The Effect of Operating Parameters on the Efficiency of an Industrial Unit to Remove Mercury Vapors from Natural Gas A thesis submitted to the… [603117]
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. Hassan Abd -El-Moneim Farag
Chemical Engineering Department
Faculty of Engineering – Alexandria University
Dr. Rania Farouk Abdou Salama
Petrochemical Engineering Department
Faculty of Engineering – Pharos University
جامعة االسكندرية
كلية الهندسة
دراسة في تأثير عوامل التشغيل على كفاءة وحدة
صناعية الزالة بخار الزئبق من الغاز الطبيعي
رسالة مقدمة من المهندس /احمد محمد محسن عبدهللا
إلى قسم الهندسة الكيميائية كمتطلب جزئى للحصول
على درجة الماجستير فى الهندسة الكيميائية
1026
المشــــرفون
أ.د. حسن عبد المنعم فرج
أستاذ متفرغ – قسم الهندسة الكيميائية
كلية الهندسة – جامعة االسكندرية
د. رانيا فاروق عبده سالمة
مدرس – قسم هندسة البتروكيماو يات
كلية الهندسة – جامعة فاروس
i
ACKNOWLEDGEMENT
Thanks to Allah that 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. Hassan Abd -El-Moneim Farag
Dr. Rania Farouk Salama
for offering me this unique chance of working under their direct supervision and for their
helpful advice, constant support, and encouragement throughout this study. Their comments
and suggestions not only provided valuable knowledge but broadened my perspect ive 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 would also like to thank my family for their support, help and their en couraging words
that lightened my way to success.
Finally, my thanks are also to everyone who helped me in this work.
ii
List of figures
Figure 2-1 Mercury Electron Configuration ………………………….. ………………………….. ………….. 3
Figure 2-2 Mercury Environmental Cycle ………………………….. ………………………….. …………….. 7
Figure 2-3 Mercury Releases and Control Measures ………………………….. …………………………. 10
Figure 2-4 Shape of Mercury Corrosion and Failure ………………………….. …………………………. 15
Figure 2-5 Brunauer Classification for Equilibrium ………………………….. ………………………….. 18
Figure 2-6 Schematic Diagram of Porous Adsorbent Particle ………………………….. …………….. 22
Figure 2-7 General Differential Mass Balance Equation over Bed Element ……………………… 23
Figure 2-8 Categories of Equilibrium Isotherms ………………………….. ………………………….. ….. 24
Figure 2-9 Configuration of Regenerative Mercury Removal Sys tem ………………………….. …. 33
Figure 2-10 Concentration Fronts through Columns and Zones Classification …………………. 35
Figure 3-1 Simple Process Flow Diagram ………………………….. ………………………….. …………… 37
Figure 3-2 Experimental Breakthrough Curves ………………………….. ………………………….. ……. 39
Figure 4-1 Applying Freundlich Isotherm for Data of 4 mm Adsorbent ………………………….. . 45
Figure 4-2 Applying Freundlich Isotherm for Data of 2 mm Adsorbent ………………………….. . 46
Figure 4-3 Applying Langmuir Isotherm for Data of 4 mm Adsorbent ………………………….. .. 47
Figure 4-4 Applying Langmuir Isotherm for Data of 2 mm Adsorbent ………………………….. .. 47
Figure 4-5 Testing Lagergren Pseudo -First Order for Data of 4 mm Adsorbent ……………….. 48
Figure 4-6 Testing Lagergren Pseudo -First Order for Data of 2 mm Adsorbent ……………….. 48
Figure 4-7 Testing Pseudo -Second Order for Data of 4 mm Adsorbent ………………………….. . 49
Figure 4-8 Testing Pseudo -Second Order for Data of 2 mm Adsorbent ………………………….. . 49
Figure 4-9 Testing Elovich Model for Data of 4 mm Adsorbent ………………………….. ………… 50
Figure 4-10 Testing Elovich Model for Data of 2 mm Adsorbent ………………………….. ………. 50
Figure 4-11 Testing Weber & Morris Model for Data of 4 mm Adsorbent ………………………. 51
Figure 4-12 Testing Weber & Morris Model for Data of 2 mm Adsorbent ………………………. 51
Figure 4-13 Testing Diffusion -Chemisorption Model for Data of 4 mm Adsorbent ………….. 52
Figure 4-14 Testing Diffusion -Chemisorption Model for Data of 2 mm Adsorbent ………….. 52
Figure 4-15 Testing Vinod & Anirudvan (Linear Driving Force Diffusion) Model for Data of
4 mm Adsorbent ………………………….. ………………………….. ………………………….. ………………….. 53
Figure 4-16 Testing Vinod & Anirudvan (Linear Driving Force Diffusion) Model f or Data of
2 mm Adsorbent ………………………….. ………………………….. ………………………….. ………………….. 53
Figure 4-17 Testing Zhang & Cheng Model for Data of 4 mm Adsorbent ……………………….. 54
Figure 4-18 Testing Zhang & Cheng Model for Data of 2 mm Adsorbent ……………………….. 54
Figure 4-19 Testing Wolborska Model for Data of 4 mm Adsorbent ………………………….. ….. 55
Figure 4-20 Testing Wolborska Model for Data of 2 mm Adsorbent ………………………….. ….. 55
Figure 4-21 Testing Clark Model for Data of 4 mm Adsorbent ………………………….. ………….. 56
Figure 4-22 Testing Clark Model for Data of 2 mm Adsorbent ………………………….. ………….. 56
Figure 4-23 Testing Bohart & Adams Model for Data of 4 mm Adsorbent ………………………. 57
Figure 4-24 Testing Bohart & Adams Model for Data of 2 mm Adsorbent ………………………. 57
Figure 4-25 Bohart & Adams Model Prediction versus Actual Breakthrough Data for 4 mm
Adsorbent ………………………….. ………………………….. ………………………….. ………………………….. . 58
Figure 4-26 Bohart & Adams Model Prediction versus Actual Breakthrough Data for 2 mm
Adsorbent ………………………….. ………………………….. ………………………….. ………………………….. . 59
Figure 4-27 Example of the Automated Use of Simulation Tool ………………………….. ………… 60
Figure 4-28 Pressure Drop Distribution across Layers of MRU Bed ………………………….. …… 62
Figure 4-29 Simulated Breakthrough Curves for Designed Bed Height and after Skimming of
Top 500 mm ………………………….. ………………………….. ………………………….. ……………………….. 63
Figure 4-30 Simulated Breakthrough Curves for Different Particle Size s ………………………… 64
iii
Figure 4-31 Simulated Breakthrough Curves for Inlet Mercury Concentration Effect ……….. 65
Figure 4-32 Simulated Breakthrough Curves for Gas Flow Rate Effect ………………………….. . 66
Figure 4-33 Simulated Breakthrough Curves for Bed Height Effect ………………………….. …… 67
Figure 5-1 Simulated Breakthrough Curves for Proposed Bed Height with Different Inlet
Mercury Concentrations ………………………….. ………………………….. ………………………….. ……….. 71
Figure 5-2 Simulated Breakthrough Curves till Saturation for Proposed Bed Height with
Different In let Mercury Concentrations ………………………….. ………………………….. ………………. 72
iv
List of tables
Table 2-1 Mercury Properties ………………………….. ………………………….. ………………………….. …. 3
Table 2-2 Mercury Compounds ………………………….. ………………………….. ………………………….. . 5
Table 2-3 Regional Average Mercury Content in Natural Gas ………………………….. …………… 13
Table 2-4 Categorization of Adsorption Systems ………………………….. ………………………….. …. 25
Table 2-5 Dynamic Models of Linear Equilibrium, Isothermal, and Trace Systems ………….. 26
Table 2-6 Dynamic Models of Irreversible Equilibrium, Isothermal, and Trace Systems ….. 27
Table 2-7 Dynamic Models of Nonlinear Equilibrium, Isothermal, and Trace Systems …….. 27
Table 2-8 Kinetic Expressions ………………………….. ………………………….. ………………………….. . 28
Table 4-1 Calculated Langmuir Constants ………………………….. ………………………….. ………….. 47
Table 4-2 R2 Values of Models Linear Regression ………………………….. ………………………….. . 58
Table 4-3 Design Basis of Mercury Guard Bed ………………………….. ………………………….. …… 61
v
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 = K ads/Kdes
BDST Bed Depth Service Time
BET Brunauer, Emmett, and Teller
BT Breakthrough
C (C o/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
CO 2 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
Coefficient s
Kg Kilograms
Kr Krypton
L Bed Length (Height)
LME Liquid Metal Embrittleme nt
LNG Liquefied Natural Gas
LUB Length of Unused Bed
m Mass of Adsorbent
MEOH Methanol
mg/m3 Milligrams per Cubic Meter
vi
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
NO X 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 Lim it
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 (q s/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/ Conc entration 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
SO 2 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
vii
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 A xial 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 Iso therm
or Lumped
Parameter in Zhang & Cheng Model
ᵵ Lumped Parameter in Bohart & Adams Model
viii
Table of contents
Ac nowl dg m nt………………………………………………………………………………i
Li t of figur …………………………………………………………………………………..ii
List of tabl …………………………………………………………………………….….…i v
Nom nclatur …………………………………………………………………………………. .v
Summary………………………………………………………………………………….……1
1. Introduction ………………………….. ………………………….. ………………………….. …………………… 2
2. Theoretical Part ………………………….. ………………………….. ………………………….. ……………… 2
2.1 Mercury Chemistry ………………………….. ………………………….. ………………………….. ….. 2
2.1.1 Mercury in Nature and History ………………………….. ………………………….. ………… 2
2.1.2 Physical Properties ………………………….. ………………………….. …………………………. 2
2.1.3 Chemical Properties ………………………….. ………………………….. ……………………….. 4
2.2 Mercury Environmental Concern ………………………….. ………………………….. ……………. 6
2.2.1 Sources of Pollution ………………………….. ………………………….. ……………………….. 6
2.2.2 Mercury Environmental Cycle ………………………….. ………………………….. …………. 6
2.2.3 Bioaccumulation and Biomagnification ………………………….. ………………………… 8
2.2.4 Mercury Health Effect s ………………………….. ………………………….. …………………… 8
2.2.5 Exposure Limits ………………………….. ………………………….. ………………………….. … 9
2.2.6 Minimization of Mercury Global Pollution ………………………….. ………………….. 10
2.2.7 Regulations ………………………….. ………………………….. ………………………….. …….. 11
2.2.8 Historical Accidents ………………………….. ………………………….. ……………………… 12
2.2.9 Situation of Egyptian Environment ………………………….. ………………………….. … 12
2.3 Mercury Impact in Energy Industry ………………………….. ………………………….. ………. 13
2.3.1 Abundance in Energy Industry ………………………….. ………………………….. ………. 13
2.3.2 Mercury Industrial Problems ………………………….. ………………………….. …………. 13
2.3.3 Mercury Corrosion ………………………….. ………………………….. ……………………….. 14
2.3.4 Failure Incidents ………………………….. ………………………….. ………………………….. . 14
2.4 Mercury Adsorption ………………………….. ………………………….. ………………………….. .. 15
2.4.1 Adsorption in Purification & Separation ………………………….. ……………………… 15
2.4.2 Adsorbents Characteristics ………………………….. ………………………….. …………….. 16
2.4.2. 1 Particles ………………………….. ………………………….. ………………………….. …… 17
2.4.2. 2 Adsorbent Aging ………………………….. ………………………….. ……………………. 17
2.4.3 Adsorption Mechanisms ………………………….. ………………………….. ……………….. 17
2.4.4 Adsorption Equilibrium ………………………….. ………………………….. ………………… 18
2.4.5 Capillary Condensation ………………………….. ………………………….. …………………. 19
2.4.6 Diffusion Transport ………………………….. ………………………….. ………………………. 19
2.4.7 Adsorption Kinetics ………………………….. ………………………….. ……………………… 20
2.4.8 Pressure Drop in Adsorption Packed Beds ………………………….. …………………… 22
2.4.9 Dynamic Modeling of Adsorption Beds ………………………….. ………………………. 22
2.4.10 Analytic al Solutions ………………………….. ………………………….. ……………………… 26
2.4.11 Mercury Adsorption and Chemisorption Processes ………………………….. ……….. 30
2.4.12 Beds Design Fundamentals ………………………….. ………………………….. ……………. 33
2.4.13 Mercury Removal Unit Operation ………………………….. ………………………….. ….. 34
3. Mathematical Modeling of Mercury Adsorption from Natural Gas in an Industrial Unit –
Case Study ………………………….. ………………………….. ………………………….. …………………………. 37
3.1 General Process Description ………………………….. ………………………….. ………………… 37
3.2 Mercury Remova l Unit (MRU) Description ………………………….. ……………………….. 37
3.3 Studying Dynamics of Industrial MRU Bed (Packed Fixed Bed) ………………………. 38
ix
3.3.1 Adsorption Equilibrium ………………………….. ………………………….. ………………… 39
3.3.1.1 Freundlich Equilibrium Isotherm ………………………….. …………………………. 39
3.3.1.2 Langmuir Equilibrium Isotherm ………………………….. ………………………….. . 39
3.3.2 Adsorption Kinetics ………………………….. ………………………….. ……………………… 40
3.3.2.1 Lagergren Pseu do-First Order ………………………….. ………………………….. …. 40
3.3.2.2 Pseudo -Second Or der Expression ………………………….. ………………………… 40
3.3.2.3 Elovich’ Mod l ………………………….. ………………………….. ……………………. 40
3.3.2.4 Weber an d Morris Model ………………………….. ………………………….. ……….. 41
3.3.2.5 Diffusion -Chemi sorption Model ………………………….. ………………………….. 41
3.3.2.6 Linear Driving Force diffus ion Expression ………………………….. ……………. 41
3.3.3 Adsorption Modeling ………………………….. ………………………….. ……………………. 41
3.3.3.1 Zhang and Cheng Model ………………………….. ………………………….. ………… 42
3.3.3.2 Wolborska Model ………………………….. ………………………….. ………………….. 42
3.3.3.3 Clark Model ………………………….. ………………………….. ………………………….. 43
3.3.3.4 Bohart & Adams Model ………………………….. ………………………….. …………. 44
4. Results and Discussions ………………………….. ………………………….. ………………………….. … 45
4.1 Studying Adsorption Equilibrium ………………………….. ………………………….. …………. 45
4.1.1 Applying Freundlich Equilibrium Isotherm ………………………….. ………………….. 45
4.1.2 Applying Langmuir Equilibrium Isotherm ………………………….. …………………… 47
4.2 Studying Adsorption Kinetics ………………………….. ………………………….. ………………. 48
4.2.1 Testing Lagergren Pseu do-First Order ………………………….. …………………………. 48
4.2.2 Testing Pseudo -Second Or der Expression ………………………….. ……………………. 49
4.2.3 Testing Elovich’ Mod l ………………………….. ………………………….. ……………….. 50
4.2.4 Testing Weber an d Morris Model ………………………….. ………………………….. …… 51
4.2.5 Testing Diffusion -Chemi sorption Model ………………………….. ……………………… 52
4.2.6 Testing Linear Driving Forc e diffusion Expression ………………………….. ………. 53
4.3 Studying Adsorption Mathematical Models ………………………….. ……………………….. 54
4.3.1 Testing Zhang and Cheng Model ………………………….. ………………………….. ……. 54
4.3.2 Testing Wolborska Model ………………………….. ………………………….. ……………… 55
4.3.3 Testing Clark Model ………………………….. ………………………….. …………………….. 56
4.3.4 Testing Bohart & Adams Model ………………………….. ………………………….. …….. 57
4.4 Bohart & Adam Mod l’ Pr dictability for th MRU Sy t m ………………………….. 58
4.5 Calculations and Building a Simulation Tool ………………………….. ……………………… 59
4.6 Simulating Original versus Current Bed Design Performance …………………………. 61
4.7 Effect of Different Operating Parameters – Sensitivity Analysis and Case Studies .. 64
4.7.1 Adsorbent Particle Size ………………………….. ………………………….. …………………. 64
4.7.2 Inlet Mercury Concentration ………………………….. ………………………….. ………….. 65
4.7.3 Feed Gas Flow Rate – Velocity ………………………….. ………………………….. ………. 66
4.7.4 Bed Height ………………………….. ………………………….. ………………………….. ……… 67
5. Conclusion s and Recommendations ………………………….. ………………………….. ……………. 68
6. Refferences ………………………….. ………………………….. ………………………….. …………………. 73
الملخصبالعربى …………………………………………… ..…..……………………………… .. 83
1
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 a n industrial u nit that removes mercury from
natural gas by a dsorption . An industrial fixed bed , which is packed with a specific
type of mercury adsorbents was modeled and tested the model prediction for the bed
design basis, and studied the effect of layer skimming on b ed performance to solve
pressure drop problem. The model was used to simulate bed performance an d predict
breakthrough and life time by varying catalyst size, different bed heights, different gas
flow rate s, and different inlet mercury concentrations.
It was observed that the mercury adsorbent is proving a high adsorption capacity as
observed from manufacturer data, modeling results, and field performance that make s
the bed last for years of lifetime.
The breakthrough data was used in testing two equilibri um isotherms, six kinetics
mathematical expressions, and four mathematical models to obtain thorough
information about b d’ charact ri tic and dynamic .
One mathematical model provides bes t fit for the breakthrough data, in which this
model assumes singl e trace -component adsorption, existence of chemical reaction,
irreversible isotherm, and neglecting of axial dispersion, accordingly, the matching
model proves adsorbent characteristics, nature of equilibrium, and the chemical
reaction mechanism (Chemisorp tion).
Based on simulation analysis, model prediction found that i ncreasing 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 , improve
treatment of feed gas , and maximize plant availability by optimization of bed design
and its configuration.
2
1. Introducti on
Natural gas is one of the most important energy sources because it is a cheap and
clean fossil fuel. It consists naturally of Methane as the main component mixed with
limited percentage of ethane, propane, butane, pentane, and hexane. In addition to
hydroc arbons, natural gas is contaminated with different impurities such as water,
acid gases, carbon dioxide, hydrogen sulfide, mercury, oxygen, and nitrogen.
Accordingly, it is subjected to different processing techniques to remove the
impurities and meet the standard specifications before consumption.
Mercury traces naturally exist in natural gas in different concentrations around the
world. Although mercury exists in traces, it has severe effects on the environment and
on the energy industry, so it has a con siderable interest in research for efficient
removal. It is primarily removed from natural gas by efficient adsorption process, and
such processes have been studied and developed extensively due to its criticality.
2. Theoretical Part
2.1. 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 thermom eters due to
high rate of thermal expansion that is fairly constant over a wide temperature range .
2.1.1. 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) ore s 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 Chines e empire as though to prolong life,
also ancient Greeks and Romans used it in cosmetics. [3,4]
In the past, it was extracted in tensively 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 a re also reports of small -scale mining
of mercury in China, Russia (Siberia), Outer Mongolia, Peru, and Mexico. [6]
2.1.2. 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
3
temperature with exceptional low est boiling point and low est melting point for a
metal . [8] The following table summarizes its main physical properties.
Table 2-1 Mercury Properties
General Information
Name Mercury
Symbol Hg
Color Silvery
Odor Odorless
Atomic Number 80
Element Category Transition Metal
Group in Periodic Table 12 (d block below Cadmium and above Copernicium)
Row/Period in Periodic Table 6 (Next to Gold and previous to Thallium)
Standard Atomic Weight 200.592 g.mol -1
Mercury Electron
Configuration Short Form :
[Xe]4f145d106s2
Long Form :
1s22s22p63s23p63d104s24p64d105s25p64f145d106s2
Shell Structure:
2 8 18 32 18 2
Atomic/Molar Volume 14.81 cm3/mole
Filling Orbital 5d10
Number of Electrons 80
Number of Protons 80
Number of Neutrons (most
common/stable nuclide) 121
Valence Electrons 6s2
Physical Properties
Phase Liquid
Melting Point −38.8290 °C [9-11]
Boiling Point 356.73 °C [9-11]
Density at Standard
Conditions 13.534 g/cm3
Molar Heat Capacity 27.983 J/(mol·K)
Specific Heat Capacity 140 J/(kg . K)
Vapor Pressure 1 bar at 356 °C
Van der Waals radius 155 pm
Thermal Conductivity 8.30 W/(m·K)
Electrical Conductivity 0.0104 106/cm Ω
Magnetic Ordering Diamagnetic
Thermal Expansion 60.4 µm/(m·K) (at 25 °C)
Flammability Class Noncombustible Liquid
Figure 2-1 Mercury Electron Configuration
4
Its unique electron configuration is the reas on 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 sod ium 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]
2.1.3. 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 (2.1)
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 eleme ntal mercury is extracted from Cinnabar by heating
abov 540˚C in presence of air followed by condensing the vapors. [14]
HgS + O 2 → Hg + SO 2 (2.2)
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 merc ury 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
5
divalent and do not react with water. They usually have the formula HgR2, which are
often volatile, or Hg RX, 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 s olids. Dimethyl –
mercury, however, is a colorless liquid. [14]
The following table summarizes many mercury compounds and their applications.
Table 2-2 Mercury Compounds [21]
Compound Applications
Elemental Mercury It is used in thermometers, barometers, manometers, sphygmomanometers,
float valves, mercury switches, mercury relays, batteries, fluorescent lamps , in
amalgam material for dental restoration fillings , and mining extraction of gold,
silver , zinc from their ores. It is a lso used as a liquid electrode in the
manufacture of chlorine and sodium hydroxide (chlor -alkali production) by
electrolysis of brine. Most of applications are slowly phased out due to health
and safety regulations. [22]
Mercury(I) or mercurous chloride
Hg 2Cl2 A colorless solid also known as calomel. It is a standard in electrochemistry and
still used in medicine occasionally as a purgative. Also used as fungicide,
maggot control in agriculture, and fireworks. [23]
Mercury(II) or mercuric chlor ide
HgCl 2 An easily sublimating white solid (corrosive sublimate – a violent poison). It is
used as an insecticide, in rat poison, as a disinfectant, tanning of leather, spray
for potato seedlings (to protect from disease), preservation of wood, embalming
fluid, textile printing, and engraving.
Mercury(II) oxide Arises when the metal is exposed to air for long periods at elevated
temperatures. It is used in skin ointments, red or yellow pigment in paints,
disinfectant, fungicide (to kill fungi), perfumes, and cosmetics.
Mercury(II) sulfide HgS It occurs in nature as the ore cinnabar and is the brilliant pigment vermillion, a
high-grade paint red pigment. Another crystalline form of the sulphide also
used as a pigment is black. [23]
Mercury(II) selenide ( HgSe) It is a grey -black crystalline solid semi -metal , occurs naturally as the mineral
Tiemannite. Selenium is used in filters in some steel plants to remove mercury
from exhaust gases. The solid product formed is HgSe . It can be used as an
ohmic contact t o wide -gap II -VI semiconductors such as zinc selenide or zinc
oxide.
Mercury(II) telluride (HgTe) as
well as various derivatives, e.g.
mercury cadmium telluride and
mercury zinc telluride Semiconductors useful as infrared detector materials. [24]
Mercury(II) salts (Hg2N+),
(HgNH+2)n), [Hg(NH3)2]Cl2,
potassium tetraiodomercurate(II)
(HgI2−4) Occasionally used to test for ammonia
Mercury(II) fulminate (Hg(ONC) 2 A detonator widely used in explosives. [7]
Mercuric arsenate (HgHAsO 4) Waterproofing pai nts
Mercuric benzoate (Hg(C 7G5O2)2) Medicine; used to treat syphilis
Mercuric cyanide (Hg(CN) 2) Germicidal soaps (soaps that kill germs), photography
Mercurous chromate (Hg 2CrO 4) Green pigment in paints
6
Mercurous iodide (Hg 2I2) Kills bacteria on the skin
Mercurochrome (Merbromin) Topical Antiseptic
Organic mercury compounds Generally in past, used in pesticides (extensive use in seed dressing among
others) and biocides in some paints, pharmaceuticals, paper industry, and
cosmetics. Use of dimethyl -mercury in small amounts as a reference standard
for some chemical tests, and thimerosal (which contains ethyl -mercury) used as
a preservative (antiseptics) in some vaccines and other medical and cosmetic
products (mascara) inc th 1930’ .
2.2. Mercury Environmental Concern
2.2.1. 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.
2.2.2. Mercury Environmental Cycle
Mercury naturally exists in environment in the most common forms of metallic
elemental, mercuric sulfide, mer curic 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 incl ude 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 s ource 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 th e 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
d po it hidd n in th arth’ cru t and r l a d into th bio ph r , it can b highly
mobil , cycling b tw n th arth’ urfac as deposits and the atmosphere as vapors .
7
Th arth’ urfac oil , wat r 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]
Methy lmercury 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 m ethylmercury . 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 immobi lized and remains undisturbed by anthropogenic or natural
activity (climatic and geological). [34]
The following sketches illustrate simply the mercury environmental cycling.
8
Figure 2-2 Mercury Environmental Cycle
2.2.3. Bioac cumulation 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 eliminati on (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 thu s 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 bio magnification , meaning the progressive build up by
successive t rophic 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 peop le, especially sensitive (such as pregnant women and young children), to
limit or avoid consumption of certain types of fish from various water bodies. [36]
2.2.4. Mercury Health Effect s
Mercury is of global concern due to extreme toxic nature of all its compoun ds 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
exposur e, and the age of poisoned person . Health effects include tremors, impaired
9
cognitive skills, mood changes, inability to concentrate, sleep disturbance , chest pa in,
dyspnea, cough, hemoptysis, impairment of pulmonary function, evidence of
interstitial pneu monitis, 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, blindn ess, 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 ar e 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 wor ldwide, 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 Internati onal 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
effec ts. [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 b e 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]
2.2.5. 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 Reco mmended Exposure Limit
(REL) are TWA: 0.05 mg/m3, Ceiling: 0.1 mg/m3, and IDLH: 10 mg/m3. Also
10
ACGIH set recommended airborne exposure limit to 0.025 mg/m3 averaged over an
8-hour work shift. [17,4 6,77]
2.2.6. Minimization of Mercury Global Pollutio n
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 technolog ies have been developed for coal combustion plants and waste
incinerators with the primary intention of addressing acidifying substances (especially
SO 2 and NO X), 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 f orms 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 2-3 Mercury Releases and Control Measures [34]
The possible measures of controlling mercury releases can be summarized as the
following:
Preventive measures
1. Reducing consumption of raw materials and products generating mercury
releases
2. Substitution by non -mercury alternatives in processes and products
Controlling measures
3. End-of-pipe techniques
4. Waste management
11
The preventive controls are generally cost -effective, unless the alternative s are
significantly mor e 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 co mplexity arises from the many different source of waste.
However, c ombination of the different measures is the opt imum for reducing mercury
emissions towards saving the environment globally. [52]
As the awareness of mercury's potential adverse impacts on h ealth 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]
2.2.7. 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, Ind ia,
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 pressu re
measuring device. [57]
There are various regional initiatives conducted by different countries such as; [58]
12
setting environmental quality standards limiting concentrations in industrial
emissions, wastes, drinking waters, water bodies, air, soil, and d iet (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.
2.2.8. 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 , occurr d in th 1950’
where orga nic 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 se ed dressing containing organic mercury compounds were used for
bread. [64]
2.2.9. 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, st udies paid attention for
the surrounding environment and recommend ed continuous monitoring of Hg in Wadi
Al-Qamar area and continuous health monitoring for residents. [65]
Regarding aquatic environment, a study concerned wit h 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 –
Burullu ≥ El -Manzallah > Mariout > Edku. [66] However, another study, for
evaluation of the extent of human exposure to methylmercury in Aldakahlia
Gover norate 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 fu rther 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 o f the additional sources of runoff to the bay from industrial wastes .
13
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, t he general populations need enrichment for the culture related to the
environment and public health and raising awareness about mercury hazards,
environmental and health effects.
2.3. Mercury Impact in Energy Industry
2.3.1. 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 2-3 Regional Average Mercury Content in Natural Gas [75]
Location Elemental Mercury Concentration µg/Nm3
North Africa 0.3 – 130
Middle East 1 – 9
Far East 58 – 193
North America 0.055 – 0.04
South America 69 – 119
Eastern US Pipeline 0.019 – 0.44
Midwest US Pipeline 0.001 – 0.10
Groningen 180
2.3.2. Mercury Industrial Problems
Mercury has either concern s 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
remov ed 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 Ind ustry, 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
14
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, a nd 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 pro gressive 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]
2.3.3. 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 Al 2O3 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 prev iously
amalgamated mercury free to continue develop again successive amalgam with the
metal in a progressive process developing weak spots (cracks) .
Al + Hg Al Hg (2.3)
2AlHg + 6 H 2O 2Al (OH) 3 + 3 H 2 + 2 Hg (2.4)
This corrosion process is called liquid metal embrittlement (LME) with reference to
liquid elemental mercury penetration into the aluminum ox ide 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 hy drates.
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 inc lude steel -Cu, stainless steel -Zn, and aluminum -Hg. Mercury often
embrittles mild steels in str essed or unstressed conditions at low temperatures . [86]
2.3.4. 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 corros ion at
Aluminum exchanger in the Moomba gas plant in South Australia lead to gas release
and fire. [76]
15
Figure 2-4 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, U SA, Indonesia, Thailand, and Holland as a result
of presence of m ercury particularly in Cryogenic plants.
2.4. Merc ury Adsorption
2.4.1. 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. A s
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. A ctivated 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 bee n 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
valuat d by ratio of H nry’ law con tant . [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 compar able with the diameter of diffusing
species, also separation may be based on difference in adsorption equilibrium or
different pore diffusion rates. [87]
16
2.4.2. Adsorbents Characteristics
Microporous adsorbents differ in pore size from a few angstrom s 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 (Al 2O3.3H 2O) 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 dryin g 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 por es. 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 su ch 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 co rrespond 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 adsor ption 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 Sie ves 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 SiO 4 and AlO 4 tetrahedral cells joined together) where Oxygen is
shared between Al & Si atoms forming rings of Oxygen members which provide the
inter-crystalline open channels str ucture 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 negati ve 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 whic h 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. Accord ingly, the structure, Si/Al ratio, and cation
types are the ruler of adsorption properties. [87]
17
2.4.2.1. Particles
The adsorbent micro -porous crystals are formed as a macro -porous pellet or
granulated bead (spherical) particles or even flakes with optimized dimens ions,
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 resist ance) but at
the cost of mechanical strength, pressure drop. In addition , 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. Differ ent
binder materials have adsorption properties, which alter the selectivity of original
crystals. [87]
2.4.2.2. 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 b y adsorbed component occupancy.
2.4.3. 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 p otential 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
interactio ns 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 fol lowing:
higher heat of adsorption,
specific ,
monolayer,
18
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 electr ic 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 t he nature of changes in
Gibb' fr n rgy ΔG, nthalpy ΔH, and ntropy ΔS with ad orption. Bri fly, mo t of
phy ical ad orption proc ar xoth rmal (g n rating h at & ΔH i n gativ ) and
accordingly observed to be promoted with lowering temperature a nd some are
endothermal (activated by heat) and thus promoted by increasing temperature. [87]
2.4.4. 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
equili brium constant which is inversely proportional to temperature. [87]
q = KC (2.5)
A general classification described by Brunauer et al. for physical adsorption schemes
of equilibrium isotherms as in below photo.
Figure 2-5 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 po re 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]
19
1- Adsorption occurs at fixed number of well defined localized sites.
2- Monolayer coverage, i.e. each site is occupied by one molecule.
3- Homogeneous sites with equivalent adsorptive energy.
4- The adsorbed molecules are isolated from each other (No mutual interaction).
(2.6)
By Langmuir model, a good fit of many experimental isotherms can be achieved just
by optimum selection of constants b (equilibrium constant = Kads/K des) and q s
(saturation capacity). Higher values of b means higher adsorption rate relative to
desorption rate and accordingly higher removal.
Another linear form is [87,89]
+
(2.7)
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 (2.8)
Also linear form is
ln q = ln K + 1/n ln C (2.9)
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 b etween 0
to 1 is indication of a favorable adsorption and 1 reduces to linear Henry law.
Accordingly, for fitting experimental data, vari ations 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, Redlic h-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]
2.4.5. 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]
2.4.6. 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)
(2.10)
20
Pore diffusion mechanism depends on pore s ize, 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 particl es, 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 fac tor 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 mechanis m 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 di ameter 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 experime ntally 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 vi scosity. [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 ef ficient separation processes
such as separation of O 2 and N 2 by carbon molecular sieves. [87]
2.4.7. Adsorption Kinetics
Kinetics of adsorption processes are very important and studied extensively in pilot
and industrial scales for many purposes not limited to t he 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 t reatment …etc.
21
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 resist ing 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]
1- External mass transfer res istance 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]
(2.11)
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/R p for spherical
shape. Mass transfer coefficient is correlated in the form of dimensionless
group Sherwood number Sh = 2R pK/D m (4.8) where Dm is molecular
diffusion, and Sherwood number is correlated with Reynolds & Schmidt
number. [87]
2- 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 diffu sion resistance more dominant than external film resistance
in evaluating mass transfer rate. [87]
3- Micropore or intra -crystalline diffusional resistance such in zeolites , which
consist of microporous crystals formed into macroporous particles in pellet o r
bead granulated shapes. When it is significant, the concentration through the
particle is uniform and uptake rate is independent of particle size. [87]
4- Axial dispersion or called also axial mixing , but it is not a mass transfer
resistance. [87]
5- Heat tran sfer 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 man y 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 gra dient through the
particle and between the particle and the fluid will be negligible and the
system is considered isothermal. [87]
22
Figure 2-6 Schematic Diagram of Porous Adsorbent Particle [87]
2.4.8. 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 low er limit
to avoid high pressure -drop and upper limit to avoid wall effects and axial dispersion
effects. [87]
2.4.9. Dyna mic 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 de scribe 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 sys tem and extract information about the process. [87]
Models derivation usually starts with differential mass balance (sometim es called
continuity equation) for an element of the bed. The general equation for fluid phase
differential mass balance of axially dispersed plug flow is :
+
+
+
(2.12)
23
Figure 2-7 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]
(2.13)
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
utiliz es 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, favorab le, and unfavorable isotherm. For distinction
between them, X -Y diagram describe their relationships. [87]
24
Figure 2-8 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 unfavor able 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]
25
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 iso therm,
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 i n
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 efficie nt 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 transpor t, 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 2-4 Categorization of Adsorption Systems
Category Basis Categories Distinction
Equilibrium
Relationship Linear
Isotherm Dispersive or proportionate pattern of concentration front and give
good approximation for low concentration ranges
Nonlinear
Favorable
Isotherm Constant pattern of concentration front in shape and width ; can be
represented by Langmuir expression particularly for strongly
adsorbed species which is extreme case of high favorable isotherm
called irreversible or rectangular isotherm ( 0 –> C0 approaches zero)
Nonlinear
Unfavorable
Isotherm Dispersive or proportionate pattern of concentration front and
observed in the desorption of favorable adsorption
Heat Transfer Isothermal When heat transfer can be considered rapid enough (high heat
transfer coefficients) relative to the adsorption rate and accordingly
the rate of heat ge neration 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 and the
profile of concentration front is due to the effects of axial
dispersion an d mass transfer resistance. It is commonly observed in
trace species system with inert carrier.
Non
Isothermal Cannot neglect heat transfer resistance where the slow heat transfer
between fluid and solid lead to additional extension in the
concentration front. In adiabatic systems where heat transfer is also
slow between column wall and surroundings will increase
complex ity due to adding thermal front and secondary mass
transfer zone. These features appear with high heat of adsorption
and high concentrations.
Adsorbate
Concentration Traces (Low
Concentration) The adsorbate is present in low concentration that will not c hange
fluid velocity during its adsorption. (Constant velocity)
High The adsorbate is present in sufficiently high concentration that will
26
Concentration lead to variation in fluid velocity during its adsorption. (consider
velocity variation)
Flow Model Ideal Plug
Flow Can neglect axial dispersion and remove -DL∂2c/∂z2 term
simplifying more the model expression
Dispersed Plug
Flow Dispersion is significant and can't be neglected
Kinetic Complexity
(Adsorption Rate
Expression) Negligible
mass transfer
resistance Instantaneous equilibrium throughout the bed
Single mass
transfer
resistance Simple linear driving force expression
(2.14)
Where the rate coefficient is overall effective lumped parameter
Single dominant diffusional resistance expression whether intra –
particle (macropore) or intra -crystalline (micropore)
Two
resistances External film + Intra -particle (macropore) diffusion
Both internal; Intra -particle (macropore) diffusion+ Intra –
crystalline (micropore) diffusion
All three
resistances Most complex, accurate, general model, which considers external
film and both internal diffusional resistances so such mode l
describes all practical adsorption processes realistically. However,
the simpler models provide reliable and sufficient approximation
for many practical systems.
2.4.10. Analytic al 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 follow ing table provides the analytic solutions for line ar equilibrium, isothermal ,
and trace systems available in literatures: [87]
Table 2-5 Dynamic Models of Linear Equilibrium, Isothermal, and Trace
Systems
System Description Model Expression
Authors
Ideal Plug Flow + Simple Linear Rate Expression
Walter
Anzelius
Furnas
Nusselt
Klinkenberg
Dispersed Plug Flow + Simple Linear Rate Expression
Lapidus &
Amundson
Levenspiel &
Bischoff
Ideal Plug Flow + Single Intra -particle (macropore) diffusion control Rosen
Dispersed Plug Flow + Single Intra -particle (macropore) diffusion
control Rasmuson &
Neretnieks
Ideal Plug Flow + Two resistances: External film & Intra -particle
(macropore) diffusion control Rosen
Disp ersed Plug Flow + Two resistances: External film & Intra -particle
(macropore) diffusion control Rasmuson &
Neretnieks
Ideal Plug Flow + All three mass transfer resistances (External film &
internal macropore & micropore diffusion resistances) Kawazoe &
Takeuchi
Dispersed Plug Flow + All three mass transfer resistances (External
film & internal macropore & micropore diffusion resistances) Rasmuson
27
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 follow ing table provides the analytic solutions for irreversible equilibrium,
isothermal , and trace s ystems available in literatures.
Table 2-6 Dynamic Models of Irreversible Equilibrium, Isothermal, and Trace
Systems [87]
System Description Model Expression Authors
Ideal plug flow + quasi -chemical rate equation
(2.15) Bohart & Adams (1920)
Ideal plug flow + linear rate solid film expression
(2.16) Cooper
Ideal plug flow + linear rate fluid film expression
(2.17) Cooper
Ideal plug flow + micropore diffusion control Cooper
Ideal plug flow + macropore diffusion control Cooper & Liberman
Ideal plug flow + both external film & macropore diffusion
control Weber & Chakravorti
Ideal plug flow + both external film & micropore diffusion
control Y.Yoshida & T. Kataoka
The follow ing table provides the analytic solutions for other cases of nonlinear
equilibrium systems, isothermal, and trace systems available in literatures: [87]
Table 2-7 Dynamic Models of Nonlinear Equilibrium, Isothermal, and Trace
Systems
System Description Model Expression Authors
A general analytic solution for Langmuir system + pseudo
second order reversible kinetic reaction originally based on ion
exchange theory
] (2.18)
Suitable for negligible mass transfer limitations Thomas (1944)
Results in graphical
forms of Thomas Model
by Hiester & Vermeulen
Ideal plug flow + Langmuir Equilibrium + linear rate f luid film
expression Zwiebel et al.
Michaels
Ideal plug flow + Langmuir Equilibrium + linear rate solid film
expression Garg & Ruthven
Hall et al.
Ideal plug flow + Freundlich Equilibrium + External film &
solid film Tien & Thodos
Ideal plug flow + Langmuir Equilibrium + micropore diffusion
expression Antonson & Dranoff
Garg & Ruthven
Graphs of Garg &
Ruthven
Numerical tabulation of
Hall et al. (Constant
Diffusion)
28
Ideal plug flow + Freundlich Equilibrium + micropore diffusion
expression Kyte
Ideal plug flow + Langmuir Equilibrium + macropore diffusion
expression Garg & Ruthven
Graphs of Garg &
Ruthven
Carter & Husain
Numerical tabulation of
Hall et al.
Quadratic driving force for micropore diffusion approximation Vermuelen
Macropore diffusion approximation Vermuelen & Quilici
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 stud ied 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 2-8 Kinetic Expressions [94-109]
System Description Expression Form Model Expression
Authors
Lagergren Pseudo -first order. It
differs from true first order because
parameter (qe – qt) does not
represent the number of available
sites, and log qe is adjustable
parameter and observed that it
doesn't equal the intercept of Log(q e
– qt) versus t. The model drawback is
that qe must be known while in many
cases it is unknown as chemisorption
tends to be slow, and the measured
adsorption is still much s maller than
qe. In most cases in the
literatures, the model doesn't match
well for the whole range of contact
time and is generally applicable over
the initial 20 to 30 minutes of the
sorption process. So, one has to find
some means of extrapolating the
experimental data to infinite t, or
treat qe as an adjustable parameter by
trial and error.
(qe – qt) (2.19)
Linear Form:
Log(q e – qt) = log q e –
(2.20) Lagergren, 1898
[97]
Pseudo -second order equation
presented for chemisorption of heavy
metals and avoids the pseudo -first
order disadvantage of prior
estimating qe.
(qe – qt)2 (2.21)
Linear Form:
+
(2.22)
Initial adsorption rate h =
Ho and McKay,
1998 [95,96 ]
29
Elovich’ mod l
A kinetic equation for chemisorption
and valid for heterogeneous surfaces .
(2.23)
Linear Form:
+
(2.24)
is initial adsorption rate
is desorption constant or
the relationship between
degree of surface coverage
and activation energy of
chemisorption Zeldowitsch,
1934 [100-102]
Weber and Morris, intra -particle
diffusion model derived from fick's
law with neglecting external film
diffus ion and considering intra –
particle diffusion is the rate limiting
step. qt = k id + C (2.25)
kid is diffusion rate constant
=
(2.26)
C intercept indicates the
external film thickness, where
the higher C, the higher
external film effects Weber and
Morris, 1962
[98]
First order equation of Bhattacharya
and Venkobachar . [91]
k1C – k2Cs (2.27) Bhattacharya &
Venkobachar
Diffusion -Chemisorption Model, an
empirical expression developed to
model adsorption of heavy metals at
heterogeneous surfaces . [99]
n K DC t n-1(qe – qt) 2/ q e 2
(2.28)
Linear form with n = 0.5 as
per Sutherland analysis:
+
(2.29)
Initial adsorption rate
Ki = K DC2 / q e Sutherland, 2004
[99]
Linear Driving Force diffusion
model by Vinod & Anirudvan . Ln (1 – ) = k t (2.30)
is the fractional attainment
of equilibrium q/q e and k is
diffusion time constant Vinod &
Anirudvan [103]
Wolborska model developed
considering mass transfer and
diffusion in axial dispersive plug
flow for the low concentration region
of breakthrough curve . [107]
c (2.31) Wolborska, 1989
[107]
Clark model combining mass
transfer expression with Freundlich
isotherm . [106] k (C – Ce) = v dC / dz (2.32) Clark, 1987
[106]
Yoon Nelson expression based on
gas adsorption kinetics and
assumption that the rate of decrease
in the probability of adsorption for
each molecule is proportional to the
probability of sorbate adsorption and
the probability of sorbate
breakthrough on the sorbent . [105]
(2.33)
where Q (%) is the
probability for adsorption, P
(%) is the pr obability for
breakthrough, an d t (s) is
time, the rate of decrease in
the probability of adsorption
is directly proportional to
fluid concentration C, and the
volumetric flow rate U, and
inversely proportional to the
weight of the carbon Wc Yoon & Nelson,
1984 [105]
30
Deactivation Kinetic (Zhang &
Cheng) model was originally
developed for gas -solid adsorption
but found also best fit applicability
with liquid systems. It considers
simultaneous physical adsorption
with linear equilibrium, first order
chemical reaction kin etic, and active
sites deactivation function in
Isothermal and Ideal Plug Flow
system. [108] Reaction rate
Φ x -Kd t)
(2.34)
K reaction rate constant
Φ is the deactivation
function (dimensionless)
Kd is the deactivation rate
constant Zhang & Cheng,
2000 [108]
2.4.11. 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 concern s, 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 ad sorption 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, a ctivated 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 f or 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 si tes 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 t he likely mechanism. [120] Skodras
et al. studied mercury and PCBs adsorption from gas phase using diff erent 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
31
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 fo r 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 mer cury 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. Mc kay compared extensively chemisorption kinetic models applied for
various pollutant sorption systems, and found r esults indicating that chemisorption
processes could be rate -limiting step, and that t he 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 b iosorption 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 fitte d 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 ki netics 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, su rface 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 cad mium 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 contribu te 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
32
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, c hemical 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 (2.35)
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 material s, 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 conde nsation 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 hydrocarb on 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 siev es,
sulfur -treated zinc oxide, chemically -modified carrier with metal sulfide, mixture of
33
basic copper carbonate, basic zinc carbonate, and simultaneous mercury removal and
drying using regenerative silver -coated molecular sieves. [176-178]
Figure 2-9 Configuration of Regenerative Mercury Removal System
2.4.12. 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 conside ring
the following aspects:
1. Knowing the gas composition, operating pressures, operating temperatures,
and the applied processes to define suitable adsorbent characteristics, and
define l ocation of the bed to avoid any process interference, avoid any proces s
incompatibility, and ensure optimum operating conditions. [115]
2. 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 specif ied lifetime in terval 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]
3. 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.
4. Ensuring good distribution of the fluid through the adsorbent bed which is
determined by the bed dimensio ns.
34
5. Flow direction either downward or upward, is generally specified according to
design velocities.
6. Support structure of the bed including; the wire meshes, and inert (ceramic)
balls to prevent fluidization, and particles escape.
Evaluation of experiment al 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 equ ations. [109, 171 -174]
Breakthrough capacity Qb =
–
(2.36)
Saturation capacity Qs =
–
(2.37)
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 cha nges from inlet concentration to approaching zero or
the designed specification,
this MTZ is approximated by [109, 171 -174]
MTZ = L Q b / Q s (2.38)
and length of unused bed (LUB) or called virgin zone to be approximated by
LUB = L ( 1 – Qb/Q s) (2.39)
2.4.13. 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 acc ommodate 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 pass ing out of the bed , which is called br eakthrough , and the unit has to be
shutdown, unload ed from the spent material, and reload ed 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]
1- The saturated zone that does not uptake mercury,
2- 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 concent ration to
desired effluent concentration,
3- Virgin zone , which is unused guard layer and free of mercury.
35
The below simple sketches explain the three zones.
Figure 2-10 Concentration Fronts through Columns and Zones Classification
36
Generally, the removal efficiency is normally affected by certain factors that can be
summarized as below. [115]
1- 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 longe r contact time, which improves adsorption
rate, but less turbulence may decrease external film mass transfer
coefficient and increase undesirable axial dispersion effects.
2- Gas mercury concentration
The higher the concentration, the more rapid bed saturatio n, 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.
3- Sorbent characteristics
Sorbent composition, porosity, pore size, surf ace area, pore volume,
tortuosity, surface pH, synthesis procedure, particle size, and crush
strength are all characteristics influence adsorption and pressure drop.
4- 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.
5- Operating pressure
Pressure has effect on physical adsorption, which is usually promoted by
higher pressure.
6- 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.
7- 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.
8- Contact time
Longer residence time normally increase s the chance of mercury uptake
and achieves higher removal and it is function of flow rate and bed
geometry.
9- Bed porosity /density and particles distribution
The lower bed porosity negatively affects pressure drop although it
indicates much dense be d and higher removal capacity, while the much
high porosity gives chance to channeling and axial dispersion effects in
addition to lower capacity.
37
3. Mathematical Modeling of Mercury Adsorption from
Natural Gas in an Industrial Unit – Case Study
3.1. General Pro cess 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 industrie s 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 b y gas pre -chilling using Brazed -Aluminum -Plate -Fin Heat Exchangers
commo nly 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 t he national gas grid just losing the
low fractions of heavy hydrocarbons.
The below simple process flow diagram explains pretreatment processes.
Figure 3-1 Simple Process Flow Diagram
3.2. Mercury Remova l 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 failu res, 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 remova l adsorbents , porous particles, designed to
38
selectively capture the elemental mercury within specified life time 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 desc ribed 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 low er
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 chem ical
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 mac ropore 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 t o 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 (3.1)
3.3. 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
follo wing graph .
39
Figure 3-2 Experimental Breakthrough Curves
3.3.1. 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.
3.3.1.1. Freundlich Equilibrium Isotherm
Equilibrium relationship: [90]
q = K C 1/n (3.2)
Linear form:
ln q = ln K + 1/n ln C (3.3)
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.
3.3.1.2. Langmuir Equilibrium Isotherm
Equilibrium relationship: [87,89]
(3.4)
Linear form:
+
(3.5)
By Langmuir model, a good fit of many experimental isotherms can be achieved just
by optimum selection of equilibrium constants b (equilibrium constant = Kads/K des)
and q s (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.
40
3.3.2. 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 ;
1- Lagergren Pseudo -first order,
2- Pseudo -second order,
3- Elovich’ mod l,
4- Weber and Morris,
5- Diffusion -Chemisorption Model,
6- Linear Driving Force diffusion
3.3.2.1. Lagergren Pseudo -First Order
The kinetic expression: [97]
(qe – qt) (3.6)
Linear form:
Log(q e – qt) = log q e –
(3.7)
Where qt is mercury concentration in solid phase Kg/m3 at time t and q e is equilibrium
concentration and k is kinetic constant (1 / time unit).
By plotting Log(q e – qt) versus time, it would be linear in case of fitting the data to the
kinetic expression.
3.3.2.2. Pseudo -Second Or der Expression
The kinetic expression: [95,96]
(qe – qt)2 (3.8)
Linear Form:
+
(3.9)
Where initial adsorption rate h = , qt is mercury concentration in solid phase
Kg/m3 at time t and q e is equilibrium concentration and k is kinetic constant (m3/Kg-
hr).
By plotting
versus time, it would be linear in case of fitting the data to the kinetic
expression.
3.3.2.3. Elovich’s Model
The kinetic expression: [102]
(3.10)
Linear Form:
+
(3.11)
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 in case of fitting the data to the kinetic
expression.
41
3.3.2.4. Weber and Morris Model
The kinetic expression: [98]
qt = k id + C (3.12)
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 in case of fitting the data to the kinetic
expression.
3.3.2.5. Diffusion -Chemi sorption Model
The kinetic expression: [99]
n K DC t n-1(qe – qt) 2 / q e 2 (3.13)
Linear form with n = 0.5 as per Sutherland analysis: [99]
+
(3.14)
Where initial adsorption rate is expressed by
Ki = K DC2 / q e (3.15)
By plotting / q t versus , it would be linear in case of fitting the data to the
kinetic expression.
3.3.2.6. Linear Dri ving Force diffusion Expression
The kinetic expression: [103]
Ln (1 – ) = k t (3.16)
Where is the fractional attainment of equilibrium q/q e and k is diffusion time
constant .
By plotting Ln (1 – ) versus t, it would be linear in case of fitting the data to the
kinetic expression.
3.3.3. 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 conce ntration in gas phase and time rearranged in
linear forms;
1- 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.
2- Wolborska Model,
tested to check the assumption of external mass transfer limitations.
3- Clark Model,
tested to check the assumption of mass transfer limitations and Fruendlich
isotherm.
4- Bohart & Adams Model
tested as the model assumes one component adsorpti on with chemical reaction
proposed by quasichemical expression and irreversible isotherm.
42
3.3.3.1. 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]
1. Physical adsorption with linear isotherm
q = K iC (3.17)
Where Ki is the adsorption constant in ( m3 gas/Kg adsorbent )
2. Simultaneous catalytic reaction assuming first order reaction equation
Φ (3.18)
3. Catalyst deactivation with covering active sites by reaction products .
Assuming first order deactivation reaction and deactivation function is
Kd Φ (3.19)
and the solution is
Φ xp ( -Kd t) (3.20)
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.
4. Ideal plug flow with no axial dispersion
5. Isothermal adsorption
6. The continuity equation of the fixed bed
+
+
+ (3.21)
Where v is superfic ial 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 = C o
At t = 0, x > 0, C = 0
The following solution was obtained:
(3.22)
Where
+ / (3.23)
= 1 +
(3.24)
and L is the length of the bed (m).
By plotting
versus t, it would be linear in case the model is fitting the
breakthrough data.
3.3.3.2. 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/C o from 10-5 to 0.05. It was assumed that adsorption rate is controlled
by the external mass transfer resistance . [107]
The continuity equation:
+
+
(3.25)
43
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 ∞ .
By introducing new variables and rearranging t hen continuity equation becomes:
(3.26)
and the initial and boundary conditions are x x τ o, and
∞ τ .
The kinetic equation for external resistance:
i (3.27)
Where Ci is concentration at gas -solid interface and by assuming quick intraparticle
diffusion then Ci << C and kinetic equation becomes:
(3.28)
Rearranging continuity equation gives:
(3.29)
The following was the linear solution form obtained by Wolborska:
(3.30)
By plotting
versus t, it would be linear in case the model is fitting the
breakthrough data.
3.3.3.3. Clark Model
The model was developed based on the following assumptions: [106]
1. Mass balance over a finite element of the bed
(3.31)
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).
2. Mass trans fer limitations
(3.32)
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).
3. Fruendlich isotherm for equilibrium
q = K C 1/n (3.33)
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 (3.34)
Where n does not equal one.
By plotting
– – 1] versus t, it would be linear in case the model is fitting
the breakthrough data.
44
3.3.3.4. Bohart & Adams Model
The model was developed based on the following assumptions: [94]
1. One trace component adsorption
2. Adsorption with simultaneous chemical reaction (Chemisorption)
3. Irreversible adsorption isotherm q = 0 at C = 0 & q = q s at C > 0
4. Ideal plug flow with negligible axial dispersion
5. Mass balance (continuity) equation for the fixed bed
+
+
(3.35)
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
concen tration in the adsorbent (Kg adsorbate/m3 adsorbent).
6. Adsorption kinetics described by quasichemical rate law
(3.36)
Where qs is the saturation capacity of q that corresponds to the equilibrium
condition at the gas/a dsorbent 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 occupie d.
The differential continuit y equation was solved by Cooney as below: [117]
z (3.37)
Where
= K C o (t –
) (3.38)
z =
(3.39)
By rearranging, the following is the linear form of the model:
+ + (3.40)
By plotting
versus
it would be linear in case of fitting breakthrough
data with the proposed model by Bohart & Adams.
45
4. Results and Discuss ions
4.1. 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 .
4.1.1. 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 4-1 Applying Freun dlich Isotherm for Data of 4 mm Adsorbent
y = 0.1679 x + 6.101
R² = 0.8637
0 0.5 1 1.5 2 2.5 3 3.5
-30 -25 -20 -15 -10 -5 0 Ln(q)
Ln(C) Ln(q) vs Ln(C) (Applying Freundlich)
Data 4 mm Sorbent
Linear (Data 4 mm
Sorbent)
y = 2267 .3x – 2203 .6
R² = 0.9971
0 5 10 15 20 25 30
0.97 0.975 0.98 0.985 0.99 q (KgHg/m3 particle)
C^1/n q vs C^ 1/n (Applying Freundlich )
Data 4 mm Sorbent
Linear (Data 4 mm
Sorbent)
46
Figure 4-2 Applying Freundlich Isotherm for Data of 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.
y = 0.0912 x + 5.6315
R² = 0.9962
3 3.2 3.4 3.6 3.8 4 4.2 4.4
-30 -25 -20 -15 -10 Ln(q)
Ln(C) Ln(q) vs Ln(C) (Applying Freundlich)
Data 2 mm Sorbent
Linear (Data 2 mm Sorbent)
y = 275.72x + 0.6376
R² = 0.9976
20 25 30 35 40 45 50 55 60 65 70
0 0.05 0.1 0.15 0.2 0.25 0.3 q (KgHg/m3 particle)
C^1/n q vs C^ 1/n (Applying Freundlich )
Data 2 mm Sorbent
Linear (Data 2 mm Sorbent)
47
4.1.2. Applying Langmuir Equilibrium Isotherm
By plotting Ce/qt versus Ce, the following charts indicate results of linear regression
where it would be linear in case of obeying Langmuir Equilibrium Isotherm.
Figure 4-3 Applying Langmuir Isotherm for Data of 4 mm Adsorbent
Figure 4-4 Applying Langmuir Isotherm for Data of 2 mm Adsorbent
Both particles show 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 q s where calculated very close to the
figures calculated from breakthrough data.
Table 4-1 Calculated Langmuir Constants
Particle Size R2 b
m3/Kg qs
Kg Hg/m3
sorbent from
Langmuir qs
Kg Hg/m3
sorbent from
breakthrough data
4 mm 0.9996 702,000,000 28.49 28.96361
2 mm 0.9994 493,333,333.3 67.57 68.1634 y = 0.0351 x + 5E-11
R² = 0.9996
0 1E-09 2E-09 3E-09 4E-09 5E-09 6E-09 7E-09 8E-09 0 5E-08 0.0000001 1.5E-07 0.0000002 2.5E-07 C / q
C Kg/m3 Gas C/q vs C (Applying Langmuir)
Data ( 4 mm Sorbent)
Linear (Data ( 4 mm
Sorbent))
y = 0.0148 x + 3E-11
R² = 0.9994
0 5E-10 1E-09 1.5E-09 2E-09 2.5E-09 3E-09 3.5E-09 0 5E-08 0.0000001 1.5E-07 0.0000002 2.5E-07 C / q
C Kg/m3 Gas C/q vs C (Applying Langmuir)
Data ( 2 mm Sorbent)
Linear (Data ( 2 mm
Sorbent))
48
4.2. 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;
1- Lagergren Pseudo -first order,
2- Pseudo -second order,
3- Elovich’ mod l,
4- Weber and Morris,
5- Diffusion -Chemisorption Model,
6- Linear Driving Force diffusion
4.2.1. Testing Lagergren Pseudo -First Order
By plotting Log(q e – qt) versus time , the following charts indicate the results of linear
regression where it would be linear in case of fitting the data to the kinetic expression.
Figure 4-5 Testing Lagergren Pseudo -First Order for Data of 4 mm Adsorbent
Figure 4-6 Testing Lagergren Pseudo -First Order for Data of 2 mm Adsorbent
Figures 4 -5 and 4-6 show that the expression gives moderate fit with R2 values of
0.9169 and 0.9462 at 4 mm and 2 mm particle size respectively.
y = -0.0009 x + 1.9271
R² = 0.9169
-1 -0.5 0 0.5 1 1.5 2 2.5
0 500 1000 1500 2000 2500 3000 Log (qe-qt)
t (Hours) Lagergren Pseudo -first order
Log (qe-qt) vs. t
4 mm Sorbent Loading
Data
Linear ( 4 mm Sorbent
Loading Data)
y = -0.0005 x + 2.6479
R² = 0.9462
-0.5 0 0.5 1 1.5 2 2.5
0 2000 4000 6000 8000 Log (qe-qt)
t (Hours) Lagergren Pseudo -first order
Log (qe-qt) vs. t
2 mm Sorbent Loading
Data
Linear ( 2 mm Sorbent
Loading Data)
49
4.2.2. Testing Pseudo -Second Order Expression
By plotting
versus time , the following charts indicate the results of linear
regression where it would be linear in case of fitting the data to the kinetic expression.
Figure 4-7 Testing Pseudo -Second Order for Data of 4 mm Adsorbent
Figure 4-8 Testing Pseudo -Second Order for Data of 2 mm Adsorbent
Figures 4 -7 & 4 -8 indicate that the expression gives poor fit with R2 values of 0.5904
and 0.7047 at 4 mm and 2 mm particle size respectively.
y = 0.0073 x + 59.166
R² = 0.5904
55 60 65 70 75 80 85 90
0 500 1000 1500 2000 2500 3000 t / qt (Hours – m3 particle / KgHg)
t (Hours) Pseudo -second order
t/qt vs. t
4 mm Sorbent Loading
Data
Linear ( 4 mm Sorbent
Loading Data)
y = 0.0042 x + 54.892
R² = 0.7047
60 65 70 75 80 85 90
0 2000 4000 6000 8000 t / qt (Hours – m3 particle / KgHg)
t (Hours) Pseudo -second order
t/qt vs. t
2 mm Sorbent Loading
Data
Linear ( 2 mm Sorbent
Loading Data)
50
4.2.3. Testing Elovich’s Model
By plotting qt versus Ln (t) , the following charts indicate the results of linear
regression where it would be linear in case of fitting the data to the kinetic expression.
Figure 4-9 Testing Elovich Model for Data of 4 mm Adsorbent
Figure 4-10 Testing Elovich Model for Data of 2 mm Adsorbent
Figures 4 -9 & 4 -10 show that 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 might be the main rate-limiting step at
smaller particles of 2 mm. y = 11.044x – 58.048
R² = 0.9148
0 5 10 15 20 25 30 35
4 5 6 7 8 9 qt (KgHg/m3 particle)
Ln(t) Elovich’s model
qt vs. Ln(t)
4 mm Sorbent Loading
Data
Linear ( 4 mm Sorbent
Loading Data)
y = 36.946x – 249.15
R² = 0.9823
15 25 35 45 55 65 75
7 7.5 8 8.5 9 qt (KgHg/m3 particle)
Ln(t) Elovich’s model
qt vs. Ln(t)
2 mm Sorbent Loading Data
Linear ( 2 mm Sorbent
Loading Data)
51
4.2.4. Testing Weber and Morris Model
By plotting qt versus , the following charts indicate the results of linear regression
where it would be linear in case of fitting the data to the kinetic expression.
Figure 4-11 Testing Weber & Morris Model for Data of 4 mm Adsorbent
Figure 4-12 Testing Weber & Morris Model for Data of 2 mm Adsorbent
Figures 4 -11 & 4 -12 show that 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. y = 0.0009 x – 0.0102
R² = 0.9786
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
0 10 20 30 40 50 60 qt (KgHg/m3 particle)
√t (Hours1/2) Weber and Morris
qt vs. √t
4 mm Sorbent Loading
Data
Linear ( 4 mm Sorbent
Loading Data)
y = 0.0014 x – 0.0288
R² = 0.9687
0.02 0.03 0.04 0.05 0.06 0.07 0.08
30 40 50 60 70 80 qt (KgHg/m3 particle)
√t (Hours1/2) Weber and Morris
qt vs. √t
2 mm Sorbent Loading
Data
Linear ( 2 mm Sorbent
Loading Data)
52
4.2.5. Testing Diffusion -Chemisorption Model
By plotting / q t versus , the following charts indicate the results of linear
regression where it would be linear in case of fitting the data to the kinetic expression.
Figure 4-13 Testing Diffusion -Chemisorption Model for Data of 4 mm Adsorbent
Figure 4-14 Testing Diffusion -Chemisorption Model for Data of 2 mm Adsorbent
Figures 4 -13 & 4 -14 show that the expression gives poor fit with R2 values of 0.7123
and 0.7567 at 4 mm and 2 mm particle size respectively. y = -0.0805 x + 5.0886
R² = 0.7123
1 2 3 4 5 6 7
0 10 20 30 40 50 60 √t / qt (Hours1/2 / KgHg/m3 particle)
√t (Hours1/2) Diffusion -Chemisorption Model
√t / qt vs. √t
4 mm Sorbent Loading Data
Linear ( 4 mm Sorbent
Loading Data)
y = -0.0147 x + 2.091
R² = 0.7567
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
30 40 50 60 70 80 √t / qt (Hours1/2 / KgHg/m3 particle)
√t (Hours1/2) Diffusion -Chemisorption Model
√t / qt vs. √t
2 mm Sorbent Loading
Data
Linear ( 2 mm Sorbent
Loading Data)
53
4.2.6. Testing Linear Driving Force diffusion Expression
By plotting Ln (1 – ) versus t, the following charts indicate the results of linear
regression where it would be linear in case of fitting the data to the kinetic expression.
Figure 4-15 Testing Vinod & Anirudvan (Linear Driving Force Diffusion)
Model for Data of 4 mm Adsorbent
Figure 4-16 Testing Vinod & Anirudvan (Linear Driving Force Diffusion)
Model for Data of 2 mm Adsorbent
Figures 4 -15 & 4 -16 show that 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. y = -0.0021 x + 1.0713
R² = 0.9169
-5 -4 -3 -2 -1 0 1 2
0 500 1000 1500 2000 2500 3000 Ln (1 – α)
t hours Linear Driving Force diffusion model by
Vinod & Anirudvan Ln (1 – α) vs. t
4 mm Sorbent Loading
Data
Linear ( 4 mm Sorbent
Loading Data)
y = -0.0011 x + 1.875
R² = 0.9462
-5 -4 -3 -2 -1 0 1
0 2000 4000 6000 Ln (1 – α)
t hours Linear Driving Force diffusion model by
Vinod & Anirudvan Ln (1 – α) vs. t
2 mm Sorbent Loading
Data
Linear ( 2 mm Sorbent
Loading Data)
54
4.3. Studying Adsorption Mathematical Models
The following models were used to test the breakthrough data using relationships
between mercury concentration in gas phase and time rearranged in linear forms;
1- Zhang and Cheng Model,
2- Wolborska Model,
3- Clark Model,
4- Bohart & Adams Model
4.3.1. Testing Zhang and Cheng Model
As per th e final linear form of solution, by plotting
versus t, the following
charts indicate the results of linear regression where it would be linear in case the
model is fitting the breakthrough data.
Figure 4-17 Testing Zhang & Cheng Model for Data of 4 mm Adsorbent
Figure 4-18 Testing Zhang & Cheng Model for Data of 2 mm Adsorbent
Figures 4-17 & 4 -18 show that 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. y = -0.0025 x + 4.0416
R² = 0.8568
-4 -3 -2 -1 0 1 2 3 4 5
0 500 1000 1500 2000 2500 3000 Lnln(Co/C)
t (Hours) Testing Zhang & Cheng Model
4 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
y = -0.0015 x + 5.512
R² = 0.9206
-6 -4 -2 0 2 4
0 2000 4000 6000 8000 Lnln(Co/C)
t (Hours) Testing Zhang & Cheng Model
2 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
55
4.3.2. Testing Wolborska Model
As per th e final linear form of solution, by plotting
versus t, the following
charts indicate the results of linear regression where it would be linear in case the
model is fitting the breakthrough data.
Figure 4-19 Testing Wolborska Model for Data of 4 mm Adsorbent
Figure 4-20 Testing Wolborska Model for Data of 2 mm Adsorbent
Figure s 4-19 & 4 -20 show that 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. y = 0.0054 x – 11.796
R² = 0.9639
-14 -12 -10 -8 -6 -4 -2 0 2 4
0 500 1000 1500 2000 2500 3000 Ln(C/Co)
t (Hours) Testing Wolborska Model
4 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
y = 0.0027 x – 13.52
R² = 0.8972
-14 -12 -10 -8 -6 -4 -2 0 2 4
0 2000 4000 6000 8000 Ln(C/Co)
t (Hours) Testing Wolborska Model
2 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
56
4.3.3. Testing Clark Model
As per th e final linear form of solution, by plotting
– – 1] versus t, the
following charts indicate the results of linear regression where it would be linear in
case the model is fitting the breakthrough data.
Figure 4-21 Testing Clark Model for Data of 4 mm Adsorbent
Figure 4-22 Testing Clark Model for Data of 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. y = -0.2664 x + 578.04
R² = 0.9639
-200 -100 0 100 200 300 400 500 600 700
0 500 1000 1500 2000 2500 3000 Ln[ (Co/C)^n -1 -1]
t (Hours) Testing Clark Model
4 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
y = -0.0267 x + 135.45
R² = 0.9035
-40 -20 0 20 40 60 80 100 120 140
0 2000 4000 6000 8000 Ln[ (Co/C)^n -1 -1]
t (Hours) Testing Clark Model
2 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
57
4.3.4. Testing 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 in
case of fitting breakthrough data to the proposed model by Bohart & Adams.
Figure 4-23 Testing Bohart & Adams Model for Data of 4 mm Adsorbent
Figure 4-24 Testing Bohart & Adams Model for Data of 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. y = -0.0067 x + 12.901
R² = 0.9925
-6 -4 -2 0 2 4 6 8 10 12 14
0 500 1000 1500 2000 2500 3000 Ln (Co – C / C)
t – L/v (Hours) Testing Bohart & Adams Model
4 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
y = -0.0035 x + 15.645
R² = 0.9843
-6 -4 -2 0 2 4 6 8 10 12 14
0 2000 4000 6000 8000 Ln (Co – C / C)
t – L/v (Hours) Testing Bohart & Adams Model
2 mm particle size
Breakthrough Data
Linear (Breakthrough Data)
58
The following table summarizes R2 values res ulted from linear regression of the
different models.
Table 4-2 R2 Values of Models Linear Regression
Model Zhang &
Cheng Model Wolborska
Model Clark Model Bohart &
Adams Model
R2
(4 mm particle) 0.8568 0.9639 0.9639 0.9925
R2
(2 mm particle) 0.9206 0.8972 0.9035 0.9843
4.4. 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, th e model has a considerable accuracy in prediction that can be utilized
ff ctiv ly in imulating th tudi d b d’ p rformanc with diff r nt op rating
condition . How v r, th following figur illu trat th mod l’ accuracy in
prediction.
Figure 4-25 Bohart & Adams Model Prediction versus Actual Breakthrough
Data for 4 mm Adsorbent
59
Figure 4-26 Bohart & Adams Model Prediction versus Actual Breakthrough
Data for 2 mm Adsorbent
4.5. Calculations and Building a Simulation Tool
The below detailed steps were followed in calculations .
1. Availability of data
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, t he 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.
2. Material balance and adsorbed mercury concentration in solid
Initially, the availab le 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).
3. Applying mathematica l expressions (Linear regression)
The available data of time (t) and corresponding mercury concentrations in gas
(C) and solid (q) phase s 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.
60
4. Linear regression results of the best fitting model (Bohart & Adams) and
calculating model constant
By plotting
versus
as per the re -arranged line ar 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 .
5. Calculating dimensionless parameters of Bohart & Adams model
The two dimensionless parameters & 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 = K C o (t –
) and z =
.
6. Using model equation in prediction with time (Simulation)
Finally, predicted mercury concentration and breakthrough profile are determined
using this equation
z with adequate time scale.
7. Simulation tool and drawing simulated breakthrough curves
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 inle t mercury
concentration). Then, the other sheets draw automatically all curves.
The following photos illustrate the use of the excel workbook.
61
Figure 4-27 Example of the Automated Use of Simulation Tool
4.6. 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 µ g/Nm3 of mercury. The following table summarizes the
bed design basis.
Table 4-3 Design Basis of Mercury Guard Bed
Maximum Feed Gas Flow Rate 1,350 MMSCFD
Maximum Inlet Mercury Concentration 20 µg/Nm3
Outlet Mercury Concentration Below 0.01 µ g/Nm3
Lifet ime 5 Years
Bed Diameter 5 Meters
Bed Height 2.50 Meters
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.
62
Figure 4-28 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 fo llowing graphs indicate the original bed height
performance and life time versus the bed after skimming the top 500 mm layer.
63
Figure 4-29 Simulated Breakthrough Curves for Designed Bed Height and after
Skimming of Top 500 mm
Figure 4 -29 shows that, 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 the same particles size .
64
4.7. Effect of Different Operating Parameters – Sensitivity
Analysis and Case Studies
The following parameters where studied and evaluated its effects:
1. Adsorbent particle size
(With constant flow rate of 1350 MMSCFD , inlet mercury concentration 20
µg/Nm3, and bed height 1.98 meters)
2. Inlet mercury concentration
(With constant flow rate of 1350 MMSCFD, bed height 1.98 meters, and
particles size of 2 -4 mm)
3. 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)
4. Bed Height
(With constant fl ow rate of 1350 MMSCFD, inlet mercury concentration 20
µg/Nm3, and particles size of 2 -4 mm)
4.7.1. Adsorbent Particle Size
Figure 4 -30 shows that the smaller particle size of 2 mm exhibits much better
performance and longer lifetime due to impr ovement 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 merit s 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 proces s.
Figure 4-30 Simulated Breakthrough Curves for Different Particle Size s
65
4.7.2. Inlet Mercury Concentration
Figure 4 -31 shows 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 4-31 Simulated Breakthrough Curves for Inlet Mercury Concentration
Effect
66
4.7.3. Feed Gas Flow Rate – Velocity
Figure 4 -32 shows 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 lead to decreasing contact time between gas and sol id phases and
accordingly reducing 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 4-32 Simulated Breakthrough Curves for Gas Flow Rate Effect
67
4.7.4. 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
secon d 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.
Howev er, 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 h eight of 2.48, 1.98, 1.5, 1.15, 0.8, and 0.4
meters respectively.
Figure 4-33 Simulated Breakthrough Curves for Bed Height Effect
68
5. Conclusion s and Recommendations
Mercury is a quite critical element that contaminates natural gas and it is subjected to
a considerable interest in research related to efficient removal by adsorption due to its
severe impacts on the environment and catastrophic incidents in natural gas industry
by its corrosive nat ure with specific metal types.
An industrial mercury removal bed was studied by evaluating bed and adsorbents
characteristics through applying of equilibrium isotherms, various kinetic expressions,
and various adsorption mathematical models to the availabl e breakthrough data.
Langmuir and Freundlich isotherm relationships were evaluated and found the system
is characterized by an irreversible isotherm that can be perfectly represented by
Langmuir isotherm.
The breakthrough data were examined by testing six kinetic equations to get some
information about the system. The first -order, pseudo -second order, Elovich, Weber
& Morris, diffusion -chemisorption model, and linear driving force diffusion model
were all assessed . Elovich’ mod l for chemis orption 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 might be the main rate -limiting step at
smaller particles of 2 mm. Weber & Morris 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 contributi ng rate -limiting step in both particles and
more at the larger 4 mm particle with neglecting external film resistance.
Based on the results of isotherm and kinetics assessment and for deeper bed analysis,
four adsorption mathematical models were selected t o test fitting of breakthrough
data. Zhang & Cheng model, Wolborska model, Clark model, and Bohart & Adams
model were all tested.
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
effectively to describe the kinetics of mercury adsorption, simulate the bed
performance , calculate predictions with different parameters, and draw concentration
profiles and breakthrough curves.
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:
1. Physical adsorption in which equilibrium can be described by irreversible
(recta ngular) isotherm and can use Langmuir isotherm to represent it
mathemat ically.
2. External film mass transfer resistance can be neglected.
3. The intraparticle diffusion is contributing as a rate-limiting step particularly
with larger particles.
4. Chemical reaction (chemisorption) takes place with active sites leading to high
fixation capacity and irreversibility (mercury cannot be desorbed) .
5. The system can be represented with ideal plug flow by neglecting axial
dispersion.
6. The adsorption can be described by isothermal as evident from field data.
7. The mercury concentration is in trace category.
8. All the above characteristics are in match with Bohart & Adams model
assumptions , and accordingly proven after finding it best fitting to breakthrough
69
data. Thus, the model can be used effectively to simulate bed performance and
predi ct breakthrough curves and life time with different parameters.
The model was used to simulate the bed performance with changes in operating
parameters. Different gas flow rates, inlet m ercury concentrations, adsorbent sizes,
and bed heights were evaluated by sensitivity analysis of different case studies.
Based on simulation analysis, model prediction found that increasing the gas flow rate
and the inlet concentration leads normally to f aster saturation and earlier breakthrough
similar to the effect of decreasing bed height (adsorbent quantity). Decreasing the size
of the adsorbent increases adsorption efficiency due to improving surface area and
mercury diffusivity into adsorbent pores.
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:
1. Option -1 Lower pressure drop and lower capital expenditures:
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
enhance s flow distribution by more ceramic balls.
2. Option -2 Maximizing plant availability and i mprovement of gas treatment :
The second valuable 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
arran gement, 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 s ubtracting 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 pack ed in one of the upstream dryer beds .
This molecular sieves bed will ac t 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. Nor mally, when water breakthrough occurs , it is
always a sudden condition 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
70
and exchangers until being difficult to be dissolved by methanol , and affect plant
productivity , and then a dry -out op eration 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 productio n during 36 -48 hours of dry -out as the need of dry -out will be
postponed for many years (maximizing plant availability) till saturating this guard b ed
which is not normally loaded with water content .
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 c ontinuous 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 hydrother mal
conditions of regeneration.
By approximate calculations for maximum continuous feed gas flow rate of 1,350
MMSCFD and maximu m 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 sie ves 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.
Another major advantage is that eliminating the need of dry -out saves the plant and its
equipment integrity . The dry -out exposes the cryogenic sections to hazardous thermal
stresses affecting the integrity and safe operation of plate -fin exchangers, thermal
disturbance in flanges leading to the hazard of hydrocarbon leakage cases, and danger
of mercury -aluminum amalgamation appears with operating conditions of dry -out.
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 mercu ry concentrations while using t he same particle size of 2 -4 mm, the
following plots , resulted from simulation , indicate 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/N m3 (350 ng/Nm3), 1, 2, 5, and 20 µg/Nm3 respectively.
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.
71
Figure 5-1 Simulated Breakthrough Curves for Proposed Bed Height with
Different Inlet Me rcury Concentrations
72
Figure 5-2 Simulated Breakthrough Curves till Saturation for Proposed Bed
Height with Different Inlet Mercury Concentrations
73
6. References
1. "hydrargyrum" . Random House Webster's Unabridged Dictionary .
2. Ehrlich, H.L. and Newman, D.K. eds., 2008. Geomicrobiol ogy. CRC Press.
p. 265. ISBN 978-0-8493 -7906 -2.
3. "Mercury and the environment — Basic facts" . Environment Canada , Federal
Government of Canada. 2004. Retrieved 27 March2008.
4. "Merc ury — Element of the ancients" . Center for Environmental Health
Sciences, Dartmouth College . Retrieved 9 April 2012.
5. Burkholder, M. and Johnson, L., 2008. Colonial Latin A merica. OUP
Catalogue. pp. 157–159. ISBN 0-19-504542 -4.
6. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report. ,
Summary of the Report, Chapter 7, paragraphs 100 to 103
7. Lide, D.R. ed., 2005. CRC handbook of chemistry and physics (Vol. 86). CRC
press.
8. http://www.ptable.com/#Property/State
9. Senese, F., 2007. Why is mercury a liquid at STP?. General Chemistry Online
at Frostburg State University. Retrieved May, 1.
10. Norrby, L.J., 1991. Why is mercury liquid? O r, why do relativistic effects not
get into chemistry textbooks?. J. Chem. Educ, 68(2), p.110.
11. 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.
12. Rustad, D.S., 1987. How soft is mercury?. J. Chem. Educ, 64(5), p.470.
13. Greenwood, N.N. and Earnshaw, A., 1997. Chemistry of the Elements 2nd
Edition. Butterworth -Heinemann. ISBN 0-08-037941 -9.
14. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 2, paragraphs 39 to 43
15. 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.
16. Soratur, S.H., 2002. Essentials of dental materials. Jaypee Bros. Medical
Publishers. p. 14. ISBN 978-81-7179 -989-3.
17. Barbalace, K., 1995. Periodic table of elements: Hg- Mercury.
EnvironmentalChemistry.com, 2016.
18. 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.
19. Goyer, R., Aposhian, V., Arab, L., Bellinger, D., Burbacher, T., Burke, T.,
Jacobso n, J., Knobeloch, L., Stern, A. and Ryan, L., 2000. Toxicological
effects of methylmercury. Washington, DC: National Research Council.
20. National Research Council (U.S.) – Board on Enviro nmental Studies and
Toxicology, 2000 . Toxicological effects of methylmercury . National
Academies Press.
21. http://www.chemistryexplained.com/elements/L -P/Mercury.html
22. Surmann, P. and Zeyat, H., 2005. Voltammetric analysis using a self –
renewable non -mercury electrode. Analy tical and Bioanalytical
Chemistry,383(6), pp.1009 -1013.
23. http://www.lenntech.com/periodic/elements/hg.htm
24. Rogalski, A., 2000. Infrared Detectors (Vol. 10). CRC Press , pp.507.
74
25. 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).
26. Pacyna, E.G., Pacyna, J.M., Steenhuisen, F. and Wilson, S., 2006. Global
anthropogenic mercury emission inventory for 2000. Atmospheric
environment, 40(22), pp.40 48-4063.
27. "What is EPA doing about mercury air emissions?" . United States
Environmental Protection Agency (EPA). Retrieved 1 May 2007.
28. Solnit, R., 2007. Winged mercury and the golden calf. Orion Magazine.
Retrieved, pp.12 -03.
29. "Mercury -containing Products" . United States Environmental Protection
Agency (EPA). Retrieved 1 May 2007.
30. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 2, paragraphs 44 to 46 and 48
31. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 2, paragraphs 49 to 52
32. http://www.us gs.gov/themes/factsheet/146 -00/
33. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Cha pter 3, paragraphs 55 & 56, Chapter 2, paragraph
47.
34. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapt er 6, paragraphs 77 to 80
35. 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.
36. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 3 & 4, paragraphs 61, 62 and 66
37. "Mercury: Spills, Disposal and Site Cleanup" . Environmental P rotection
Agency. Retrieved 11 August 2007.
38. 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.
39. Liang, Y.X., Sun, R.K., S un, 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.
40. McFarland, R.B. and Reigel, H., 1978. Chroni c mercury poisoning from a
single brief exposure. Journal of Occupational and Environmental Medicine,
20(8), pp.532 -534.
41. Mercury, W.I., 1991. Environmental Health Criteria 118. Geneva: World
Health Organization, 107.
42. World Health Organization, 1991. Inorganic mercury —Environmental health
criteria. WHO, Geneva, Switzerland, 118.
43. Senn, E., 1996. Controlling metallic mercury exposure in the workplace: A
guide for employers. DIANE Publishing.
44. 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.
45. http://www.pca.state.mn.us/index.php/topics/mercury/health -and-
environmental -effects -of-mercury.html
75
46. Sheet, H.S.F., 1996. New Jersey Dep artment of health and senior
services.Trenton, NJ.
47. 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.
48. Chemicals, U.N.E.P ., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 6, paragraphs 87 to 90
49. 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 com bustion
flue gas in a bench -scale entrained -flow reactor. Fuel processing
technology, 85(6), pp.463 -472.
50. 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.
51. 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.
52. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 8
53. "Minamata Convention Agreed by Nations" . United Nations Environment
Program . Retrieved 19 January 2013.
54. http://www.un.org/apps/news/story.asp?NewsID=43963&Cr=mercury&Cr1=#
.UPt4ikfdP5M
55. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 8, paragraphs 112 to 115
56. http://www.globalcement.com/magazine/articles/845 -global -cement -emissions –
standards
57. http://www.who.int/mediacentre/factsheets N°361
58. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report ,Chapter 9, paragraphs 128 to 133
59. Lee, J.D., 2008. Concise inorganic chemistry. John Wiley & Sons.
60. Waldron, H.A., 1983. Did the Mad Hatter have mercury poisonin g?. British
medical journal (Clinical research ed.), 287(6409), p.1961.
61. "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 .
62. "Minamata Disease The History and Measures" . Ministry of the Environment,
Government of Japan. Retrieved 7 July 2009.
63. Normile, D., 2013. In Minamata, mercury still divides. Science, 341(6153),
pp.1446 -1447.
64. Chemicals, U.N.E.P., Global Mercury Assessment (2002). Ref Type: Report.,
Summary of the Report, Chapter 3, paragraphs 55 & 56; and C hapter 2,
paragraph 47.
65. 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.
66. Mohamed A, S., Abd el, 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.
76
67. Assassa, M.F., Mohammd, Y.S., Aglan, M.A. and Farahat, M.I., 2015.
ESTIMATION O F HAIR MERCURY LEVELS FOR ASSESSMENT OF
METHYL MERCURY EXPOSURE IN ALDAKAHLIA -EGYPT. AL-
AZHAR ASSIUT MEDICAL JOURNAL, 13(1).
68. El-Nady, F.E., 1986. Bioaccumulation of mercury in some coastal marine fish
from Alexandria waters. In The FAO/ UNEP/WHO/IOC/IAEA M eeting on
the Biogeochemical Cycle of Mercury in the Mediterranean, Siena, Italy, 27 –
31 August 1984, FAO FISH. REP. no. 325 (pp. 74 -77).
69. http://www.greenprophet.com/2011/10/egypt -mercury -disposal/
70. Agency for Toxic Substances and Disease Registry, "Mercury" –
http://www.atsdr.cdc.gov/tfacts46. html
71. API, R., 2003. 571 –Damage Mechanisms Affecting Fixed Equipment in the
Refining Industry. American Petroleum Institute, pp.30 -53.
72. McIntyre, D.R., Case, R.P., Ballantyne, T. and Woodward, S., 2013, March.
Environmenta lly Assisted Cracking of Nickel Steels In Liquid Mercury,
Hydrogen and Methanol. In CORROSION 2013. NACE International.
73. Al-Alf, H.K.A. and Fl -Behairy, S., Mercury Removal from Natural Gas in
Egypt.
74. Shafawi, A., Ebdon, L., Foulkes, M., Stockwell, P. and Cor ns, W., 1999.
Determination of total mercury in hydrocarbons and natural gas condensate by
atomic fluorescence spectrometry. Analyst, 124(2), pp.185 -189.
75. Im abbott, Paul openshaw, "Mercury Removal Technology and Its
Application", Synetix, Belasis Av., Billmgham, Cleveland, TS23 1LB, UK
76. "Moomba Plant Update" (Press release). Adelaide, South Australia: Santos.
2004 -03-05. Archived from the original on 2004 -03-05. Retrieved 2013 -01-
18.
77. http://EnvironmentalChemistry.com/yogi/periodic/Hg.html
78. Kolman, D.G., E nvironmentally Induced Cracking. Liquid Metal
Embrittlement, ASM Handbook, 13, pp.381 -392.
79. Robertson, W.M., 1966. PROPAGATION OF A CRACK FILLED WITH
LIQUID METAL. North American Aviation Science Center, Thousand Oaks,
Calif..
80. 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.
81. Stoloff, N.S. and Johnston, T.L., 1963. Crack propagation in a liquid metal
environment. Acta Meta llurgica, 11(4), pp.251 -256.
82. 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.
83. Gordon, P. and An, H.H., 1982. The mechanisms of crack ini tiation and crack
propagation in metal -induced embrittlement of metals. Metallurgical
Transactions A, 13(3), pp.457 -472.
84. Lynch, S.P., 1988. Environmentally assisted cracking: overview of evidence
for an adsorption -induced localised -slip process. Acta Metallurgica, 36(10),
pp.2639 -2661.
85. 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.
77
86. http://www.twi -global.com/technical -knowledge/faqs/material -faqs/faq -what -is-
liquid -metal -embrittlement -and-what -are-common -embrittling -metal -couples/
87. Ruthven, D.M., 1984. Principles of adsorption and adsorption processes. John
Wiley & Sons.
88. Srivastava, V.C., Mall, I.D. and Mishra, I.M., 2008. Adsorption of toxic metal
ions onto activated carbon: Study of sor ption behaviour through
characterization and kinetics. Chemical Engineering and Processing: Process
Intensification, 47(8), pp.1269 -1280.
89. Langmuir, I., 1918. The adsorption of gases on plane surfaces of glass, mica
and platinum. Journal of the American Che mical society, 40(9), pp.1361 -1403.
90. Freundlich, H.M.F., 1906. Over the adsorption in solution. J. Phys.
Chem,57(385471), pp.1100 -1107.
91. 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).
92. Temkin, M.I. and Pyzhev, V., 1940. Kinetics of ammonia synthesis on
promoted iron catalysts. Acta physiochim. URSS, 12(3), pp.217 -222.
93. Yang, R.T., 1987. Gas separation by adsorption processes. Butterworth –
Heinemann.
94. 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.
95. 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.
96. Ho, Y.S. and McKay, G., 1998. Kinetic model for lead (II) sorption on to
peat.Adsorption science & technology, 16(4), pp.243 -255.
97. Lagergren, S., 1898. About the theory of so -called adsorption of soluble
substances.
98. 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.
99. Sutherland, C., 2009. Removal of heavy metals from water using low -cost
adsorbents: process development (Doctoral dissertation).
100. Low, M.J.D., 1960. Kinetics of Chemisorption of Gases on
Solids. Chemical Reviews, 60(3), pp.267 -312.
101. 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.
102. Zeldowitsch, J., 1934. Über den mechanismus der katalytischen
oxydation von CO an MnO2. Acta physicochim. URSS, 1, pp.364 -449.
103. Vinod, V.P. and Anirudhan, T.S., 2002. Sorption o f tannic acid on
zirconium pillared clay. Journal of Chemical Technology and
Biotechnology, 77(1), pp.92 -101.
104. Thomas, H.C., 1944. Heterogeneous ion exchange in a flowing
system.Journal of the American Chemical Society, 66(10), pp.1664 -1666.
105. 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.
106. Clark, R.M., 1987. Evaluating the cost and performance of field -scale
granular activated carbon systems. Environmental science & technology,21(6),
pp.573 -580.
78
107. Wolborska, A., 1989. Adsorption on activated carbon of p -nitrophenol
from aqueous solution. Water Research, 23(1), pp.85 -91.
108. Zhang, H. and Cheng, D., 2000. Mathematical model for a fixed bed
adsorptive reactor. Carbon, 38(6), pp.877 -880.
109. 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 u sing rice husk ash. Journal of Environmental
Chemical Engineering, 3(3), pp.1827 -1836.
110. 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 Technolo gy, 192(1), pp.47 -53.
111. 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.
112. Gray don, 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.
113. Camargo, C.L.M., de Resende, N.S., de Oliveira, A.G., Salim, V.M.M.
and Ta vares, F.W., 2014. Investigation of adsorption -enhanced reaction
process of mercury removal from simulated natural gas by mathematical
modeling.Fuel, 129, pp.129 -137.
114. Bingham, M.D., 1990. Field detection and implications of mercury in
natural gas. SPE Prod uction Engineering, 5(02), pp.120 -124.
115. Spiric, Z., 2001. Innovative approach to the mercury control during
natural gas processing. Proceedings of ETCE.
116. Wilhelm, S.M. and Bloom, N., 2000. Mercury in petroleum. Fuel
Processing Technology, 63(1), pp.1 -27.
117. Cooney, D.O., 1999. Adsorption design for wastewater treatment, CRC
Pres.INC., Boca Raton, Florida, USA.
118. 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 Ma terials, 146(1), pp.1 -11.
119. 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.
120. 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.
121. Wade, C.B., Thurman, C., Freas, W., Student, J., Matty, D. and
Mohanty, D.K., 2012. Preparation and characterization of high effic iency
modified activated carbon for the capture of mercury from flue gas in coal –
fired power plants. Fuel processing technology, 97, pp.107 -117.
122. Guo, P., Guo, X. and Zh ng, C., 010. Rol of γ -Fe 2 O 3 in fly ash
for mercury removal: results of density fu nctional theory study. Applied
Surface Science, 256(23), pp.6991 -6996.
123. 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 fo r high -temperature capture
of mercury and arsenic from fuel gas. Fuel, 89(6), pp.1323 -1325.
79
124. 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).
125. 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.
126. 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.
127. 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.
128. 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.
129. 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.
130. Scala, F., 2001. Simulation of mercury capture by activated carbon
injection in incinerator flue gas. 2. F abric filter removal. Environmental
science & technology, 35(21), pp.4373 -4378.
131. Flora, J.R., Hargi , R.A., O’Dowd, W.J., P nnlin , H.W. and Vidic,
R.D., 2003. Modeling sorbent injection for mercury control in baghouse
filters: I —model development and sensi tivity analysis. Journal of the Air &
Waste Management Association, 53(4), pp.478 -488.
132. Flora, J.R., Hargi , R.A., O’Dowd, W.J., P nnlin , H.W. and Vidic,
R.D., 2003. Modeling sorbent injection for mercury control in baghouse
filters: II —Pilot -scale studie s and model evaluation. Journal of the Air &
Waste Management Association, 53(4), pp.489 -496.
133. Ho, Y.S. and McKay, G., 1998. A comparison of chemisorption kinetic
models applied to pollutant removal on various sorbents. Process Safety and
Environmental Prot ection, 76(4), pp.332 -340.
134. 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.
135. 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.
136. 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 manageme nt, 118, pp.135 -143.
137. 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
80
activated alumina. Water quality research journal of Canada, 41(2), pp.147 –
156.
138. Ho, N., 2012. Modeling hydrogen sulfide adsorption by activated
carbon made from anaerobic digestion by -product (Doctoral dissertation).
139. Sugier, A. and Villa, F.L., Institut Francais Du Petrole, 1978. Process
for removing mercury from a gas or a liquid by absorpt ion on a copper sulfide
containing solid mass. U.S. Patent 4,094,777.
140. 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.
141. 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.
142. Eswaran, S., Stenger, H.G. and Fan, Z., 2007. Gas -phase mercury
adsorption rate studies. Energy & fue ls, 21(2), pp.852 -857.
143. 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.
144. 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.
145. 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 A MERICAN
CHEMICAL SOCIETY(Vol. 220, pp. U386 -U386). 1155 16TH ST, NW,
WASHINGTON, DC 20036 USA: AMER CHEMICAL SOC.
146. Vidic, R.D., 2002. Adsorption of elemental mercury by virgin and
impregnated activated carbon (pp. 15 -44). Lewis Publishers: Boca Raton, FL.
147. 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.
148. Huawei, Z., Xiuli, L., Li, W. and Peng, L., 2014. Cha racteristics and
Stability of Mercury Vapor Adsorption over Two Kinds of Modified
Semicoke. The Scientific World Journal, 2014.
149. Borderieux, S., Wu, C.Y., Bonzongo, J.C. and Powers, K., 2004.
Control of elemental mercury vapor in combustion systems using Fe 2O3
nanoparticles.Aerosol Air Qual. Res, 4(1), pp.74 -90.
150. Liu, Y., 2014. Elemental Mercury Removal from Flue Gas by
Metal/Metal Oxide Decorated Graphene Oxide Composites (Doctoral
dissertation, University of Alberta).
151. 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.
152. 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.
81
153. 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.
154. Wdowin, M., Wiatros -Motyka, M.M., Panek, R., Steven s, L.A.,
Franus, W. and Snape, C.E., 2014. Experimental study of mercury removal
from exhaust gases. Fuel, 128, pp.451 -457.
155. Ballestero, D., Gómez -Giménez, C., García -Díez, E., Juan, R., Rubio,
B. and Izquierdo, M.T., 2013. Influence of temperature and rege neration
cycles on Hg capture and efficiency by structured Au/C regenerable sorbents.
Journal of hazardous materials, 260, pp.247 -254.
156. Olson, E.S., Dunham, G.E., Sharma, R.K. and Miller, S.J., 2000.
Mechanisms of mercury capture and breakthrough on activat ed carbon
sorbents. ACS Fuels, 45, pp.886 -889.
157. 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.
158. Daza, L., Mendioroz, S. and Pajares, J., 1991. Mercury adsorption by
sulfurized fibrous silicates. Clays and Clay Minerals, 39(1), pp.14 -21.
159. Padak, B., 2011. Mercury reaction chemistry in combustion flue gases
from experiments an d theory (Doctoral dissertation, Stanford University).
160. 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.
161. 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.
162. 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.
163. 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.
164. Xie, J., Xu, H., Q u, 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.
165. You, X., 2002. Sulfur Impr egnation of Activated Carbon Through
Hydrogen Sulfide Oxidation (Doctoral dissertation, University of Pittsburgh).
166. 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 -monoli th
regenerable sorbents. Journal of hazardous materials, 193, pp.304 -310.
167. 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 mercu ry. Chemical engineering
journal, 174(1), pp.86 -92.
168. Eldefrawy, M.M., Kenawy, I.M. and El -Tabey, R.M., 2014.
Application of Nanosized Mesoporous Aluminosilicates for Adsorption of Hg
82
from Aqueous Solutions: Kine tic, Isotherm and Thermodynamic
Studies. International Journal, 2(12), pp.764 -781.
169. 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.
170. 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.
171. Perry, R.H. and Green, D.W., 1999. Perry's chemical engi neers'
handbook. McGraw -Hill Professional .
172. Younger, A.H., 2004. Natural Gas Processing Principles and
Technology -Part I. Gas Processors Association, Tulsa Oklahoma .
173. McKetta Jr, J.J. ed., 1992. Unit Operations Handbook: Volume 2 (In
Two Volumes) (Vol. 2). C RC Press .
174. McCabe, W.L., Smith, J.C. and Harriott, P., 1993. Unit operations of
chemical engineering (Vol. 5, p. 154). New York: McGraw -Hill.
175. Biscan, D.A., Gebhard, R.S. and Matviya, T.M., 1986, June. Impact of
process conditions on mercury removal from nat ural gas using activated
carbon. In 8 th International Conference on Liquefied Natural Gas .
176. Mokhatab, S., Mak, J.Y., Valappil, J.V. and Wood, D.A., 2013.
Handbook of liquefied natural gas. Gulf Professional Publishing .
177. Mokhatab, S. and Poe, W.A., 2012. Hand book of natural gas
transmission and processing. Gulf Professional Publishing.
178. Eckersley, N., 2010. Advanced mercury removal
technologies. Hydrocarbon processing, 89(1).
179. 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.
83
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 باإلضافة 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 اإلمتزاز . وحدة
صناعية عبارة عن Fixed Bed معبأة بنوع خاص من الحبيبات الشرهة إلمتزاز 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 و اختالف ارتفاع طبقة
الحبيبات واختالف معدل سريان الغاز وتركيزات مختلفة للزئبق بالغاز عند دخول الوحدة .
لوحظ ان المادة المستخدمة الزالة الزئبق تثبت سعة امتزاز عالية للزئبق من خالل ما ذكر في معلومات الشركة
المصنعة ومن خالل نتائج المحاكاة ومن خالل اداء الوحدة بالحقل مما يعطي الوحدة استمرارية للعمل لسنين
طويلة.
البيانات المتاحة الختراق الزئبق لوحدة امتزاز تم استخدامها في إختبار نموذجي إتزان وستة نماذج لدينياميكية
االمتزاز و اربعة نماذج محاكاة رياضية للحصول على معلومات عميقة عن خصائص وديناميكية الوحدة.
احد النماذج الرياضية يعطي افضل تطابق لبيانات االختراق حيث يفترض النموذج عملية امتزاز لعنصر واحد
قليل التركيز ووجود عملية تفاعل كيميائي وإتزان غير قابل لإلنعكاس عند ثبوت الحرارة وتجاهل التشتت
المحوري وبالتالي النموذج المطابق يثبت خصائص المادة وطبيعة إتزان العملية ووجود عملية تفاعل كيميائي مع
اإلمتزاز.
بناء على تحليل نتائج المحاكاة , تنبؤ النموذج الرياضي وجد ان زيادة معدل سريان الغاز وتركيز الزئبق يؤدي
طبيعياً كما هو متوقع إلى إختراق الزئبق بشكل مبكر للوحدة وتشبع الوحدة بالزئبق بشكل اسرع . استخدام
مقاسات اصغر للحبيبات يرفع من كفائة عملية اإلمتزاز بفضل تحسين المساحات السطحية وإنتشار ابخرة الزئبق
خالل مسام الحبيبات. بشكل عام وُجِد ان ارتفاعات اقل من طبقة الحبيبات تستطيع استيعاب معدالت 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 خالل مقترحين موفرين للنفقات وأحد المقترحَيْن يُحَسِنْ من معالجة الغاز
ويعَظِمْ من توافر المحطة لإلنتاج وذلك من خالل 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… [603117] (ID: 603117)
Dacă considerați că acest conținut vă încalcă drepturile de autor, vă rugăm să depuneți o cerere pe pagina noastră Copyright Takedown.