical formulas for three common biofuels: Biodiesel, Ethanol, and Renewable Diesel ................................ ................................ ................................ .......................... 3 2 This diagram, which was extracted from Bambang etal, 2011, shows the stoichiometry of the renewable diesel reaction ................................ ................................ .................. 4 3 Summary of total expenses in a current HDO renewable diesel processing plant. ........ 5 4 Stoichiometric equation for the decarboxylation reaction ................................ .............. 6 5 This diagram, extracted and modi fied by from Kubickova, 2007, is of the Parr 4590 bench top reactor setup used in the experiment ................................ .......................... 9 6 These figures show the yield of both total fuel range alkanes and the major product heptadecane. ................................ ................................ ................................ .............. 12 7 GC chromatograms of alkane reaction products at 673 K ................................ ............ 13 8 Concentration of liquid Alkanes in the liquid product ................................ .................. 14 9 This chart shows the selectivity to C 17 heptadecane in the total alkane product ........... 14 10 Real time data of produced gas from the reaction. ................................ ..................... 15 11 This chart show the concentration of CO 2 in the g aseous products from GC ............ 16 12 Process Flow Diagram ................................ ................................ ................................ 20 13 Global Warming Potential Indicator ................................ ................................ ........... 25 14 Carcinogenic Effects Indicator ................................ ................................ ................... 26 15 Aquatic Eco Toxicity Indicator ................................ ................................ ................... 26 16 This chart show the life cycle GHG emissions from cradle to grave of three pathways to fuel production normalized to petroleum diesel. ................................ .................. 27

PAGE 10

x 17 CO2e emissions from only the processing stage of three possible p athways to diesel fuel. ................................ ................................ ................................ ........................... 28

PAGE 11

1 CHAPTER I E XPERIMENTAL I ntroduction Background and P roblem "If you are in a shipwreck and all the boats are gone, a piano top . that comes along makes a fortuitous life preserver. But this is not to say that the best way to design a life preserver is in the form of a piano top. I think that we are clinging to a great many piano tops in accepting yesterday's fortuitous contrivings." Buckminster Fuller [Critical Path, 1981] Society is at a transition point and now we have an opportunity to re design our energy future. With the uncertainty of the world 's oil supply, economic crisis, and global climate change all on the horizon a suitable fuel replacement must be identified. Wind, solar, fuel cells, geo thermal and other renewable energy sources wil l play a part in a diverse energy portfolio but liquid fuels which w ork with current infrastructure must be utilized during the transition Bio fuels are a potential fit for a liquid fuel replacement, but rigorous study needs to be done to determine the true costs. An exponential increase in the consumption of such biofuels throughout the world has taken place in the past few years [ Serrano Ruiz etal, 2012]. Bio fuels have existed for thousands of years. The first fuel source for humans was wood, whic h we utilized for heat and to cook our food. Wood or cellulose, contains carbon, which is sequestered from the air using photosynthesis. As technology and human ingenuity grew we figured out how to extract fats and oils from plants and

PAGE 12

2 animals. This b ecame our primary light source, which was a great advancement for humanity because it meant we did not have to use fire in the summer time just for a little light. Health wise candles were a great improvement over the soot and smoke from a fire [Barnes etal 2005]. Human kind however, did not abandon the use of fire once they discovered candles. There was a need and a use for both. It was not until thousands of years later, with the discovery of fossil fuels that civilization resorted to a one size fits all fuel source. Rudolf Diesel observed and disapproved of this back in the 1890's. His engine did not use petroleum but locally grown peanut oil. "The use of vegetable oils for engine fuels may seem insignificant today," Diesel said in 1912, "but such oils may become in the course of time as important as the petroleum and coal tar p roducts of the present time" [Paula etal 2011 ]. In the late 1800s other engines were released on the market that ran on gasoline alone. However, the diesel engine could operate on many different fuels sources including coal dust, animal fats, vegetable oil, and petroleum. The problem that we face today is a growing demand for fossil fuels with a shrinking supply. Our economy is dependent on growth, which results in increased energy consumption. Since the early 20 th century we have been designing our systems to be reliant on one source of fuel, oil, unlike Rudoulf Diesel 's engine. The t ransporta tion sector of our society heavily relies on petroleum, which accounts for essentially all (96%) of the transportation energy [ Serrano Ruiz etal 2012]. The natural world evolved to have biodiversity and resilien ce, t he measure of which is an ecosystems a bility to bounce back after a catastrophic event like a hurricane or a volcano. This ability to bounce back is now our great struggle as a human nation. W e are now faced with the challenge of coming up with replacements to a

PAGE 13

3 cheap hydrocarbon based syste m or face the slow sinking of our great ship and the frantic clinging to piano tops as we struggle to survive Biomass Based Diesel /Renewable Diesel The commercial market for biodiesel is approximately 2.5 billion gallons per year and will grow rapidly in the near future due to the Renewable Fuels Standards (RFS 2), which establishes renewable fuel consumption and production mandate s in the United Stat es [EPA, 2012] The global market for biofuels is expected to reach $280 billion by the year 2022 [Solecki etal 2011; REN21, 2011] Renewable diesel or Biomass Based Diesel, refers to fuel produced from biomass that has as least a 50% reduction in life cycle CO2 emissions as compared to petroleum and can be used as a direct substitute (not blend ed) in internal combustion engines s uch as diesel, gasoline, or jet [EPA, 2012]. The chemical makeup of renewable diesel is exactly identical to petroleum unli ke ot her biofuels such as biodiesel and ethanol (see figure 1) Figure 1 Chemical formulas for three common biofuels: Biodiesel, Ethanol, and Renewable Diesel The most commonly know n biofuel today is biodiesel or F.A.M.E (fatty acid methyl esters). This fuel has a chemical makeup that is compatible with diesel engines, but not identical to petroleum diesel [Mikulec etal 2010] In recent years a new biofuel based on pure hydrocarbon s has been heavily researched. This differs from biodiesel significantly in molecular structure and production process. From an environmental

PAGE 14

4 standpoint b oth renewable diesel and biodiesel outperform petroleum diesel in all categories of emissions, including particulates, unburned hydrocarbons, NOx, SOx, and carbon monoxide [ Kalnes etal 2009 ] This new class of second gen er ation biofuels is referred to as Biomass Based Diesel or Renewable Diesel in the EPA's Re newable Fuel Standards (RFS2) [ EPA 2012 ]. These fuels can come from diverse feedstocks such as vegetable oil, fats, algae oil, and waste oil. As opposed to biodiesel, which can only be used in a di esel engine, renewable diesel can be refined to make jet fuel, gasoline, green napthala, and other hydrocarbon based products [ Knothe etal 2010 ] There are three major chemical pathways to remove oxygen from the fatty acids founds in plant based lipids a nd hence produce renewable hydrocarbons Those pathways are hydrodeoxygentation, decarboxylation, and decarbony lation (see figure 2 ). Figure 2 This diagram which was extracted from Bambang etal, 2011 s hows the stoichiometry of the renewable diesel reaction The typical cascade reaction for upgrading triglycerides to hydrocarbon liquid fuels with three main routes: Decarboxylation, Decarbonylation, and Hydrodeoxygen tation The m ajor body of research on renewable diesel, and all current commercial endeavors, has been focused on the hydrodeoxygentation pathway since it is most similar Ni/SiO 2 Al 2 O 3 ,Pd/ c -Al 2 O 3 ,Pt/ c -Al 2 O 3 ,Ru/Al 2 O 3 ,CoMo/ c -Al 2 O 3 andNiMo/ c -Al 2 O 3 .Fig.S2showsTEMimagesofthemetal-supportedcatalystsandTableS2liststhesurfaceareas,porediameters,porevolumes,metalparticlesizes,andmetalloadingsofthe catalysts,whichwerecharacterizedusingBET,ICP-ES,andTEM measurements. Theeffectsofthecatalystsonthehydrotreatingefciencyand productcompositionwereexaminedatacatalyst/oilweightratio of0.044andtheresultsareshownin Figs.3and4 andlistedin Table1 .Thecurrentbatchreactorrequiredanhourtoreachthe experimentallydesiredtemperatureof400 C(see Fig.1 ),after whichthereactionproceededforanotheronehourat400 C.The distillationprolesofthesoybeanoilandpetroleumdieselare shownin Fig.3 forcomparisonpurposes.Thepuresoybeanoil wasdistilledmainlyat590615 C.Over85wt.%ofthepetroleum dieselwasdistilledatthenarrowboilingpointsof183359 C. WhenthesoybeanoilwashydrotreatedusingthePdorNiMocatalyst,thedistillationproleswereatoverawiderecoveryrange atthenarrowboilingpointsof290330 C,whichfallintothedieselfuelboilingpointrange.Thefractionsofhigherboilingpoints from380 Cto450 Cmaybepartiallyreactedintermediatesbetweentriglyceridesandalkanes.Thesimilarityofthenalboiling pointsofthehydroprocessedproductstothatofsoybeanoilmay betheresultofunreactedtriglycerides.Anotherpossibilityfor theheavyfractionintheliquidproductmightberelatedtooligomerizationoraromaticizationofreactionintermediatescontaining doublebondsintheirmolecularstructures [15] .Incontrast, approximately80wt.%oftheliquidproductusingtheRucatalyst wasoverthedieselfuelboilingpointrange.Thisindicatesthepredominateformationofhigh-molecular-weightspecies,probablyas aresultofthepolymerizationofthedoublebondspresentinthe triglycerides.ThelowhydroprocessingactivityoftheRucatalyst Fig.2. Possiblereactionpathwaysoftriglyceridesoverhydrotreatingcatalyst. Fig.3. Simulateddistillationcurvesofhydrotreatedproductsofvariouscatalysts. Thecatalyst/oilweightratiowas0.044. Fig.4. (a)Effectsofcatalystsonconversionandselectivityand(b)effectsof catalystsonthedrygascomposition.Thecatalyst/oilweightratiowas0.044. B.Veriansyahetal./Fuel94(2012)578585 581

PAGE 15

5 to current refinery technology and has provided the highest yields of hydrocarbons [ im‡ # ek etal 2009; Huber etal 2007; Veriansyah etal 2012 ] Other studies have looked just at the decarboxylation route or the decarbonylation route [ Na etal 2012; Rozmyszowicz etal 2011 ] One appeal to the decarboxylation r oute is that it does not consume hydrogen gas in the reaction whi ch is expensive and mainly produced from non renewable sources [Abbasi etal 2011] Figure 3 clearly shows the high cost of hydrogen gas in the br eakdown of capex and opex for a current commercial renewable di esel plant [ Pearlson, 2011 ] Figure 3 Summary of total expenses in a current HDO renewable diesel processing plant. For each tonne of hyd rogen produced from natural gas approximately 2.5 tonnes of carbon is released as CO2. When hydrogen is produced from coal, approximately 5 tonnes of carbon is emitted to the atmosphere per tonne of hydrogen. Currently over 90% of our hydrogen production comes from fossil fuels (coal natural gas, and petroleum) [Abbasi etal 2011 ]. Purpose of the Project The goal of the project is to d etermine a safe, economic, and environmentally fr iendly means to convert biomass into renewable diesel that does not require large CAPEX 25% Direct OPEX 26% Variable OPEX 19% H2 Expense 30%

PAGE 16

6 a mounts of hydrogen gas an d hence is less expensive and safe r The pro ject aims to assess the benefit s of catalytic deoxygenation to convert l ipids from non food oil crops (e.g. camelina, castor bean, jatropha, and algae and waste biomass) into renewable diesel fuel through an energy densification process using thermal decarboxylation, which does not utilize molecular hydrogen, to deoxygenate li pids to hydrocarbon fuels. The reaction scheme for decarboxylation is presentenced in figure 4. Figure 4 Stoichiometric equation for the decarboxylation reaction This research looked at the impacts of lipid saturation on the extent of the decarboxylation reaction. Since feedstocks vary dramatically in fatty aid makeup, it was important for me to understand how saturation will impact the selectivity of the reaction towards our desired product. With this understanding it will allow operators to select appropriate feedstocks in their area as well as look into possible pre treatment steps that end in desired saturation levels (see table 1) Algae oil also has the potential to be breed for desired traits such as saturatio n level [ Schlagermann etal, 2010]

PAGE 17

7 Table 1 This table show levels of saturation in some common feedstock oils In addition to the laboratory research I also compare d the environmental impacts of this process to hydrodeoxygenation and petroleum diesel using life cycle a nalysis and literature review The main difference between the two pathways to renewable diesel is the use or non use of hydrogen gas in the bio fuel production process The focus of the compariso n was on energy requirements and GHG emissions of each process as well as other inputs and outputs deemed important by the industry and EPA Experimental Method s Chemicals and Catalysts Oleic acid, stearic acid, and l inoleic acid was purchased from Sigma A ldrich and used without further purification. A hydrocalcite solid phase powdered catalyst with

PAGE 18

8 magnesium to alumina ratio of 70:30 (MG70) was purchased from Sasol. Palladium on activated c arbon (Pd/C) with a 10% loading rate was purchased from Acros che micals. Decarboxylation Reaction Several studies have previously explored the decarboxylation approach to renewable diesel production. The methods for this experiment were collected through rigorous literature review [ Na etal 2010; Na etal 2012; Rozmyszowicz etal 2011; SnÂŒre etal 2007 ; Huber etal, 2007 ]. The decarboxylation reaction was performed using a Parr bench top stirred reac tor (4590 micro reactor with a 100 ml reactio n vessel and a 4848 controller). This batch reactor was customized fo r this reaction and fit underneath a chemical fume hood to evacuate any harmful gasses during sample ejection. The reactor was designed for a max imum pre ssure of 345 bar and 500 C with a flat graphite gasket. A multi blade impeller was purchased for mixi ng a liquid reactant with a solid phase catalyst. The temperature and pressure were plotted using the Parr software provided with the reactor. In one batch reaction 28g of reactant was place d in the reactor with 1.4g of catalyst (reactant catalyst=20:1 ). After loading the reactant and catalyst the reactor was purge d with nitrogen to remove any oxygen from the vessel. The purge method was based on standard procedures from literature [ Kinsley etal 2001 ]. After purging the reactor was heated to 400 C and maintained until the reaction was completed. The reaction was monitored in real time using the pressure curve from the Parr software and was determined to be complete when the curve reached the upper asymptote (about 3 hours at 400 C and 4 hours total) Stirring speed was maintained at 150 r otations p er m inute during the reaction.

PAGE 19

9 Figure 5 This diagram extracted and modified by from Kubickova 2007, is of the Parr 4590 bench top reactor setup used in the experiment After completion the reactor was cooled down to 70 C using a customized cold water jacket (below this te mperature unreacted stearic acid is solid) Then the gaseous products were collected in a te dlar bag and the liquid product s were collected in a glass vial. A 1 ml sa mple of the liquid product was taken and d issolved in 25 ml of solvent (50/50 Dicholomethane and Acetone) and centrifuged at 5000 rpm for 5 minutes to separate any remaining catalyst. A polar and non polar solvent was chosen because the unreacted starting material, fatty acids, are polar and the product hydrocarbons, are non polar. Analysis of Products The liquid products were first characterized using a Gas Chromatograph Mass Spec ( Shimadzu GC/MS QP2010 ) equipped with a Restek 30m x 0.25mm x 0.5 $ m Rtx Wax column. The carrier gas was helium and the flow rate was 1.0 ml/min. The column temperature was programmed to increase from 40 C to 240 C at 5 C/min. Samples were deri vitized to protect the column from unreacted carboxylic acids using N2 Gas bag

PAGE 20

10 Dimethylformamide dimethyl acetal and then dissolved in D ichloromethane before injection. After characterization of the products pure hydrocarbons standards, alkanes from C 6 to C 30 were obtained and calibration curves were developed for each individual alkane using a GC FID Peak areas and retention times were associated with know concentrations of pure samples. The liquid pro ducts from the react ion were then analyzed using a Thermo tra ce GC FID equipped with a Restek 30m x 0.18mm x 0.18$m Rxi 1ms column for quantification of desired products. The carrier gas was helium and the flow rate was 2.75 ml/min. The column temperature was programmed to increase from 40 C to 330 C at a 20 C/min rate. The final temperature was held for 20 minutes and the FID base temperature was held at 350 C. The gas products were analyzed using a gas chromotograph (SRI 8610 C) equipped with an FID de tector. Results Effects of Feedstock Saturation on Liquid Products The chromatograms ( GC results ) of the liquid product analyses are presented in figure 7 The major product was heptadecane, C 17 and other minor products were alkanes from C 6 to C 17 Other hydrocarbon s may be present in the product but were not considered in this study Alkanes are viewed as more desirable over aromatics due to toxicity and suitability to diesel fuel [ Senate, 2013 ; Kubic kova etal 2005 ] Straight chain alkanes also have a higher cetane number than aromatics, however isomers and cracked hydrocarbons have physical properties desirable for fuels such as lower cloud points [Huber etal 2007]. For the purpose of this research I only looked at straigh t chain alkanes. Since the feedstocks used have a carbon count of C 18 the main product was C 17 heptadecane, due to the decaboxylation reaction removing on e carbon along with the O 2

PAGE 21

11 molecule (see fig 4 ). Equation 1 show s how yield was calculated in the analysis. The reactants we re fatty acids and the products were either heptadecane or total alkanes in the fuel range. Both results are presented in figure 6. Equation 2 shows how selectivity was calculated. Selectivity shows what percentage of the product was alkanes in the fuel range and then what percentage of the alkanes was the major product heptadecane To do this the mass concentrations from the GC were converted to volume % based on the mass and density of the different hydrocarbons. Both yield and selectivity were looked at in this study and the results are presented in figures 6 8, 9 Yield = moles of recovered p roducts / moles of reactants (1) Selectivity = volume of desired products / volume of total products (2) The yield and selectivity for C 17 heptadecane was much higher when the Pd/C catalyst was used when compared to MG70 The results of the MG70 catalyst showed an even distribution among the fuel range hydrocarbons with a slightly higher concentration in the lighter range hydrocarbons. When Pd/C was used as the catalyst there was a clear increase in diesel range alkanes as feedstock saturation increased. Selectivity towards C 17 also increased when stearic acid was used a s a feed stock when compared to oleic and linoleic acids In addition to an overall increase in alkane yield and higher selectivity towards heptadecane stearic acid had no peak for unreacted parent mate rial in the chromatorgram (fig 7 ). Overall Pd/C produced a higher concentration of diesel range hydrocarbons when compared to MG70 The concentrations of alkanes in the total liquid products (selectivity) from various exper iments are presented in table 2 and figure 8

PAGE 22

12 Figure 8 represents various jar s of product and shows what percentage of the jar is diesel range alkanes It can be seen that as feedstock saturation increases the percentage of the liquid product that is alkanes increased. Figure 9 is a similar chart but looks at the selectivity of heptadecane amongst the alkanes. Again it can be seen as feedstock saturation increases the percentage of the alkanes that is h eptadecane increases. The selectivity for heptadecane when MG70 was used as a catalyst was non existent. The product yield is also significantly lower then the product catalyzed by Pd/C Table 2 Concentration in mg/l of straight chain alkanes from C 7 to C 17 in the liquid products. Figure 6 These figures show the yield of both total fuel range alkanes and the major product heptadecane

PAGE 23

13 Figure 7 GC chromatograms of alkane reaction products at 673 K (1:hexane; 2:heptane; 3:octane; 4:nonane; 5: decane; 6: dodecane, 7: undecane; 8: tridecane; 9: tetradecane; 10: pentadecane; 11: hexadecane; 12: heptadecane; 13: oleic and stearic acid).

PAGE 24

14 Figure 8 Concentration of liquid Alkanes in the liquid product Each of the bars represent on e liter of liquid product. The blue section of the bar is identified and quantified diesel range alkanes present in the product converted from mass per volume to volume per volume using the density of each component. The red sections are other compound s most likely made up of alkenes, aromatics, isomers, and unreacted feedstock. Figure 9 This chart shows the se lectivity to C 17 heptadecane in the total alkane product Each of the bars represent on e liter of alkane product The blue section of the bar is percentage of heptadecane present in the alkane product converted from mass per volume to volume per volume using the density of heptadecane The red sections are other compound s made up of diesel range alkanes.

PAGE 25

15 Effects of F eedstock Saturation on Gaseous P roducts The gas eous products were collected, characterized, and quantified using GC and the re sults are presented in figures 10 and 11 After the gas was collected it was left for 12 hours to cool and acclimate to room temperature and pressure. The total quantity was measured using the ideal gas l aw: P1*V1 / T1 = P2*V2 / T2 Figure 10 Real time data of produced gas from the reaction. This graph shows pressure data from the reactor with varying levels of feedstock saturation and using two different catalysts (MG 70 and Pd/C). The reactor was cleared of all gasses before the reaction so this pressure curve relates to quantity of total g asses produced from the reaction. [ blue =stearic (Pd/C), red = oleic (Pd/C), green = linoleic (Pd/C), purple = oleic (MG70)] As you can see from figure 10 the pressure in the reactor increased more quickly with increasing feedstock saturation when the Pd/C c atalyst was used. When the MG70 catalyst was used the overall quantity was much lower. The concentration of CO 2 in the

PAGE 26

16 products was quantified by first injecting a known sample of CO 2 into the GC to develop a calibration curve. Then based on peak area I was able to determine the concentration of CO 2 in the gaseous products. This value decreased as unsaturation in the feedstock increased showing that the decarboxylation reaction, which produces CO2, was more prevalent when a saturated feedstock was use d When MG70 was used as a catalyst the CO 2 levels were much lower showing less decarboxylation occurred. This confirms that unsaturation of the feedstock leads to deactivation of the catalyst as well as a suppression of the decarboxylation reaction [ Rozmyszowicz etal 2011 pg 2818] Figure 11 This chart show the concentration of CO 2 in the gaseous products from GC analysis Discussion From the results we see that the decarboxylation reaction was more active based on both the liquid phase and gas phase analyses when the feedstock was more saturated This is because the reaction order goes from hydrogenation to isomerization to

PAGE 27

17 decarboxylation. When double bonds are present in the feedstock they are susceptible to isomer ization, hydrogenation, and a romatization reactions [ Rozmyszowicz, 2011 ]. These reactions inhibit decarboxylation and therefore reduce the quantity of CO 2 and suppress the formation of desired alkanes. When oleic acid was used (one double bond) the tendency is for the molecule to become hydrogenated from H 2 produced in situ from side reactions to stearic acid and then to saturated alkanes. This is even more apparent with the use of linoleic acid (two double bonds). When a fully saturated feedstock such as stearic acid is used the decarboylation reactio n happens much more quickly with fewer side reactions. This is also why stearic acid has higher selectivity to C 17 heptadecane when Pd/C is used as a catalyst. When the f eedstock is saturated the decarboxylati on reaction happens much more quickly so cracking and isomerization are less prevalent leading to a more predictable product based on the stoichiometric reaction. The fact that there is no remaining unreacted parent material in the product when stearic acid was used with Pd/C catalyst also show s a more complete reaction scenario There are still many undefined liquids in the sample s that are most likely some combination of alkenes and aromatics as well as iso mers of the parent material. Overall when MG70 was used the reaction was much less complete and there was almost no selectivity to C 17 heptadecane. It was therefore concluded that PD/C is a much more effective catalyst for this reaction and no further st udy was done using MG70

PAGE 28

18 CHAPTER II ENVIRONMENTAL ASSESMENT Introduction This chapter takes a brief look at the environmental impacts associated with renewable diesel production. This was not intended to be a full life cycle analysis on the production of renewable diese l I coupled data found in the literature on life cycle a nalysis with data on yields from experimental research to compare two pathways for renewable diesel This study was designed to see the value proposition and environmental potential in a hydrogen free renewable diesel scenario The main difference between the two pathways is the use or non use of hydrogen gas in the bio fuel production process. I focus ed on the non hydrogen method because the hydrogen p rocess has been well documented [ Pearlson etal 2011 ]. I have shown through laboratory experiments and environmental assessment that a viable alternative drop in replacement fuel can be produced with out the us e of hydrogen gas, which could reduce cost and environmental impacts. The decarboxylation process is currently on a laboratory scale. Data was collected in the lab and the yield of renewable hydrocarbons fuel was reported (see chapter 1) The hydrogenati on p rocess (HDO) data was gathered from an LCA study done by the Helsinki University of Technology on the Next BTL process at the Neste Oil plant in Kilpilahti [Nikander, 2008 ].

PAGE 29

19 Goal and Scope This chapter is not designed to be a complete life cycle assess ment, but rather a lit erature review and look at the potential environmental and economic benefits of a non hydrogen gas process to produce renewable hydrocarbons. The purpose of this chapter is to better understand the potential impacts of two different paths to renewable diesel production. One is hydrodeoxygentation (HDO) process, which is well known, and the other is a fragmen t of that reaction decarboxylation, whic h has been isolated in this study. Can this one reaction suffice in a co mmercial scale renewable diesel operation? Will the yields be great enough to make this technology cost competitive both environmentally and economically, with the better researched HDO reactor? The stakeholders being addressed through this study are ve nture capitalists and the fuel production industry. The goal is to compare the energy efficiency and environmental impacts of both HDO and non hydrogen processes. System Boundaries The syst em boundary for this study is from agricultur e to fuel transportat ion ( fig 11 ). Both processes start w ith agricultural production of s oybean oil and end with renewable diesel (hydrocarbon mix). One process uses only decarboxylation (no hydrogen), and the other uses Hydrogenation/Hydrodeoxygentaion (HDO) The functiona l unit for this LCA is the amount of fuel needed to travel 1 vehicle mile travelled (VMT) This unit is based on previous LCA studies in this field and is an easy to understand unit For the purpose of this study I gathered data from an LCA study done by the University of Helsinki for all data sets relating to the H DO process [Nikander, 2008 ]. Lab scale data was used to extrapolate the yields of the decarboxylation process

PAGE 30

20 A previous LCA study on biodiesel was used for the data for unit processes on agr icul ture, crushing, and transport [ Sheehan 1998 ]. Process flow diagram Figure 12 Process Flow Diagram This flow diagram show s the system boundary of the study from agricultural to fuel transportation Hydrogen Conversion Oil Transport Non Hydrogen Conversion Soybean Agriculture Soybean Transport Soybean Crushing Fuel Transport Fuel Transport

PAGE 31

21 Assumptio ns and Limitations Many assumptions were made while compiling this LCA study. These assumptions will be looked at in two major categories: experimental lab data and feedstock selection. Experimental Lab D ata The single biggest limitation was moving from the laboratory data collected on the non hydrogen process to a viable comparison with a commercial scale process I assumed that all inputs for the non hydrogen process at commercial scale would be the same as reported for the NexBTL process minus the energy and emissions associated with hydrogen gas. Further modeling would be needed for a more accurate representation Also all experimental data was gathered using fatty acids as model compound s in li e u of using soybean oil. These compound s are similar to soyb ean oil in composition, but they do no t contain gly cerin backbones. It was assumed that a hydrolysis step would be added to the industrial reaction scheme to first produce fatty acids from the soybean oil. Fat splitting, or h ydrolysis of triglycerides produces fatty acids and glycerol as seen in reactio n below. RCOOCH 2 CHOOCR CH 2 OCOR + 3 H 2 O % 3 RCOOH + HOCH 2 CHOH CH 2 OH My data assumes high yield of fatty acids from soybean oil and therefore might limit the accuracy of the study but provides a good comparison Since the conditions for hydrolysis is very similar to decarboxylation (350 C in an aqueous environment with a catalyst) I have considered the GHG emissions associated with hydrolysis to be equivalent to the emissions from the reactor itself [ Sinclair etal 2007 ]. Another

PAGE 32

22 limitation of this work was that the carbon emissions from the conversion processes were not consider ed emissions to the air. Since the emission s could be captured during the process it was assumed that they would be sequestered in some way. Fee dstock The only feedstock considered was soybean oil. All data on agriculture to refined oil came from data on soybean oil. Other feedstocks would offer different LCA results. All upstream data on unit processes other then the conversion step was gather ed from a n NREL LCA study on biodiesel [ Sheehan 1998 ]. Life Cycle Inventory Allocations In the LCA study from Helsinki a ll beneficial co products were assessed on their energy and GHG implications by mass balance, or boun dary expansion. All impacts were assigned to the primary product. Credits were assigned based on the amount of co product production. The two major co products relevant to this analysis are soy meal from the agriculture phase, and fuel gas from the processing phase. The amount of ener gy and GHG it takes to make an equivalent amount of co product using the traditional methods was credited to each unit process {Nikander, 2008] Method of Data Collection Each unit process presented in the process flow diagram is represented in it s own sep arate table (see appendix). The d ata for the unit processes for agriculture, t ransport, and crushing were taken from an LCA study conducted by NREL on biodies el production [Sheehan, 1998 ]. Here we assume that these processes will remain the same for

PAGE 33

23 rene wable diesel production. All data on input flows relating to the HDO process was acquired from the LCA conducted by University of Helsinki. Out put data on CO 2 e from the HDO process was reported from the previ ously mentioned LCA study conducted Helsinki Un iversity and based on an 84% yield from pre treated so ybean oil to hydrocarbon fuel [Nikander, 2008 ]. Other output data for the HDO process was obtained from the NREL LCI database [ U.S. LCI Database, 2010] For the non hydrogen reaction process lab scal e data on potential yields of hydrocarbon fuel was coupled with the industrial scale process data from the Helsinki LCA study to model production of 1 VMT of fuel quantity. There are many limitations to this approach, but it is the best option available t o me since there is no commercial scale data to date on a non hydrogen process All data has been normalized using the functional unit of one vehicle mile traveled (1 VMT). All data streams have been chosen because they are relevant to both processes. All emissions streams were chosen based on a previous LCA study, which examined biodiesel production. Below the data is represented in graphical form (figures 13 15 ) Data from the five unit processes were taken to mid point indicators using IMPACT 2002. Decarboxylation Reaction Methods I used a Parr bench top reactor with a 100 ml reaction vessel to collect data on the non hydrogen process During each reaction the vessel was loaded with oleic acid stearic acid or linoleic acid and catalyst (2 0:1 ratio), and the headspace purged with nitr ogen [Kinsley, 2001 ]. At the conclusion of the r eaction all gaseous products were collected in a tedl ar bag and analyzed using GC. The liquid products were collected and

PAGE 34

24 analyzed using GD FID. Based on the fin dings the median conversion of a mixture of fatty acids t o fuel range hydrocarbons was 45 %. Using these numbers and standardizing to 1 VMT of fuel produced the results that are presented in table 3 All data sets on material flows are presented in the a ppendix. Conversion Methods During the course of the study it was appropriate to use different units based on the process being analyzed. All data was eventually converted to 1 vehicle mile traveled. Included below is a chart referencing all relevant co nver sions made during the study. All numbers were based off of the 25 mpg l ight truck average in the US [US DOT, 2011 ] Table 3 A collec tion of calculations needed to compute environmental impacts of biofuels. Variable value source Miles per gallon of petroleum 25 MPG [10] 1 gallon of Highway Diesel 135.8 MJ [ORNL, 2008 ] 1 VMT Needs 5.432 MJ [ORNL, 2008 ] 1 kg soy beans 0.18 kg soy oil [USDA, 2010 ] 0.147 kg soy oil (165.7 ml) 1 VMT calculation 0.139 kg renewable Diesel 1 VMT [ORNL, 2008 ] Denver electricity CO2e emission factor 0.8 kgCO2e/kWh [Ramaswami, 2008 ] Denver electricity emission factor 0.22 kg CO2 e/MJ calcu lation 49 MJ 1 kg Natural gas NREL database 1 kWh 3.6 MJ [Cengal, 2011 ] Environmental Impact Assessment Emission factors for the unit processes of agriculture, transportation, and crushing were gathered from the NRE L LCA study conducted on biodiesel [Sheehan, 1998 ]. Emission factors for the hydrogen NexBTL process we re gathered from two different LCA studies on Neste Oil's pl ant in Proovo Finland [Nikander, 2008; Reinhardt, 2006 ].

PAGE 35

25 All data for the non hydrogen process was scaled up using the study from Finland and laboratory data on yield and catalysts Characterization factors came fro m IMPACT 2002. This study only analyze d mid point categories and did not address damage categories. Mid Point Impacts analyzed: global warming potential, carcinogenic effects, and aquatic ecotoxicity. The data is reported in the figures below. Figure 13 G lobal Warming Potential Indicator GWP is reported in carbon dioxide equivalents. All life cycle transportation emissions were combined into one bar The processing stage has the HDO process in blue on the bottom and the non hydrogen process in red on top.

PAGE 36

26 Figure 14 Carcinogenic Effects Indicator Carcinogenic effects are reported in chloroethylene equivalents. The unit processes are listed separately and all li fe cycle transportation data is shown in one bar The processing stage has the HDO process in blue on the bottom and the non hydrogen process in red on top. Figure 15 Aquatic Eco Toxicity Indicator Aqua t i c eco toxicity is shown as Tr iethylene glycol equivalents. A ll unit processes are show separately and life cycle transportation is collected into one bar. The processing stage has the HDO process in blue on the bottom and the non hydrogen process in red on top.

PAGE 37

27 Table 4 Cradle to grave comparison of petroleum diesel, HDO renewable diesel and non hydrogen renewable diesel Figure 16 This chart show the life cycle GHG emissions from cradle to grave of three pathways to fuel production normalized to petroleum diesel The chart shows that the HDO process has a 20% decrease in life cycle GHG emissions and the un optimized non H2 process having a 30% decrease when compared to petroleum diesel.

PAGE 38

28 Fi gure 17 CO2e emissions from only the processing stage of three possible pathways to diesel fuel. Interpretation Goal and Scope The goal of the analysis was to determine which unit process has the greatest impact on the three mid point categories selected: global warming potential (GWP), carcinogenic effects, and aquatic eco toxicity In addition to determining which unit process had the greatest impact a direct comparison of two approaches to renewable diesel conversion was ana lyzed using energy inputs and GHG emissions

PAGE 39

29 Results and Discussion This study found that the agricultural phase of the biofuels industry is the largest producer of GHG emissions and therefore has the highest global warming potential That is comparible with other LCA studies in this field [Sheehan, 1998 ] In addition to emitting GHG emissions, agriculture also has the largest impact on ecosystem toxicology and cance r causing substances [figures 14 and 15 ]. This is mainly due to the energy an d fossil fuel intense nature of industrial agriculture as well as runoff to rivers and streams This study shows that the agriculture industry has more work to be done before being considered a viable alternative to the current petroleum industry. Accord ing to the LCI even with 40 % lower yields the non hydrogen process had less of an impact on global warming potential and aquatic e co toxicit y then did the HDO proces s. The renewable diesel plant in this study uses natural gas reforming to produce hydrogen as is the case with most hydrogen production in the world today [Abassi etal 2011]. Hydrogen production uses large quantities of fossil energy and is responsibl e for the high GWP and aquatic e co toxicit y of the HDO process in addition to contri buting to 30% of the financial costs The carcinogenic mid point indicator showed the HDO process being slightly less harmful then the non hydrogen process. This is attributed to the increased usage of electricity both from coal and natural gas needed in the non hydroge n process to produce an equivalent yield in biof uel as the HDO process. This resulted in the increase d release of cancer causing gasses to the air and eventually into humans. The results from comparing the GHG emissions of the HDO pr oce ss, the non h ydrogen process, and petroleum diesel ar e presented in figures 15 and table 4 The data for the petroleum life cycle energy inputs and emission was collected and compiled from an

PAGE 40

30 NREL study comparing petroleum to first generation biodiesel [S heehan, 1998]. The data for the HDO process and the non hydrogen process was taken from the Helsinki study [Nikander, 2008]. All of the upstream energy inputs and emissions for the HDO process and the non hydrogen process are exactly the same. T he diffe rence that can be seen between the two renewable diesel processes starts in the processing and pre treatment stage highlighted in yellow in table 4 Here it can be seen that hydrogen has been removed from the process resulting in a decrease in energy inpu ts as well as GHG emissions. This decre ase is seen even when a 40 % drop in biofuel yield from the non hydrogen process. It can also be seen from figure 16 that the current commercially available HDO renewable diesel processes has higher GHG emissions the n even petroleum diesel This means that the fuel would not qualify as Biomass Based Diesel which call s for a 50% reduction in life cycle GHG emissions [EPA 2012]

PAGE 41

31 Chapter III C ONCLUSIONS Significant Findings The journey towards developing a viable alternative to cheap energy dense fossil fuels is an incremental one. Many research groups and industries are working simultaneousl y at a multitude of routes for achieving that end. This study has also added to that body of knowledge. T hrough laboratory experiments, literature review, and environmental assessment I have verified that a viable alternative drop in replacement fuel can be produced with out the use of hydrogen gas, which in turn could reduce cost and environmental impacts of biofu els I have shown that a significant quantity of saturated hydrocarbons can be produced from fatty acids found in soybean oil us ing the decarboxylation route This means that the potential for producing biofuels with increasingly lower green house gas emi ssions and reduced fossil energy inputs is possible. In addition this study has also directly compared the energy inputs and GHG emissions of petroleum diesel, commercial renewable diesel, and renewable diesel without the use of hydrogen gas. These prel iminar y findings show that even with an un optimized reaction a non hydrogen process could have half of the emissions of current commercial renewable diesel in the processing stage As we continue to optimize the reaction and increase yields energy demands and GHG emission will look even more promising This shows that it is possible to eliminate the need for large amounts of environmentally de trimental hydrogen gas in commercial scale renewable diesel production

PAGE 42

32 Next Steps The results pr esented in thi s thesis are exciting and hopeful. But this is just the beginning. The most promising aspect is the capacity building that took place over the course of the research. The Green Engineering and Systems Analysis Lab at CU Denver now has the foundation to build on this research and continue producing state of the art research in advanced biofuels. At the beginning of the research there was a reactor sitting in a corner of a dusty unused lab space. Now the lab is fully equipped to use both ba tch and semi batch processes to produce biofuels as well as to analyze and report on the products and results of the reaction s. Yield optimization of the reaction is underway looking at temperature and pressure as well as catalyst activation/ deactivation At 400 C cracking and isomerization is prevalent, so decreasing the reaction temperature is highly favorable for both product selectivity and energy reduction [Huber, 2007]. More analysis and research needs to be done to further characterize the comple te liquid product to better understand the reaction mechanics. In addition research on a catalytic process to hydrogenate unsaturated feedstock in the reaction vessel in underway. The use of a proton donating solvent has been reported to be effective [Kubic kova etal, 2008]. In addition I plan to employ a food industry process called catalytic transfer hydrogenation as a pre treatment step Differing from the c lassical hydrogenation techniques using molec ular hydrogen, hydrogen don ors are used as a s ource of hy drogen in a catalytic transfer reduction to saturate the feedstock The gene ralized e quation 1 represents this process: DH+ A catalyst % D+ AH Here "A" represents the acceptor and "D" represents the hydrogen donor [Naglic, 1998

PAGE 43

33 pg 1; Dumesic, 2011]. This process could greatly decrease the cost and environmental issues with using molecular hydrogen while increasing the yields of fuel range hydroc arbons from the decarboxylation reaction presented in this paper. Based on these findings we plan to work on catalytic transfer hydrogenation to saturate feedstocks before the decarboxylation reaction. Little work has been done in this real m and it could produce significant increases in yields. In addition we plan to look into wet algae processing for biofuels production. This paper showed that the most significant contributor to life cycle environmenta l impacts is the agricultural phase. In order to reach the 50% reduction in life cycle GHG emission mandated by the RFS2 fuel standards much work must be done in the feedstock production real m Algae ar e being considered as a possible solution to environmental problems of in ground agricu lture I have started cultivating st r ains of algae in the lab and am look ing into oil extraction methods as well as wet processing. Now that we have biofuels research capacity at UCD there is limitless directions that we can take the research. Cheers.

PAGE 44

34 R EFERENCES Abbasi, T., & Abbasi, S. a. (2011). "Renewable" hydrogen: Prospects and challenges. Renewable and Sustainable Energy Reviews 15 (6), 3034 3040. doi:10.1016/j.rser.2011.02.026 Barnes, B., Mathee, A., & Moiloa, K. (2005). Assessing ch ild time activity patterns in relation to indoor cooking fires in developing countries: a methodological comparison. International journal of hygiene and environmental health 208 (3), 219 25. doi:10.1016/j.ijheh.2005.01.022 Cengel, Yunus A. Boles, Michael A. (2011) Thermodynamics: An Engineering Approach. New York, NY, McGraw Hill Chia, M., & Dumesic, J. a. (2011). Liquid phase catalytic transfer hydrogenation and cyclization of levulinic acid and its esters to & valerolactone over metal oxide catalysts. Ch emical communications (Cambridge, England) 47 (44), 12233 5. doi:10.1039/c1cc14748j EPA 1320 SUMMARY': Regulation of Fuels and Fule Additives: 2012 Renewable Fuel Standards (2012). Huber, G. W., Iborra, S., & Corma, A. (2006). Synthesis of transportation fuels from biomass: chemistry, catalysts, and engineering. Chemical reviews 106 (9), 4044 98. doi:10.1021/cr068360d Huber, G. W., O'Connor, P., & Corma, A. (2007a). Processing biomass in conventional oil refin eries: Production of high quality diesel by hydrotreating vegetable oils in heavy vacuum oil mixtures. Applied Catalysis A: General 329 120 129. doi:10.1016/j.apcata.2007.07.002 Huber, G. W., O'Connor, P., & Corma, A. (2007b). Processing biomass in conve ntional oil refineries: Production of high quality diesel by hydrotreating vegetable oils in heavy vacuum oil mixtures. Applied Catalysis A: General 329 120 129. doi:10.1016/j.apcata.2007.07.002 Kalnes, T. N., Koers, K. P., Marker, T., & Shonnard, D. R. (2009). A Technoeconomic and Environmental Life Cycle Comparison of Green Diesel to Biodiesel and Syndiesel, 28 (1), 111 120. doi:10.1002/ep Kinsley, G. R. J. (2001). Properly Purge and Imert Storage Vessels. Chemical Engineering Progress 97 (2), 57 61. Knothe, G. (2010a). Biodiesel and renewable diesel: A comparison. Progress in Energy and Combustion Science 36 (3), 364 373. doi:10.1016/j.pecs.2009.11.004

PAGE 45

35 Knothe, G. (2010b). Biodiesel and renewable diesel: A comparison. Progress in Energy and Combustion Science 36 (3), 364 373. doi:10.1016/j.pecs.2009.11.004 Koloini, T., & Andrej, S. (1998). Kinetics of Catalytic Transfer Hydrogenation. Journal of American Oil Chemists Society 75 (5), 629 633. Kubi#kov‡, I., SnŒre, M., ErŠnen, K., MŠki Arvela, P., & Murzi n, D. Y. (2005). Hydrocarbons for diesel fuel via decarboxylation of vegetable oils. Catalysis Today 106 (1 4), 197 200. doi:10.1016/j.cattod.2005.07.188 Kubickova, I., SnŒre, M., Era, K., Murzin, D. Y., & Turku, F. (2007). Catalytic Deoxygenation of Fat ty Acids and Their Derivatives. Energy and Fuel 21 (16), 30 41. Mikulec, J., Cvengro(, J., Jor’kov‡, )., Bani#, M., & Kleinov‡, A. (2010). Second generation diesel fuel from renewable sources. Journal of Cleaner Production 18 (9), 917 926. doi:10.1016/j.jc lepro.2010.01.018 Na, J. G., Han, J. K., Oh, Y. K., Park, J. H., Jung, T. S., Han, S. S., Yoon, H. C., et al. (2012). Decarboxylation of microalgal oil without hydrogen into hydrocarbon for the production of transportation fuel. Catalysis Today 185 (1), 31 3 317. doi:10.1016/j.cattod.2011.08.009 Na, J. G., Yi, B. E., Kim, J. N., Yi, K. B., Park, S. Y., Park, J. H., Kim, J. N., et al. (2010). Hydrocarbon production from decarboxylation of fatty acid without hydrogen. Catalysis Today 156 (1 2), 44 48. doi:10.1 016/j.cattod.2009.11.008 Nikander, S. (2008). GREENHOUSE GAS AND ENERGY INTENSITY OF PRODUCT CHAIN!: CASE TRANSPORT BIOFUEL Helsinki University of Technology. Oak Ridge National Laboratory (2008), Conversion Factors for Bioenergy, retrieved from www.ces.ncsu.edu/forestry/biomass.html Paula, Matthew de (2011). Veggie Oil Smells Better Than Diesel, But It's No Slam Dunk. Forbes,. Retrieved from http://www.forbes.com/sites/matthewdepaula/2011/05/27/veggie oil might smell better than diesel but its no slam dunk/ Pearls on, M. N. (2011). A Techno Economic and Environmental Assessment of Hydroprocessed Renewable Distillate Fuels by MIT. Ramaswami, A., Hillman, T., Janson, B., Reiner, M., & Thomas, G. (2008). Policy Analysis A Demand Centered Hybrid Life Cycle Methodolog y for City Scale Greenhouse Gas Inventories. American Chemical Society 42 (17), 6455 6461.

PAGE 46

36 Reinhardt, G. (2006). Final Report An Assessment of Energy and Greenhouse Gases of NExBTL (pp. 1 30). Porvoo, Findland. REN21. (2011). Renewables 2011 Global Satus R eport (pp. 1 116). Retrieved from www.ren21.net Roh, H. S., Eum, I. H., Jeong, D. W., Yi, B. E., Na, J. G., & Ko, C. H. (2011). The effect of calcination temperature on the performance of Ni/MgO Al2O3 catalysts for decarboxylation of oleic acid. Catalysis Today 164 (1), 457 460. doi:10.1016/j.cattod.2010.10.048 Rozmyszowicz, B., Lestari, S., Simakova, O., Er, K., Salmi, T., & Murzin, D. Y. (2011). Catalytic Deoxygenation of Tall Oil Fatty Acid over Palladium Supported on Mesoporous Carbon. Energy and Fuels (25), 2815 2825. Schlagermann, P., Gšttlicher, G., Dillschneider, R., Rosello Sastre, R., & Posten, C. (2012). Composition of Algal Oil and Its Potential as Biofuel. Journal of Combustion 2012 1 14. doi:10.1155/2012/285185 Senate Low Aromatic Fuel Act 2 013 (2013). Retrieved from www.comlaw.gov.au/ *enol, O., Ryymin, E. M., Viljava, T. R., & Krause, a. O. I. (2007). Effect of hydrogen sulphide on the hydrodeoxygenation of aromatic and aliphatic oxygenates on sulphided catalysts. Journal of Molecular Catal ysis A: Chemical 277 (1 2), 107 112. doi:10.1016/j.molcata.2007.07.033 Serrano Ruiz, J. C., Pineda, A., Balu, A. M., Luque, R., Campelo, J. M., Romero, A. A., & Ramos Fern‡ndez, J. M. (2012). Catalytic transformations of biomass derived acids into advanced biofuels. Catalysis Today 195 (1), 162 168. doi:10.1016/j.cattod.2012.01.009 Sheehan, J. et al. (1998). Life Cycle Inventory of Biodiesel and Petroleum Diesel for Use in an Urban Bus (pp. 1 314). "im‡#ek, P., Kubi#ka, D., "ebor, G., & Posp’(il, M. (2009). Hydroprocessed rapeseed oil as a source of hydrocarbon based biodiesel. Fuel 88 (3), 456 460. doi:10.1016/j.fuel.2008.10.022 Sinclair, C., & Manager, E. (2007). Cremation of Human Remains!: A Comparison of Alkaline Hydrolysis versus Combustion (pp. 1 6). Retrieved from www.resomation.com SnŒre, M., Kubi#kov‡, I., MŠki Arvela, P., Chichova, D., ErŠnen, K., & Murzin, D. Y. (2008). Catalytic deoxygenation of unsaturated renewable feedstocks for production of diesel fuel hydrocarbons. Fuel 87 (6), 933 945. doi :10.1016/j.fuel.2007.06.006

PAGE 47

37 SnŒre, M., Kubi#kov‡, I., MŠki Arvela, P., ErŠnen, K., WŠrnŒ, J., & Murzin, D. Y. (2007). Production of diesel fuel from renewable feeds: Kinetics of ethyl stearate decarboxylation. Chemical Engineering Journal 134 (1 3), 29 34. doi:10.1016/j.cej.2007.03.064 Solecki, M., & Epstein, B. (2011). Advanced Biofuel Market Report 2011 Meeting the California LCFS Authors Table of Contents (pp. 1 22). Tat, M. E., Wang, P. S., Gerpen, J. H., & Clemente, T. E. (2007). Exhaust Emissions from an Engine Fueled with Biodiesel from High Oleic Soybeans. Journal of the American Oil Chemists' Society 84 (9), 865 869. doi:10.1007/s11746 007 1109 6 U.S. Department of Agriculture (2010), Soybeans and Oil Crops: Background. Retrieved from: http://www.ers.usda.gov/Briefing/SoybeansOilcrops/background.htm U.S. Department of Transportation, Federal Highway Administration, (2011) Highway Statistics (Washington, DC: Annual Issues), table VM 1, available at http://www.fhwa.dot.gov/policyinformation/statistics.cfm "U.S. Life Cycle Inventory Database." (2012). National Renewable E nergy Laboratory, 2012. Accessed November 19, 2012: https://www.lcacommons.gov/nrel/search Veriansyah, B., Han, J. Y., Kim, S. K., Hong, S. A., Kim, Y. J., Lim, J. S., Shu, Y. W., et al. (2012). Production of renewable diesel by hydroprocessing of soybean oil: Effect of catalysts. Fuel 94 578 585. doi:10.1016/j.fuel.2011.10.057 Wang, W. C., Roberts, W. L., & Stikeleather, L. F. (2012). Hydrocarbon Fuels From Gas Phase Decarboxylation of Hydrolyzed Free Fatty Acid. Journal of Energy Resources Technology 1 34 (3), 032203. doi:10.1115/1.4006867 Wang, W. C., Thapaliya, N., Campos, A., Stikeleather, L. F., & Roberts, W. L. (2012). Hydrocarbon fuels from vegetable oils via hydrolysis and thermo catalytic decarboxylation. Fuel 95 622 629. doi:10.1016/j.fuel.2011.12.041

PAGE 48

38 APPENDIX Soybean Agriculture Unit Process Data This includes fuels, fertilizers, and inputs used on the farm and their associated emissions, as well as the upstream emission s associated with those inputs used to grow soybeans. These upstream environmental flows are combined with the flows associated with the actual soybean growing and harvesting to calculate the total emissions associate d with soybean agriculture [Sheehan, 199 8 ]. All data has been normalized to 1 VMT. Table 5 Soybean agriculture raw materials data Raw Materials for Soybean Agriculture Units kg of raw material/ 1 VMT Coal (in ground) kg 0.0474376 Oil (in ground) kg 0.2439248 Natural Gas (in ground) kg 0.0928704 Uranium (U, ore) kg 0.000002604 Phosphate Rock (in ground) kg 0.335776 Potash (K2O, in ground) kg 0.1578192 Perlite (SiO2, ore) kg 0.000047544 Limestone (CaCO2, in ground) kg 0.0061376 Sodium Chloride (NaCl) kg 0 Water Use kg 3081.96 Elec from Coal kg 2.44048 Elec from Natural gas kg 2.44048 Table 6 Soybean agriculture emissions data Emissions from Soybean Agriculture Units Total emissions / 1 VMT AIR Carbon Dioxcide (CO2, fossil) kg 1.024072 Carbon Dioxcide (CO2, biomass) kg 0 Methane (CH4) kg 0.001012592 Nitrous Oxcide (N2O) kg 0.00004676 Carbon Monoxcide (CO) kg 0.004890312 Hydrocarbons kg 0.006069504 Benzene kg 4.6928E 08 Formaldehyde kg 6.328E 07 Particulates, unspecified kg 0.001043084 Sulfur oxides kg 0.003357558 Nitrogen oxides kg 0.007168 Hydrogen chloride kg 0.00000952 Hydrogen fluoride kg 4.3848E 07 Ammonia (air) kg 0.0026208

PAGE 49

39 WATER Agrochemicals kg 0.00037352 BOD5 kg 0.000189896 COD kg 0.00134624 Metals kg 0.000005824 Ammonia kg 0.000022512 Nitrates kg 9.576E 08 Soybean Transport Unit Process Data The transport of soybeans from the field to the crusher is based on current practices in the soybean farming and soybean crushing sectors. The actual data used in this study were based on a combination of conversations with industry representatives and modeled data. They ignored transporting soybeans for the export market, which represents the longest transportation distances in the industry [ Sheehan, 1998 ]. Table 7 Soybean transport raw materials data Raw Materials for Soybean Transport Units Kg of material / 1 VMT Coal (in ground) kg 0.00059472 Oil (in ground) kg 0.01903664 Natural Gas (in ground) kg 0.00163464 Uranium (U, ore) kg 1.316E 08 Phosphate Rock (in ground) kg 0 Potash (K2O, in ground) kg 0 Perlite (SiO2, ore) kg 0.000004088 Limestone (CaCO2, in ground) kg 0.00011312 Sodium Chloride (NaCl) kg 0 Water Use kg 0.00264936 Elec from Coal kg 0.126017889 Elec from Natural gas kg 0.126017886 Table 8 Soybean transport emissions data Emissions from Soybean Transport Units Total emissions / 1 VMT AIR

PAGE 50

40 Carbon Dioxcide (CO2, fossil) kg 0.0634228 Carbon Dioxcide (CO2, biomass) kg 0 Methane (CH4) kg 2.27276E 05 Nitrous Oxcide (N2O) kg 6.30168E 06 Carbon Monoxcide (CO) kg 0.00021512 Hydrocarbons kg 0.0000714 Benzene kg 4.2728E 09 Formaldehyde kg 5.712E 08 Particulates, unspecified kg 8.54661E 05 Sulfur oxides kg 0.000093408 Nitrogen oxides kg 0.00059673 Hydrogen chloride kg 3.192E 07 Hydrogen fluoride kg 3.9928E 08 Ammonia (air) kg 8.12E 11 WATER Agrochemicals kg 0 BOD5 kg 0.000012768 COD kg 0.000108024 Metals kg 5.2808E 07 Ammonia kg 1.86704E 06 Nitrates kg 5.0568E 10 Soybean Crushing Unit Process Data The crushing process in this study is assumed to be representative of typical soybean crushing operations in the United States. The major distinguishing feature among the types of processes used in the U.S. crushing plants is in the oil extraction method. Our model facility uses solvent extraction to recover the oil. The other major method for extraction is mechanical crushing. Only 1% to 2% of the soybeans processed in the Unit ed States are recovered via mechanical extraction methods [ Sheehan, 1998 ] Table 9 Soybean crushing raw materials data Raw Materials for Soybean Crushing Units Raw materials / 1 VMT Coal (in ground) kg 0.019754224 Oil (in ground) kg 0.003182424 Natural Gas (in ground) kg 0.053525584 Uranium (U, ore) kg 4.7348E 07 Phosphate Rock (in ground) kg 0 Potash (K2O, in ground) kg 0 Perlite (SiO2, ore) kg 0 Limestone (CaCO2, in ground) kg 0.00376656

PAGE 51

41 Sodium Chloride (NaCl) kg 0 Water Use kg 0.004359208 Elec from Coal kg 0.508763304 Elec from Natural gas kg 0.508763304 Table 10 Soybean crushing emissions data Emissions from Soybean Crushing Units Total emissions from Agg/ 1 VMT AIR Carbon Dioxcide (CO2, fossil) kg 0.21234768 Carbon Dioxcide (CO2, biomass) kg 0 Methane (CH4) kg 0.000395656 Nitrous Oxcide (N2O) kg 1.92284E 06 Carbon Monoxcide (CO) kg 6.43876E 05 Hydrocarbons kg 0.001979569 Benzene kg 0 Formaldehyde kg 3.18E 15 Particulates, unspecified kg 0.000281171 Sulfur oxides kg 0.001516627 Nitrogen oxides kg 0.000398267 Hydrogen chloride kg 1.06201E 05 Hydrogen fluoride kg 1.32712E 06 Ammonia (air) kg 4.8972E 08 WATER Agrochemicals kg 0 BOD5 kg 2.6818E 07 COD kg 2.19314E 06 Metals kg 6.2858E 08 Ammonia kg 3.53404E 07 Nitrates kg 1.6854E 08 Soybean Oil Transport Unit Process Data The NREL LCA practitioners studied average transport distances for soybean oil in depth. This data includes average fuel emissions from wells to pump and pump to wheels [ Sheehan, 1998] Table 11 Soybean transport data raw materials Raw Materials for Soybean Oil Transport Units Kg of raw material / 1 VMT

PAGE 52

42 Coal (in ground) kg 0.00021702 Oil (in ground) kg 0.006945164 Natural Gas (in ground) kg 0.000596416 Uranium (U, ore) kg 5.1072E 09 Phosphate Rock (in ground) kg 0 Potash (K2O, in ground) kg 0 Perlite (SiO2, ore) kg 1.57472E 06 Limestone (CaCO2, in ground) kg 4.11768E 05 Sodium Chloride (NaCl) kg 0 Water Use kg 0.000966573 Elec from Coal kg 0.0459648 Elec from Natural gas kg 0.0459648 Table 12 Soy bean transport emissions data Emissions from Soybean Oil Transport Units Total emissions from Oil Transport / 1 VMT AIR Carbon Dioxcide (CO2, fossil) kg 0.023296044 Carbon Dioxcide (CO2, biomass) kg 0 Methane (CH4) kg 8.37718E 06 Nitrous Oxcide (N2O) kg 2.4839E 07 Carbon Monoxcide (CO) kg 7.77499E 05 Hydrocarbons kg 2.80794E 05 Benzene kg 1.5476E 09 Formaldehyde kg 2.0776E 08 Particulates, unspecified kg 1.42888E 05 Sulfur oxides kg 3.39094E 05 Nitrogen oxides kg 0.00038425 Ammonia (air) kg 2.9574E 11 WATER Agrochemicals kg 0 BOD5 kg 4.64074E 06 COD kg 3.92678E 05 Metals kg 1.9193E 07 Ammonia kg 1.8338E 10 Nitrates kg 1.5052E 08 Soybean Oil Conversion to Renewable Diesel using Hydrogen Process This data set includes the energy use fro m the HDO reaction at the NESTE Oil plant in Finland plus all inputs from production of hydrogen gas [Nikander, 2008]

PAGE 53

43 Table 13 HDO raw materials data assuming an 84% yield Raw Materials for HDO conversion Units Kg of raw material / 1 VMT Treated Soybean Oil kg 0.14699 Natural Gas (for hydrogen) kg 0.00588 Fuel Gas (for hydrogen) kg 0.00196 Process Water kg 0.02861 Cooling water Use kg 0.0004887 Steam MJ 0.003801 Elec from Coal kWh 0.00332 Elec from Natural gas kWh 0.00332 Table 14 HDO emissions data assuming an 84% yield Emissions Units Total emissions from Agg/ 1 VMT AIR CO2e (from electricity) kg 0.001 1 CO2e (from Steam) kg 0.00396 CO2e (from hydrogen) kg 0.04561 Methane (CH4) kg 2.37478E 05 Nitrous Oxcide (N2O) kg 0 Carbon Monoxcide (CO) kg 3.31169E 06 Hydrocarbons kg 0 Benzene kg 1.87249E 09 Formaldehyde kg 2.33507E 09 Particulates, unspecified kg 2.95796E 07 Sulfur oxides kg 0.002156816 Nitrogen oxides kg 3.11198E 06 Ammonia (air) kg 0 0 WATER 0 Agrochemicals kg 1.7376E 08 BOD5 kg 0 COD kg 2.37478E 05 Metals kg 0 Ammonia kg 3.31169E 06 Nitrates kg 0

PAGE 54

44 Soybean Oil Conversion to Renewable Diesel using Non Hydrogen Process This data set looks at all inputs and relevant outputs from the lab scale yield data overlaid on the process data from the Neste Oil plan t [Nikander, 2008] Table 15 Non hydrogen raw materials data assuming a 45 % yield Raw Mate rials for Non H2 conversion Units Kg of raw material / 1 VMT Treated Soybean Oil kg 0.1940268 Palladium on Carbon kg 0.0096888 Process Water kg 0.0377652 Cooling water Use kg 0.000645084 Steam MJ 0.00501732 Elec from Coal kWh 0.0043824 Elec from Natural gas kWh 0.0043824 Table 16 Non hydrogen emissions data assuming a 45 % yield Emissions Units Total emissions from Agg/ 1 VMT AIR CO2e (from electricity) kg 0.00151 CO2e (from Steam) kg 0.00554 CO2e (from Palladium) kg 0.0000015 Methane (CH4) kg 2.99515E 05 Nitrous Oxcide (N2O) kg 0 Carbon Monoxcide (CO) kg 4.20173E 06 Hydrocarbons kg 0 Benzene kg 2.35595E 09 Formaldehyde kg 2.93798E 09 Particulates, unspecified kg 2.60605E 05 Sulfur oxides kg 0.002713707 Nitrogen oxides kg 4.35272E 06

PAGE 55

45 Ammonia (air) kg 0 WATER Agrochemicals kg 0 BOD5 kg 0 COD kg 0 Metals kg 0 Ammonia kg 2.18624E 08 Nitrates kg 7.10528E 12