University of Southern California Dissertations and Theses

Lab-scale and field-scale study of siloxane contaminants removal from landfill gas

(USC Thesis Other)

Lab-scale and field-scale study of siloxane contaminants removal from landfill gas

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Lab-scale and Field-scale Study of Siloxane Contaminants
Removal from Landfill Gas

By

Alireza Divsalar

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMICAL ENGINEERING)

December 2017

i

Dedication

This thesis is dedicated to my father Mr. Gholamreza Divsalar and my mother Mrs. Kafi Divsalar
for their unlimited love and support in every stage of my life.

ii

Acknowledgment
I would like to express my sincere regards to my advisers Professor Theodore T. Tsotsis and
Professor Fokion Egolfopoulos for giving the chance to do my PhD in their research group and
under their supervision. This dissertation is a result of their valuable help, abundant support and
remarkable patience. I also wish to thank Professor Kathrine Shing for serving on my qualifying
and dissertation committee. I would like to thank Prof. Edward Goo and Prof. Jayakanth
Ravichandran from Mork family department of chemical engineering and materials science for
serving as my qualifying exam committee members as well.
This work could not be done with generous help of Mr. Richard Prosser, our partner from ES
Engineering Services, with high years of experience in Landfill gas industry, whose advice and
recommendation always accelerated the field-test study process.
My special thanks also go to PhD colleagues and friends in our research group: Dr. Sasan Dabir,
Dr. Ashkan Garshasbi, Dr. Zhongtang Li, Dr. Nazanin Entesari, Dr. Nitin Nair, Dr. Majid Monji,
Mr. Lin Sun, Mr. Matthew Dods, Ms. Ying Zhou, Mr. Linghao Zhao, Mr. Mingyuan Cao and MR.
Zhuofan Shi. They all helped me in my PhD courses and experiments. Special acknowledgement
to Dr. Kasra Khosoussi from MIT, Longlong Chang and Ashkan Davani from USC Mechanical
Engineering Department for their aid in simulation and modeling part of the thesis. Special
appreciation and gratitude to my brother Mr. Hasan Divsalar not only for his assist for the field-
test study analysis part, but also for his innumerable support and companionship in all of the hard
moments.

iii

Many thanks to Ms. Tina Silva, the Department’s Instructional Lab Manager and Mr. Shokry
Bastorous, Department’s Instructional Lab Assistant, and Mr. Philip Sliwoski, the director of the
USC glass shop for their unlimited helps in every way possible throughout the past four years. I
would also like to express my gratitude to Mork Family Department staff, Ms. Laura Carlos, Ms.
Heather Alexander, Ms. Angeline Fugelso, Ms. Karen Woo, Mr. Andy Chen, and Mr. Martin
Olekszyk, the Department Business Analyst, who is no longer with us. He helped me a lot through
ordering parts and instruments for my lab-scale and field-test studies. May his soul rest in peace.
Above all, I would like to express my deepest appreciation to my parents, Gholamreza Divsalar
and Kafi Divsalar. They have been my great mentors and motivators in my entire life and I want
to thank them from the bottom of my heart for their indefinite love and support; no matter where
they are, they are close to me in my heart.

iv

Contents

Dedication ........................................................................................................................................ i
Acknowledgment ............................................................................................................................ ii
List of Tables ................................................................................................................................. vi
List of Figures ............................................................................................................................... vii
Abstract ........................................................................................................................................... x
Chapter 1 Introduction .................................................................................................................... 1
1.1. Landfills and Landfill Gas................................................................................................ 2
1.2. Siloxanes .......................................................................................................................... 6
1.3. The Problems Caused by the NMOC and the Siloxane Compounds ............................... 8
1.4. NMOC and Siloxanes Removal Techniques .................................................................. 10
1.4.1. Adsorption technique .............................................................................................. 11
1.4.2. Absorption technique .............................................................................................. 13
1.4.3. Other methods ......................................................................................................... 14
Chapter 2. UV Photodecomposition of Siloxanes in Biogas ........................................................ 16
2.1. Introduction ........................................................................................................................ 17
2.2 Experimental ....................................................................................................................... 22
2.3 Modeling ............................................................................................................................. 26
2.4. Results and Discussion ...................................................................................................... 35
v

2.4.1. Experimental ............................................................................................................... 35
2.4.2 Modeling ...................................................................................................................... 45
2.5. Conclusion ......................................................................................................................... 50
Chapter 3. Field Testing of the PhoR for Siloxanes Removal in Landfill Gas ............................. 51
3.1. Introduction ........................................................................................................................ 52
3.2. Experimental ...................................................................................................................... 56
3.3. Results and Discussion ...................................................................................................... 62
3.4. Conclusion ......................................................................................................................... 65
Chapter 4 Scale-up Design and Economic Analysis..................................................................... 66
4.1 Introduction ......................................................................................................................... 67
4.2 Simulation and Modeling .................................................................................................... 69
4.3 Scale-up and Design ........................................................................................................... 74
4.4 Conclusion .......................................................................................................................... 79
Chapter 5 Ideas for Future Work .................................................................................................. 81
References ..................................................................................................................................... 84

vi

List of Tables
Table 1.1 Components frequently present in landfill gas………...………………………………...5
Table 1.2 Siloxane compounds known to be present in LFG……………………………………....7
Table 2.1 Parameters fitting results for all mixtures of L2 and D4 with air……………………….46
Table 3.1 Landfill gas analysis at Whittier landfill plant………………………………………….57
Table 4.1 BENA flare station gas analysis for organic silicon NMOC compounds……………..68
Table 4.2 Pipe and Lamp specifications used for simulation……………………………………..70
Table 4.3 Siloxane concentrations and total conversion in each stage……………………………74
Table 4.4 Scaled-up system specifications……………………………………………………….76
Table 4.5 Costs for UV photodecomposition and adsorption methods……..…………………….77
Table 4.6 Specifications of UVC LED bulb vs. Mercury lamps………………………………….80

vii

List of Figures
Figure 1.1 Structures of two most common siloxanes available in LFG…………………………...6
Figure 1.2 Spark plug (left) and engine head (right) damaged by silica……………………………9
Figure 1.3 Silica particles on turbine blade wheel (left) and cylinder head (right) ………………...9
Figure 2.1 Schematic of the UV photo-decomposition reactor (PhoR)…………………………...22
Figure 2.2 Schematic of the experimental set-up…………………………………………………24
Figure 2.3 Schematic of the UV lamp geometry with the walls at certain distances for intensity
validation…………………………………………………………………………………………28
Figure 2.4 Incident radiation profiles on a line along the lamp at 5mm distance for different solid
angles………………………………………………………………………………………….…29
Figure 2.5 Incident radiation profiles for a line on a Wall-04 placed at 5mm from the lamp for
different mesh size distributions………………………………………………………………….30
Figure 2.6 Laser/Power meter (left) and UV meter (right) instruments for intensity
measurements…………………………………………………………………………………….31
Figure 2.7 Measured surface intensity profile of the UV lamp …………………………………...32
Figure 2.8 Simulation results and experimental measurements of surface incident radiation along
the length of the lamp at different radial distances from the surface………………………………32
Figure 2.9 UV lamp surface temperature measurements …………………………………………33
viii

Figure 2.10 Simulated and experimental temperature profiles at different radial distances from the
lamp………………………………………………………………………………………………34
Figure 2.11 Measured and Simulated temperature of the outer quartz tube………………………34
Figure 2.12. Effect of feed concentration and residence time on the conversion of L2 in air……...36
Figure 2.13. Effect of feed concentration and residence time on D4 conversion in air……………37
Figure 2.14. Effect of feed concentration and the residence time on the mixture of L2 and D4 in
air………………………………………………………………………………………………...38
Figure 2.15. Effect of feed concentration and residence time on the L2 conversion in SLFG…….39
Figure 2.16. Effect of feed concentration and residence time on the D4 conversion in SLFG…….40
Figure 2.17. Effect of feed concentration and residence time on L2 and D4 mixture in SLFG....…41
Figure 2.18 L2 and D4 conversions mixed with LFGV cylinder as function of flow rate….…..…42
Figure 2.19 Effect of UV reactor on other VOC components presence in LFGV cylinder……......43
Figure 2.20 Molecular structures of L2 (right) and L3 (left)……………………………………...44
Figure 2.21 Conversion of L2 and L3 mixed with SLFG cylinder as function of flow rates……...44
Figure 2.22 Simulation and experimental results for of L2 and D4 conversions as function of flow
rates……………………………………………………………………………………………....47
Figure 2.23 Proposed schematic of the scale-up reactor where the lamps are not in contact with the
gas ……………………………………………………………………………………………….48
Figure 2.24 L2 mole fraction profile along the reactor for different inner diameters ……………..48
ix

Figure 2.25 L2 mole fraction profile along the UV reactor for different number of UV lamps
employed…………………………………………………………………………………………49
Figure 3.1 Schematic of field-test experimental set-up…………………………………………...58
Figure 3.2 18W and 41W UV lamps utilized for the field-test study…………………………...…60
Figure 3.3 Reactor configuration where the lamps placed outside of the quartz (a) and the case
where they are in direct contact of the gas (b)…………………………………………………….61
Figure 3.4 L2 and D4 conversions for 18W UV reactor vs. residence time for different feed
concentrations……………………………………………………………………………………62
Figure 3.5 Field-test study of L2 and D4 conversions as function of flow rates for 41W UV
reactor……………………………………………………………………………………………63
Figure 3.6 41W UV reactor inlet (left) and outlet (right) sides …………………………………...64
Figure 4.1 Mesh distribution of geometry with one lamp inside a 33.5 ID pipe ………………….70
Figure 4.2 D4 conversion as function of flow rate ………………………………………………..71
Figure 4.3 Contours of the D4 mole fraction (left) and incident radiation (right) on surface normal
to lamp……………………………………………………………………………………………71
Figure 4.4 Mesh distribution for the geometry with 169 lamps…………………………………...72
Figure 4.5 Flow rate per lamp as function of number of lamps to attain a given concentration …..73
Figure 4.6 schematic of the scale-up system …………………………………………..................75
Figure 5.1 Conversion of VOC compounds using FTCMR technique……………………………83
x

Abstract
In this study, the feasibility of using a photochemical process for landfill gas (LFG) clean-up has
been investigated. The process involves the use of an ultraviolet (UV) photodecomposition reactor
(PhoR) that is primarily employed for removing Si-containing trace compounds known as
siloxanes frequently found in LFG, but also happens to simultaneously remove a number of other
contaminants in LFG known as non-methane organic compounds or NMOCs. The study includes
both lab-scale experiments as well as a field-test of the process at a landfill located in the City of
Whittier. In the lab-scale experiments, the effectiveness of PhoR in siloxane removal was
investigated in both air and simulated landfill gas with (SLFGV) and without (SLFG) other NMOC
being present. High siloxanes conversions were attained using the UV reactor that was shown
capable to convert the siloxanes into silica (SiO2) particles. A reactor model was also developed
that properly accounts for the transport and reaction phenomena that take place. The model was
used to fit the lab-scale experimental data. A commercial CFD software, specifically, the ANSYS
Fluent package was utilized together with companion codes written in other programming
languages (Matlab and C) in order to analyze the model and to fit the experimental data. The data-
validated was subsequently to design the reactors for the field-test study and for hypothetical
process design and economic evaluation. The focus of the field-test was to evaluate the ability of
the PhoR to function properly in the presence of real LFG, which turned out to be the case. Another
goal was to test a number of different reactor configurations.
1

Chapter 1 Introduction

2

1.1. Landfills and Landfill Gas
One of the most common methods for disposing of the solid wastes is landfilling which is burying
the wastes under the soil. Landfilling is the most efficient way for disposing the wastes and has
negligible effect on the surrounding environment and ground-water if the process is properly
designed and operated. Leachate and landfill gas (LFG) are two streams that result from landfilling
of the wastes. Unlike leachate that is in the liquid phase and has no known beneficial uses, landfill
gas contains a high methane concentration (~40% or higher) and is considered as a potential
renewable resource for power and electricity generation.
Unfortunately, despite its high methane concentration, LFG does not find today widespread use
for power generation, and instead a large fraction of it either escapes to the environment as fugitive
emissions or is being flared with the by-products of its combustion being released to the
atmosphere. As a result, landfill gas is considered as the third largest anthropogenic source of
methane emissions in the United States and accounts for 18% of the total methane emissions
according to Environmental Protection Agency (Inventory of U.S. greenhouse gas, EPA, 2016).
Out of all methane emissions contributed by the landfills (active and closed sites), around 91% of
the emissions are caused by open landfills (Powell et al. 2015).
The most efficient way to control the fugitive methane emissions generated in landfills is to
combust the gas to produce electricity. Studies on the generation and use of LFG first began in the
late 1970’s, and their number increased significantly during the 1990’s due to the promise of LFG
as a potential renewable source for power and electricity generation, and as a means to help avoid
global climate change due to methane emissions. The interest in LFG is also driven by newly
enacted regulations and legislature. For example, a law enacted in California mandates that by
3

2020 1/3 of the State’s power generation must come from renewable sources (including LFG and
biogas), and 50% of it by 2030.
According to EPA (http://www.epa.gov/outreach/sources.html), each person in the USA generates
~4.4 lb/day of landfilled waste; based on their estimates that ~100 m
3
of CH4 is generated per Mg
of waste, this translates into 8x10
11
ft
3
of CH4 produced yearly from landfills in the USA alone,
equivalent roughly to 138 million barrel (Mbl) of crude oil, Sludge digesters in municipal waste-
water treatment plants (MWTP) contribute significantly as well in the production of biogas,
equivalent to ~21% of the gas from landfills, bringing the total to ~ 170 Mbl of crude oil equivalent.
According to EPA, in September 2014, 636 LFG-based electricity generating projects were
operating in 48 US States, producing ~16 billion kWh of electricity, enough to power ~1.6 million
homes. Similar potential for LFG utilization exists in other countries as well. For example, the
potential for power generation from landfills in Mexico has been estimated to be ~7.6 x10
5
MW/h
(Aguilar-Virgen et al 2014). Numerous other unexplored opportunities for the use of LFG exist
worldwide. For example, there is no waste-to-energy facility in Makkah (Saudi Arabia) to take
advantage of the landfill gas. Makkah hosts millions of people each year and existence of such
system would generate significant income from landfill gas diversion for power generation
(Nizami et al. 2017).
As noted above, one of the potential LFG uses is to generate electricity. Combustion of the LFG
in internal combustion (IC) engines, gas turbines and micro-turbines are the three most commonly
used techniques to generate electricity from LFG. Most of the existing facilities today utilize IC
engines that are suitable for up to 3 MW size projects. Micro-turbines and gas turbines are also
used for projects in the range 30-250 kW and 5 MW and higher, respectively. Less common
technologies to generate electricity from LFG include boiler/steam turbines and also combined
4

systems. In steam turbines, LFG is combusted first in the boiler to produce steam that then powers
a turbine to generate electricity. In the combined systems, LFG is combusted in the gas turbine
first, but the turbine’s hot flue-gas exhaust is also used to generate steam that then powers an
auxiliary turbine to generate additional electricity.
Landfill gas can also be used directly in heating equipment like boilers and commercial building
furnaces as an alternative fuel to natural gas (NG). Some facilities are intentionally located near
landfills in order to utilize the LFG produced. In other instances, pipelines have been built to
transport the LFG from a landfill to a nearby such facility. LFG also is beginning to find use as a
high-BTU fuel for direct injection into the NG pipeline system. For that purpose, the LFG needs
to undergo significant and careful treatment. It must be purified of a number of toxic contaminants
that it contains, which are known collectively as non-methane organic compounds (NMOC). In
addition, its methane content must increase to match the NG specs (such gas is then known as
renewable natural gas or RNG). RNG may also be used in processes that generate alternative liquid
fuels such as methanol, ethanol, or DME.
When solid wastes are buried, some biodegradation processes occur in order to generate the gas.
This process may continue for up to 20 to 50 years after the initial landfilling of the waste. The
biodegradation process that generates the LFG is thought to take place in four phases. In the first
phase (Phase I) of waste decomposition, long molecular chains of carbohydrates, proteins and
lipids are broken down by aerobic bacteria and carbon dioxide is generated as the by-product. This
phase lasts until all the oxygen present is consumed by the aerobic bacteria. In the second phase,
anaerobic bacteria generate acetic, lactic and formic acids and alcohols like methanol and ethanol
using the compounds created in the previous phase. The produced acids react with the moisture
that is present and generate carbon dioxide and hydrogen. In Phase III, anaerobic bacteria produce
5

acetate by consuming the organic acids generated in Phase II. The produced acetate and carbon
dioxide are then consumed by methane-producing bacteria. Methane concentration starts
increasing in this stage till the last phase where it becomes almost steady. LFG during this phase
is, typically, composed of approximately 50-55% CH4, 45-50% CO2, and 5% N2, and smaller
amounts of NMOC’s such as benzene, vinyl chloride, chloroform, 1,1-dichloroethene, carbon
tetrachloride, etc., including a particularly troublesome class of Si-containing compounds known
as siloxanes – see further discussion below. In addition, non-organic species such as hydrogen
sulfide and vapor-phase mercury are often found in LFG. Table 1-1 summarizes typical
compositional information for LFG.
Table 1.1 Components frequently present in landfill gas
Component Volume (%) Characteristic
CH 4 45-60 CH 4

is a naturally occurring, colorless, and odorless gas. Its
concentration in ambient air is 0.0002%.
CO 2 40-60 CO 2 is a colorless and slightly acidic gas that occurs naturally at
a small concentration (0.03%) in the atmosphere.
N 2 2-5 N 2 comprises approximately 79% of the atmosphere. It is
odorless, tasteless, and colorless
O 2 0.1-1 O 2 comprises approximately 21% of the atmosphere. It is
odorless, tasteless, and colorless.
Ammonia 0.1-1 Ammonia is a colorless gas with a pungent odor. Atmospheric
concentrations are less than 0.0001%.
NMOC 0.01-0.6 NMOC are organic compounds (i.e., compounds that contain
carbon) excluding methane. NMOC may occur naturally or be
formed by synthetic chemical processes.

Sulfides

0-1
Sulfides (e.g., hydrogen sulfide, dimethylsulfide, mercaptans) are
naturally occurring gases that gives the landfill gas mixture its
rotten egg smell. Sulfides can cause unpleasant odors even at very
low concentrations. Ambient air concentrations are less than
0.001%
Hydrogen 0-0.2 Hydrogen is an odorless and colorless gas. Atmospheric
concentrations are less than 0.00005%.
CO 0-0.2 CO is an odorless and colorless gas. Atmospheric concentrations
are less than 0.00001%.

6

1.2. Siloxanes
Siloxanes are organosilicon compounds containing alternating silicon and oxygen atoms (Si-O-Si)
in their backbone, with various attached groups. The name siloxane is derived from their elements:
Silicon + Oxygen + Alkane (Ricaurte Ortega and Subrenat, 2009). The Si–O bond is 1.64 Å (vs. a
Si–C distance of 1.92 Å) and the Si–O–Si angle is rather open at 142.5
o
. This geometric
consideration is the basis of the useful properties of some siloxane-containing materials, such as
their low glass transition temperatures. They are made primarily through the hydrolysis of
chlorosilanes (which are, themselves, produced via the chlorination of Si by CH3Cl), the process
resulting in the basic raw materials (monomers) from which a host of products are made. Most
siloxanes volatize quickly into the atmosphere, during use, where eventually they degrade into
carbon dioxide, silica, and water (Dow Corning, 1997). Table 1.2 shows information about all the
siloxane compounds that are known to be present in LFG and biogas, including their formula,
molecular weights, boiling point (B.P.), and vapor pressure (V.P.).
Among all different types of siloxanes, the most abundant ones in LFG are Hexamethyldisiloxane
(L2), which is a linear siloxane, and Octamethylcyclotetrasiloxane (D4), a cyclic siloxane, with
their structures shown in Figure 1.1. The main focus in this research was then put on these two
types of siloxanes.

Hexamethyldisiloxane (L2) Octamethylcyclotetrasiloxane (D4)
Figure 1.1 Structures of two most common siloxanes available in LFG
7

Table 1.2 Siloxane compounds known to be present in LFG
Name CAS
Number
Formula MW B.P.,
o
F
V.P.,
mmHg
Hexamethylcyclotrisiloxane (D 3) 541-05-09 C 6H 28O 3Si 3 222 275 10
Octamethylcyclotetrasiloxane (D 4) 556-67-2 C 8H 24O 4Si 4 297 348 1.3
Decamethylcyclopentasiloxane (D 5) 541-02-6 C 20H 30O 5Si 5 371 412 0.4
Dodecamethylcyclohexasiloxane(D 6) 540-97-6 C 12H 36O 6Si 6 445 473 0.02
Hexamethyldisiloxane (L 2) 107-46-0 C 6H 18OSi 2 162 224 31
Octamethyltrisiloxane (L 3) 107-51-7 C 8H 24O 2Si 3 236 304 3.9
Decamethyltetrasiloxane (L 4) 141-62-8 C 10H 30O 3Si 4 310 381 0.55
Dodecamethylpentasiloxane (L 5) 141-63-9 C 12H 36O 4Si 5 384 444 0.07

D3 and D4 are the most common siloxane compounds used in cosmetic products. Since most of
the siloxanes have high vapor pressures, they are classified as volatile compounds, therefore users
of the cosmetic products are in direct exposure of the vaporized siloxane compounds. Siloxanes
are also available in silicone sealants, which are commonly used as adhesive agents in the
construction area. They are also used in other areas such as in the food industry, in the production
of cleaning and paint products and even in health-care products.
Siloxanes have some impressive properties in their use in cosmetic products giving them a
combination of soft and smooth feeling on the skin. Substituting the siloxanes with other
alternatives would not be straightforward and the price for the same product might be quite
expensive compared the one that contains siloxanes ingredients.
There has not been sufficient research on all the types of siloxanes in term of hazards they may
pose. However, to date, among all types of siloxanes that have been tested, only D5 has been
reported to potentially have a carcinogenic effect based on preliminary results, and the rest have
not been reported to show any significant genotoxic signs (some research, however, indicates
8

reproductive toxicity of D4 on male and female animals (Lassen et al. 2005)). The aforementioned
test on D5 was done by Dow Corning and was reported to EPA and shows that D5 is potentially
carcinogenic for rats.

1.3. The Problems Caused by the NMOC and the Siloxane Compounds
As noted previously, a major problem with LFG, preventing its widespread use for energy and
power production is its miscellaneous contaminants, especially the NMOC. During combustion,
these NMOC generate acids (HCl, H2SO4, etc.) and the siloxanes are decomposed to silanols (Si-
OH) and other compounds that are subsequently oxidized into CO2 and silicon dioxide (SiO2).
Silicon dioxide, in the form of silica microparticles, creates thin layers and films on the internal
surfaces of pieces of equipment like boilers, engines and furnaces, which then results in their
corrosion and failure, thus, necessitating frequent servicing (Ajhar et al., 2010; Popat and
Deshusses, 2008; Ohannessian et al., 2008). For instance, the silica films coating the boiler tubes
and the combustion chambers cause reduction in the heat transfer coefficients, necessitating
frequent and expensive maintenance (Augenstein and Pacey, 1992; Pacy, Doorn and Thorneloe,
1994). Figure 1-2 shows the silica particles coated on the engine spark plug and engine head
(Sevimoglu et al., 2013). Also, Figure 1-3 demonstrates the turbine blade and cylinder head
covered by silica micro-particulates. (Urban et al., 2009).
9

Figure 1-2 Spark plug (left) and engine head (right) damaged by silica

Figure 1-3 Silica particles on turbine blade wheel (left) and cylinder head (right)

Silica microparticles may also be emitted into the environment from the exhaust gas of the power
generating equipment and of various flares, and potentially pose danger for human health and the
environment. The aforementioned California law that mandates the use of renewable fuels means
that larger quantities of RNG may one day find their way into the NG pipeline network. It is
important, therefore, that one removes from the RNG all the NMOC, in general, and the siloxanes,
in particular. The current concern is that the common technologies used in order to remove these
10

NMOC, in general, and siloxanes, in particular, are not quite adequate, which may then result in
trace amounts of siloxane compounds entering the NG pipeline system; these, then, would turn
into silica microparticulates, when combusted in home appliances such as stoves and heaters,
potentially, causing serious environmental emissions and hazardous effects on human health
(Abatzoglou and Boivin, 2009; El-Fadel et al., 1997; Nair et al., 2012, 2013).

1.4. NMOC and Siloxanes Removal Techniques
According to Table 1-1, methane and carbon dioxide are the two most common compounds found
in landfill gas. In order to further use the gas for power and electricity generation, one should
separate carbon dioxide from the gas to enhance its methane content. Several techniques have been
proposed for this purpose. One of the methods is the use of membranes, for example, supported
ionic liquid membranes (SLIM), which preferentially transport the CO2 over the other components
of the LFG. In these membranes, the CO2 dissolves in the ionic liquid (Hojniak et al., 2014) that
infiltrates the membrane structure and prevents other molecules from permeating through. Saedi
et al. (2014) investigated the ability of polyethersulfone (PES) and polyethersulfone/polyurethane
(PU) composite membranes to separate CO2 from CH4. Adding PU results in membranes with
reduced CO2 permeance but enhanced CO2/CH4 separation selectivity. Simcik et al. (2016) studied
thin-film polyamide composite (TFC) membranes for this purpose. In addition to membrane
separation adsorption (Zhao et al., 2010; Gomes et al., 2001; Montanari et al., 2011; Scholz et al.
2013; Andriani et al., 2014) and absorption (Petersson et al., 2009; Ryckebosch et al., 2011;
Läntelä et al., 2012; Tippayawong et al., 2010; Baciocchi et al., 2012) processes find application
for the removal of CO2 from LFG to increase its energetic content.
11

NMOC removal from LFG prior to its combustion is a difficult problem, because of their wide
range and presence at trace amounts, and presents challenges for the conventional clean-up
technologies. A number of techniques, including adsorption, absorption and refrigeration, have
been investigated in the past for the removal of the halogenated and sulfided NMOC from LFG
and biogas, but significantly less is known on their applicability to rid the gas of the volatile
siloxanes. The latter are key components of many commercial and consumer products, such as
detergents, shampoos, deodorants, and other cosmetics which unfortunately find their way into
various landfills where such products are discarded. Because of their increased use, siloxanes have
emerged in recent years as one of the most difficult NMOC contaminants to control in LFG and
biogas. As with the treatment of the other trace NMOC constituents in LFG, adsorption, absorption
and refrigeration are the techniques currently utilized for siloxanes removal; these techniques are
expensive to apply, but remain commonly in use because there are no other commercially available
processes to replace them.

1.4.1. Adsorption technique
Adsorption is the most common technique used to remove NMOC impurities including siloxanes
from LFG. Adsorbents specifically used to remove gaseous siloxanes, include activated carbon
(AC), activated alumina (AA), and silica gel (SG). Activated carbon is one of the materials used
in many industrial gas separation processes due to its high adsorption capacity and large surface
area (Pradhan et al., 1999). Nam and Namkoong (2013) compared different types of carbon
adsorbents in removing L2, D4 and D5. They observed that the carbon adsorbents provided up to
80% removal rates for L2; their conclusion was that while carbon adsorbents seem to be effective
in removing siloxanes, removal rates are highly dependent on the type of siloxane molecule, and
12

the pore structure of the adsorbents used. Cabrera-Codony et al. (2014) reported that the presence
of humidity and other compounds in biogas will decrease adsorption of D4 on commercial
activated carbons. Ricaurte Ortega and Subrenat (2009) compared the application of different
porous media including AC, zeolite and silica gel in adsorption processes for siloxane removal.
They observed that the mass transfer within the porous material was more rapid for the AC rather
than for the zeolite and the silica gel. Also, they found that light siloxanes (L2) with a lower B.P.
break through much faster than their heavier counterparts, thus necessitating more frequent
regeneration which is relatively expensive, and because of its process complexity carries with it
the risk for system failure (Ajhar et al., 2012).
Typically, the adsorption media utilized are not particularly selective towards siloxanes, and will
adsorb most other trace contaminants in the LFG; thus, the adsorption of the rest of the NMOC
present in the gas significantly reduces the bed capacity for siloxanes. Normally adsorption
capacities (defined by the siloxane adsorbed when breakthrough is detected) for various AC media
range from 2,000-17,000 mg/lb (0.4 - 3.7 wt%), with virgin media having the highest capacity
(Wheless and Gary, 2002). The adsorption capacity of regenerated media is less than that of the
virgin media, and successive on site regenerations yield a progressively lower capacity, until media
replacement is necessary, which comes at a great cost. According to Noshadi et al. (2016), the
capacity loss of a hydrophobic mesoporous adsorbent was measured to be around 7% after one
cycle (100 min), which leads to a capacity loss of 54% after 10 cycles. Another key challenge with
adsorption is that it does not change the siloxane (or the NMOC, in general) molecules, which
when released from the beds during regeneration are still the same as when entering the beds.
Siloxane disposal, during regeneration, involves burning the off-gas which releases silica oxide
particles to the atmosphere. Sigot et al. (2014) observed that the siloxane adsorption (L2 and D5)
13

by silica gel (SG) decreased by almost 20% when humidity was added to the gas stream. Thus,
removing the humidity prior to reaching the silica gel adsorbent is required to attain acceptable
siloxane removal.

1.4.2. Absorption technique
Another approach for the removal of gaseous siloxanes is physical absorption using solvents like
Selexol
TM
and methanol operating at relatively high pressures in order to be effective. In a study
conducted by Schweigkofler et al. (2001), hot concentrated nitric acid was used in the absorption
process to remove siloxanes. They found that an elevated temperature was necessary in this
application, and 70 – 75% removal efficiency was achieved. Absorption of siloxanes from LFG
by alkaline solutions seemed to have moderate success due to presence of large concentrations of
CO2 that reacts with such solutions. Since the siloxanes are not soluble in water, methods like
water scrubbing are not that effective to remove trace amounts of such impurities in LFG.
Hydrophobic absorbents are not faring significantly better either. For example, Devia and Subrenat
(2013) compared different hydrophobic absorbents for L2 and D4 removal, including motor oil,
cutting oil, and a water-cutting oil mixture, as absorbents. They reported that the best absorbent
was motor oil, attaining removal rates for L2 and D4 of 60% and 80%, respectively, which are not
quite as high as those attained by the adsorption technique. A problem with using organic solvents,
is that in addition to the siloxanes they absorb most of the other organic NMOC in biogas, making
their application quite ineffective (Dewil et al., 2006).
Effective solvent regeneration is key to absorption processes in order to reduce solvent disposal
and system operating costs, and to assure long-term operation without significant consumption of
14

the scrubbing solution (solvent recycle impacts the removal efficiency, however, due to the
inadvertent accumulation of siloxane in the solvent). The capital/operating costs of the absorption
processes are similar to those for adsorption, but performance, as noted above, is inferior and
significant process improvement is, therefore, needed.

1.4.3. Other methods
One of the methods that can be utilized to separate condensable impurities including siloxanes and
other NMOC from landfill gas is a chilling process. Because of the relatively low concentrations
of siloxanes found in biogas and LFG, high removal rates using this method can only be obtained
at very low temperature (McBean et al., 2008). However, attaining such low temperatures requires
high energy consumption, which makes the method uneconomical (Urban et al., 2009). To improve
on that, Schweigkofler et al. (2001) used refrigeration combined with adsorption, and were able to
remove 98% of siloxanes from biogas from a water treatment plant. They still report, however,
that the high energy consumed by the refrigeration equipment for cooling the large flows of sewage
gas is a major drawback. Due to the lack of success by the conventional methods, a number of
non-traditional methods for siloxane removal have also been studied over the years. Popat et al.
(2008) used biological treatment to remove the siloxanes. A low removal efficiency (~10%) was,
however, reported in their study. Appels et al. (2008) studied peroxidation as a method to reduce
the siloxane content of biogas. A reduction in siloxane concentration of 50 – 85% was observed,
which is not so effective when compared with existing technologies. Ajhar et al. (2006) used the
AspenPlus  simulation code to investigate gas permeation via membranes as a potential new
method for siloxane removal. However, we know of no experimental data validating such an
approach.
15

It should be clear from the discussion, so far, that the common techniques like adsorption,
absorption and refrigeration are not all that effective for siloxane removal from LFG and biogas,
and more novel techniques like bioremediation and membrane separation face significant technical
and economic hurdles as well. There is a need, therefore, to develop more effective and efficient
techniques to rid these components from LFG, and this is a key focus for this study. UV
photodecomposition is the method that has been selected here for further study for the removal of
siloxanes from LFG and biogas. It is a promising technique in that regard, as it has been previously
shown effective in the treatment of a variety of VOC at trace amounts in air and other contaminated
gas streams. UV photodecomposition has, in fact, been investigated by a number of previous
studies in order to decompose a number of silicon containing compounds (Dalton, 1985; Ouchi et
al. 1999; Urbanova et al., 2001; Yingling and Garrison, 2003; Schmidt 2013). Urbanova et al.
(2001), for example, compared IR laser thermolysis and UV laser photolysis for the decomposition
of 1,3-diethyldisiloxane ([H2(C2H5)Si]2O). Their study revealed that typical hydrocarbon products
of UV photolysis are ethane (C2H6, a major product) accompanied by smaller amounts of ethylene
(C2H4), methane (CH4), propane (C3H8), and butane (C4H10). They reported that the cleavage of
the Si-O bond does not take place under photolysis. Dalton also showed that Si-O bond cleavage
is not a feasible photochemical step. Even though their findings provide great insight into the
photodecomposition of siloxanes, their work focuses on the reaction pathway of photoreaction and
there is no reaction rate information available to systematically evaluate the removal of siloxanes.

16

Chapter 2. UV Photodecomposition of Siloxanes in Biogas

17

2.1. Introduction
Because of its large methane content, biogas has been investigated in recent years as a potential
alternative renewable energy source (Amini et al., 2011; Havukainen et al., 2014). Biogas is
commonly produced from the biodegradation of organic waste in landfills and from biological
sludge in anaerobic digesters in waste-water treatment plants (WWTP). Although biogas is mostly
composed of CH4 and CO2, together with smaller concentrations of O2 and N2, it also contains a
variety of trace NMOC impurities that include sulfided compounds and halogenated organic
compounds, as well as several silicon-containing compounds known as siloxanes (Thorneloe et
al., 2003) Siloxanes are commonly found in commercial and consumer products such as
detergents, shampoos, deodorants, and other cosmetics and find their way into landfills and
WWTP as these products are consumed or discarded. The increased usage of siloxane-containing
products in recent years has resulted in increasing siloxane concentrations in LFG.
The presence of the NMOC in biogas creates technical challenges that presently hinder its
widespread use as a renewable fuel for power generation (Jalali et al., 2013). For example, the
sulfur- and halogen-containing NMOC when combusted generate mineral acids which corrode the
power generating equipment, and if released to the environment may contribute to acid rain.
Siloxanes, such as L2 and D4, have been shown to generate silica microparticulates during
combustion, which have been found to damage the equipment, and also present a potent threat to
the environment, if released (Augenstein et al., 1992; Pacey et al., 2002; Urban et al., 2009; Nair
et al., 2012). Thus, prior to using the biogas for power and energy production, its NMOC trace
constituents must be removed. For the use of biogas in renewable energy production to become
more economical and less harmful to the environment, cost-effective remediation techniques for
its NMOC constituents, including the siloxane pollutants, must therefore be developed. To date,
18

considerable effort has been expended on the remediation of the sulfur- and halogen-containing
NMOC, but relatively little research has focused on the removal of siloxanes from biogas (Ho et
al., 2013; Lakhouit et al., 2016), which is the primary focus of this study.
Currently, the most common techniques employed commercially for remediating siloxanes in
biogas, including the most frequently encountered ones, such as L2 and D4, are adsorption and
absorption (Ajhar et al., 2010). Adsorption is, typically, an effective technique for the removal of
trace contaminants from gas waste-streams and is commonly utilized for the clean-up of biogas as
well. However, it is not all that effective for the clean-up of siloxanes because their concentration
in LFG is, typically, quite low compared to those of the other NMOC in the LFG, some of which
also show a lot more affinity towards activated carbon, the common adsorbent utilized, rather than
the siloxanes. Thus, in practical settings the adsorption capacity of the adsorbents towards the
siloxane molecules is greatly hindered by the other (especially the non-volatile) NMOC, therefore
diminishing the effectiveness of the adsorption technique (Urban et al., 2009; Montanari et al.,
2010; Jafari et al., 2016) for siloxane remediation. Light siloxanes, like L2, are more difficult to
deal with as they break through the adsorbent beds much quicker than do heavier siloxanes (e.g.,
D4), as a result necessitating frequent and often expensive regeneration of the adsorbent (Ortega
et al., 2009). Additionally, adsorption processes do not alter typically the chemical structure of the
siloxane molecules. Hence, during adsorbent regeneration, which involves the incineration of the
off-gas, the siloxanes are decomposed into silica microparticles that are released into the
atmosphere.
Absorption is another technique that has been studied for the removal of siloxanes. Reactive
absorption processes are capable to convert siloxanes into less volatile compounds, but they
require the use of harsh absorption agents (e.g., nitric acid and/or sulfuric acid) something that
19

significantly increases the cost and technical complexity associated with these processes
(Schweigkofler et al., 2001). Physical absorption techniques involve less toxic chemicals but, on
the other hand, are not all that effective for the removal of light siloxanes (e.g., L2) due to their
high volatilities. Such volatile siloxane pollutants often require relatively high pressures in order
to achieve high removal efficiencies (Rasi et al. 2014; Ghorbel et al., 2014). Additionally, physical
absorption processes also suffer from similar disadvantages with the competitive adsorption
techniques: (i) the need for absorbent regeneration, which can reduce siloxane removal efficiencies
over time due to the accumulation of siloxanes inside the absorbing medium from successive
regeneration cycles; (ii) the fact that the siloxanes are released intact during the regeneration
process and must be properly disposed of.
In addition to adsorption/absorption, other methods have also been investigated for the removal of
siloxanes from biogas. They include refrigeration, biofiltration and membrane separation. Because
they require a relatively large capital investment and have high operating costs, refrigeration
processes are only economically effective for high biogas flow rates and at elevated siloxane loads.
The energy costs associated with refrigeration are a major drawback (Schweigkofler et al., 2001),
as are concerns about ice formation in humid LFG streams (Urban et al., 2009; Ajhar et al., 2010).
A novel idea is to use refrigeration in combination with another technique, such as adsorption, as
did Schweigkofler et al. (2001) who reported a 98% siloxane removal efficiency, but high energy
costs still remain a concern even for the hybrid process. Biofiltration is an effective technique for
the removal of volatile organic compounds from gas waste-streams; however its application for
siloxane remediation in biogas has not proven successful, because of the low removal efficiencies
attained, due to mass transfer limitations and possibly due to the low biodegradability of siloxane
compounds (Wang et al., 2014; Soreanu et al., 2016). Membrane separation technologies are
20

continuous processes which is an advantage over their adsorption/absorption counterparts (which
require adsorbent/absorbent regeneration). Current membrane processes are not all that effective,
however, in dealing with the dilute siloxane concentrations in biogas and exhibit significant
methane parasitic losses of ~7%, which hurts their economic efficiency (Ajhar et al., 2012).
Another idea that has been tried is to diminish siloxane concentration at the point of biogas
generation. Appels et al. (2008) have studied the removal of siloxane compounds in waste
activated sludge via peroxidation, and have demonstrated siloxane removal efficiencies of 50-85%,
which are moderate when compared to existing siloxane removal technologies (such an approach
is, however, not feasible for LFG).
In summary, conventional (adsorption/desorption) and emerging (refrigeration, biofiltration,
membrane separation) techniques have not, as yet, proven effective for the removal of siloxanes
in biogas. What is proposed for study here, instead, is a simple and potentially cost-effective
technique for remediating siloxane pollutants in LFG that involves the use of a UV
photodecomposition reactor (PhoR). The use of such reactors for the removal of trace
contaminants from gaseous waste-streams is not new. In the last decade or so several studies have
demonstrated the ability of UV reactors to decompose VOC in contaminated air streams into less
harmful compounds (Zhuo-wei et al., 2013; Huang et al., 2011; Pengyi et al., 2003). UV radiation
at wavelengths of ~200 nm can cleave the O=O bond in O2, thus ozone (O3) is likely to form under
such conditions via the reaction of atomic oxygen with a neighboring O 2 molecule (Kockarts,
1976). Zhang and Anderson (2013) have demonstrated that the presence of O3 in UV
photodecomposition reactions of VOC can have profound effects on the quantum efficiencies of
these reactions, perhaps due to the fact that O3 strongly absorbs the UV photons. Li et al. (2012)
21

and Mahajan et al. (2015) have even demonstrated the potential for these UV/O3 processes to
remediate hydrogen sulfide (H2S), a pollutant that is commonly found in LFG.
Though we are not aware of the use of UV photodecomposition to remove siloxanes from biogas,
a number of past studies have demonstrated the potential for UV light to decompose silicon-
containing compounds (Dalton 1985; Ouchi et al., 1999; Urbanová et al., 2001; Yingling et al.,
2003) to form Si-containing solid materials. By irradiating the siloxane molecules with UV light
at wavelengths less than 200 nm, it is possible to cleave the Si-C bonds of the siloxane molecule
to yield Si and CH3 radicals. When oxygen is present, the UV light also helps to cleave the O=O
bond yielding two oxygen radicals, which are then available to react with the Si radicals from the
siloxane decomposition to form SiO2 solid materials (Dalton 1985; Ouchi et al., 1999; Kockarts,
1976; Pola et al., 2002; Ouyang et al., 2000; Brinkmann et al., 2001). By analyzing the UV
absorption spectrum of a given siloxane molecule, one can identify the different wavelengths that
are potentially optimal for the cleaving of the Si-C bonds.
Though not presently practiced for removing siloxanes, LFG is an ideal environment to apply the
UV PhoR to decompose and remove such NMOC. One reason for that, is that LFG in addition to
CH4, and CO2 commonly contains small concentrations of O2 (1-3%), which are likely to
participate in the siloxane photodecomposition reaction. In addition, because the absorption
spectrum of the CH4 molecule (the main component of LFG) is located in the IR region, it is
unlikely that UV radiation, when used to photo-decompose the siloxanes, will decompose the CH4
molecules as well. This is important in terms of diminishing parasitic reactions that may degrade
the energetic content of LFG. The key focus of this study, therefore, is to investigate the application
in the lab of a UV PhoR to decompose two model siloxanes, namely L2, a linear siloxane, and D4,
a cyclic siloxane, in a LFG stream (the choice of L2 and D4 is because they are the two most
22

common siloxanes found in biogas and LFG). We utilize a homogeneous UV PhoR in our
research, as photocatalytic reactors (e.g., using TiO2 as a catalyst) are not well-suited for the
application, as siloxane photo-decomposition on surfaces deposits Si-containing solid films that
can quickly degrade photocatalytic activity. In addition, though TiO2 has been used as a
photocatalyst to decompose various VOC in contaminated air streams (Liang et al. 2010, 2012;
Moulis et al., 2013, Neti et al., 2010; Hussain et al., 2011; Cheng et al., 2013), the mineral acid
by-products of decomposition of the halide- and sulfur-containing NMOC found in LFG are likely
to also be quite harmful to catalyst lifetime.

2.2 Experimental
Figure 2.1 shows the schematic of the UV photo-decomposition flow reactor used in our lab-scale
experiments. It is a simple configuration employing a single low-pressure 18W (G18T5VH) UV-
C lamp (connected to an electric ballast which acts as a transformer to control the power to the
lamp) installed inside a quartz tube. Low-pressure UV-C lamps made of high purity quartz emit
their maximum energy output at a wavelength of 254 nm and 185 nm.

Figure 2.1 Schematic of the UV photo-decomposition reactor (PhoR)
23

Figure 2.2 shows a schematic of the overall experimental set-up, which is placed inside an
enclosure equipped with a powerful inline suction fan in order to prevent potential explosion in
case of any system leaks. The system consists of three key parts: (i) the gas delivery system, (ii)
the PhoR (see Figure 2.1), and (iii) the analysis section. In the laboratory studies, we utilize
simulated LFG with two different compositions. The first type of LFG (hereinafter referred to as
SLFG) consists of 52% methane, 38% carbon dioxide, 9% nitrogen, 1% oxygen, and is purchased
premixed in a high-pressure cylinder from the Matheson Company. The second type of simulated
LFG (hereinafter referred to as SLFGV, purchased premixed from the Air Liquide Company, has
the same composition as the SLFG but in addition contains five model NMOC. Specifically, it
contains Carbonyl Sulfide (50 ppm), 1,3-Dichlorobenzene (5 ppm), Dimethyl Sulfide (49.8 ppm),
Trichloroflouromethane (52.4 ppm) and Vinyl Chloride (47.9 ppm). The high-pressure cylinder
containing the simulated LFG (SLFG or SLFGV) is connected to the system with two different
stainless steel lines via a pressure regulator (that controls the gas delivery pressure), each line being
connected to a separate Mass Flow Controller (MFC) that is used to set the flow of gas. The
temperature of the lines downstream of the MFC is kept constant at a desired value by utilizing
heating tapes and temperature controllers, and by insulating the lines well with glass wool held in
place wrapped with aluminum foil.
To prepare the (simulated) LFG feed mixtures containing a pre-determined concentration of
siloxanes (L2 and D4 in this Thesis), we use a syringe pump (Harvard Apparatus PHD 2000)
equipped with a 500 microliter (µL) syringe (1750 TLL Hamilton syringe) that is able to deliver
very small and consistent flow rates of liquid siloxanes. Using a heated tee-fitting, whose
temperature is controlled by a controller, the liquid siloxane flow is delivered into one of the two
24

lines noted above carrying the heated simulated LFG stream, where it is then vaporized and
transferred into the heated gas at the tee-junction.

Figure 2.2 Schematic of the experimental set-up

Since only trace (ppm-level) concentrations of siloxanes are typically found in real landfill gas,
and there are limitations on how low a flow rate of liquid siloxane the syringe pump can deliver in
a stable manner, to be able to generate the low siloxane concentrations we use in our experiments
we need to mix and dilute the siloxane-containing LFG stream with the other heated LFG stream.
Typically, during the experiments the syringe pump is set at its lowest stable flow rate, and the
flow rates of the two heated LFG streams are adjusted in such a manner as to generate the combined
LFG stream with the desired siloxane concentration. In order to adjust the residence time in the
reactor, only part of the flow of the combined LFG stream is directed to the photodecomposition
reactor, while the rest is diverted to the fume-hood using a by-pass line. Teflon (PTFE) tubing is
used from the mixing point onward to the reactor and for the lines to the analyzer – see below – in
25

order to minimize the potential for adsorption of the siloxanes on the tubing walls. In addition, all
lines to the reactor and to the analyzer are also heated with heating tapes, with their temperature
being controlled by controllers in order to further diminish the potential for adsorption.
The feed and exit siloxane compositions are analyzed using a GC/MS (7890A Agilent GC and
5975C MS), via the use of two PTFE lines to the 6-port sampling valve of the GC-MS, which is
equipped with a 250 µL sampling loop. For the analysis, we used Agilent GC Column 30𝑚 ×
250𝜇𝑚 × 0.25𝜇𝑚 . The operating settings for the analysis were: GC temperature-30 °C rising @
15.0 °C/min, final temperature- 120 °C; GC total flow 13.2 mL/min, split ratio 1:5, purge flow 3
mL/min and vent flow 8.5 mL/min. MS quad and MS source temperatures are 150 °C and 230 °C,
respectively.

26

2.3 Modeling
The main goal of this research was to provide a “proof-of-concept” of the proposed PhoR
technology in the laboratory prior to its field-testing at a real landfill site. An additional important
goal was to develop a data-validated model, which properly accounts for all the phenomena that
take place inside the PhoR, including fluid flow, mass and heat transfer, UV radiation and the
photodecomposition reaction. This model was then utilized to design, construct and optimize the
experimental system used in the field-testing component of this study. In our research, we have
utilized the ANSYS® Fluent package as the CFD software to model the experimental behavior of
the UV photodecomposition reactor. For the experimental validation of the model, we have also
developed additional programing software that links with the ANSYS® Fluent software, and
which allows us to perform non-linear fitting regression of the laboratory data to estimate the
various system parameters, including the UV lamp power characteristics and the rate parameters
of the global photodecomposition reaction rate expression.
To analyze the UV radiation profiles in the PhoR, we employ the Discrete Ordinates (DO) model
in the ANSYS® Fluent. The radiative transfer equation (RTE) used in the DO model for an
absorbing, emitting and scattering medium at 𝑟⃗ in the direction 𝑠⃗ is defined as follows:
𝑑 𝐼 (𝑟 ⃗,𝑠 ⃗)
𝑑 𝑠 + (𝑎 + 𝜎 𝑠 ) 𝐼 (𝑟 ⃗, 𝑠 ⃗) = 𝑎 𝑛 2
𝜎 𝑇 4
𝜋 +
𝜎 𝑠 4𝜋 ∫ 𝐼 (𝑟⃗, 𝑠⃗
′
)
4𝜋 0
Φ (𝑠⃗ . 𝑠⃗
′
)𝑑 Ω
′
(1)
In Equation 1, 𝑠 , 𝑟 ⃗, 𝑠 ⃗ and 𝑠⃗
′
are the path length, position vector, direction vector and scattering
direction vector, respectively. 𝐼 (𝑟 ⃗, 𝑠 ⃗) is the radiation intensity, which is a function of 𝑟⃗ 𝑎𝑛𝑑 𝑠⃗. 𝑎
and 𝜎 𝑠 in the 2
nd
term on the left are the absorption and scattering coefficients, respectively. In the
first term on the right side of the equation, n, T and 𝜎 are the refractive index, local temperature
27

and the Stefan-Boltzmann constant (5.62 × 10
−8
𝑊 𝑚 2
𝐾 4
), respectively. Finally, Φ and Ω
′
are the
phase function and the solid angle.
The DO model solves Equation 1 via a finite-differences discretization of the solid angles each
associated with a vector direction 𝑠 ⃗ fixed in the global Cartesian system (x, y, z). Therefore, the
radiative transfer Equation 1 can be recast as:
∇. (𝐼 (𝑟⃗, 𝑠⃗)𝑠⃗) + (𝑎 + 𝜎 𝑠 ) 𝐼 (𝑟⃗, 𝑠⃗) = 𝑎 𝑛 2
𝜎 𝑇 4
𝜋 +
𝜎 𝑠 4𝜋 ∫ 𝐼 (𝑟⃗, 𝑠⃗
′
)
4𝜋 0
Φ (𝑠⃗ . 𝑠⃗
′
)𝑑 Ω
′
(2)
ANSYS® Fluent also allows the modeling of non-gray radiation using a gray-band model. The
RTE for the spectral intensity 𝐼 𝜆 (𝑟 ⃗, 𝑠 ⃗) can be written as:
∇. (𝐼 𝜆 (𝑟⃗, 𝑠⃗)𝑠⃗) + (𝑎 𝜆 + 𝜎 𝑠 ) 𝐼 𝜆 (𝑟 ⃗, 𝑠 ⃗) = 𝑎 𝜆 𝑛 2
𝐼 𝑏𝜆
+
𝜎 𝑠 4𝜋 ∫ 𝐼 𝜆 (𝑟⃗, 𝑠⃗
′
)
4𝜋 0
Φ (𝑠⃗ . 𝑠⃗
′
)𝑑 Ω
′
(3)
In Equation 3, 𝜆 is the wavelength, 𝑎 𝜆 is the spectral absorption coefficient, and 𝐼 𝑏𝜆
is the black
body intensity given by the Planck function. As the UV lamp emits radiation in a range of
wavelengths, Equation 3 is used in our simulations.
In the numerical implementation of the DO radiation model, each octant of the angular space 4𝜋 at
any spatial location is discretized into 𝑁 𝜙 × 𝑁 𝜃 solid angles of magnitude 𝜔 𝑖 , called the control
angles. Here, 𝜃 and 𝜙 refer to the polar and azimuthal angles, respectively, defined in the global
Cartesian system (𝑥 , 𝑦 , 𝑧 ). As a result of the discretization, at each spatial position the space is
divided into 𝑁 𝜙 × 𝑁 𝜃 directions, and Equation 3 is solved for each cell specifying a given
direction. As one may expect, the accuracy of the problem solution depends on the degree of solid
angle discretization. However, as the 𝑁 𝜃 and 𝑁 𝜙 values increases, more memory is needed for the
computation, therefore, the cost of computation increases. Hence, it is important to find optimum
values for these parameters to minimize the computational time needed. According to Ho and
28

Khalsa (2011) from Sandia National Laboratories and also the ANSYS® Fluent manual itself for
the DO radiation model, “Theta Pixels” and “Phi Pixels” should be set at 3*3. Therefore, we kept
these two parameters constant and varied 𝑁 𝜃 𝑎𝑛𝑑 𝑁 𝜙 .
In order to select the appropriate 𝑁 𝜃 𝑎𝑛𝑑 𝑁 𝜙 , the cylindrical reactor geometry was generated and
a hexahedral mesh distribution was assigned using the ANSYS® ICEM. Figure 2.3 shows the
geometry and the hexahedral mesh distribution for solid angles optimization. The reactor geometry
consists of an inner cylinder representing the UV lamp an outer cylindrical surface representing
the reactor wall, and other concentric cylindrical surfaces where the model is used to calculate the
temperature profiles and incident radiation (Due to the cylindrical reactor symmetry, in the
simulations, only one quarter of the geometry was used).

Figure 2.3 Schematic of the UV lamp geometry with the walls at certain distances for intensity validation

UV lamp surface
Wall-04 (5mm)
Wall-03 (10mm)
Wall-02 (15mm)
Wall-01 (20mm)
29

Selecting the optimum values of 𝑁 𝜃 𝑎𝑛𝑑 𝑁 𝜙 involves varying the values for 𝑁 𝜃 𝑎𝑛𝑑 𝑁 𝜙 and
running the simulation for each pair of values and recording changes, if any, in the radiation
profiles. The optimum pair was selected when no further changes were observed (less than
0.005%) in the calculated radiation profiles upon further increase in the value of 𝑁 𝜃 and/or 𝑁 𝜙 .
For improved simulation accuracy, a larger 𝑁 𝜙 value is required as increasing the value of 𝑁 𝜃 did
not noticeably affect the simulation results. Once the appropriate values for 𝑁 𝜙 , 𝑁 𝜃 are chosen the
number of hexahedral cells in the mesh distribution is increased and the radiation intensity profiles
are recalculated. An optimal number of such cells was selected beyond which no noticeable (less
than 0.005%) changes in the profiles were recorded. Figure 2.4 demonstrates the intensity profiles
along a line on Wall-04 (5 mm distance from the lamp) for different solid angles values. According
to Fig 2.4, one can conclude that 𝑁 𝜙 = 14 𝑎𝑛𝑑 𝑁 𝜃 = 3 not only account accurately for the
radiation emission, but also solve the problem with less time and memory utilization.

Figure 2.4 Incident radiation profiles on a line along the lamp at 5 mm distance for different solid angles
Theta Divisions= 3
Phi Divisions= 14
Theta Divisions= 14
Phi Divisions= 14
Theta Divisions= 16
Phi Divisions= 16
Theta Divisions= 10
Phi Divisions= 10
30

Figure 2.5 Incident radiation profiles for a line on a Wall-04 placed at 5 mm from the lamp for different
mesh size distributions

Figure 2.5 shows the radiation profiles at 5 mm distance from the lamp for different mesh size
distributions and fixed optimized solid angles (𝑁 𝜙 = 14 𝑎𝑛𝑑 𝑁 𝜃 = 3). According to Fig 2.5,
200,000 cells are adequate for the software to attain precise get convergence while employing less
memory.
Prior to using ANSYS® Fluent in deriving the global photodecomposition rate expression and for
the design of the pilot-scale PhoR unit, the ability of the model to describe the UV intensity (W/m
2
)
characteristics and the temperature profiles inside the reactor was validated. For validating the
ability to properly describe the UV radiation characteristics, we experimentally measured the
intensity profiles generated by the 18W UV lamp. In these measurements, we employed two
different measuring devices, specifically a laser power/energy meter (Coherent LabMax-TO) and
207756 Hexahedral cells 1406496 Hexahedral cells
42134416 Hexahedral cells 10448956 Hexahedral cells
31

a UV meter (EIT Power Puck Low Power). The former (the UV power meter) measures the
radiation (W/m
2
) at a wavelength of 254 nm, which is in the UVC range, while the latter (the laser
power/energy meter) measures the power for a broad range of wavelengths (>248 nm). Figure 2.6
shows the Laser/Power meter and UV meter instruments used to measure the intensity of the lamp.

Figure 2.6 Laser/Power meter (left) and UV meter (right) instruments for intensity measurements

The surface intensity radiation of the lamp was measured using the Power/Laser meter and is
shown in Figure 2.7. The experimental profile was then incorporated (inserted) into the ANSYS®
Fluent code via a user-defined function (UDF) code as the radiation flux on the surface, and then
the incident radiation at different axial and radial distances from the lamp was computed and
compared with the experimentally measured values. Figure 2.8 shows the measured and simulated
incident radiation along the length of lamp at a radial distance of 5 mm, 10 mm, 15 mm and 20
mm from the lamp. The points are the experimental data and the lines are the fits provided by the
ANSYS® Fluent code. As one can conclude from Figure 2.8, the model provides an adequate fit
of the experimental intensity data.
Detector analyzer screen
Detector
32

Figure 2.7 Measured surface intensity profile of the UV lamp
Figure 2.8 Simulation results and experimental measurements of surface incident radiation along the length
of the lamp at different radial distances from the surface
0
50
100
150
200
250
300
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Intensity (W/m2)
Axial Direction (m)
Simulation 5mm
Simulation 10mm
Simulation 15mm
Simulation 20mm
Experiment 5mm
Experiment 10mm
Experiment 15mm
Experiment 20mm
0
50
100
150
200
250
300
350
400
450
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Intensity (W/m2)
Axial direction (m)
33

For testing the model’s ability to properly describe the heat transfer characteristics, we measured
the temperature at the surface of the UV lamp in the laboratory atmosphere (see Figure 2.9) and
also at different radial distances from the lamp along its length. A UDF code was then used to
import the measured surface temperature data into the software in order to perform the heat transfer
simulations. Assuming natural convection as the primary mechanism for the heat transmission,
and treating air as an ideal gas, the temperature along the length of the lamp was calculated at
different radial distances from the lamp surface.
Figure 2.10 shows the experimental temperatures and the model fit. The model does an adequate
job in describing the experimental data.

Figure 2.9 UV lamp surface temperature measurements

273
283
293
303
313
323
333
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Temperature (K)
UV lamp length (m)
34

Figure 2.10 Simulated and experimental temperature profiles along the lamp length at different radial
distances from the lamp

Figure 2.11 Measured and Simulated temperature of the outer quartz tube
Figure 2.11 shows the simulated temperature profile at the outer surface of the quartz tube and also
the measured temperature at four points during an experiment running for a mixture of D4 and
315
320
325
330
335
340
345
350
355
360
365
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Temperature (K)
Lamp length (m)
Experiment
Simulation
295
300
305
310
315
320
325
330
335
340
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Temperature (K)
Lamp length (m)
Experiment- r = 0
Experiment- r = 5mm
Experiment- r = 15mm
Simulation - r = 0
Simulation- r = 5mm
Simulation- r = 15mm
35

LFG. As can be seen from this figure, the simulated heat transmission inside the quartz tube can
be used reliably for further parameter fitting, design and scale-up.

2.4. Results and Discussion
2.4.1. Experimental
We have studied here the photodecomposition of two different siloxanes, namely the linear
siloxane L2 and the cyclic siloxane D4. They were chosen because they are the siloxanes most
commonly encountered in biogas and landfill gas. The photodecomposition reaction was studied
with the siloxanes being present in trace amounts in three different carrier media, specifically air,
and the two aforementioned model LFG (SLFG and SLFGV) with the compositions previously
noted. In the experiments, we have studied the effect of varying the feed siloxane concentration
and the residence time on the conversion of the siloxanes. A key objective of these experiments
was to generate the experimental data required to determine the overall global rates of
decomposition for the two tested siloxanes, and to use these rates for sizing the field-scale
experimental reactor.
In the first series of experiments air, laced with the two model siloxanes, was used as the carrier
gas. The reasons for carrying out the air experiments are two-fold: First, a fundamental one to
determine the differences in reactivity among the two different carrier media (i.e., air and LFG).
In addition, siloxane photodecomposition studies in air are of practical importance, in their own
right, during the regeneration of active carbon beds saturated with siloxanes and other NMOC. In
our studies, L2 and D4 either individually or as a mixture were added to an air stream and were
then passed through the reactor to undergo UV photodecomposition. In the experiments, the impact
36

of siloxane feed concentration and the residence time through the reactor were investigated. Figure
2.12 shows the conversion of L2 in air as a function of the air stream flow rate (i.e., for different
reactor residence times) for various siloxane feed concentrations.

Figure 2.12 Effect of feed concentration and residence time on the conversion of L2 in air

The reactor conversion decreases as the flow rate increases (i.e., as the residence time decreases)
as expected. On the other hand, the feed concentration seems to have a relatively small impact on
reactor conversion, which is consistent with an overall global rate expression which is near-linear
in the L2 concentration. This is further validated by employing the PhoR model to calculate the
global rate parameters, i.e., the rate constant and order of reaction. The order of the L2
decomposition reaction in air is estimated to be ~0.82, see Table 2.1 and further discussion to
follow.
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10
Conversion (%)
Flow rate (cc/sec)
L2 concentration = 15.2 ppm
L2 concentration = 35.8 ppm
L2 concentration = 53.2 ppm
L2 concentration = 79.9 ppm
37

Figure 2.13 shows the conversion of D4 in air as a function of the air stream flow rate (i.e., for
different reactor residence times) for various feed concentrations. The order of reaction of D4, is
quite similar to that for L2. As Table 2.1 indicates, the activation energy for D4 decomposition is
a bit higher (10.91 kJ/mol vs. 9.84 kJ/mol) that that for L2, however, the pre-exponential factor is
somewhat larger as well, with the net outcome that the photodecomposition rates for the two
siloxanes being fairly similar to each, though that for D4 is a bit lower.

Figure 2.13 Effect of feed concentration and residence time on D4 conversion in air

0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8
Conversion (%)
Flow rate (cc/sec)
D4 concentration= 13.5ppm
D4 concentration= 19.6ppm
D4 concentration= 28.5ppm
D4 concentration= 34.5ppm
38

Figure 2.14 shows the conversion of (L2+D4) mixtures in air as a function of feed flow of air and
various siloxane concentrations, as indicated on the Figure. The goal of these experiments was to
investigate the effect that the linear siloxane (L2) may have on the conversion of the cyclic one
(D4) and vice versa, since these two siloxanes are typically found together in landfill gas and
biogas. At the time of submission of this Thesis the investigation of the global reaction kinetics
of the mixtures is continuing and results will be presented in an upcoming publication. However,
given the dilute concentrations of these mixtures and the fairly similar global reaction kinetics
observed with the individual siloxanes we do not expect to see significant differences observed
between the values of these parameters in the mixture case and those calculated for the individual
siloxanes.
Figure 2.14 Effect of feed concentration and the residence time on the mixture of L2 and D4 in air

0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/sec)
L2 concentration = 7.5 ppm
L2 concentration = 14.5 ppm
L2 concentration = 30 ppm
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
flow rate (cc/sec)
D4 concentration = 6.2 ppm
D4 concentration = 13.4 ppm
D4 concentration = 27.3 ppm
39

Figure 2.15 shows the conversion of the L2 intermixed in SLFG as a function of the flow rate of
the carrier gas at different trace concentrations of the siloxane. As can be seen in Figure 2.15, there
is again not a significant dependence of conversion on siloxane concentration, which is in
agreement with Figure 2.12 reporting the results of L2 photodecomposition in air. Therefore,
though the investigation of the global reaction kinetics is still ongoing, the expectation is that the
calculated reaction rate order for the case of siloxane photodecomposition in SLFG will be again
near one. When comparing the experimental conversions of L2 in SLFG versus those that in air
for the same residence times in the PhoR, those in air are higher. Such differences in the
photocomposition rates for L2 in SLFG and air are to be expected, given the role that oxygen plays
in ozone generation during UV photodecomposition processes, as previously discussed.

Figure 2.15 Effect of feed concentration and residence time on the L2 conversion in SLFG

Figure 2.16 shows the conversion of D4 intermixed in SLFG as a function of the flow rate of the
carrier gas at different trace concentrations of the siloxane. As can be seen in Figure 2.16, there is
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8
Conversion (%)
Flow Rate (cc/s)
L2 concentration = 20 ppm
L2 concentration = 40 ppm
L2 concentration = 72.4 ppm
40

again not a significant dependence of conversion on siloxane concentration, which is in agreement
with Figure 2.13 reporting the results of D4 photodecomposition in air. Therefore, though the
investigation of the global reaction kinetics is still ongoing, again the expectation is that the
calculated reaction rate order for the case of siloxane photodecomposition in SLFG will be again
near one. When comparing the experimental conversions of D4 in SLFG versus those that in air
for the same residence times in the PhoR, those in air are higher. Again, such differences in the
photocomposition rates for D4 in SLFG and air are to be expected, given the role that oxygen plays
in ozone generation during UV photodecomposition processes, as previously discussed for the case
of L2. As was the case with air as a carrier gas, the photodecomposition rate of D4 in SLFG is
lower to that of L2, but the differences in conversion are more pronounced in SLFG case.
. Figure 2.16 Effect of feed concentration and residence time on the D4 conversion in SLFG
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8
Conversion (%)
Flow rate (cc/sec)
D4 concentration = 1.59ppm
D4 concentration = 3.6ppm
D4 concentration = 5.5ppm
D4 concentration = 11ppm
D4 concentration = 13.2ppm
41

Figure 2.17 Effect of feed concentration and residence time on L2 and D4 mixture in SLFG

Figure 2.17 shows the conversion of (L2+D4) mixtures in SLFG as a function of the feed flow of
the carrier gas and various siloxane concentrations, as indicated on the Figure. The goal of these
experiments was again to investigate the effect that the linear siloxane (L2) may have on the
conversion of the cyclic one (D4) and vice versa. At the time of submission of this Thesis the
investigation of the global reaction kinetics of the siloxane mixtures in SLFG is continuing.
However, given the dilute concentrations of these mixtures and the fairly similar global reaction
kinetics observed with the individual siloxanes, as in the case with air as the carrier medium, we
do not expect to see significant differences observed between the values of these parameters in the
mixture case and those calculated for the individual siloxanes.

0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/sec)
L2 concentration = 2.39ppm
L2 concentration = 3.36ppm
L2 concentration = 5ppm
L2 concentration = 11.9ppm
L2 concentration = 12.6ppm
L2 concentration = 19ppm
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/sec)
D4 concentration = 1.46ppm
D4 concentration = 2.25ppm
D4 concentration = 4.63ppm
D4 concentration = 9.52ppm
D4 concentration = 12.08ppm
D4 concentration = 18.2ppm
42

In order to investigate the effect that the presence of other NMOC present in LFG may have on
siloxane photodecomposition, experiments were carried out in which the siloxanes were
intermixed with SLFGV as a carrier gas. As a reminder, SLFGV has the same composition as
SLFG, other than the fact that it contains, in addition, five model trace NMOC that are typically
encountered in real LFG. Figure 2.18a shows the conversion of L2 mixed in SLFGV as function
of carrier flow rate for various siloxane concentrations. Figure 2.18b is the corresponding Figure
for D4. Comparing the experimental conversions for L2 and D4 in the two different model LFG
(SLFG and SLFGV) one observes that the presence of the other NMOC have little impact on the
conversion of siloxanes.

Figure 2.18 L2 and D4 conversions mixed with LFGV cylinder as function of flow rate

0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/s)
L2 concentration = 25ppm
L2 concentration = 70 ppm
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/s)
D4 concentration = 17ppm
D4 concentration = 100ppm
43

Figure 2.19 Effect of UV reactor on other VOC components presence in LFGV cylinder
Figure 2.19 shows the conversion for the other contaminants present in the SLFGV as a function
of the carrier gas flow rate. The result shows that the UV radiations not only helps to decompose
the siloxane compounds, but also has significant effect on other NMOC compounds as well, for
which the UV reactor appears as effective in decomposing.
In order to investigate the effect of molecular structure and size on the conversion of the siloxanes,
another linear siloxane compound namely Octamethyltrisiloxane (L3) has also been investigated
in the UV photocatalytic reactor Figure 2.20 compares to L2, the L3 molecular structures, with the
L3 having an extra 𝑂 − 𝑆𝑖 bond substituted with a methyl group. Moreover, L3 is less volatile than
L2 as Table 1.2 shows (the vapor pressure of L2 is 31 mmHg, while for L3 it is 3.9 mmHg).

0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8
Conversion (%)
Flow rate (cc/s)
Vinyl Chloride-- 47.9ppm
Trichlorofluoromethane--52.4ppm
Dimethylsulfide--49.8ppm
1,3-Dichlorobenzene--5ppm
Carbonyl Sulfide--50ppm
44

Figure 2.20 Molecular structures of L2 (right) and L3 (left)
L3 and L2 were injected to the system and mixed with SLFG as the carrier gas. Figure 2.21 shows
the conversion of L2 and L3 at different initial feed concentrations and flow rates. According to
the conversions obtained for L2 and L3, one can state that the size of the molecule does not seem
to have effect on the conversion when using the UV photodecomposition reactor.
Figure 2.21 Conversion of L2 and L3 mixed with SLFG cylinder as function of flow rates

0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/s)
L3 concentration = 10 ppm
L3 concentration = 25 ppm
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/s)
L2 concentration = 11 ppm
L2 concentration = 30 ppm
45

Finally we have also carried out a systematic analysis and search for various gas phase by-products
resulting from the decomposition of the three model siloxanes. Though some small secondary
peaks were observed in the MS spectra (we are continuing the efforts to identify such peaks) their
sizes are quite small, indicative that the UV photodecomposition technique is quite effective in
completely decomposing the siloxane compounds.

2.4.2 Modeling
The ultimate goal of this project is to field-test this technology in a real landfill. For that purpose,
a data-validated model for the reactor is required that properly accounts for all transport
phenomena including the fluid flow in the reactor, its heat transmission characteristics, and the
UV radiation spatial distribution. In addition, one must have good knowledge of the global reaction
kinetics. In this study, the ANSYS® Fluent package was utilized as the CFD software, and as
discussed above it properly accounts for all the aforementioned transport phenomena in the UV
reactor. Here, this simulation package is utilized to generate global rate expressions for the
photodecomposition of the siloxanes from the fitting of the experimental data. For that purpose, a
programing language was developed that links with the ANSYS® Fluent software, and which
allows one to perform the non-linear fitting regression.
The global reaction rate utilized is given by Equation 4 below.
−𝑟 𝑆𝐿
= 𝑘 0𝑆𝐿
𝑒 −𝐸 𝑆𝐿
𝑅𝑇
𝐶 𝑆𝐿
𝑛 𝐼 (𝑥 , 𝑦 , 𝑧 ) (4)
In Equation 4, 𝑘 0𝑆𝐿
, 𝐸 𝑆𝐿
𝑎𝑛𝑑 𝑛 are the pre-exponential factor, the activation energy and order of
the reaction, respectively. The assumption is made here that the rate of the siloxane photo-
degradation process depends linearly on 𝐼 (𝑟 ) which is the fluence rate, defined as the total radiant
46

power incident from all directions at that location. The reaction rate also depends on the n-th power
of the siloxane concentration. Because oxygen is in substantial excess, its concentration
dependence is not accounted explicitly in the global rate expression, but its effect is implicitly
embedded in the value of the effective rate constant.
Table 2.1 shows the calculated values of the rate constant parameters from the fitting of the
experimental data previously described for the (L2+air) and (D4+air) experiments. Figure 2.22
shows an example case (at a select siloxane feed concentration) of the experimental measurements
and simulation result for the conversion of L2 and D4 mixed with air as a function of flow rate.
As one can observe, the model accurately predicts the conversion based on the values computed
by the parameter fitting regression method. At the time of submission of this Thesis, simulations
are currently in progress in order to calculate the values for the rate constants for all the other sets
of experimental data reported in Section 2.4.1 above. When all such kinetic parameters are
identified they will, hopefully, be in line with the following conclusions drawn from the
experimental observations: (i) the global photodecomposition rates of siloxane in air are faster than
those in the model LFG, as expected given the positive influence the presence of oxygen has on
those rates; (ii) the photodecomposition rate of L2 is faster than that of D4 in both air and in the
LFG; (iii) the global reaction rate parameters for each siloxane (L2 or D4) remain unaffected by
the presence of the other siloxane.
Table 2.1 Parameters fitting results for all mixtures of L2 and D4 with air
Mixture Pre-exponential factor

Activation Energy
(kJ/mol)
Reaction order
L2 + Air 4.66 x10
-4
9.84 0.822
D4 + Air 5.6 x10
-4
10.91 0.826
47

Figure 2.22 Simulation and experimental results for of L2 and D4 conversions as function of flow rates

The availability of a data-validated UV photodecomposition reactor model allows for the optimal
design and scale-up of a larger field-scale unit. One of the PhoR configurations being evaluated is
one in which the UV lamps are not in direct contact with the LFG stream. A schematic of such a
reactor configuration employing eight such 36W UV lamps is shown in Figure 2.23 below (the
inner side of the outside tube enclosing the UV lamps is coated with a reflective material that
prevents the radiation from penetrating through).

0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8 10
Conversion (%)
Flow rate (cc/s)
L2 concentration = 80ppm - Simulation
L2 concentration = 80ppm - Experiment
0
10
20
30
40
50
60
70
80
90
100
0 2 4 6 8
Conversion (%)
Flow rate (cc/s)
D4 concentration = 13.5 ppm - Simulation
D4 concentration = 13.5 ppm - Experiment
48

Figure 2.23 Proposed schematic of the scaled-up reactor where the lamps are not in contact with the gas
Figure 2.24 shows the simulation results of the impact on L2 conversion of employing internal
tubes with differing diameters (in these simulations Equation 4 with the following kinetic rate
parameters 𝑘 0𝑆𝐿
= 0.0014 , 𝐸 𝑆𝐿
= 4.86 𝑘𝐽 /𝑚𝑜𝑙 𝑎𝑛𝑑 𝑛 = 1 is utilized). As expected, employing
a larger diameter tube increases the residence time of the siloxane in the PhoR, which in turn results
in higher siloxane conversions.

Figure 2.24 L2 mole fraction profile along the reactor for different inner diameters

Fig. (a) Whole geometry Fig. (b) One quarter of the geometry
UV lamp
LFG
Outside tube
0.0E+00
2.0E-06
4.0E-06
6.0E-06
8.0E-06
1.0E-05
1.2E-05
0 0.2 0.4 0.6 0.8 1
L2 Mole Fraction
Reactor Length (m)
quartz ID= 2cm
quartz ID= 2.5cm
quartz ID= 3cm
49

Figure 2.25 show the impact of employing, 2, or 4 or all 8 of the UV lamps. Again, it is clear the
impact of providing the adequate radiation has for attaining the desired PhoR performance.

Figure 2.25 L2 mole fraction profile along the UV reactor for different number of UV lamps employed

0.0E+00
2.0E-06
4.0E-06
6.0E-06
8.0E-06
1.0E-05
1.2E-05
0 0.1 0.2 0.3 0.4
L2 Mole Fraction
Reactor Length (m)
8 Lamps
4 Lamps
2 Lamps
84.8% conversion
60.8% conversion
97.67% conversion
50

2.5. Conclusion
Lab-scale experimental studies for different L2, L3 and D4 siloxane concentrations in both air and
simulated LFG at different flow rates show that the siloxanes are highly responsive to the UV
radiation. Linear siloxanes (L2 and L3) due to their relatively simple structures absorb more UV
radiation than the cyclic ones (D4), and the conversion for such linear siloxanes should, therefore,
be higher than that of the cyclic compounds. Our experimental results for mixtures of (L2+SLFG)
and (D4+SLFG) generally corroborate this theoretical hypothesis.
Not only is high siloxane removal attained by this technology, but also other NMOC compounds
undergo a photocatalytic decomposition process as well, and thus high conversions were also
attained for those contaminants. Since other methods for NMOC removal are either expensive or
face technical difficulties, and they also do not really destroy the siloxanes and other contaminants,
the UV photodecomposition technology can be a good and inexpensive alternative.
The models that were developed also proved to account accurately for all phenomena that take
place inside the reactor. Reaction kinetics parameters that were computed by utilizing the non-
linear fitting regression toolbox option in Matlab linked with ANSYS® Fluent can precisely
regenerate the measured experimental data for mixtures of siloxanes in air (with studies presently
under way to validate the model use for fitting the siloxane in LFG experimental data). Such
models can, therefore, be applied for optimization of the field-test study apparatus in order to
compute the optimal number of lamps, pipe size length and diameter.

51

Chapter 3. Field Testing of the PhoR for Siloxanes Removal in
Landfill Gas

52

3.1. Introduction
In recent years, there has been a significant attention to renewable resources for power and
electricity generation. According to a 2012 California law, by 2020, 1/3rd of the State’s power
generation must come from renewables. Biogas and landfill gas (LFG) are potential alternative
renewable sources of energy, and attention to them, as a result, has increased in recent years.
Biogas contains 35-45% methane, 30-35% CO2, 5-10% N2, 1-3% O2 and some traces of
contaminants. A major problem with LFG preventing its widespread use is its miscellaneous
contaminants, commonly known as non-methane organic compounds or NMOC. They include
sulfided species (e.g., H2S, CS2, COS), halogenated organics and other volatile organic compounds
such as Si-containing compounds known as siloxanes, which are typically harmful and toxic to
human, animal, and plant life (He et al. 1997). The most common siloxanes in LFG and biogas are
the volatile methylsiloxanes (VMS), among which one finds both cyclic
(hexamethylcyclotrisiloxane (D3), octamethylcyclotetrasiloxane (D4),
decamethylcyclopentasiloxane (D5), and dodecamethylcyclohexasiloxane (D6)), as well as linear
(hexamethyldisiloxane (L2), octamethyltrisiloxane (L3)) siloxanes (Soreanu et al., 2011).
During power/energy production, these NMOC generate acids (HCl, H2SO4, etc.) and silica micro-
particulates which result in corrosion and equipment failure (Madi et al., 2015). Therefore,
frequent device maintenance and replacement is required. Also in the flare station at landfill plants,
SiO2 is generated when the landfill gas is flared which results in air pollution and may endanger
human health. According to the International Agency for Research on Cancer, silica particulates
are classified as a human carcinogen, which may cause lung cancer (Liu et al., 2013; Lin et al.,
2006; Vida et al. 2010; Tse et al., 2011) upon exposure. Siloxane compounds must, therefore, be
53

removed from LFG prior to its use for power generation. Common techniques including
adsorption, absorption and the chilling process are not quite efficient for siloxane removal.
In the adsorption processes, for example, commonly used adsorbents are not particularly selective
for siloxanes among all the other NMOC, According to Urban (2009), the presence of non-volatile
and sulfur-containing compounds can reduce the adsorbent’s capacity for the siloxanes. Also, the
presence of CH4, CO2 and water vapor can further decrease the adsorbent’s capacity for the
siloxane compounds (Cabrera-Codony et al., 2014). Some adsorbents that are capable to remove
the siloxanes cannot be readily regenerated (Ajhar et al., 2010). For adsorbents for which
regeneration is possible, the process often results in the reduction of the adsorbent’s capacity, so
frequent replacement is required at great expense (Gislon et al., 2013). Further, adsorption does
not change the toxic NMOC, which are, typically, released intact from the beds during
regeneration. Therefore, these compounds must either be flared upon adsorbent regeneration or
disposed in some other way (Ajhar et al., 2010).
The second common approach for siloxane removal is absorption. In this process, a solvent is used
to separate the impurities from the LFG. In absorption processes, either a strong acid (such as
H2SO4) is used to destroy the siloxane (chemical absorption) or water and some other organic
solvent can be used to physically absorb the siloxanes (physical absorption). No siloxane removal
was observed for the latter method (Schneider 2001; Rasi et al., 2008; Rossol et al., 2003). On the
other hand, high siloxane (L2, D4 and D5) compound elimination was observed using strong acids
as the chemical absorbents (Schweigkofler and Niessner, 2001). However, regeneration is costly
and results in solvent loss.
Deep chilling (i.e., refrigeration/condensation) is another method that can be effective for siloxane
removal but requires the use of very low temperatures due to the high vapor pressures of some of
54

the siloxanes (e.g., L2, L3, and D3). However, attaining these low temperatures requires a high
energy consumption, which is not-cost effective (Urban 2009; Rossol 2003). And since it is very
difficult to condense the most highly volatile siloxanes (L2, L3 and D3), no significant removal
rates are observed for these siloxanes (Ajhar et al., 2010).
Biological processes for siloxane removal consist of the use of aerobic and anaerobic filters in
order to remove the siloxane compounds. For example, 45 mg/m
3
of D4 intermixed in humid air
was tested using biological filters by Popat et al. (2008), and 43% conversion was obtained at a
residence time of 20 min. Due to the large residence times required, the method is not practical to
be applied in real landfills characterized by high LFG flow rates. A laboratory-scale biotrickling
filter (BTF) for siloxane removal was studied by Soreanu et al. (2010). The BTF was operated
under anaerobic conditions with low flowrates of biogas for three months by feeding siloxane-rich
biogas at the bottom of the BTF, at different flowrates and siloxane concentrations, while counter-
currently feeding a nutrient solution at the top. The BTF exhibited generally low performance,
which Soreanu et al. (2010) attributed to low substrate availability and low siloxane
biodegradability. Because of the low efficiency (maximum of 60% with very small flow rates) this
method (use of BTF) does not appear promising for siloxane removal.
There are only a few field-testing studies of landfill gas clean-up using new methods and
technologies. Most of the methods used are conventional and not that much effective. For instance,
Shin et al. (2002) studied the effect of moisture and LFG composition on the removal
characteristics of trace compounds in the LFG via adsorption onto activated carbon in the
Sudokwon landfill. They concluded that the breakthrough time and adsorption capacity of
benzene, toluene, and ethylbenzene decreased rapidly when the relative humidity of the LFG was
over 60%. Yan et al. (2016) combined the bubble column and the UV-Fenton reaction to degrade
55

VOCs in landfill gas. They focused on the effective degradation of BTEX which is a group of
toxic and typical air pollutants in landfill gas and found the UV-Fenton process was an effective
method for BTEX degradation. However, the removal efficiency will be affected by many factors,
including the amount dissolved oxygen and initial conditions.
A field-test study of a catalytic/sorption hybrid process (CSHP) was performed by our research
group (He et al., 1997) in order to remove halogenated and sulfided compounds from LFG. The
test was done on a landfill site in Anoka County, Minnesota. Although the conversions were
promising for the halogenated and sulfided compounds, no analysis was performed for siloxane
contaminants removal.
Lakhouit et al. (2016) investigated the effectiveness of two novel biofilters in reducing VOC
emissions through a field study conducted at a municipal solid waste landfill site in Quebec,
Canada. The removal efficiency was shown to be practical, in the range of 54% to 100% for many
of the VOCs found in the LFG. The problem with this method is that the flow rate of the gas
feeding to the bio-filter was quite low (between 12 to 18 L/h) in order to become an alternative
method to be replaced with the conventional techniques.
In conclusion, commonly used techniques for siloxanes removal are not significantly effective and
do not have the potential to be used as a scaled-up system for use in wastewater treatment plants
and landfills for siloxane removal from biogas. What is proposed in this study, instead, is the use
of a UV photodecomposition reactor that has been tested in the lab and has shown promise in the
remediation of the landfill gas and for decomposing the siloxane compounds into silica
particulates. The reactor consists of a UV lamp that emits its radiation at wavelengths of 185 nm
and 254 nm. This high energy then helps to cleave oxygen bonds which, in turn, results in the
production of ozone.
56

As noted in Chapter 2 of this Thesis, lab-scale siloxane removal was tested in three different carrier
media, namely, air and two different simulated LFG (SLFG and SLFGV) whose compositions
were previously provided. High siloxane conversions were observed. In addition, the UV
photodecomposition reactor proved effective in decomposing a number of model VOC found in
one of the model LFG (SLFGV). In this Chapter, we describe preliminary results of studies
focusing on the field-testing of the technology (at the time of the submission of this Thesis the
field studies are ongoing with results to be reported in an upcoming publication). The key goal of
such studies is to validate the ability of the PhoR system to operate successfully while treating real
LFG.

3.2. Experimental
Figure 3.1 shows the schematic of experimental set-up for the field-test which is installed at the
City of Whittier Landfill plant. The system is connected to the blower outlet of the pipeline via
Teflon tubing. Since the pressure of the gas inside the pipe is rather low, a brushless diaphragm
pump (TD-4X2N) from Brailsford & Co, Inc. has been installed too boost up the pressure. Since
LFG is saturated with water, a 0.5” pipe was utilized to collect the water droplets in order to prevent
the pressure drop in the system. The system is equipped with a low range pressure regulator to
keep the pressure constant inside the system. The total LFG flow rate is measured by a rotameter
ranging from 1 to 5 SCFH.
The composition of the LFG at the landfill is measured on a regular basis, and one such
measurement is shown in Table 3.1.
57

Table 3.1 Landfill gas analysis at the Whittier landfill plant
Siloxane compounds are not monitored on a regular basis at the landfill and at the start of our field-
testing such information was not available. Analysis by our Group indicated that siloxanes are
present, but their amounts are atypically low, less than 1 ppm total. It was decided, therefore, for
the field-testing at the site employing a side stream of the real LFG to spike the gas with two
siloxane compounds (L2 and D4). These were injected into the system via an accurate syringe
pump (Harvard Apparatus PHD 2000) that is able to deliver the liquid siloxanes at very low flow
58

rates (microliters per hour). The liquid siloxanes are vaporized using a heated tee whose
temperature is controlled by thermocouples and a temperature controller (Omega CN9000a). The
mixture of LFG and siloxane compounds are separated into two streams whose flow rates are
controlled by rotameters and needle valves. Since siloxane removal rates are measured at different
flow rates, one of the lines is used as a by-pass to return the gas to the suction port of the blower.

Figure 3.1 Schematic of the field-testing experimental set-up

The other portion of the gas is then directed into the reactor via the second line. In order to collect
the samples for analysis, two outlets are considered prior and after the reactor each of which is
equipped with a ball valve, a needle valve and a rotameter. SKC Tedlar bags (1L) with
Polypropylene fitting were used to collect the samples, because they proved much more accurate
D4 analysis compared to other types of sampling bags (Flexfoil Plus). The flow rate of the gas
59

collected at the inlet for sampling is measured carefully by the rotameter in order to calculate the
actual flow rate of the gas going to the reactor.
The analysis was performed at USC using a GC/MS (7890A Agilent GC and 5975C MS)
instrument. The Gas Chromatograph is equipped with Agilent GC Column 30𝑚 × 250𝜇𝑚 ×
0.25𝜇𝑚 . A gas-tight syringe (1ml 1001 Hamilton syringe) was used to inject the sample into the
GC/MS. Since the LFG contains numerous NMOC components (see Table 3.1) and the focus of
the field-testing is on the ability of the PhoR to remove the siloxanes, the Selected Ion Monitoring
(SIM) acquisition method was used to be able to selectively detect the specific siloxane
compounds. The operating settings for the analysis were: GC temperature, 30 °C rising @ 15.0
°C/min, final temperature, 120 °C; GC total flow 13.2 mL/min, split ratio 1:5, purge flow 3 mL/min
and vent flow 8.5 mL/min. MS quad and MS source temperatures are 150 °C and 230 °C,
respectively.
Four different rector configurations are being tested on the field. In our initial efforts we tested
single-lamp PhoR systems. Two different PhoR were tested, one containing a single low-pressure
18 W UV lamp (G18T5VH Atlantic Ultraviolet Corp.) and the other a single 41W UV lamp
(G36T5VH Atlantic Ultraviolet Corp). For safety reasons, in the field studies the pins at the two
ends and also the wires that connect the lamp to the electric ballast were placed outside of the main
body of the quartz reactor. Figure 3.2 shows the 18 W and 41 W PhoR used in the field-test study.

60

Figure 3.2 18 W and 41W UV lamps utilized for the field-testing study

Another proposed configuration we study (these studies are ongoing at the time this Thesis is being
submitted) use multiple 41W lamps placed around a quartz tube inside of which the LFG is
flowing. Figure 3.3 (left) shows the reactor with four lamps placed at equal distance from the
quartz tube. As our final investigation, the UV lamps are placed inside a 3 in chlorinated polyvinyl
chloride (CPVC) pipe through which the LFG is passed. Figure 3.3 (right) shows the reactor body
made of CPVC with 4 lamps placed inside the pipe. Unlike the previous reactor set-up where the
gas is not in direct contact with the UV radiation, in this configuration, the gas is in direct contact
with the lamps.

61

Figure 3.3 Reactor configuration where the lamps are placed outside of the quartz tube (a) and the case
where they are in direct contact with the gas (b)

Fig. a Fig. b
62

3.3. Results and Discussion
As noted above, for the field-test study, two of the most commonly encountered siloxanes namely
L2 and D4 were spiked into the real LFG at the site for investigating the effect of the UV reactor
has on siloxane decomposition in a real biogas stream. Figure 3.4 shows the conversion of (L2+D4)
mixtures in LFG as a function of residence time for different siloxane concentrations when an 18
W UV lamp (G18T5VH Atlantic Ultraviolet Corp.) was used inside the reactor.
Figure 3.4 L2 and D4 conversions for a 18W UV lamp reactor vs. residence time for different feed
concentrations

As can be seen in Fig 3.4, the conversion seems to be fairly independent of feed concentration, as
was also the case with the lab-scale studies, indicating a global reaction rate order very near unity.
Figure 3.5 shows the conversion of the (L2+D4) mixtures in LFG as a function of residence time
for the case of using the more powerful 41W lamp (note that volumes of these two reactors are
different from each other, specifically the volume of the PhoR using the 41W lamp is 220 cm
3

while that employing the 18 W is 87 cm
3
. Therefore, for a certain residence time the LFG flow
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80
Conversion (%)
Retention time (sec)
L2 concentration = 9.5 ppm
L2 concentration = 20 ppm
L2 concentration = 36 ppm
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80
Conversion (%)
Retention time (sec)
D4 concentration = 9.5 ppm
D4 concentration = 22 ppm
D4 concentration = 40 ppm
63

processed by the reactor with the larger lamp is almost three times greater than that of the PhoR
with the smaller lamp) As can be seen in Fig 3.4 and Fig 3.5, the conversion of L2 at a certain
residence time is generally higher than that of D4 for the same residence time. This is in line with
the experimental results obtained in the laboratory with the (L2+D4) mixtures in SLFG. Also note,
that the 41W PhoR shows higher conversion than the 18 W PhoR for the same corresponding
residence times and feed concentrations.

Figure 3.5 Field-test L2 and D4 conversions as a function of flow rates for the 41W UV reactor

Figure 3.6 shows the inlet and outlet of the 41 W UV reactor after the field-test experiments. As
can be seen in Fig 3.6, unlike the inlet where no particles were observed on the surface of the lamp
and the quartz tube, silica particles are formed on the inner surface of the quartz tube at the outlet
which is consistent with the expectation that the siloxanes decompose into silica microparticulates.
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200
Conversion (%)
Retention time (sec)
L2 concentration = 2.15ppm
L2 concentration = 9.5ppm
L2 concentration = 38ppm
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200
Conversion (%)
Retention time (sec)
D4 concentration = 2.2ppm
D4 concentration = 8ppm
D4 concentration = 40ppm
64

However, no silica particles or other contaminants are deposited on the UV lamps themselves to
reduce their intensity.

Figure 3.6 41 W UV reactor inlet (left) and outlet (right) sides

In addition to the single-lamp configurations, we are also testing two multiple-lamp configurations.
For this purpose, two configurations were designed and fabricated as can be seen in Fig. 3.3. The
experiments are currently under way for these two configurations. The results for these
experiments will be presented in subsequent publications.

65

3.4. Conclusion
As expected from the lab-scale experimental results, higher L2 conversions compared to D4 were
obtained for the field-test study as well. The main reason for this observation is the simpler L2
molecular structure unlike D4, which has a complex cyclic molecular structure. Also, more
siloxane removal is attained by employing a more powerful UV lamp (41W) for the reactor.
High conversions attained for L2 and D4 removal using the photo-decomposition technique at a
real landfill plant indicate the efficiency and advantage of this method compared to more
commonly used technologies for landfill gas clean-up. The results also prove that siloxane
compounds decomposition occurs in the presence of other NMOC compounds including the
sulfided species (e.g. H2S, COS, CS2) and halogenated ones (Vinyl Chloride).
Oxygen has an important effect on the photodecomposition process. The higher the oxygen present
in the system is, the more ozone is produced which results in higher conversion of the siloxanes
and the NMOC compounds. According to the daily test analysis performed at the Whittier Landfill
plant, there is around 3-4% oxygen in the gas. Higher siloxane conversions were obtained at the
landfill when compared to the lab experiments for similar residence times, and this most likely can
be attributed to the higher concentration of oxygen present in the real LFG (as a reminder, the
simulated landfill gas used in the lab contains only 1% oxygen as compared to ~3-4% of the real
LFG in the field).

66

Chapter 4 Scale-up Design and Economic Analysis

67

4.1 Introduction
Results from the lab-scale study and the field-testing of the UV photodecomposition reactor have
shown it to be effective for siloxane contaminants removal (with the added benefit that it also
removes other contaminants found in LFG, including halogenated and sulfided compounds). The
successful field-testing of the technology at the Whittier landfill plant, in particular, has motivated
us to perform hypothetical design calculations for a scaled-up system that consists of a cascade of
multi-lamp PhoR to be used as an alternative technique to conventional methods, specifically
adsorption which, as noted in more than occasion in this Thesis, is the most common method
currently used on landfill sites for siloxane (and other NMOC) removal.
Such a PhoR system should be able to attain high siloxane removal rates (~95%) which is typically
the required value for the commonly methods used in landfill plants. According to the test analysis
performed on most landfill plants for siloxane compounds compositions, D4 has found to be the
most frequent siloxane present in the landfill gas (e.g., Table 4-1 shows the gas analysis results for
siloxane compounds at the BENA flare station that indicates the relatively high concentration of
D4 compared to other siloxanes). Also, according to field-test and lab-scale experimental results,
lower conversion was attained for D4 compared to L2 (or L3) due to its cyclic and more recalcitrant
molecular structure. As a result, one can conclude that if a high conversion (>95%) is obtained for
D4, definitely other siloxane (particularly the linear ones) components will be decomposed as well
to a similar or higher degree). Therefore, only D4 was used as a model siloxane for the simulations
reported here.

68

Table 4.1 BENA flare station gas analysis for organic silicon NMOC compounds

As noted during the experiments, the reactor residence time plays an important role in siloxane
removal. Higher residence times (i.e., using lower flow rates) result in higher conversion for the
siloxane compounds investigated. For the lab-scale and field-test studies, the highest flow rate of
the gas mixture employed was around 1 L/min. However, the usual flow rate of the gas produced
in a landfill that must be dealt with is around 1500 SCFM. Therefore, the optimized scaled-up
system designed should be able to attain high siloxane conversion while treating these significantly
higher flow rates. Furthermore, economic analysis should be performed to validate the efficiency
of the designed system in different aspects, including the cost for building the system, the annual
power consumption, and maintenance/replacement costs. This is because for the new technology
69

to be adopted by the landfill industry its economics must be competitive with the state of the art
(SOTA) remediation technology (i.e., adsorption).

4.2 Simulation and Modeling
As noted above, in the simulations to support the design of a scaled-up PhoR system we focused
attention only on D4 (as a model siloxane) removal. We employ in the simulations the reaction
kinetic parameters for D4 conversion in simulated landfill gas computed from the lab-scale
experimental data. In the scale-up simulations, we employ a multi-lamp PhoR employing 41W
UV lamps (G36T5VH Atlantic Ultraviolet Corp.). Information about the characteristics of the lab
and of the reactor itself are shown in Table 4.2.
Because of symmetry, in the simulations of the multi-lamp PhoR we deal with only one quarter of
the whole geometry. Figure 4.1 demonstrates the mesh distribution of the geometry with one lamp
in the middle of the reactor tube. The tube size diameter is 33.5 cm (~1ft) and the length of the
reactor is equal to the length of the lamp (1.5 m). For the geometry in Fig 4.1 (employing a single
lamp) and an inlet D4 concentration of 10 ppm, the simulation was run for different LFG flow
rates. (The surface intensity of the lamp was calculated based on the UV output provided by the
Atlantic Ultraviolet Corp. divided by the surface area of the lamp). Figure 4.2 shows the conversion
of D4 as function of the flow rate.

70

Figure 4.1 Mesh distribution of geometry with one lamp inside a 33.5 ID pipe

Table 4.2 Pipe and Lamp specifications used for simulation
Specifications Values
Reactor Length (cm) 141.2
Lamp Diameter (cm) 1.5
Inner pipe size diameter (cm) 33.5
Reactor Volume (cc) 124000
Intensity of the lamp (W/m
2
) 800
Wall reflection (%) 90

71

Figure 4.2 D4 conversion as function of flow rate

Figure 4.3 shows the contours of the D4 mole fraction and the intensity radiation on a surface
normal to the UV lamp for a flow rate of 2695 cc/s (5.7 scfm).
Figure 4.3 Contours of the D4 mole fraction (left) and incident radiation (right) on surface normal to lamp
We then carried out additional simulations in which the diameter of the tube was kept constant and
the number of lamps was increased (up to 169 lamps for that particular reactor size tube). Figure
4.4 shows the mesh distribution of the corresponding geometry.
0
10
20
30
40
50
60
70
80
90
100
0 1000 2000 3000 4000 5000 6000 7000
Conversion (%)
Flow rate (cc/s)
72

Figure 4.4 Mesh distribution for the geometry with 169 lamps
There are two factors one needs to consider when attempting to optimize the number of lamps
employed inside a given size PhoR: (i) increasing the number of lamps inside the pipe, would
result in higher conversion due to the increase in incident radiation inside the reactor; (ii) on the
other hand, increasing the number of lamps would decrease the reactor volume available to the
gas, and thus the residence time for the gas mixture would decrease, which in turn would result in
D4 conversion reduction.. Figure 4.5 shows the results of such an optimization exercise whereby
for a given PhoR diameter, we calculate the flow rate per lamp required to attain a certain
conversion (in this case 40%) as a function of total number of lamps placed inside the reactor.
73

Figure 4.5 Flow rate per lamp as function of number of lamps to attain a given conversion

According to Fig 4.5, the best performance for this size PhoR can be attained using only one lamp
placed in the center of the pipe. Placing a higher number of lamps inside the PhoR has an obvious
negative impact on performance. Such calculations as those shown in Fig. 4.5 must then be
repeated with successively larger tube dimeters in order to establish a global performance
optimum.

0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0 20 40 60 80 100 120 140 160 180
Flow rate/lamps
number of lamps
74

4.3 Scale-up and Design
In order to obtain 95% conversion for treating landfill gas with a flow rate of 1500 scfm, we
propose an integrated system of PhoR placed in parallel and in series. Based on the simulation and
optimization studies outlined above, we have concluded that the proposed system should have four
stages with each stage attaining 50% conversion. For illustrative purposes, Table 4.3 shows the
siloxane conversion attained at each stage, assuming a feed siloxane concentration of 1 ppm.
Table 4.3 Siloxane concentrations and total conversion in each stage
Stage number Siloxane concentration at the
end of each stage (ppm)
Total Conversion at end of
each stage (%)
1 0.5 50
2 0.25 75
3 0.125 87.5
4 0.0625 94

Each single stage consists of nine reactors placed in parallel, with each PhoR containing 11 UV
lamps (therefore, the total flow rate of the gas to the reactor is divided into nine separate streams,
with each reactor receiving around 150 scfm of the gas). Figure 4.6 shows the schematic of the
proposed scale-up design. Each PhoR is made from a high-density polyethylene (HDPE) pipe (48
in OD and 45 in ID) which is 1.5 m long. The inner side of the pipe facing the lamps is assumed
to be covered by heavy duty aluminum foil in order to reflect the radiation and prevent it from
being lost (and also potentially causing health impacts for landfill personnel). The two ends of the
pipe are sealed with flat (either HDPE or steel) plates. In order to be able to readily access the
75

inner space of each reactor, one plate is bolted to the reactor pipe while the other end is permanently
welded.

Figure 4.6 Schematic of the scaled-up system

As Figure 4.6 shows, the overall system for landfill gas clean-up contains four such stages in series,
for a total of 36 multi-lamp PhoR. We envision that each stage will be mounted and run on a
separate process skid with a dimension of 12 ft wide × 30 ft long ×9 ft tall. HDPE pipes 4 in and
12 in diameter (OD) are used as distribution inlet/outlet pipes and as the main inlet and outlet,
conduits, respectively. Table 4.4 further summarizes the specifications of the lamps, reactors and
individual stages for the process.

76

Table 4.4 Scaled-up system specifications
Number of stages 4 Area per reactor (square feet) 11.0044
Number of reactors per stage 9 Area per stage (square feet) 100
Number of reactors in total 36 Reactor Volume (cubic feet) 492.126
Number of lamps per reactor 11 Total Volume (cubic feet) 1968.5
Number of lamps per stage 99 Total length of the reactor (feet) 19.685
Number of lamps in total 396 Residence time (sec) 78.74

Cost estimation and economic analysis is performed here to investigate costs for the process and
to compare such costs to the conventional methods currently being used for landfill gas clean up.
Specifically, the comparison has been made between the proposed scale-up design and the
adsorption method which is considered as the most common technique employed for the LFG
remediation process. The comparison is made based on fabrication and start-up costs, annual
power consumption, maintenance and replacements cost, and finally the quantity of LFG lost for
each method. Table 4.5 shows the costs for the LFG clean up using UV photodecomposition
method as well as the corresponding costs for the adsorption method.

77

Table 4.5 Costs for the UV photodecomposition and adsorption methods
UV System Adsorption System
System cost $850,000 $750,000
Annual kwh required (kwh/yr) 402,960 660,000
Annual methane loss (btu/yr) 0 13,000,000
Cost of power (per kwh) $0.15 $0.15
Value of gas (per mmbtu) $10.00 $10.00
Annual power cost $58,000 $100,000
Annual maintenance and replacement $55,000 $20,000
Total operating cost $105,000 $120,000

In the cost estimation shown in the Table, each PhoR (the lamps not included) is assumed to cost
~$20,000. Based on the number of PhoR required (36), the total cost would be ~$720,000. To that
one should add the lamps cost, which is estimated to be ~$40,000. For the adsorption system the
estimated capital cost is calculated to be ~$750,000, according to our industrial partner ES
Engineering Services. Their system consist of two treatment vessels, each filled with a treatment
media to clean-up the gas. In their adsorption system, regeneration is performed using temperature
swing adsorption (TSA) (one vessel is online while the other regenerates, typically, once per day).
During regeneration, the gas is heated using an electric heater and circulated through the
regenerating vessel with a multistage blower. As the media is heated a small stream of the gas and
contaminants is sent to a flare to be destroyed. After the vessel is regenerated, it is cooled down
and placed back online.
78

The system capital cost including design, fabrication and start-up for these two methods are fairly
close to each other, as Table 4.5 indicates. Assuming that each UV lamp consumes an average of
~115 W, the annual power cost for the PhoR system would be ~$58,000 (based on a price of $0.15
per kwh). On the other hand, the power cost for the adsorption system would be rather high due to
the heater for the regeneration process. For such a system, an 180 kW heater is used at
approximately 80% capacity, for 12 hours a day in addition to a 60 hp blower running for the same
process. Therefore, the annual power cost for this system (assuming a cost of $0.15 per kwh) would
be ~$100,000, which is around twice the power cost for the PhoR system.
Other cost for the UV system includes replacing any malfunctioning UV lamps and cleaning the
PhoR inner sides that maybe covered by the silica particles. Each UV lamp has a minimum life-
time of 10,000 hours (maximum life time is 13,000 hours) provided by Atlantic Ultraviolet Corp.
Taking into account the worst case scenario, which is that all the lamps should be replaced after
10,000 hours (~ 1 year), the replacement of the lamps and cleaning of the reactors would cost
approximately $55,000. However, if all the lamps would work for 13,000 hours, the annual cost
for the UV system would decrease to ~$40,000. For the adsorption process, the main cost would
be the bed replacement which normally occurs every 5 years. The total cost for the bed replacement
and maintenance would be ~$20,000.
Also, when utilizing a UV system, there will be no need to flare the gas which results in no methane
loss during the process. However, in the adsorption method, the gas is being flared during the
regeneration process. Although the methane loss is economically small, the siloxane compounds
being directed to the flare and being released to the environment as silica particulates could
potentially endanger the human health and the environment.
79

So in summary, considering the total cost for these two systems, one can state that the UV system
is more efficient economically compared to adsorption method which is currently used on the
landfill plants.

4.4 Conclusion
The simulation results show that although the number of lamps is an important factor in order to
attain high conversion, it decreases the volume of the reactor available to the gas which results in
conversion reduction by lowering the residence time for the landfill gas in the reactor. The
proposed scaled-up design, therefore, is based on the optimization of the number of lamps required
performed via modeling and simulation.
We have presented in this Chapter costs for the PhoR and adsorption systems which are based on
the simulation results, quotes from various vendors and data provided by ES Engineering Services,
our industrial partner. Considering the total cost including the annual power consumption and
maintenance costs, both methods are rather similar, but the UV system is somewhat more efficient.
The advantage of adsorption system is that 99% siloxane removal can be attained compared to
94% siloxane removal with UV reactor. On the other hand, the siloxanes removed by the
adsorption system must be flared, potentially posing significant environmental hazards (it should
be noted that ~95% siloxane removal would satisfy present regulatory limits for most US landfills).
The main reason for the UV system to have a high annual cost is the power that each lamp requires.
The UV system would be quite more preferable if it required less energy. That could be
accomplished if the conventional germicidal UV lamps presently in use could be replaced by UV
LED lamps which are much more efficient in power consumption. Table 4.6 provided by
80

Aquisense Technologies Corp. shows the comparison of current UVC bulb with the new UVC
LED lamps. As it can be seen, the power required for the LED lamp is around 8 times less than of
the mercury type UV lamp. The other advantage of UVC LED lamp is preventing the use of
mercury inside the bulbs which is highly dangerous in case of lamp break. So such lamps, if
proven effective in siloxane decomposition, would offer a great advantage for the application of
the PhoR technology for LFG clean-up.
Table 4.6 Specifications of UVC LED bulb vs. Mercury lamps
Attribute Conventional Mercury Lamp UV-C LED Product Implication
Mercury Content 20-200 mg None Safe disposal - no special
handling
Lifetime 5,000 - 15,000 hours 10,000 hours Flexible operation
On/Off Cycles Max. 4 per day Unlimited Intermittent-flow friendly
Warm-up Time Up-to 15 minutes Instantaneous Extended replacement
intervals
Operating Surface
Temp.
100-600° C Same as process
water
Zero-flow friendly does not
promote fouling
Architecture Cylindrical tube Point source Versatile implementation
Durability Fragile glass tube Rugged
semiconductor
Versatile operation
Wavelength Polychromatic (200-300nm)
Monochromatic (254 nm)
Selectable (250-300
nm)
No wasted energy & targeted
performance
Power Supply 110-240V AC 6-30V DC Battery/Solar option

81

Chapter 5 Ideas for Future Work

82

The experimental results from the study of the PhoR system both in the laboratory but also on the
field show that this technology is quite effective for siloxane compound removal but also helps to
remove other model NMOC compounds. The other method that was previously tried in our
research group was the use of a flow-through catalytic membrane reactor (FTCMR) technique. It
was also shown promising to remove various NMOC components in simulated LFG. Figure 5.1
shows, for example, recent experiments employing this method to convert various VOC
compounds in such simulated LFG. As can be seen, total removal occurred for all the VOC
compounds at ~300 ºC for these lab-scale experiment. A combined system of the PhoR and
FTCMR should, therefore, be field-tested to investigate its performance under real conditions and
to examine the added advantages that the combined system should be able to offer. .

83

Figure 5.1 Conversion of VOC compounds using FTCMR technique

0
20
40
60
80
100
120
0 100 200 300 400
Conversion (%)
Temperature ºC
Carbonyl Sulfide
0
20
40
60
80
100
120
0 100 200 300 400
Conversion (%)
Temperature ºC
Vinyl Chloride
0
20
40
60
80
100
120
0 100 200 300 400
Conversion (%)
Temperature ºC
Trichlorofluoromethane
0
20
40
60
80
100
120
0 100 200 300 400
Conversion (%)
Temperature ºC
Dimethylsulfide
0
20
40
60
80
100
120
0 100 200 300 400
Conversion (%)
Temperature ºC
1,3-Dichlorobenzene
84

References

 Abatzoglou, N., Boivin, S., A review of biogas purification processes, Biofuels
Bioproducts and biorefining, Volume 3, Issue 1, January/February 2009
 Aguilar-Virgen, Q., Taboada-Gonzalez, P., Ojeda-Benítez, S., “Analysis of the feasibility
of the recovery of landfill gas: a case study of Mexico,” Journal of Cleaner production 79,
pp. 53-60, 2014.
 Ajhar , M., Bannwarth, S., Stollenwerk, K-H, Spalding, G., Yüce, S., Wessling,
M., Melin, T., Siloxane removal using silicone–rubber membranes, Separation and
Purification Technology, Volume 89, Pages 234–244, 2012
 Ajhar, M., Travesset, M., Yüce, S., and Melin, T., “Siloxane removal from landfill and
digester gas – A technology overview,” Bioresource Technol., vol. 101, pp. 2913-2923,
2010.
 Ajhar, M., Melin, T., Siloxane removal with gas permeation membranes, Desalination,
Volume 200, Issue 1, 2006
 Amini, H.R., and Reinhart, D.R., “Regional prediction of long-term landfill gas to energy
potential,” Waste Manage., vol. 31, pp. 2020-2026, 2011.
 Andriani, D., Wresta, A., Atmaja, T. D., Saepudin, A., “A Review on Optimization
Production and Upgrading Biogas Through CO2 Removal Using Various Techniques,”
Appl Biochem Biotechnol (2014) 172:1909–1928.
 ANSYS® Fluent, Discrete Ordinate Radiation model theory,
http://aerojet.engr.ucdavis.edu/fluenthelp/html/ug/node580.htm
85

 Appels, L., Baeyens, J., Dewil, R., Siloxane removal from biosolids by peroxidation,
Energy Conversion and Management, Volume 49, Issue 10, 2008
 Augenstein, D., and Pacey, J., “Landfill Gas Energy Utilization: Technology Options and
Case Studies,” U.S. Environmental Protection Agency, Washington, D.C., Rep.
EPA/600/R-92/116 (NTIS PB92-203116), 1992.
 Augenstein, D., The greenhouse effect and US landfill methane, Global Environmental
Change, Volume 2, Issue 4, 1992
 Baciocchi, R., Carnevale, E., Corti, A., Costa, G., Lombardi, L., Olivieri, T., Zanchi, L., &
Zingaretti, D. (2012). Innovative process for biogas upgrading with CO2 storage: results
from pilot plant operation. Biomass and Bioenergy, 1–10.
 Baker, L., Capouya, R., Cenci, C., Crooks, R., and Hwang, R., “The Landfill Testing
Program; Data Analysis and Evaluation Guidelines”, CAPCOA/CARB Report, Sept 1990.
 Bhowmick, M., Semmens, M.J., Ultraviolet Photooxidation for the destruction of VOCs in
air, Wat. Res. Vol. 28, No. 11, pp. 2407 2415, 1994
 Bolton, J.R., Ultraviolet Applications Handbook, third edition, Bolton Photosciences 2013
 Brinkmann M., Chan V.Z.-H., Thomas E.L., Lee V.Y., Miller R.D., Hadjichristidis N., and
Avgeropoulos 2001, Room-Temperature Synthesis of a-SiO2 Thin Films by UV-Assisted
Ozonlysis of a Polymer Precursor, Chem.Mater., 13, 967-972, 2001
 Cheng, Z.-W., Feng, L., Chen, J.-M., Yu, J.-M., Jiang, Y.-F., “Photocatalytic conversion
of gaseous ethylbenzene on lanthanum-doped titanium-dioxide nanotubes,” J. Hazard.
Mater., vols. 254-255, pp. 354-363, 2013.
 Dalton J.C. 1985, Organic Photochemistry, ed. Padwa A., Dekker M., New York, vol 7,
ch3. 1985
86

 Dalton, J.C., “Silane Photochemistry,” in Organic Photochemistry, vol. 7, A. Padwa, Ed.,
New York, NY: Marcel Dekker, Inc., 1985, ch. 3.
 Dewil, R., Appels, L., Baeyens, J., Energy use of biogas hampered by the presence of
siloxanes, Energy Conversion and Management, Volume 47, Issue 13, 2006
 Dow Corning, An Overview of Volatile Methylsiloxane (VMS) Fluids in the Environment,
Environmental Information Update, September 1997
 Dow Corning, Basic Silicon Production and Siloxane Polymerization
www.dowcorning.com/content/sitech/sitechbasics/siloxane_polymerization.asp
 Dow Corning. 1997, Environmental information updates. “An overview of volatile
Methylsiloxane (VMS) fluids in the environment”. 1997
 El-Fadel, M., Findikakis, A.N., Leckie, J.O., Environmental Impacts of Solid Waste
Landfilling, Journal of Environmental Management, Volume 50, Issue 1, May 1997
 EPA (2014), Overview of greenhouse gases, Accessed: 2014
 EPA (2015), Green power from Landfill Gas, Accessed: May 2015
 Feiyan, C., Pehkonen, S.O., Ray, M.B., Kinetics and mechanisms of UV-photodegradation
of chlorinated organics in the gas phase, Water Research 36 (2002) 4203–4214
 Ghorbel, L., Tatin, R., and Couvert, A., “Relevance of an organic solvent for absorption of
siloxanes,” Environ. Technol., vol. 35, pp. 372-382, 2014.
 Gomes, V. G., & Hassan, M. M. (2001). Coalseam methane recovery by vacuum swing
adsorption. SeparPurifTechn, 24, 189–196.
 Havukainen, J., Uusitalo, V., Niskanen, A., Kapustina, V., and Horttanainen, M.,
“Evaluation of methods for estimating energy performance of biogas production,” Renew.
Energ., vol. 66, pp. 232-240, 2014.
87

 He, C., Herman, D., Minet, R.G., and Tsotsis, T.T., “A Catalytic/Sorption Hybrid Process
for Landfill Gas Clean-up,” Ind. Eng. Chem. Res., 36, 4100, 1997.
 Ho, C.K., Khalsa, S.S., Computational Fluid Dynamics Based Models for Assessing UV
Reactor Design and Installation, Water research foundation, Sandia National Laboratories,
2011
 Ho, C.K., Radiation Dose Modeling in FLUENT, Sandia National Laboratories, 2009
http://www.sandia.gov/cfd-water/uvdisinfection.htm
 Ho, K.-L., Lin, W.-C., Chung, Y.-C., Chen Y.-P., and Tseng, C.-P., “Elimination of high
concentration hydrogen sulfide and biogas purification by chemical-biological process,”
Chemosphere, vol. 92, pp. 1396-1401, 2013.
 http://www3.epa.gov/climatechange/ghgemissions/gases/ch4.html,
 http://www3.epa.gov/lmop/documents/pdfs/green_power_from_landfill_gas.pdf ,
 Huang H., and Li, W., “Destruction of toluene by ozone-enhanced photocatalysis:
Performance and mechanism,” Appl. Catal. B-Environ., vol. 102, pp. 449-453, 2011.
 Hussain, M., Russo, N., and Saracco, G., “Photocatalytic abatement of VOCs by novel
optimized TiO2 nanoparticles,” Chem. Eng. J., vol. 166, pp. 138-149, 2011.
 Jafari, T., Jiang, T., Zhong, W., Khakpash, N., Deljoo, B., Aindow, M., Singh, P., and
Suib, S.L., “Modified Mesoporous Silica for Efficient Siloxane Capture,” Langmuir, vol.
32, pp. 2369-2377, 2016.
 Jalali, A., Motamedhashemi, M.M.Y., Egolfopolous, F., and Tsotsis, T.T., “Fate of
Siloxane Impurities during the Combustion of Renewable Natural Gas,” Combust. Sci.
Technol., vol. 185, pp. 953-974, 2013.
88

 Kockarts, G., “Absorption and Photodissociation in the Schumann-Runge Bands of
Molecular Oxygen in the Terrestrial Atmosphere,” Planet. Space Sci., vol. 24, pp. 589-
604, 1976.
 Koyer-Gołkowska, A., Musialik-Piotrowska, A., and Rutkowski, J. D., “Oxidation of
Chlorinated Hydrocarbons over Pt–Pd-based Catalyst: Part 1. Chlorinated Methanes”,
Catalysis Today, 90, 133, 2004.
 Lakhouit, A., Cabral, A.R., and Cabana, H., “Two Novel Biofilters to Remove Volatile
Organic Compounds Emitted by Landfill Sites,” Water Air Soil Pollut., vol. 227, 2016.
 Läntelä, J., Rasi, S., Lehtinen, J., & Rintala, J. (2012). Landfill gas upgrading with pilot-
scale water scrubber: performance assessment with absorption water recycling. Applied
energy, 92, 307–314
 Li, X., Zhang, G., and Pan, H., “Experimental study on ozone photolytic and photocatalytic
degradation of H2S using continuous flow model,” J. Hazard. Mater., vols. 199-200, pp.
255-261, 2012.
 Liang, W., Li, J., and Jin, Y., “Photo-catalytic degradation of gaseous formaldehyde by
TiO2/UV, Ag/TiO2/UV, and Ce/TiO2/UV,” Build. Environ., vol. 51, pp. 345-350, 2012.
 Liang, W.J., Li, J., and Jin, Y.Q., “Photocatalytic degradation of gaseous acetone, toluene,
and p-xylene using a TiO2 thin film,” J. Environ. Sci. Health A Tox. Hazard. Subst.
Environ. Eng., vol. 45, pp. 1384-1390, 2010.
 Mahajan, D., Tonjes, D.J., Mamalis, S., Boudreaux, R., Hasty, J.K., Danhui, X., Youcai,
Z., Jianglin, C., Wentao, Z., and Xiaoli, C., “Effective landfill gas management strategies
for methane control and reuse technology,” J. Renew. Sustain. Ener., vol. 7, 2015.
89

 Mazzarino, I., and Barresi, A.A., “Catalytic Combustion of VOC Mixtures in a Monolithic
Reactor”, Catal. Today, 17, 335, 1993.
 McBean, E. A., Siloxanes in biogases from landfills and wastewater digesters, Canadian
Journal of Civil Engineering, Volume 35, Issue 4, 2008
 Montanari, T., Finocchio, E., Bozzano, I., Garuti, G., Giordano, A., Pistarino, C., and
Busca, G., “Purification of landfill biogases from siloxanes by adsorption: a study of silica
and 13X zeolite adsorbents on hexamethylcyclotrisiloxane separation,” Chem. Eng. J., vol.
165, pp. 859-863, 2010.
 Montanari, T., Finocchio, E., Salvatore, E., Garuti, G., Giordano, A., Pistarino, C., &
Busca, G. (2011). CO2 separation and landfill biogas upgrading: a comparison of 4A and
13X zeolite adsorbents. Energy, 36, 314– 319.
 Moulis, F., and Krysa, J., “Photocatalytic degradation of several VOCs (n-hexane, n-butyl
acetate and toluene) on TiO2 layer in a closed-loop reactor,” Catal. Today, vol. 209, pp.
153-158, 2013.
 Nair, N., Novel Methods for Landfill Gas and Biogas Clean-up, PhD dissertation, 2013
 Nair, N., Vas, A., Zhu, T., Sun, W., Gutierrez, J., Chen, J., Egolfopoulos, F., Tsotsis, T.T.,
Effect of Siloxanes Contained in Natural Gas on the Operation of a Residential Furnace,
52 (18), pp 6253–6261, 2013
 Nair, N., Zhang, X., Gutierrez, J., Chen, J., Egolfopoulos, F., Tsotsis, T.T., Impact of
Siloxane Impurities on the Performance of an Engine Operating on Renewable Natural
Gas, Industrial & Engineering Chemistry Research, 51 (48), pp 15786–15795, 2012
90

 Nam, S., Namkoong, W., Kang, J-H, Park, J-K. Adsorption characteristics of siloxanes in
landfill gas by the adsorption equilibrium test, Waste Management, Volume 33, Issue 10,
Pages 2091–2098, 2013
 Neti, N.R., Parmar, G.R., Bakardjieva, S., and Šubrt, J.,“Thick film titania on glass
supports for vapor phase photocatalytic degradation of toluene, acetone, and ethanol,”
Chem. Eng. J., vol. 163, pp. 219-229, 2010.
 NIST, Meeting airborne molecular contamination challenges in laser pattern generators,
http://micromagazine.fabtech.org/archive/04/01/ekberg.html
 Nizami, A.S., Shahzad, K., Rehan, M., Ouda, O.K.M., Khan, M.Z., Ismail, I.M.I.,
Almeelbi, T., Basahi, J.M., Demirbas, A., “Developing waste biorefinery in Makkah: A
way forward to convert urban waste into renewable energy,” Applied Energy, Vol. 186,
pp. 189-196, 2017.
 Ohannessian, A., Desjardin, V., Chatain, V., Germain, P., Volatile organic silicon
compounds: the most undesirable contaminants in biogases, Water science and technology
: a journal of the International Association on Water Pollution Research, Volume 58, Issue
9, 2008
 Ortega, R., Subrenat, A., Siloxane treatment by adsorption into porous materials,
Environmental Technology, 09/2009, Volume 30, Issue 10, No. 10, September 2009
 Ouchi, A., Koga, Y., Bastl, Z., and Pola, J., “UV-Laser-Induced Photolysis of
Trimethyl(vinyloxy)silane for Chemical Vapour Deposition of Polysiloxane Films,” Appl.
Organometal. Chem., vol. 13, pp. 643-647, 1999.
91

 Ouyang M., Yuan C., Muisener R.J., Boulares A., and Koberstein J.T. 2000, Conversion
of Some Siloxane Polymers to Silicon Oxide by UV/Ozone photochemical Processes,
Chem. Mater., 12, 1591-1596. 2000
 Pacey, J., Doorn, M., and Thorneloe, S., “Landfill Gas Energy Utilization: Technical and
Non-Technical Considerations,” U.S. Environmental Protection Agency, Washington,
D.C., Rep. EPA/600/A-94/139.
 Pengyi, Z., Fuyan, L., Gang, Y., Qing, C., and Wanpeng, Z., “A comparative study on
decomposition of gaseous toluene by O3/UV, TiO2/UV, and O3/TiO2/UV,” J. Photoch,
Photobio. A, vol. 156, pp. 189-194, 2003.
 Petersson, A., & Wellinger, A. (2009). Biogas upgrading technologies—developments and
innovations. IEA Bioenergy
 Pola J., Galikova A., Galik, A., Blechta V., Bastl Z., Subrt J., and Ouchi A. 2002, UV Laser
Photolysis of Disiloxanes for Chemical Vapor Deposition of Nano-Textured Silicones,
Chem. Mater., 14, 144-153. 2002
 Popat, S. C., Deshusses, M. A., Biological removal of siloxanes from landfill and digester
gases: opportunities and challenges, Environmental science & technology, Volume 42,
Issue 22, 2008
 Powell, J.T., Townsend, T.G., Zimmerman, J.B., “Estimates of solid waste disposal rates
and reduction targets for landfill gas emissions,” Nature climate change,
DOI:10.1038/NCLIMATE2804, 2015.
 Pradhan, B.K., Sandle, N.K. Effect of different oxidizing agent treatments on the surface
properties of activated carbons, Carbon, Volume 37, Issue 8, Pages 1323–1332, 1999
92

 Rasi, S., Läntelä, J., and Rintala, J., “Upgrading landfill gas using a high pressure water
absorption process,” Fuel, vol. 115, pp. 539-543, 2014.
 Rojas Devia, C., Subrenat, A., Absorption of a linear (L2) and a cyclic (D4) siloxane using
different oils: application to biogas treatment, Environmental technology, Volume 34,
Issue 21-24, 2013
 Rothman, L.S., et al., “The HITRAN 2008 molecular spectroscopic database,” Journal of
Quantitative Spectroscopy and Radiative Transfer, V100, 533-572, 2009.
 Ryckebosch, E., Drouillon, M., & Vervaeren, H. (2011). Techniques for transformation of
biogas to biomethane. Biomass and Bioenergy, 35, 1633–1645
 Schmidt, J.A., Johnson, M.S., Schinke, R., Carbon dioxide photolysis from 150 to 210 nm:
Singlet and triplet channel dynamics, UV-spectrum, and isotope effects, Proceedings of the
National Academy of Sciences of the United States of America, Volume 110, Issue 44,
2013
 Scholz, M., Melin, T., & Wessling, M. (2013). Transforming biogas into biomethane using
membrane technology. Renewable and Sustainable Energy Reviews, 17, 199–212
 Schweigkofler, M., Niessner, R., Removal of siloxanes in biogases, Journal of Hazardous
Materials, Volume 83, Issue 3, Pages 183–196, 2001
 Shin, H.-C., Park, J.-W., Park, K., and Song, H.-C., “Removal characteristics of trace
compounds of landfill gas by activated carbon adsorption,” Environ. Poll., vol. 119, pp.
227-236, 2002.
 Soreanu, G., “Insights into siloxane removal from biogas in biotrickling filters via process
mapping-based analysis,” Chemosphere, vol. 146, pp. 539-546, 2016.
93

 Thorneloe, S.A., “U.S. EPA’s Field Test Programs to Update Data on Landfill Gas
Emissions,” presented at the 9
th
International Waste Management and Landfill Conference,
Sardinia, Italy, 2003.
 Tippayawong, N., & Thanompongchart, P. (2010). Biogas quality upgrade by simultaneous
removal CO2 and H2S in a packed column reactor. Energy, 35, 4531–4535.
 U.S. Environmental Protection Agency, “Inventory of U.S. Greenhouse Gas, Emission and
Sinks:1990-2014”, 2016.
 Urban, W., Lohmann, H., Salazar Gómez , J.I., Catalytically upgraded landfill gas as a
cost-effective alternative for fuel cells, Journal of Power Sources, Volume 193, Issue
1,Pages 359–366, 2009
 Urbanova M., Bastl Z., Subrt J. and Pola J. 2001, IR laser-induced thermolysis and UV
laser-induced photolysis of 1,3-diethyldisiloxane: chemical vapor deposition of
nanotextured hydridoalkylsilicones. Journal of Materials Chemistry, 11, 1557-1562. 2001
 Wang, J., Zhang, W., Xu, J., Li, Y., and Xu, X., “Octamethylcyclotetrasiloxane removal
using an isolated bacterial strain in the biotrickling filter,” Biochem. Eng. J., vol.91, pp.
46-52, 2014.
 Wheless, E. and Gary, D., 2002, “Siloxanes in Landfill and Digester Gas,” Proceedings
from the Solid Waste Association of North America’s 25th Annual Landfill Gas
Symposium, Publication #GR-LG-00325, pp. 29-41. 2002
 Yan, L., Liu, J., Feng, Z., & Zhao, P., “Continuous degradation of BTEX in landfill gas by
the UV-Fenton reaction,” RSC Advances, 6(2), 1452-1459, 2016.
 Yingling, Y.G., and Garrison, B.J., “Photochemical ablation of organic solids,” Nucl. Instr.
and Meth. in Phys. Res. B, vol. 202, pp. 188-194, 2003.
94

 Zhang, L., and Anderson, W.A., “Kinetic Analysis of the photochemical decomposition of
gas-phase chlorobenzene in a UV reactor: Quantum yield and photonic efficiency,” Chem.
Eng. J., vol. 218, pp. 247-252, 2013.
 Zhao, Q., Leonhardt, E., MacConnell, C., Frear, C., & Chen, S. (2010). Purification
technologies for biogas generated by anaerobic digestion. CSANR Research Report
 Zhuo-wei, C., Peng-fei, S., Yi-feng, J., Jian-ming, Y., and Jian-meng, C., “Ozone-assisted
UV254 nm photodegradation of gaseous ethylbenzene and chlorobenzene: Effects of process
parameters, degradation pathways, and kinetic analysis,” Chem. Eng. J., vol. 228, pp. 1003-
1010, 2013.

Abstract (if available)

Abstract In this study, the feasibility of using a photochemical process for landfill gas (LFG) clean-up has been investigated. The process involves the use of an ultraviolet (UV) photodecomposition reactor (PhoR) that is primarily employed for removing Si-containing trace compounds known as siloxanes frequently found in LFG, but also happens to simultaneously remove a number of other contaminants in LFG known as non-methane organic compounds or NMOCs. The study includes both lab-scale experiments as well as a field-test of the process at a landfill located in the City of Whittier. In the lab-scale experiments, the effectiveness of PhoR in siloxane removal was investigated in both air and simulated landfill gas with (SLFGV) and without (SLFG) other NMOC being present. High siloxanes conversions were attained using the UV reactor that was shown capable to convert the siloxanes into silica (SiO2) particles. A reactor model was also developed that properly accounts for the transport and reaction phenomena that take place. The model was used to fit the lab-scale experimental data. A commercial CFD software, specifically, the ANSYS Fluent package was utilized together with companion codes written in other programming languages (Matlab and C) in order to analyze the model and to fit the experimental data. The data-validated was subsequently to design the reactors for the field-test study and for hypothetical process design and economic evaluation. The focus of the field-test was to evaluate the ability of the PhoR to function properly in the presence of real LFG, which turned out to be the case. Another goal was to test a number of different reactor configurations.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Novel methods for landfill gas and biogas clean-up

PDF

The use of carbon molecule sieve and Pd membranes for conventional and reactive applications

PDF

Process intensification in hydrogen production via membrane-based reactive separations

PDF

A high efficiency, ultra-compact process for pre-combustion CO₂ capture

PDF

A flow-through membrane reactor for destruction of a chemical warfare simulant

PDF

Performance prediction, state estimation and production optimization of a landfill

PDF

Dynamic modeling and optimization of a landfill

PDF

The purification of contaminated air streams via recative separation techniques

PDF

Flame ignition studies of conventional and alternative jet fuels and surrogate components

PDF

On the use of membrane reactors in biomass utilization

PDF

Experimental and kinetic modeling studies of flames of H₂, CO, and C₁-C₄ hydrocarbons

PDF

CFD design of jet-stirred chambers for turbulent flame and chemical kinetics experiments

PDF

Biogas reforming: conventional and reactive separation processes and the preparation and characterization of related materials

PDF

A study of the application of membrane-based reactive separation to the carbon dioxide methanation

PDF

Studies of Swiss-roll combustors for incineration and reforming applications

PDF

Investigations of fuel effects on turbulent premixed jet flames

PDF

Studies of combustion characteristics of heavy hydrocarbons in simple and complex flows

PDF

Experimental studies of high pressure combustion using spherically expanding flames

PDF

Accuracy and feasibility of combustion studies under engine relevant conditions

PDF

Methanol synthesis in a contactor-type membrane reactor

Asset Metadata

Creator Divsalar, Alireza (author)

Core Title Lab-scale and field-scale study of siloxane contaminants removal from landfill gas

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Chemical Engineering

Publication Date 09/26/2017

Defense Date 08/30/2017

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag landfill gas,OAI-PMH Harvest,process design,reactor modeling,siloxane,UV photodecomposition

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Tsotsis, Theodore (committee chair), Egolfopoulos, Fokion (committee member), Shing, Katherine (committee member)

Creator Email alireza.divsalar@gmail.com,divsalar@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-436805

Unique identifier UC11264107

Identifier etd-DivsalarAl-5751.pdf (filename),usctheses-c40-436805 (legacy record id)

Legacy Identifier etd-DivsalarAl-5751.pdf

Dmrecord 436805

Document Type Dissertation

Rights Divsalar, Alireza

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA