The Effect of Operating Parameters on the Efficiency of an Industrial Unit to Remove Mercury Vapors from Natural Gas A thesis submitted to the… [301786]
Alexandria University
Faculty of Engineering
A Study on
The Effect of Operating Parameters on the Efficiency of an Industrial Unit to Remove Mercury Vapors from Natural Gas
A thesis submitted to the Chemical Engineering Department for partial fulfillment of the requirement
for the Degree of Master of Science in Chemical Engineering
By
Eng. Ahmed Mohamed Mohsen Abd Allah
2016
Supervisors
Prof. Dr. [anonimizat]-[anonimizat]. Prof. [anonimizat] – Pharos University
جامعة الاسكندرية
كلية الهندسة
دراسة في تأثير عوامل التشغيل على كفاءة وحدة صناعية لازالة بخار الزئبق من الغاز الطبيعي
رسالة مقدمة من المهندس/ احمد محمد محسن عبدالله
إلى قسم الهندسة الكيميائية كمتطلب جزئى للحصول
على درجة الماجستير فى الهندسة الكيميائية
2016
المشــــرفون
أ.د. حسن عبد المنعم فرج
أستاذ متفرغ – قسم الهندسة الكيميائية
كلية الهندسة – جامعة الاسكندرية
أ.م.د. رانيا فاروق سلامة
أستاذ مساعد – قسم الهندسة الكيميائية
كلية الهندسة – جامعة فاروس
[anonimizat] I had finished this thesis only by his guidance and will.
I wish to express my deepest appreciation and gratitude to my thesis supervisors:
Prof. Dr. [anonimizat]-Moneim Farag
Asst. Prof. [anonimizat], [anonimizat]. Their comments and suggestions not only provided valuable knowledge but broadened my perspective in practical applications.
I wish to express my sincere appreciation to my work management and my colleagues in operation department for their support and encouragement.
Also, I [anonimizat].
Finally, my thanks are also to everyone who helped me in this work.
List of figures
Figure 1-1 Mercury Electron Configuration 3
Figure 2-1 Mercury Environmental Cycle 7
Figure 2-2 Mercury Releases and Control Measures 11
Figure 3-1 Shape of Mercury Corrosion and Failure 16
Figure 4-1 Pore Volume and Pore Size Distribution of Adsorbents 19
Figure 4-2 Brunauer Classification for Equilibrium 21
Figure 4-3 Adsorption Isotherms and Mathematical Forms. 22
Figure 4-4 Schematic Diagram of Porous Adsorbent Particle 25
Figure 4-5 General Differential Mass Balance Equation over Bed Element 26
Figure 4-6 Categories of Equilibrium Isotherms 27
Figure 4-7 Configuration of Regenerative Mercury Removal System 37
Figure 4-8 Concentration Fronts through Columns and Zones Classification 39
Figure 4‑9 Flow Diagram of Natural Gas Sampling System 42
Figure 4‑10 Natural Gas Sampling System in Field 43
Figure 4‑11 Mercury Analyzer in Field (Merlin Detector) 43
Figure 4‑12 Simple Sketch of Mercury Analyzer (Merlin Detector) 44
Figure 5-1 Simple Process Flow Diagram 45
Figure 5-2 Experimental Breakthrough Curves 47
Figure 6-1 Freundlich Isotherm Linear Regression for 4 mm Adsorbent 53
Figure 6-2 Freundlich Isotherm Linear Regression for 2 mm Adsorbent 54
Figure 6-3 Langmuir Isotherm Linear Regression for 4 mm Adsorbent 55
Figure 6-4 Langmuir Isotherm Linear Regression for 2 mm Adsorbent 55
Figure 6-5 [anonimizat] 4 mm Adsorbent 56
Figure 6-6 [anonimizat] 2 mm Adsorbent 56
Figure 6-7 Pseudo-Second Order Linear Regression for 4 mm Adsorbent 57
Figure 6-8 Pseudo-Second Order Linear Regression for 2 mm Adsorbent 57
Figure 6-9 Elovich Model Linear Regression for 4 mm Adsorbent 58
Figure 6-10 Elovich Model Linear Regression for 2 mm Adsorbent 58
Figure 6-11 Weber & Morris Model Linear Regression for 4 mm Adsorbent 59
Figure 6-12 Weber & Morris Model Linear Regression for 2 mm Adsorbent 59
Figure 6-13 Diffusion-Chemisorption Model Linear Regression for 4 mm Adsorbent 60
Figure 6-14 Diffusion-Chemisorption Model Linear Regression for 2 mm Adsorbent 60
Figure 6-15 Vinod & Anirudvan (Linear Driving Force Diffusion) Model Linear Regression for 4 mm Adsorbent 61
Figure 6-16 Vinod & Anirudvan (Linear Driving Force Diffusion) Model Linear Regression for 2 mm Adsorbent 61
Figure 6-17 Zhang & Cheng Model Linear Regression for 4 mm Adsorbent 62
Figure 6-18 Zhang & Cheng Model Linear Regression for 2 mm Adsorbent 62
Figure 6-19 Wolborska Model Linear Regression for 4 mm Adsorbent 63
Figure 6-20 Wolborska Model Linear Regression for 2 mm Adsorbent 63
Figure 6-21 Clark Model Linear Regression for 4 mm Adsorbent 64
Figure 6-22 Clark Model Linear Regression for 2 mm Adsorbent 64
Figure 6-23 Bohart & Adams Model Linear Regression for 4 mm Adsorbent 65
Figure 6-24 Bohart & Adams Model Linear Regression for 2 mm Adsorbent 65
Figure 6-25 Bohart & Adams Model Prediction versus Actual Breakthrough Data for 4 mm Adsorbent 66
Figure 6-26 Bohart & Adams Model Prediction versus Actual Breakthrough Data for 2 mm Adsorbent 67
Figure 6-27 Example of the Automated Use of Simulation Tool 68
Figure 6-28 Pressure Drop Distribution across Layers of MRU Bed 70
Figure 6-29 Simulated Breakthrough Curves for Designed Bed Height and after Skimming of Top 500 mm 71
Figure 6-30 Simulated Breakthrough Curves for Particle Size Effect 72
Figure 6-31 Simulated Breakthrough Curves for Inlet Mercury Concentration Effect 73
Figure 6-32 Simulated Breakthrough Curves for Gas Flow Rate Effect 74
Figure 6-33 Simulated Breakthrough Curves for Bed Height Effect 75
Figure 8-1 Simulated Breakthrough Curves for Proposed Bed Height with Different Inlet Mercury Concentrations 78
Figure 8-2 Simulated Breakthrough Curves till Saturation for Proposed Bed Height with Different Inlet Mercury Concentrations 79
List of tables
Table 1-1 Mercury Properties 2
Table 1-2 Mercury Compounds 5
Table 3-1 Regional Average Mercury Content in Natural Gas 14
Table 4-1 Characteristics of Zeolite Types 18
Table 4-2 Categorization of Adsorption Systems 29
Table 4-3 Dynamic Models of Linear Equilibrium, Isothermal, and Trace Systems 30
Table 4-4 Dynamic Models of Irreversible Equilibrium, Isothermal, and Trace Systems 31
Table 4-5 Dynamic Models of Nonlinear Equilibrium, Isothermal, and Trace Systems 31
Table 4-6 Kinetic Expressions 32
Table 4-7 Mercury Analytical Techniques 41
Table 6-1 Calculated Langmuir Constants 55
Table 6-2 R2 Values of Models Linear Regression 66
Table 6-3 Design Basis of Mercury Guard Bed 70
Table 10‑1 Breakthrough Profile for 4 mm Particles 90
Table 10‑2 Breakthrough Profile for 2 mm Particles 91
Table 10‑3 Simulation Results of Case #1 92
Table 10‑4 Simulation Results of Case #2 93
Table 10‑5 Simulation Results of Case #3 94
Table 10‑6 Simulation Results of Case #4 95
Table 10‑7 Simulation Results of Case #5 96
Table 10‑8 Simulation Results of Case #6 97
Table 10‑9 Simulation Results of Case #7 98
Table 10‑10 Simulation Results of Case #8 99
Table 10‑11 Simulation Results of Case #9 100
Table 10‑12 Simulation Results of Case #10 101
Table 10‑13 Simulation Results of Case #11 102
Table 10‑14 Simulation Results of Case #12 103
Table 10‑15 Simulation Results of Case #13 104
Table 10‑16 Simulation Results of Case #14 105
Table 10‑17 Simulation Results of Case #15 106
Table 10‑18 Simulation Results of Case #16 107
Table 10‑19 Simulation Results of Case #17 108
Table 10‑20 Simulation Results of Case #18 109
Table 10‑21 Simulation Results of Case #19 110
Table 10‑22 Simulation Results of Case #20 111
Table 10‑23 Simulation Results of Case #21 112
Table 10‑24 Simulation Results of Case #22 113
Nomenclature
a External Surface Area per Unit Particle Volume
A Column Cross Sectional Area (m2) in Clark Model
ACGIH The American Conference of Governmental Industrial Hygienists
Al Aluminum
Al2O3 Aluminum Oxide
b Langmuir Equilibrium Constant = Kads/Kdes
BDST Bed Depth Service Time
BET Brunauer, Emmett, and Teller
BT Breakthrough
C (Co/Ce/Ct)
Adsorbate Concentration in Fluid Phase in mass of adsorbate / fluid volume units or mass of adsorbate / fluid mass units (Initial Concentration at Inlet / Concentration at Equilibrium / Concentration at Time t)
Ca+2 Calcium Ion
CB Ceramic Balls
CO2 Carbon Dioxide
Cu Copper
D Diffusivity and Diffusion Coefficients
d-block Elements Group in Periodic Table
DL Axial Dispersion Coefficient
Dm Molecular Diffusion
DNA Deoxyribonucleic Acid
DP Differential Pressure or Pressure Drop
G (ΔG) Gibb's Free Energy (Changes in Gibb's Free Energy)
H (ΔH) Enthalpy (Changes in Enthalpy)
H Bed Height
h Initial Adsorption Rate in Pseudo-Second Order Kinetic Expression h =
H2 Hydrogen
HC Hydrocarbons
Hg Mercury
HgS Cinnabar (Mercuric Sulfide)
I Iodine
IARC International Agency for Research on Cancer
IDLH Immediately Dangerous to Life and Health
J Flux in Fick's law or The Mass Transfer Rate per Unit Reactor Volume in Clark Model
k / K Constants of Equilibrium and Kinetic Rate Expressions and Mass Transfer Coefficients
Kg Kilograms
Kr Krypton
L Bed Length (Height)
LME Liquid Metal Embrittlement
LNG Liquefied Natural Gas
LUB Length of Unused Bed
m Mass of Adsorbent
MEOH Methanol
mg/m3 Milligrams per Cubic Meter
mg/Nm3 Milligrams per Normal Cubic Meter
mm Millimeters
MMSCFD Million Standard Cubic Feet per Day
MRU Mercury Removal Unit
MS Metal Sulfide
MTZ Mass Transfer Zone
n Constant in Freundlich Isotherm
N2 Nitrogen
NeoC5 Neo-Pentane
NG Natural Gas
ng Nano Gram (10-9 Grams)
NGL Natural Gas liquids
NIOSH National Institute for Occupational Safety and Health in United States
Nm3 Normal Cubic Meters
NOX Nitrous Oxides
O2 Oxygen
OSHA Occupational Safety and Health Administration in United States
P Pressure
P The Probability for Breakthrough in Yoon Nelson Model
PCBs Polychlorinated Biphenyl
PEL Permissible Exposure Limit
pH Potential Hydrogen, a scale representing the relative acidity (or alkalinity) of a solution, in which a value of 7.0 is neutral, below 7.0 is acid, and above 7.0 is alkaline.
PM Particulate Matter
ppm Part Per Million
Psig Pound per Square Inch (Gauge Pressure)
q (qs/qo/qe/qmax/qt)
Adsorbate Concentration in Solid Phase (Adsorbent) in mass of adsorbate / adsorbent volume units or mass of adsorbate / adsorbent mass units (Saturation / Initial / Equilibrium / Maximum Concentration/ Concentration at Time t)
Q Flaw Rate or the Probability for Adsorption in Yoon Nelson Model
Qb Bed Breakthrough Capacity
Qs Bed Saturation Capacity
R Reaction Rate in Zhang & Cheng Model
REL Recommended Exposure Limit
Rp Particle Radius
S (ΔS) Entropy (Changes in Entropy)
S Sulfur
SH- Sulfhydryl Group or called also Thiol Group
Sh Sherwood Number
Si Silicon
SO2 Sulfur Dioxide
t, tb, ts Time, Breakthrough Time, Saturation Time
TWA Time Weighted Average
U Volumetric Flow Rate in Yoon Nelson Model
UAE United Arabic Emirates
UNEP United Nations Environment Program
USA United States of America
v ( Fluid Velocity
Vp Pore Volume
Wc Weight of the Carbon Adsorbents in Yoon Nelson Model
WHO World Health Organization
x Distance and Axial Coordinates
z Distance and Axial Coordinates or Bed Height in Clark Model or Lumped Parameter in Bohart & Adams Model
Zn Zinc
µg Micro Gram (10-6 Grams)
˚C Centigrade (Celsius)
˚A Angstrom
Equilibrium Factor or Desorption Constant or The Relationship between Degree of Surface Coverage and Activation Energy of Chemisorption in Elovich Model or Kinetic Constant in Wolborska Model
ϵ Porosity
Initial Adsorption Rate in Elovich Model or The Fractional Attainment of Equilibrium q/qe in Vinod & Anirudhan Model of Linear Driving Force Diffusion Model or Lumped Parameter in Zhang & Cheng Model
Φ Deactivation Function (dimensionless) in Zhang & Cheng Model
ρp Particle Density
ρb Bed Density
θ Adsorbent Coverage in Langmuir Isotherm or Lumped Parameter in Zhang & Cheng Model
ᵵ Lumped Parameter in Bohart & Adams Model
Table of contents
Acknowledgement………………………………………………………………………………i
List of figures…………………………………………………………………………………..ii
List of tables…………………………………………………………………………….….…iv
Nomenclature…………………………………………………………………………………..v
Summary………………………………………………………………………………….……1
1. Introduction – Mercury Chemistry 2
1.1 Mercury in Nature and History 2
1.2 Physical Properties 2
1.3 Chemical Properties 4
2. Mercury Environmental Concern 6
2.1 Sources of Pollution 6
2.2 Mercury Environmental Cycle 6
2.3 Bioaccumulation and Biomagnification 8
2.4 Mercury Health Effect 9
2.5 Exposure Limits 10
2.6 Minimization of Mercury Global Pollution 10
2.7 Regulations 12
2.8 Historical Accidents 12
2.9 Situation of Egyptian Environment 13
3. Mercury Impact in Energy Industry 14
3.1 Abundance in Energy Industry 14
3.2 Mercury Industrial Problems 14
3.3 Mercury Corrosion 15
3.4 Failure Incidents 15
4. Literature Survey: Mercury Adsorption 17
4.1 Adsorption in Purification & Separation 17
4.2 Adsorbents Characteristics 17
4.2.1 Zeolite Types 18
4.2.2 Particles 19
4.2.3 Pore Diameters and Pore Volume 19
4.2.4 Adsorbent Aging 20
4.3 Adsorption Mechanisms 20
4.4 Adsorption Equilibrium 21
4.5 Capillary Condensation 23
4.6 Diffusion Transport 23
4.7 Adsorption Kinetics 24
4.8 Axial Dispersion 25
4.9 Pressure Drop in Adsorption Packed Beds 25
4.10 Dynamic Modeling of Adsorption Beds 26
4.11 Analytic Solutions 30
4.12 Mercury Adsorption and Chemisorption Processes 34
4.13 Beds Design Fundamentals 37
4.14 Mercury Removal Unit Operation 38
4.15 Mercury Detection and Laboratory Apparatus 41
5. Studying Mercury Adsorption from Natural Gas in an Industrial Unit – Case Study 45
5.1 General Process Description 45
5.2 Mercury Removal Unit (MRU) Description 45
5.3 Studying Dynamics of Industrial MRU Bed (Packed Fixed Bed) 46
5.3.1 Adsorption Equilibrium 47
5.3.1.1 Freundlich Equilibrium Isotherm 47
5.3.1.2 Langmuir Equilibrium Isotherm 47
5.3.2 Adsorption Kinetics 48
5.3.2.1 Lagergren Pseudo-First Order 48
5.3.2.2 Pseudo-Second Order Expression 48
5.3.2.3 Elovich’s Model 48
5.3.2.4 Weber and Morris Model 49
5.3.2.5 Diffusion-Chemisorption Model 49
5.3.2.6 Linear Driving Force diffusion Expression 49
5.3.3 Adsorption Modeling 49
5.3.3.1 Zhang and Cheng Model 50
5.3.3.2 Wolborska Model 50
5.3.3.3 Clark Model 51
5.3.3.4 Bohart & Adams Model 52
6. Results and Discussions 53
6.1 Studying Adsorption Equilibrium 53
6.1.1 Applying Freundlich Equilibrium Isotherm 53
6.1.2 Applying Langmuir Equilibrium Isotherm 54
6.2 Studying Adsorption Kinetics 56
6.2.1 Lagergren Pseudo-First Order Linear Regression 56
6.2.2 Pseudo-Second Order Expression Linear Regression 57
6.2.3 Elovich’s Model Linear Regression 58
6.2.4 Weber and Morris Model Linear Regression 59
6.2.5 Diffusion-Chemisorption Model Linear Regression 59
6.2.6 Linear Driving Force diffusion Expression Linear Regression 60
6.3 Studying Adsorption Mathematical Models 62
6.3.1 Linear Regression of Zhang and Cheng Model 62
6.3.2 Linear Regression of Wolborska Model 63
6.3.3 Linear Regression of Clark Model 64
6.3.4 Linear Regression of Bohart & Adams Model 65
6.4 Bohart & Adams Model’s Predictability for the MRU System 66
6.5 Calculations and Building a Simulation Tool 67
6.6 Simulating Original versus Current Bed Design Performance 70
6.7 Effect of Different Operating Parameters – Sensitivity Analysis and Case Studies 72
6.7.1 Adsorbent particle size 72
6.7.2 Inlet mercury concentration 73
6.7.3 Feed gas flow rate – velocity 74
6.7.4 Bed Height 75
7. Conclusion 76
8. Recommendations 77
9. Refferences 80
10. Appendices 90
10.1 Appendix A – Breakthrough Tables 90
10.2 Appendix B – Simulation Results 92
الملخص بالعربى ………………………………………………………………………………..…114
Summary
Mercury is one of the most critical impurities that contaminate all types of fossil fuels and exists in variable ranges of concentration in natural gas; however, it is a critical impurity due to its hazardous environmental and health impacts in addition to its corrosive nature for some metals that affects oil & gas industries by catastrophic failure accidents.
The purpose of this thesis is to study an industrial unit that removes mercury from natural gas by adsorption. An industrial fixed bed packed with specific type of mercury adsorbents was modeled and tested model prediction for the bed design basis, and studied the effect of layer skimming on bed performance to solve pressure drop problem. The model was used to simulate bed performance and predict breakthrough and lifetime by varying catalyst size, different bed heights, different gas flow rates, and different inlet mercury concentrations.
It was observed that mercury adsorbent is proving a high adsorption capacity as observed from manufacturer data, modeling results, and field performance that make the bed last for years of lifetime. The matching model also proves adsorbent characteristics, nature of equilibrium, and the chemical reaction mechanism (Chemisorption). Increasing the gas flow rate and the inlet concentration leads normally to faster saturation and earlier breakthrough. Decreasing the size of the adsorbent increases adsorption efficiency due to improving surface area and mercury diffusivity into adsorbent pores. However, lower bed heights can accommodate high flow rates and moderate mercury concentrations up to 2 µg/Nm3 (2,000 ng/Nm3) and this opportunity can be seized in the upcoming years by two proposals that save capital expenditures and improve treatment of feed gas by optimization of bed design and its configuration.
Introduction – Mercury Chemistry
Mercury is a chemical element with 80 atomic number. It is commonly called also as quicksilver and has a symbol of Hg in chemistry referring to being historically named hydrargyrum, which is a Latinized form of Greek word meaning water-silver as it is shiny silver liquid. [1] Also mercury word has a correspondence to an ancient Roman god, and Mercury planet. Occasionally, such heavy, silvery d-block element is the only metallic element that is liquid at standard conditions of temperature and pressure. Mercury has unique properties resulted in utilizing it in different applications such as barometers and manometers due to its high density, and used in thermometers due to high rate of thermal expansion that is fairly constant over a wide temperature range.
Mercury in Nature and History
Mercury is existing naturally as one of the rare metallic elements in Earth's crust with mass concentration of 0.08 parts per million (ppm). [2] Mercury exists in deposits all over the world commonly in the form of cinnabar (HgS) ores and rarely in elementary form but rather within compounds and inorganic salts. Cinnabar usually occurs as a dark red powder. It is often called by the common name of vermillion or Chinese vermillion.
Historically, Mercury was known by the ancient civilizations such Egyptians as found in Egyptian tombs, and used by the ancient Chinese empire as though to prolong life, also ancient Greeks and Romans used it in cosmetics. [3,4]
In the past, it was extracted intensively in wide range in Spain and Southern America after inventing a process to extract silver from ores by mercury. [5]
The largest producer of mercury outside the United States is Spain, Kyrgyzstan, Algeria, China, Italy, Yugoslavia, and Finland. In the United States, mercury is produced as a by-product of gold mining. There are also reports of small-scale mining of mercury in China, Russia (Siberia), Outer Mongolia, Peru, and Mexico. [6]
Physical Properties
Mercury is generally a heavy metal with shiny silver color and a certain capability to conduct electricity while poor heat conductor. [7] In addition, it is characterized by advantageous thermal expansion properties, and it is a liquid metal in ambient temperature with exceptional lowest boiling point and lowest melting point for a metal. [8] The following table summarizes its main physical properties.
Table 1-1 Mercury Properties
Figure 1-1 Mercury Electron Configuration
Its unique electron configuration is the reason of its properties, where all electron orbits of mercury are full by electrons from 1s till 6s giving stability like the noble gases resisting electrons removal, movement, or sharing leading to weak bonds and low melting points unlike other metals such as gold, which is only one shift to lift side in periodic table, where 6s shell has one less electron and thus exhibiting easily more removal and sharing of electrons between atoms and give rise to stronger metallic bonds and different usual metallic properties. [9,10] Sharing electrons between metal atoms results in electricity conduction, hardness, sliding and formation …etc. For example, magnesium has a higher melting point than sodium because Mg2+ centers are glued together by an electron sea with 2 electrons for every atom, while each atom in sodium metal contributes only one electron. Mercury hangs on to its valence 6s electrons very tightly. Mercury-mercury bonding is very weak because its valence electrons are not shared readily. (In fact, mercury is the only metal that does not form diatomic molecules in the gas phase). Heat easily overcomes the weak binding between mercury atoms, and mercury boils and melts at lower temperatures than any other metal and mercury's ability to conduct electricity and heat much poorer than expected for a metal at that position in the periodic table. The s electrons are able to come very close to the nucleus. They swing around very massive nuclei at speeds comparable to that of light. When objects move at such high speeds, relativistic effects occur. The s electrons behave as though they were more massive than electrons moving at slower speeds. The increased mass causes them to spend more time close to the nucleus. This relativistic contraction of the 6s orbital lowers its energy and makes its electrons much less likely to participate in chemistry- they're buried deep in the atomic core. [10,12]
Chemical Properties
The natural common form of mercury is the mercuric sulfide "Cinnabar", a red pigment vermilion stable form, and formed by reaction of elemental form with sulfur.
Hg + S → HgS (1.1)
Accordingly, this reaction is utilized to capture mercury such as solid sulfur flakes, which are used in mercury spill kits to absorb mercury. Activated carbon and powdered zinc are also used. [13]
On the other hand, the elemental mercury is extracted from Cinnabar by heating above 540˚C in presence of air followed by condensing the vapors. [14]
HgS + O2 → Hg + SO2 (1.2)
Mercury is moderately active and does not react easily with Oxygen in the air and does not react with most cold and diluted acids such as diluted sulfuric acid but does react with concentrated sulfuric acid, nitric acid, and aqua regia to produce sulfates, nitrates, and chlorides. Mercury also reacts with hydrogen sulfide.
The major activity of mercury is the capability to dissolve other metals to form amalgams (alloys) which is called amalgamation. Such metals are Gold, Silver, Aluminum, Manganese, Copper, Sodium, Zinc, and Platinum, but the latter is not easily amalgamated. Iron is one of the important exceptions. [15,16]
Incompatibilities include Acetylene, ammonia, chlorine dioxide, azides, calcium (amalgam formation), sodium carbide, lithium, rubidium, and copper. [17]
The elemental form of mercury has a symbol Hg or Hg(0) while the two main oxidation inorganic forms or salts are monovalent Hg(I) and most common divalent Hg(II) or Hg2+ form. Higher forms are rare such as Mercury fluoride HgF4. [18] When mercury combines with carbon, the compounds formed are called "organic" mercury compounds or organomercurials. Organomercury compounds are always divalent and do not react with water. They usually have the formula HgR2, which are often volatile, or HgRX, which are often solids called Organo-ionic compounds, where R is aryl or alkyl and X is usually halide or acetate. Methyl-mercury, a generic term for compounds with the formula CH3HgX, is a dangerous family of compounds. [19,20] They arise by a process known as biomethylation. There are a potentially large number of organic mercury compounds (such as dimethyl-mercury, phenyl-mercury, ethyl-mercury, and methyl-mercury); however, by far the most common organic mercury compound in the environment is methyl-mercury. Like the inorganic mercuric compounds, both methyl-mercury and phenyl-mercury exist as "salts" (for example, methyl-mercuric chloride, or phenyl-mercuric acetate). When pure, most forms of methyl-mercury and phenyl-mercury are white crystalline solids. Dimethyl-mercury, however, is a colorless liquid. [14]
The following table summarizes many mercury compounds and their applications.
Table 1-2 Mercury Compounds [21]
Mercury Environmental Concern
Sources of Pollution
Mercury releases are resulted from natural sources and numerous human activities. The natural sources include any natural mobilization such as volcanic eruptions, and normal breakdown of minerals in rocks and soil through exposure to wind and water [25], while human activities include the following man made mobilization: [26-28]
The majority is from combustion of fossil fuels; mostly the coal in power plants, oil, and also natural gas
gold and silver mining
non-ferrous metal production, typically smelters
cement production (mercury in lime)
waste disposal, including municipal and hazardous waste, crematoria, industrial wastewater, and sewage sludge incineration [29]
caustic soda (chlor-alkali) production
iron and steel production
mercury production
manufacturing of products containing mercury
uses of mercury containing products
It is estimated that one to two thirds of total mercury releases are from human activities, like mining and fossil fuel-burning and still mercury concentrations in the environment are increasing by human activity.
Mercury Environmental Cycle
Mercury naturally exists in environment in the most common forms of metallic elemental, mercuric sulfide, mercuric chloride, and methyl-mercury. Most of human mercury releases are direct in the air while some releases are direct in soil such by fertilizers, and direct in water such by wastewater disposal. The recipients of mercury releases to the environment include the atmosphere, water environments (aquatic) and soil environments (terrestrial). There are continuing interactions – fluxes of mercury – between these compartments. The speciation – the chemical form – of the released mercury varies depending on the source types and other factors. Elemental mercury in the atmosphere can undergo transformation into inorganic mercury forms, providing a significant pathway for deposition of emitted elemental mercury. Some microorganisms and natural processes can change the mercury in the environment from one form to another. The most common organic mercury compound that microorganisms and natural processes generate from other forms is methyl-mercury. Once mercury has been liberated from either ores or from fossil fuel and mineral deposits hidden in the earth’s crust and released into the biosphere, it can be highly mobile, cycling between the earth’s surface as deposits and the atmosphere as vapors. The earth’s surface soils, water bodies, and bottom sediments are thought to be the primary biospheric sinks for mercury. [30]
The different forms of mercury are called mercury species with different properties. Those properties play a key role in defining; physical state of exposure, transport nature through organisms, degree of toxicity, nature of accumulation, modification, detoxification, emission control, and extent of environmental pollution relative to emission source (transport distances). [31]
Methylmercury can be formed in the environment mainly by microbial metabolism (biotic processes), such as by certain bacteria particularly in earth water bodies and tend to concentrate in marine creatures that consume those bacteria or planktons adsorbing the methylmercury. Methylation process was found affected by acidity degree, dissolved organic carbon content, and can be reversed by exposure to ultraviolet sunrays. [32,33]
Mercury emissions are transported locally (regional) and globally by air and by ocean currents, so any local source contributes to global background.
The only long-term sinks for removal of mercury from the biosphere are deep-sea sediments and, to a certain extent, controlled landfills, in cases where the mercury is physio-chemically immobilized and remains undisturbed by anthropogenic or natural activity (climatic and geological). [34]
The following sketches illustrate simply the mercury environmental cycling.
Figure 2-1 Mercury Environmental Cycle
Bioaccumulation and Biomagnification
In the earth water bodies, lakes, rivers, seas, and deep oceans; the marine creatures naturally concentrate mercury, typically Methylmercury absorbed easily more than other forms, in their bodies. Methylmercury is soluble in fats and concentrate mainly in muscle tissue, and viscera. [35] Most of the methylmercury in fish tissue is covalently bound to protein sulfhydryl groups. This binding results in a long half-life for elimination (about two years).The mercury exists in marine microorganisms and eaten by small fishes and marine creatures which are eaten by a predator, the mercury is accumulated in fish tissues increasingly with time due to much less efficiency to depurate and thus the bioaccumulation occurs. Consequently, as long as the predators are larger on the top of food chain, larger in size, the much mercury is concentrated in its tissues which is called the biomagnification, meaning the progressive build up by successive trophic levels till the top of food chain. Generally, mercury level in fishes exceeds the level in the water they live in due to quick absorbing great amounts of mercury. The available data in researches indicate the presence of mercury all over the world (especially in fish) in levels that negatively affects human and wildlife. These levels have led to consumption advisories for fish in a number of countries, warning people, especially sensitive (such as pregnant women and young children), to limit or avoid consumption of certain types of fish from various water bodies. [36]
Mercury Health Effect
Mercury is of global concern due to extreme toxic nature of all its compounds and must be treated carefully in different handling and leakage cases with specific procedures aiming to avoid extremely hazardous exposure. [37] Spill cases can be collected by effective amalgamation with applicable powders such as zinc, and sulfur. Toxicity differs between the compounds, and the most toxic forms are the organic methyl, and dimethylmercury.
Liquid exposure can be absorbed by the skin and mucous membranes while vapors are inhaled and poisoning may be chronic or acute.
The target organs are generally eyes, skin, respiratory system, central nervous system, and kidneys.
Exposure to mercury generally had shown many health effects differing with mercury chemical form, concentrations, time of exposure, frequency of exposure, the way of exposure, and the age of poisoned person. Health effects include tremors, impaired cognitive skills, mood changes, inability to concentrate, sleep disturbance, chest pain, dyspnea, cough, hemoptysis, impairment of pulmonary function, evidence of interstitial pneumonitis, profound central nervous system effects, including psychotic reactions characterized by delirium, hallucinations, suicidal tendency, erethism, irritability, excitability, excessive shyness, insomnia, violent muscular spasms, loss of memory, blindness, discoloration of the cornea & lens of the eye, disturbances of vision, vivid dreams, depression, nerve, brain, thyroid, kidney damage, eye irritation, skin rashes, vomiting and diarrhoea. [38-43]
The effects can generally be simplified into the following main effects:
Disruption of the nervous system
Damage to brain functions
DNA damage and chromosomal damage
Allergic reactions, resulting in skin rashes, tiredness and headaches
Negative reproductive effects, such as sperm damage, birth defects and miscarriages
Damaged brain functions can cause degradation of learning abilities, personality changes, tremors, vision changes, deafness, muscle incoordination, and memory loss. Chromosomal damage is known to cause mongolism. [23]
Methylmercury is a well-documented neurotoxicant, which may in particular cause adverse effects on the developing brain. Moreover, this compound readily passes both the placental barrier and the blood-brain barrier; therefore, exposures during pregnancy are of highest concern. In addition, some studies suggest that even small increases in methylmercury exposures may cause adverse effects on the cardiovascular system, thereby leading to increased mortality. Given the importance of cardiovascular diseases worldwide, these findings, although yet to be confirmed, suggest that methylmercury exposures need close attention and additional follow-up. Moreover, methylmercury compounds are considered possibly carcinogenic to humans (group 2B) according to the International Agency for Research on Cancer (IARC, 1993), based on their overall evaluation. [33]
Mercury has profound cellular, cardiovascular, hematological, pulmonary, renal, immunological, neurological, endocrine, reproductive, and embryonic toxicological effects. [44]
The greatest risk, however, is for fetuses and young children because their nervous systems are still developing. They are four or five times more sensitive to mercury than adults. Damage occurring before birth or in infancy can cause a child to be late in beginning to walk and talk and may cause lifelong learning problems. Unborn children can be seriously affected even though the methylmercury causes no symptoms in their mothers. [45]
Exposure Limits
Routes of Exposure are generally through inhalation, skin absorption, Ingestion, Skin and/or eye contact.
Humans can be exposed to mercury through polluted atmosphere, polluted water, occupational exposure, accidental exposure directly or by leakage or spill from products, eating crops sprayed by mercury containing material in agriculture, and main exposure by eating contaminated fishes. The larger the fish the higher potential mercury level. Also, mercury from soils can accumulate in mushrooms.
OSHA Permissible Exposure Limit (PEL) and NIOSH Recommended Exposure Limit (REL) are TWA: 0.05 mg/m3, Ceiling: 0.1 mg/m3, and IDLH: 10 mg/m3. Also ACGIH set recommended airborne exposure limit to 0.025 mg/m3 averaged over an 8-hour work shift. [17,46,77]
Minimization of Mercury Global Pollution
As combustion of fossil fuels is increasing in order to meet the growing energy demands of both developing and developed nations, mercury emissions can be expected to increase accordingly in the absence of the deployment of control technologies or the use of alternative energy sources.
Mercury concentration in incineration flue gases is commonly in the range of 0.5 mg/Nm3, which is much higher than the regulation accepted emission limit to the extent of 0.05 mg/Nm3 in some countries. [47]
Control technologies have been developed for coal combustion plants and waste incinerators with the primary intention of addressing acidifying substances (especially SO2 and NOX), and particulate matter (PM). Such existing technologies may provide some level of mercury control, but when viewed at the global level, currently these controls result in only a small reduction of mercury from these sources. Many control technologies are significantly less effective at reducing emissions of elemental mercury compared to other forms due to being volatile and water insoluble. Optimized technologies for mercury control are being developed and demonstrated, but are not yet commercially deployed. [48-51]
The figure below shows mercury release categories with main types of possible control mechanisms.
Figure 2-2 Mercury Releases and Control Measures [34]
The possible measures of controlling mercury releases can be summarized as the following:
Preventive measures
Reducing consumption of raw materials and products generating mercury releases
Substitution by non-mercury alternatives in processes and products
Controlling measures
End-of-pipe techniques
Waste management
The preventive controls are generally cost-effective, unless the alternatives are significantly more costly or there are other limitations. Also, end of pipe treatment for control of mercury emissions as a control strategy still leads to mercury wastes which are other possible source of re-emissions and accordingly must be treated, and disposed in adequate safe environmentally techniques. Mercuric waste management may include one or more of the techniques for a controlled landfill, controlled underground storage, and pretreatment of waste by keeping the mercury in stable inert form and management complexity arises from the many different source of waste. However, combination of the different measures is the optimum for reducing mercury emissions towards saving the environment globally. [52]
As the awareness of mercury's potential adverse impacts on health and the environment has been rising, the recorded virgin mercury production has been decreasing from about 6000 to about 2000 metric tons per year during the last two decades, and consequently, related releases from mining and usage of mercury may also be declining.
Anthropogenic emissions from a number of major sources have decreased during the last decade in North America and Europe due to reduction efforts. Also, total anthropogenic emissions to air have been declining in some developed countries in the last decade. [48]
Regulations
Internationally, on the 10th of October 2013, agreement was signed by 140 countries on Minamata Convention on Mercury by the United Nations Environment Program (UNEP) to obligate countries for control mercury emissions generally and phase-out mercury products. [53,54] The Unites States, and European Union have initiatives within the last decade regulating mercury emissions, and restricting mercury products and trade.
The application of different strategies for controlling mercury emissions differ from country to country according to government and local priorities, information and education about possible risks, the legal framework, enforcement, implementation costs, perceived benefits and other factors. [55]
Usually, more developed countries generally have more stringent limitations than developing ones with some certain notable exceptions. Mercury emissions from cement plants are regulated at 0.05-0.10 mg/Nm3 in Europe, Egypt, Brazil, Nigeria, Australia, Chile, and South Africa.
Pakistan (10 mg/Nm3), Colombia (0.03 mg/Nm3) and Germany (0.03 mg/Nm3) are the extremes in mercury emissions limits. Also, there are many countries that have no limit for mercury emissions at all including the major markets of China, India, Turkey, UAE, Saudi Arabia, and Lebanon. China is in the process of implementing new regulations, effective 1 March 2014. [56]
The World Health Organization, OSHA, and NIOSH all treat mercury as an occupational hazard, and have established specific occupational exposure limits.
The World Health Organization leads many initiatives by publishing the health impacts of the different forms of mercury, guidance of risk from mercury exposure, tools to reduce mercury exposure, and guidance on the replacement of mercury-containing thermometers & blood pressure measuring devices in health care, projects to promote the management and disposal of health-care waste and has facilitated the development of an affordable, validated, non-mercury-containing blood pressure measuring device. [57]
There are various regional initiatives conducted by different countries such as; [58]
setting environmental quality standards limiting concentrations in industrial emissions, wastes, drinking waters, water bodies, air, soil, and diet (fish),
restriction for mercury uses, mercury products, and trade,
regulations on workplace exposures,
advisories for fish consumption,
promotion of developing and introducing safer alternatives and cleaner technology,
and raising people awareness.
Historical Accidents
Historically, from the mid-18th to mid-19th centuries, highly toxic mercuric nitrate was used in making felt hats in a process called carroting. The symptoms appeared on hat makers of nervous moods and shaking are the reason of the common expression of mad hatter. [59,60]
The most popular mercury release accidents were at Minamata Bay in Japan, Colex plant at Oak Ridge in Tennessee. [61]
Colex plant, a lithium-isotope separation plant, 1950s – 1960s, was responsible for one of the largest historical releases of mercury. Another disaster of dumping mercury compounds (Methylmercury) into Minamata Bay, Japan, occurred in the 1950’s where organic mercury by-products of industrial-scale acetaldehyde production were discharged in the local bay resulted in thousands of severe poisoning and death cases known as Minamata disease. [62,63] Also, the Iraqi poisoning events where wheat treated with a seed dressing containing organic mercury compounds were used for bread. [64]
Situation of Egyptian Environment
Regarding emissions restrictions, Egypt is one of the countries that put restrictions on emissions from cement production at maximum of 0.05 mg/Nm3, which is one of the most strict limitations in the world.
Another major pollution source was chlor-alkali plant in Alexandria utilizing mercury cells with adverse environmental and health impacts and fortunately changed to alternative safe process using membrane cells. Accordingly, studies paid attention for the surrounding environment and recommended continuous monitoring of Hg in Wadi Al-Qamar area and continuous health monitoring for residents. [65]
Regarding aquatic environment, a study concerned with mercury and methylmercury in sediments of northern lakes reported results with a conclusion that they can be considered not polluted with mercury and the average total mercury concentration was found in the following order; Mariout > El-Manzallah > El-Burullus > Edku > El-Bardaweel, while the methylmercury was found in the order; El-Bardaweel > El-Burullus ≥ El-Manzallah > Mariout > Edku. [66] However, another study, for evaluation of the extent of human exposure to methylmercury in Aldakahlia Governorate by examining mercury concentration in hair samples, revealed that hair mercury content levels were generally higher than the internationally accepted levels and in El-Matareye city specifically where most of people work in fishery and recommended further researches to detect sources leading to exposure to mercury. [67]
The high values of mercury contamination in fishes from Abu Kir Bay, Alexandria, which was reported in Mercury bioaccumulation study, concluded that, it is reflected to the influence of the additional sources of runoff to the bay from industrial wastes.
In general, it was reported through various studies, a contamination in Egyptian coastal marine ecosystem. [68]
In Egypt, attention is needed for millions of fluorescent tubes produced and improperly disposed either ending at landfill or broken releasing mercury to environment. Those tubes shall be phased out and at least subjected to waste management strategy. [69]
However, the general populations need enrichment for the culture related to the environment and public health and raising awareness about mercury hazards, environmental and health effects.
Mercury Impact in Energy Industry
Abundance in Energy Industry
Mercury is one of the common trace components found naturally in fossil fuels including natural gas, oil, condensates, tar sands, coal, and bitumen accordingly it is available in different concentrations in production, processing systems, end products, and combustion exhausts. [70]
The mercury is known to have more accompaniment tendency with lighter hydrocarbons such as propane and butane, so can be existing in increased concentrations in these streams. [71,72]
Mercury exists in Natural Gas reserves worldwide in different concentrations predominantly in elemental form with low concentrations much below saturation limit, i.e. no liquid phase exists. Content usually expressed in micrograms per normal cubic meters typically in the range of less than 1 to <300 µg/Nm3 differing with geological origin. It could be present also in other forms; inorganic (such as HgCl2), organic (such as CH3HgCH3, C2H5HgC2H5), and organo-ionic (such as ClHgCH3) compounds. [73,74]
The highest gas concentrations reported in Southeast Asia, Eastern Europe, and North Africa. [75]
Table 3-1 Regional Average Mercury Content in Natural Gas [75]
Mercury Industrial Problems
Mercury has either concerns related to process problems or environmental hazard or for both. If the concern is environmental as in coal and cement plants, the removal technology may be applied at the raw material itself or at the combustion flue gases based on the feasibility, but if the concern is process problems, it is certainly must be removed from raw material before passing to process units of concern.
Mercury has corrosion problems with some metals of amalgamation potential between the two metals, and known as Liquid Metal Embrittlement (LME). The most common metal used in Natural Gas Industry, and suffers from severe corrosion, is the Aluminum. Accordingly, mercury must be generally removed before contacting any material of corrosion potential in equipment, pipes, and valves to avoid mechanical failure. Also, mercury may be removed to avoid any incompatibility with processes or poisoning of metal catalysts used in the reactions during hydrocarbon processing. [73, 178]
Significant mercury removal is mandatory particularly in cryogenic gas plants such in LNG, and NGL industries to avoid corrosion and mechanical failure of Aluminum exchangers that are always utilized in the process of those plants. Although, mercury exists in very low concentrations in natural gas, but it affects Aluminum in cumulative way by progressive amalgamation. The removal is commonly achieved by mercury removal units (MRU) or called guard beds located upstream cryogenic units. The treated gas has commonly specification of maximum allowable limit of 0.01 µg/Nm3 or generally undetectable level. [176-178]
The removal system also may be applied to the end product if it will be delivered to a petrochemical process sensitive to mercury. [76]
Mercury Corrosion
Mercury amalgamates (forming alloy/amalgam) with the surface layer of metal it contacts. The produced amalgam is weaker than the base metal. In case of mercury in contact with Aluminum surface, aluminum diffuses into mercury and transforms to Aluminum oxide Al2O3 in presence of air or water acting like inducing bores in the Aluminum and replacing with brittle oxide layers. The amalgamation process proceeds with corrosion by removing the oxide layer in presence of a catalyst or an aqueous media resulting Aluminum Hydroxide, Hydrogen, and leaving the previously amalgamated mercury free to continue develop again successive amalgam with the metal in a progressive process developing weak spots (cracks).
Al + Hg Al Hg (3.1)
2AlHg + 6 H2O 2Al (OH)3 + 3 H2 + 2 Hg (3.2)
This corrosion process is called liquid metal embrittlement (LME) with reference to liquid elemental mercury penetration into the aluminum oxide protective coating and coming in direct contact with the aluminum. This phenomenon happens above the melting point of elemental mercury (approximately -40 ˚C) in presence of water; such as in warm feed gas and during maintenance when cryogenic equipment is shutdown and warmed up for repairs or dry-out operations to remove accumulated hydrates. When equipment are put back in service, stresses may cause metal separation at the weak spots (cracks) leading to catastrophic failure. [71,75,76]
Many theories provide proposed description for the mechanism of LME. [78-85] Generally, it can be described by the loss of ductility in normally ductile metals when stressed under contact with liquid metal. Some significant examples of embrittling couples include steel-Cu, stainless steel-Zn, and aluminum-Hg. Mercury often embrittles mild steels in stressed or unstressed conditions at low temperatures. [86]
Failure Incidents
Mercury effects were recognized by 1973 through investigating the failure incidents like the catastrophic failure happened at Aluminum exchanger in Algerian Skikda LNG plant on January 2004. Also, mercury content in the range of 0.001 to 180 µg/Nm3 was reported, in a study, responsible for corrosion in gas gathering system in Groningen fields in Holland.
Another catastrophic incident, on 1 January 2004, occurred by mercury corrosion at Aluminum exchanger in the Moomba gas plant in South Australia lead to gas release and fire. [76]
Figure 3-1 Shape of Mercury Corrosion and Failure
It is reported that gas plants problems are almost from the common existing elemental mercury form, not suspected to be caused from organic or non-organic forms. Failure cases were reported from Algeria, USA, Indonesia, Thailand, and Holland as a result of presence of mercury particularly in Cryogenic plants.
Literature Survey: Mercury Adsorption
Adsorption in Purification & Separation
Adsorption now is the most effective process in terms of cost, efficiency, and flexibility. The most common adsorption process is utilizing an adsorbent column/bed packed with a hydrophilic porous material to dry fluids. However, there are many other specific adsorption processes in large scale deigned to remove undesirable impurities such as acid gases, mercury, organic and metallic pollutants in water. As long as traces or low concentration component that being removed from a fluid, it is categorized in Purification processes that get rid of undesirable components, or low economic value components, and maximize value of fluid by increasing its purity. Development of adsorption process opened the category of separation adsorption by being much more economic choice than distillation of close boiling components by utilizing sorbent with adsorption separation factor much greater than the relative volatility. Activated carbon and silica gel were in common use before the development of molecular sieve adsorbents particularly synthetic zeolites in the late 1950s which lead to the potential of separation by adsorption process and various zeolite structures have been tailored.
Adsorption separation factor usually differs by temperature and composition accordingly operating conditions are significantly considered during process design. However, in ideal Langmuir system it has no relation with composition rather being evaluated by ratio of Henry’s law constants. [87]
Separation may be based on kinetic separation as in molecular sieves where selectivity depends on relative micropore diffusivities of both components and this requires tailoring of micropore diameter comparable with the diameter of diffusing species, also separation may be based on difference in adsorption equilibrium or different pore diffusion rates. [87]
Adsorbents Characteristics
Microporous adsorbents differ in pore size from a few angstroms to tens of Angstroms and each sorbent suits specific process conditions. In traditional sorbents, there is a distribution for pore size and the average pore size and range of distribution is controlled by the synthesis process procedures but the zeolites almost have no size distribution and size is controlled by crystal structure.
Silica Gel (> 20 A°), contain hydroxyl groups which is responsible for degree of polarity to have affinity for polar molecules such as mainly water, as a desiccant, also amines, alcohols, phenols, and unsaturated hydrocarbons. [87]
Activated Alumina (Al2O3.3H2O) is much polar than Silica Gel and shows comparable affinity for water and capacity difference depend on temperature which makes Alumina much efficient desiccant in drying at warmer temperatures.
Impregnated or chemically modified Activated Alumina also has chemisorption applications.
Activated Carbon is originally Graphite or a carbonaceous material activated by steam or Carbon Dioxide at high temperatures to open its pores. All of its characteristics (porosity, surface area, and capacity) vary with synthesis procedures of pyrolysis (carbonization) and activation. It is almost nonpolar so it has no affinity for water and rather selective for organics. [87] Hetero-atoms such as oxygen, nitrogen, hydrogen, sulfur, and phosphorous are functional groups that can be found and play key roles in capacity, surface pH, electrostatic interactions, and chemisorption. Its wide range of pore sizes <20 to >500 A° and characteristics correspond multi applications. Larger pore diameters are used with liquid phase than gas phase to decrease mass transfer resistance.
Impregnated/ chemically modified Activated Carbon has also applications by utilizing deposited active sites to increase adsorption capacity. Sulfur Impregnated Carbon is a common example and investigation studies proven that sulfur has no significant effect on the initial pore structure and characteristics but has a significant effect on adsorption capacity.
Carbon Molecular Sieves are Activated Carbons prepared by specific activation procedure to obtain narrow range of pore size distribution to enable some kind of sieving selectivity. [87]
Clay Minerals with potential high adsorption, and ion exchange properties are widely utilized to decontaminate aqueous solutions.
Biosorbents and lignocellulosic agricultural wastes are promising low cost adsorbents.
Metal Oxides have also applications in water treatment.
Zeolites, whether natural or synthesized, are porous crystals of Aluminosilicates (lattice structures of SiO4 and AlO4 tetrahedral cells joined together) where Oxygen is shared between Al & Si atoms forming rings of Oxygen members which provide the inter-crystalline open channels structure or pores which is penetrated by the adsorbed molecule. Shared Oxygen configuration results in various structure arrangements with different pore dimensions based on the number of Oxygen atoms in the ring, 6, 8, 10, and 12. Aluminum atoms give negative charges to be balanced by exchangeable cations. Each cation can change the adsorption properties and may obstruct channels leading to reducing effective pore size and changing inter-crystalline diffusivity. Si/Al ratio is a very important parameter which is at least =1 or higher or even reaching a Silica pure. Rich Al zeolites are selective for polar molecules such as water, amines, alcohols, phenols, and unsaturated hydrocarbons while lean Al zeolites (Pentasil) adsorb the non-polar hydrocarbons. Accordingly, the structure, Si/Al ratio, and cation types are the ruler of adsorption properties. [87]
Zeolite Types
Table 4-1 Characteristics of Zeolite Types [87]
Particles
The adsorbent micro-porous crystals are formed as a macro-porous pellet or granulated bead (spherical) particles or even flakes with optimized dimensions, porosity, and mechanical strength. The particle has two diffusion resistances to mass transfer; one micro-pore intra-crystalline resistance and one macro-pore intra-particle resistance. Size reduction achieves enhancement in diffusivities (low resistance) but at the cost of mechanical strength, pressure drop. Also, reducing crystal size may enhance micro-pore resistance but will increase macro-pore resistance. [87]
Binders, such as, clay are added to stick crystals and enhance strength. Different binder materials have adsorption properties, which alter the selectivity of original crystals. [87]
Pore Diameters and Pore Volume
The following figure summarizes briefly the ranges of pore diameter and pore volume of some adsorbents.
Figure 4-1 Pore Volume and Pore Size Distribution of Adsorbents [87]
Adsorbent Aging
Aging or deactivation of adsorbent is the loss of equilibrium capacity or increase in mass transfer resistance by;
loss of crystalline structure (some hydrothermal conditions in Zeolite X), [87]
pores closure partially (some hydrothermal conditions in Zeolite A), [87]
pores blockage by poisonous impurities (fouling),
pores blockage by coke formation,
pores blockage by capillary condensation,
Losing active sites in chemisorption either by leaching or flushing active phase or normally by adsorbed component occupancy.
Adsorption Mechanisms
Adsorption is a process through which porous particles capture certain chemical component, selectively, within its pore structure either physically, called physical adsorption, or chemically, called chemisorption. Physical adsorption occurs primarily by weak intermolecular forces called Van der Waals forces (dispersion and repulsion) and other potential contributions by electrostatic interactions (polarization, dipole, and quadrupole interactions) due to coulombic forces of attraction between charges of molecules and surfaces, and it has almost exothermic nature. [87] The electrostatic interactions importance are involved significantly by the ionic structure of zeolites and also can be a key factor in optimizing pH of aqueous solutions to enhance adsorption by certain changes in adsorbent surface charges with pH. [88] Chemical adsorption or chemisorption includes additionally a chemical reaction forming bond between the molecule and specific active site at the adsorbent surface.
For comparison, physical adsorption main features are summarized in the following: [87]
lower heat of adsorption,
nonspecific,
possibility of mono and multi layers,
usually promoted at lower temperatures,
rapid,
reversible,
no electron transfer, no dissociation, but polarization may occur
On the other hand, chemisorption characterized by the following:
higher heat of adsorption,
specific,
monolayer,
may involve dissociation,
applicable in a wide range of temperature,
slow and activated,
irreversible,
involve electron transfer in the reaction
However, in ionic adsorbents characterized by an electric field at the surface, such as zeolites, the significant contribution of electrostatic interactions in physical adsorption exhibits some of the features of chemisorption such as high heat of adsorption, slow, specific process, and activated by diffusion. [87]
Gas adsorption is usually characterized by relatively fast mass transfer than liquid adsorption. Gas adsorption also progressive from monolayer to multilayer and may exhibit capillary condensation while liquid adsorption is primarily monolayer and pores are already filled with liquid. Gas desorption may be achieved by temperature swing, pressure swing, and gas purges while liquid desorption is achieved by temperature swing with liquid purge.
Adsorption thermodynamics are addressed by investigating the nature of changes in Gibb's free energy ΔG, enthalpy ΔH, and entropy ΔS with adsorption. Briefly, most of physical adsorption processes are exothermal (generating heat & ΔH is negative) and accordingly observed to be promoted with lowering temperature and some are endothermal (activated by heat) and thus promoted by increasing temperature. [87]
Adsorption Equilibrium
The basic equilibrium relationship is the linear equilibrium between fluid phase concentration and adsorbed phase concentration, which is considered in cases of low concentrations and described by Henry's law and constant of proportionality is Henry adsorption equilibrium constant which is inversely proportional to temperature. [87]
q = KC (4.1)
A general classification described by Brunauer et al. for physical adsorption schemes of equilibrium isotherms as in below photo.
Figure 4-2 Brunauer Classification for Equilibrium [87]
Class I describes equilibrium in microporous sorbents with pore size matching the adsorbed molecule diameter and characterized by a saturation limit. Classes II & III describe the adsorbents with a range of pore sizes leading to progressive loading from monolayer to multi layers and capillary condensation. Class IV for two surface layers and class V for significant intermolecular attraction effects. [87]
One of the most common and simple models formulating the monolayer adsorption is the Langmuir Isotherm (Langmuir, 1918) which describes the highly favorable (irreversible or called rectangular isotherms) equilibrium (class I) and originally formulated to represent chemisorption or strongly adsorbed species as it is based on the below assumptions: [87,89]
Adsorption occurs at fixed number of well defined localized sites.
Monolayer coverage, i.e. each site is occupied by one molecule.
Homogeneous sites with equivalent adsorptive energy.
The adsorbed molecules are isolated from each other (No mutual interaction).
(4.2)
By Langmuir model, a good fit of many experimental isotherms can be achieved just by optimum selection of constants b (equilibrium constant = Kads/Kdes) and qs (saturation capacity). Higher values of b means higher adsorption rate relative to desorption rate and accordingly higher removal.
Another linear form is [87,89]
(4.3)
Freundlich isotherm (Freundlich, 1906) is empirical expression appropriate for highly heterogeneous surfaces with no restriction to monolayer coverage [87,90]
q = K C 1/n (4.4)
Also linear form is
ln q = ln K + 1/n ln C (4.5)
Where K is equilibrium constant indicative for adsorption capacity and n is a constant representing the adsorption intensity or the degree of favorability where 1/n between 0 to 1 is indication of a favorable adsorption and 1 reduces to linear Henry law. Accordingly, for fitting experimental data, variations in the shape of the breakthrough curve are fit by adjusting constant n while variations in the equilibrium adsorption capacity are fit by adjusting the parameter K. The adsorption capacity is inversely related to K. [87,90]
In cases of multilayer adsorption, a reference to the model developed by Brunauer, Emmett, and Teller (BET Isotherm) is preferred. [87,91]
There are other two or three parameters expressions such as; Temkin (1940) isotherm which considers adsorbate-adsorbent interactions, Redlich-Paterson isotherm which combines between Langmuir and Freundlich characteristics, Sips isotherm, Flory-Huggins, Dubinin-Raduskevich expression, Koble-Corrigan, Halsey isotherm, and Toth isotherm model. [91,92]
Figure 4-3 Adsorption Isotherms and Mathematical Forms [91]
Capillary Condensation
In the small pores of porous adsorbents, the effect of surface tension may be significant to the extent of continuous loading from multilayer adsorption to capillary condensation where the pores are filled with liquid phase because the vapor pressure in a small pore is lowered by the surface tension. Accordingly; pore size selection for adsorbents is very critical when such phenomenon is expected and undesired to avoid active sites leaching or avoid blockage of active sites by condensed molecules rather than the desired sorbate molecules and resulting lower capacity. [87]
Diffusion Transport
The transport of adsorbed molecules through the pores of adsorbent is a major controller for the rate of adsorption and desorption. There is no flow through the pores but there is a diffusion process by the driving force of chemical potential gradient rather than the concentration gradient as may be deduced from the formula of Fick's law. [87]
J = – D(c) (4.6)
Pore diffusion mechanism depends on pore size, pore shape, tortuosity, microcrystal structure, molecule size, molecule shape, molecule concentration, and other conditions. [87]
For distinction, there is intraparticle, or pore, or macropore diffusion process in the larger pores of adsorbent particles, and there is intracrystalline or micropore diffusion process in the much small micropores of Zeolite crystals. [87]
The effective macropore diffusivity involves contributions from four mechanisms and depends on the pore structure where a tortuosity factor is considered and defined through correlations with porosity (inverse proportion) and such factor considers the effects that reduce effective diffusivity such as random orientation of pores, and variation in pore size, or shape. [87]
The first mechanism is the molecular diffusion resulting from collisions between molecules and inversely proportional to pressure and directly proportional to temperature. This mechanism dominates when mean free path between molecular collisions is much smaller than pore diameter while at the contrary conditions in small pores and low pressures the molecular collisions with the pore wall will be much frequent and dominate the Knudsen diffusive resistance. Such mechanism is known as Knudsen diffusion and directly proportional to pore size and temperature slightly, while inversely proportional to molecular weight. Indeed, there will be transient conditions where both resistances will be significant. Other mechanism is the surface diffusion resulting from the flux of physically adsorbed layer on the macropore surface, and such contribution depends on the thickness of such layer and accordingly the temperature. Also, it depends on concentration significantly. Another one mechanism is called Poiseuille flow, and considered experimentally while neglected in packed beds. It is a contribution to flux from forced laminar flow through the macropores resulting from a difference in total pressure across a particle. It is directly proportional to pressure and pore size and inversely with viscosity. [87]
In micropore diffusion, the process is activated, relatively slow, and temperature dependent, and the activation energy is correlated with molecular diameter relative to micropore window diameter (Direct proportion). In general, micropore diffusivity is inversely proportional to crystal diameter, and directly proportional to temperature, pressure, adsorbed phase concentration, and percentage of Ca+2 cations. Differences in micropore diffusivities resulted in the development of efficient separation processes such as separation of O2 and N2 by carbon molecular sieves. [87]
Adsorption Kinetics
Kinetics of adsorption processes are very important and studied extensively in pilot and industrial scales for many purposes not limited to the following;
basically to understand adsorption mechanisms and resistances,
define mass transfer resistances, and relative importance of each individual resistance to specify the rate controlling resistance or called rate limiting step,
adsorption process design,
study dynamics of a process by modeling,
process optimization from operating conditions, and sorbent characteristics,
examining new synthesized sorbents experimentally,
examining kinetics of certain processes using certain adsorbents,
defining capacities, breakthrough time, equilibrium nature, equilibrium constants, mass transfer constraints and mathematical coefficients,
researches for purification processes serving the environment such as water treatment, waste water treatment, exhaust gases treatment …etc.
The kinetics are referred to rate expressions describe uptake rate and accordingly define the residence time of adsorption. Kinetics are studied by defining the extent or the significance of the effect of each individual constraint resisting the mass transport process, for each system with particular conditions, the process may be controlled by one or more resistance, the following are the major constraints which may contribute in the kinetics of any system; [87]
External mass transfer resistance in the film or laminar fluid boundary layer surrounding the adsorbent particle when the fluid is binary or mixture. This resistance may be expressed in terms of fluid phase concentrations or solid adsorbed phase concentrations. The film or laminar layer surrounding the particles contains mass transfer by molecular diffusion and the thickness of the film or mass transfer coefficient depends on hydrodynamic conditions. Transport rate is usually represented by a linear driving force equation: [87]
(4.7)
Where k is the effective mass transfer coefficient, C & C* are fluid phase concentration in fluid and in equilibrium respectively, q is the averaged solid adsorbed phase concentration, Rp is the particle radius, and a is the external surface area per unit particle volume which equivalent to 3/Rp for spherical shape. Mass transfer coefficient is correlated in the form of dimensionless group Sherwood number Sh = 2RpK/Dm (4.8) where Dm is molecular diffusion, and Sherwood number is correlated with Reynolds & Schmidt number. [87]
Macropore or intra-particle diffusional resistance. When it is significant, there will be concentration profile through the particle while uniform concentration in micros crystals and particle size will be a factor in the uptake rate. Investigation literatures conclude that adsorbent particle internal concentration gradient is higher than external gradient so it is common practice to consider intra-particle diffusion resistance more dominant than external film resistance in evaluating mass transfer rate. [87]
Micropore or intra-crystalline diffusional resistance such in zeolites, which consist of microporous crystals formed into macroporous particles in pellet or bead granulated shapes. When it is significant, the concentration through the particle is uniform and uptake rate is independent of particle size. [87]
Axial dispersion or called also axial mixing, but it is not a mass transfer resistance. [87]
Heat transfer resistances in non-isothermal systems; because of the exothermic nature of adsorption, the heat of adsorption will be transferred and temperature difference will exist between the particle and fluid as long as adsorption occurs but this depends on many conditions from which rates of mass and heat transfer. When heat transfer can be considered rapid enough (high heat transfer coefficients) relative to the adsorption rate and accordingly the rate of heat generation then the temperature differences or gradient through the particle and between the particle and the fluid will be negligible and the system is considered isothermal. [87]
Figure 4-4 Schematic Diagram of Porous Adsorbent Particle [87]
Axial Dispersion
Axial dispersion is also called axial mixing which is necessary to be minimized through design stages as it affects separation efficiency negatively. It is represented by a single lumped coefficient reflecting the contributions of each mechanism. In case applicable to neglect dispersion effects, it is called ideal plug flow. The contributing mechanisms include molecular diffusion, mixing or flow by-passing the particles, and also wall effect dispersion in cases of bad packing distribution (non-uniform) or small bed or tube diameters, or high bed porosity. Ratio of bed to particle diameter shall be large enough not less than 20. The decrease in particle sizes below 3 mm diameter has indeed the advantage of minimizing pore diffusional resistance but investigation literatures indicate offset of such advantage by increasing axial dispersion that may be due to the tendency of smaller particles to stick together. [87]
Pressure Drop in Adsorption Packed Beds
Pressure drop is a major design parameter to be considered for minimizing pumping or re-compression costs in cases, and may also affect productivity in some other cases when the pressure is a factor. The pressure drop depends mainly on fluid velocity, and of course governed by; fluid distribution, solid suspension in fluid, particles size, particles crush strength & possibility of dust generation, particles packing & distribution, bed porosity, bed dimensions, fluid density, and fluid viscosity. Such parameter had been investigated thoroughly in literatures and usually correlated with dimensionless friction factor, which depends on Reynolds number. Flow direction has direct relationship with flow velocity limitations and downward flow is preferred to extend the tolerance limit of velocity while upward flow is much limited much below bed fluidization velocity. In industrial scale, its preferred not to over decrease adsorbent particle size relatively below 1 mm diameter to maintain reasonable pressure drop. In addition, bed porosity or voidage shall be controlled with lower limit to avoid high pressure-drop and upper limit to avoid wall effects and axial dispersion effects. [87]
Dynamic Modeling of Adsorption Beds
The overall dynamics of the adsorption beds reflect all the mentioned objectives for studying adsorption kinetics, controlling the system design and defining efficiency of processes. Extracting information about adsorption kinetics, and equilibrium is commonly achieved by fitting experimental response curves of a system to match theoretical response calculated from the most appropriate dynamic model derived in literatures to describe certain process features. The selection of the model depends on the main assumptions that approximates the system features and once found a mathematical model fitting the actual data to a reasonable extent, this model can be used to simulate the system and extract information about the process. [87]
Models derivation usually starts with differential mass balance (sometimes called continuity equation) for an element of the bed. The general equation for fluid phase differential mass balance of axially dispersed plug flow is:
(4.9)
Figure 4-5 General Differential Mass Balance Equation over Bed Element [87]
The adsorption rate expression produced from mass balance over the adsorbent particle is generally written as: [87]
(4.10)
It is a general equation, but the rate equation is usually expressed by equations including one or more diffusional resistance with relative boundary conditions and utilizes the appropriate equilibrium isotherm. The dynamic response of a bed will be the solution of c (z,t), and q(z,t) from the differential equations within the specified initial and boundary conditions. [87]
The features of mass transfer zone and concentration front depend on the equilibrium isotherm, but, of course, proven that the concentration profile may change by system kinetics. The equilibrium relationship between fluid and solid concentrations can be categorized generally to linear, favorable, and unfavorable isotherm. For distinction between them, X-Y diagram describe their relationships. [87]
Figure 4-6 Categories of Equilibrium Isotherms [87]
By defining equilibrium factor similar to relative volatility or separation factor [87]
When the adsorbent is initially concentration free c'0 = q'0 = 0
For desorption, the situation is reversed and accordingly, favorable sorption will be desorbed in unfavorable scheme, and unfavorable sorption will be desorbed in favorable scheme. While adsorption and desorption curves are mirror images in linear equilibrium systems. [87]
For linear isotherm 0 –> C0 = 1 and less than 1 for favorable and higher than 1 for unfavorable. [87]
The mass transfer zone in linear and unfavorable equilibrium systems is continuously increasing in its width as long as concentration front travels through the bed and such profile is called dispersive or proportionate pattern. In favorable equilibrium isotherm, the mass transfer zone propagates in width in the initial region and after certain short distance, it maintains constant pattern in shape and width and so called constant pattern isotherm. The distance, at which mass transfer zone starts propagate in constant pattern with no further changes in shape, this distance depends on or the extent of nonlinearity of isotherm and on kinetics, and it is very small in many practical systems. [87]
However, the sharp concentration front is mandatory for efficient separation, while dispersed front reflects resistances to mass and/or heat transfer within the particle.
Instantaneous equilibrium between fluid and solid is commonly validated assumption for strongly rapid adsorbed species, not restricted in transport, or long enough contact time. [93]
For non-equilibrium modeling, the expression considers mass transfer resistance, and includes constants such mass transfer coefficient, pore diffusivity constants; effective, molecular, Knudsen, surface and these values are estimated through correlations in literatures.
The dynamic features of adsorption processes can be categorized based on the behavior of mass transfer front (equilibrium nature) and the extent of complexity in the mathematical model with regard to concentrations range, adsorption rate expression, and nature of flow model whether dispersed or can neglect dispersion. The following is a general table for adsorption system classification. [87]
Table 4-2 Categorization of Adsorption Systems
Analytic Solutions
For linear equilibrium, many expressions for the breakthrough curve and concentration front are available in literatures with considering different kinetic cases.
The linear isotherm assumption is an accepted approximation in case of low concentration changes. [87]
The following table provides the analytic solutions for linear equilibrium, isothermal, and trace systems available in literatures: [87]
Table 4-3 Dynamic Models of Linear Equilibrium, Isothermal, and Trace Systems
In many practical systems, there is no practical value from the complexity of the more comprehensive models and the simple linear model provides accepted error by using effective mass transfer coefficient as a lumped parameter from the different mass transfer resistances.
For the cases of strongly adsorbed species, equilibrium is nonlinear and represents extreme limit of highly favorable isotherm approaches zero. It is important limiting case commonly described by irreversible and rectangular isotherm, which is the simplest favorable case. The earliest solution of breakthrough curve was derived by Bohart & Adams utilizing semi-chemical kinetic rate equation. [87,94]
The following table provides the analytic solutions for irreversible equilibrium, isothermal, and trace systems available in literatures.
Table 4-4 Dynamic Models of Irreversible Equilibrium, Isothermal, and Trace Systems [87]
The following table provides the analytic solutions for other cases of nonlinear equilibrium systems, isothermal, and trace systems available in literatures: [87]
Table 4-5 Dynamic Models of Nonlinear Equilibrium, Isothermal, and Trace Systems
It was concluded in literatures that the simple linear rate model is a valid approximation in unfavorable, linear, and moderate favorable isotherms 1 > > 0.5 while breakdown with high favorable isotherms.
The effect of axial dispersion on favorable (constant pattern) isotherms is also available in literatures by Acrivos, Garg & Ruthven, and Rhee & Amundson. In addition, the non-isothermal behavior was studied by Garg, Ruthven, and Crawford for a Langmuir equilibrium with solid film linear rate expression. [87]
The most common kinetic models that were widely examined for experimental sorption systems are listed in the below table:
Table 4-6 Kinetic Expressions [94-109]
Mercury Adsorption and Chemisorption Processes
Although mercury adsorption, in literatures, are emphasized for the case of mercury control from coal combustion power plants due to global environmental concerns, but the removal from natural gas has often similar importance and has the same adsorption principles, however few studies were concerned with natural gas.
Mercury is conventionally removed by adsorption onto porous particles such as activated carbon. The most common used carriers or support materials are activated carbon, activated alumina, and zeolite materials. The fluid, called sorbate or solute, is typically contacting solid adsorbents packed in fixed bed reactors. In addition, sorbents injection techniques, in ducts of flue gases, were also studied extensively.
Commercial trapping products differ in the nature of the trapping agent and/or the support material.
The principle types of trapping agents are elemental sulfur, metal sulfides, metal oxides, activated carbon, iodine, silver, and gold. [118-123] Sliver and gold active phases are simply based on effective amalgamation removal with regeneration potential.
In mercury adsorption, the fixation of mercury in stable form is a necessary characteristic desired in adsorbents to consider the environmental hazard of disposing the spent adsorbents after being saturated. Accordingly, some recent studies are available for adsorbents capable of keeping the adsorbed mercury fixed in stable form. [112] In addition, some studies focused on the modeling of the adsorption mechanism of mercury such as Chung et al. model, which was developed for mercury adsorption onto active sites of impregnated activated carbon. [110] Another study by Ren et al. who utilized a successful representative model considering surface equilibrium and mass transfer mechanisms to investigate mercury adsorption from flue gases by activated carbon and fly ash. [111] Also Carla et al. investigated the stable chemisorption of mercury from natural gas by using laboratory synthesized hydroxyapatites modified with copper sulfide as active sites for chemical reaction and utilized a model considering adsorption, diffusion, and chemical reaction. [113] Sasmaz et al. investigated the chemistry of mercury adsorption onto brominated activated carbon where chemisorption was found the likely mechanism. [120] Skodras et al. studied mercury and PCBs adsorption from gas phase using different types of activated carbon where they found increase in efficiency by adding impregnated sulfur active agent for chemisorption mechanism. [124] The study included experimental testing supported by a developed mathematical model that considers fundamental equilibrium and mass transfer expressions with a successful representation. Meserole et al. studied mercury removal from flue gases in ducts by sorbent injection, and presented a theoretical model that combines sorbent characteristics (extracted from experimental results), mass transfer characteristics of the system, Freundlich equilibrium isotherm, the surface area available for sorption, and the residence time, but did not incorporate any terms to account for intraparticle diffusion. [125] The model considers adsorption in two steps; gas-phase mass transfer of the mercury to the sorbent surface, followed by surface reaction that is simply modeled as physical adsorption. The model was used to determine the conditions under which either mass-transfer limitations or sorbent capacities are the rate-limiting step of mercury removal when injecting sorbents into the duct. Flora et al. performed studies for mercury removal from flue gases by activated carbon injection, which are similar to the previous mentioned study by Meserole et al. in work and results, but they employed a two-stage model followed by sensitivity analysis. The first stage considers sorption in the duct, while the second stage models the additional sorption due to the retention of carbon particles on the filter in the form of a fixed bed. This model incorporates key mass transfer and equilibrium processes that govern adsorption of mercury vapors on activated carbon in the duct and on the fabric filter. [131-132] Similar experimental work and modeling attempts were investigated by Chen and co-workers, Serre et al., and Scala. [126-130]
Based on the fact that the principles of adsorption, chemisorption processes, equilibrium isotherms, kinetics, and all governing theories are the same for whatever process, the literature survey is extended for other various chemisorption studies. Y. S. HO and G. Mckay compared extensively chemisorption kinetic models applied for various pollutant sorption systems, and found results indicating that chemisorption processes could be rate-limiting step, and that the pseudo-second order equation has high potential for application in many chemisorption processes more than the pseudo-first order rate equation. [133] Sutherland and Venkobachar studied chemisorption of copper salts using a biological forest product (fungus) and found the controlling mechanisms are film diffusion combined with surface adsorption during the initial stages followed by combination of diffusion and chemisorption for the subsequent 98% of the reaction period. Accordingly, his empirical diffusion-chemisorption model simulated successfully the entire process of biosorption kinetics. [134] This model had been developed and used in his study of heavy metals chemisorption from waters by utilizing low-cost adsorbents. [99] Gusmão et al. evaluated the adsorption of cationic dyes by modified sugarcane bagasse, where they found chemisorption is the controlling mechanism, and that the adsorption process could be described by the pseudo-second-order kinetic model, and well fitted by the Langmuir isotherm. [136] Singh and Pant investigated kinetics and mass transfer limitations in the adsorption of arsenic onto activated alumina and the modified iron oxide impregnated activated alumina. They found that the first-order Lagergren kinetics fits to the adsorption of arsenic over activated alumina, while the pseudo-second order equation describes the behavior of arsenic (III) adsorption over the modified activated alumina. Regarding diffusional limitations, during the initial period, surface diffusion was predominant, but as the adsorption progresses, pore diffusion dominated the rate of adsorption. [137] Odoemelam et al. assessed the adsorption capability of bamboo dust and the bamboo-based activated charcoal for removal of lead and cadmium ions from aqueous solutions by experimental work and kinetics studies. The pseudo second-order model represented the process indicating that the rate-limiting step could be a chemical reaction, and particle-diffusion limitations were found to contribute in adsorption. [135] Deokar and Mandavgane studied adsorption of 2,4-dichlorophenoxyacetic acid using rice husk ash. They tried to apply different mathematical models, including bed-depth service time, Bohart & Adams, Wolborska, Thomas, Clark, Yoon–Nelson, and deactivation kinetic (Zhang & Cheng Model), to experimental data for breakthrough curve prediction and to study the optimal bed parameters. They found that more than one rate-controlling step is involved in the adsorption. According to the Adams–Bohart and Wolborska models, the kinetics, in the early stages of adsorption, are dominated by external mass transfer. The Thomas model is more suitable than Clark model because the system involves monolayer adsorption. Because the deactivation kinetic model best fits to experimental data over the entire breakthrough curve, it was utilized for simulating this solid–liquid adsorption system. [109] Similar study was performed by Natalie Ho to simulate hydrogen sulfide adsorption, for biogas treatment, by using activated carbon material, where the deactivation kinetic, Zhang & Cheng model, was utilized successfully in representing the system and simulating its performance. [138]
Many other researches thoroughly deal with experimental evaluation for the capability of various adsorbents to remove effectively the mercury species. [118-123, 139-170]
Hence, chemical trapping of mercury in the form of cinnabar, a stable non-volatile mercury ore, is the most commonly used mercury removal method for natural gas. Whichever the technique employed, either the porous carriers are chemically modified with sulfur by a chemical bond or the carriers are just impregnated, the objective is to achieve the following rapid reaction, which increases adsorption capacities significantly. [114, 176-178]
Hg + S HgS (4.32)
The resulted mercury sulfide is trapped within the structure of the adsorbents (fixed) until being saturated and replaced usually within more than 5 years lifetime according to the load.
Accordingly, as there is no commercial value from the spent materials, so they must undergo strict environmental friendly disposal technology within waste management strategy, either through acceptable landfills, or through mercury recovery by heating followed by condensation to be recycled for commercial uses instead of additional mercury mining.
Generally, materials shall be selected carefully in design, and bed location to be optimized as there are limitations in some applications. Although the sulfur impregnated active carbon has a great performance, the following problems are addressed in its application: [175-178]
not effective for treating condensates or crude oils,
feed gas should be free of entrained liquids and any liquid carryover to be avoided,
both zeolite materials and activated carbon are small pore materials which are subjected to capillary condensation potential if used near dew point conditions,
irreversible damage or loss of active agent occurs due to sulfur solubilization by the hydrocarbon liquid leading to flushing and leaching it from particles,
Operators' health problems due to the sulfur sublimation and condensation, [75]
reducing pores accessibility by increasing mass transfer and diffusion resistances when the pore volume is filled with liquid,
Consequently, much less capacity, much less efficiency, and early breakthrough occurs.
Alternative applications are also utilized such as; adsorption onto a solid ion-exchange resin containing chemically bound active -SH groups, molecular sieves, sulfur-treated zinc oxide, chemically-modified carrier with metal sulfide, mixture of basic copper carbonate, basic zinc carbonate, and simultaneous mercury removal and drying using regenerative silver-coated molecular sieves. [176-178]
Figure 4-7 Configuration of Regenerative Mercury Removal System
Beds Design Fundamentals
Each vendor has criteria for designing beds for their material but some rules of thumb are that the bed should be sized for a superficial flow velocity of about 50 ft/min and a residence time of 10 seconds. Generally, the fixed beds are designed by considering the following aspects:
Knowing the gas composition, operating pressures, operating temperatures, and the applied processes to define suitable adsorbent characteristics, and define location of the bed to avoid any process interference, avoid any process incompatibility, and ensure optimum operating conditions. [115]
Packing with a sufficient quantity of adsorbents to provide the required removal efficiency, typically reaching undetectable level as much as below 0.01 µg/Nm3 during a specified lifetime interval for changing out the spent material. The amount are estimated based on the sorbent characterized capacity against the fluid properties; maximum and minimum flow rate (velocity/contact time), and maximum mercury concentration (load). [176-178]
Pressure drop is affected by many parameters including; fluid velocity, fluid distribution, solid suspension in fluid, particles size, particles crush strength & possibility of dust generation, particles packing & distribution, bed porosity, bed dimensions, fluid density, and fluid viscosity. By considering all the fixed and optimum input parameters, the pressure drop determines the diameter of the bed.
Ensuring good distribution of the fluid through the adsorbent bed which is determined by the bed dimensions.
Flow direction either downward or upward, is generally specified according to design velocities.
Support structure of the bed including; the wire meshes, and inert (ceramic) balls to prevent fluidization, and particles escape.
Evaluation of experimental breakthrough curves enables defining bed capacity, range of mass transfer zone, adsorbent usage rate, and volume of fluid treated by mass balance, and integrating the area under the curve, which is proportional to capacity according to the following equations. [109, 171-174]
Breakthrough capacity Qb = (4.33)
Saturation capacity Qs = (4.34)
where tb and ts are breakthrough and saturation times respectively, m mass of adsorbent, Co inlet concentration, and Q flaw rate. [109]
According to the above evaluation and known bed height L, the mass transfer zone where adsorption occurs and propagates through the bed in which:
bed saturation varies from 100% saturated to fresh free adsorbents,
fluid concentration changes from inlet concentration to approaching zero or the designed specification,
this MTZ is approximated by [109, 171-174]
MTZ = L Qb / Qs (4.35)
and length of unused bed (LUB) or called virgin zone to be approximated by
LUB = L ( 1 – Qb/Qs) (4.36)
Mercury Removal Unit Operation
Operation commences by introducing the natural gas into fresh sorbate-free bed where the selective mass transfer occurs between mercury molecules and the surface of solid sorbents. Accordingly, the first layers of the bed uptakes mercury from the gas in the beginning till holding the maximum amount of mercury that can be trapped i.e. maximum sorption capacity, and thus being saturated and can't uptake anymore. The subsequent layers are still active and accommodate the uptake of upcoming mercury molecules, and the position of those active layers is moving with time in the direction of flow till reaching the bottom layers, which when saturated, the mercury will start passing out of the bed, which is called breakthrough, and the unit has to be shutdown, unloaded from the spent material, and reloaded with new fresh charge of adsorbents. [171-174]
From the above scenario, three distinct zones in the bed are commonly defined as the following: [171-174]
The saturated zone that does not uptake mercury,
The mass transfer zone (MTZ) where adsorption occurs and propagates through the bed in which bed saturation varies from 100% saturated to fresh free adsorbents, and fluid concentration changes from inlet concentration to desired effluent concentration,
Virgin zone, which is unused guard layer and free of mercury.
The below simple sketches explain the three zones.
Figure 4-8 Concentration Fronts through Columns and Zones Classification
Generally, the removal efficiency is normally affected by certain factors that can be summarized as below. [115]
Gas flow rate (velocity)
The higher the flow rate, the more rapid bed saturation, and hence the earlier breakthrough occurs.
Higher flow rate than the design leads to general shorter lifetime, and shorter contact time, which adversely influences uptake rate, in addition to higher pressure drop.
Lower flow rate leads to longer contact time, which improves adsorption rate, but less turbulence may decrease external film mass transfer coefficient and increase undesirable axial dispersion effects.
Gas mercury concentration
The higher the concentration, the more rapid bed saturation, MTZ propagates more rapidly, and hence the earlier breakthrough occurs with shorter lifetime.
Higher concentration than the design leads to breakthrough and general shorter lifetime.
Sorbent characteristics
Sorbent composition, porosity, pore size, surface area, pore volume, tortuosity, surface pH, synthesis procedure, particle size, and crush strength are all characteristics influence adsorption and pressure drop.
Gas composition
Adsorption is influenced by gas composition such in cases of co-adsorption potential for other species, existing poisoning impurities that may block pores (fouling), capillary condensation potential by water or heavy hydrocarbons, and liquid carryover potential.
Operating pressure
Pressure has effect on physical adsorption, which is usually promoted by higher pressure.
Operating temperature
Operating temperature affects both physical and chemical adsorption according to thermodynamic features of the adsorption process whether exothermic that will be promoted by lower temperature or endothermic promoted by higher temperature.
Bed length
The higher the bed height, the higher the capacity, the longer the lifetime, the later the breakthrough, but the higher the pressure drop.
Contact time
Longer residence time normally increases the chance of mercury uptake and achieves higher removal and it is function of flow rate and bed geometry.
Bed porosity/density and particles distribution
The lower bed porosity negatively affects pressure drop although it indicates much dense bed and higher removal capacity, while the much high porosity gives chance to channeling and axial dispersion effects in addition to lower capacity.
Mercury Detection and Laboratory Apparatus
Generally, mercury detection techniques are various and differ in application, limitations, and range of level can be detected.
Colorimetric analysis technique passes the mercury-contaminated gas into Iodide solution such Potassium Iodide to produce mercuric iodide and estimate mercury content. Another colorimetric method used for high concentrations is based on passing the gas through Draeger tube to detect mercury content by the resulting color changes by formation of copper-mercury complex.
Another technique works for low concentrations is based on recovery of mercury in elemental and volatile form and expose to ultraviolet radiation to estimate the mercury level by measuring the absorption of ultraviolet radiation. This technique is known as cold vapor atomic absorption sepectrophotometry.
Often, a technique depends on amalgamation collection is utilized by using metals such as gold and silver in the form of thin wire or film to capture mercury and form amalgams. The use of gold film to form amalgam with mercury, which is characterized by a higher electrical resistance to be the guide to estimate mercury content based on resistance change. [114]
The below table shows different analytical techniques and detection level.
Table 4-7 Mercury Analytical Techniques [116]
The MRU bed under study in the plant is normally under continuous monitoring for mercury adsorption efficiency and pressure drop across the bed. The monitoring of removal efficiency is simply achieved by sampling from inlet and outlet gas in periodical basis to ensure the bed is working well.
The system used to detect mercury level is a combination between two techniques. The detection system used is PSA SIR GALAHAD II based on atomic fluorescence principles (Merlin Detector) for detection combined with gold trapping system PSA 10.547 (Dual Adsorber) in sampling. Atomic fluorescence is characterized by advantageous signal sensitivity and low detection levels below the current regulation levels imposed in America and Europe. The system suits both natural gas application and urban air monitoring and allows using low volume samples in accurate measurements.
Briefly, the system selectively captures the atomic mercury by trapping in gold-sand adsorber tube, then heating the sample to vaporize the mercury and being detected quantitatively by produced fluorescence signals.
The sampling system includes the following;
inlet sample connection from process pipeline,
pressure indicator for introduced sample,
isolation valve between high pressure process side and regulated sample side
primary bypass connection to provide higher sample flow rates 15-20 liters/minute bypassing the system to reduce residence time in sampling lines and avoid mercury retention/losses providing higher accuracy,
heated regulation valve to reduce sample pressure from 0 to 30 psig and recommended 5 – 10 psig. The regulator is electrically heated to avoid condensation and minimize mercury adsorption to system stainless steel surfaces,
pressure indicator for regulated sample
three way valve used to optionally configure sampling in series or in parallel,
safety relief valve at regulated sample side to prevents pressure build up over 25 psig,
injection port with gas-tight syringe,
mercury adsorber tube heated electrically to keep sample temperature above dew point (~ 95 ˚C) to prevent hydrocarbon condensation which affects trapping efficiency. The heating coil is also used to evaporate the trapped mercury for detection. The trap has a trapping capacity of a few micrograms enough for most of applications.
thermocouple indicate adsorber temperature,
sample outlet port,
flow meter to control and indicate sample flow through adsorber tube 0.1 – 1 liters per minute (recommended 0.4 – 0.5 Liters per minute specially in low concentration detection and 2 liters/minute leads to mercury breakthrough).
The following diagram explains the sampling system simply with its main components.
Figure 4-9 Flow Diagram of Natural Gas Sampling System
Also the following photos are captured for the sampling system in the field and laboratory.
Figure 4-10 Natural Gas Sampling System in Field
In the detection system, a carrier gas entrained with the evaporated mercury gives fluorescence that measured by photomultiplier tube (signal). The carrier gas is preferred to be Argon rather than the Nitrogen or Air as they affect the detected signal and accordingly measuring accuracy.
The following photos are captured for the analyzer system in the laboratory and also drawn the following sketch to simply illustrate the main components.
Figure 4-11 Mercury Analyzer in Field (Merlin Detector)
Figure 4-12 Simple Sketch of Mercury Analyzer (Merlin Detector)
In conclusion without details, the overall cycle starts with introducing the sample in the gold-sand trap to adsorbs all mercury traces, then flushing the gas out, then the trap is heated by a coil up to 500 ˚C approximately to evaporate the mercury and being entrained with the carrier gas and introduced to fluorescence detector, then the trap is cooled by a flow of inert gas to be ready for receiving further samples, and finally a software translates the fluorescence signal into a calculated mercury concentration using a configured calibration curve which can be verified by using certain calibration procedure using standard mercury source to revalidate the curve.
Studying Mercury Adsorption from Natural Gas in an Industrial Unit – Case Study
General Process Description
In order to maximize economic value of Egyptian natural gas reserves, derivatives are extracted from the gas supplied from producers to serve the local or international markets in addition to providing feedstock for petrochemical industries to additionally magnify the economic value. Accordingly, the NGL plant is established to achieve those objectives. The plant receives feed gas already treated to a degree satisfactory for the set specifications of local market consumers as of homes, power plants, and the national gas grid. According to the cryogenic nature of NGL processes, the gas has to undergo much advanced processing to suit the cryogenic process. Thus, the received feed gas is first subjected to pretreatment based on gas composition including; liquid separation, solid filtration, coalescing, extremely efficient dehydration, and finally mercury removal. Liquids recovery is a cryogenic process achieved by gas pre-chilling using Brazed Aluminum Plate Fin Heat Exchangers commonly known as cold box, followed by isentropic expansion of high pressure cold gases through turbo-expanders to reach extremely low temperature by the effect of enthalpy loss to maximize liquids recovery, in addition to gaining free power to recompress the low pressure dry gas with saving in energy consumption. The mixed liquids recovered are fractionated to separate products using conventional distillation columns while the dry gas is recompressed back to the national gas grid just losing the low fractions of heavy hydrocarbons.
The below simple process flow diagram explains pretreatment processes.
Figure 5-1 Simple Process Flow Diagram
Mercury Removal Unit (MRU) Description
The plant feed gas is already contaminated with very low concentrations making the removal mandatory process wise to eliminate the hazard of Aluminum corrosion in the downstream exchangers by the effect of mercury (LME) and avoid risk of catastrophic failures, which globally experienced in similar processes. Thus, the concern is process safety much more than environmental concern of emissions.
The unit is composed of only one fixed bed and not subjected to regeneration. The bed is packed with mercury removal adsorbents, porous particles, designed to selectively capture the elemental mercury within specified lifetime for the bed, then when bed is saturated, it shall be shutdown, unloaded, and reloaded with fresh charge.
The bed includes an inlet distributor, supporting inert (ceramic) balls, mesh screens, the adsorbents, and bottom outlet collector. In addition, it is equipped with side connections for sampling to monitor the concentration through the bed and measuring pressure drop across the bed interlayers.
Feed gas is introduced flowing from top to bottom in contact with adsorbent particles that uptake the mercury by adsorption, and effluent gas is discharged on spec almost free of mercury.
Adsorption of mercury can be described in two steps; the first step is the physical adsorption of mercury from the gas phase to the solid surface by low energy Van-der-Waals forces. [179] This process is normally reversible due to weak linkage and can be considered slightly exothermic so the equilibrium is promoted by lower temperature and high sorbate concentration. The second step is the trapping (fixation) of adsorbed mercury by chemical reaction with the active phase (agent) supported within the porous structure of particles (carrier). This process is irreversible chemical reaction, under mild conditions, due to stronger linkage involving an exchange of electrons and for this reason, the whole process is called chemisorption process. [179] The chemical reaction is claimed to be promoted by temperature accordingly, the temperature shall be low enough within range that does not to affect equilibrium negatively, while also high enough to promote the reaction kinetics. [179]
For much clear description to improve understanding of process, the carrier is chosen to be macropore particle in beads shape of pore diameter > 80 Angstroms in order to avoid capillary condensation even near dew point conditions in addition to minimizing intraparticle (macropore) diffusion resistances. Often, the active phase was chosen to be metal sulfide supported within the pore structure of the carrier. The active phase is unlike usual impregnated agents to avoid activity loss in cases of entrained liquids carryover; but rather it is bounded to the carrier by a chemical bond giving suitability to liquids without losing activity. Hence, in brief summary, the mercury molecules are adsorbed on surface, diffuse through pores with much less resistance, and chemically react with metal sulfide resulting in mercury sulfide which is stable, non-volatile, and not hydrocarbon soluble so it is readily trapped. [179]
Hg(0) + 2MS → HgS (Mercury Sulfide) + M2S (5.1)
Studying Dynamics of Industrial MRU Bed (Packed Fixed Bed)
In order to study the dynamics of MRU bed, all available bed information and adsorbent characteristics were collected for better initial assumptions regarding the equilibrium relationship and adsorption kinetics then obtained breakthrough data from the adsorbent manufacturer to use it in testing various known mathematical models by linear regression and perform all possible calculations to help in studying the system. The data obtained for two particle sizes 4 mm, and 2 mm as the bed is packed with a range of 2-4 mm adsorbents and the breakthrough curves are shown below in the following graph.
Figure 5-2 Experimental Breakthrough Curves
Adsorption Equilibrium
Both Langmuir and Freundlich isotherm relationships were chosen to be applied to the available numerical data to get insight into adsorption equilibrium of the system.
Freundlich Equilibrium Isotherm
Equilibrium relationship: [90]
q = K C 1/n (5.2)
Linear form:
ln q = ln K + 1/n ln C (5.3)
Where K is equilibrium constant indicative for adsorption capacity and n is a constant representing the adsorption intensity or the degree of favorability where 1/n between 0 to 1 is indication of a favorable adsorption and 1 reduces to linear Henry law.
By plotting Ln(q) versus ln(C), it would be linear in case of obeying the assumed empirical form of Freundlich Isotherm.
Langmuir Equilibrium Isotherm
Equilibrium relationship: [87,89]
(5.4)
Linear form:
(5.5)
By Langmuir model, a good fit of many experimental isotherms can be achieved just by optimum selection of equilibrium constants b (equilibrium constant = Kads/Kdes) and qs (saturation capacity). Higher values of b means higher adsorption rate relative to desorption rate and accordingly higher removal.
By plotting Ce/qt versus Ce, it would be linear in case of obeying Langmuir Equilibrium Isotherm.
Adsorption Kinetics
The following kinetic expressions were applied to check curve fit using relationships between mercury concentration in solid phase and time rearranged in linear forms;
Lagergren Pseudo-first order,
Pseudo-second order,
Elovich’s model,
Weber and Morris,
Diffusion-Chemisorption Model,
Linear Driving Force diffusion
Lagergren Pseudo-First Order
The kinetic expression: [97]
(qe – qt) (5.6)
Linear form:
Log(qe – qt) = log qe – (5.7)
Where qt is mercury concentration in solid phase Kg/m3 at time t and qe is equilibrium concentration and k is kinetic constant (1 / time unit).
By plotting Log(qe – qt) versus time, it would be linear curve in case of fitting the data to the kinetic expression.
Pseudo-Second Order Expression
The kinetic expression: [95,96]
(qe – qt)2 (5.8)
Linear Form:
(5.9)
Where initial adsorption rate h = , qt is mercury concentration in solid phase Kg/m3 at time t and qe is equilibrium concentration and k is kinetic constant (m3/Kg-hr).
By plotting versus time, it would be linear curve in case of fitting the data to the kinetic expression.
Elovich’s Model
The kinetic expression: [102]
(5.10)
Linear Form:
(5.11)
Where is initial adsorption rate, is desorption constant or the relationship between degree of surface coverage and activation energy of chemisorption, qt is mercury concentration in solid phase Kg/m3 at time t.
By plotting qt versus Ln (t), it would be linear curve in case of fitting the data to the kinetic expression.
Weber and Morris Model
The kinetic expression: [98]
qt = kid + C (5.12)
Where kid is diffusion rate constant = and C intercept indicates the external film thickness, where the higher C, the higher external film effects.
By plotting qt versus , it would be linear curve in case of fitting the data to the kinetic expression.
Diffusion-Chemisorption Model
The kinetic expression: [99]
n KDC t n-1(qe – qt) 2 / qe 2 (5.13)
Linear form with n = 0.5 as per Sutherland analysis: [99]
(5.14)
Where initial adsorption rate is expressed by
Ki = KDC2 / qe (5.15)
By plotting / qt versus , it would be linear curve in case of fitting the data to the kinetic expression.
Linear Driving Force diffusion Expression
The kinetic expression: [103]
Ln (1 – ) = k t (5.16)
Where is the fractional attainment of equilibrium q/qe and k is diffusion time constant.
By plotting Ln (1 – ) versus t, it would be linear curve in case of fitting the data to the kinetic expression.
Adsorption Modeling
The adsorption models were chosen based on the probability of matching the system available information with the assumptions that were proposed to derive the mathematical model expression.
The following models were considered to test the breakthrough curves using relationships between mercury concentration in gas phase and time rearranged in linear forms;
Zhang and Cheng Model,
tested as the model assumes linear isotherm and first order chemical reaction with simultaneous first order deactivation function by the effect of reaction product.
Wolborska Model,
tested to check the assumption of external mass transfer limitations.
Clark Model,
tested to check the assumption of mass transfer limitations and Fruendlich isotherm.
Thomas Model,
considered as it assumes Langmuir isotherm (favorable) and second order reversible reaction kinetics but not tested due to complexity of the model with cumbersome expressions and finding very bad fit with second order kinetic expression in addition to describing the chemisorption with irreversible process in manufacturer operation manual.
Bohart & Adams Model
tested as the model assumes one component adsorption with chemical reaction proposed by quasichemical expression and irreversible isotherm.
Zhang and Cheng Model
This model was developed for a catalytic reaction in a fixed bed packed with carbon, and was based on the following assumptions: [108]
Physical adsorption with linear isotherm
q = KiC (5.17)
Where Ki is the adsorption constant in (m3 gas/Kg adsorbent)
Simultaneous catalytic reaction assuming first order reaction equation
R = KCΦ (5.18)
Catalyst deactivation with covering active sites by reaction products. Assuming first order deactivation reaction and deactivation function is
Kd Φ (5.19)
and the solution is
Φ = exp (-Kd t) (5.20)
where R is the reaction rate (Kg/m3 – hr), K is the reaction rate constant (hr-1) , C is the adsorbate concentration in the gas stream (Kg/m3), t is time (hr), Φ is the deactivation function (dimensionless), and Kd (hr-1) is the deactivation rate constant.
Ideal plug flow with no axial dispersion
Isothermal adsorption
The continuity equation of the fixed bed
(5.21)
Where v is superficial velocity (m/hr), x is the distance from the bed inlet (m), ε is bed porosity (dimensionless), and q is the adsorbate concentration in the adsorbent (Kg adsorbate/m3 adsorbent).
The partial differential equation of continuity was solved using the boundary conditions:
At x = 0, t > 0, C = Co
At t = 0, x > 0, C = 0
The following solution was obtained:
(5.22)
Where
(5.23)
Ɵ = 1 + (5.24)
and L is the length of the bed (m).
By plotting versus t, it would be a linear curve in case the model is fitting the breakthrough data.
Wolborska Model
The model was developed by considering two regions in breakthrough curve where the adsorption front is migrating through the column in different ways. The model was developed to describe the low concentration region of the breakthrough curve in the range of C/Co from 10-5 to 0.05. It was assumed that adsorption rate is controlled by the external mass transfer resistance. [107]
The continuity equation:
(5.25)
where C is the adsorbate concentration in the gas phase (Kg adsorbate/m3 gas), t is time (s), v is the superficial velocity (m/s), q is the adsorbate concentration in the solid phase (Kg adsorbate/m3 adsorbent), D is axial diffusion coefficient, and x is the distance from the column inlet (m).
The initial condition at t = 0 is C (z,0) = 0, q (z,0) = 0.
The boundary conditions are at z = 0, C(0,t) = Co (inlet concentration), and at z = ∞, c(∞,t) = 0.
By introducing new variables and rearranging then continuity equation becomes:
(5.26)
and the initial and boundary conditions are C(x,0) = 0, q(x,0) = 0, C(0,τ) = Co, and C(∞,τ) = 0.
The kinetic equation for external resistance:
(5.27)
Where Ci is concentration at gas-solid interface and by assuming quick intraparticle diffusion then Ci << C and kinetic equation becomes:
(5.28)
Rearranging continuity equation gives:
(5.29)
The following was the linear solution form obtained by Wolborska:
(5.30)
By plotting versus t, it would be a linear curve in case the model is fitting the breakthrough data.
Clark Model
The model was developed based on the following assumptions: [106]
Mass balance over a finite element of the bed
(5.31)
Where J is the mass transfer rate per unit reactor volume (Kg adsorbate/(s-m3)), v is the superficial velocity of gas per unit of cross-sectional area (m/(s-m2)), A is the column cross sectional area (m2), and C is the gas phase adsorbate concentration into the differential element volume (Kg/m3), and z is the bed height (m).
Mass transfer limitations
(5.32)
Where Kt is the mass transfer coefficient in (s-1), and Ce is the equilibrium adsorbate concentration at the gas-solid interface (Kg adsorbate/m3 adsorbent).
Fruendlich isotherm for equilibrium
q = K C 1/n (5.33)
Where K is the equilibrium constant and 1/n is the slope of the isotherm.
By simplifying and rearranging, the final solution obtained by Clark becomes:
– 1] = -rt + LnA (5.34)
Where n does not equal one.
By plotting – 1] versus t, it would be a linear curve in case the model is fitting the breakthrough data.
Bohart & Adams Model
The model was developed based on the following assumptions: [94]
One trace component adsorption
Adsorption with simultaneous chemical reaction (Chemisorption)
Irreversible adsorption isotherm q = 0 at C = 0 & q = qs at C > 0
Ideal plug flow with negligible axial dispersion
Mass balance (continuity) equation for the fixed bed
(5.35)
Where C is the gas phase adsorbate concentration (Kg adsorbate/m3 gas), t is time (hr), v is superficial velocity of the gas stream (m/hr), x is distance from the column inlet (m), ε is bed porosity (dimensionless), and q is the adsorbate concentration in the adsorbent (Kg adsorbate/m3 adsorbent).
Adsorption kinetics described by quasichemical rate law
(5.36)
Where qs is the saturation capacity of q that corresponds to the equilibrium condition at the gas/adsorbent interface (Kg adsorbate/m3 adsorbent), and k is the kinetic constant. The rate of adsorption (quasichemical rate law) is proportional to the gas-phase adsorbate concentration, and to the remaining fraction of the adsorbent capacity, which still not occupied.
The differential continuity equation was solved by Cooney as below: [117]
(5.37)
Where
ᵵ = K Co (t – ) (5.38)
z = (5.39)
By rearranging, the following is the linear form of the model:
(5.40)
By plotting versus it would be a linear curve in case of fitting breakthrough data with the proposed model by Bohart & Adams.
Results and Discussions
Studying Adsorption Equilibrium
Both Langmuir and Freundlich isotherm relationships were applied to the available numerical data to get insight into adsorption equilibrium of the system by comparing the results and evaluating them.
Applying Freundlich Equilibrium Isotherm
By plotting Ln(q) versus ln(C), the following charts indicate the results of linear regression where it would be linear in case of obeying the assumed empirical form of Freundlich Isotherm.
Figure 6-1 Freundlich Isotherm Linear Regression for 4 mm Adsorbent
Figure 6-2 Freundlich Isotherm Linear Regression for 2 mm Adsorbent
For 4 mm particle size, best fit is obtained by setting 1/n from 0.05 to 0.0000001 which is indicative of extremely high favorable adsorption (irreversible isotherm) and for 2 mm particle size, best fit is obtained with 1/n = 0.0912 (n = 10.9649) which is indicative of also highly favorable adsorption.
Applying Langmuir Equilibrium Isotherm
By plotting Ce/qt versus Ce, the following charts indicate the results of linear regression where it would be linear in case of obeying Langmuir Equilibrium Isotherm.
Figure 6-3 Langmuir Isotherm Linear Regression for 4 mm Adsorbent
Figure 6-4 Langmuir Isotherm Linear Regression for 2 mm Adsorbent
Both particles shows ideal fit to Langmuir isotherm as expected due to matching system's information with Langmuir assumptions. R2 values are 0.9994 & 0.9996 and equilibrium constants "b" are calculated with obtaining much high values indicating irreversible isotherm while saturation capacity qs where calculated very close to the figures calculated from breakthrough data.
Table 6-1 Calculated Langmuir Constants
Studying Adsorption Kinetics
The following kinetic expressions were applied to check curve fit using relationships between mercury concentration in solid phase and time rearranged in linear forms;
Lagergren Pseudo-first order,
Pseudo-second order,
Elovich’s model,
Weber and Morris,
Diffusion-Chemisorption Model,
Linear Driving Force diffusion
Lagergren Pseudo-First Order Linear Regression
By plotting Log(qe – qt) versus time, the following charts indicate the results of linear regression where it would be linear curve in case of fitting the data to the kinetic expression.
Figure 6-5 Lagergren Pseudo-First Order Linear Regression for 4 mm Adsorbent
Figure 6-6 Lagergren Pseudo-First Order Linear Regression for 2 mm Adsorbent
As can be seen from plots, the expression gives moderate fit with R2 values of 0.9169 and 0.9462 at 4 mm and 2 mm particle size respectively.
Pseudo-Second Order Expression Linear Regression
By plotting versus time, the following charts indicate the results of linear regression where it would be linear curve in case of fitting the data to the kinetic expression.
Figure 6-7 Pseudo-Second Order Linear Regression for 4 mm Adsorbent
Figure 6-8 Pseudo-Second Order Linear Regression for 2 mm Adsorbent
As can be seen from plots, the expression gives poor fit with R2 values of 0.5904 and 0.7047 at 4 mm and 2 mm particle size respectively.
Elovich’s Model Linear Regression
By plotting qt versus Ln (t), the following charts indicate the results of linear regression where it would be linear curve in case of fitting the data to the kinetic expression.
Figure 6-9 Elovich Model Linear Regression for 4 mm Adsorbent
Figure 6-10 Elovich Model Linear Regression for 2 mm Adsorbent
As can be seen from plots, the expression gives moderate fit with R2 value of 0.9148 at 4 mm particle size and much better fit with R2 value of 0.9823 at 2 mm particle size. This may indicate that the chemical reaction (chemisorption) is not the sole rate-limiting step at 4 mm size while it is the main rate-limiting step at smaller particles of 2 mm.
Weber and Morris Model Linear Regression
By plotting qt versus , the following charts indicate the results of linear regression where it would be linear curve in case of fitting the data to the kinetic expression.
Figure 6-11 Weber & Morris Model Linear Regression for 4 mm Adsorbent
Figure 6-12 Weber & Morris Model Linear Regression for 2 mm Adsorbent
As can be seen from plots, the expression gives better fit with R2 values of 0.9786 and 0.9687 at 4 mm and 2 mm particle size respectively. This may indicate that intraparticle diffusion is a contributing rate-limiting step in both particles and more at the larger 4 mm particle.
Diffusion-Chemisorption Model Linear Regression
By plotting / qt versus , the following charts indicate the results of linear regression where it would be linear curve in case of fitting the data to the kinetic expression.
Figure 6-13 Diffusion-Chemisorption Model Linear Regression for 4 mm Adsorbent
Figure 6-14 Diffusion-Chemisorption Model Linear Regression for 2 mm Adsorbent
As can be seen from plots, the expression gives poor fit with R2 values of 0.7123 and 0.7567 at 4 mm and 2 mm particle size respectively.
Linear Driving Force diffusion Expression Linear Regression
By plotting Ln (1 – ) versus t, the following charts indicate the results of linear regression where it would be linear curve in case of fitting the data to the kinetic expression.
Figure 6-15 Vinod & Anirudvan (Linear Driving Force Diffusion) Model Linear Regression for 4 mm Adsorbent
Figure 6-16 Vinod & Anirudvan (Linear Driving Force Diffusion) Model Linear Regression for 2 mm Adsorbent
As can be seen from plots, the expression gives moderate fit with R2 values of 0.9169 and 0.9462 at 4 mm and 2 mm particle size respectively.
According to the above kinetics analysis, the most fitting kinetic expressions are Elovich model for chemisorption and Weber & Morris model for intraparticle diffusion as a rate-limiting step with neglecting external film resistance.
Studying Adsorption Mathematical Models
The following models were used to test the breakthrough curves using relationships between mercury concentration in gas phase and time rearranged in linear forms;
Zhang and Cheng Model,
Wolborska Model,
Clark Model,
Bohart & Adams Model
Linear Regression of Zhang and Cheng Model
As per the final linear form of solution, by plotting versus t, the following charts indicate the results of linear regression where it would be linear curve in case the model is fitting the breakthrough data.
Figure 6-17 Zhang & Cheng Model Linear Regression for 4 mm Adsorbent
Figure 6-18 Zhang & Cheng Model Linear Regression for 2 mm Adsorbent
As can be seen from plots, the model is not fitting the breakthrough data with R2 values of 0.8568 and 0.9206 at 4 mm and 2 mm particle size respectively.
Linear Regression of Wolborska Model
As per the final linear form of solution, by plotting versus t, the following charts indicate the results of linear regression where it would be linear curve in case the model is fitting the breakthrough data.
Figure 6-19 Wolborska Model Linear Regression for 4 mm Adsorbent
Figure 6-20 Wolborska Model Linear Regression for 2 mm Adsorbent
As can be seen from plots, the model is not giving the best fit to the breakthrough data with R2 values of 0.9639 and 0.8972 at 4 mm and 2 mm particle size respectively.
Linear Regression of Clark Model
As per the final linear form of solution, by plotting – 1] versus t, the following charts indicate the results of linear regression where it would be linear curve in case the model is fitting the breakthrough data.
Figure 6-21 Clark Model Linear Regression for 4 mm Adsorbent
Figure 6-22 Clark Model Linear Regression for 2 mm Adsorbent
For 4 mm particle best fit is obtained be setting n = 50 however it was calculated with much great value from Fruendlich isotherm plots and R2 value is 0.9639 while for 2 mm particle, best fit is obtained by setting n = 10.9649 as calculated from Fruendlich isotherm plots and R2 value is 0.9035.
Linear Regression of Bohart & Adams Model
As per the final linear form of solution, by plotting versus, the following charts indicate the results of linear regression where it would be linear curve in case of fitting breakthrough data to the proposed model by Bohart & Adams.
Figure 6-23 Bohart & Adams Model Linear Regression for 4 mm Adsorbent
Figure 6-24 Bohart & Adams Model Linear Regression for 2 mm Adsorbent
Bohart & Adams model indicates best fit to breakthrough data by R2 values of 0.9925 and 0.9843 at 4 mm and 2 mm particle size respectively. Accordingly, it can be used to describe the kinetics of mercury adsorption, simulate the bed performance, calculate predictions with different parameters, and draw concentration profiles and breakthrough curves.
The following table summarizes R2 values resulted from linear regression of the different models.
Table 6-2 R2 Values of Models Linear Regression
Bohart & Adams Model’s Predictability for the MRU System
After the remarkable success of Bohart & Adams model in best fitting to the breakthrough data by obtaining the highest R2 values of 0.9925 & 0.9843 in linear regression, the model has a considerable accuracy in prediction that can be utilized effectively in simulating the studied bed’s performance with different operating conditions. However, the following figures illustrate the model’s accuracy in prediction.
Figure 6-25 Bohart & Adams Model Prediction versus Actual Breakthrough Data for 4 mm Adsorbent
Figure 6-26 Bohart & Adams Model Prediction versus Actual Breakthrough Data for 2 mm Adsorbent
Calculations and Building a Simulation Tool
Generally, the following parameters are either directly known, or simply calculated from other known parameters using fundamental equations. Anyway, feed gas flow rate, velocity, operating temperature, operating pressure, quantity of adsorbents, particle density, particle diameter, bed density, bed diameter, bed length, bed porosity, particle porosity, pore volume, pore size, particle surface area, and, of course, the breakthrough profile of inlet and outlet mercury concentration with time, are all available for using in calculations. The breakthrough profile is shown in appendix-A section 10.1.
Initially, the available data were used to make consistency at all measuring units, then, by material balance and integration, calculated the mass of mercury adsorbed at time (t), cumulative mass of mercury adsorbed with time, mercury concentration in particle (loading) with time at both volume and mass basis (qt), and maximum particle fixation (saturation/equilibrium) capacity for mercury (qs or qmax or qe).
The available data of time (t) and corresponding mercury concentrations in gas (C) and solid (q) phases were used in the calculations of each mathematical expression of the studied equilibrium isotherms, kinetic equations, and adsorption models. The maximum particle fixation capacity for mercury (qs) is verified using the slope and the intercept of Langmuir plot.
By plotting versus as per the re-arranged linear form of Bohart & Adams model , the calculated slope of the line equals , and in turn, the kinetic constant K is calculated by knowing mercury inlet gas-phase concentration .
The two dimensionless parameters ᵵ & z are then calculated by knowing the kinetic constant K, mercury inlet gas-phase concentration , bed length L, gas velocity , maximum particle fixation capacity for mercury qs, and bed porosity using each parameter equation ᵵ = K Co (t – ) and z = .
Finally, predicted mercury concentration and breakthrough profile are determined using this equation with adequate time scale.
Finally, a simple simulation tool was created by excel workbook to facilitate the simulation of different case studies and study the effect of different parameters in much easier way, where only one data sheet is used for data entry (inputs), and one calculations sheet is automated by set of equations for each cell of all calculations with just manual selection for adequate time scale that is suitable for the data entered for examination (flow rate, bed height, and inlet mercury concentration). Then, the other sheets draw automatically all curves.
The following photos illustrate the use of the excel workbook.
Figure 6-27 Example of the Automated Use of Simulation Tool
Simulating Original versus Current Bed Design Performance
The bed is basically designed to treat a maximum of 1,350 MMSCFD feed gas flow rate contaminated with 20 µg/Nm3 of mercury. The following table summarizes the bed design basis.
Table 6-3 Design Basis of Mercury Guard Bed
The loading of the adsorbent actually ended at 2.48 meters bed height, and after start-up, the pressure drop across the bed increased to 2.75 bar at 1,165 MMSCFD and found 75% of the pressure drop is across the top layer as shown in below graph.
Figure 6-28 Pressure Drop Distribution across Layers of MRU Bed
Consequently, a decision taken and officially approved to skim the top 500 mm layer of the mercury adsorbent and reload instead 500 mm of 2" inert ceramic balls to get better gas distribution, and lower velocities, and vortices at the top of the catalyst.
By applying simulation of bed performance using Bohart & Adams model expressions, available bed information, data extracted from model linear regression, and mass balance calculations; the following graphs indicate the original bed height performance and lifetime versus the bed after skimming the top 500 mm layer.
Figure 6-29 Simulated Breakthrough Curves for Designed Bed Height and after Skimming of Top 500 mm
As shown from plots, at 1,350 MMSCFD feed gas contaminated with 20 µg / Nm3 of mercury; the model predicted breakthrough at 1,893 days (5.18 years) with bed height of 2.48 meters and 2-4 mm adsorbents, which is very close to manufacturer data sheet. After skimming, the model predicted breakthrough at 1,458 days (about 4 years) with bed height of 1.98 meters and same particles size.
Effect of Different Operating Parameters – Sensitivity Analysis and Case Studies
The following parameters where studied and evaluated its effects:
Adsorbent particle size
(With constant flow rate of 1350 MMSCFD, inlet mercury concentration 20 µg/Nm3, and bed height 1.98 meters)
Inlet mercury concentration
(With constant flow rate of 1350 MMSCFD, bed height 1.98 meters, and particles size of 2-4 mm)
Feed gas flow rate – velocity
(With constant inlet mercury concentration 20 µg/Nm3, bed height 1.98 meters, and particles size of 2-4 mm)
Bed Height
(With constant flow rate of 1350 MMSCFD, inlet mercury concentration 20 µg/Nm3, and particles size of 2-4 mm)
Adsorbent particle size
It is evident from plots that the smaller particle size of 2 mm exhibits much better performance and longer life time due to improvement of the mercury diffusivity into the active sites compared to 4 mm size by the effect of increasing surface area, decreasing pore depth, decreasing tortuosity which lead to decreasing intraparticle diffusion resistance, and increase adsorbent capacity. The model predicted breakthrough after 5.55 years, 3.98 years, and 2.25 years by using 2 mm, 2-4 range mm, and 4 mm particles respectively. However, due to pressure drop limitations; the bed is packed with a range of 2-4 mm adsorbents to compensate between the merits of small size particles (better mass transfer, longer life time) and the disadvantages of lower crush strength, lower bed porosity, and higher dust formation which all contribute to increase the pressure drop across the bed which is not desired in process.
Figure 6-30 Simulated Breakthrough Curves for Particle Size Effect
Inlet mercury concentration
It is evident from plots that the higher mercury concentration quickly propagates through the bed with saturating the adsorbents and earlier breakthrough occurs in contrast to lower concentration. Model predicted breakthrough after 2.63 years, 3.16 years, 3.98 years, 5.34 years at inlet mercury concentration of 30, 25, 20, 15 µg / Nm3 respectively, while breakthrough is expected after 16.5 years for concentrations below 5 µg / Nm3.
Figure 6-31 Simulated Breakthrough Curves for Inlet Mercury Concentration Effect
Feed gas flow rate – velocity
It is evident from plots that the effect of flow rate is similar to inlet mercury concentration where the higher flow rates quickly propagate through the bed with saturating the adsorbents and earlier breakthrough occurs in contrast to lower flow rates. Also increasing flow rate (velocity) lead to decreasing contact time between gas and solid phases and accordingly reduce the available chance of mercury molecules to transport from gas phase to the surface of the sorbent and then diffuse into the pores and react with active sites. However, only flow rates within the bed design range were evaluated by the model.
Breakthrough is predicted to occur after 3.98 years, 5.06 years, 6.34 years, 8.37 years, and 10.84 years at 1350, 1100, 900, 700, and 550 MMSCFD feed gas flow rate respectively.
Figure 6-32 Simulated Breakthrough Curves for Gas Flow Rate Effect
Bed Height
Higher bed length means two effects; the first one is much adsorbents quantity which lead to much available adsorption capacity and accordingly longer life time, while the second effect is increasing contact time between gas and solid phases and accordingly maximize the available chance of mercury molecules to transport from gas phase to the surface of the sorbent and then diffuse into the pores and react with active sites. However, various bed lengths are evaluated with the maximum of 2.48 meters for original bed design due to pressure drop limitations.
Model prediction of breakthrough to occur after 5.17 years, 3.98 years, 2.83 years, 1.99 years, 1.17 years, and 78 days at bed height of 2.48, 1.98, 1.5, 1.15, 0.8, and 0.4 meters respectively.
Figure 6-33 Simulated Breakthrough Curves for Bed Height Effect
Conclusion
The mercury adsorbent is proving a high mercury adsorption capacity as observed from manufacturer data, modeling results, and field performance.
It can be characterized by the following:
Physical adsorption in which equilibrium can be described by irreversible (rectangular) isotherm and can use Langmuir isotherm to represent it mathematically.
External film mass transfer resistance can be neglected.
The intraparticle diffusion is contributing as a rate-limiting step particularly with larger particles.
Chemical reaction (chemisorption) takes place with active sites leading to high fixation capacity and irreversibility (mercury cannot be desorbed).
The system can be represented with ideal plug flow by neglecting axial dispersion.
The adsorption can be described by isothermal as evident from field data.
The mercury concentration is in trace category.
All the above characteristics are in match with Bohart & Adams model assumptions, and accordingly proven after finding it best fitting to breakthrough data. Thus, the model can be used effectively to simulate bed performance and predict breakthrough curves and lifetime with different parameters.
Recommendations
In spite of strict design basis due to Mercury criticality, and proven high adsorption capacity; fortunately the current feed gas resources are contaminated with mercury in the range of 0.01 – 0.02 µg/Nm3 concentration, which is much below design load and also feed gas flow rate has often been below the 1,350 MMSCFD. The maximum throughput of 1,350 MMSCFD did not last for long time and the rate is declining naturally with resources depletion. Briefly, the bed is not loaded with mercury contamination and the model predicts too long years of lifetime.
This opportunity can be seized in the upcoming years by analyzing mercury in new potential natural gas resources and in case of the same range of mercury concentrations from 0.01 -0.02 µg/Nm3 or even higher up to the extent of 1 µg/Nm3 (1000 ng/Nm3); the opportunity can be seized by the following two options:
For upcoming charges of the same adsorbent, to purchase only a quantity sufficient for bed height of 400 mm and increase quantity of 1" ceramic balls to increase the height of the top ceramic balls layer from 150 mm to 700 mm. This option is an example for the overall idea and can be modified regarding the height and size of the adsorbents, or the height and size of ceramic balls. However, this example saves about 243,429.85 Euro by calculating the saving resulted from decreasing catalyst purchased quantity and also subtracting the extra charges of additional quantities of ceramic balls as per last available prices. This option has two main advantages of saving capital expenditures and also will decrease pressure drop across the bed and enhances flow distribution by more ceramic balls.
The second option is replacing the removed 1,600 mm height of mercury adsorbents by 1,600 mm of molecular sieves. This option also is an example for the overall idea and can be modified regarding overall bed configuration, packing arrangement, and size of adsorbents. However, this example is more attractive as it saves directly about 169,135.23 Euro by calculating the saving resulted from decreasing catalyst purchased quantity and also subtracting the extra charges of purchasing a quantity of molecular sieves as per last available prices. The quantity of the molecular sieves is equivalent to 81.34% of the quantity packed in one of the upstream dryer beds. This molecular sieves bed will act as a guard for water breakthrough cases that occur occasionally by many reasons such as process upsets in plant, process upsets in upstream conditioning plants, or upsets in regeneration, high ambient temperatures in summer, and other various causes. Normally, when water breakthrough occurs, it is always sudden conditions and proper action is taken after short time either by skipping some of regeneration time or by decreasing feed gas flow rate. Such water breakthrough cases accumulate hydrate with time in cryogenic sections like strainers and exchangers until being difficult to be dissolved by methanol, and affect plant productivity, and then a dry-out operation for 36-48 hours is mandatory to remove all accumulated hydrate in cryogenic units. This molecular sieve guard is accordingly will save indirectly the high revenue losses of stopping production during 36-48 hours of dry-out as the need of dry-out will be postponed for many years till saturating this guard bed which is not normally loaded with water content as the maximum breakthrough water content will be 0.1 – 0.2 ppm(v) and last for maximum of 1 hour and occurs occasionally during the year while the molecular sieve dryers are designed to accommodate much continuous high water content of 300 – 400 ppm(v) for 22 -24 hours with outlet dry gas of 0.01 ppm(v). This guard also has an advantage of being not exposed to aging by hydrothermal conditions of regeneration. By approximate calculations for maximum continuous feed gas flow rate of 1,350 MMSCFD and maximum breakthrough water content of 0.2 ppm(v), the example quantity of molecular sieves can accommodate this continuous load for 530 days of operation by assuming sieves capacity of 0.126 Kg water/Kg sieves and 662 days with assuming capacity of 0.157 Kg water/Kg sieves. These rough calculations just prove that the need of dry-out operation will be postponed for many years.
By using the model for prediction of the shortened bed (400 mm) performance for mercury removal with maximum feed gas flow rate of 1,350 MMSCFD and different inlet mercury concentrations while using the same particle size of 2-4 mm; the following plots resulted from simulation indicates predicted mercury breakthrough after 35 years, 10 years, 4.4 years, 1.4 years, and 78 days with inlet mercury concentration of 0.35 µg/Nm3 (350 ng/Nm3), 1, 2, 5, and 20 µg/Nm3. Consequently the modification of the bed configuration to be 400 mm height of mercury adsorbents and 1,600 mm of molecular sieves as an example; would work efficiently for both mercury removal up to inlet concentration of 2 µg/Nm3 (2,000 ng/Nm3) and capturing all water breakthrough as a guard for years based on the maximum feed gas flow rate of 1,350 MMSCFD.
Figure 8-1 Simulated Breakthrough Curves for Proposed Bed Height with Different Inlet Mercury Concentrations
Figure 8-2 Simulated Breakthrough Curves till Saturation for Proposed Bed Height with Different Inlet Mercury Concentrations
References
"hydrargyrum". Random House Webster's Unabridged Dictionary.
Ehrlich, H.L. and Newman, D.K. eds., 2008. Geomicrobiology. CRC Press. p. 265. ISBN 978-0-8493-7906-2.
"Mercury and the environment — Basic facts". Environment Canada, Federal Government of Canada. 2004. Retrieved 27 March2008.
"Mercury — Element of the ancients". Center for Environmental Health Sciences, Dartmouth College. Retrieved 9 April 2012.
Burkholder, M. and Johnson, L., 2008. Colonial Latin America. OUP Catalogue. pp. 157–159. ISBN 0-19-504542-4.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 7, paragraphs 100 to 103
Lide, D.R. ed., 2005. CRC handbook of chemistry and physics (Vol. 86). CRC press.
http://www.ptable.com/#Property/State
Senese, F., 2007. Why is mercury a liquid at STP?. General Chemistry Online at Frostburg State University. Retrieved May, 1.
Norrby, L.J., 1991. Why is mercury liquid? Or, why do relativistic effects not get into chemistry textbooks?. J. Chem. Educ, 68(2), p.110.
Lide, D. R., ed., 2005. CRC Handbook of Chemistry and Physics (Vol. 86). Boca Raton (FL): CRC Press. pp. 4.125–4.126. ISBN 0-8493-0486-5.
Rustad, D.S., 1987. How soft is mercury?. J. Chem. Educ, 64(5), p.470.
Greenwood, N.N. and Earnshaw, A., 1997. Chemistry of the Elements 2nd Edition. Butterworth-Heinemann. ISBN 0-08-037941-9.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 2, paragraphs 39 to 43
Gmelin, Leopold, 1852. Hand book of chemistry. Cavendish Society. pp. 103 (Na), 110 (W), 122 (Zn), 128 (Fe), 247 (Au), 338 (Pt). Retrieved 30 December2012.
Soratur, S.H., 2002. Essentials of dental materials. Jaypee Bros. Medical Publishers. p. 14. ISBN 978-81-7179-989-3.
Barbalace, K., 1995. Periodic table of elements: Hg- Mercury. EnvironmentalChemistry.com, 2016.
Wang, X., Andrews, L., Riedel, S. and Kaupp, M., 2007. Mercury is a transition metal: the first experimental evidence for HgF4. Angewandte Chemie International Edition, 46(44), pp.8371-8375.
Goyer, R., Aposhian, V., Arab, L., Bellinger, D., Burbacher, T., Burke, T., Jacobson, J., Knobeloch, L., Stern, A. and Ryan, L., 2000. Toxicological effects of methylmercury. Washington, DC: National Research Council.
National Research Council (U.S.) – Board on Environmental Studies and Toxicology, 2000. Toxicological effects of methylmercury. National Academies Press.
http://www.chemistryexplained.com/elements/L-P/Mercury.html
Surmann, P. and Zeyat, H., 2005. Voltammetric analysis using a self-renewable non-mercury electrode. Analytical and Bioanalytical Chemistry,383(6), pp.1009-1013.
http://www.lenntech.com/periodic/elements/hg.htm
Rogalski, A., 2000. Infrared Detectors (Vol. 10). CRC Press, pp.507.
Krabbenhoft, D. and Schuster, P., 2002. Glacial ice cores reveal a record of natural and anthropogenic atmospheric mercury deposition for the last 270 years (No. 051-02).
Pacyna, E.G., Pacyna, J.M., Steenhuisen, F. and Wilson, S., 2006. Global anthropogenic mercury emission inventory for 2000. Atmospheric environment, 40(22), pp.4048-4063.
"What is EPA doing about mercury air emissions?". United States Environmental Protection Agency (EPA). Retrieved 1 May 2007.
Solnit, R., 2007. Winged mercury and the golden calf. Orion Magazine. Retrieved, pp.12-03.
"Mercury-containing Products". United States Environmental Protection Agency (EPA). Retrieved 1 May 2007.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 2, paragraphs 44 to 46 and 48
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 2, paragraphs 49 to 52
http://www.usgs.gov/themes/factsheet/146-00/
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 3, paragraphs 55 & 56, Chapter 2, paragraph 47.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 6, paragraphs 77 to 80
Cocoros, G., Cahn, P.H. and Siler, W., 1973. Mercury concentrations in fish, plankton and water from three Western Atlantic estuaries. Journal of Fish Biology, 5(6), pp.641-647.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 3 & 4, paragraphs 61, 62 and 66
"Mercury: Spills, Disposal and Site Cleanup". Environmental Protection Agency. Retrieved 11 August 2007.
Ngim, C.H., Foo, S.C., Boey, K.W. and Jeyaratnam, J., 1992. Chronic neurobehavioural effects of elemental mercury in dentists. British journal of industrial medicine, 49(11), pp.782-790.
Liang, Y.X., Sun, R.K., Sun, Y., Chen, Z.Q. and Li, L.H., 1993. Psychological effects of low exposure to mercury vapor: application of a computer-administered neurobehavioral evaluation system. Environmental Research, 60(2), pp.320-327.
McFarland, R.B. and Reigel, H., 1978. Chronic mercury poisoning from a single brief exposure. Journal of Occupational and Environmental Medicine, 20(8), pp.532-534.
Mercury, W.I., 1991. Environmental Health Criteria 118. Geneva: World Health Organization, 107.
World Health Organization, 1991. Inorganic mercury—Environmental health criteria. WHO, Geneva, Switzerland, 118.
Senn, E., 1996. Controlling metallic mercury exposure in the workplace: A guide for employers. DIANE Publishing.
Rice, K.M., Walker, E.M., Wu, M., Gillette, C. and Blough, E.R., 2014. Environmental mercury and its toxic effects. Journal of preventive medicine and public health, 47(2), pp.74-83.
http://www.pca.state.mn.us/index.php/topics/mercury/health-and-environmental-effects-of-mercury.html
Sheet, H.S.F., 1996. New Jersey Department of health and senior services.Trenton, NJ.
Jurng, J., Lee, T.G., Lee, G.W., Lee, S.J., Kim, B.H. and Seier, J., 2002. Mercury removal from incineration flue gas by organic and inorganic adsorbents. Chemosphere, 47(9), pp.907-913.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 6, paragraphs 87 to 90
Zhuang, Y., Zygarlicke, C.J., Galbreath, K.C., Thompson, J.S., Holmes, M.J. and Pavlish, J.H., 2004. Kinetic transformation of mercury in coal combustion flue gas in a bench-scale entrained-flow reactor. Fuel processing technology, 85(6), pp.463-472.
Pacyna, E.G. and Pacyna, J.M., 2002. Global emission of mercury from anthropogenic sources in 1995. Water, Air, and Soil Pollution, 137(1-4), pp.149-165.
Ling, L.I.U., DUAN, Y.F., WANG, Y.J., Hui, W.A.N.G. and YIN, J.J., 2010. Experimental study on mercury release behavior and speciation during pyrolysis of two different coals. Journal of Fuel Chemistry and Technology,38(2), pp.134-139.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 8
"Minamata Convention Agreed by Nations". United Nations Environment Program. Retrieved 19 January 2013.
http://www.un.org/apps/news/story.asp?NewsID=43963&Cr=mercury&Cr1=#.UPt4ikfdP5M
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 8, paragraphs 112 to 115
http://www.globalcement.com/magazine/articles/845-global-cement-emissions-standards
http://www.who.int/mediacentre/factsheets N°361
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 9, paragraphs 128 to 133
Lee, J.D., 2008. Concise inorganic chemistry. John Wiley & Sons.
Waldron, H.A., 1983. Did the Mad Hatter have mercury poisoning?. British medical journal (Clinical research ed.), 287(6409), p.1961.
"Introduction". Y-12 Mercury Task Force Files: A Guide to Record Series of the Department of Energy and its Contractors. United States Department of Energy.
"Minamata Disease The History and Measures". Ministry of the Environment, Government of Japan. Retrieved 7 July 2009.
Normile, D., 2013. In Minamata, mercury still divides. Science, 341(6153), pp.1446-1447.
Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report., Summary of the Report, Chapter 3, paragraphs 55 & 56; and Chapter 2, paragraph 47.
Rashad, M., Darwish, K.M., Salam, H.E.A. and Shalaby, E.A., 2013. Environmental monitoring of mercury in urban soils, air and plants of a Mediterranean area, Alexandria, Egypt. Int J Energy Environ, 4, pp.311-320.
Mohamed A, S., Abdel, S.A., Asia Abd El Samie, T., El, M.M.A. and Hamada BI, H., 2012. Mercury and Methyl Mercury in Sediments of Northern Lakes-Egypt. Journal of Environmental Protection, 2012.
Assassa, M.F., Mohammd, Y.S., Aglan, M.A. and Farahat, M.I., 2015. ESTIMATION OF HAIR MERCURY LEVELS FOR ASSESSMENT OF METHYL MERCURY EXPOSURE IN ALDAKAHLIA-EGYPT. AL-AZHAR ASSIUT MEDICAL JOURNAL, 13(1).
El-Nady, F.E., 1986. Bioaccumulation of mercury in some coastal marine fish from Alexandria waters. In The FAO/ UNEP/WHO/IOC/IAEA Meeting on the Biogeochemical Cycle of Mercury in the Mediterranean, Siena, Italy, 27-31 August 1984, FAO FISH. REP. no. 325 (pp. 74-77).
http://www.greenprophet.com/2011/10/egypt-mercury-disposal/
Agency for Toxic Substances and Disease Registry, "Mercury"-http://www.atsdr.cdc.gov/tfacts46. html
API, R., 2003. 571–Damage Mechanisms Affecting Fixed Equipment in the Refining Industry. American Petroleum Institute, pp.30-53.
McIntyre, D.R., Case, R.P., Ballantyne, T. and Woodward, S., 2013, March. Environmentally Assisted Cracking of Nickel Steels In Liquid Mercury, Hydrogen and Methanol. In CORROSION 2013. NACE International.
Al-Alf, H.K.A. and Fl-Behairy, S., Mercury Removal from Natural Gas in Egypt.
Shafawi, A., Ebdon, L., Foulkes, M., Stockwell, P. and Corns, W., 1999. Determination of total mercury in hydrocarbons and natural gas condensate by atomic fluorescence spectrometry. Analyst, 124(2), pp.185-189.
Im abbott, Paul openshaw, "Mercury Removal Technology and Its Application", Synetix, Belasis Av., Billmgham, Cleveland, TS23 1LB, UK
"Moomba Plant Update" (Press release). Adelaide, South Australia: Santos. 2004-03-05. Archived from the original on 2004-03-05. Retrieved 2013-01-18.
http://EnvironmentalChemistry.com/yogi/periodic/Hg.html
Kolman, D.G., Environmentally Induced Cracking. Liquid Metal Embrittlement, ASM Handbook, 13, pp.381-392.
Robertson, W.M., 1966. PROPAGATION OF A CRACK FILLED WITH LIQUID METAL. North American Aviation Science Center, Thousand Oaks, Calif..
Glikman, E.É. and Goryunov, Y.V., 1978. Mechanism of embrittlement by liquid metals and other manifestations of the rebinder effect in metal systems. Materials Science, 14(4), pp.355-364.
Stoloff, N.S. and Johnston, T.L., 1963. Crack propagation in a liquid metal environment. Acta Metallurgica, 11(4), pp.251-256.
Westwood, A.R.C. and Kamdar, M.H., 1963. Concerning liquid metal embrittlement, particularly of zinc monocrystals by mercury. Philosophical Magazine, 8(89), pp.787-804.
Gordon, P. and An, H.H., 1982. The mechanisms of crack initiation and crack propagation in metal-induced embrittlement of metals. Metallurgical Transactions A, 13(3), pp.457-472.
Lynch, S.P., 1988. Environmentally assisted cracking: overview of evidence for an adsorption-induced localised-slip process. Acta Metallurgica, 36(10), pp.2639-2661.
Popovich, V.V. and Dmukhovskaya, I.G., 1987. The embrittlement of metals and alloys being deformed in contact with low-melting alloys (a review of foreign literature). Materials Science, 23(6), pp.535-544.
http://www.twi-global.com/technical-knowledge/faqs/material-faqs/faq-what-is-liquid-metal-embrittlement-and-what-are-common-embrittling-metal-couples/
Ruthven, D.M., 1984. Principles of adsorption and adsorption processes. John Wiley & Sons.
Srivastava, V.C., Mall, I.D. and Mishra, I.M., 2008. Adsorption of toxic metal ions onto activated carbon: Study of sorption behaviour through characterization and kinetics. Chemical Engineering and Processing: Process Intensification, 47(8), pp.1269-1280.
Langmuir, I., 1918. The adsorption of gases on plane surfaces of glass, mica and platinum. Journal of the American Chemical society, 40(9), pp.1361-1403.
Freundlich, H.M.F., 1906. Over the adsorption in solution. J. Phys. Chem,57(385471), pp.1100-1107.
Zhao, G., Wu, X., Tan, X. and Wang, X., 2010. Sorption of heavy metal ions from aqueous solutions: a review. The open colloid science journal, 4(1).
Temkin, M.I. and Pyzhev, V., 1940. Kinetics of ammonia synthesis on promoted iron catalysts. Acta physiochim. URSS, 12(3), pp.217-222.
Yang, R.T., 1987. Gas separation by adsorption processes. Butterworth-Heinemann.
Bohart, G.S. and Adams, E.Q., 1920. Some aspects of the behavior of charcoal with respect to chlorine. 1. Journal of the American Chemical Society, 42(3), pp.523-544.
Ho, Y.S. and McKay, G., 1998. Kinetic models for the sorption of dye from aqueous solution by wood. Process Safety and Environmental Protection,76(2), pp.183-191.
Ho, Y.S. and McKay, G., 1998. Kinetic model for lead (II) sorption on to peat.Adsorption science & technology, 16(4), pp.243-255.
Lagergren, S., 1898. About the theory of so-called adsorption of soluble substances.
Weber, W.J. and Morris, J.C., 1963. Kinetics of adsorption on carbon from solution. Journal of the Sanitary Engineering Division, 89(2), pp.31-60.
Sutherland, C., 2009. Removal of heavy metals from water using low-cost adsorbents: process development (Doctoral dissertation).
Low, M.J.D., 1960. Kinetics of Chemisorption of Gases on Solids. Chemical Reviews, 60(3), pp.267-312.
Chien, S.H. and Clayton, W.R., 1980. Application of Elovich equation to the kinetics of phosphate release and sorption in soils. Soil Science Society of America Journal, 44(2), pp.265-268.
Zeldowitsch, J., 1934. Über den mechanismus der katalytischen oxydation von CO an MnO2. Acta physicochim. URSS, 1, pp.364-449.
Vinod, V.P. and Anirudhan, T.S., 2002. Sorption of tannic acid on zirconium pillared clay. Journal of Chemical Technology and Biotechnology, 77(1), pp.92-101.
Thomas, H.C., 1944. Heterogeneous ion exchange in a flowing system.Journal of the American Chemical Society, 66(10), pp.1664-1666.
Yoon, Y.H. and NELSON, J.H., 1984. Application of gas adsorption kinetics I. A theoretical model for respirator cartridge service life. The American Industrial Hygiene Association Journal, 45(8), pp.509-516.
Clark, R.M., 1987. Evaluating the cost and performance of field-scale granular activated carbon systems. Environmental science & technology,21(6), pp.573-580.
Wolborska, A., 1989. Adsorption on activated carbon of p-nitrophenol from aqueous solution. Water Research, 23(1), pp.85-91.
Zhang, H. and Cheng, D., 2000. Mathematical model for a fixed bed adsorptive reactor. Carbon, 38(6), pp.877-880.
Deokar, S.K. and Mandavgane, S.A., 2015. Estimation of packed-bed parameters and prediction of breakthrough curves for adsorptive removal of 2, 4-dichlorophenoxyacetic acid using rice husk ash. Journal of Environmental Chemical Engineering, 3(3), pp.1827-1836.
Chung, S.T., Kim, K.I. and Yun, Y.R., 2009. Adsorption of elemental mercury vapor by impregnated activated carbon from a commercial respirator cartridge. Powder Technology, 192(1), pp.47-53.
Ren, J., Zhou, J., Luo, Z., Zhong, Y. and Xu, Z., 2007. Fixed-bed experiments and mathematical modeling for adsorption of mercury vapor. InChallenges of Power Engineering and Environment (pp. 843-849). Springer Berlin Heidelberg.
Graydon, J.W., Zhang, X., Kirk, D.W. and Jia, C.Q., 2009. Sorption and stability of mercury on activated carbon for emission control. Journal of hazardous materials, 168(2), pp.978-982.
Camargo, C.L.M., de Resende, N.S., de Oliveira, A.G., Salim, V.M.M. and Tavares, F.W., 2014. Investigation of adsorption-enhanced reaction process of mercury removal from simulated natural gas by mathematical modeling.Fuel, 129, pp.129-137.
Bingham, M.D., 1990. Field detection and implications of mercury in natural gas. SPE Production Engineering, 5(02), pp.120-124.
Spiric, Z., 2001. Innovative approach to the mercury control during natural gas processing. Proceedings of ETCE.
Wilhelm, S.M. and Bloom, N., 2000. Mercury in petroleum. Fuel Processing Technology, 63(1), pp.1-27.
Cooney, D.O., 1999. Adsorption design for wastewater treatment, CRC Pres.INC., Boca Raton, Florida, USA.
Yang, H., Xu, Z., Fan, M., Bland, A.E. and Judkins, R.R., 2007. Adsorbents for capturing mercury in coal-fired boiler flue gas. Journal of Hazardous Materials, 146(1), pp.1-11.
Hsi, H.C. and Chen, C.T., 2012. Influences of acidic/oxidizing gases on elemental mercury adsorption equilibrium and kinetics of sulfur-impregnated activated carbon. Fuel, 98, pp.229-235.
Sasmaz, E., Kirchofer, A., Jew, A.D., Saha, A., Abram, D., Jaramillo, T.F. and Wilcox, J., 2012. Mercury chemistry on brominated activated carbon.Fuel, 99, pp.188-196.
Wade, C.B., Thurman, C., Freas, W., Student, J., Matty, D. and Mohanty, D.K., 2012. Preparation and characterization of high efficiency modified activated carbon for the capture of mercury from flue gas in coal-fired power plants. Fuel processing technology, 97, pp.107-117.
Guo, P., Guo, X. and Zheng, C., 2010. Roles of y-Fe 2 O 3 in fly ash for mercury removal: results of density functional theory study. Applied Surface Science, 256(23), pp.6991-6996.
Baltrus, J.P., Granite, E.J., Pennline, H.W., Stanko, D., Hamilton, H., Rowsell, L., Poulston, S., Smith, A. and Chu, W., 2010. Surface characterization of palladium–alumina sorbents for high-temperature capture of mercury and arsenic from fuel gas. Fuel, 89(6), pp.1323-1325.
Skodras, G., Diamantopoulou, I., Natas, P., Orfanoudaki, T., Amarantos, P., Stavropoulos, G.G. and Sakellaropoulos, G.P., 2003, September. Mercury and PCBs adsorption on activated carbons: Experimentation and modelling. In Proc. 8th Int. Conf. Env. Sci. & Technol (pp. 8-10).
Meserole, F.B., Chang, R., Carey, T.R., Machac, J. and Richardson, C.F., 1999. Modeling mercury removal by sorbent injection. Journal of the Air & Waste Management Association, 49(6), pp.694-704.
Chen, S., Rostam-Abadi, M. and Chang, R., 1996. Mercury removal from combustion flue gas by activated carbon injection: mass transfer effects.PREPRINTS OF PAPERS-AMERICAN CHEMICAL SOCIETY DIVISION FUEL CHEMISTRY, 41, pp.442-446.
Rostam-Abadi, M., Chen, S.G., Hsi, H.C., Rood, M., Chang, R., Carey, T., Hargrove, B., Richardson, C., Rosenhoover, W. and Meserole, F., 1997. Novel vapor phase mercury sorbents. Proceedings of the EPRI-DOE-EPA Combined Utility Air Pollutant Control, pp.25-29.
Serre, S.D., Gullett, B.K. and Ghorishi, S.B., 2001. Entrained-flow adsorption of mercury using activated carbon. Journal of the Air & Waste Management Association, 51(5), pp.733-741.
Scala, F., 2001. Simulation of mercury capture by activated carbon injection in incinerator flue gas. 1. In-duct removal. Environmental science & technology, 35(21), pp.4367-4372.
Scala, F., 2001. Simulation of mercury capture by activated carbon injection in incinerator flue gas. 2. Fabric filter removal. Environmental science & technology, 35(21), pp.4373-4378.
Flora, J.R., Hargis, R.A., O’Dowd, W.J., Pennline, H.W. and Vidic, R.D., 2003. Modeling sorbent injection for mercury control in baghouse filters: I—model development and sensitivity analysis. Journal of the Air & Waste Management Association, 53(4), pp.478-488.
Flora, J.R., Hargis, R.A., O’Dowd, W.J., Pennline, H.W. and Vidic, R.D., 2003. Modeling sorbent injection for mercury control in baghouse filters: II—Pilot-scale studies and model evaluation. Journal of the Air & Waste Management Association, 53(4), pp.489-496.
Ho, Y.S. and McKay, G., 1998. A comparison of chemisorption kinetic models applied to pollutant removal on various sorbents. Process Safety and Environmental Protection, 76(4), pp.332-340.
Sutherland, C. and Venkobachar, C., 2010. A diffusion-chemisorption kinetic model for simulating biosorption using forest macro-fungus, fomes fasciatus.International Research Journal of Plant Science, 1(4), pp.107-117.
Odoemelam, S.A., Onwu, F.K., Sonde, C.U. and Chinedu, M.A., 2014. Adsorption Kinetics of Cd (ll) and Pb (ll) Ions from Aqueous Solutions by Bamboo-Based Activated Charcoal and Bamboo Dust. Orbital-The Electronic Journal of Chemistry, 6(4), pp.223-232.
Gusmão, K.A.G., Gurgel, L.V.A., Melo, T.M.S. and Gil, L.F., 2013. Adsorption studies of methylene blue and gentian violet on sugarcane bagasse modified with EDTA dianhydride (EDTAD) in aqueous solutions: kinetic and equilibrium aspects. Journal of environmental management, 118, pp.135-143.
Singh, T.S. and Pant, K.K., 2006. Kinetics and mass transfer studies on the adsorption of arsenic onto activated alumina and iron oxide impregnated activated alumina. Water quality research journal of Canada, 41(2), pp.147-156.
Ho, N., 2012. Modeling hydrogen sulfide adsorption by activated carbon made from anaerobic digestion by-product (Doctoral dissertation).
Sugier, A. and Villa, F.L., Institut Francais Du Petrole, 1978. Process for removing mercury from a gas or a liquid by absorption on a copper sulfide containing solid mass. U.S. Patent 4,094,777.
YU, X., ZHU, L., GUO, B. and HE, S., 2008. Adsorption of mercury on laterite from Guizhou Province, China. Journal of Environmental Sciences,20(11), pp.1328-1334.
Luo, J., Hein, A.M. and Hwang, J.Y., 2004. Adsorption of vapor phase mercury on various carbons. Journal of Minerals and Materials Characterization and Engineering, 3(01), p.13.
Eswaran, S., Stenger, H.G. and Fan, Z., 2007. Gas-phase mercury adsorption rate studies. Energy & fuels, 21(2), pp.852-857.
Steijns, M., Peppelenbos, A. and Mars, P., 1976. Mercury chemisorption by sulfur adsorbed in porous materials. Journal of colloid and interface science,57(1), pp.181-186.
Furuta, A., Sato, K., Sato, K., Matsuzawa, T. and Ito, H., Jcg Corporation, 1991. Process for removal of mercury from a liquid hydrocarbon. U.S. Patent 5,037,552.
Panagiotou, T., Momncy, J.R. and Senior, C.L., 2000, August. ZEOLITE-BASED MERCURY SORBENT-LABORATORY TESTING AND MODELING. In ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY(Vol. 220, pp. U386-U386). 1155 16TH ST, NW, WASHINGTON, DC 20036 USA: AMER CHEMICAL SOC.
Vidic, R.D., 2002. Adsorption of elemental mercury by virgin and impregnated activated carbon (pp. 15-44). Lewis Publishers: Boca Raton, FL.
Ghorishi, B. and Gullett, B.K., 1998. Experimental study on mercury sorption by activated carbons and calcium hydroxide. US Environmental Protection Agency, Air Pollution Prevention and Control Division.
Huawei, Z., Xiuli, L., Li, W. and Peng, L., 2014. Characteristics and Stability of Mercury Vapor Adsorption over Two Kinds of Modified Semicoke. The Scientific World Journal, 2014.
Borderieux, S., Wu, C.Y., Bonzongo, J.C. and Powers, K., 2004. Control of elemental mercury vapor in combustion systems using Fe2O3 nanoparticles.Aerosol Air Qual. Res, 4(1), pp.74-90.
Liu, Y., 2014. Elemental Mercury Removal from Flue Gas by Metal/Metal Oxide Decorated Graphene Oxide Composites (Doctoral dissertation, University of Alberta).
Holmes, M.J., Pavlish, J.H., Zhuang, Y., Benson, S.A. and Fritze, M.J., 2004. Pilot-scale evaluation of activated carbon-based mercury control options for utilities burning lignite coal. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem, 49(1), p.281.
Fryxell, G.E., Skaggs, R. and Parker, K.E., 2004, March. Novel nanoporous sorbents for removal of mercury. In ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY (Vol. 227, pp. U1088-U1089). 1155 16TH ST, NW, WASHINGTON, DC 20036 USA: AMER CHEMICAL SOC.
Crocker, C.R., Benson, S.A., Holmes, M.J., Zhuang, Y., Pavlish, J.H. and Galbreath, K.C., 2004. Comparison of sorbents and furnace additives for mercury control in low-rank fuel combustion systems. Prepr. Pap.-Am. Chem. Soc., Div. Fuel Chem, 49(1), p.289.
Wdowin, M., Wiatros-Motyka, M.M., Panek, R., Stevens, L.A., Franus, W. and Snape, C.E., 2014. Experimental study of mercury removal from exhaust gases. Fuel, 128, pp.451-457.
Ballestero, D., Gómez-Giménez, C., García-Díez, E., Juan, R., Rubio, B. and Izquierdo, M.T., 2013. Influence of temperature and regeneration cycles on Hg capture and efficiency by structured Au/C regenerable sorbents. Journal of hazardous materials, 260, pp.247-254.
Olson, E.S., Dunham, G.E., Sharma, R.K. and Miller, S.J., 2000. Mechanisms of mercury capture and breakthrough on activated carbon sorbents. ACS Fuels, 45, pp.886-889.
Wilcox, J., Rupp, E., Ying, S.C., Lim, D.H., Negreira, A.S., Kirchofer, A., Feng, F. and Lee, K., 2012. Mercury adsorption and oxidation in coal combustion and gasification processes. International Journal of Coal Geology, 90, pp.4-20.
Daza, L., Mendioroz, S. and Pajares, J., 1991. Mercury adsorption by sulfurized fibrous silicates. Clays and Clay Minerals, 39(1), pp.14-21.
Padak, B., 2011. Mercury reaction chemistry in combustion flue gases from experiments and theory (Doctoral dissertation, Stanford University).
De, M., Azargohar, R., Dalai, A.K. and Shewchuk, S.R., 2013. Mercury removal by bio-char based modified activated carbons. Fuel, 103, pp.570-578.
Bae, K.M., Kim, B.J. and Park, S.J., 2014. Overlook of carbonaceous adsorbents and processing methods for elemental mercury removal. Carbon letters, 15(4), pp.238-246.
Bisson, T.M., Ong, Z.Q., MacLennan, A., Hu, Y. and Xu, Z., 2015. Impact of Sulfur Loading on Brominated Biomass Ash on Mercury Capture. Energy & Fuels, 29(12), pp.8110-8117.
O'Dowd, W.J., Hargis, R.A., Granite, E.J. and Pennline, H.W., 2004. Recent advances in mercury removal technology at the National Energy Technology Laboratory. Fuel Processing Technology, 85(6), pp.533-548.
Xie, J., Xu, H., Qu, Z., Huang, W., Chen, W., Ma, Y., Zhao, S., Liu, P. and Yan, N., 2014. Sn–Mn binary metal oxides as non-carbon sorbent for mercury removal in a wide-temperature window. Journal of colloid and interface science, 428, pp.121-127.
You, X., 2002. Sulfur Impregnation of Activated Carbon Through Hydrogen Sulfide Oxidation (Doctoral dissertation, University of Pittsburgh).
Izquierdo, M.T., Ballestero, D., Juan, R., García-Díez, E., Rubio, B., Ruiz, C. and Pino, M.R., 2011. Tail-end Hg capture on Au/carbon-monolith regenerable sorbents. Journal of hazardous materials, 193, pp.304-310.
Abad-Valle, P., Lopez-Anton, M.A., Diaz-Somoano, M. and Martinez-Tarazona, M.R., 2011. The role of unburned carbon concentrates from fly ashes in the oxidation and retention of mercury. Chemical engineering journal, 174(1), pp.86-92.
Eldefrawy, M.M., Kenawy, I.M. and El-Tabey, R.M., 2014. Application of Nanosized Mesoporous Aluminosilicates for Adsorption of Hg from Aqueous Solutions: Kinetic, Isotherm and Thermodynamic Studies. International Journal, 2(12), pp.764-781.
Zakria, M.H., Omar, A.A. and Bustam, M.A., 2016. Mercury Removal of Fluctuating Ethane Feedstock in a Large Scale Production by Sulphur Impregnated Activated Carbon. Procedia Engineering, 148, pp.561-567.
Shafawi, A., Ebdon, L., Foulkes, M., Stockwell, P. and Corns, W., 2000. Preliminary evaluation of adsorbent-based mercury removal systems for gas condensate. Analytica Chimica Acta, 415(1), pp.21-32.
Perry, R.H. and Green, D.W., 1999. Perry's chemical engineers' handbook. McGraw-Hill Professional.
Younger, A.H., 2004. Natural Gas Processing Principles and Technology-Part I. Gas Processors Association, Tulsa Oklahoma.
McKetta Jr, J.J. ed., 1992. Unit Operations Handbook: Volume 2 (In Two Volumes) (Vol. 2). CRC Press.
McCabe, W.L., Smith, J.C. and Harriott, P., 1993. Unit operations of chemical engineering (Vol. 5, p. 154). New York: McGraw-Hill.
Biscan, D.A., Gebhard, R.S. and Matviya, T.M., 1986, June. Impact of process conditions on mercury removal from natural gas using activated carbon. In 8th International Conference on Liquefied Natural Gas.
Mokhatab, S., Mak, J.Y., Valappil, J.V. and Wood, D.A., 2013. Handbook of liquefied natural gas. Gulf Professional Publishing.
Mokhatab, S. and Poe, W.A., 2012. Handbook of natural gas transmission and processing. Gulf Professional Publishing.
Eckersley, N., 2010. Advanced mercury removal technologies. Hydrocarbon processing, 89(1).
Jubin, C. and Ducreux, O., 2014, November. Mercury Removal Units Operation at Front-end Location. In Abu Dhabi International Petroleum Exhibition and Conference. Society of Petroleum Engineers.
Appendices
Appendix A – Breakthrough Tables
Table 10-1 Breakthrough Profile for 4 mm Particles
Table 10-2 Breakthrough Profile for 2 mm Particles
Appendix B – Simulation Results
Table 10-3 Simulation Results of Case #1
Table 10-4 Simulation Results of Case #2
Table 10-5 Simulation Results of Case #3
Table 10-6 Simulation Results of Case #4
Table 10-7 Simulation Results of Case #5
Table 10-8 Simulation Results of Case #6
Table 10-9 Simulation Results of Case #7
Table 10-10 Simulation Results of Case #8
Table 10-11 Simulation Results of Case #9
Table 10-12 Simulation Results of Case #10
Table 10-13 Simulation Results of Case #11
Table 10-14 Simulation Results of Case #12
Table 10-15 Simulation Results of Case #13
Table 10-16 Simulation Results of Case #14
Table 10-17 Simulation Results of Case #15
Table 10-18 Simulation Results of Case #16
Table 10-19 Simulation Results of Case #17
Table 10-20 Simulation Results of Case #18
Table 10-21 Simulation Results of Case #19
Table 10-22 Simulation Results of Case #20
Table 10-23 Simulation Results of Case #21
Table 10-24 Simulation Results of Case #22
الملخص العربى
الزئبق هو احد اخطر الشوائب التي تلوث كل انواع الوقود الحفرية الطبيعية ويتواجد بتركيزات مختلفة في الغاز الطبيعي وخطورته تكمن في تأثيره على البيئة والصحة بخطورة شديدة بالإضافة إلى طبيعته المهاجمة بالتأكل لبعض المعادن المستخدمة في الصناعات البترولية وصناعة الغاز وتؤثر على الصناعة بحيث تؤدي إلى حوادث كارثية.
الهدف من هذا العمل هو دراسة لوحدة صناعية قائمة تزيل ابخرة الزئبق من الغاز الطبيعي بتقنية الإمتزاز. وحدة صناعية عبارة عن Fixed Bed معبأة بنوع خاص من الحبيبات الشرهة لإمتزاز ابخرة الزئبق تم تمثيلها رياضياً وتم اختبار تنبؤ النموذج الرياضي للوحدة على اساس تصميمها ودراسة تأثير ازالة طبقة من الحبيبات على اداء الوحدة وذلك للتغلب على مشكلة فرق ضغط كبير على الوحدة. تم استخدام النموذج لمحاكاة اداء الوحدة والتنبؤ باختراق الزئبق للوحدة وعمر خدمة الوحدة بإستخدام مقاسات مختلفة للحبيبات واختلاف ارتفاع طبقة الحبيبات واختلاف معدل سريان الغاز وتركيزات مختلفة للزئبق بالغاز عند دخول الوحدة.
لوحظ ان المادة المستخدمة لازالة الزئبق تثبت سعة امتزاز عالية للزئبق من خلال ما ذكر في معلومات الشركة المصنعة ومن خلال نتائج المحاكاة ومن خلال اداء الوحدة بالحقل مما يعطي الوحدة استمرارية للعمل لسنين طويلة. النموذج المطابق يثبت ايضاً خصائص المادة وطبيعة إتزان العملية ووجود عملية تفاعل كيميائي مع الإمتزاز. زيادة معدل سريان الغاز وتركيز الزئبق يؤدي طبيعياً كما هو متوقع إلى إختراق للزئبق مبكر للوحدة وتشبعها الوحدة بالزئبق بشكل اسرع. استخدام مقاسات اصغر للحبيبات يرفع من كفائة عملية الإمتزاز بفضل تحسين المساحات السطحية وإنتشار ابخرة الزئبق خلال مسام الحبيبات. بشكل عام وُجِد ان ارتفاعات اقل من طبقة الحبيبات تستطيع استيعاب معدلات سريان عالية من الغاز وتركيزات متوسطة من ابخرة الزئبق تصل إلى 2 µg/Nm3 (2,000 ng/Nm3) وهذه الفرصة من الممكن إنتهازها في السنين المقبلة من خلال مقترحين موفرين للنفقات وأحد المقترحَيْن يُحَسِنْ من معالجة الغاز وذلك من خلال تغيير في تصميم الوحدة وترتيبها بشكل امثل.
Copyright Notice
© Licențiada.org respectă drepturile de proprietate intelectuală și așteaptă ca toți utilizatorii să facă același lucru. Dacă consideri că un conținut de pe site încalcă drepturile tale de autor, te rugăm să trimiți o notificare DMCA.
Acest articol: The Effect of Operating Parameters on the Efficiency of an Industrial Unit to Remove Mercury Vapors from Natural Gas A thesis submitted to the… [301786] (ID: 301786)
Dacă considerați că acest conținut vă încalcă drepturile de autor, vă rugăm să depuneți o cerere pe pagina noastră Copyright Takedown.