BENCH SCALE ANALYSIS OF BIOGAS PRODUCTION FROM BIODIESEL
SLUDGE USING A.D. KARVES ANAEROBIC DIGESTER DESIGN
Lindsay Elisabeth Voss
B.S., Worcester Polytechnic Institute, 2000
A thesis submitted to the
University of Colorado at Denver and Health Sciences Center
In partial fulfillment
Of the requirements for the degree of
Master of Science
This thesis for the Master of Science
Lindsay Elisabeth Voss
has been approved
Voss, Lindsay Elisabeth (M.S., Environmental Sciences)
Bench Scale Analysis of Biogas Production from Biodiesel Sludge Using A.D.
Karves Anaerobic Digester Design
Thesis directed by Professor Anuradha Ramaswami, Ph.D.
This study tests claims of high methane yields (0.475 cm3 CH4/mg
CODred) and high methane content (>90%) in biogas produced from an
innovative, open-system anaerobic digester developed by A.D. Karve in India.
High methane purity in Karves design was attributed to selective removal of
carbon dioxide (CO2) by equilibration with the atmosphere.
This thesis replicates the Karve design at the bench scale, testing
methane yields and purity in biogas obtained from three substrates: a simple
sugar, simulated kitchen wastewater, and biodiesel sludge. Methane yields
ranged from 0.231 cm3 CH4/mg CODred for simple sugar to 0.057 cm3 CHVmg
CODred for biodiesel sludge, much less than that reported by Karve. The
highest methane content in biogas produced was 74%, and depended more
upon substrates than digester design. Abiotic experiments confirmed the CO2
loss pathway to the atmosphere but found the rate to be insufficient to
achieve methane purity >90%.
Although bench scale biodigester results did not approach Karves
stated high levels of performance, they were consistent with results published
by other researchers. These lab findings were scaled to the plant level to
estimate renewable energy (BTUs) that can be obtained through anaerobic
digestion of biodiesel sludge waste at a local biodiesel plant in Colorado.
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
I would like to acknowledge my advisor, Dr. Anu Ramaswami, and my
research partner, Mark Pitterle, for their contributions and support of my
I would also like to thank Aaron Perry and Luke Eisenhauer from Rocky
Mountain Sustainable Enterprises for approaching us with this project
I also wish to thank the members of my committee for their valuable
participation and insights.
TABLE OF CONTENTS
1.2 Biogas Plant Technology.........................................3
1.3 Biogas from Biodiesel Sludge....................................6
2. Literature Review...............................................8
2.1 A.D. Karves Biodigester Design.................................8
2.2 Results of Comparable Research.................................12
3.1 Experimental Design............................................14
4. Materials and Methods..........................................15
4.1 Digester Design and Construction...............................15
4.2.1 Sucrose (Table Sugar).....................................18
4.2.2 Simulated Wastewater......................................18
4.2.3 Biodiesel Sludge..........................................20
4.2.4 Nutrient Mix..............................................20
4.2.5 Alkalinity Buffer.........................................21
4.3 Feed Procedure and Frequency...................................22
5. Analytical Methods.............................................24
5.3 Volatile Fatty Acids (VFAs)...................................24
5.4 Chemical Oxygen Demand (COD)..................................25
5.5 Biogas Volume.................................................25
5.6 Biogas Composition............................................25
5.6.1 Carbon Dioxide (C02).....................................26
5.6.2 Hydrogen Sulfide (H2S)...................................26
5.6.3 Methane (CH4)............................................26
5.7 Methane Production from COD Measurements......................27
5.8 Calculation of Actual Biogas and Methane Production...........28
5.9 Abiotic Testing Methods.......................................29
6.1 Abiotic Results...............................................32
6.1.1 Carbon Dioxide Loss Experiments..........................32
6.1.2 High Pressure Carbon Dioxide Loss Experiment.............36
6.1.3 Methane Loss Experiments.................................37
6.2 Biotic Results................................................37
6.2.1 Biotic Tests from Sucrose................................39
6.2.2 Biotic Tests with Simulated Wastewater...................42
6.2.3 Biotic Tests with Biodiesel Sludge.......................43
6.3 Plant Scale Energy Yield Using Karves Design.................45
7.1 Biogas Composition............................................47
7.2 Biogas Yield..................................................49
LIST OF FIGURES
1: Illustration of Anaerobic Metabolism (Grady et al., 1999)...............2
2: Schematic of a Biogas Production Plant, Including Gas (Hansen, 1992)....5
3: Photo of Biodigester Used in Karves 2005 Field Trials (Pitterle, 2005).9
4: Schematic of Karves Biodigester........................................16
5: Karve Digester Used in Biotic Experiments (Kocman, 2005)................17
6: Feed Composition Change Methodology.....................................23
7 : Carbon Dioxide Loss Using Small Gap....................................33
8: Carbon Dioxide Loss Using Medium Gap....................................34
9: Carbon Dioxide Loss Using Large Gap.....................................34
10: Comparison of Carbon Dioxide Loss Using Different Base Sizes.......36
11: Methane Loss Results Using Medium Gap..............................38
12: Methane Loss Results Using Large Gap...............................39
13: Methane Production from Sucrose Feedstock..........................41
14: pH Fluctuation during Sucrose Feed Trial with and without Alkalinity
15: Methane Production from Simulated Wastewater Feedstock with pH
16: Methane Production from Biodiesel Sludge Feedstock with pH Trend.... 44
17: Correlation between Organic Content Loaded and Biogas Production ...45
18: Methane Content Found in Biogas Produced from Different Feeds..........48
LIST OF CALCULATIONS
1: Biogas Composition Estimation for Simple Carbohydrate..............3
2: Theoretical Methane Yield at Pune, India Temperature and Pressure..12
3: Feedstock Alkalinity Buffer........................................21
4: Expected Methane Yield per Milligram of COD Loaded, at Lab
Temperature and Pressure.......................................28
5: Potential Monthly Energy Generation Resulting from Anaerobic
Digestion of Biodiesel Sludge Using A.D. Karves Digester Design ...46
6: Estimation of BOD Loading in Biodiesel Sludge Feedstock.............50
LIST OF TABLES
1: Literature review, percent methane in biogas as seen in past
2: Literature review, percent of organic content degradation as
seen in past studies...................................13
3: Bench scale anaerobic digester parts list (Kocman, 2005)...17
4: Simulated wastewater feed components, solid portion...19
5: Lays nutrient mix (Lay, 1999).............................21
6: Abiotic experiments completed..............................30
7: Gap size tested in abiotic trials..........................32
Biogas is a combustible gas generated by the anaerobic
decomposition of organic matter. When burned, it produces a blue flame,
devoid of smoke and soot, and its intensity is easily controlled (Biogas
Support Programme, 2000). Biogas is an alternate source of energy, and can
be applied to various domestic and commercial applications. It is considered
a renewable energy source because it is generated from waste sources that
are essentially inexhaustible.
As shown in Figure 1, three separate microbial metabolic processes
(hydrolysis, acidogenesis, and methanogenesis) produce biogas having equal
quantities of methane (CH4) and carbon dioxide (CO2) via internal redox
reactions. This chemical reaction for a simple carbohydrate is shown in
Calculation 1. In general, biogas collected from anaerobic digesters generally
contains 65% methane and 35% carbon dioxide, rather than a composition
with equal parts as derived in Calculation 1. This is because C02 is soluble in
water (1.45 g/L at STP) while CH4 has a relatively low solubility (0.02 g/L at
STP). A portion of the C02 produced dissolves in the liquid compartment of
the digester while CH4 is released and is collected in the gas compartment.
Biogas BTU content is directly proportional to CH4 content. For
example, pure methane provides approximately 1000 BTUs per cubic foot at
STP while biogas consisting of 65% methane provides 650 BTUs per cubic
foot (Krich et al., 2005). A digester that produces biogas containing a very
high percent of methane is efficient because the gas is can be directly
utilized. Not having to remove CO2 also produces cost savings. This has
important implications relative to sustainability development.
Figure 1: Illustration of Anaerobic Metabolism (Grady et al., 1999)
QHiaOe + 6 H20 ->60O2 + 24 H+ + 24 e
COa + 8 I1*+* e>* CH4 + 2 H20 * ,-.4 ..
'&^06 + 4Hi0^6C02 + 24H+ + 24e-
* 3C02 + 24H+ + 24e^3CH* + 6H20
' v . ,.Gii*2gi6 ^icifc+^co* _ ;
...* * 4rr- *.*# ' '*4Sf-r " r\C. '
.............- ....________X..... ............-_________ . .....
Calculation 1: Biogas Composition Estimation for Simple Carbohydrate
1.2 Biogas Plant Technology
A biodigester consists of two components: a digester (or fermentation
tank) and a gas holder. Biodigester technology takes advantage of the energy
that is naturally present in household and animal waste. The digester is a
waterproof container with an inlet into which the fermentable mixture (organic
waste) is introduced in liquid slurry form. As these wastes break down, they
release a biogas that contains methane, carbon dioxide, and several other
trace gases such as hydrogen sulfide. The gas holder is normally an airproof
steel container that, by floating like a ball on the fermentation mix, cuts off
oxygen to the digester and collects the gas generated. A gas outlet with an
airtight valve allows the stored biogas to be released for use. On a household
scale, the biogas pipe can be connected to cooking stoves to provide a direct
fuel source. The digester is provided with an overflow pipe to lead displaced
digested sludge out into a drainage pit. Because the process is anaerobic,
more than 90% of most disease-causing pathogens are killed in the digestion
process (Frame, 2001). The sterilized effluent provides high quality organic
fertilizer that can be spread on fields to enrich the soil
Biogas power plants are a combination of a biodigester with associated
electricity generators such as gas turbines or gas engines. The raw biogas
contains water vapor that is separated out through condensation. The gas
portion is dried (refrigerated), compressed, and filtered to remove dust
particles and organosilicon compounds. Hydrogen sulfide (usually <2% of gas
volume) should be removed since it is toxic and releases are restricted by air
quality standards. The gas is then reheated and used to drive a gas turbine or
generator, which produces electrical energy. By-products of this process are
steam and hot water. The hot water can be recycled in a combined heat and
power cycle to increase the temperature of the digesters to optimal
conditions. A schematic of a typical biogas power plant is shown in Figure 2.
It is estimated that 16 million small-scale biogas plants are in operation
worldwide (REN21,2005). These inexpensive models are generally used on
a household scale to produce low cost energy for daily activities. Countries
like India, China, Germany, Denmark, Austria, Ireland, Scotland, Australia,
Arabia, Israel, and Japan have developed biogas markets and have
employed large, centralized biogas plants that provide energy to communities
and towns. Motivations for implementing biogas plants vary from country to
country. Israel, Australia, and Japan treat source-separated waste from
households and industry with a goal of reducing total volume of waste sent to
landfill. Middle Eastern countries use digester effluent to fertilize and condition
sandy soils. Biogas plants are common in highly populated Asian countries,
like China and India, because they provide low-cost energy for cooking and
lighting to poor, rural areas. Political pressures in Western Europe are
encouraging a shift to cleaner and renewable energy sources. Farmers
worldwide view biogas plants on their farms as a supplemental source of
income generated from readily available farm wastes.
Figure 2: Schematic of a Biogas Production Plant, Including Gas (Hansen, 1992)
In 2004, 2.8% of the energy consumed in the United States was
produced from biomass and waste sources. This equates to almost 3
quadrillion BTU and represents almost half of the total consumption of energy
from renewable sources (Energy Information Administration, 2004). Although
biogas (or renewable energy in general) is not a major source of energy yet,
biogas is viewed as a viable source of energy and will likely continue to grow
as movement toward renewable energy sources gains momentum in the
For biogas plant construction, important criteria are: a) the amount of
gas required for a specific use or uses, b) the amount of waste material
available for processing and, c) the food to microorganism ratio (F/M) in the
digester. These variables dictate the volume of the biodigester, the hydraulic
retention time (HRT) of material inside the digester, and ultimately, the
volume of feed that can be consumed each day. Regular monitoring of
digester is necessary to ensure that the digester it is functioning as efficiently
1.3 Biogas from Biodiesel Sludge
Biodiesel is produced through a process in which organically derived oils
(animal fats or vegetable oils) are combined with alcohol (ethanol or
methanol) in the presence of a catalyst to form ethyl or methyl ester
Biodiesel is an environmentally safe, low polluting fuel that can be used
in most diesel internal combustion and turbine engines.
Rocky Mountain Sustainable Enterprises (RMSE) plays an important
role in Colorados bio-fuel supply chain. One role of their business is the
collection of used cooking oil from Front Range restaurants. The processes
undertaken in their Berthoud plant remove food and waste solids and dewater
the oil, returning it to a 99%+ pure vegetable oil state. They then sell the oil to
their partners for the production of commercial grade biodiesel. As of May
2007, 38,200 gallons of wastewater and sludge is removed during the oil
purification process each month. This material is currently sent to compost.
RMSE has plans to scale up their processing capacity and predicts a
maximum wastewater/sludge volume of 192,000 gallons per month
RMSE is working toward becoming a zero-waste company and would
like to find a way to make their bio-sludge work for them. They are interested
in implementing a low-cost internal process that will allow them convert the
biodiesel sludge to biogas, which can then fuel various mechanical processes
at their plant.
The purpose of this study is to test the effectiveness of a novel
biodigester design developed by A. D. Karve and assess biogas yields from
biodiesel sludge using this design.
2. Literature Review
2.1 A.D. Karves Biodigester Design
A.D. Karve is a researcher affiliated with the Smt. Kashibai Navale
College of Engineering and is President of the Appropriate Rural Technology
Institute (ARTI), both operating in India (www.arti-india.org). ARTI is a
charitable trust, founded in 1996 by a group of scientists, technologists and
social workers. The biogas project, started in 2003, is one of over ten different
projects run by ARTI. Karves research is aimed at bringing sustainable,
clean, and safe energy production systems to rural Indian households. He
has designed several biogas production systems for use as energy
production systems rather than waste disposal systems, as they have
historically been used in India. In his 2005 paper titled, A New Compact
Biogas System Based on Sugary/Starchy Feedstock published in the Energy
for Sustainable Development journal (Karve, 2005), Karve published the
preliminary composition and yield results of his biogas trials.
Karves system, called the compact biogas plant, is essentially a
scaled-down version of the floating-dome-type conventional biogas plant and
is shown in Figure 3. It consists of two cylindrical drums that telescope into
one another. Both the drums are open at one end. The larger drum acts as
the digester while the smaller drum, inverted and inserted into the larger one,
acts as the gas-holder. The digester drum has a diameter of 85 cm, height of
85 cm, and a total internal volume of about 480 liters. It is provided with an
inlet pipe for letting the feedstock into the digester and an outlet near the top
of the liquid compartment for displacing effluent slurry. The biogas
accumulates in the gas-holder and lifts it up. The gasholder is provided with a
gas tap for removing the gas accumulated in it. Figure 3 shows the
experimental system used by Karve in his 2005 field trials with the discussed
features labeled. Karves upright cylinder digester design used cheap or free
inedible waste materials as microbial feedstock for generating biogas. Its
simple design is inexpensive to construct, is modular, and has a small
Figure 3: Photo of Biodigester Used in Karves 2005 Field Trials (Pitterle, 2005)
In his 2005 paper, A. D. Karve reported on the biogas composition
produced during two separate feed trials. Karves first feed trial used a
sugary/starchy feedstock consisting of 100% cereal flour. Based on the
calculation for a simple carbohydrate, 50% methane would be expected in the
biogas produced (see Calculation 1). His second feed trial used a mixed
starchy/oily feedstock that was 40% oilseed cake and 60% cereal flour. For
fats and vegetable oils (triglycerides), a typical CH4:C02 ratio for the biogas
produced is 70:30 as shown in the following equation.
C54H106O6 + 28H20 => 40CH4 + 17C02) (Krich et al 2005) (1)
Substituting this composition for the oilseed cake and the numbers
derived in Calculation 1 for the cereal flour, 58% methane would be expected
in the biogas produced from the mixed feedstock.
The biogas during Karves study was chemically analyzed to find the
proportion of C02 contained in it. This was done using a standard titration
method routinely used for C02 measurement in an undergraduate chemical
engineering laboratory (Karve, 2005). Presumably, the biogas was bubbled
through an indicator solution, causing the formation of carbonic acid,
indicated by a color change. He then likely added a basic solution, bringing
the indicator solution back to its original color, and quantifying the amount of
carbon dioxide that was introduced in the first step. However, since the
process used by Karve is not explained in his paper, this is only an
assumption. Karve reported that the biogas produced during each trial
contained only 5% carbon dioxide. He assumed that the remaining 95%
consists mainly of methane and some quantity of water vapor. Thus, an
exceptionally high purity (>90%) methane concentration in the biogas was
determined. Karve did not expect such highly efficient biogas to result. He
hypothesized that most of the C02 produced during digestion dissolves in the
liquid compartment and then diffuses out into the atmosphere through the 1
cm gap between the liquid and gas compartments. Karves publication
indicates that the yield and composition of the gas depends on the relative
populations of the starch-digesting and oil-digesting bacteria present in the
digester and that the system need to be balanced in order to continue
producing good-quality biogas every day (Karve, 2005).
Karves 2005 results also showed that feeding his compact biogas
plant 50 g of cereal flour per day yielded 25 L of biogas per day (Karve,
2005). He states that gas volume was measured based on tube rise and gave
no indication that the daily volume reported was adjusted for local
temperature and pressure once released from the gas compartment of the
biodigester. Since the pressure caused by the weight of the gas compartment
is unknown, in order to calculate theoretical yield, the assumption is made
here that the biogas volume is the volume under local temperature and
pressure. The average temperature in Pune, India is 298.45 K (range is 286
to 311 K) and the atmospheric pressure based on altitude is 91.56 kPa. At
95% methane content, 23.75 L of methane was produced per day.
Calculation 2 shows the theoretical uncompressed methane yield that would
result from chemical oxygen demand (COD) digestion based on conditions in
Pune, India where Karves field trials took place. The theoretical daily yield
based on Calculation 2 should be 22 liters of methane per day. This assumes
that the feed is 100% digestible and that 100% of the organic content is
degraded. In order to produce 23.75 L of methane per day, 54 g of COD
would need to be loaded each day. This is relatively close to Karves reported
feeding rate and the difference could be due to errors in the measurement of
feed amounts or gas volumes. If Karves design has the ability to consistently
biodegrade 100% of the organic material loaded into it each day, it would be
an extraordinarily efficient design.
Balanced reducing half-reactions relating COD to methane generation:
2 02 + 8H+ + 8^4H20 V.--' :
C02 + 8 * +8 CH4 + 2 HaO
2 Moles Q2 1 nw.
2 Moles 02
64 g 62: 64
64 g COD:
At Pune, India Tm =
298.45 K, Paw,.
64 g COD: 28.376 L CH4
64 g COD : 28.376 Umol CH4* 1 mol CH4
Calculation 2: Theoretical Methane Yield at Pune, India Temperature and Pressure
2.2 Results of Comparable Research
A review of available literature was completed in order to gain an
understanding of how Karves claims compare with the results of other
completed studies. The two areas reviewed were methane content in biogas
produced and the percent of organic content degraded using anaerobic
Table 1: Literature review, percent methane in biogas as seen in past studies
Feedstock ch4 Content " " T 7 f 1 T Source
Fresh Vegetable Waste & Wastewater sludge high feed frequency 67% Carucci, 2005
Separated liquid fraction from organic household waste 61% Held, 2002
Animal waste (manure) 60% Balsam, 2006
Fresh Vegetable Waste & Wastewater sludge low organic load 57% G. Carucci, 2005
Dehydrated potato flake waste 56% Kempter-Regel, 2000
Fresh Vegetable Waste & Wastewater sludge high organic load 37% G. Carucci, 2005
Table 2: Literature review, percent of organic content degradation as seen in past
..-Vi* Feedstock % of Organic Material Degraded (COD reduction) -.r s Source
French fry (potato) waste 70% Burke et al., 1997
Separated liquid fraction from organic household waste 68% Held, 2002
Municipal solid waste, 75% organic content 63% Kempter-Regel, 2000
Seafood processing plant wastewater, tuna condensate 60% Prasertsan, 1993
Anoxic lake sediments, natural attenuation 39% Molongoski, 1980
As seen in Tables 1 and 2 above, no prior biological process has resulted in
COD degradation above 90%, or a methane percentage in biogas greater than 70%.
However, Karves open system design may facilitate carbon dioxide removal from
biogas, which will be evaluated in this study.
With this background in mind, the objective of this study consists of three
a. To prove or disprove that Karves design is in fact capable of
producing biogas that contains > 90% methane, using abiotic
b. To test the ability of Karves biodigester design to convert loaded
organic material, expressed as COD, to methane, in a bioactive
c. To test the biodegradability of RMSEs biodiesel sludge in terms of
BTUs produced per unit of volume in a bioactive system. These
results will be scaled up to the plant level in order to determine if
biogas is cost effective and should be integrated into their internal
3.1 Experimental Design
The experimental design used for this study consists of two distinct methods.
Abiotic test of Karves digester design
o Carbon dioxide loss
o Methane loss
Biotic tests using three different feedstock compositions
o Simulated wastewater
o Biodiesel sludge
4. Materials and Methods
4.1 Digester Design and Construction
Two small-scale anaerobic digesters were constructed according to
Karves design principles for use in this research. The design consists of two
cylindrical PVC pipes that telescope into one another. Both the pipes are
open at one end. The larger diameter pipe acts as the fermenter/digester and
the smaller diameter pipe, inverted and inserted into the larger one, acts as
the biogas collector. The liquid compartment has a diameter of 8.9 cm. (3.5
in.), a height of 71 cm. (28 in.), and a total volume of 4.4 liters. The system
has a total gas storage capacity of 2.8 liters. The biogas accumulates in the
gas-holder and causes the inner tube to rise. The gas-holder is provided with
a gas tap for removing the gas accumulated in it. Figure 4 shows a schematic
and Figure 5 shows a photograph of the constructed bench scale system
used in the biotic portion of this study.
The parts were bought at the hardware store and constructed. The
PVC parts were set in place using glue and silicon sealant and were bolted to
a large base board. The threaded hose barbs and valves were added to the
system last. The holes were drilled and a tap and die were used to thread the
PVC for fitting the attachments. The fittings were leak tested with soapy
water. The complete parts list used in construction is listed in Table 3.
Denver Metro Wastewater Districts Central Treatment Plant in
northeast Denver provided a 500 mL seed sludge sample from their
anaerobic digester. This represented 12% of the total digester liquid
The liquid compartment was completely filled with water and the gas
tube was inserted into it. The air was evacuated from the tube by opening the
valve and pressing the tube down until it was submerged, in order to create
an anaerobic environment.
A 500 ml volume of anaerobic sludge was mixed with 500 ml of room
temperature tap water (25 C) and loaded the mixture into the digester
through the base inlet pipe, allowing displaced clean water to escape through
the outlet pipe. The first feeding of the digester occurred two days later, in
order to first allow the microbial population to settle and take hold.
Figure 4: Schematic of Karves Biodigester
Figure 5: Karve Digester Used in Biotic Experiments (Kocman, 2005)
Table 3: Bench scale anaerobic digester parts list (Kocman, 2005)
Floatina aas comoartment: Liauid compartment/main Inlet and outlets: body:
3 schedule 40 black PVC pipe PVC stand % inch polyethylene tubing
PVC threaded connector and stop valve 4 schedule 40 PVC black pipe Hose barb with threaded connector
V2 inch threaded brass connector and reducer 4 schedule 40 PVC connector Slip clamp
T-valve with thread and compression fitting Black rubber connector with metal clamps
1/4 inch polyethylene tube
Karves simple design does not include mechanical or electrical
devices for stirring or heating the digester contents. Some mixing occurs
when slurry is piped into the digester and effluent is displaced and moving the
gas compartment around in the liquid compartment provides some temporary
mixing. The digester operates at room or ambient temperature so its
efficiency is affected by changing seasons and local climates. During this
research, the laboratory temperature remained stable at 22.5 C (72.5 F).
Some indigestible scum had to be skimmed off the top of the liquid
compartment periodically as it started to block the gas outlet. This required
removing the gas compartment for a brief period of time.
4.2.1 Sucrose (Table Sugar)
The first type of feedstock tested for biodegradability and biogas
production was a mixture of table sugar (sucrose, C12H22O11), the nutrient mix
shown in Table 6, and tap water. To jump-start biogas production, 140 grams
of sucrose was initially added two days after the seed sludge was added to
the biodigester. Feedstock consisting of 2 grams of sucrose, 2.0 ml of nutrient
solution, and tap water was added each day for 13 Days. The CODin for this
feed was 5,690 mg/L. Feeding a volume of 200 ml each day resulted in a
HRT of 22 days. The daily organic loading rate was 1,138 mg/day.
4.2.2 Simulated Wastewater
A document used for an environmental engineering competition at the
University of California Davis contained a recipe for a simulated wastewater
solution (University of California Davis ASCE Student Chapter, 1998). It
was determined that this would serve as a more complex, multi-source,
organic feedstock for use in the second feed trial. The recipe for this
feedstock is shown in Table 4. The ingredient amounts listed are intended to
be mixed with 10 gallons of tap water. All of the ingredients listed were mixed
thoroughly in a blender. This listed ingredients served as the solids portion
of the wastewater feedstock. When mixing batches of feed, a proportionate
amount of the solids was combined with tap water to create the strength of
The alkalinity for the simulated wastewater feed was found to be 0.0
g/L as CaCo3l a negligible amount. To buffer acidic conditions, the baking
soda buffer shown in Calculation 3 was added to the wastewater feedstock.
The nutrient mix was not included in order to determine if using a more
complex feedstock sufficiently supplied all the metabolic requirements of the
The 7-day feed change method shown in Figure 6 was used to
transition the digester from the sucrose to wastewater feedstock. The daily
feed volume varied between 300 and 400 mL/day for this feed trial, which
resulted in an HRT range of 14.67 to 11 days. The CODjn for this feed was in
the range of 2,236 to 5,871 mg/day.
Table 4: Simulated wastewater feed components, solid portion (UC-Davis, 1998)
whole leaves 1/2 lb.
grass clippings 1/2 lb.
Silt (fine) 1 oz.
tomato sauce 8 oz.
pasta (cooked w/pasta water) 8 oz.
crushed garlic 2 tbsp.
coffee grounds (used) 1 cup
coffee filter (used) 1
Vinegar 1/2 cup
Vegetable oil 1/4 cup
lemon dish soap 2 oz.
Red table wine 2 cups
Tofu 1/2 lb.
banana peel 1
crushed banana 1
beets, shredded 1 cup
carrot, shredded 1 cup
4.2.3 Biodiesel Sludge
The final phase of biotic testing measured the biogas yields produced
from biodiesel sludge. Several samples were provided by RMSE, each of
which included oil, water, and a high concentration of food solids removed
from used cooking oil collected from restaurants. A low sample was
retrieved from the lower depths of the tank, and a high sample was skimmed
off the top. The high sample contained more oil while the low sample
contained more heavy solid particles. Both sample types are comparable to
the 60% starchy, 40% oily feedstock Karve tested in his 2005 field trials.
The feedstock created contained 10-12% sludge by volume (variation
between batches), mixed with 90% water by volume. The standard baking
soda concentration and 2.0 mLA of Lays nutrient solution provided the
alkalinity buffer and nutritional supplements. The digester was transitioned to
this new feed using a 7 day ramp-up method as shown in Figure 6.
A variety of feed volumes were tested, varying from 250 to 400
mL/day. This change in volume was completed gradually and was aimed at
testing the effect of a longer and shorter HRT. The range of HRTs tested was
between 17.6 to 11.0 days. A variety of feed strengths was also tested in
order to see the effect on gas production. The CODin range for this feed was
between 6,438 and 9,430 mg/day.
4.2.4 Nutrient Mix
A nutrient mix was added to each feed type in order to supply a variety
of necessary elements that the feed itself could not provide. The nutrient mix
used is shown in Table 5 (Lay, 1999). The amounts listed in the table were
mixed with 2 liters of de-ionized water for use over the course of the research.
Table 5: Lays mix for 2L nutrient solution (Lay, 1999)
Nutrient Name Chemical Formula Amount
Ammonium Bicarbonate NH4HCOa 200 g
Potassium Phosphate kh?po4 100 g
Magnesium Sulfate MQSO4 10 g
Sodium Chloride NaCI 1 g
Sodium Molybdate Na2Mo04 1 g
Calcium Chloride CaCI2 1 g
Manganese Sulfate MnS04 1.5 q
Iron (II) Chloride FeCI? 0.278 g
Nickelous Chloride NiCI2 1 mg
Cobalt Chloride O0OI2 1 mg
Yeast Extract 100 mg
4.2.5 Alkalinity Buffer
A buffer was added to the daily feed solution in order to increase the
alkalinity and maintain a neutral pH. A common and inexpensive buffer is
pure baking soda (NaHC03). A normal, healthy range for alkalinity is between
1 and 5 g/L as CaC03 (Seereeram, 2004). An alkalinity goal of 4 g/L as
CaC03 was used for the duration of this research. The amount of baking soda
(NaHC03) to be added in order to maintain this alkalinity is shown in
g/L NaHC03 needed =
6.7 g/L NaHCOs eg.
..... # .
Calculation 3: Feedstock Alkalinity Buffer
4.3 Feed Procedure and Frequency
The steps taken to develop a variety of feedstock solutions used over
the course of this research project are outlined in the section below.
Determine optimal HRT
Divide the liquid volume of the digester by the HRT in order to
determine the daily feed volume.
Determine the level of total solids (TS) solid to liquid ratio of the
feed solution (10-12% solids was used in this research).
Using a container with a known volume, mix a feedstock with 10%
solids and fill the remaining volume with tap water (90%).
If necessary, use a blender to liquefy the feed solution in order to
break up larger solid particles.
Run a COD analysis on the feed solution to determine the mass
organic daily loading rate (CODin).
Mix a large batch of feedstock for use over several weeks in order
to decrease variation in the feed composition day to day.
A well-mixed and room temperature feed solution of a pre-determined
volume was fed into the bottom inlet tube each day using a funnel. The gas
compartment value was left closed while feeding and the displaced digester
effluent was collected through the outlet tube for analysis. The type
(composition) of feed, feed batch number, and volumetric amount fed were
recorded after each feeding.
Because microbial populations grow and change quickly, it is possible
for them to adapt to one type of feed through natural selection. An abrupt
change in feed can halt gas production completely and may even require re-
seeding and re-stabilizing the digester. When changing from one feed type to
another, a 7-day ramp-up period was used. This method gradually but
simultaneously reduced the solid amount of the old feed and increased the
solid amount of the new feed. The total feed volume stayed the same, but the
percent of that volume that was made up of the old and new feedstock
changed each day. An example of the calculation is shown in Figure 6.
Anaerobic Digester Feed Change Methodology
Figure 6: Feed Composition Change Methodology
5. Analytical Methods
The stability of the biodigester was monitored by measuring the
The optimal pH for the biodigester is in the range of 6.6 to 7.6. The pH
was monitored daily using a Hach One digital pH meter (model 43800) with a
detection range of 1 to 14. The instrument was calibrated weekly using Hach
standard buffer solution.
The optimal alkalinity for the digester is in the range of 1.0 to 5.0 g/L as
CaC03. The alkalinity of the feedstock was measured whenever a new batch
was mixed. The alkalinity of the digester effluent was measured weekly or
whenever the feedstock was changed. Alkalinity was measured with a Hach
One Digital Titrator, alkalinity reagent cartridges, and a digital pH meter
according to Hach standard method #8203. The pH meter was calibrated
weekly using a standard buffer solution.
5.3 Volatile Fatty Acids (VFAs)
The optimal concentration of VFAs in the digester effluent is in the
range of 500 to 3500 mg/L. VFA concentrations were measured monthly
using an Electromantle MA Solid State Stirrer and a Hach One Digital Titrator
according to Hach standard methods 8291 and 8218. There is no detection
limit for this procedure.
5.4 Chemical Oxygen Demand (COD)
The optimal COD concentration in the digester is in the range of 4400
to 8800 mg/L. The COD of the feed and the COD of the effluent were
monitored at least weekly as well as whenever a feedstock change occurred.
Instruments used included a Hach COD reactor (model 45600), a Hach D/R
2000 spectrophotometer, a Hach D/R 2010 spectrophotometer, and High-
range COD digester vials. Proper dilution of samples was necessary to
ensure that concentrations would fall within the detection limits of 0 to 1650
mg/L. Hach standard method #8000 was used. Due to variation between
spectrophotometer readings, the average of four readings was taken for each
sample in order to determine the final recorded value. The COD concentration
for a standard with a known value (sucrose) was analyzed to determine error.
Error was found to be +/-1.71% (mg/L).
5.5 Biogas Volume
Biogas volume was measured daily using a level and a tape measure.
The volume of gas in the collector was calculated using the equation below.
Tube rise (h) pi r2 gas volume multiplier (2.275) (2)
This measurement does not have any limits. However, error due to
visual approximation of tube rise measurements is inherent. The estimated
error per measurement is +/- 0.025 liters of biogas.
5.6 Biogas Composition
The chemical makeup of the biogas produced by the digester was
determined using the following methods.
5.6.1 Carbon Dioxide (CO2)
The carbon dioxide content of the biogas produced by the digester was
measured bi-weekly using a Draeger pump and 5% CO2 tubes (P/N
CH20301). The instruments have detection limits of 5 to 60% of gas volume.
Inherent error in the measurements is due to the visual approximation of the
colorimetric reaction in the tubes since hash-marks on the tube are every 5
units of volume. The maximum error possible is +/- 5% per reading.
5.6.2 Hydrogen Sulfide (H2S)
The hydrogen sulfide content of the biogas produced by the digester
was measured bi-weekly using a Draeger pump and 1/d H2S tubes (P/N
8101831). The instruments have detection limits of 1 to 2000 ppm. Readings
from the tube in ppm were converted to percent of gas volume. Inherent error
in the measurements is due to the visual approximation of the colorimetric
reaction in the tubes since hash-marks on the tube are every 200 ppm. The
maximum error in taking tube readings is 200 ppm (or 0.02%) per reading
5.6.3 Methane (CH4)
The methane content of the biogas produced by the digester was
measured bi-weekly according to the equation below since Draeger does not
make tubes that measure methane concentrations directly.
Methane % = 100% C02 % + H2S % (3)
Due to detection limits and error for the tubes discussed in sections
5.6.1 and 5.6.2, the detection limit for this measurement is between 5 and
95% and the potential measurement errors is +/- 5.02%.
5.7 Methane Production from COD Measurements
As mentioned previously, the research laboratory does not have the
equipment needed to determine concentrations of specific types of volatile
solids, only total volatile solids. Therefore, for the purposes of this research,
the expected biogas yield was calculated based on the net decrease between
the mass of COD that was loaded into the digester via the feedstock (CODin,
mg), and the mass of COD that remained in the effluent at the end of each
feed trial (CODout, mg). Error exists in this calculation since some percentage
of COD reduced is used to build new microbial biomass. Bacterial biomass in
the effluent also makes up some percentage of the CODout concentration.
However, since both the CODinand CODout measurements include a similar
margin of error (theoretically), net error is minimized and should not have a
significant effect on the calculation of methane yields.
The theoretic methane yield calculated for Pune, India (Calculation 2)
was adjusted to reflect the conditions in the Denver, CO (USA) laboratory
used for this research. The laboratory temperature stayed constant at 22.5 C
(295.5 K). Atmospheric pressure based on Denvers altitude is 83.426 kPa.
As shown in Calculation 4, theoretically, methane is produced at a rate of
0.46 cm3/mg COD reduced. COD reduction was calculated by taking the total
amount of COD loaded during the feed trial and subtracting the total
concentration of COD remaining one day following the end of the feed trial. A
conservative methane concentration of 50% of biogas volume was used and
theoretical methane volume was doubled to get the theoretical biogas
At Denver, CO (USA) T**. = 295.5 K, = 83.426 kPa):
/ 64 g COD: itnote CH4* 29.44 L/mol
64 mg COD: 0.02944 L CH4
1 mg COD: 0.00046 L CH4
: 1 mg COD: 0.46 cm3 CH4
, Feed vofcJnjie ^ x COD
Effluent volume (L) x COD (mg/L) = CODout (mg)
CODjn (roKJfcCODreducad (mg)
lyy?.-%* ^ %. ;
CODraduced (n^)_x.0.45 cm3/mg COD=CH*<*t (cm3)
CH4 egi (cm3) x 0.001 CH4, (L)
.. P : - ' ............ r
Calculation 4: Expected Methane Yield per Milligram of COD Loaded, at Lab
Temperature and Pressure
5.8 Calculation of Actual Biogas and Methane Production
The gas compartment of the digester had an internal diameter of 7.62
cm. (3 in.). The surface area of the gas compartment was 45.6 cm2. The
volume of compressed biogas produced was measured by multiplying this
surface area by the height of the gas compartment tube rise.
Because the gas in the collector is under pressure that is greater than
Denvers atmospheric pressure (due to the compartments mass), a multiplier
was calculated in order to convert compressed gas volumes to volumes under
local pressure. This multiplier was obtained through a volume displacement
experiment. An empty tedlar bag and a weight were submerged into a water
bath containing a known volume of water in order to find the volume of the
empty bag and the weight. A volume of gas was released from the gas
compartment into the tedlar bag. The tube-rise before and after the release
was recorded in order to calculate the compressed gas volume. The filled bag
was then submerged in the water bath and the volume of the uncompressed
gas was recorded as the displaced water volume less the volume of the
empty bag and the weight. The ratio of the uncompressed volume to the
compressed volume was calculated as 2.275. These results were later
confirmed through the use of an Ashcroft 30 psi gage. The pressure inside
the gas compartment was measured as 29 psi, which results in a ratio of
uncompressed to compressed volume of 2.314. Because the percent
difference between these ratios was so small, yields were not recalculated to
reflect the gauged result. This multiplier was used to convert from the
compressed gas volume measured by collector tube rise to uncompressed
gas volume shown in Calculation 4.
5.9 Abiotic Testing Methods
A second Karve digester of similar scale was constructed for use in
abiotic experiments. As shown in Table 6, two types of abiotic experiments
were completed, carbon dioxide loss and methane loss.
Carbon dioxide release experiments were completed using three
different diameter liquid compartments in order to create different ratios
between the gas compartment and liquid compartment surface areas. This
variation was aimed at testing whether a larger gap between the two
compartments resulted in a lower concentration of CO2 in the biogas. The
same 8.9 cm. (3.5 in.) diameter gas compartment was used for all abiotic
To initiate each C02 test, a decompressor was used to remove all of
the air from the gas compartment. Using visual approximation based on tube
rise, the gas compartment volume was filled 50% with either commercial
grade natural gas (-98% methane), or 99.97% pure compressed methane
gas (balance Nitrogen). The other 50% of the tube was filled with 99.98%
compressed bone dry carbon dioxide (balance Nitrogen). A sample was taken
immediately and a Draeger tube analysis was used to establish the baseline
CO2 concentration. To measure the CO2 loss, a gas sample was taken each
day and the remaining CO2 content was measured using a Draeger tube.
For the final CO2 loss experiment (test #6), ten pounds of additional
weight was anchored to the gas compartment, tripling the weight of the gas
compartment (from 5 lbs. to 15 lbs.). The goal of this experimental variation
was to determining if increased pressure caused an increase in the rate of
carbon dioxide loss from the gas compartment to the liquid compartment.
Table 6: Abiotic experiments completed
# Test Type Duration Liquid vOivipai iiiiQnl Base Diameter, cm Gas Mix Used ^ '
1 C02 loss 5 days 10.2 36.0% C02/ 64.0% CH4J
mm C02 loss 7 days 36.8 41.5% C02/ 58.5% CH4
C02 loss 7 days 56.5 33.5% C02/ 66.5% CH4
mm CH4 loss 5 days 36.8 100% Natural Gas, -98% CH4 '
CH4 loss 5 days 56.5 100% Natural Gas, -98% CH4"
H C02 loss w/ weight added 3 days 36.8 25% C02/75% CH4
Purity of gases uses was on the order of 99.9% according to the specification on the
From UCD Technology Laboratory gas line, commercial grade natural gas is 98% methane
on average (Tetlow-Smith, 1995)
During the CH4 loss experiments, the height of the gas compartment
rise was measured using a level and tape measure. Some error in volume
measurements exists due to the fact that the gas compartment bobs up and
down in the liquid compartment and it is somewhat difficult to keep steady
while measuring. Statistics were recorded daily to calculate the volume of CH4
that was leaving the gas compartment over each twenty-four hour period.
Results are in terms of percent volume change per day.
6.1 Abiotic Results
Because many variables exist within the biotic experiment design,
abiotic tests were completed in order to eliminate these living-system
variables and to test whether it was possible to get to 5% carbon dioxide after
beginning with a typical biogas mix (50% methane, 50% carbon dioxide).
The methods used to conduct both the carbon dioxide loss and
methane loss experiments have been discussed in a previous section and are
summarized in Table 7. The results of the five trials are shown in this section.
A methane release experiment was not completed for the 10.2 cm base as
results using the two larger bases made further experimentation unnecessary.
The gap sizes resulting from the digester bases of three different
diameters are shown in Table 7.
Table 7: Gap size tested in abiotic trials
Base ID. , Gas compartment, outside i diameter, cm Liquid compartment, inside i diameter, cm I"? I.W Mill. II ii in I Difference, cm tattmep b-wpiaiiiigii Gap, cm
Large Base 8.9 56.5 47.6 23.8
Medium Base 8.9 36.8 27.9 13.95
Small Base 8.9 10.2 1.3 0.65
6.1.1 Carbon Dioxide Loss Experiments
Over a test period of 5 days, the C02 concentration on the small base
dropped from a starting concentration of 36.0% to 23.0% C02 for a total
decrease of 13.0%. The concentration dropped sharply between Day 1 and
Day 2 and then evened off, losing 1 -2% per day over the last 4 days. These
results are shown in Figure 7.
Abiotic CO^ Loss Small Base, 5 Days
%OG2 % Decrease'day
Figure 7 : Carbon Dioxide Loss Using Small Gap
Over a test period of 7 days, the CO2 concentration in the medium
base system dropped from a starting concentration of 33.5% CO2 to 8.0%
CO2 for a total decrease of 25.5%. Again, the concentration dropped sharply
between Day 1 and Day 2 and then evened off, losing 1-3% per day over the
last 4 days. These results are shown in Figure 8.
Over a test period of 7 days, the CO2 concentration in the large base
system dropped from a starting concentration of 41.5% CO2 to 6.5% C02 for
a total decrease of 35.0%. The same trend was apparent with the largest
C02 loss occurring between Day 1 and Day 2 and then gradually evening off,
with a 3-5% loss per day over the last 5 days. These results are shown in
Based on these three figures, it is evident that a digester base with a
larger gap between the liquid and gas compartments allows more CO2 to
dissolve and then escape into the open atmosphere. This supports Karves
theory that CO2 loss occurs using his biodigester design. However, the
lowest CO2 concentration obtained was 6.5% (on the large base) and it took 7
full days to get to that concentration from a starting concentration of 41.5%.
This is approximately the percentage of CO2 expected in the biogas produced
from his 60% starchy/40% oily feedstock. Since Karve claims that his biogas
contained only 5% carbon dioxide on a daily basis, it is determined that his
results were not able to be replicated using this abiotic experiment. It does not
appear to be possible to reduce the C02 concentration by 37% in a single
day, regardless of the gap size used.
Because each of the three carbon dioxide loss experiments did not
have the same C02 starting concentration and were conducted over different
periods of time, the data was normalized in order to show a true comparison
among the results. In Figure 10, C02 loss is shown as a percentage of the
initial concentration and only data over the first 5 days of the medium and
large base experiments is included.
As shown in Figure 10, a larger digester base and gap does in fact
allow for a higher percentage of total CO2 to be removed from the biogas.
However, the percent change between the medium base and large base
system is minimal (0.9%). The optimal gap size is likely to exist between 1.0
cm. and 14.0 cm. The gains in biogas efficiency obtained when a larger arid
larger base size is used likely do not balance with the increase in costs that
comes with decreased mobility, footprint size, maintenance, and material
costs. Furthermore, while larger gap sizes increase gas loss pathways, they
also pose a risk of oxygen entering the system from the atmosphere, making
Abiotic CQ> Loss, Normalized over 5 days
Small Base MadiumBase Large Base
Digester Base Diameter
Figure 10: Comparison of Carbon Dioxide Loss Using Different Base Sizes
6.1.2 High Pressure Carbon Dioxide Loss Experiment
When the C02 loss experiment was repeated on the medium base with
ten pounds of weight added to the gas compartment, the carbon dioxide
content dropped quickly from 25% of volume to 5% of volume within a 24
hour period. This represents a 80% reduction of the starting C02 volume. The
amount of carbon dioxide lost from the system in one day during this
experiment can be compared to the 63% loss over five days as shown in
Figure 10 (medium base). Through this comparison, it is apparent that high
gas compartment pressure results in lower carbon dioxide content in the
biogas. In the weighted experiment, the volume of gas in the gas
compartment was depleted to an immeasurable amount within three days.
Because of the rapid loss of gas, only two daily C02 measurements could be
taken. This large decrease in overall gas volume indicates that methane, as
well as carbon dioxide, was lost from the gas compartment.
6.1.3 Methane Loss Experiments
In order to verify that methane was not being lost through the gap along
with the carbon dioxide, the experiment was repeated with only methane
filling the gas compartment. The methane loss was measured in terms of
volume rather than by percent loss because the composition of the gas in the
tube was approximately 100% methane over the course of the experiment.
As shown in Figure 11, over a test period of 5 days, the CH4 gas volume
on the medium base dropped from 2.394 liters to 2.325 liters on Day 3 and
then remained steady over the next two days. This volume decrease is
equivalent to a 2.88% decrease in volume. Because natural gas with a
methane concentration of -98% was used for this experiment, the loss in
volume can mostly be attributed to the non-methane components of the gas
that are more soluble in water (carbon dioxide, nitrogen, hydrogen sulfide). It
was determined that significant methane loss was not occurring.
The experiment was repeated on the large base to verify the results.
As shown in Figure 12, the volume of CH4 dropped from 2.235 liters to 2.189
liters on Day 5. This volume decrease is equivalent to a 2.04% decrease in
volume. Pure methane (99.97%) was used for this experiment so the volume
decrease can not be fully accounted for. It may be due to errors in the
measurement of the tube rise as discussed in section 5.9. Regardless, the
volume decrease is small enough to determine that methane loss from the
system was not occurring.
6.2 Biotic Results
During Karves 2005 field trials, a compact biogas plant with a 1 cm. gap
between the gas compartment and the liquid compartment was used. He
claims that the 95% methane content found in his biogas is at least partially
due to carbon dioxide dissolving into the liquid and then escaping to the
atmosphere through this gap. The anaerobic digester used for the biotic
portion of this research had a gap of 1.5 cm. A comparable gap size,
combined with similar feedstock used, allows for the gas composition and
yield results from this research to be compared to Karves results.
Abiotic CH, Loss Mecfim Base, 5 Days
Figure 11: Methane Loss Results Using Medium Gap
CH4 Loss % per Day
Abiotic Loss-Large Base, 5 Days
\folume CH4 (L) % Decnease/day
Figure 12: Methane Loss Results Using Large Gap
6.2.1 Biotic Tests from Sucrose
Biogas production stopped on Day 14 of the feed trial which is why
only 13 days of biogas data are available. A total mass of 93,316 mg of
organic material was loaded during over the first 13 days of this feed trial. On
Day 14, a CODout measurement was taken for the displaced digester effluent.
This value was averaged with the other readings taken over the 13 day period
for an average CODout of 65,780 mg. Therefore, the COD reduction during the
feed trial was 27,536 mg. This represents a 30% reduction, which is
conservative since microbial biomass in the effluent is contributing to the
CODout measurement. Using a conversion factor of 0.46 mL CH4/mg COD
reduced (see Calculation 4), 12.67 L of methane is the expected yield based
on COD reduction.
During this trial, 13.65 L of biogas was produced. Methane content in
the biogas was 46.67% on average, resulting in 6.37 L of methane. This was
57.3% of the expected methane yield. Actual and expected methane yields
are compared in Figure 13.
Although the simple carbohydrate feedstock and digester design used
in this trial were comparable to those used in Karves first trial (100% cereal
flour), the results were not comparable. The biogas produced had a methane
content of 46.67%, which is less than half of Karves biogas. The organic
content was degraded by 30%, which is less than one third of Karves results.
Up until Day 14 when biogas production stopped completely, the pH of
the digester effluent had been within the optimal range (6.6 7.6) but had
been decreasing gradually. Regular digester feeding and monitoring
continued until Day 23 when the effluent pH reached a low point of 5.89.
When volatile acid production occurs more rapidly than volatile acid
consumption (methane production), an upset condition occurs in the digester.
The digester becomes acidic or sour. Because methane-forming bacteria are
very sensitive to acidic conditions, methane production decreases as volatile
acid concentration increases. A measurement of volatile fatty acid (VFA)
concentrations taken on Day 8 of the feed trial showed levels that were 5
times higher than optimal levels (500-3,500 mg/L). Methane production
usually terminates when the pH drops below 6.0, which was the case here.
The alkalinity of the feedstock was measured and the feed was found
to only contribute 0.26 g/L of alkalinity. Since optimal levels are 1.0 5.0 g/L,
an alkalinity buffer had to be added to neutralize the acids that were causing
the pH to drop. The buffer amount was determined using Calculation 3 and
6.72 g/L NaHC03 was added to each feed batch. After feeding the buffered
feed solution for an additional 18 days, pH levels increased to within the
optimal range but biogas production did restart. The pH trend discussed is
shown in Figure 14.
Methane Production ResiJts- Sucrose Feedstock
from OOD reduction
Figure 13: Methane Production from Sucrose Feedstock
pH Huctu^ion Sucrose Feed Trial
Feed Trial Day#
|-EfeilypH without Atkatirity after ^-DaHypHwithAlkaiirity Buffer
Figure 14: pH Fluctuation during Sucrose Feed Trial with and without Alkalinity Buffer
It was hypothesized that using table sugar as the sole source of
organic material was resulting in low or no biogas production because the
feed composition was too simple. The results of the second biotic feed trial
that utilized a more complex feed composition are discussed in the next
6.2.2 Biotic Tests with Simulated Wastewater
A 7-day ramp-up period from the sugar feedstock to the simulated
wastewater feedstock was used as not to further upset the surviving microbe
population. However, after a month of feeding and monitoring, biogas was still
not being produced. A small amount of the initial seed obtained from Metro
Water remained and it was added to the digester. The following day, biogas
had started to accumulate again. The results of the wastewater feed trial
include only the 32 consecutive days when biogas production occurred.
A total mass of 162,732 mg of organic material was loaded during this
32 day feed trial. The CODout measurement taken on Day 33 was averaged
with the other readings taken over the 32 day period for an average CODout of
36,383 mg. The COD reduction during the feed trial was 126,350 mg. This
represents a 78% reduction. Using Calculation 4, 58.12 liters of methane is
the expected yield based on COD reduction
During this trial, 8.14 L of biogas was produced with an average
methane content of 70.75%, resulting in 5.76 L of methane produced. This
was 11.3% of the expected methane yield of 58.12 L. The expected and
actual methane yields over the length of the feed trial are shown in Figure 15.
The methane content resulting from this feed trial was significantly
higher than the content seen in the single-source trial. However, since this
feedstock included carbohydrates, oils, and proteins, it would be expected
that the methane content in the biogas produced would be higher than 50%.
The 70% methane content is also in line with what has been demonstrated by
typical anaerobic digesters in past applications. Because the methane content
in the biogas produced was consistently 21-27% below the content seen by
A.D. Karve, it was determined that Karves gas composition results could not
be replicated using this more complex feedstock.
Methane Production Results Simulated Whstewater
9.00 Expected from OCD reduction )
7.00 - Aotud Methane
4.00 Effluert pH
Feed Trial Day#
Figure 15: Methane Production from Simulated Wastewater Feedstock with pH Trend
6.2.3 Biotic Tests with Biodiesel Sludge
A total mass of 200,578 mg of organic material was loaded through the
biodiesel sludge feedstock during this 29 day trial. The mass loading of
organic content was increased from the previous two trials order to test
whether a relatively high loading rate would produce results closer to those
seen in Karves trials. After each daily feeding, the gas compartment was
spiraled around in the liquid compartment in order to mix the contents over
the full length of the column.
The CODout measurement taken on the last day of the feed trial was
averaged with the other readings taken over the 29 day period for an average
CODout of 111,021 mg. Therefore, the COD reduction during the feed trial was
89,557 mg. This represents a 45% reduction. Using a conversion factor of
0.46 ml_ CH4 gas per mg COD reduced (see Calculation 4), 41.20 liters of
methane is the expected yield based on COD reduction.
During this trial, 9.61 L of biogas was produced with an average
methane content of 53.79%, resulting in 5.17 L of methane. This was 12.5%
of the expected methane yield of 41.20 L. A comparison of actual and
expected methane yields over the length of the biodiesel sludge feed trial is
shown in Figure 16.
The relationship between organic loading (as COD) and biogas
production is shown in Figure 17. The two data sets are paired so that the
COD amount loaded on Day 3 is paired with the biogas production results for
Day 4, for example. This assumes that COD is converted to biogas within a
Methane Production Results-BkxEesel Sludge
14.CC 13.00 12.00 11.00 10.00 9.00 8.00 _ Methane Expected from OCD reduction L
7.00 - Actual Methane
4.00 > Effluent pH
Feed Trial Day#
Figure 16: Methane Production from Biodiesel Sludge Feedstock with pH Trend
Daily Biogas Yield Produced from Biocfiesel Sludge
Figure 17: Correlation between Organic Content Loaded and Biogas Production
6.3 Plant Scale Energy Yield Using Karves Design
The results of this research can be scaled up to reflect the expected
monthly energy yield that can be obtained by implementing a similar process
at the RMSE production plant. This scale-up, which includes plant production
statistics obtained from RMSE employees, is shown in Calculation 5. By
initializing this research project, RMSE was interested in determining if the
energy that can be captured by the anaerobic degradation of their processing
waste would be sufficient to fuel a major component of their internal
manufacturing process, such as a boiler.
RMSE waste volume
% volumefiat removablesolids (estimated)
If diluted to 10%aludge
V* -W \,f ... j
* . .> .2M
SfJii ^ * .1, *# ; -*. P .1 10,143 gal/mo
t4 \ : y' ', V Or 38,395 Umo.
tÂ§|f X M V : 383,951 L/mo.
Average COD cdhtent df .diluteb f eed, 1rom COD analysis in UCDlab
"f.-r/ *44,560 mg/L
u, ' .
Total COD loading available
'& ;7 ' t -.'6 H i S' "
, . rt 4 17,108,679,503 tng
Percent COD recluctiohexpeded, based on observed UCD lab tests
'' M ''U : ' - 45% ? '*
Total COD reduction ^ ^ )
Cm3 CH4 produced per mg COD reduced, based on UCD lab results
. * * -. ^ a -m, y
Total Methane produced S' y
..-s.~x.v X 'fix
T" Sy */ Xr^^X? ^ . Or 15,566ft3/mo.
' S - , 5 - , ' Â£ \i '
Energy value expected if system performs as in UCD tests
< '*r '-fs ' ^ 'S
Calculation 5: Potential Monthly Energy Generation Resulting from Anaerobic
Digestion of Biodiesel Sludge Using A.D. Karves Digester Design
7.1 Biogas Composition
Through the use of digester designs and feed types that were
comparable to those used in Karves trials, biogas containing > 90% methane
could not be obtained. A graph comparing the methane content from samples
taken during the biotic portion of this study to Karves results is shown in
Figure 18. The error bars indicate standard deviation for the gas composition
readings taken. The highest methane content obtained during this study was
73.9%, which is 21.1% lower than Karves claims. Abiotic tests proved that
carbon dioxide is lost from the system as Karve predicted. However, the rate
of CO2 loss seen in experiments using an un-weighted gas compartment was
not rapid enough to consistently provide the high efficiency biogas that he
reported on a daily basis. When weight was added to the gas compartment,
the rate of C02 loss was comparable to Karves results, indicating that a gas
compartment having heavy weight may be responsible for the high methane
content seen in his biogas. However, a large volume of methane gas was
also lost from the gas compartment during this experiment, indicating that
overall methane yields are also decreased when the pressure inside the gas
compartment is increased, which is not desirable. Based on these results, it
does not appear possible for high yields of biogas having >90% methane
content to be produced simultaneously using this anaerobic digester design.
Achieving high efficiency sacrifices methane yield volumes, and vice versa.
Methane content in the biogas remains dependent on feed composition and
not on the digester design.
Since the biotic portion of this study only tested methane content for a
digester with one base size, a continuation of this study could combine the
biotic and abiotic methods in order to test the effect of different base sizes on
biogas carbon dioxide content when a consistent feedstock is used. One
consideration is that if the base used is too large, excess oxygen will be
dissolved into the liquid compartment and the system will no longer be
Methane Content in Biogas Roduced from Various Feedstocks
0 0 c
Sugar Feed Sim. WW Feed BcxSesel Sludge
Figure 18: Methane Content Found in Biogas Produced from Different Feeds
7.2 Biogas Yield
During the biodiesel sludge feed trial, digester stability was good and,
although actual volumes were low compared to expected, a relatively
consistent amount of biogas was produced daily (0.10 to 0.40 L/day). Based
on measurements taken over the feed trial, 45% of the COD loaded through
the feed was digested, however this is not evident in the gas volume results
since only 12.5% of the methane expected from that COD reduction was
collected. The actual rate of methane production in terms of COD reduction at
UCD, using Karves biodigester design, was 0.057 cm3 CHVmg COD
(compared to theoretical yield of 0.46 cm3 CHU/mg COD reduced).
One possible explanation for this discrepancy is that some amount of
undigested organic solids may have settled in the bottom of the liquid
compartment or solidified and floated on top of the water level under the gas
compartment. Effluent samples were taken at the top of the column (outside
of the gas compartment). Because this undigested but degradable matter
would not be mixed into the sample volume drawn, CODout measurements
taken may have been lower than they would have been in a fully mixed
system. This would in turn make the COD reduction appear higher than it
actually was. Accumulated floating scum was periodically removed from the
gas compartment when it started to block the gas outlet valve. It was
assumed that this was indigestible material but it is possible that some of the
removed material was organic, and its removal contributed to lower than
expected biogas yields.
Another explanation is that the expected yields displayed were
calculated under the assumption that 100% of the COD loaded is degradable
and represent the maximum yields possible. COD represents the total
concentration of all content in the water that can be chemically oxidized and
biological oxygen demand (BOD) measures the amount of organic content
that can be oxidized by bacteria specifically. Therefore, one explanation for
the relatively low biogas yields is that the COD loaded has a low
biodegradability. RMSE sent a sample of their sludge and wastewater mixture
to Warren Analytical Labs for analysis in April 2007. The results indicate that
the sample has a COD 7,980,000 mg/L and BOD of 132,000 mg/L. This
means that for that sample, the BOD concentration is only 0.0165 of the COD
concentration (COD is 60.5 times higher by concentration). This is indicative
of the low biodegradability of the biodiesel sludge. Using these analytical
results, an estimation of the BOD concentration that would exist in the
biodiesel sludge feedstock used in this study is shown in Calculation 6.
Concentration of BOD in wastewater/sludge sample sent for analysis:
Percent of entire volume that is sludge (estimated):
l ^ 26.6%
.'i. '-He** "4.?'. '%A *
Average BOD removed with solid pbrtion'^estimated):
- ' > -.W-*: 32.5%
BOD of sludge portico only:
BOD of feedstock when diluted to 10% sludge:
ft \K 1s
l 7 "-V- f t . 4290ntg/L
BOD as % of ODD of feed with same tiieftion: . J
_JtÂ£:___**-..............-V-- .^___________*w' '*%*>______________
Calculation 6: Estimation of BOD Loading in Biodiesel Sludge Feedstock
Based on these estimates, total BOD loaded through the feed is
32,175 mg, the remaining BOD after the feed trial would be 11,381 mg, and
the total BOD reduction would be 20,794 mg. This represents a 65%
reduction in BOD. Using Calculation 4, the new estimated methane yield
would be 9.57 liters. In this case, the actual yield of 5.17 liters is 54.0% of
expected versus 12.5% of expected using COD concentrations.
An additional adjustment that can be made to expected methane yields
involves the generally accepted assumption that 10% of COD or BOD is used
to build new microbial biomass and is not converted to methane (NACWA,
2007). It can also be assumed that 10% of the COD concentration in the
effluent is made up of microbial biomass, not undigested BOD (or
biodegradable COD). By applying this correction, the actual methane yields
as a percentage of expected increase to 14% using the COD calculation, and
to 60% using the BOD calculation.
When these adjustments are applied, organic content degradation
figures achieved in this research are in line with literature review findings
summarized in Table 2. However, the degradation rates are still not close to
Karves claims of close to 100% degradation, as discussed in Chapter 2.
Based on the literature review and the results of this research, it does not
appear to be possible to reach a rate of 100% degradation using Karves
simple digester design.
The efficiency of the simple digester design tested could be improved
by adding automated mixing and heating devices, which would increase the
quantity organic matter that is digested per unit of time. This would increase
the biogas yield and provide more energy via the volume produced rather
than by increasing the percent of the biogas that is methane. However, these
changes would increase the cost to operate the digester and would reduce
the net energy gain resulting from the implementation of a biodigester. The
adjustment of HRT and feed strength did not have a significant effect on
biogas yield during this study, so it is unclear whether further experimentation
with these ranges would increase biogas yield from a digester built using
Karves original design.
Balsam, J. (2006). Anaerobic Digestion of Animal Wastes: Factors to Consider.
NCAT Energy Specialists, ATTRA Publication #IP219.
Biogas Support Programme. (Feb 2000). Hands on: A Pat Solution. Television
Trust for the Environment. Apr 2007 <
Burke, D.A., Butler, R., and Hummel, S. (1997), "An Assessment of the AGF (Anoxic
Gas Flotation) High Rate Anaerobic Digestion Process." Water Environment
Federation 12th Annual Residuals and Biosolids Management Conference,
Bellevue, WA pp 423-430.
Carucci G. et al. (2005). Anaerobic Digestion of Food Industry Wastes: Effect of
Codigestion on Methane Yield. Journal of Environmental Engineering, 131
Eisenhauer, Luke and Aaron Perry. Personal interview. June 1,2006.
Energy Information Administration. (2004). Renewable energy trends. Official Energy
Statistics from the U.S. Government.
Frame, Dennis and Fred Madison. (2001). Anaerobic digesters and methane
production. University of Wisconsin-Extension, prepared for Discovery Farms
Hansen, R.W. (1992). Methane generation from livestock wastes. Colorado State
University. Figure from unpublished manuscript.
Held, C. et al. (2002). Two-stage anaerobic fermentation of organic waste in CSTR
and UFAF-reactors. Bioresource Technology, 81 (1), 19-24.
Grady, C.P., G.T. Daigger and H.C. Lim. (1999). Biological wastewater treatment,
2nd ed. Marcel Dekker, Inc., NY.
Karve, A.D. (2005). A new compact biogas system based on sugary/starchy
feedstock. Energy for Sustainable Development, Volume 4 (1), pages
Kempter-Regel, Brigitte & Trosch, Walter (2000). Processing and anaerobic
digestion of municipal solid waste. Fraunhofer IGB.
Kocman, Shawna. (2005). Comparison of biogas options for Rural Village in India.
Unpublished manuscript. University of Colorado, Denver, CO.
Krich, Ken et al. (2005). Biomethane from dairy waste: a sourcebook for the
production and use of renewable natural gas in California. Modesto, CA
Western United Diarymen
Lay, Jiunn-Jyi et al. (1999). Feasibility of biological hydrogen production from organic
fraction of municipal solid waste. Water Resources, 33 (9), 2579-2586.
Molongoski, John J. & KLUG, Michael J. (1980) Anaerobic metabolism of particulate
organic matter in the sediments of a hypereutrophic lake. Freshwater Biology
10 (6), 507-518.
National Association of Clean Water Agencies (NACWA). (2007). NACWA
Comments on Wastewater Treatment Emissions Estimates in EPAs
Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2005, Draft for
Expert Review. Unpublished manuscript. Available online at www.nacwa.org.
Prasertsan, P. et al. (1993). Optimization for growth of Rhodocyclus gelatinosus in
seafood processing effluents. World Journal of Microbiology and
Biotechnology, 9, 593-596.
Pitterle, Mark. (2005). Photo of Biodigester Used in Karves 2005 Field Trials.
REN21 Renewable Energy Policy Network. 2005. Rural (off-grid) renewable energy.
Renewables 2005 Global Status Report. Washington, DC:Worldwatch
Saifudden, N. and K.H. Chua. (2004). Production of Ethyl Ester (Biodiesel) from used
Frying Oil: Optimization of Transesterification Process using Microwave
Irradiation. Malaysian Journal of Chemistry, 2004, Vol. 6, No. 1, 077 082
Seereeram, Shanta. (2004). Aqua Enviro anaerobic digestion trials final report.
Prepared for SLR Consulting, .
Tetlow-Smith, A. (1995) Environmental factors affecting global atmospheric methane
concentrations. Prog. Phys. Geog. 19, 336-50.
University of California Davis ASCE Student Chapter. (1998). Environmental
Engineering Competition Rules and Guidelines: water treatment from your
kitchen. Unpublished manuscript.