Caltex Process Plant Corrosion

SECTION 3
PROCESS PLANT
CORROSION
REFINERY MATERIALS MANUAL
October 1999
SECTION 3
PROCESS PLANT CORROSION
TABLE OF CONTENTS
1.0
ABSTRACT ................................................................
................................................................................................
...............................................
............... 1
2.0
INTRODUCTION ................................................................
................................................................................................
......................................
...... 1
3.0
REFINERY OPERATING
OPERATING OBJECTIVES ................................................................
............................................................................
............ 1
3.1
3.2
3.3
3.4
3.5
3.6
SAFETY, THE ENVIRONMENT, AND RELIABILITY............................................................................2

REFINING PROCESS OVERVIEW.....................................................................................................3
REFINERY PROCESS FLOW ............................................................................................................5
PROCESS INTERACTIONS WITH CORROSION.............................................................................11
ROLE OF PERSONNEL IN EQUIPMENT INTEGRITY.......................................................................11
CONCLUSION............................................................................................................................12
4.0
CRUDE UNIT................................
UNIT ................................................................
................................................................
..........................................
.......... 13
4.1
4.2
4.3
4.4
CRUDE UNIT CORROSION AND CORROSION CONTROL ...........................................................13

CRUDE UNIT PROCESS DESCRIPTION.........................................................................................13
CRUDE DESALTING ....................................................................................................................17
MATERIALS OF CONSTRUCTION................................................................................................19
4.4.1 Columns .............................................................................................................................19
4.4.2 Exchangers and Piping........................................................................................................20
4.4.3 Fired Heaters.......................................................................................................................21
4.5
MATERIALS AND CORROSION PROBLEMS..................................................................................22
4.5.1 Inorganic Salts ....................................................................................................................22
4.5.2 Sulfur Compounds ..............................................................................................................22
4.5.3 Organic Acids .....................................................................................................................23
4.5.4 Organic Chlorides ...............................................................................................................24
4.5.5 Corrosion Control Measures................................................................................................24
4.5.6 Corrosion Monitoring .........................................................................................................27
4.6
INSPECTION...............................................................................................................................29
5.0
5.1
5.2
5.3
FLUID CATALYTIC CRACKING UNIT ................................................................

.........................................................................
......... 31
ABSTRACT ..................................................................................................................................31
INTRODUCTION ........................................................................................................................31
FLUID CATALYTIC CRACKING UNIT PROCESS DESCRIPTION ......................................................31
5.3.1 Cracking Reaction ...............................................................................................................31
5.3.2 Catalyst / Catalyst Circulation .............................................................................................32
5.3.3 Cat Cracker Hardware / Process Flow ..................................................................................32
5.3.4 Riser/Reactor.......................................................................................................................33
5.3.5 Regenerator ........................................................................................................................34
5.3.6 Catalyst Transfer System .....................................................................................................34
5.3.7 Flue Gas System ..................................................................................................................35
5.3.8 Main Fractionator ...............................................................................................................35
5.4
5.4.1 Reactors..............................................................................................................................41
5.4.2 Regenerator ........................................................................................................................43
5.4.3 Catalyst Transfer Piping System...........................................................................................45
October 1999
5.4.4 Reaction Mix Line, Main Fractionator and Bottoms Piping...................................................45

5.4.5 Flue Gas System ..................................................................................................................46
5.5
CORROSION / METALLURGICAL DAMAGE MECHANISMS AND CONTROL MEASURES...............46
5.5.1 High Temperature Oxidation ..............................................................................................47
5.5.2 High Temperature Sulfidation .............................................................................................48
5.5.3 High Temperature Carburization.........................................................................................49
5.5.4 Polythionic Acid Stress Corrosion Cracking (PASCC)............................................................49
5.6
CATALYST EROSION ...................................................................................................................50
5.6.1 High Temperature Graphitization........................................................................................51
5.6.2 Sigma Phase Embrittlement ................................................................................................51
5.6.3 885oF Embrittlement...........................................................................................................52
5.6.4 Creep Embrittlement...........................................................................................................52
5.6.5 High Temperature Creep ....................................................................................................53
5.6.6 Thermal Fatigue ..................................................................................................................53
5.7
INSPECTION/MONITORING METHODS .....................................................................................53
6.0
6.1
CATALYTIC LIGHT ENDS RECOVERY UNIT ................................................................

................................................................ 59
ABSTRACT ..................................................................................................................................59
6.1.1. CLER Process Description ....................................................................................................59
6.2
6.2.1. Columns .............................................................................................................................59
6.2.2 Exchangers .........................................................................................................................60
6.2.3 Piping .................................................................................................................................60
6.3
CORROSION PROBLEMS ............................................................................................................60
6.3.1 Hydrogen Induced Damage ................................................................................................61
6.3.2 Ammonia Stress Corrosion Cracking....................................................................................62
6.3.3 Carbonate Stress Corrosion Cracking ..................................................................................62
6.3.4 Fouling/Corrosion of Reboiler Circuits .................................................................................62
6.4
CONTROL MEASURES ...............................................................................................................62
6.4.1 Water Washing ...................................................................................................................62
6.4.2 Polysulfide Injection ............................................................................................................64
6.4.3 Corrosion Inhibitors ............................................................................................................64
6.4.4 Corrosion Monitoring .........................................................................................................64
6.4.5 Corrosion Probes ................................................................................................................65
7.0
7.1
7.2
CATALYTIC REFORMING
REFORMING UNIT ................................................................
................................................................................
................ 69
CORROSION ..............................................................................................................................69
MATERIALS .................................................................................................................................70
8.0
HYDROPROCESSING UNITS ................................................................

....................................................................................
.................... 75
8.1
8.2
8.3
8.4
ABSTRACT ..................................................................................................................................75
INTRODUCTION ........................................................................................................................75
PROCESS DESCRIPTIONS............................................................................................................76
MATERIALS SELECTION ..............................................................................................................76
8.4.1 Reactor System ...................................................................................................................76
8.4.2 Reactor Feed System ...........................................................................................................77
8.4.3 Reactors..............................................................................................................................77
8.4.4 Reactor Effluent System.......................................................................................................78
8.4.5 Reactor Effluent - Distillation Feed Exchangers.....................................................................78
8.4.6 Effluent Air Coolers .............................................................................................................78
8.4.7 Effluent Air Cooler Inlet and Outlet Piping...........................................................................79
8.4.8 Separator Vessels.................................................................................................................80
8.4.9 Recycle Hydrogen System ...................................................................................................80
8.4.10
Distillation Section ..........................................................................................................80
October 1999
8.5
CORROSION PHENOMENA IN HYDROPROCESSING UNITS........................................................81

8.5.1 High Temperature Hydrogen Attack....................................................................................81
8.5.2 High Temperature Hydrogen Sulfide / Hydrogen Corrosion ................................................82
8.5.3 High Temperature Hydrogen Sulfide Corrosion in Areas with Negligible Hydrogen .............82
8.5.4 Naphthenic Acid Corrosion .................................................................................................83
8.5.5 Ammonium Bisulfide Corrosion...........................................................................................83
8.5.6 Chloride Stress Corrosion Cracking .....................................................................................85
8.5.7 Polythionic Acid Stress Corrosion Cracking (PASCC)............................................................86
8.5.8 Wet H2S Cracking ...............................................................................................................86
8.5.9 Material Property Degradation Mechanisms ........................................................................88
9.0
ALKYLATION UNIT ................................................................

................................................................................................
.................................. 95
9.1
9.2
ABSTRACT ..................................................................................................................................95
PROCESS DESCRIPTION..............................................................................................................95
9.2.1 Reaction Section .................................................................................................................95
9.2.2 Treating Section..................................................................................................................96
9.2.3 Fractionation Section ..........................................................................................................96
9.2.4 Refrigeration Section...........................................................................................................96
9.3
MATERIALS OF CONSTRUCTION..............................................................................................101
9.4
MATERIALS AND CORROSION PROBLEMS................................................................................101
9.4.1 Sulfuric Acid Corrosion......................................................................................................102
9.4.2 Acid and Neutral Esters .....................................................................................................103
9.4.3 Acid Carry-over .................................................................................................................104
9.4.4 Corrosion Under Insulation ...............................................................................................104
9.4.5 Fouling Problems ..............................................................................................................105
9.5
CORROSION CONTROL MEASURES..........................................................................................105
9.5.1 Reactor Section Corrosion .................................................................................................105
9.5.2 Tower Overhead Corrosion ...............................................................................................105
9.5.3 Reboiler Corrosion and Fouling Control ............................................................................106
9.5.4 Acid Tanks ........................................................................................................................106
9.5.5 Corrosion Control During Unit Shutdowns........................................................................107
9.5.6 Corrosion Under Insulation ...............................................................................................107
9.5.7 Corrosion Monitoring .......................................................................................................107
10.0 AMINE TREATING UNIT ................................................................

........................................................................................
........................ 111
10.1
10.2
10.3
ABSTRACT ................................................................................................................................111
INTRODUCTION ......................................................................................................................111
TYPES OF AMINES USED...........................................................................................................111
10.3.1
Monoethanolamine (MEA) ............................................................................................111
10.3.2
Diethanolamine (DEA)...................................................................................................112
10.3.3
Methyldiethanolamine (MDEA) .....................................................................................112
10.3.4
Diisopropanolamine (DIPA) ...........................................................................................112
10.3.5
Diglycolamine Agent (DGA) ..........................................................................................112
10.3.6
Specialty Amines ...........................................................................................................112
10.4 REFINERY AMINE UNIT PROCESS DESCRIPTION .......................................................................113
10.5 CORROSION PHENOMENA ......................................................................................................117
10.6 CORROSIVE SPECIES.................................................................................................................118
10.7 AMINE DEGRADATION.............................................................................................................121
10.8 CRACKING PHENOMENA .........................................................................................................122
10.9 CORROSION INHIBITORS .........................................................................................................122
10.10 MATERIALS OF CONSTRUCTION..............................................................................................123
10.11 CORROSION MONITORING .....................................................................................................123
10.12 CORROSION CONTROL MEASURES..........................................................................................124
October 1999
11.0 MEROX TREATING UNITS ................................................................

.....................................................................................
..................... 126
11.1
11.2
MEROX EXTRACTION...............................................................................................................126
MEROX LIQUID/LIQUID SWEETENING......................................................................................126
11.2.1
Merox Solid Bed Sweetening.........................................................................................126
11.2.2
Merox Minalk Sweetening.............................................................................................127
11.2.3
Pretreatment.................................................................................................................127
11.2.4
Post Treatment..............................................................................................................128
11.3 MATERIAL OF CONSTRUCTION................................................................................................128
11.4 TYPES OF CORROSION.............................................................................................................128
11.5 CORROSION CONTROL MEASURES..........................................................................................129
11.5.1
Corrosion Monitoring ...................................................................................................129
11.6 INSPECTION.............................................................................................................................129
12.0 SOUR WATER STRIPPING

STRIPPING UNIT ................................................................
..............................................................................
.............. 133
12.1
12.2
12.3
12.4
PROCESS ..................................................................................................................................133
CORROSION PROBLEMS ..........................................................................................................133
CORROSION CONTROLS .........................................................................................................133
MATERIALS SELECTION ............................................................................................................133
13.0 SULFUR RECOVERY UNITS................................

UNITS ................................................................
.....................................................
..................... 137
13.1
13.2
PROCESS ..................................................................................................................................137
MATERIALS ...............................................................................................................................137
14.0 SELECTION OF CORROSION

CORROSION MONITORING METHODS -REFINERY APPLICATIONS....
APPLICATIONS .... 141
14.1
14.2
14.3
ABSTRACT ................................................................................................................................141
INTRODUCTION ......................................................................................................................141
TYPES OF CORROSION MONITORING TECHNIQUES................................................................141
14.3.1
Corrosion Coupons .......................................................................................................142
14.3.4
Electrochemical Methods ..............................................................................................144
14.3.5
Linear Polarization Resistance Method ...........................................................................144
14.3.6
Other Types Of EC Monitoring......................................................................................145
14.3.7
Hydrogen Flux Monitoring ............................................................................................146
14.4 CORROSION MONITORING .....................................................................................................148
14.4.1
Chemical Injection ........................................................................................................148
14.4.2
Dewpoint......................................................................................................................149
14.5 SUMMARY OF CORROSION MONITORING CONSIDERATIONS IN REFINERY UNITS .................149
14.5.1
Atmospheric Distillation Unit.........................................................................................149
14.5.2
Vacuum Distillation Unit................................................................................................150
14.5.3
Fluid Catalytic Cracking Unit .........................................................................................150
14.5.4
Amine Treating Unit......................................................................................................150
14.5.5
Sour Water Stripping Units............................................................................................151
14.5.6
Sulfuric Acid Alkylation Unit ..........................................................................................151
14.6 AUTOMATED ON-LINE MONITORING AND DATA ANALYSIS ...................................................151
14.7 NDT and Corrosion Monitoring Techniques..............................................................................152
14.8 CONCLUSIONS ........................................................................................................................152
TABLE OF TABLES
Table 3-1
Table 3-2
Table 3-3
Table 3-4
Table 3-5
Related Standards and Regulations ........................................................................................2

Common Refining Processes ..................................................................................................4
Typical Refinery Process Selection ..........................................................................................5
FCC Unit Reactor, Regenerator & Main Fractionator Damage Mechanisms ..........................47
Inspection & Control Measures for FCCU Reactor, Regenerator & Main
Table 3-6
Table 3-7
Table 3-8
Table 3-9
Table 3-10
Table 3-11
Table 3-12
Table 3-13
Fractionator Damage Mechanisms (1) ...................................................................................57

Common Corrosion Probe Locations in Sulfuric Acid Alkylation Units.................................108
Common Stream Analyses for H2SO4 Alkylation..................................................................108
Acid Gas Absorption Reactions...........................................................................................117
Chemical Data on Selected Substances ..............................................................................118
Corrosion Reactions in Amine Systems ...............................................................................119
Activity of Sweetening Catalyst ..........................................................................................127
Matrix of Available Corrosion Monitoring Techniques ........................................................153
Application of NDT Techniques to Detect and Monitor Corrosion......................................154
TABLE OF FIGURES
Figure 3-1
Figure 3-2
Figure 3-3
Figure 3-4
Figure 3-5
Figure 3-6
Figure 3-7
Figure 3-8
Figure 3-8
Figure 3-9
Figure 3-10
Figure 3-11
Figure 3-12
Figure 3-13
Figure 3-14
Figure 3-15
Figure 3-16
Figure 3-17
Figure 3-18
Figure 3-19
Figure 3-20
Figure 3-21
Figure 3-22
October 1999
Simplified Refinery Flow Diagram.. 9

Generic Crude Unit.... 15
Generic Fluid Catalytic Cracking Unit - Process Flow Diagram. 37
Generic Fluid Catalytic Cracking Unit - Materials of Construction Diagram 39
Generic Fluid Catalytic Cracking Unit - Inspection Summary Diagram.55
Catalytic Cracking Light Ends Recovery Unit..67
Semi-Regenerative Catalytic Reforming Unit..73
Page 1 Hydrodesulfurizers and Hydrofiners..91
Page 2 Hydrocracking..93
Typical Auto-Refrigeration Alkylation Plant with Stirred Reactors ..97
Typical Effluent Refrigeration Alkylation Plant with Contactor-Type Reactor....98
Typical Caustic and Water Wash Facility.99
Typical Fractionation Facility.99
Refinery Primary Amine Systems with Multiple Absorbers.... 115
Quench Tower and Tail Gas Unit .116
Merox Mercaptan Extraction Unit 131
Liquid-Liquid Merox Sweetening Unit ..131
Conventional Fixed-Bed Merox Sweetening Unit. 132
Minalk Fixed-Bed Merox Sweetening Unit 132
Sour Water Strippers (Acidified) 135
Sour Water Strippers (Non-Acidified)135
Typical Conventional Refluxed Sour Water Stripper136
Sulfur Plant.139
October 1999
1.0
October 1999
ABSTRACT
Optimization of corrosion control in petroleum refineries, which includes metallurgical
upgrading, addition of corrosion inhibitors or water washing, and other operational or
maintenance procedures, is a function of a variety of factors. A number of these factors relate to
the basic operating objectives of the individual refinery and what processes and feed stocks are
used to achieve those objectives. Maintaining the integrity of process equipment and achieving
safe and profitable operations, requires a balance of these objectives and an understanding of
operations and corrosion interactions. Inspection, engineering, operations and maintenance
personnel all play roles in maintaining equipment integrity and implementing corrosion control
measures.
2.0
INTRODUCTION
The purpose of this overview of refinery operations is to put in perspective for both technical staff
and operations personnel how refineries operate and how these operations may influence
corrosion control. This section discusses the general aspects of refinery operations as an
introduction to the more detailed process unit reviews in the following sections. An overview of
the interaction of refinery objectives, safety, process change and various refinery personnel with
corrosion control is covered.
3.0
REFINERY OPERATING
OPERATING OBJECTIVES
Individual oil companies and refineries have individual operating objectives that are major
considerations when determining what feed stocks to run, which processes to employ and what
products to produce. However, in a broad sense, a petroleum refinery's fundamental goal is to
maximize its contribution to corporate profitability consistent with safe, environmentally
responsible operations. These goals can be achieved in many ways.
In some cases, the use of light, clean feed stocks may provide the correct product blend required
for a target market. This could minimize capital investment as sophisticated upgrading and cleanup processes would not be needed. In a market with a high demand for motor gasoline and
distillate fuels, higher-investment and more complex upgrading facilities are often justified.
Where the investment in such upgrading is made, heavy, low-quality, low-cost feed stocks are
economically attractive. These feed stocks are often corrosive and may require the use of high
alloy materials or other costly corrosion controls.
A company's marketing goals establish the extent to which they refine crude oil. A refinery
committed to the lubricating oil business requires specialized processing. Other companies may
produce and sell unfinished products which can be used by others in fuels or lubes upgrading,
the production of petrochemicals, and so forth. All such decisions are driven by economic
considerations which ultimately influence the level of investment in equipment, the process
employed, the feed stocks to be processed, the products made and the design basis of
equipment.
Page 3 - 1
3.1
October 1999
SAFETY, THE ENVIRONMENT,

ENVIRONMENT, AND RELIABILITY
Operating safety and the protection of the environment are critical concerns to all refiners and
impact how plants are designed and operated. The costs and risks associated with a safety-related
incident and the increasing legislation and industry standards surrounding protection of
employees, neighboring citizens and the environment, all have an impact on how refineries
operate. The condition of operating equipment can have major impact on the reliability and
operational integrity of a plant as they impact safety and the environment. Table 3-1 lists several
of the regulations and standards which are important in this regard.
TABLE 3-1
RELATED STANDARDS AND REGULATIONS
REGULATION/STANDARD
SUBJECT
OSHA 1920.119j
Mechanical integrity programs
API 510
Inspection, repair and re-rating of pressure vessels
API 570
Inspection and repair of piping
API 650, 651, 652, 653
Design, inspection, repair, cathodic protection of tankage
API RP 530
Design of fired heaters
NBIC
National Board code covering inspection and repair of pressure vessels
ASME Boiler and Pressure Vessel

Code
New pressure vessel design, fabrication and inspection
ASME Piping Code
New piping design, fabrication and inspection
NACE RP0170
Prevention of polythionic acid stress corrosion cracking
NACE RP0472
Prevention of cracking of CS welds
NACE RP0296
Inspection, fabrication and repair of equipment in wet H2S service
South African Standards
New vessel design, fabrication and inspection
Australian Standards
Japanese high Pressure Gas Law
Korean High pressure Gas Law
British Standard 5500
Corrosion and materials problems are important contributors to equipment condition. The role of
the inspector and the materials, corrosion, equipment or facilities engineer is to take an
understanding of these problems and develop it into monitoring, inspection and corrosion
control programs which comply with the prevailing regulatory requirements and refinery
objectives. These programs can then be used to identify problems before they affect safe
operations or the environment.
The reliability of equipment may be an important factor impacting operations even though it
may not directly impact safety or the environment. Companies or operating plants which
strongly depend upon a high and predictable service factor of equipment, low operating costs,
and/or a high level of process integration may be unable to afford an unplanned shutdown due
to an equipment materials or corrosion problem. Such operating demands may warrant the use
of higher alloys, increased monitoring, and/or the use of process additives in order to ensure
reliable operations. In other cases, however, pressures to minimize capital investment, the
Page 3 - 2
October 1999
availability of spare capacity, and/or the influence of cyclical business environment may warrant
the acceptance of somewhat lower reliability. In order to address these kinds of issues, it is
necessary to develop an understanding of the factors which influence equipment reliability.
3.2
REFINING PROCESS OVERVIEW

Petroleum refineries vary widely in complexity and processes employed in order to manufacture
the required products. Table 3-2 lists many of the various processes employed in the refining
industry. Simple refineries may produce fuels from basic crude distillation and limited upgrading
and product clean-up. Complex refineries may utilize various conversion and upgrading
processes to make larger quantities of valuable lighter fuels from relatively heavy, low-cost
streams. Other specialty processes are used to manufacture lubricating oils in some plants. Nearly
all refineries operate supporting utilities processes which provide steam, cooling water and clean
fuels for internal use, such as for firing process heaters, as well as disposal of waste waters. Feed
stocks and products are handled in what are often a complex arrangement of tankage and
piping.
Page 3 - 3
October 1999
TABLE 3-2 COMMON REFINING PROCESSES
Basic Distillation
Atmospheric Distillation
Crude Light Ends Separation
Vacuum Distillation
Light Product Upgrading

Catalytic Reforming
Alkylation
Isomerization
Desulfurization
Hydrotreating
Naphtha
Distillates
Gas Oils
Residuum
Heavy Oil Upgrading

Hydrockracking
Fluid Catalytic and Light Ends Separation
Coking and Light Ends Separation
Thermal Cracking
Visbreaking
Supporting Processes
Amine Treating
Caustic Treating
Sour Water Stripping
Sulfur Recovery
H2 Manufacture (Steam Reforming), partial oxidation
Pressure Swing Absorption
MTBE Production
Utilities
Cooling Water
Boiler Feed Water Treatment
Steam Generation
Waste Water Treatment
Flare Systems
Cogeneration Facilities
Lube Processing
Lube Extraction
Dewaxing
Deasphalting
Lube Hydrotreating
Oil Movement and Storage

Product Blending
Piping
Tanks
Pressurized Spheres
A refinery which uses only limited processes to produce basic fuels (gas, LPG, motor gasoline,
distillates) is called a 'hydroskimming refinery'. Hydroskimming refineries do not upgrade heavy
residual oils to lighter products, and they generally endeavor to limit the production of heavy fuel
oils by running lighter crude feeds which are often higher cost. Table 3-3 lists the common
refinery process found in a hydroskimming refinery.
A refinery which has more complex processes to convert heavy oil fractions to more valuable
lighter products is often referred to as a 'conversion refinery'. Conversion refineries usually run
heavy, lower quality crudes or crude blends. They not only require special heavy oil upgrading
processes, but they also generally operate additional light product upgrading processes for
increasing the octane of motor gasoline blending components. Several product cleanup facilities
are provided as the crudes they run are generally high in sulfur and other contaminants. Table 33 lists refining processes common to conversion refineries. Some conversion refineries may not
operate cokers. Others may use thermal crackers or Visbreakers in their place.
Page 3 - 4
October 1999
TABLE 3-3 TYPICAL REFINERY PROCESS SELECTION

Hydroskimming Refinery
Conversion Refinery
Vacuum Distillation
Vacuum Distillates
Hydrotreating
Hydrotreating
Naphtha
Naphtha
Distillates
Distillates
Catalytic Reforming
Gas Oils
Merox Treating
Catalytic Reforming
Amine Treating
Caustic Treating
Fluid Catalytic Cracking
Sulfur Plant
Coking
Utilities
Alkylation
Hydrocracking
Steam Reforming
Partial Oxidation
Amine Treating
Sulfur Plant
Utilities
3.3
REFINERY PROCESS FLOW

Figure 3-1 is a flow diagram showing an overview of a full conversion refinery operations. It
illustrates how processes are typically inter-linked and what products they produce. The processes
illustrated in this flow diagram are the principal ones utilized to produce the basic petroleum
products of such a refinery. Utility units are not shown. The types and order of processes used will
vary from refinery-to-refinery, and this process selection will dictate the kinds of products
obtained.
Crude oil or blends of crudes are first fractionated at elevated temperature in atmospheric
distillation units which separate them into basic products such as gas, naphtha, light distillates
(diesel and kerosene), light gas oils and atmospheric residuum. The residuum from the
atmospheric distillation unit is usually sent to a vacuum distillation unit which primarily separates
a range of heavy gas oils or lube feed stocks. As their names imply, atmospheric and vacuum
distillation take place at pressures slightly above atmospheric and at vacuum respectively.
The light streams from the crude distillation unit are separated in light ends facilities into fuel gas
(methane and ethane), LPG, butane and a variety of other hydrocarbons such as C3, C4 and C5
products. Some of these streams are used as feeds to petrochemical plants. The fuel gas is often
Page 3 - 5
October 1999
used within the refinery as fuel for the plant's fired process heaters and boilers. The fuel gas may
be used in cogeneration operations to generate steam and electric power.
The naphtha stream is often processed in a catalytic reforming unit following hydrotreating to
reduce sulfur which is a reforming catalyst poison. Reforming processes increase the octane value
of these fuels. A by-product of catalytic reforming is hydrogen which is used in hydrotreating
processes as they consume hydrogen as part of their reactions.
In refineries which have a high hydrogen demand due to sulfur removal (i.e. hydrodesulfurization) and hydro-conversion needs, hydrogen is produced in steam reforming units or,
in some cases, partial oxidation units. In hydrogen plants, a light hydrocarbon, such as methane,
natural gas or naphtha, is reacted with steam in catalyst-filled fired heater tubes to produce
hydrogen and carbon dioxide as a by-product. Gas treating or Pressure Swing Absorption (PSA)
processes are used to remove the CO2 from the hydrogen produced to provide higher purity
hydrogen for hydrotreating and hydrocracking processes. Partial oxidation units use gas, heavy
gas oils, resid and coke or coal plus oxygen to produce a synthesis gas which is water gas
reaction shifted to produce hydrogen. The hydrogen clean up is similar to the clean up facilities
on a steam methane reformer except that H2S must be handled.
Distillate fuels, which are heavier than naphtha, from both crude distillation and upgrading
processes, are usually hydrotreated to reduce their sulfur content. They are then blended to
produce kerosene, diesel and jet fuels.
The gas oil streams from atmospheric and vacuum distillation may be blended into fuel oils at
refineries which have no further processing. In some cases, these fuels will be hydrotreated to
reduce the level of sulfur present. Many refineries, however, use these gas oils as feeds to catalytic
cracking units and hydrocracking units to produce additional gasoline and middle distillate fuels.
Many refiners also charge reduced crude to the FCCs. Catalytic cracking unit feeds are
sometimes hydrotreated.
Products from catalytic cracking units are 'unsaturated' or hydrogen deficient, meaning they have
fewer hydrogen atoms per carbon atom which results in undesirable properties. The lightest
products are separated and blended into fuel gas. Gasoline-range materials may be reformed by
processing in alkylation, polymerization or isomerization units to provide octane-improving
blending components for motor fuel. Heavier products may be hydrotreated, used as
hydrocracker feeds or blended into fuel oil.
Vacuum fractionator bottoms may be blended into fuel oils or upgraded in coking processes.
Coking processes are thermal cracking processes which operate on the principal of carbon
rejection in which the coking process releases excess carbon in the form of coke. Coker feeds are
not desulfurized prior to processing. In coking, heavy hydrocarbons are cracked producing lighter
hydrocarbons and petroleum coke. This coke is essentially carbon which contains some
concentration of metals like sodium and vanadium which occur naturally in crude oil. This coke is
also a saleable product. Coker products are usually processed in fractionation and light ends
facilities similar to those used for fluid catalytic cracking unit products.
Vacuum distillation unit products may also be used as feed stocks for lube oil production. The
high value lube base stocks (raffinate) are extracted from these streams using solvent extraction.
The low value extract is often used as feed to catalytic cracking units while the raffinate is
hydrotreated, dewaxed and then blended with various additives to make a range of lubricating
oils with a variety of properties.
Page 3 - 6
October 1999
Hydroprocessing is widely used throughout refineries and covers two basic types of operations:
(a) hydrodesulfurization (and denitrification)
(b) and hydroconversion.
In both types of processes, the stream being processed and hydrogen are heated, combined and
reacted in a catalyst-filled vessel or reactor. The hydrogen and sulfur react to form hydrogen
sulfide which ends up in both the recycle hydrogen stream and the sour water streams as well as
the main reactor effuent stream. Generally, heavier oil streams are processed at higher pressures
and temperatures than lighter oil streams. Nitrogen removal is also accomplished to some extent
in heavy oil desulfurization. Hydrocracking is the most common hydroconversion process. In
hydrocracking, not only is sulfur removed, but heavier oils are also converted to lighter, higher
value products.
A number of supporting processes are part of the basic operation in most refineries. One such
process is amine treating which removes hydrogen sulfide (H2S) from fuel gas, recycle hydrogen
gas from hydrotreating units, and LPG. Sour water strippers are used to reduce the H2S and
ammonia contents of refinery process condensates so these waters can be re-used in the plant or
discarded. The H2S recovered in these units is converted to elemental sulfur in sulfur plants. This
sulfur is often sold to manufacturers of sulfuric acid or fertilizers.
Refineries also operate a number of important utilities units. Cooling water systems consisting of
one or more cooling towers, water treatment facilities and a network of piping distribute water
throughout the plant to supply some of the needed process cooling capacity. Steam generation
facilities, although sometimes integrated with the major process units, often consist of fired
boilers although co-generation facilities using gas turbines to generate steam and electricity are
becoming common. Steam is used throughout refineries to aid distillation and product stripping,
to drive pump and compressor power turbines and for other purposes. In order to provide high
quality water for the boilers, water treatment plants are usually part of the boiler plant. These
units provide demineralized, deaerated and chemically treated water necessary for high quality
steam without boiler fouling and corrosion problems. A flare system collects and burns process
releases from pressure relief systems. Waste water from the plant is treated in a series of steps to
remove environmental pollutants prior to discharge.
Page 3 - 7
October 1999
Page 3 - 8
FUEL GAS
LIGHT ENDS
RECOVERY
AMINE
TREATING
LPG
SULFUR
PLANT
H2S
CRUDE
OIL
CATALYTIC
REFORMING
HYDROTREATING
NAPHTHA
SOUR
TREAT GAS
ATMOSPHERIC
DISTILLATON
HS2O4 /HF
ALKYLATION
SOUR
TREAT GAS
KEROSENE
HYDROTREATING
MIDDLE
DISTILLATES
MOTOR
GASOLINE
AMINE
TREATING
H2
CRUDE
DESALTING
SULFUR
FCCU
FRACTIONATION
& LIGHT ENDS
HYDROCRACKING
NATURAL GAS
STEAM
HYDROGEN
PLANT/CO2
REMOVAL
KEROSENE
H2
JET FUEL
DIESEL FUEL
HEATING OIL
ATMOSPHERIC
GAS OILS
CATALYTIC
CRACKING
HYDROTREATING
LUBE
EXTRACT OILS
REDUCED
CRUDE
VACUUM
DISTILLATION
LUBE EXTRACTION
VA C U U M
GASOILS
FUEL OIL
LUBE
RAFFINATE
HYDROTREATING
LUBE BASE
STOCKS
DEWAXING
DEASPHALTING
WAX
VACUUM
RESUDUUM
ASPHALT
COKER
FRACTIONATION
& LIGHT ENDS
SIMPLIFIED REFINERY FLOW DIAGRAM

Figure 3-1
DELAYED COKING/
FLUID COKING/
FLEXICOKING
COKE
October 1999
Page 3 - 10
3.4
October 1999
PROCESS INTERACTIONS
INTERACTIONS WITH CORROSION
As Figure 3-1 illustrates, refinery processes are closely linked to each other. Such a high level of
process integration carries with it potentially far-reaching effects when feed or process changes
are considered or when corrosion control measures are being evaluated.
Changes in feed stocks can have effects in many units. A refinery can not arbitrarily elect to run
crude oil feeds higher in sulfur, for example, without considering the broad process and corrosion
design issues. From a process standpoint, higher sulfur levels may risk overloading the sulfur
removal capabilities of hydrotreating, amine treating and sulfur recovery plants. Such a change
also has far reaching corrosion implications. Corrosion can increase in the distillation units and
downstream processes like hydrotreaters, amine treating units, catalytic crackers, cokers and sour
water strippers due to the increased loads of corrosive sulfur compounds like H2S. Even a slight
increase in feed chlorides can result in increased corrosion, not only in the crude distillation area,
but also in downstream reforming and hydrotreating processes.
Changes in process operating conditions are often a source of corrosion problems. Gradual
changes are often the most troublesome since, based on small incremental step changes, the
impact can go unnoticed. For example, a gradual increase in fired heater temperatures in crude
distillation units to improve product cut points can ultimately have a dramatic influence on
corrosion rates. Increased pressures may increase the solubility of corrosive species in water or
raise hydrogen partial pressures to the point where high temperature hydrogen attack becomes a
concern in hydrotreating units. Periodic reviews of operations are warranted to ensure that the
operating limitations of materials of construction are not exceeded unknowingly.
As corrosion develops in processes, it is important to recognize that corrosion control measures
also can not be arbitrarily applied without considering the effects throughout the refinery. For
example, if not carefully applied, use of a neutralizing amine to control corrosion by acidic
chloride condensates may result in fouling and corrosion of equipment by the neutralized salt.
Filming inhibitors which successfully arrest a corrosion problem in one unit can become a serious
catalyst poison in downstream processes due to the high nitrogen content of most of this kind of
inhibitor. The use of water wash to mitigate sour water corrosion is a common practice.
However, if the water is oxygenated, corrosion can be made worse. Also, water which can not be
adequately disengaged can be carried into downstream units and cause corrosion there.
Process improvements can also have an important impact on corrosion. An improved catalyst in a
hydrotreating process can increase the amount of H2S and ammonia which affects corrosion in
both the high temperature systems and in the sour water systems. Operational changes and
catalyst improvements can also increase the denitrogenation of the feeds resulting in a drastic
increase in corrosion due to the higher ammonium bisulfide levels in the sour water system. For
example, increasing sour water corrosion in diesel oil hydrotreating unit effluent systems has
occurred as a result of process improvements and more severe operations which reduce the sulfur
and nitrogen contents of the product stream. This sulfur and nitrogen is converted to higher
ammonium bisulfide levels in the sour water system.
3.5
ROLE OF PERSONNEL
PERSONNEL IN EQUIPMENT INTEGRITY
Equipment integrity affects refinery operations in a number of ways. Reliable equipment results in
a safer working environment, prevention of environmental releases and ultimately impacts unit
Page 3 - 11
October 1999
service factor and profitability. Similarly, refinery operations affect equipment integrity. The
quality of feed stocks and the operating conditions at which they are processed influence the
measures used to control corrosion, such as chemical treatment and alloy selection, and
ultimately the deterioration rate of the equipment.
It is clear that a petroleum refinery can be a complex entity influenced by many operating
objectives. As such, the need to maintain ongoing communications among all plant personnel is
essential in support of these objectives to ensure that the impact of changing operations can be
evaluated and programs developed to address concerns.
The roles of equipment engineers such as metallurgists, inspectors and corrosion and mechanical
engineers include the following:
Understanding the factors which affect reliability and equipment degradation

Ensuring materials are selected and installed correctly
Establishing corrosion monitoring and control programs
Assessing and reporting equipment condition
Ensuring compliance with codes and standards
Plant operators and supporting process engineers also have roles in equipment integrity. These
responsibilities include the following:
Identifying operating unit operating basis and constraints

Operating within equipment design limits
Operating within agreed conditions established by materials degradation concerns
(which are often more restrictive than mechanical design limits)
Communicating changes in operating conditions, feed stocks, etc.
Carrying out necessary corrosion monitoring and control measures
Maintenance personnel also have critical responsibilities to ensure overall equipment integrity.
These responsibilities may include engineering repairs to overcome problems related to
equipment design, carrying out repairs and maintenance according to specifications and
providing information on equipment failures.
3.6
CONCLUSION
By understanding the relationships between refinery operations and the important materials and
corrosion issues in specific processes, engineers, inspectors, planners, operating and maintenance
personnel can all be important contributors to safe, reliable and profitable refinery operations.
Page 3 - 12
October 1999
4.0
CRUDE UNIT
4.1
CRUDE UNIT CORROSION

CORROSION AND CORROSION CONTROL
This section reviews fundamental corrosion issues concerning the Crude Unit process. It contains,
in concise form:
4.2
a description of the process and major equipment found in the Crude Unit;
types of corrosion and where they occur;
corrosion monitoring and inspection advice.
CRUDE UNIT PROCESS

PROCESS DESCRIPTION
In the petroleum refining process, the Crude Unit is the initial stage of distillation of the crude oil
into useable fractions, either as end products or feed to downstream units. The major pieces of
equipment found on crude units will vary depending on factors such as the assay of the design
crude, the age of the refinery and other downstream units. The unit discussed in this paper has
all of the major pieces of equipment found on crude units including double desalting, a preflash
section, an atmospheric section, a vacuum section and a stabilization section.
Cold crude from storage is transferred from tankage by the unit charge pump and is preheated in
a series of heat exchangers. It then passes through the desalters and another series of heat
exchangers. The operation of the desalter is special enough to warrant a separate section
following this crude unit process overview description.
There may be a flash drum in the middle of the desalted crude preheat, which will allow lighter
vapors and water to be removed from the crude and sent into the upper part of the crude tower.
This design helps to prevent accumulation of water hardness on the preheat tubes as a precursor
to fouling. If there is no flash drum, at 230oC to 285oC (450oF to 550oF) the crude may enter a
Preflash Column.
The Preflash Column typically has no reboiler section or bottoms stripping steam, so with no
upward moving vapors from any source other than the crude preheat, the crude will enter below
the bottom tray. In the Preflash Column, most of the light naphtha and all of the lighter
components are removed from the crude oil, yielding a 'flashed crude'. Preflashing the crude
unloads the top of the Atmospheric Column and the Crude Heater, thus increasing throughput
and reducing heater coking. A Preflash Column will often operate at a temperature low enough
that condensation of water can occur inside the tower, which often leads to corrosion of the
tower internals. Some units process a crude heavy enough that they do not have a Preflash
Column, and crude is directly fed to the Atmospheric Column following the preheat and crude
heater.
In a Preflash Column, butane and lighter fractions will go overhead. Liquids which distill
overhead are most often sent to a Debutanizer and the gases are sent to the Saturate Gas Plant or
plant fuel gas. Light naphtha may be drawn off the side of the Preflash Column, or taken
overhead, depending on the refinery configuration. The light naphtha and naphtha from the
atmospheric column overhead may be combined and sent to a Naphtha Splitter. On units which
do not include a Preflash Column or Flash Drum, it is common for the unit design to include a
Stabilizer Column which will remove pentane and lighter material from the Atmospheric Column
naphtha product.
Page 3 - 13
October 1999
The flashed crude from the bottom of the Preflash Column or Flash Drum is passed through a
series of heat exchangers and enters the Atmospheric Heater at about 260oC to 285oC (500oF to
550oF.) Leaving the heater at 345oC to 380oC (650oF to 720oF), it enters the Atmospheric
Column flash zone. Naphtha vapors off the top of the tower are condensed and the naphtha
liquid may combine with either a light or heavy naphtha and then go to the Splitter. The flow
plan, shown in Figure 3-2, shows a simplified single stage overhead system with one set of
condensers and a reflux drum. The tower top temperatures of this type of system are typically in
the range of 120oC to 140oC (250oF to 285oF). While somewhat uncommon, there are units with
the Atmospheric Column top temperature near or below the water dew point and hence water
condensation occurs in the tower. Some units have a two stage overhead system with tower top
temperatures above 140oC (285oF). Often, these two-stage systems will condense part of the
naphtha in the first stage and the remaining naphtha plus the water in the second stage. The first
stage of such a unit may have problems related to shock condensation, due to low tube-wall
temperatures, and salt deposition in the absence of a bulk water phase. The second stage of a
two-stage system has similar types of corrosion problems to a single stage overhead. To
maximize heat recovery, some units may have a more complex three- or four-stage condensation
scheme. The process and design variations described above make universal corrosion control
schemes impractical.
Kerosene is drawn off the upper part of the column, sent to a stripper and then to hydrotreating
or #2 fuel oil product storage. Diesel is drawn off the middle of the column, sent to a stripper,
and then to hydrotreating, Hydrocracker feed, or diesel or #2 fuel oil product storage.
Atmospheric gas oil is drawn off the lower portion of the column, stripped, and sent to Fluid
Catalytic Cracking (FCCU) feed or Hydrocracker feed. The atmospheric residuum from the
bottom of the column is sent to the Vacuum Heater.
The residuum enters the Vacuum Heater 10oC to 20oC (15oF to 30oF) below the temperature that
crude enters the Atmospheric Column, leaves at about 370oC to 395oC (700oF to 740oF), and is
fed to the flash zone of the Vacuum Column. The vapor off the top of the column goes to a series
of vacuum condensers which provide the necessary vacuum for column operation. Light vacuum
gas oil is drawn off the upper portion of the column and goes to FCCU feed or Hydrocracker
feed. Heavy vacuum gas oil is drawn off the lower part of the column, is combined with the light
vacuum gas oil, and sent as feed to the Cat Cracker. The column bottoms residuum is sent to a
Coking Unit or to plant fuel oil. Vacuum towers may also be used to produce feedstocks for lube
oil plants.
In the stabilization section, the Naphtha Splitter bottoms go to the Catalytic Reformer and the
overhead liquid to the Debutanizer Column. From the Debutanizer Column, the overhead liquid
goes to the Saturate Gas Plant or to plant fuel and the bottoms to the Isomerization Unit.
Page 3 - 14
GENERIC CRUDE UNIT

Figure 3-2
October 1999
Page 3 - 16
4.3
October 1999
CRUDE DESALTING
Crude oils are complex mixtures obtained from many parts of the world and all crudes contain
varying degrees of impurities. These impurities consist of naturally occurring water, salts, solids
and metals as well as added contamination from well stimulants, gathering methods, storage and
transportation. Adverse effects of these impurities are excessive corrosion, fouling and unit
upsets. These effects can result in shortened unit run lengths and reduced equipment reliability.
To minimize these effects, the refiner often washes the crude oil with water, and uses a desalting
vessel to remove the added water and most of the inorganic contaminants from the crude prior
to distillation in the crude unit.
Common desalter types and a brief description of them are given below:
Electrical desalting
an electric field is induced by AC or DC current in the

oil and water mixture to enhance water coalescence.
Chemical desalting
surfactant chemicals are used to aid water coalescence
Chemical and electrical

desalting
a hybrid of electrical and chemical methods
Gravitational separation
typically a large tank or drum which allows water and

water borne contaminants to separate due to density
difference between the water and oil phases.
The type, size and series stages of desalting facilities chosen is dictated by the individual refiner
based on refinery specific requirements and limitations.
The fundamental functions of desalters are:
1. Remove chloride salts - typically calcium, magnesium, and sodium - to minimize corrosion in
the crude unit overhead system. This corrosion is caused by hydrochloric acid which is
formed by hydrolysis of the magnesium and calcium salts during the distillation process.
2. Remove solids and sediment that cause erosion or abrasion of equipment. Deposition of
solids in the preheat exchanger train can lead to plugging of tubes or fouling which results in
reduced heat transfer and higher energy consumption.
3. Minimize unit upsets by preventing water slugs from tankage to be charged directly to the
distillation column.
A detailed description of how desalters operate is beyond the scope of this section. However, a
summary of the major variables and their expected effect on the desalter operation follows:
(i) Crude oil properties
Because desalters rely on the density difference between oil and water, lower gravity (higher
density), higher viscosity crudes make it more difficult to separate water from the crude, and
hence more difficult to desalt.
(ii) Desalting temperature and pressure
Page 3 - 17
October 1999
Generally desirable desalting temperatures are in the range of 121o to 149oC (250o to 300oF). The
upper temperature limit is to avoid vaporization of the crude oil in the desalter, or to prevent
damage to the electrical grid insulator bushings.
(iii) Residence time
Adequate residence time is essential for oil-water separation. Heavier crudes require longer
residence time because the gravity difference between the oil and water is reduced. For low
gravity crudes, the required water residence time can be 2 hours. Chemical emulsion breaker
selection may have a significant effect on oil undercarry in the water which is caused by
inadequate residence time.
(iv) Wash water quality and rate
Variables in water quality, particularly pH can affect the effectiveness of desalting and the
transport of water and ammonia into the crude or oil into the desalter brine water. Sufficient
added water must be provided to ensure good coalescence of the water in the crude.
(v) Wash water mixing
To ensure the added water is dispersed well so that it can be available to combine with the
contaminants in the crude, a controllable mixing is required. This is typically accomplished by a
mixing valve with adjustable pressure drop. Location of the wash water injection may vary,
normally into one or more places between the raw crude charge pump and the mix valve.
Injecting desalter water into the suction of a crude pump is not recommended because this
mixing can not be controlled. Over mixing can prevent adequate water coalescence.
Some of these items are discussed in more detail in the paragraphs which follow.
The source of desalter wash water is governed by the refiner's needs, environmental
requirements and availability of reusable process waters. However, the purer the water, the
easier it is to wash the crude. The volume of water can be from 3 -10% with typical usage at
5% based on total crude charge. Lowering the wash water rate below 3% of total charge
reduces the rate of coalescence and often makes water removal more difficult. A low water
rate in conjunction with high mixing energy will likely further degrade desalter performance.
For the wash to be effective in removing the impurities from the crude, it must make good
contact with the crude. Controlled mixing is achieved most often by use of a mixing valve
which permits varying degrees of water/oil contact. The higher the pressure drop (^P) across
the mix valve, the greater the mixing energy. However, if the ^P is excessive, a tight
emulsion will form which cannot be easily resolved in the desalter. Poor water separation
increases the BS&W (Basic Sediment and Water) carryover with the crude and high oil
entrainment in the effluent brine. If the ^P is too low, the crude/water contact will be
insufficient for good desalter efficiency. Typical mixing valve ^P is 10 to 20 psi. The only sure
method for determining the optimum ^P for operation is by testing and adjusting while
monitoring the desalter. Many variables dictate mixing valve ^P requirements and must be
considered before adjustments are made.
To minimize fouling in the raw crude preheat train from the deposition of salts in the crude,
it is advisable to utilize a portion of the wash water for injection immediately downstream of
the crude charge pump. This water is referred to as 'primary' water. While testing and
adjusting to find and maintain the optimum primary water rate is advisable, a good starting
point is an even split. For example, 5% (v) wash water would be split into 2.5% each for
primary and secondary locations, with an injection at the desalter mix valve as the most
common secondary location. Tighter control on mixing valve adjustments is required while in
this mode of operation.
Page 3 - 18
4.4
October 1999
Another item that may improve desalting is wash water pH control. Chemical desalters are
more efficient with high pH water, somewhere around 8.0 - 9.5. While electrical desalters
function much better in the 5.5 - 7.0 pH range. Low pH's result in excessive corrosion while
high pH permits ammonia to migrate into the crude. Excessively high pH can aid in stable
emulsion formation. Typical pH control, if required, is done with sulfuric acid or caustic into
the water as far upstream as possible, with pH controllers monitoring results. Quills for
injection of the acid or caustic are necessary to avoid mix point corrosion. Spent acids and
caustic are not advisable. They may return impurities to the crude stream that can promote
equipment fouling and corrosion.
Chemical assistance is extremely important for the desalting operation. Chemicals, when
properly applied, will not only enhance the speed of separation, but will assist residence time,
improve solids removal, minimize water carryover/oil undercarry, and reduce the emulsion
layer (cuff, or rag) to a manageable thickness. The chemicals used are termed emulsion
breakers, wetting agents, and/or demulsifiers. They can be oil soluble, water soluble, or
water/oil dispersible and in varying forms of chemistry. However, they all serve a common
goal, to enhance separation of impurities from the crude. The chemicals are surfactants
which migrate to the oil/water interface to rupture the stabilizing film around the water
droplets that allows them to merge and coalesce.
Chemical usage rates vary widely with crude type, equipment and operating parameters. It is
usually in the range of 1 pint (3v ppm) to 1 gallon (25v ppm) per thousand barrels of crude.
Several test methods are available for chemical selection on a cost/performance basis.
Chemical vendors are best equipped to assist with these evaluations, as refinery laboratories
are not normally equipped for these tests.
For the chemical to be effective, it must be well dispersed before it arrives in the desalting
vessel. The oil soluble/dispersible types are normally injected into the crude charge pump
suction or upstream of the wash water inlet point. Water soluble/dispersible chemistry
injection is preferred with the wash water, in the wash water pump suction, or before the
flow controller.
MATERIALS OF CONSTRUCTION
CONSTRUCTION
The majority of the equipment in a Crude Unit is made of carbon steel regardless of whether the
crude slate is 'sweet' or 'sour'. The term 'sour' refers to the release of H2S, but is often applied to a
crude oil based on its sulfur content, with less than 0.5 wt% sulfur called 'sweet' or greater than
1.0% called 'sour'. This use of carbon steel is possible because at temperatures below about
230oC (450oF), except for the preflash and atmospheric column overhead systems, the streams
are essentially non-corrosive to carbon steel. Where temperatures exceed 230oC (450oF),
problems with high temperature sulfur attack and naphthenic acid corrosion may begin. In the
overhead system, the formation of acidic deposits of condensates occurs below about 120oC
(250oF) and often necessitates the use of one or more highly alloyed materials.
The purpose of this section is to point out where problems occur in major equipment and
systems, and to discuss the materials commonly used to alleviate those problems.
4.4.1
Columns
In a Crude Unit, the Preflash Column will typically have Monel 400 cladding in the top zone
which operates near or below the dew point. The remainder of the shell will be bare carbon steel.
Page 3 - 19
October 1999
This Ni-Cu alloy will be corroded by sulfur compounds above about 177o to 204oC (350oF to
400oF). Since the inlet temperature is about 260oC (500oF), crude units typically have a 12%
chrome lining in the bottom to protect against sulfur corrosion.
The Atmospheric Column is commonly lined more extensively than the Preflash Column because
the feed, at about 365oC (690oF), is not only hotter, but also contains larger amounts of HCI and
H2S. The top of the column is often lined with Monel to protect against condensing HCI. Even
though the top temperature may be above the water dew point, the addition of 'cold' reflux can
cause localized condensation and conditions extremely corrosive to carbon steel. Typically, the
lower two-third to three-fourth of the column will be lined with 12% chrome cladding to protect
against high temperature sulfur corrosion.
In the area of the feed inlet, or flash zone, Type 316L stainless steel may be required in those
plants processing crudes high in naphthenic acid content.
It is important to make sure that the nozzles in each area are lined with the same material as the
shell. After exposure to service, the shell immediately above and below the cladding should be
closely monitored for wall loss. It is not uncommon to find that the original clad areas need to be
extended. For that reason most new units will have the 12Cr or 316L cladding extend up the
column to the 232o to 260oC (450o to 500oF) temperature.
In the absence of naphthenic acids, the Vacuum Column is typically lined with 12% chrome, with
the exception of the top few feet and head. The flash zone is often one of the worst naphthenic
acid problem areas. For highly naphthenic crudes, Type 316L or 317L stainless steel cladding
may be required in all areas of the column operating above 230oC (450oF). For very high TAN
crudes, higher molybdenum alloys may be required.
In sweet crude plants, the side-stream strippers are usually unlined even though the diesel and
atmospheric gas oil feeds are 285oC (550oF) and 345oC (650oF). In plants running sour crude,
these hot strippers might require a 12% chrome lining. The Naphtha Splitter and Debutanizer
are normally not lined.
While many units were built with a sweet-sour crude distinction, most modern units have to be
materialed to handle almost any kind of crude oil. This is to take advantage of opportunity
crudes which may be difficult to process from a corrosion standpoint. It is not uncommon to use
316L cladding for the major columns and the vacuum transfer line in order to be able to run
some of the market crudes that have naphthenic acid corrosion problems. There may be
considerable cost advantage to be able to run some opportunity crudes.
4.4.2
Exchangers and Piping

Heat exchanger metallurgy varies with stream composition and temperature. The majority of the
exchangers are 100% carbon steel. In fresh water cooled exchangers, admiralty brass tubes have
been used to prevent water-side fouling and corrosion. Due to the cost of the brass bundles and
improvements in cooling water treatment, many brass bundles are being replaced with carbon
steel. Where seawater or brackish water are used, admiralty, aluminum brass, cupro-nickel,
Monel, titanium, and some of the super ferritic (e.g., 26 Cr-l Mo) and duplex stainless steels have
been used successfully where carbon steel failed to perform. The use of austenitic stainless steels
has been limited in water service due to their susceptibility to chloride stress cracking and underdeposit pitting.
In hot hydrocarbon service, the use of 5% chrome materials in heat exchangers is common. As
the sulfur content in the crude increases, the use of high chrome tubes and 12% chrome shell
Page 3 - 20
October 1999
and channel linings becomes necessary. Austenitic stainless steels are also used to great
advantage in this service.
Generally, the most severe corrosion problems are in the areas of initial condensation in the
atmospheric column and preflash column overhead systems. This may include the top of the
column, the overhead vapor line, the naphtha exchangers, coolers and interconnecting piping.
As mentioned previously, these are the areas where HCl vapor, formed by the hydrolysis of the
magnesium and calcium chloride salts in the preheat, dissolves in the condensing water to form
hydrochloric acid. HCl, along with hydrogen sulfide which is also present, creates a very corrosive
environment. Usually the Monel lining and trays in the tops of the columns are effective in
resisting the acid attack unless chloride salt deposits form. The overhead vapor line, which is
typically carbon steel, can be severely attacked if unneutralized condensate is present. A good pH
control program in conjunction with corrosion inhibitors can be very effective in protecting the
bare steel line.
The chlorides in the overhead receiver water should be kept below 25 WPPM, which can usually
be accomplished with effective desalting of the crude oil and judicious use of caustic addition to
the desalted crude. This will go a long way toward solving the acidic condensate corrosion
problems. However, if the chlorides exceed about 30 PPM, the solution to the problem may be
quite difficult. It is sometimes necessary to install a Monel, or Monel clad, vapor line. The heat
exchangers closest to the point of initial condensation or chloride salt deposition may require
alloy tubes, ranging from admiralty brass to titanium. Where chloride salt fouling and corrosion
occurs, titanium exchanger tubes have worked well. The unlined carbon steel exchanger shells
may be attacked, particularly around the inlet nozzles. This may require Monel cladding or weld
overlay in this area. If the pH of the system is well controlled, as measured at the overhead
receiver, and inhibitors are properly used, the remainder of the piping and exchangers
downstream can be carbon steel with few serious problems.
The overhead vacuum condensers may have admiralty brass tubes and Type 316 stainless steel
lined shells and Type 316 SS outlet lines. The stainless steel may be needed because of CO2 and
H2S in the condensing vapors. However, carbon steel is often used successfully. 90-10 or 70-30
copper-nickel tubes are sometimes used in the vacuum condensers as part of standard vendor
steam ejector packages. In some cases, 90-10 can experience accelerated corrosion since coppernickel alloys are not highly resistant to H2S. Materials such as Admiralty or Aluminum brasses may
be considered, although the Cu-Zinc alloys are susceptible to ammonia stress corrosion cracking.
4.4.3
Fired Heaters
The fired heaters have corrosion and material problems due to the elevated temperatures
experienced both on the process side and in the fire-box. The Atmospheric Heater receives
flashed crude at about 260oC (500oF) and sends it to the Atmospheric Column at about 365oC
(690oF). For sweet crude, the radiant tubes and lower rows of convection tubes are typically 5%
chrome with carbon steel in the upper rows of the convection section. In the Vacuum Heater,
with a 360oC (680oF) inlet and 380oC (720oF) outlet the radiant tubes and convection tubes
would be 5% chrome for sweet, 9% chrome for sour crudes. Some plants running very sour
crudes have Type 316 austenitic stainless steel radiant tubes. This material would also be used
where naphthenic acid attack is severe.
High fire-box temperatures of >815oC (1500oF) also create materials problems. Tube supports
and hangers suffer excessive oxidation and premature failure if they are not sufficiently alloyed.
Historically, HH casting alloy (25% chrome - 12% nickel) was the industry standard. This material
did well in the cooler convection section, but failed in the radiant section. The substitution of HK
alloy (25% chrome - 20% nickel) added extra life in the hot areas. Higher nickel materials give
Page 3 - 21
October 1999
excellent service where low sulfur fuel is burned. However, where sulfur is high, these alloys suffer
from sulfidation. This is also true for the high nickel welding electrodes commonly used to
fabricate or repair the Cr-Ni castings. Units which burn fuel oil high in sodium and vanadium may
have refractory lined HK alloy or solid stabilized 50 Cr-50 Ni supports to resist fuel ash corrosion.
The transfer lines from the heaters to the columns are usually alloyed much the same as the
heater tubes. The vacuum heater outlet piping and transfer line may be severely attacked by
naphthenic acid, requiring the use of Type 316 stainless steel.
4.5
MATERIALS AND CORROSION

CORROSION PROBLEMS
Crude oil is a mixture of many different chemical compounds, generally combinations of carbon
and hydrogen, all with their own unique physical properties. Crude oil as such is not considered
to be corrosive to carbon steel.
However, crude oils all contain some impurities, several of which can be extremely corrosive
under crude unit operating conditions. The more common of these potentially damaging
impurities are:
4.5.1
Inorganic Salts
Sulfur Compounds
Organic Acids
Organic Chlorides
Inorganic Salts
Inorganic salts are present in brine produced with the crude oil or picked up as a contaminant
from tanker ballast. The bulk of the salts present in the water are sodium chloride (NaCl),
magnesium chloride (MgCl12) and calcium chloride (CaCl12), commonly reflecting the
composition of sea water (85%, 10%, 5%, respectively). However, these ratios can vary widely.
The total salt content by weight can vary from less than three pounds per thousand barrels of
crude oil (PTB) to 300 PTB or more.
When the crude oil is preheated, most of the MgCl12 and a small amount of the CaCl12 begin to
hydrolyze at about 120oC (250oF) and form hydrogen chloride (HCl) vapor. At 370oC (700oF),
approximately 95% of the MgCl12 and 15% of the CaCl12 have hydrolyzed. (Refer to Section 3
5.2.9.)
A similar reaction occurs for the CaCl12. The NaCl, being more temperature stable, does not
hydrolyze to any appreciable extent.
4.5.2
Sulfur Compounds
Some forms of sulfur are found in virtually all crude oils. Sulfur contents up to 6 wt% are not
unusual, but most crudes fall within the range of 0.5 - 3.0 wt%. The most important sulfur
related corrosion problems are caused by hydrogen sulfide (H2S), both below the water dew
point (aqueous) and above 260oC (500oF). While small amounts of naturally occurring H2S may
survive the journey to the crude unit in some crudes, the bulk of the H2S present in the unit is the
Page 3 - 22
October 1999
result of the thermal decomposition of reactive organic sulfur compounds which occurs in the
heaters between 260oC (500oF) and 480oC (900oF).
It is difficult to predict the corrosivity of a crude oil based entirely on its sulfur content. Generally,
the dividing line between non-corrosive and corrosive crudes lies somewhere between 0.5% and
1.0%. However, the determining factor is quite often not the amount of sulfur compounds, but
rather the extent to which these compounds thermally decompose to form H2S. This
phenomenon requires evaluating each crude individually.
At temperatures in excess of about 260oC (500oF), H2S reacts with iron to form iron sulfide scale.
The rate at which this reaction occurs is dependent on the H2S concentration, the temperature,
the stream velocity, and the composition of the material. Generally, an increase in sulfide
concentration, temperature, or velocity will increase the rate of metal loss. An increase in the
chromium content of the material will decrease the rate, with 5% chromium being a practical
minimum threshold level required for corrosion protection. Lower chrome alloys like 1-1/4Cr1/2Mo and 2-1/4Cr-1Mo do not have significantly enhanced corrosion resistance to justify their
increased cost over carbon steel. High temperature sulfur attack is a serious problem in the hot
portions of the atmospheric column, preflash column, the vacuum column, fired heater tubes,
hot heat exchangers and associated piping. The problem is alleviated by the use of proper alloy
materials.
Aqueous phase H2S corrosion is widespread in the predominantly carbon steel equipment where
water can condense. While a specific mechanism to cover all situations is not available, it is
known that three important variables in determining its severity are pH, chloride ion
concentration, and sulfide ion concentration. The types of corrosion control programs described
in the discussion on inorganic salts also apply to corrosion control in H2S containing sour water.
(Refer to Section 2 5.1.5 and 5.3.6)
4.5.3
Organic Acids
Many crude oils contain organic acids, but seldom do they constitute a serious corrosion
problem. However, a few crudes contain sufficient quantities of organic acid, generally
naphthenic acids, to cause severe problems in those parts of the crude unit operating over 230oC
(450oF). Thus, naphthenic acid attack often occurs in the same places as high temperature sulfur
attack such as heater tube outlets, transfer lines, column flash zones, and pumps. In sour crude
units a crude TAN (Total Acid Number) of 1.0 (mg KOHg) is sufficient to be concerned about
potential naphthenic acid corrosion. In sweet units, a TAN of 0.5 may be high enough to cause
corrosion. Some light southeast asian crudes with very low sulfur contents have naphthenic acid
corrosion problems at TANs as low as 0.1. Both high temperature sulfur and naphthenic acid
mechanisms are strongly affected by velocity. Whereas sulfur corrosion is characterized by a
smooth surface with a sulfide scale deposit, naphthenic acid corrosion results in sharp edged,
smooth grooves, gouges, or holes with no corrosion scale or deposit. Those materials commonly
used to prevent high temperature sulfur corrosion, primarily 5 - 12% Cr steels, can be severely
attacked by naphthenic acid. The most commonly used material is type 316 stainless steel, which
does well because of its molybdenum content. Type 304, which contains no molybdenum, has
some resistance to lower levels of naphthenic acids, but in most cases it is no better than carbon
steel. Higher TAN crudes/cuts may require higher molybdenum alloys to resist corrosion. (Refer
to Section 2 5.2.11)
Page 3 - 23
4.5.4
October 1999
Organic Chlorides
Organic chlorides constitute a contaminant in crude oil, often resulting from the carry-over of
chlorinated solvents which are used in the oilfields. They can also be picked up by the crude
during transportation in contaminated tanks or lines. Organic chlorides are not removed in the
desalters. Some of them can decompose in the heaters, forming HCl, causing erratic pH control
and accelerated corrosion in the crude unit overhead system as well as in downstream units.
(Refer to Section 2 5.2.12)
4.5.5
Corrosion Control Measures

The crude unit overhead system can benefit from corrosion control measures other than materials
selection as described earlier. Several steps can be taken to reduce the severity of acid attack in
the crude unit overhead circuit:
Blending
Desalting
Caustic addition
Overhead pH control
Use of corrosion inhibitors
Water washing
(i) Blending
Perhaps the most commonly used technique for corrosion control is the blending of problem
crudes with nonproblem crudes. Sometimes the flexibility may not exist, or blending may not
provide enough reduction of the problems, and in those cases more attention needs to be placed
on the following options.
(ii) Desalting
As the name implies, the primary purpose of a desalter is to reduce the amount of salt in the
crude oil, less than 3 ppm (1 PTB) being a commonly targeted level. Removal of the salt reduces
the amount of HCl produced from hydrolysis in the preheat and flash zone of the crude tower. In
addition to salt removal, the desalting process also removes entrained solids such as sand, salt,
rust, and paraffin wax crystals which may be present in the crude. Removal of these
contaminants helps decrease plugging and fouling in heaters and preheat exchangers.
(iii) Caustic Addition
The addition of a small amount of dilute caustic (NaOH) to the desalted crude is often an
effective way to reduce the amount of HCl released in the preheaters. The caustic converts the
HCl to thermally stable NaCl, thus reducing the amount of free HCl produced. While the results
of caustic addition can be quite beneficial, there is a risk of crude preheat train fouling,
accelerated atmospheric, vacuum, and visbreaker or coker coking, caustic stress corrosion
cracking, and catalyst contamination problems in downstream units, like FCC's, if it is not
controlled properly. A typical limit for avoiding coking problems in furnaces is to inject no more
than necessary based on downstream chloride (20 to 30 ppm in the Atmospheric Column
overhead water) or sodium limits (20-50 wppm in the Vacuum Tower bottoms).
Fresh caustic is preferred over spent caustic for two major reasons. Spent caustic tends to have
variable amounts of free or available NaOH to neutralize the HCl formed. As a result, proper
control is very difficult. Also, spent caustic, depending on its source, can be a significant
promoter of preheat exchanger fouling.
Page 3 - 24
October 1999
To minimize the negative effects of caustic injection and maximize its efficiency, thorough mixing
is necessary. To achieve good mixing, the caustic is often added to suction of the crude booster
pumps after desalting. Some refineries will mix by injecting the dilute caustic into a slipstream of
desalted crude oil prior to its injection into the main process. Injection of caustic upstream of the
desalters is not recommended because high desalter water pH can result in the formation of
emulsions and can drive ammonia into the crude. Also, the caustic will be unavailable to react
where the salt hydrolysis takes place since it will typically be removed in the desalter brine. For
units without a desalter, to minimize potential for caustic cracking, if possible caustic should be
added to the preheat train at or about desalter outlet temperature.
(iv) Overhead pH Control
The desired result of an overhead pH control program is to produce an essentially non-corrosive
environment by neutralizing the acidic components in the overhead liquid. This is done by
injecting ammonia, an organic neutralizing amine, or a combination of the two. The desired pH
control range depends on the concentrations of the various components of the corrosive
environment.
Usually, this range is 5.5 to 6.5. However, it is important to recognize that neutralizers may have
a different effect on the pH at the initial condensation point. At this point, the pH could be
higher or lower, depending on the product selected. A pH above 8 must be avoided if brass
alloys are used in the overhead system as they are vulnerable to stress corrosion cracking and
accelerated corrosion at high pH.
The preferred injection point for the neutralizer is the subject of some debate. In single overhead
drum systems, some chemical vendors advocate injecting the neutralizer into the column reflux
stream to help protect the tower internals. Others discourage this practice because neutralizerchloride salts, similar to ammonia salts, that form in the tower may be corrosive especially to
copper bearing alloys such as Monel, and may be trapped in a section of the tower. Because
stability of neutralizer-chloride salts vary depending on the type of neutralizer used, the various
options and their risks should be discussed with the chemical vendor prior to implementing a
chemical treatment program.
In two-stage overhead systems, the neutralizer or ammonia (or both) is normally injected
upstream of the second stage condensers. Generally neutralizers are not used in the first stage if it
operates without water condensation due to concerns with forming corrosive neutralizer-chloride
salts which may also be refluxed to the tower. Wet first stage systems, however, may benefit from
neutralizer addition if there is a continuous water draw from the first stage drum. Neutralizers are
sometimes used in vacuum tower overhead systems as well, using an application point that
minimizes or eliminates the possibility of introducing neutralizer-chloride salts into the tower.
A variety of neutralizers and blends of neutralizers are available for pH control. Some neutralizer
components in widespread use today include an aqueous solution of ammonia (NH3),
morpholine, ethylene diamine (EDA), monoethanolamine (MEA), and methoxypropylamine
(MOPA). All of the neutralizer salts are water soluble. MOPA and MEA form liquid neutralizer salts
with chlorides at elevated temperatures. NH3, morpholine and EDA form solid salts. Liquid salts
may be less prone to fouling, but they may also flow better and result in more widespread salt
corrosion if they are returned to the atmospheric tower.
(v) Corrosion Inhibitors
Most overhead corrosion control programs include the injection of proprietary film forming
organic inhibitors, commonly referred to as filmers. These inhibitors establish a continuously
Page 3 - 25
October 1999
replenished thin film which forms a protective barrier between acids in the system and the metal
surface underneath the film. For maximum results, proper pH control of the system is essential.
Filming inhibitor injection rates will vary with time and between refineries. There is a surface
adsorption/desorption steady state established which varies based on the aggressiveness of
corrosion in the system, and the inhibitor concentration. Factors which affect inhibitor solubility
in the liquids, such as pH, and affect the inhibitor's ability to adsorb onto the surface, such as
temperature, will affect the effective dosage for a given situation. A typical injection rate is of the
order of 3 to 5 vppm for normal operations. During startups or unit upsets, injection rates may
be temporarily increased to levels such as 12 vppm to help establish or re-establish the protective
film. Inhibitors also could have a cleaning effect so they may remove some iron sulfide deposits,
particularly at the higher injection rates.
Because these inhibitors have high molecular weights, they are non-volatile and will follow the
path of other liquids present following their injection. Therefore, they must be independently
injected into both stages of a two-stage overhead system. Filming inhibitors should normally not
be injected in concentrated form. Inhibitors are noncorrosive to equipment at treatment dosage
dilutions, but near 100% concentration they may be corrosive to injection equipment. This
should be kept in mind when designing an injection system. Typically, naphtha dilution is
provided to help the dispersion at the injection point.
In the feed to the Atmospheric and Vacuum Columns, as well as in the columns themselves,
naphthenic acid corrosion can occur. There has been some success with the use of corrosion
inhibitors purported to be effective in the 260oC (500oF) to 370oC (700oF) temperature range
and for this type of corrosion. These inhibitors may offer some economic advantage over alloy
when the acidic crudes are charged intermittently, but their effectiveness is hard to determine.
Additionally, most of the inhibitors available contain phosphorus, which may be considered to be
a poison to some hydrotreating catalysts.
(vi) Water Washing
Since the products of the above discussed neutralization reactions, ammonium chloride or amine
chloride, can be highly corrosive and also cause fouling, it is common practice to recirculate
water from the overhead receiver back into the column overhead vapor line. Stripped sour water
and/or other water condensates are also recirculated by some refineries. Water which contains
dissolved oxygen, however, can dramatically accelerate corrosion and should be avoided. Water
washing can be quite effective, but must be carefully engineered to prevent the creation of more
corrosion problems and cause significant loss of heat exchange in the overhead naphtha coolers.
Water washing the vapor line can prove to be beneficial or disastrous. Too little water can just
add to the acid-making process; too much can cause grooving of the line. The path of the
grooves can be unpredictable and difficult to locate with normal U.T. surveys. A proper spray
nozzle is necessary to prevent impingement corrosion of the pipe downstream of the injection
point. When the wash water is injected directly upstream of the condensers, a good distribution
system is necessary to ensue evenly divided flow among the different banks of exchangers. An
intermittent wash is difficult to optimize, may be neglected, and may actually increase corrosion
of otherwise dry and non-corrosive salts. Therefore, use of water on an intermittent basis should
be considered only when a continuous wash is not possible due to process constraints, or when a
continuous wash has been shown to create erosion problems.
The ideal water injection rate is 5 -10% of the overhead stream. Excessive water rates, however,
can result in poor water separation in the overhead drum. Poor separation can result in water
being returned to the tower in the reflux and resultant corrosion both in the tower and the
Page 3 - 26
October 1999
overhead line. With the proper mechanical design and chemical balance, the water wash can be
an important part of the overhead corrosion control program.
4.5.6
Corrosion Monitoring
Several methods of evaluating the effectiveness of crude unit corrosion control programs are
employed:
Water analyses - for pH, metals, chlorides, and hardness

Hydrocarbon analyses - Inhibitor residual and metals
Corrosion rate measurement by:
Electrical resistance probe
Weight loss coupon
Linear polarization resistance probe

On stream non-destructive examination by UT or RT
(i) Water Analyses (Overhead Corrosion Control)

The most important monitoring parameter for good overhead corrosion control is receiver pH.
The system pH can shift from an acceptable pH to an aggressively corrosive pH in a matter of
minutes, so the overhead receiver pH should be measured as frequently as possible on the
Atmospheric Column. The Preflash Column and Vacuum Column pH will usually not shift as
rapidly. Continuous pH monitor reliability is poor relative to most other instruments used in
refining, and so most refineries still rely on manual readings. Although pH measurements can
capture a corrosive event and prevent extended damage, even holding the pH in an acceptable
range does not always assure the lowest possible corrosion rate. See the section on Corrosion
Control Measures, Overhead pH Control for more information.
Routine analysis of the overhead receiver water for metals can be of value in some cases,
particularly when used in conjunction with other methods of measurement. Iron, copper and
zinc are typically measured, but this depends on the materials used in the overhead system. If no
brass, copper, nickel or Monel alloys are used, for example, there is little value in determining
copper, nickel or zinc concentrations. Much reliance has been put on the iron content of the
water and very often the results are misleading. Since iron solubility is quite dependent on pH,
the iron concentration in the receiver water may not be indicative of the amount of iron going
into solution somewhere upstream where the pH may be lower. The only source of copper and
zinc in a typical system would be brass or Monel exchanger bundles. If their levels increase in the
water (particularly zinc in brass containing systems), there probably is a corrosion problem. This
has been seen very dramatically in the FCCU and cokers.
Overhead receiver water chlorides are a very useful parameter to measure. Since aqueous
corrosion is almost always related to the quantity of hydrochloric acid or chloride salts, measuring
chlorides can help to confirm when a corrosion event began, and how long it was sustained. A
regular measurement of chlorides can also be used to optimize caustic addition or blending of
crudes which result in minimum corrosivity.
An often overlooked measurement which can be useful for corrosion control measurement is
hardness. The hardness of water condensing in an overhead should be zero. If any hardness is
detected, it generally will mean that a leak has occurred in a cooling water exchanger. If a
recycled water wash is in use, a cooling water leak means that oxygenated water is being
recycled. Oxygen can accelerate corrosion. Additionally, the hardness from the water can
Page 3 - 27
October 1999
precipitate when the water is injected into the overhead, causing severe fouling. If hardness is
detected, it is possible that adjustments will need to be made to the corrosion control program,
and that repairs need to be scheduled.
(ii) Hydrocarbon Analyses

For filming inhibitors used in an overhead to control aqueous corrosion, depending on the
inhibitor formulation, it is sometimes possible to run a 'residual test' on a stream to detect the
presence of the corrosion inhibitor. As mentioned earlier, there is an adsorption/desorption
steady state which is affected by the environment. There must be sufficient inhibitor present to
continuously replenish the film. This is often seen as a residual of 3 - 5 ppm. However, for many
inhibitors the nearest available test is total nitrogen which is not specific enough to quantify
inhibitor residual.
For naphthenic acid corrosion control measurement, sometimes the only tool for measuring the
aggressiveness of the environment is metals analysis of the oils. For this measurement, historical
data is very useful as a check on current conditions. The absolute value of the metals content will
change when naphthenic crudes are processed. Some of that metal comes from 'tramp' metals in
the crude oil. Some of these metal-naphthenates will distill, which can make even a relative
determination of the rate of corrosion difficult. The ratio of iron to nickel has been used with
some success as a relative measure of the effectiveness of naphthenic acid corrosion inhibitors. In
most systems, the presence of nickel is from 'tramp' sources, because the nickel alloys which are
used corrode very little. The measure of iron will include both the tramp iron and the iron from
active corrosion. If the iron/nickel ratio declines, it is then assumed to be due to inhibitor
effectiveness. What cannot be determined using this technique is the uniformity of the
protection, and localized corrosion zones remain a concern.
(iii) Corrosion Rate Measurement

Electrical resistance corrosion rate probes are widely used but with varied success. These devices
measure the change of cross section of the measuring element by measuring the change in
resistance to electric current flow. It is necessary to take a series of readings over a period of time
to establish a curve, the slope of which is indicative of the corrosion rate. This device is used with
good success in many instances. However, like all such devices, it is only indicating the corrosivity
of the measured stream at the point where the probe is located. It is not always possible to relate
the probe readings to a pipe wall or the condensing surfaces of exchanger tubes. They lend
themselves well to the evaluation of a corrosion control program which changes the environment
through pH control and inhibitor injection. They have the advantage of being read on stream.
Also, they can be designed to be retracted through a packing gland and replaced on-stream.
Electrical resistance probes are most commonly used in the tower overhead systems. They are
often used at both the inlet and outlet of overhead exchangers and may also be installed in the
bulk sour water draw-off from the overhead.
There are high temperature electrical resistance probe designs which have been used to measure
naphthenic acid corrosion. The same limitations with location exist for these probes. The biggest
impediment to their use are the serious safety issues related to inserting and extracting an
instrument at the temperatures and pressures involved.
Weight-loss coupons yield a calculated corrosion rate based on initial surface area and weight.
They lend themselves to visual examination as well as giving rate data. They have a disadvantage
in that they must be removed to give information, and they can not represent heat transfer
surfaces. They can often be replaced on stream and are used in high temperature sections of
Page 3 - 28
October 1999
crude units with the same safety concerns as for electrical resistance probes. They are commonly
used in overhead systems.
Linear polarization resistance probes give an instantaneous corrosion rate based on a
measurement of the probe element corrosion current. This type of probe will work only in a
conductive medium. It is for on-stream measurements and lends itself well to bulk water systems
like cooling water streams. Applications in the overhead receiver water drum are limited but
feasible. These probes can be made retractable through a packing gland and replaced on-stream.
NDE is normally not used for extensive routine corrosion monitoring because of its cost. It is most
often used onstream on an exception basis, when there is a confirmed or suspected problem
which is being watched closely. Since NDE is often used as an inspection tool, its use on and off
line is detailed in 4.6.
4.6
INSPECTION
Whenever possible, corrosion rate information should be verified by direct measurement of
equipment. This may not be possible on-stream for items like tube bundles, but piping and
vessels can be checked for changes in wall thickness using ultrasonics (UT) or radiography (RT).
UT readings can be taken easily and quickly on most surfaces which can be reached by the
inspector. These readings are accurate and reproducible when taken on clean, relatively smooth
surfaces. Readings can be taken on-stream at metal temperatures as high as 400oC (750oF). This
method allows the monitoring of a particular spot over a period of time, providing good data for
corrosion rate calculations at that point. In this manner, the routine monitoring of a relatively few
representative points in a large piping system may yield an accurate picture of that system. UT
readings are the most important on-stream data obtainable on a plant-wide basis. The selection
of representative UT points must consider differences in flow rates, turbulence and fouling
tendencies which affect corrosion.
Scanning UT methods are limited to lower temperatures than spot, manual UT measurements.
However, they permit making a permanent record for future comparisons. These methods are
particularly well suited to areas where localized corrosion can occur such as at high turbulence
areas in the hot or overhead systems or in areas of the overhead system vulnerable to underdeposit corrosion or impingement.
While sometimes limited by accessibility and geometry, RT is also an important on-stream
inspection tool. It can be used to measure wall thickness, indicate the presence of pitting, and
under some circumstances, show thickness of deposits on pipe walls. RT provides as permanent
visual record, unlike some of the other measuring methods.
Another useful on-stream inspection method is the use of infrared to measure temperatures of
heater tubes, vessel shells, electrical equipment, heat exchangers, insulation damage, etc. The
ability to perform reliable remote temperature measurement is extremely important both from
the standpoint of equipment reliability and economy of energy.
When the unit is shut down and equipment opened up, visual inspection can be performed. This
includes not only looking, but also checking the depths of pits with pit gauges, calipering the
O.D. and I.D. of exchanger tubes, looking for cracks using dye-penetrant or one of the several
magnetic particle methods, plus extensive use of UT and RT methods.
Page 3 - 29
October 1999
Since heat exchanger tubes, due to their geometry and arrangement in bundles, do not lend
themselves to visual inspection along their full length, several inspection tools have been devised.
The simplest is the borescope which allows visual examination along the I.D. of a tube. Since no
measurement can be made, this is of limited value.
Eddy-current equipment allows a record to be made of the ID of the tube wall, indicating cracks,
pitting and general wall thickness. Equipment based on ultrasonic principles is also used for this
purpose.
Ultrasonic wall thickness measurements should be routinely made at predetermined points on all
piping systems in the unit. UT measurements should also be taken at a number of representative
points on vessel shells and nozzles, exchanger shells and nozzles and at one or more points along
the length of each heater tube. If coking of heater tubes is a problem, radiographic techniques
can be developed to evaluate this.
While not as common as with FCC light ends, equipment in the overhead systems of the Preflash,
Atmospheric Vacuum, and light ends towers are vulnerable to wet H2S cracking. Therefore this
equipment should be included in a wet H2S inspection program.
Page 3 - 30
5.0
October 1999
FLUID CATALYTIC CRACKING UNIT

REACTOR, REGENERATOR, AND MAIN FLUID CATALYTIC CRACKING UNIT CORROSION AND
METALLURGICAL DAMAGE MECHANISMS: FRACTIONATOR
5.1
ABSTRACT
This section describes the many types of corrosion and metallurgical damage that can occur in
the hot sections of Fluid Catalytic Cracking (FCC) Units; both in service and during shutdowns.
High temperatures, corrosive liquids and gases, and erosive solids create environments in which
serious metal loss and metallurgical damage can occur. Primary areas of focus are the
riser/reactor, regenerator, flue gas system and main fractionator. An overview of the process,
general materials of construction, corrosion/metallurgical damage mechanisms, and monitoring
/inspection techniques are presented. Wet corrosion associated with the main fractionator
overhead and gas recovery section is discussed in a separate section entitled "Catalytic Cracking
Light Ends Recovery Unit Corrosion and Corrosion Control" - Section 3 6.0
5.2
INTRODUCTION
Fluid Catalytic Cracking (FCC), is a refining process in which heavy oils or residuum feedstocks of
little commercial value are broken down or cracked into lighter more useful products through the
use of elevated temperature, relatively low pressure and catalyst.
Although the exact composition of feedstocks processed in an FCC Unit varies from one refinery
to another, it usually includes straight run heavy gas oils and Coker gas oils. Since the
introduction of more advanced catalysts in the 1960s, atmospheric residuum and vacuum tower
bottoms have also been processed at an increasing rate.
Generally, operating conditions, catalyst, and hardware design are varied to maximize
production of high octane gasoline. The cracking process also yields useful quantities of
isobutane and light olefins suitable for downstream production of premium gasoline blending
components such as methyl tertiary butyl ether (MTBE) and alkylate. The evolution of the FCC,
which began in the early 1940's, was initiated by the technological developments required to
meet the demand for high-octane motor gasoline and alkylate feedstocks created by World War
II.
5.3
FLUID CATALYTIC
CATALYTIC CRACKING UNIT PROCESS DESCRIPTION
5.3.1
Cracking Reaction
The cracking reaction, the key to obtaining desired yields, is accomplished by subjecting a
vaporous feed stream or charge of heavy, long-chain hydrocarbon molecules to fluidized catalyst
at 480oC (900oF) to 540oC (1000oF) for a matter of seconds. As the large molecules break or
crack, a full range of smaller molecules are formed; including cracked paraffinic/olefinic gas,
gasoline, and light cycle oil. More densely packed, higher boiling constituents such as heavy
cycle oil, slurry recycle (or fractionator bottoms), and coke are also formed. However, because
the catalysts are "selective," they promote desirable reactions to form valuable gasoline and
Page 3 - 31
October 1999
olefins. This characteristic distinguishes the FCC process from "thermal cracking," which produces
a significantly higher yield of light gas and coke.
5.3.2
Catalyst / Catalyst Circulation

The FCC process presently relies on synthetic zeolitic catalysts, commercialized over 25 years ago.
Individual composite catalyst particles consist of a mixture of fragile crystalline aluminosilicate
materials, or zeolites, dispersed in an amorphous matrix of active alumina, silica, clays, etc. The
zeolites provide the primary cracking function. The matrix offers desirable physical properties
such as size, strength, hardness, and density; facilitates heat transfer during operation; and
promotes some degree of added cracking of the heaviest feed components.
Zeolites are effective in actively promoting cracking reactions because they have a highly porous
structure and, therefore, a tremendous amount of internal surface area. Their numerous microcavities and channels facilitate intimate contact between vaporized hydrocarbon feed molecules
and catalyst active cracking sites. It is for this reason, that zeolites are frequently referred to as
'molecular sieves'.
Prior to the introduction of zeolites in the 1960's, the active ingredient in FCC catalyst was
amorphous alumina. These early catalysts were temperature sensitive and, in comparison,
suffered from low activity and poor gasoline selectivity. Moving to zeolitic catalysts permitted
shorter cracking reaction times and eliminated the need to recycle unconverted reactor effluent
back into the feed stream to obtain desired product yields. In effect, zeolitic catalysts
revolutionized the FCC process, substantially increasing plant throughputs and markedly
influencing hardware design/modifications.
The name Fluid Catalytic Cracking Unit is derived from the manner in which the catalyst is
handled in the process. FCC Unit catalyst consists of very fine 40 to 100 micron (0.0016" to
0.0039") diameter microspheres , resembling talcum powder, which move through the plant in a
fluidized state. In fluidization, gas in the form of air, steam, or vaporized hydrocarbon, travels
through the powdered catalyst at a velocity sufficient to suspend it. The aerated solid-gas mixture
acts as a boiling, bubbling fluid that can he continuously circulated between the regenerator and
reactor, greatly simplifying the equipment necessary for catalyst handling and the control of
catalyst flow.
Fluidized catalyst exhibits a hydrostatic or head pressure just as liquids do. However, unlike a true
liquid, the density of fluidized catalyst depends on the velocity of the gas used to suspend it.
Consequently, increasing the gas velocity decreases the density and head pressure exerted by a
given height of fluidized catalyst. Catalyst transport, therefore, is controlled primarily by
differential gas pressure between the regenerator and reactor, differential catalyst-gas mixture
densities (through aeration) and slide valves.
5.3.3
Cat Cracker Hardware / Process Flow

The four principal component systems in the hot section of the FCC Unit are the riser/reactor,
regenerator, flue gas system, and main fractionator. The riser/reactor is the area in the cat cracker
where the cracking reaction takes place. The regenerator restores catalyst activity by burning
catalyst coke deposits, and provides the heat required by the endothermic cracking reaction. To
maintain the process, continuous circulation of fluidized catalyst is needed from the reactor to
the regenerator. The flue gas system is responsible for heat recovery, and purifies regenerator
waste gas for discharge to atmosphere. The main fractionator cools the cracked reactor effluent
Page 3 - 32
October 1999
gas and separates the light and heavy cycle oils from the lighter fractions (cracked gasoline,
olefins, etc.).
Figure 3-3 provides a simplified flow diagram for one possible FCC Unit configuration and is the
basis for the detailed discussion below of each principal component system.
5.3.4
Riser/Reactor
The riser/reactor is the area in the FCC where the cracking reaction takes place. Gas oil feed,
preheated from 260oC (500oF) to 425oC (800oF) using heat exchangers and/or a feed furnace, is
introduced at the bottom of the vertical riser through a single pipe inlet or multiple feed nozzles
(atomizing steam is often introduced in conjunction with the feed to disperse it evenly
throughout the riser). In the riser, intimate contact between the heated hydrocarbon charge and
hot 675oC (1250oF) to 730oC (1350oF) regenerated catalyst causes the feed to vaporize rapidly
and rise. The cracking reaction begins as soon as the vaporized hydrocarbon is adsorbed onto the
catalyst and enters the pores to contact active cracking sites. The mixture of hydrocarbon charge
vapors continues to crack as it moves up the riser, and can be assisted by lift gas (e.g. typically
steam). During the cracking reaction, carbon is deposited on the catalyst in the form of coke. As
the coke builds, it decreases the ability of the catalyst to crack hydrocarbon molecules,
deactivating it. By the time the vaporized charge reaches the reactor, the cracking process is
essentially complete and the catalyst is spent. This entire process occurs very rapidly, typically on
the order of 2 to 5 seconds.
Contemporary reactors, such as that shown in Figure 3-3, which typically operate in the 480oC
(900oF) to 540oC (1000oF) range, do little more than separate cracked hydrocarbon vapors from
the catalyst (hence, they are often much smaller than early reactor designs). Nearly all cracking
takes place in the riser. In early designs, intended for use with lower activity amorphous alumina
catalysts, very little cracking took place in the riser (the primary function of early risers was simply
to mix and vaporize the catalyst and feed). Consequently, it was necessary for cracking to
continue in the reactor. This was accomplished by maintaining a fluidized bed of fixed height in
the reactor, termed the dense phase. The cracked gas/catalyst mixture above the dense phase
was termed the dilute phase. Most commonly in bed cracking reactors, the aerated hydrocarbon
stream entered the reactor through a plate grid which supported the fluidized bed. With the
advent of zeolitic catalysts in the 1960s, riser cracking became the norm. Many fixed bed units
were modified to permit riser cracking by extending the original vertical catalyst transfer line to
increase residence time. Some units originally designed for riser cracking; however, continued to
have the flexibility of bed cracking.
Heat provided by the hot catalyst and continued contact between the catalyst and hydrocarbon
gas keeps the cracking reaction going. To prevent over-cracking (e.g. of gasoline into lighter
ends) in the dilute phase and carryover of catalyst downstream to the main fractionator,
centrifugal separators known as cyclones are required. The cyclones separate the spent catalyst
from the hydrocarbon vapors using centrifugal force. The catalyst is thrown against the cyclone
walls and drops to the bottom of the reactor through an extended pipe called a dip leg. Cracked
hydrocarbon vapors exit the top of the cyclones and are transported from the reactor to the main
fractionator by way of the reaction mix line. The cyclones can be located either inside or external
to the reactor and typically consist of two or more stages to improve separation efficiency.
Before leaving the reactor, spent catalyst passes through a stripper section, where any remaining
adsorbed hydrocarbon is separated from the catalyst using a combination of stripping steam and
baffles/shed trays. Catalyst flow from the reactor is regulated by a slide valve in the spent catalyst
transfer line. This slide valve also regulates the reactor pressure.
Page 3 - 33
5.3.5
October 1999
Regenerator
The function of the regenerator is to burn coke deposits that have accumulated on the spent
catalyst so that it can be reused. The regenerator can be located either adjacent to the reactor
(often at different elevations), as shown in Figure 3-3, making it a side-by-side FCC Unit, or it can
be mounted below the reactor (or vice versa), in which case it is called a stacked FCC Unit.
The design of the initial FCC Unit regenerators was limited by the early amorphous catalysts,
which were not capable of maintaining catalyst activity if heated above 590oC (1100oF). As
catalyst quality improved, temperatures increased. With the high 650oC (1200oF) to 760oC
(1400oF) regeneration temperatures currently employed, residual carbon on regenerated catalyst
has been lowered to less than 0.1%, while catalyst holding times have improved to 3 - 4 minutes
(as compared to 10 - 15 minutes for units which operated at 590oC (1100oF).
The regeneration process begins when spent catalyst from the reactor enters the regenerator
through the spent catalyst standpipe. It is propelled up the spent cat standpipe into the
regenerator using lift gas (e.g. steam, air, etc.). Once in the regenerator, the hot catalyst is
contacted by oxygen, supplied continuously through an air distributor (shown in Figure 3-3 to be
a perforated grid) at the bottom of the regenerator, and combustion commences. Some units
have pipe grids and others have an air ring/nozzle air distribution arrangement. Coke is
consumed in the combustion (oxidation) process, producing regenerated catalyst, flue gas (e.g.
CO, CO2, SOx, NOx, etc.) and heat. The major reaction is the coke plus air (oxygen) going to
CO and then CO2. This reaction generates considerable heat. The heat is retained by the catalyst
to sustain cracking in the reactor. The majority of combustion occurs in an area at the bottom of
the regenerator above the air distributor, where catalyst concentration is greatest (called the
dense phase). Little combustion occurs in the upper part of the regenerator (or dilute phase),
which consists primarily of flue gas and entrained catalyst.
The regenerator employs cyclone separators, similar to those previously described for reactors, to
disengage catalyst carried upward by the rising flue gas. The flue gas escapes out of the top of
the cyclones into the flue gas system, while the recovered catalyst is directed back down to the
dense phase of the regenerator through the cyclone dip legs.
Following regeneration, catalyst is directed to the regenerated catalyst standpipe and returned to
the riser to again participate in the cracking reaction.
5.3.6
Catalyst Transfer System

Catalyst transfer piping used to continuously carry fluidized catalyst from the reactor to the
regenerator and back again can have one of two basic arrangements: U-shaped lines (U-bends)
or vertical (possibly sloped) standpipes and risers. By far, the most common arrangement today
consists of standpipes and risers.
The application of the U-bend concept, used in conjunction with some fairly early fixed bed
reactors, simplified catalyst circulation by eliminating the need for valve-controlled flow.
Although valves were present to prevent potential catastrophic upsets such as catalyst flow
reversal, normal catalyst flow was achieved by changes in differential pressure between the
reactor and regenerator and by changes in the density of fluidized spent catalyst entering the
regenerator (using aeration).
Systems employing spent catalyst standpipes and regenerated catalyst risers use slide valves to
control flow. The regenerated catalyst slide valve controls the amount of hot catalyst entering the
Page 3 - 34
October 1999
reactor, and is, therefore, responsible for maintaining the appropriate reactor temperature. The
spent catalyst slide valve controls the reactor catalyst level in the stripper.
5.3.7
Flue Gas System

The flue gas system purifies regenerator waste gas for discharge to the atmosphere. This involves
cooling the gas, removing catalyst fines and removing pollutants.
Waste flue gas leaves the regenerator at 675oC (1250oF) to 760oC (1400oF) through cyclone
separators, which are responsible for removing the majority of the entrained catalyst. In most
units, the gas then passes through a steam generator or vertical shell and tube heat exchanger
known as a flue gas cooler to produce additional steam for the refinery. After cooling, fine
catalyst particles (known as fines), too small to be removed by the regenerator's cyclone
separators, are typically removed by electrostatic precipitators or wet gas scrubbers. The
precipitator imparts an electrical charge to the particles and then pulls them from the gas by way
of magnetic attraction using oppositely charged plates. Stack scrubbers are used to remove
pollutants (NOx, SOx, etc) in addition to fines. The flue gases are then either discharged to the
atmosphere or burned in a CO (carbon monoxide) boiler for further heat recovery. Many newer
units take advantage of the hot flue gas to generate power and drive the main air blower. The
power recovery turbines take the hot flue gas off the regenerator after suitable catalyst removal
and discharge it to the downstream cooldown and clean up system after extracting considerable
horsepower.
5.3.8
Main Fractionator
Cracked hydrocarbon gases exit the reactor cyclones between 510oC (950oF) to 540oC (1000oF).
They pass, without cooling, into the main fractionator column where heavy cat cracked gasoline
and light/heavy cycle oils are separated from the lighter fractions. (The fractionator functions
much like an atmospheric crude distillation column in a Crude Unit). The cat cracked gasoline
makes a good motor blending component, the light cycle oil makes a good blending stock for
No. 2 domestic heating oil or diesel fuel, and the heavy cycle oil is fed to a Coker (thermal
cracker), Hydrocracker or used as a residual fuel component.
Unlike other refinery plants, the main fractionator does not require a reboiler. Heat can be
supplied solely from the hot gas leaving the reactor. Stripping steam is often used at the
fractionator inlet to drive the hydrocarbon molecules farther apart (making it easier to
fractionate) and assist carrying the lighter gases up the tower. Most main fractionators have a
bottoms temperature in the 340oC (650oF) to 400oC (750oF) range.
The overhead stream from the fractionator, typically 95oC (200oF) to 120oC (250oF), is piped to a
gas recovery section and subjected to further fractionation, caustic treating, and H2S removal.
This yields both light and heavy gasolines, and considerable quantities of propane, butane and
light gas. A detailed process description and discussion of aqueous corrosion mechanisms that
occur in these sections is provided in a 6.0 - Catalytic Cracking Light Ends Recovery Unit.
Page 3 - 35
October 1999
Page 3 - 36
MAIN
FRACTIONATOR
Paraffinic / Olefinic and Light Cracked Gasoline
Reaction Mix Line
o
200 -250
Fo
900 -1000 F
TO FRACTIONATOR OVERHEAD SYSTEM
Flue Gas
Slide Valve
Reflux
Treated
Gas
Cracked
Gas Outlet
Heavy Cracked Gasoline
Flue Gas Line

o
o
1275 -1425 F
o
275 -300
Fo
Reflux
o
900 -1000
F
Multistage
Cyclone
Seperators
LCO
STRIPPER
o
500 -600
F
Heavy Cycle Oil

(HCO)
Reflux
Light Oil Cycle
(LCO)
o
o
375 -450 F
REACTOR
Trickle or
Flapper
Valve
Stripping
Steam
RISER
Feed
Nozzle(s)
CHARGE
TANK
Hot Feed
o
450 F
Spent Cat
Transfer Line
(StandPipe)
Normal Cat Flow
2nd Stage
1st Stage
Electrostatic
Precipitator
Dilute Phase
Fract. Bottom Pumps
650 -700 F
Plenum
o
1250 -1400
Fo
STRIPPER
Bottoms
Diplegs
Cat Level
Baffle
Plates
FLUE GAS
COOLER
REGENERATOR
Air
Grid
Expansion
Bellows
Spent
Catalyst
Slide Valve
Fract.
Bottoms
Cold
Feed
Charge
Pump
Charge
Preheat
Exchanger
Regenerated
CatSlide Valve
PREHEATER
Atomizing
Steam
GENERIC FLUID CATALYTIC CRACKING UNIT
PROCESS FLOW DIAGRAM
Figure 3-3
Lift Steam
Lift Air
Main Air
Stack
October 1999
Page 3 - 38
KEY
1
2
3
4
5
6
7
- Carbon Steel
- 1 Cr Low alloy Steel
- 5 Cr Low Alloy Steel
- 9 Cr Low Alloy Steel
- 12 Cr Stainless Steel
- 3xx Series Stainless Steel
- 4xx Stainless Steel
= Bare Alloy
= Internal Erosion Resistant
Refractory Required in Conjunction
with Metal (Hot Wall Design)
= Internal Insulating / Erosion Resistant
Refractory Required In Conjunction
With Metal (Cold Wall design)
GENERIC FLUID CATALYTIC

CRACKING UNIT
MATERIALS OF CONSTRUCTION DIAGRAM
Figure 3-4
October 1999
Page 3 - 40
5.4
October 1999
CONSTRUCTION
FCC technology and design have evolved almost continuously since its initial introduction in the
mid-1940's. Consequently, many of the over 350 FCC Units which are in operation today worldwide are essentially unique with their own set of design and materials problems. Most of these
units were designed and run primarily on distillate feed with some resid. Some new units are
designed to run primarily on resid feed and have a different design and materials problems.
There is no such thing as a standard configuration in FCC Units. Despite the differences,
however, certain basic principles provide a unifying materials philosophy. These principles, along
with common materials of construction, are presented below. Figure 3-4 provides a graphical
summary of FCC Unit materials of construction, using the simplified FCC Unit configuration
discussed previously in Section 3 - 5.3.
5.4.1
Reactors
The following damage mechanisms must be considered when designing, modifying or inspecting
a FCC Unit reactor. A detailed explanation of each damage mechanism is provided in the
Section 2 - Corrosion and Damage Mechanisms.
Materials exposed to full reactor temperatures must resist high temperature sulfidation and
carburization. Some reactors are thermally insulated with refractory linings to avoid exposure
of the shell to the process environment.
Both metals and refractory linings must be capable of resisting catalyst erosion. The severity
of erosion varies with location in the reactor.
Metals must not suffer unacceptable metallurgical changes that would lead to embrittlement,
deformation, internal fissuring or early failure.
Design should account for thermal expansion to avoid mechanical distress/cracking of
constrained components.
(i) Shells
Historically, reactors have been divided into hot wall and cold wall designs. The hot wall reactors
are typically constructed of low alloy steel, such as 1-1/4Cr-1/2Mo. This material is primarily
selected over carbon steel for its improved high temperature strength and freedom from
graphitization. However, low alloy reactors require pre and post heat treatment when welding
repairs are made, which complicates any repair considerably. Many older hot shell reactors are
carbon steel and some have been in service approaching 450,000 hours. These reactors operate
in the creep range and are candidates for remaining life studies.
Hot wall reactors may be refractory lined entirely or selectively for erosion resistance (i.e. only in
the dense bed area). Erosion-resistant linings used in hot wall reactors consist of a relatively thin,
hard, highly dense refractory (typical density is greater than 115 lb/ft3 or 1840 kg/m3). The
linings, about 1in (25 mm) in thickness, are usually phosphate-bonded and hand-packed, and are
anchored by 12 Cr stainless steel (Type 410) hexmesh that is welded directly to the shell. The
choice of 12 Cr stainless steel is dictated by thermal expansion considerations, to prevent
breaking the hexmesh-to-shell attachment welds. More recently, 'S-bar' anchors have been
introduced. These anchors can be used where it would be difficult to properly install hexmesh;
such as where the hexmesh must be forced to fit on extremely curved surfaces.
Cold wall reactors are constructed with carbon steel shells. They have to be internally insulated to
take advantage of the lower design temperature allowed by ASME Boiler and Pressure Vessel
Code. As a result, they are thinner walled than a similar materialed hot wall reactor. Castable
refractories with good insulating characteristics, however, are very soft and easily eroded. To
Page 3 - 41
October 1999
achieve a combination of good erosion resistance and insulating properties, two cold wall
refractory systems can be employed in the reactor:
dual layer (composite) linings or

single-layer (monolithic) intermediate-density castables.
In the early years, dual-layer linings were used exclusively. The dual-layer linings consisted of a 4
in (100 mm) insulating layer of soft, less than 75 lb/ft3 (1200 kg/m3 ) density refractory, against
the shell, protected from erosion by a 1 in (25 mm) thick, hard layer of high density refractory
packed into 12 Cr stainless steel hexmesh. The hexmesh was attached to the shell by metal studs
that extended the full thickness of the insulating layer. This type of lining, however, proved
expensive to install and difficult to maintain. One of the major problems was coking between or
behind the refractory.
An alternative lining system, used more frequently today, consists of a single, thick layer of
medium-weight, intermediate-density castable, supported by Type 304 stainless steel v-anchors
rather than hexmesh. Refractory densities range from 75 lb/ft3 (1200 kg/m3) to 115 lb/ft3 (1840
kg/m3). This type of lining does not provide as much insulation as the light-weight insulating
refractory nor as much resistance to erosion as the hard, high density refractory, but is generally
proving to be satisfactory. The intermediate density castables, like the low density castables, are
typically thick, on the order of 4 in (100 mm) to 5 in (125 mm). They are also hydraulically
bonded (require dry-out) and gun-applied. Stainless steel fibers are often added to the refractory
mix to hold it together should cracks develop as a result of thermal expansion.
As previously noted, reactor shells are built of either carbon steel or 1-1/4Cr-1/2Mo low alloy
steel. Although the question of which material to use is a factor of design, hot or cold wall, the 11/4Cr-1/2Mo does have some disadvantages. To achieve soft, ductile weldments bearing good
toughness. 1-1/4Cr-1/2Mo usually requires postweld heat treatment (PWHT) following
construction or repair. This involves heating the material to 730oC (1350oF) for a period of 1 to 2
hours, and can add significant cost, difficulty and time to a job.
Although the hot reactor environment is very high in hydrogen sulfide (H2S), industry experience
indicates a surprising absence of H2S corrosion on carbon steel and 1-1/4Cr-1/2Mo. The reason
for this is not entirely clear, since both high temperature and H2S concentrations would lead one
to expect relatively high corrosion rates. Case history data, presented as early as 1974, suggests
that H2S corrosion rates rise with temperatures up to about 455oC (850oF) and then decrease.
The low corrosion rates at very high temperatures in the reactor may be due to the formation of
a protective coke layer.
(ii) Internals
Reactor internals such as cyclones, grids and stripping section baffles are typically constructed of
carbon steel, which has generally proven to be satisfactory for distillate FCCUs. The new purpose
designed resid FCCs generally require higher alloy internals to combat the higher design
temperature.
Carbon steel cyclones and dip legs characteristically suffer from erosion damage and must be
internally protected by an erosion-resistant lining of hard, high density refractory supported using
12 Cr stainless steel hexmesh. Because cyclones have hot process gas on all sides, they cannot be
kept cool using insulating refractory. Consequently, the carbon steel reactor cyclones can also
suffer a slow metal loss due to carburization and to sulfidation. To prevent this, cyclones have
been upgraded to 12 Cr stainless steel in some plants. However, although 12 Cr offers added
Page 3 - 42
October 1999
resistance to carburization and to sulfidation, it may embrittle at the reactor operating

temperature, making subsequent maintenance and repair very difficult.
Reactor internals are often complicated in design, making it difficult to account for all needed
allowances to permit free thermal expansion and contraction on heating and cooling. Resulting
mechanical distress can cause cracking that can lead to rapid catalyst erosion if a pressure
differential exists across the crack.
5.4.2
Regenerator
In the early days, regenerator temperatures were low, around 620oC (1150oF), due to limitations
with amorphous catalysts. With the advent of zeolitic catalysts in the 1960's and complete
combustion some years later, regenerators began operating as high as 730oC (1350oF) to 760oC
(1400oF). Higher demands have been put on construction materials as a result. Damage
mechanisms which should be considered when designing, modifying or inspecting an FCC Unit
regenerator follow. Detailed explanations of damage mechanisms are provided in Section 2 "Corrosion and Damage Mechanisms".
Materials exposed to the full regenerator temperature and complete combustion must resist
high temperature oxidation. Materials not having sufficient oxidation resistance must be
thermally insulated with refractory linings.
Materials exposed to the full regenerator temperature and partial combustion must resist
high temperature carburization.
Both metals and refractory linings must be capable of resisting catalyst erosion. The severity
of erosion varies with location within the regenerator; typically most mild along straight
vessel walls and most severe inside cyclones.
Metals must not suffer unacceptable metallurgical changes that could result in
embrittlement, internal fissuring, or early failure.
Component design should include allowance for free thermal expansion to avoid mechanical
distress/cracking.
(i) Shell
Regenerator shells are commonly constructed of carbon steel. Internal refractory linings are used
to keep the shell cool enough to avoid an unacceptable loss of strength, prevent the occurrence
of graphitization and protect against erosion where needed. The linings also protect against
oxidation and carburization, a factor that has become increasingly important as regenerator
temperatures have risen over the years.
The basic refractory systems used in the regenerator are similar to those used in the reactor (see
5.4.1). Either a dual layer refractory lining is used, where the light, soft inner layer acts as
thermal insulator and the outer dense layer prevents erosion; or a single-layer, medium weight
refractory is used with intermediate properties. In early plants, the dual layer lining was used
exclusively. The studs which supported the 12 Cr (Type 410) stainless steel hexmesh were carbon
steel. As regenerator temperatures rose, stud failures occurred by oxidation, causing loss of the
entire refractory system. To solve this problem, stainless steel (Type 410) studs were specified in
lieu of carbon steel. However, at the high operating temperatures 12Cr has marginal oxidation
resistance.
Today, intermediate-density refractories have become popular for use on regenerator shells. They
do not insulate as well as the light insulating refractories, nor do they provide as good an erosion
barrier, but they do represent a good compromise with substantial cost savings and ease of
Page 3 - 43
October 1999
application. The typical system now used on regenerator shells is a single-layer intermediate
refractory that is applied by gunning and supported by 304 stainless steel vee studs.
(ii) Internals
Cyclones in early plants were usually carbon steel (although C-1/2Mo steel was used in a few
instances). These had reasonable life at regenerator temperatures up to about 620oC (1150oF),
but had to be upgraded as operating temperatures rose for improved strength and oxidation
resistance. The material of choice today for cyclones and the cyclone support structure is Type
304H stainless steel. Cyclones are internally protected by a 1 in (25 mm) thick erosion-resistant
refractory linings, supported by Type 304 hexmesh (though gradual refractory thickness loss by
erosion should almost always be anticipated). Recently, 'S-bar' anchors have begun to be used
instead of hexmesh, especially for repairs. These offer an advantage on curved surfaces when
bending hexmesh to fit is difficult. The top few feet of dip legs in the cyclones may also be lined
with refractory.
Many different systems have been employed to introduce catalyst and air into the regenerator.
The predominant air distribution system used to be perforated grids. However, multi-nozzle air
distributors and air rings are now common. Because grid temperatures are lower than those in
the catalyst bed above the grid, due to the cooling effect of regeneration air introduced below
the grid and the combustion of coke above it, lesser alloys can often be used for the grid than for
some of the other internals. Plants commonly use grids of 1Cr-1/2Mo to 5Cr-1/2Mo low alloy
steel. To accommodate thermal-expansion differences between the grid and shell and maintain a
pressure drop across the grid (i.e. to prevent flue gases/catalyst from leaking by), flexible
members called 'grid seals' are used. Grid seals are typically 13 Cr (Type 405) stainless steel. With
respect to air ring or multi-nozzle air distributors, metal temperatures are hot enough to require
the use of stainless steel (Type 304H). Because of the high thermal gradient between the air ring
and the regenerator atmosphere, it is usually necessary to refractory line the OD of the air ring.
In many FCC Unit designs, the spent catalyst enters through the bottom of the regenerator vessel
and passes upward through the grid by way of an internal refractory lined low alloy (i.e. 1-1/4Cr1/2Mo) or 3xx Series stainless standpipe. Because thermal expansion considerations make it
impractical to weld the spent and regenerated catalyst standpipes to the regenerator shell and
grid, expansion bellows are used where these lines passed through the grid. Expansion bellows
are typically 3xx stainless steel or nickel-base Alloy 625 (UNS N06625). Alloy 625 has more
elevated temperature strength, but can embrittle at regenerator operating temperatures.
It is often desirable to install refractory linings on the outside of the spent catalyst standpipe, in
addition to its use inside, if it is exposed to turbulent catalyst flow inside the regenerator (i.e. it
extends into the dense bed for the case of a regenerator with a grid air distribution system). In
instances where air distribution rings are employed, external refractory linings are also common.
In fact, much of the regenerator internals subject to erosion are lined using a intermediatedensity or phosphate-bonded castable. Such linings should contain metal fiber for reinforcement
and are generally 1 in (25 mm) to 2 in (50 mm) thick.
Although Type 304H is commonly used to construct regenerator internals in today's high
temperature regenerators, it can become susceptible to polythionic acid stress corrosion during
shutdowns from previous high temperature exposure during normal operation. Because
experience has shown that polythionic acid cracking failures in the regenerator are rare, however,
most refiners continue to view Type 304H as the material of choice for the regenerator
environment. The Type 304H material becomes less sensitive to polythionic acid cracking
problems with exposures of over 10,000 hours at greater than 621oC (1150oF).
Page 3 - 44
October 1999
The regenerator environment, when cooled through the dew point, is a very potent polythionic
acid stress corrosion cracker. Welded 304H or 304 that does not see sufficient time at
temperatures >621oC (1150oF) will be vulnerable to cracking while in service. Typically, this
would include the solid austenitic stainless steel slide valve bonnets and the catalyst withdrawal
lines. The regenerator environment is also a potent carburizing environment. The more CO, the
more potent the carburizing potential. The 2-1/4Cr-1Mo and the 5Cr-1/2Mo regenerator
internals used in the transition from carbon steel to austenitic stainless steel in the 1960s and
1970s had major carburization problems. With modern FCCU operation, the regenerator
environment will carburize the 304H internals currently in use.
5.4.3
Catalyst Transfer Piping System

Catalyst transfer piping is usually constructed of carbon or low alloy (1-1/4Cr-1/2Mo to 9Cr1Mo) steel with an internal refractory lining. In the early days, a dual-layer refractory lining was
used, supported by hexmesh. As refractory technology improved, the refractory of choice
became the single-layer, intermediate-density refractory, supported by vee studs. The quality of
this lining was further improved by reinforcing it with stainless steel (Type 304) needles.
Many of the early regenerated-cat slide valves had cast or wrought Type 304 stainless steel
bodies and erosion resistant refractory linings on parts exposed to flow. Steam was often used as
a purge to keep catalyst from collecting in the valve body. If the valve body was not externally
insulated to keep it hot, water condensation could form at the valve end remote from flow. The
combination of water and sulfur oxides from the process gas created an aqueous acidic condition
that often led to polythionic acid stress corrosion cracking of wrought 3xx Series stainless steel
valve bodies. Cast stainless steel slide valves generally did not polythionic crack, but were
susceptible to sigma phase embrittlement. These problems associated with stainless slide valves
can be avoided by using internally insulated and erosion-resistant refractory-lined carbon steel or
low alloy steel cold shell slide valves.
5.4.4
Reaction Mix Line, Main

Main Fractionator and Bottoms Piping
The reaction mix line, which carries cracked gas from the reactor to the main fractionator, has
been constructed from a variety of alloys over the years, depending on perceived corrosion and
metallurgical problems. Materials used in the past include internally insulated carbon steel or
non-insulated 1Cr-1/2Mo, 1-1/4Cr-1/2Mo, 5Cr -1/2Mo and 3xx Series stainless steel. Materials
selection is based primarily on the need for strength and resistance to high temperature
graphitization. Localized attack by high temperature H2S is possible at 'cool' spots where heat is
drawn away by external supports (i.e. a protective coke barrier does not form at the lower
temperatures). However, the potential for H2S attack in this environment is not a strong driver for
upgrading to more sulfidation resistant alloys. Although normally solved through design rather
than materials upgrading, thermal fatigue cracking has occurred in reaction mix lines, especially
at miters. The source of stress is the differential thermal growth between the reactor overhead
and the fractionator inlet nozzle which can place high strains on the piping each time the reactor
is cycled.
Fractionator shells are typically carbon steel, clad with 12 Cr stainless steel (Type 405, 410 or
410S) in the areas hot enough to be susceptible to sulfidation corrosion above 285oC (550oF).
Trays are typically 12 Cr (Type 405, 410 or 410S) stainless steel in the hotter areas and 12 Cr or
carbon steel further up in the column. Fractionator maximum temperatures are usually less than
370oC (700oF) to minimize coking, however the inlet nozzle can run hot enough, at 480oC
(900oF) to 540oC (1000oF), to be susceptible to high temperature graphitization. A concern has
Page 3 - 45
October 1999
been the strength of carbon steel at the feed nozzle entry to the column. Some people have used
clad low alloy steels in the feed area.
The hot 340oC (650oF) to 370oC (700oF) oil fractionator bottoms system needs to resist erosion
from catalyst slurry, as well as corrosion from high temperature H2S. Process fluids entering the
main fractionator contain catalyst fines, which often cause local erosion in the column bottoms
system. The erosion in the lower part of the main fractionator is generally minor (not a serious
problem); however, higher velocity areas in downstream bottoms piping / equipment (i.e.
pumps, ells, etc.) can be significant. Piping and valves are typically 5Cr-1/2Mo or 9Cr-1Mo for
sulfidation resistance. Downstream heat exchanger shell / channel claddings and tubes are often
12 Cr or 3xx Series stainless steel. Hardfacing alloys/vapor diffusion coatings are often needed to
resist erosion in pressure let down valves and bottoms pumps. Bottoms pump cases have ranged
from carbon steel to 12Cr. High chrome, erosion resistant irons are also used for bottoms pumps.
The bottoms pumps often suffer from erosion problems even when operated at slow speed.
Carbon steel cases have the advantage of easier weld repair.
5.4.5
Flue Gas System

With today's high temperature regenerators, flue gas exits the regenerator cyclones at 675oC
(1250oF) to 775oC (1425oF). Erosion from catalyst fines, oxidation resistance, carburization
resistance, and the need for high temperature strength are the primary concerns. In flue gas
ducts, erosion is more of a problem at elbows than in straight runs. It is particularly severe in and
just downstream of restriction orifices and the slide valve. Piping materials are typically refractory
lined carbon steel. When a power recovery turbine is used, inlet piping is typically uninsulated
3xx stainless steel to avoid refractory particles entering the turbine.
Flue gas coolers (a vertical shell and tube heat exchanger with boiler feed water shell side) have
refractory lined carbon steel in the inlet to protect against erosion and overheating. Steam
generation heat exchanger tubes are carbon steel and depend on the cooling from the boiler
feed water to keep tube metal temperatures low.
5.5
CORROSION / METALLURGICAL
METALLURGICAL DAMAGE MECHANISMS AND CONTROL MEASURES
In FCC Units, high temperatures, corrosive liquids and gases, and erosive solids create
environments in which serious metal loss can occur. High temperatures also result in
metallurgical problems not often seen in most other refinery plants. This section discusses each
corrosion/metallurgical damage mechanism in detail, provides plant equipment locations where
material problems are likely to be encountered and offers solutions to prevent future problems.
For reference, a summary of possible damage mechanisms for each FCC Unit component is
provided in Table 3-4. Table 3-5 provides control measures to prevent damage. Figure 3-5
shows the probable location of some of these mechanisms.
Page 3 - 46
October 1999
TABLE 3-4 FCC UNIT REACTOR, REGENERATOR & MAIN FRACTIONATOR DAMAGE MECHANISMS
COMPONENT
Feed Riser
Reactor Shell
Reactor Internals
Reactor Cyclones
Reaction Mix Lines (overhead piping)
Catalyst Transfer Lines
Slide Valves
Regenerator Shell
Regenerator Internals
Regenerator Cyclones
Flue Gas Lines and Coolers
Fractionator and Side Cut Piping,

Exchangers
Fractionator Bottoms Piping, Valves,
Exchangers
5.5.1
EXPECTED DAMAGE MECHANISM

Catalyst erosion of feed nozzle and riser downstream of feed nozzle
Catalyst Erosion
Refractory Damage
High Temperature Sulfidation
High Temperature Carburization
Creep
Creep Embrittlement
Catalyst Erosion
Refractory Damage
High Temperature Graphitization
885oF Embrittlement
Catalyst Erosion
Refractory Damage
Creep
Catalyst Erosion
Thermal Fatigue
Catalyst Erosion
Refractory Damage
Catalyst Erosion
Refractory Damage
Sigma Phase Embrittlement
Polythionic Acid Stress Corrosion Cracking
Catalyst Erosion
Regenerator Damage
Creep
High Temperature Oxidation (Complete Combustion)
High Temperature Carburization (Partial Combustion)
Catalyst Erosion
Refractory Damage
Catalyst Erosion
Refractory Damage
Creep
Catalyst Erosion
Refractory Damage
885oF Embrittlement
Catalyst Erosion
High Temperature Oxidation

In a FCC plant, air, at times oxygen enriched, is used in the regenerator to burn carbon (coke) off
of the catalyst. During this process, oxygen at temperatures over 540oC (1000oF) reacts with steel
and other iron-based alloys to form the iron oxide Fe3O4. Continued reaction of the iron with
Page 3 - 47
October 1999
oxygen causes the formation of a hard, uniform, black metallic looking scale that tends to be
quite smooth. When oxidation rates are low, it may be difficult to see that the scale is present.
Although higher oxidation rates result in a thicker scale, it can still be difficult to tell that
significant attack has occurred, unless the scale begins to crack and spall off as a result of thermal
cycling or because of the internal stresses that build up in the scale as it forms (the scale is
approximately 5 times more voluminous than the metal consumed). An energetic beating with a
hammer to remove the oxide layer, followed by ultrasonic inspection reveals the thickness of the
scale and the extent of the corrosion (but also exposes any fresh, non-scaled metal surfaces to
accelerated oxidation rates on plant start-up).
High temperature oxidation presents a potential FCC plant problem only in the regenerator and
its flue gas system. As FCC technology has improved, oxygen concentrations and temperatures in
FCC regenerators have increased, resulting in more severe service conditions. Protection against
high temperature oxidation is achieved by either insulating metal surfaces to keep them cool or
by using alloys with increasing chromium content. The shell of an FCC regenerator, for example,
is kept cool and free of oxidation by using an insulating refractory. Internal parts of the
regenerator that cannot be insulated are protected through the use of alloys. Although lowchrome steels, like 1-1/4Cr-1/2Mo, are not measurably better in oxidation resistance than carbon
steel, 5Cr-1/2Mo oxidizes at somewhat reduced rates and 12 Cr is even better. However, for
parts operating at full regenerator temperatures, austenitic (3xx Series) stainless steels such as
Type 304 /304H which contain approximately 18 wt% chromium are needed. Cyclones and the
hexmesh S-bars supporting refractory are typical examples of Type 304 stainless steel use in the
regenerator.
5.5.2

Hydrogen sulfide (H2S) is formed in the FCC preheater and reactor by thermal decomposition of
organic sulfur compounds in the plant feed. It is corrosive to steel at a temperature above
approximately 285oC (550oF) in concentrations greater than 1 ppm. The high temperature
corrosion reaction of iron with H2S, known as 'sulfidation', produces a uniform, and sometimes
tenacious, black iron sulfide (FeS) scale in the FCC Unit. High temperature sulfidation occurs
mainly in the preheater, feed piping downstream of the preheater, reactor, reaction mix line
(transfer line from the reactor to main fractionator), main fractionator, and fractionator bottoms
/side cut piping and exchangers.
Sulfidation corrosion is a function of temperature and H2S concentration. Changing to a high
sulfur feed, for example, almost inevitably leads to higher corrosion rates. Raising temperature,
however, does not always result in increased sulfidation. Industry experience indicates that
sulfidation rates increase up to about 425oC (800oF) to 455oC (850oF) and then decrease with
further increases in temperature. Sulfidation corrosion in FCC reactors occurs at even lower rates
than one would normally expect using general industry H2S corrosion data, allowing many plants
to use carbon steel cyclones in FCC reactors operating as high as 480oC (900oF) to 510oC
(950oF). It may be that the sulfide scale layer becomes plugged with coke, which acts to retard
sulfide diffusion through the scale.
Resistance to high temperature sulfidation in iron-base alloys is achieved by chromium additions.
Of the iron-based alloys, 5Cr-1/2Mo is the least alloyed to have significantly better resistance
than carbon steel. For example, hot side cut/bottoms piping and heat exchanger tubing
downstream of the main fractionator are 5Cr-1/2Mo steel. However, carbon steel reactor
cyclones and 1-1/4Cr-1/2Mo reactor effluent lines/hot wall reactor shells have demonstrated
acceptable sulfidation rates. The 12 Cr steels (Types 405 or 410) are corroded very slowly by H2S
under any conditions likely to be encountered in an FCC Unit. Some clad reactors still have
considerable clad left after 250,000 hours service. The clad loss seems due to sulfidation and
erosion. Some plants have upgraded reactor cyclones to 12 Cr stainless steel. The main
Page 3 - 48
October 1999
fractionator 'hot zone' cladding and internals are often 12 Cr (Types 405 or 410) to resist
sulfidation. The 3xx Series stainless steels (Type 304, 321, 347) are also totally corrosion-resistant
to high temperature H2S and are widely used in preference to 12 Cr for cyclones because of
superior room temperature mechanical properties following periods of extended high
temperature exposure.
As with high temperature oxidation, in the absence of alloy, sulfidation resistance can be
achieved by insulating metal surfaces internally to keep them cool. Cold wall reactors, for
example, are internally insulated carbon steel.
High temperature sulfidation corrosion does not occur rapidly enough in FCC Units to create the
probability of a catastrophic failure. Attack is easily found during turnarounds by UT because
rates are nominal and attack is generally quite uniform (although highly turbulent areas can
aggravate corrosion).However, the combination of sulfidation and erosion in the slurry lines has
caused some unanticipated failures. (Refer to Section 2 5.1.5)
5.5.3

Carburization in FCC Units involves a series of steps. Initially, carbon (coke) is deposited on the
metal surface. The carbon then reacts with the metal to form metal carbides. Diffusion (interatomic movement) of carbon and/or metal atoms within the material control the extent to which
the reaction product (metal carbide) penetrates the metal. As the metal carbide layer forms, it
experiences a high compressive stress, since it occupies a greater volume than the unaffected
metal. The metal carbide may bulge away from the unaffected substrate or simply flake off,
gradually reducing the metal thickness. Coke formation followed by carburization appears to be
the mechanism on the reactor and fractionator side of the reaction section. The regenerator
environment has enough CO present to be carburizing to the low Cr-Mo steels and the austenitic
stainless steels. Both 2 1/4Cr-1Mo and 304H will carburize under partial combustion operation.
This has become more of a problem since resid is being fed to the distillate units and resid
crackers have come onstream.
In FCC units, carburization used to be classified as a phenomenon of minor concern and was
largely ignored. For example, the slow thinning of carbon steel reactor cyclones was attributed to
carburization, but material upgrades were generally not made for this reason. However, in recent
years, a number of problems have been reported in regenerators operating at higher
temperatures with partial combustion (excess CO). Under such circumstances, CO in the
regenerator and flue gas system has carburized even 3xx Series stainless steel.
As a general rule, chromium additions seem to retard carburization in oxidizing or sulfiding
(where H2S is present) environments, but not in reducing environments. However, 1-1/4Cr1/2Mo reactor shells have not been found to carburize significantly.
5.5.4
Polythionic Acid Stress Corrosion Cracking (PASCC)

Alkthough the modified 3xx Series do offer improved resistance to sensitization, they will
eventually sensitize to some degree at following long term operation at regeneration
temperatures. Although the potential exists for PASCC, standard Type 304 or 304H has been
used successfully in the vast majority of plants for many years without cracking. Some laboratory
studies, backed up by plant experience, indicate that service at >621oC (1150oF) for at least
10,000 hours tends to diffuse Cr back into the grain boundary areas and will minimize cracking
problems. The experience with slide valve bonnet insulation seems to verify the same reasoning.
Page 3 - 49
October 1999
The major area of concern for PASCC is the regenerator system where Type 304/304H has
become widespread since the introduction of high temperature regeneration in the 1960s.
(Reactors generally do not contain much 3xx Series stainless steel). Although cracking of
regenerator refractory anchors/hexmesh and cyclones has occurred, most reported PASCC cases
have been in equipment external to the regenerator, such as in 3xx Series catalyst withdrawal
nozzles, wrought slide valves, flue gas lines and expansion bellows.
Most PASCC problems that have occurred have been solved by means other than alloying. For
example, elimination of wet steam purges (or purging with nitrogen rather than steam) to
prevent catalyst accumulation in slide valves, using internally insulated carbon steel slide valves
rather than cast or wrought 3xx Series stainless steel, specifying packed and insulated expansion
joints, and eliminating water washing during shutdowns to keep dust levels down have all proved
successful. If water washing cannot be eliminated, sodium carbonate (soda ash) can be added to
wash water to neutralize the polythionic acids as they form. 'NACE International Recommended
Practice RP0170-93' provides guidelines for soda ash washing equipment. The addition of
ammonia to susceptible areas during shutdown has minimized PASCC in some instances. The
low carbon grades have a strength problem that usually prevents their use around the
regenerators. (Refer to Section 2 5.3.4)
5.6
CATALYST EROSION
Erosion, the largest single problem in the hot dry sections of FCC Units, is defined as the loss of
material due to the impact and cutting action of solid particles. Unlike corrosion, which is the loss
of metal from chemical reaction, erosion does not involve any reaction of the metal or refractory
with the environment. It is a mechanical phenomenon. Eroded surfaces typically appear bright
(absence of corrosion product) with sharp edges and distinctive flow patterns.
The rate of catalyst erosion is influenced by the properties of the material surface being eroded.
With respect to metals, there is little practical difference in erosion resistance between carbon
steel, alloy steels, Type 410 stainless steel or 3xx Series stainless steel. Attempts to increase
surface hardness through heat treatment or cold work have not increased erosion resistance
significantly. Hard facing materials, deposited by welding or vapor diffusion, can provide
improved erosion resistance. However, if the hard facing alloy consists of hard particles (i.e.
tungsten carbide) in a soft matrix, the fine FCC catalyst eventually erodes the matrix material
between the hard particles allowing them to drop out. Concerning refractory materials, erosion
resistance generally increases with increasing refractory density, though compressive strength,
binder type/amount, and aggregate hardness/distribution are also important parameters.
The rate of catalyst erosion is also influenced by catalyst particle size, hardness, velocity,
loading/concentration and impingement angle. Particle size/hardness is normally outside an
Inspector's and Corrosion/Materials Engineer's control. In theory, larger particles are more erosive
than smaller ones because they strike the metal or refractory surface with more energy. In
practice, the size of catalyst particles does not fluctuate widely. In theory, a harder catalyst causes
more erosion than a soft one. In practice, refiners do not typically optimize hardness to prevent
erosion when purchasing catalyst. A commonly accepted threshold below which experience
indicates little or no erosion occurs is 50 ft/s (15 m/s) to 60 ft/s (18 m/s). Once this threshold is
exceeded, metal loss can increase dramatically. Erosion is a function of dilute phase catalyst
concentration (loading). The more particles that strike a surface per unit time, the more metal
that will be removed. It is interesting to note that fluidized catalyst, which differs from a liquid in
that it does not exert a constant density across its cross-sectional area, exhibits increased erosion
(and particle concentration) toward the outside of a bend when making a turn. Most FCC Unit
equipment is eroded most quickly when catalyst particles strike the surface at a 45o angle (in
Page 3 - 50
October 1999
contrast, brittle materials would experience their worst erosion damage at a 90o impingement
angle).
Erosion is a potential problem in the reactor and regenerator vessels, particularly in:
cyclones, cyclone dip leg flapper or trickle valves, and grid holes,
catalyst and flue gas piping, primarily at ells, in slide valves, and surrounding feed/steam
injection inlets,
and in the main fractionator bottoms system (pumps, piping, heat exchanger tubes).
Protection, outside the fractionator bottoms system, is generally afforded using erosion-resistant
refractories. Hard facings provide protection as needed in the main fractionator bottoms where
carryover of catalyst fines causes most problems. They are also used on a limited basis to protect
selected parts of the catalyst slide valves. Experience suggests that erosion is most prevalent at
changes in direction (i.e. elbows), locations of pressure drop (i.e. control valves), or locations of
extreme turbulence/velocity (i.e. pumps, thermowells, cyclones, feed/steam injection inlets in cat
transfer lines). Mild to moderate erosion is generally experienced on vessel wall or pipe straights.
Small cracks or gaps in reactor or regenerator internals warrant special concern, as differential
pressure can propel catalyst through them at high velocity. This situation has caused many
serious erosion incidents, including through wall vessel leaks.
5.6.1

Graphitization requires exposure from 440oC (800oF) to 700oC (1300oF) for long periods and is
most rapid around 565oC (1050oF). To minimize the risk of graphitization, carbon steels are
normally limited to a maximum temperature of 425oC (800oF) . These temperature limits are
normally applied to pressure-containing parts. For vessel internals, graphitization is often risked,
because cracking will not result in a leak to atmosphere, and plate steels used to fabricate much
of the carbon steel internals (such as reactor cyclones, etc.) are typically not as susceptible to
chain graphitization as pipe steels. This is due to differences in deoxidization practice during steel
making.
Besides reactor cyclones, graphitization is possible in the main fractionator inlet nozzle, shell
immediately adjacent to the inlet nozzle, and any location where the internal thermal insulation
for carbon steel is damaged. Chromium contents of 1/2% or greater provide immunity.
Consequently, chrome-moly steels such as 1-1/4Cr-1/2Mo, are sometimes used for construction
of FCC Unit reactor shells, do not graphitize. [Refer to Section 2 5.5.1 (ii)]
5.6.2

Sigma is a brittle intermetallic compound or phase that forms in stainless steels over long periods
of time at temperatures above 590oC (1100oF), which are commonly experienced by the
regenerator. It results in brittleness and loss of impact strength, most apparent when the plant is
down and cool. The brittleness due to sigma phase formation tends to disappear when the metal
is heated above approximately 260oC (500oF) and to reappear on cooling below this temperature
depending on the amount of sigma formed. Consequently, embrittlement is not likely to cause
an on-stream failure, but care should be taken when carrying out maintenance work to minimize
shock loadings that might trigger a brittle fracture. Significant sigma phase, as can be
experienced in high alloy castings, may seriously affect the high temperature properties of a
piece of equipment or a component.
Sigma, itself, is an iron-chromium intermetallic compound that can form in stainless steels
containing more than approximately 17% chromium. Consequently, sigma embrittlement is
Page 3 - 51
October 1999
primarily associated in FCC Units with 3xx Series (austenitic) stainless steels, which nominally
contain about 18% chromium. It forms much more readily from a ferritic structure than an
austenitic structure. ('Ferritic' and 'austenitic' refer to the arrangement or packing of iron atoms in
a steel, with those in an 'austenitic' structure more closely spaced). As such, most problems are
associated with austenitic castings, such as hot wall slide valve bodies or weldments, which
contain about 3 to 10% ferrite to avoid hot cracking during solidification. It is for this reason that
cast stainless steel slide valves should not be insulated [heat loss keeps the metal below the 590oC
(1100oF) threshold for sigma embrittlement]. However, not insulating the austenitic slide valves
may cause operational problems such a slide sticking due to expansion differences with normal
valve clearances.
Certain chemical elements tend to promote sigma embrittlement. For example, Types 316 and
347, which contain molybdenum and columbium, respectively, embrittle more rapidly than Type
304. However, Type 304 is not immune. Sigma embrittlement failures of Type 304 components
in FCC regenerators have been reported.
Sigma embrittlement can be controlled by limiting the ferrite content of weld metal or castings
to 10%, avoiding shock loads when the metal is cold, and use of Type 304 stainless steel in the
regenerator system in preference to other austenitic stainless steel grades (Type 321, 347). Type
304H will form up to about 10% sigma in the base metal away from the welds. The combination
of sigma and carbide precipitation can result in a signification loss of room temperature ductility
when compared to annealed Type 304H. This loss of ductility means that more care must be
taken during weld repair. Unless there is significant carburization, there is usually enough room
temperature ductility to successfully weld repair.
5.6.3
885oF Embrittlement
4xx Series stainless steels (Type 405 or 410), also known as the family of ferritic and martensitic
stainless steels, are subject to embrittlement in the 370oC (700oF) to 540oC (1000oF) range. It is
called 885oF embrittlement because it occurs most rapidly at this temperature. Like sigma phase,
885oF embrittlement is also caused by the precipitation of an embrittling constituent in the steel.
This phenomenon prevents the use of 4xx series stainless steel for pressure containing
components in the hotter locations of FCC units.
The higher the chromium content, the more susceptible the alloy. For example, Type 410 (12%
Cr) is marginally susceptible, Type 405 (13% Cr) is slightly more susceptible, and Type 430 (17%
Cr) is very susceptible.
Although most 3xx Series stainless steels do not suffer from 885oF embrittlement, the ferrite
structure in 3xx Series weld metal and castings, necessary to avoid hot cracking during
solidification, is subject to 885oF embrittlement.
5.6.4
Creep Embrittlement
Embrittlement
Creep embrittlement is a marked lowering of creep rupture ductility in the weld heat affected
zone of highly stressed C-1/2Mo, 1 Cr-1/2Mo, and 1-1/4Cr-1/2Mo steel components (nozzles,
etc.). This means that a material can fail by creep [refer to Section 2 5.7.2 (i)] without the
characteristic 'stretching' that normally occurs. During high temperature operation above about
455oC (850oF), the coarse grained heat affected zone has a tendency to crack at the weld fusion
line. Although there is no metallurgical effect at room temperature, the fracture toughness at
room temperature would be lowered proportionally to the size of the cracks that occurred at
high temperature.
Page 3 - 52
October 1999
Although the potential for creep embrittlement exists, it has not yet become an issue in 1-1/4Cr1/2Mo FCC components, presumably because they are not as highly stressed as low alloy steel
components in other refinery plants. In other plants where creep embrittlement has been seen,
use of higher purity steels or an alloy upgrade to modern high purity 2-1/4Cr-1Mo have been
acceptable solutions.
5.6.5
High Temperature Creep

Much equipment in a FCC Unit operates in the creep range: hot wall reactor vessels (1-1/4Cr1/2Mo), carbon steel reactor cyclones, stainless steel regenerator cyclones, feed preheater tubes,
etc. Although internally insulated carbon steel reactors and regenerator shells do not operate in
the creep range, creep failures can occur if the insulating refractory fails. FCC Unit creep failures
can be prevented by proper equipment operation, surveillance (i.e. visual inspection for bulging
and temperature surveying using infrared thermography), temperature reduction using thermal
insulation, or alloy upgrades (creep limits generally increase with chromium content).
5.6.6
Thermal Fatigue
In FCC Units, fatigue cracking is most prevalent in reaction mix lines, especially at miters. The
source of the fatigue stress is the differential thermal growth between the reactor overhead and
the fractionator inlet nozzle. A high stress is placed on the mix line each time the reactor
temperature is cycled. After sufficient cycles, fatigue cracks initiate and grow.
5.7
INSPECTION/MONITORING
INSPECTION/MONITORING METHODS
The Inspection Summary Diagram provided in Figure 3-5 was developed to illustrate where FCC
Unit corrosion and metallurgical damage mechanisms are likely to occur. Table 3-5 describes the
inspection methods most successful in discovering problems. Used in conjunction, these tools
provide the necessary guidance to assist in the development of more detailed inspection plans for
the hot sections of the FCC Unit and offer a good quick overview of the material covered in this
section.
The majority of the inspections performed in the hot section of the FCC Unit occur during
planned turnarounds. Of the shutdown inspections presented in Table 3-5, it is vital to assess
refractory damage as near to the start of the shutdown as possible. Extensive repairs, which
require curing and dry-out, can easily become critical path if not properly scheduled. It is also
very important, when documenting refractory loss, to provide detailed descriptions of erosive
patterns so that proper repairs, upgrades or design changes can be made.
Between planned shutdowns, inspection efforts should focus particularly on the main fractionator
bottoms piping system. Erosion and corrosion are common in bare alloy steel slurry systems due
to the presence of catalyst fines and H2S. Inspection of this piping system for thickness on a
regular planned frequency is a proactive way to prevent problems.
Routine monitoring of critical process variables, primarily temperature, is also very important.
Most corrosion and metallurgical damage mechanisms experienced by the FCC reactor,
regenerator, main fractionator, as well as their associated piping, are highly temperature
dependent. In most cases, damage worsens with temperature increases. Routine surveillance can
identify when equipment is operating outside of normal parameters and prevent failures.
Page 3 - 53
October 1999
Examples include monitoring critical board thermocouples (i.e. temperatures in vicinity of

regenerator cyclones) in addition to regularly scheduled furnace tube and cold wall vessel/piping
system infrared thermography.
Page 3 - 54
KEY
1 - Catalyst Erosion
2 - Refractory Damage
3 - High Temp Oxidation
4 - High Temp Sulfidation
5 - High Temp Carburization
6 - High Temp Creep
7 - High Temp Graphitization
8 - Sigma Phase Embrittlement
o
9 - 885 F Embrittlement
10 - Creep Embrittlement
11 - Thermal Fatigue
12 - Polythionic Acid Stress Corrosion Cracking
= General / Applies to All Materials

= Specific to 3xx Series Stainless Steel
= Specific to Carbon / Low Alloy Steel
= Specific to 4xx Series Stainless Steel
GENERIC FLUID CATALYTIC

CRACKING UNIT
INSPECTION SUMMARY DIAGRAM
Figure 3-5

1999
October
Page 3 - 56
October 1999
TABLE 3-5 INSPECTION & CONTROL MEASURES FOR FCCU REACTOR, REGENERATOR
& MAIN FRACTIONATOR DAMAGE MECHANISMS (1)
DAMAGE MECHANISM
INSPECTION (2)
LOCATION
CONTROL
MEASURE
High Temperature Oxidation
Regenerator internals and flue gas system (e.g.

where metal temperatures exceed 1000oF /
540oC)
Visual (use hammer test to remove oxide scales

and revel damage). UT to determine
remaining wall thickness.
Use a resistant alloy containing sufficient

chromium (resistance improves from 5 Cr, 9 Cr
to SS). Insulate the metal surfaces to keep
them cool
Preheater furnace tubes, feed piping, reactor

internals, reaction mix line, sections of main
refractionator above 550oF (285oC),
fractionator bottoms piping and pump,
fractionator side cut piping and exchangers
which experience a metal temperature >550oF
(285oC)
Attack is quite easily found by UT or RT

because rates are generally predictable and
attack is quite uniform. Pay particular
attention to hot areas of fractionator just
beyond the 12Cr cladding.
Use a base metal or cladding/weld overlay

with sufficient chromium (resistance improves
from 5 Cr, 9 Cr to SS) to resist attack. Insulate
the metal surfaces internally with refractory to
keep them cool
Reactor and re-generator internals (with

incomplete combustion, CO can exist above
the dense phase in the regenerator)
UT to identify wall thinning.
Polythionic Acid Stress Corrosion

Cracking
Regenerator internals, slide valves, refractory

anchors, catalyst withdrawal lines, flue gas
lines and expansion bellows constructed of 3xx
Series stainless steel,
Cracking can occur online. Not normally part

of routine inspection program. If detected
visually, inspect other similar weld/base metal
locations using PT.
Take precautions during shutdowns to prevent

polythionic acid formation. Prevent water from
condensing on 3xx Series stainless steel that
exceeds 700oF (370oC) in service. Avoid water
washing dust removal; use packed and
insulated expansion joints; change to internally
insulated carbon steel slide valves rather than
stainless steel (or purge with nitrogen rather
than steam). Use low carbon or stabilized
varieties of 3xx Series stainless steel.
Catalyst Erosion
Reactor and regenerator shell and internals

(especially cyclone separators); catalyst transfer
lines; thermowells; slide valves; flue gas lines
and coolers; and fractionator bottom pumps,
heat exchangers, valves and piping
Visual for majority of equipment and internals.

UT and RT thickness measurement for piping,
tees, elbows, valves, reducers, pump,
discharges, etc. Focus first on high velocity
areas > 50 ft/s (15 m/s). Damage can be
highly localized.
Design to minimize turbulence of catalyst and

catalyst carryover. Use erosion resistant
refractory linings and hardfacing. Use SS
ferrules in inlet of flue gas coolers or
fractionator bottom exchangers.
Feed Nozzle Erosion
Riser pipe downstream of the regenerated

catalyst entry point and feed spray nozzles.
Visual or RT.
Design to minimize turbulence on the riser

wall. Use erosion resistant materials to extend
life of feed spray nozzles.
Refractory Damage
Reactor and regenerator systems, internals and

associated piping (e.g. thermal cycling cracks;
loss of anchors; spalling from poor installation;
Visual during shutdowns or survey cold wall

equipment onstream using thermography
(e.g. pyrometers or infrared analyzers) to
Proper refractory selection, application, dryout/curing, reinforcement (e.g. metal fibers)

and anchoring.
Page 3 - 57
DAMAGE MECHANISM
October 1999
INSPECTION (2)
LOCATION
CONTROL
MEASURE
insufficient dry out; coking)
identify failure of insulating refractory
High Temperature
Graphitization
CS reactor cyclones; fractionator inlet nozzle

and adjacent shell; and any location where the
thermal insulation is damaged (e.g. reactor
and regenerator internals, catalyst transfer
lines) so that metal temperatures exceed
800oF/425oC (if carbon steel) and 850oF/455oC
(if carbon-molybdenum steel)
RT, shear wave UT and field metallography of

weldments
Use chrome-molybdenum steels rather than

carbon steel or carbon-molybdenum steels for
pressure containing components. Insulate the
metal surface with refractory to lower metal
temperatures.
Welded 3xx Series stainless steel regenerator

internals or flue gas system components and
cast 3xx Series stainless steel slide valves
exposed to temperatures between 1100oF
(590oC) to 1700oF (925oC).
PT for cracks of field metallography to identify

presence and distribution of sigma phase.
Control ferrite content of weld metal to 3 to

10%. Exercise caution when performing
maintenance work at ambient temperature.
Minimize shock loading to potentially
embrittled material. In the case of slide valves,
move to internally-insulated carbon or low
alloy steel.
885oF (475oC) Embrittlement
4xx Series stainless steels exposed to 700oF to

1000oF (370oC to 540oC). 3xx Series stainless
steel weld and cast components can also
experience embrittlement depending on
ferritic content.
PT for cracks or field metallography to identify

presence of and distribution of embrittling
phase.
Do not use 4xx Series stainless steels in

pressure-containing high temperature
environments.
Creep Embrittlement
Highly stressed welded components

constructed of C- Mo, 1 Cr and 1 Cr steels
at > 850oF /455oC (e.g. nozzle welds)
PT or shear wave UT of highly stressed

weldments for cracks in the base metal heat
affected zone.
Creep embrittlement has not yet become an

issue for 1 Cr components in FCCs.
Specifying higher purity 1 Cr steel or 2 Cr
steel are means to prevent embrittlement.
High Temperature Creep
Hot wall reactor vessels, carbon steel reactor

cyclones and hangers, and stainless steel
regenerator cyclones and hangers.
Regenerators or cold wall reactors can
experience creep if the insulating refractory
fails.
Visual and PT to look for cracking and

distortion in structural and pressure containing
components;. Infrared thermography to check
refractory integrity while on-line
Ensure actual service metal temperatures do

not exceed design metal temperatures (e.g.
prevent overheating). In areas which exhibit
metal deformation, use stress-analysis
techniques to ensure thermal expansion
stresses are accounted for in design.
Thermal Fatigue
Reaction mix line; especially at miters.
Visual or PT for cracks and distortion.
Best to eliminate risk of cracking through

design. Eliminate mitered joints where stresses
concentrate.
(1)
CS = carbon steel; 1 Cr = 1 Cr - Mo alloy steel; 1 Cr = 1 Cr - Mo alloy steel; 2 Cr = 2 Cr -1Mo alloy steel; 5Cr = 5Cr - Mo alloy steel; 9Cr = 9Cr - 1Mo alloy steel,
either 12% Cr (4xx Series) or 18%Cr - 8%Ni (3xx Series)
(2)
RT = Radiographic Testing, UT = Ultrasonic Testing, and PT = Dye Penetrant Testing
Page 3 - 58
October 1999
6.0
CATALYTIC LIGHT ENDS RECOVERY UNIT
6.1
ABSTRACT
This section reviews fundamental corrosion issues concerning the Light Ends Recovery section of
a Fluid Catalytic Cracking Unit (FCCU). Cat Light-Ends Recovery (CLER) units process the material
from the overhead system of the FCCU main fractionator to recover propane and heavier
components and to separate light boiling fractions. This section summarizes: a description of the
process, major equipment found in the CLER, types of corrosion and where they occur, corrosion
control and monitoring used.
6.1.1. CLER
CLER Process Description
Gases from the FCCU main fractionator are condensed to allow collection and separation of light
cracked naphtha and off gases. The off gases from the main fractionator reflux drum are then
compressed and cooled in one or more stages. The hydrocarbon liquid condensate streams go to
a stripper (deethanizer) tower while the remaining non-compressed gases are typically sent to an
absorber tower. In many cases these are combined as one tower structure. The deethanizer
removes fuel gas components (C1's and C2's). The absorber uses chilled condensate from the
main fractionator reflux drum (wild gasoline) as lean oil to absorb remaining C3's and heavier
components allowing the fuel gas components to go overhead. The resulting rich oil is combined
with the stripped condensate from the deethanizer and sent to a debutanizer and depropanizer
(or naphtha splitter). These towers separate these streams into propane, butane, light cracked
naphtha and heavy cracked naphtha. (Figure 3-6)
6.2
MATERIALS OF
OF CONSTRUCTION
All components in CLER units are usually made from carbon steel. Carbon steel can be used
because essentially all the hydrocarbon streams are below 150oC (300oF) and because carbon
steel forms a semi-protective sulfide film when exposed to ammonia bisulfide containing sour
waters. Fractionator internals are thinner and corrode from both sides, so they are typically
constructed of 405 or 410(S) stainless steels. Tubes for overhead condensers and compressor
aftercoolers can be admiralty brass, Monel, duplex stainless steel or titanium depending on
cooling water corrosion considerations.
In recent years, special Hydrogen Induced Cracking (HIC) resistant steels have been used to
mitigate hydrogen induced damage concerns. Stainless steel clad equipment has also been used
to remove the risk of hydrogen induced damage altogether.
The purpose of the following section is to point out where problems occur in major equipment
and systems, and to discuss the materials commonly used to alleviate those problems.
6.2.1. Columns
Most columns (absorber, deethanizer, debutanizer, depropanizer, naphtha splitter) are
constructed of carbon steel. As discussed in 4.5, the most common problem is hydrogen
induced cracking and blistering due to exposure to active ammonia bisulfide and cyanide
solutions. Many columns are therefore constructed of special carbon steels (HIC resistant) that
Page 3 - 59
October 1999
improves the resistance to hydrogen damage. In some cases, due to the size and complexity of
the columns, stainless steel cladding (typically 304L) is used.
Tray internals of the columns can be carbon steel particularly in the drier back end towers.
Stainless steel (400 Series) is often used in the wetter first columns to provide alkaline sour water
corrosion resistance for these thinner components.
6.2.2
Exchangers
The majority of exchangers in these units are either coolers, condensers or tower reboilers.
Carbon steel is the material of choice for the process (usually shell) side of the coolers but the
cooling water medium may dictate other needs. Given the alkaline, (NH3 or rich sour water), the
use of copper based alloys (admiralty brass, aluminum brass, copper-nickels) may have a risk of
corrosion or possibly ammonia stress corrosion cracking. Therefore, other water and sour water
resistant alloys such as Titanium (Grades 2,12) and duplex stainless steels are often used. The
carbon steel shells of these exchangers are subject to the same hydrogen induced cracking and
blistering risk as are columns so HIC resistant steels or even stainless steel clad shells have been
used.
Reboilers are also usually all carbon steel unless dictated by the corrosivity of the tubeside heating
medium (steam, hot fractionator streams). The primary problem with reboilers is the collection of
upstream corrosion products in the bottom of the exchanger that cause under deposit corrosion.
Some users have removed the bottom rows of tubes to allow for this.
6.2.3
Piping
It is unusual to have any piping metallurgy other than carbon steel.
6.3
CORROSION PROBLEMS
Corrosion is caused by a combination of aqueous hydrogen sulfide, ammonia and hydrogen
cyanide (sour water corrosion). The rate of corrosion can vary extensively, depending on the
concentration of the above compounds and on process specifics. The amount of H2S, NH3 and
CN formed in the FCCU is usually a function of the amount of S and N in the FCCU feed. In
addition, the actual operation of the FCCU reactor system (reactor temperature, extent of catalyst
burn) may affect the amount formed of H2S, NH3, and CN for a given feed.
In the absence of hydrogen cyanide, aqueous sulfide solutions with pH values above 8 do not
generally corrode carbon steel because a protective iron sulfide film will form on the surface. This
iron sulfide, however, is soft and can be disrupted by flow effects such as turbulence or very high
velocities. Hydrogen cyanide, if present in significant quantity, destroys this protective FeS film
and converts it into soluble ferrocyanide [Fe(CN)6-4] complexes. As a result, the now
unprotected steel can corrode very rapidly. The corrosion rate depends primarily on the bisulfide
ion (HS) concentration and, to a lesser extent, on the cyanide (CN) concentration. For practical
purposes, the bisulfide and cyanide concentrations found in CLEF units, usually do not cause
severe corrosion of carbon steel. Units with excess amounts of chlorides in the fractionator
(enough to cause ammonia chloride salting) may have acidic shock condensation occur in the
first condensation zone of the fractionator overhead.
Page 3 - 60
October 1999
If excess NH3 is generated and the pH rises above 8 to 8.5, copper based alloys are subject to
accelerated corrosion and/or ammonia stress corrosion cracking. Corrosion is also caused by the
formation of soluble cyanide complexes that react with the Cu based materials. Monel has been
successfully used in these services, generally since the temperatures are low enough to sustain
protective sulfide scales.
Cr containing materials generate more stable complex sulfide films and hence improve the
resistance against sour water corrosion. For this reason various forms of stainless steels have been
used subject to fabrication and cooling water considerations. At very high ammonia bisulfide
levels, complexing by CN's can be a problem even for the stable Cr based sulfide scale and
corrosion of stainless steel can occur. Generally the levels of ammonia bisulfide found in the CLER
are not high enough to cause this.
Titanium generates a very stable oxide that is virtually immune to sulfides and hence has been
used particularly in conjunction with sea water cooling. Titanium, however, can become
embrittled due to hydrogen generated as part of ongoing system corrosion reactions. The
hydrogen reacts directly with titanium to form hydrides that reduce the toughness of the material
substantially. This damage is accelerated by temperature and galvanic coupling with other
metals.
6.3.1
Hydrogen Induced Damage

The amount of hydrogen penetrating in ammonia bisulfide solutions into the steel is typically a
function of pH. Acidic solutions will generate higher hydrogen permeation, while a neutral pH
will show a decrease. Excessive pH's though above 8 will show a steady increase in permeation.
Alone at typical CLEF pH's, ammonia bisulfide would generate nominal hydrogen damage
potential.
The HCN, disrupts the FeS scale and increases corrosion and, as a result, greatly increases the
hydrogen available for damage. The effect is so great that apparent corrosion rates may still be
quite low but sufficient hydrogen enters the steel to cause extensive damage.
Hydrogen damage of carbon steel has caused damage to coolers, separator drums,
absorber/stripper towers and overhead condenser shells. Usually, the attack occurs in interstage
and high-pressure separator drums and in absorber/stripper towers. Vapor/liquid interface areas
often show most of the damage, probably because ammonia, hydrogen sulfide and hydrogen
cyanide concentrate in thin water films or in water droplets that collect at these areas.
As a result of the extensive experience with hydrogen induced cracking (HIC) in CLER units,
inspections are generally carried out to monitor for this problem. Common techniques include
Wet Fluorescent Magnetic Particle inspections for surface cracking on equipment interiors and
ultrasonics used to detect both subsurface blistering and cracking. Acoustic emission may be used
to screen vessels for cracking activity during pressurization cycles.
Blistering can be vented to prevent crack growth. Cracks can be ground out and weld repairs are
performed as needed. The extent of repairs is assessed by appropriate engineering support and
code requirements. Heat treatment prior to welding to bakeout absorbed atomic hydrogen to
prevent further cracking during repairs is often done. Post weld heat treatment to temper
hardenable welds and heat affected zones and to reduce residual stresses are also often used. In
severe cases of hydrogen induced damage, equipment replacement may be required. Special
carbon steels with lower S levels, inclusion shape controlling of the remaining S, normalized heat
treatment and hardenability limits are often specified for this service. In some cases, the use of
stainless steel cladding is specified to totally eliminate the problem by stopping the corrosion.
Page 3 - 61
October 1999
More detailed information on hydrogen induced damage, inspection practices, repair techniques
and construction practices have been well summarized in Section 2 5.3.6.
6.3.2
Ammonia Stress Corrosion Cracking

Admiralty or aluminum brass tubes in overhead condensers are exposed to high levels of NH3. As
a result, it is not uncommon for tubes to fail as a result of ammonia stress-corrosion cracking.
Admiralty metal tubes also can corrode by severe localized attack. Admiralty tubes in compressor
aftercoolers have lasted only a few months on some units. For improved service life, replacement
with duplex stainless steel or titanium tubes is often necessary.
6.3.3
Carbonate Stress Corrosion Cracking

The FCCU generates CO2 with small amounts being carried through with the light ends into the
Catalytic Light-End Recovery (CLER) unit. The CO2 is soluble in the condensing waters and can
result in carbonates in the solution. A carbonate rich solution when exposed to the residual
stresses usually associated with welds can cause intergranular stress corrosion cracking of the heat
affected zone. This has been reported in vessels and piping in CLER. Since post weld heat
treatment reduces welding residual stresses considerably, it is effective in reducing this problem.
6.3.4
Fouling/Corrosion of Reboiler Circuits

It is commonly reported that reboiler exchangers accumulate upstream corrosion products. This
leads to under deposit corrosion on the tube surfaces in particular. The tube surface tends to
evaporate the water present and to concentrate and precipitate ionic species causing the under
deposit corrosion.
6.4
CONTROL MEASURES
MEASURES
Certain process modifications have been found to effectively reduce or prevent corrosion and
hydrogen induced damage in CLER units. These include water washing of certain process streams
to dissolve and dilute corrosives (H2S, NH3, CN), polysulfide injection into wash water to lower
HCN content, and corrosion inhibitor injection. While these measures are useful in reducing
blistering, none of these measures will significantly reduce or prevent stress cracking of hard
welds and heat affected zones. High strength bolting, used typically in floating head covers of
exchangers, will also be susceptible to stress cracking. The hard welds and heat affected zones
must be addressed and minimized during fabrication. The bolting problems can be minimized
by using lower strength bolts but they may be a problem to maintenance in tightening up to set
gaskets.
6.4.1
Water Washing
Considering extensive field experience, continuous water washing of sour gas/vapor streams can
be an important method of controlling corrosion and hydrogen entry into steel. Water washing
can be done by contacting the gas/vapor with water in a scrubbing tower, or injecting the water
directly into process piping. The most efficient method of contacting the gas is a scrubbing
vessel. However, many plants use a combination of large water volume rates and a distribution
nozzle to wash the gas in-line. Water washing primarily dilutes the concentration of NH3 and
HCN in process water. The greatest benefits of water washing are seen in the high-pressure
Page 3 - 62
October 1999
section where the partial pressures and hence, the concentration of dissolved NH3 and HCN are
highest. Water is generally injected into the main fractionator overhead, upstream of
intermediate compression stage coolers and/or upstream of the final compression stage coolers.
It is important that the process water, including wash water, not be returned from the highpressure section to the main fractionator reflux drum at the FCCU prior to disposal. This would
cause H2S, NH3 and HCN to flash off as the pressure was reduced at the reflux drum and cause
their concentration to build up in the compression loop.
It is also important that carry-over of corrosive water into downstream equipment be minimized.
This means that sufficient cooling capacity needs to be provided for compressor aftercoolers to
maintain separator drum temperatures as low as possible. On some units, additional drum
capacity may be needed, along with water draw facilities for certain fractionator towers.
Wash water should be injected through a Type 304 or 316 stainless steel distributor or quill that
is located at the center of the piping. There should be at least 15 ft (4.6 m) of piping
downstream of the injection point to ensure proper mixing ahead of coolers and condensers. The
same applies to piping bends, elbows and tees that otherwise would experience impingement
attack. Where parallel heat exchanger banks are being washed, care must be taken to ensure
even water distribution either with balanced piping or individual controlled injections into each
bank of exchangers.
Only high quality water, with a low solids content, should be used. Water quality should be
balanced against availability and cost. Typical water sources are one or more of the following,
listed in order of increasing cost:
Sour Water Condensate (pH 6 to 8.5)

Stripped Sour Water
Boiler Feeder Water
Demineralized Water, Steam Condensate or Steam
If water washing is to be combined with polysulfide injection, alkaline sour water is preferred. The
wash water pH should then be 8. It is common to cascade waters from the main fractionator
through the intercoolers and the aftercoolers. Since the water is pumped to higher pressures it
can absorb more of the corrodents while at the same time minimizing the net quantity of sour
water made.
The amount of wash water depends on the gas/vapor flow rate, the amount of water vapor
present, and the amount and type of corrosives present. Ideally, the amount of wash water
should be the minimum needed to meet one or more of the following typical criteria, listed in
order of decreasing importance:
HCN content of all water draws less than 20 to 25 ppm by weight

pH value of all water draws between 6.2 and 85
76 l/m (20 gpm) or 1545 kg/h (10,000 lb/hr) per MSCF/SD of vapors from the top of
the main fractionator of the FCCU.
Depending on the system, several different types of wash water may have to be injected to meet
these criteria. For example, a slightly acidic sour water stream may be required to depress pH
values. Polysulfide may have to be added to the wash water to decrease HCN levels.
Page 3 - 63
6.4.2
October 1999
Polysulfide Injection
Continuous injection of polysulfide solution into the wash water lowers the HCN content of sour
water condensate by forming harmless thiocyanates (SCN). Polysulfide also reacts with sulfide
corrosion products to produce a more protective film on steel surfaces. Polysulfide injection
should be considered if water washing by itself does not decrease the HCN content below the
recommended 20 to 25 ppm by weight. While several types of polysulfide solutions are
available, most refiners prefer to use commercial 55% by weight ammonium polysulfide
([NH4]2Sx) solution containing 35% by weight polysulfide sulfur. Sodium polysulfide solution is
not recommended because it increases the pH of sour water condensate and reacts more slowly
with HCN in comparison to ammonium polysulfide. It is also considerably more expensive than
ammonium polysulfide solution. Polysulfide solution should be stored and handled in carbon
steel equipment. To avoid sulfur deposition, the solution should be diluted by a factor of ten with
a slipstream of alkaline sour water. The diluted polysulfide solution is then injected into the
various wash water streams, using a simple T-connection. As a rule, the injection rate is designed
so that the amount of polysulfide sulfur added is 50 percent more than the stoichiometric
amount required for conversion of CN to SCN. Actual injection rates are adjusted to ensure some
excess polysulfide that is usually monitored by observing the color of the condensed waters. A
straw yellow color indicates excess polysulfide. The actual amount of free HCN and thiocyanates
(with a target to reach the free HCN of 20 to 25 wppm) can also be measured. Most analytical
techniques though, tend to be either inaccurate or imprecise.
6.4.3
Corrosion Inhibitors
Commercial film-forming amines have reduced hydrogen blistering of steel provided the inhibitor
concentration was sufficiently high. In practice, this means at least 30 ppm by volume versus the
normal 10 ppm.
For this reason, inhibitor injection is relatively uneconomic and recommended only for problem
areas and short-term protection until other measures, such as water washing or polysulfide
injection, can be implemented.
6.4.4
Hydrogen-activity probes and periodic chemical tests are recommended for monitoring the
effectiveness of corrosion control measures. Hydrogen-activity probes use a pressure gauge to
measure the amount of hydrogen that has diffused through a tubular carbon steel specimen.
Recommended key locations for hydrogen-activity probes include different elevations of the
absorber/stripper tower, and the vapor/liquid interface area of the high-pressure separator drum.
To avoid faulty readings due to leaks, hydrogen-activity probes must be pressure-tested with
hydrogen or helium gas and a residual pressure of hydrogen gas should be maintained in the
probes at all times. Changes in pressure due to hydrogen activity are of greater interest than the
actual pressure itself. To facilitate reading and adjusting, pressure gauges and bleed-off valves of
elevated probes may be kept at ground level and connected to the probes by stainless steel
capillary tubing. Depending on the sensitivity of the hydrogen-activity probes used, increases in
reading of less than 1 to 2 psig/day indicate satisfactory control.
Other hydrogen activity measurement techniques are also available. For example, a sealed patch
can be mounted on the exterior surface of the equipment item in areas of suspected high
hydrogen rates. The hydrogen passes through the steel wall and is collected within a sealed
patch. Measurement of the hydrogen build-up can be by various means including vacuum loss or
by reactions with solid state or wet chemistry detectors.
Page 3 - 64
October 1999
Chemical tests for cyanide and thiocyanate content of waste-water streams should be carried out
to determine if any changes occur due to feed and operations change. They can also be used to
monitor water wash and polysulfide injection systems. As indicated above, the actual chemistry
analysis may be a difficult technique and care must be taken to account for air exposure to obtain
consistent results. Air will convert ever-present sulfides to polysulfides and then gradually convert
CN to thiocyanates. As a result, particularly in polysulfide injected systems, the color monitoring
on a shift or daily basis is a simple sampling test often used.
Water pH sampled from high pressure condensates is also commonly used to monitor water
wash rates as indicated above. Care must also be taken since H2S and NH3 will flash off when
depressurized and affect the pH readings. Samples should be collected in pressurized sample
bombs to obtain meaningful results.
6.4.5
Corrosion Probes
Corrosion probes can be used to monitor ongoing corrosion in CLER units. The probes are
especially useful for monitoring high pH corrosion when Cu-based alloys are used in
condenser/cooler bundles.
Page 3 - 65
October 1999
Page 3 - 66
GAS COMPRESSION
Low Pressure
Water
Wash
Water
Wash
PRIMARY
FRACTIONATOR
High Pressure
Water
Wash
ABSORBER
DEETHANIZER
DEBUTANIZER
To Fuel Gas
NAPHTHA
SPLITTER
Lean Oil
DISTILLATE TO TREATING
LCN TO CATALYTIC
REFORMING
HCN TO
BLENDING
SOUR WATER TO RECYCLE
OR DISPOSAL
C3 / C4 PRODUCT TO
POLY OR CHEM PLANT
LEGEND
H
P
Hydrogen Probe Location
Corrosion Probe Location
CATALYTIC CRACKING LIGHT ENDS

RECOVERY UNITS
Figure 3-6
October 1999
Page 3 - 68
7.0
October 1999
CATALYTIC REFORMING
REFORMING UNIT
Catalytic reforming is a process that employs a platinum plus other usually noble metal chloride
activated catalyst. The catalyst converts low quality naphtha in the presence of hydrogen into
high grade motor fuel or aviation gasoline blending stocks. Reforming also produces an aromatic
rich feed stock for extraction and excess hydrogen. The noble metal catalyst is sensitive to lead
from rerun leaded gasoline, sulfur and nitrogen. As a result, the feed is hydrotreated to minimize
sulfur and nitrogen. The hydrotreating is run at roughly 371oC (700oF) and at pressures from
150-900 psi, depending on the unit design.
Typically, the older catalytic reformer units were built when the motor fuel demand was rapidly
increasing. Higher octane motor fuels were required by the higher horsepower automobiles. The
old units were designed to run at roughly 482o-524oC (900o-975oF) end of run and in the 500 psi
pressure range. Newly designed catalytic reformers may run 538o-543oC (1000o-1010oF) at a
total pressure of 150 psi because of better catalysts and a changing product demand. UOP, IFP,
and Chevron used a semiregenerative radial downflow fixed bed, three or four reactor in series
system. The reactors are regenerated when activity decreases significantly because of
accumulated carbon blocking the active catalyst sites. The regeneration essentially burns off the
carbon and the catalyst is reactivated with chloride. Exxon and Amoco used a swing reactor
when the online reactors began to loose activity so that the required number of reactors was
onstream to keep the throughput up. UOP then introduced a continuous regeneration stacked
reactor unit that regenerated a slipstream of catalyst and returned it to the reactor stack. This
setup allowed the operators to run the unit at high activity producing maximum octane without
having to shutdown and regenerate.
A typical catalytic reformer may contain a feed pretreat section such as an oxygen stripper to
remove oxygen from stored feed, a reaction section that includes feed/effluent exchange, charge
and intermediate furnaces, reactors, interconnecting piping, an effluent cooling system and a
product separator. In addition, there is a fractionation section and a recycle gas compression
section. The excess hydrogen is sent to the refinery hydrogen system for use in hydroprocessing
units.
7.1
CORROSION
The early units used a more sulfur tolerant catalyst and, in fact, some had no upstream
hydrotreater. The result was significant H2/H2S corrosion on the 1-1/4Cr-1/2Mo, 2-1/4Cr-1Mo
and C-1/2Mo piping, exchangers, furnace tubes and reactors. These units suffered high metal
losses and rapidly plugged the fixed bed reactors with sulfide scale. After hydrotreating was
added upstream, the major catalytic reformer materials problem was High Temperature
Hydrogen Attack (HTHA). The alloys that were used were C-1/2Mo, 1-1/4Cr-1/2Mo, and 21/4Cr-1Mo. The early reformers had considerable problems with chloride being stripped off the
catalyst and causing corrosion anywhere downstream of the reactors where water condensed
out. The loss of chloride caused some catalyst activity problems that could be solved by
rechloriding the catalyst onstream. Enough chloride was in the system that some refiners found it
necessary to change out some carbon steel piping in areas and replace it with Monel. After
several years, the operation was improved and chloride corrosion problems were minimized. For
several years, the refiners had no chloride corrosion problems until the continuous units began to
operate. Because the new units require continuous chloriding to keep the catalyst activity high,
there is considerable chloride in the system to cause corrosion.
Page 3 - 69
7.2
October 1999
MATERIALS
The reactor section of the catalytic reformers has hydrogen up to about 536oC (1000oF) and
partial pressures up to about 28 kg/cm2 (400 psi). The hydrogen attack situation has dictated
use of 1-1/4Cr-1/2Mo for the hotter areas of the reactor section. The older units used both hot
shell and cold shell reactor designs. The new continuous units use stacked hot shell reactors.
Some of the early hot shell reactors were 1Cr-1/2Mo. At end of run conditions, some of these
reactors were marginally over the "API 941" curve for the material. Considerable inspection effort
by UT and surface inplace metallography has, in general, found no attack other than some
decarburization. It is difficult to tell whether the decarburization was due to hydrogen attack or
was a result of heating during fabrication. Blistering on shell and/or head plates has been fairly
common. It does not seem to be hydrogen attack but from hydrogen diffusing through the wall,
recombining and building up pressure at inclusions in this older normally fairly dirty steel. Later
hot shell reactors have been made from 1-1/4Cr-1/2Mo until the creep embrittlement problems
became fairly common. The replacement reactors have been low impurity 1-1/4Cr-1/2Mo or low
impurity 2-1/4Cr-1Mo. The 2-1/4Cr-1Mo tends to have considerably better rupture ductility.
Recently there have been a few instances of hydrogen attack on 1-1/4Cr-1/2Mo reported to the
API 941 Panel.
The cold shell reactors are refractory lined with an internal stainless steel shroud. A least part of
the nozzles are 1-1/4Cr-1/2Mo or higher alloy for hydrogen attack resistance. The shells have
been C-1/2Mo or carbon steel depending on the specific refiners safety factor in case of gas
bypassing or refractory problems producing hot spots. A typical problem area that causes
bypassing and hot spots is at the shroud to nozzle tie in.
Many of the earlier units used considerable C-1/2Mo for hydrogen attack resistance. Numerous
failures have resulted in the recommendation not to use C-1/2Mo. The problem appears to be
due to the formation of attack prone carbides during steel processing and/or fabrication.
Hydrogen attack on C-1/2Mo can be very spotty and difficult to find. Some advanced UT
methods, such as velocity ratio/backscatter combination, seem to work fairly well but are slow
and, therefore, expensive. The C-1/2Mo situation continues to be a major industry problem since
there is considerable equipment in service and the replacement cost is very high. Some refiners
have replaced piping in high risk (of hydrogen attack) areas because it is cheaper to replace than
to inspect. In general, this does not hold true for exchangers and vessels.
The rest of the unit, outside the reaction section, is generally carbon steel. The major problem
has been migrating chlorides throughout the effluent and fractionation section. Some of the
newer continuous units use chloride traps on both the liquid and gas streams to minimize
downstream equipment chloride corrosion problems.
The charge and intermediate furnaces have typically used 2-1/4Cr-1Mo tubes with years of
successful service. Many of the furnaces are low pressure drop parallel pass designs. Typical
problems have been an occasional ruptured tube due to a low flow, hot spots during startup or
an occasional burner problem. When the continuous units became popular, the tube material
was changed to 9Cr-1Mo to give some added oxidation resistance and therefore allow a higher
limiting skin temperature. As the skin temperatures went from the 593o-621oC (1100o -1150 oF)
range to 621o-649oC (1150 o -1200 oF), the 9Cr-1Mo tubes began to have carburizing problems
that the lower alloy tubes operating at somewhat lower temperature did not have. The
carburization has made the 9Cr-1Mo tubes very sensitive to low flow/startup problems and has
resulted in a number of tube failures. Some of the failures have been primarily operational but
others have been due to some complex carburization and even in some cases metal dusting
problems.
Page 3 - 70
October 1999
If the refinery cooling water system requires copper based alloys, like aluminum brass or
admiralty, they may be very vulnerable to ammonia stress corrosion cracking and must be
thoroughly water washed before the bundles are pulled or exposed to air. The Cu-Ni alloys are
not susceptible to ammonia stress corrosion cracking and have worked for many years as effluent
coolers and or trim coolers. Most modern units use carbon steel air fans instead of shell and tube
condensers/coolers.
There may be some risk of hydrogen embrittlement of hardened compressor parts due to
ammonium chloride or iron chloride exposure. Some refiners use the same restrictions as for wet
sulfide. Keep the hardness below HRc22 and the tensile strength below 90 ksi (620 mPa).
Compressor cracking problems did occur in the early units when migrating chlorides were a
problem. Based on that experience, it would seem prudent to apply the same groundrules to the
new continuous units.
With the semi-regenerative units, a major problem is controlling the corrosion that occurs during
regeneration. Very specific procedures have been written to minimize these corrosion problems
by proper neutralization. The other more recent problem with the semi-regenerative units is the
need to regenerate frequently in order to keep octane up in competion with the continuous
units. The quick carbon laydown and subsequent carbon burning can cause serious internal
distress and result in operational problems and possible long downtime replacing the internals.
Page 3 - 71
October 1999
Page 3 - 72
NET HYDROGEN
CW
GAS
HYDROGEN RECYCLE
COMPRESSOR
250 to 500 psig
350 to 650 psig
100 Fo
925 to
o
975 F
925 too
975 Fo
925 too
975 Fo
STEAM OR
HOT OIL
100 F
CW
CW
HEATERS AND REACTORS
REFORMATE
HYDROGEN
SEPARATOR
STABILIZER
NAPHTHA
CHARGE
SEMI-REGENERATIVE
CATALYTIC REFORMING UNIT
Figure 3-7
October 1999
Page 3 - 74
8.0
October 1999
HYDROPROCESSING UNITS
CORROSION IN HYDROPROCESSING UNITS
8.1
ABSTRACT
This section contains an overview of hydroprocessing unit design and operations followed by
descriptions of the corrosion mechanisms and other materials degradation problems associated
with these processes. The mechanisms include high temperature hydrogen attack, high
temperature H2S corrosion, stress corrosion cracking by chlorides, sulfur acids, and sulfides,
aqueous corrosion by ammonium bisulfide and others. For each mechanism, a description of the
conditions causing or accelerating attack, the locations within the unit where attack may occur,
and the methods of prevention are discussed. A walk-through of a materials selection guide is
also given.
8.2
INTRODUCTION
INTRODUCTION
Hydroprocessing units are used in a refinery to upgrade hydrocarbon feedstocks by removing
undesirable constituents and/or converting heavier feeds into more valuable, lighter products.
The feedstocks range from naphtha to vacuum residuum. The reactions occur under a hydrogenrich environment at elevated temperatures and high pressures, in the presence of catalysts. Types
of hydroprocessing units and their primary purposes are:
Hydrotreaters including hydrodesulfurizers - remove sulfur and/or nitrogen

Hydrocrackers - crack heavier feeds into lower boiling point products
Hydrogenators - add hydrogen to unsaturated or other hydrogen-deficient hydrocarbons
Hydrofiners - remove color bodies
The chemical reactions which accomplish these objectives all occur more or less simultaneously,
but processing conditions are varied somewhat, in order to maximum the rate of the primarily
desired reaction.
Some units consist of two stages with the first stage completing the hydrotreating reactions and
the second stage doing the hydrogenation and cracking. From the corrosion viewpoint, the
important distinction between these stages, is that the feed to a hydrotreater contains
appreciable amount of sulfur and nitrogen while the feed to the second stage hydrocracking
section may not.
Sulfur and nitrogen react with hydrogen to form hydrogen sulfide and ammonia within a
hydrotreater reactor system. These compounds have a significant effect on corrosion and
materials selection regardless of whether the hydrotreater is a separate plant, the first stage of a
two-stage hydroprocessing unit, or a single-stage hydrocracker. Since sulfur, nitrogen and
ammonia typically reduce the activity of second stage catalyst, most are removed in the
hydrotreating stage. Hence, there are fewer corrosion considerations and less upgraded materials
in second stage hydrocrackers compared to first stages or single stage designs. Single stage
hydrocrackers are a high pressure operation that not only hydrotreats, but also converts heavier
hydrocarbons into lighter products and hydrogenates the converted hydrocarbons.
Page 3 - 75
8.3
October 1999
PROCESS DESCRIPTIONS
DESCRIPTIONS
A simplified flow diagram of one type of a typical hydrotreater is shown in Figure 3-8 Page 1. The
reactor contains catalyst(s) and typically operates between 42 to 141 kg/cm2 (800 to 2000 psi)
and 371 o to 454 oC (700 oF to 850 oF). Hydrogen is injected into the feed which is heated in
feed/effluent exchangers and a furnace. In the reactor(s), sulfur and nitrogen compounds are
converted to hydrogen sulfide and ammonia. The reactor effluent is cooled through various heat
exchangers and typically one or more air coolers, and then is sent to the separator vessels. Water
is typically injected for fouling/corrosion control upstream of the air coolers.
The gas phase from the separators consists primarily of hydrogen with some very light
hydrocarbons and a high percentage of the H2S generated in the reactors. Gas from the
separator is recycled back to the feed through a compressor, with some make-up hydrogen also
being added. The liquid hydrocarbon phase from the separators is sent through pressure letdown valves to the fractionation section of the unit.
The water phase from the separators contains almost all of the ammonia formed in the reactors.
The dissolved H2S in this water combines with the NH3 to form ammonium bisulfide (NH4HS) as
well as inorganic salts,. such as ammonium chloride. Traces of cyanide may also be present.
In a single stage Hydrocracking Unit (see Figure 3-8 page 2), the feed is mixed with hydrogen,
heated, and passed through catalyst-filled reactors. The reactor effluent is cooled, the gas phase
(consisting mostly of hydrogen) is recycled back to the feed, and the liquid hydrocarbon phase is
sent to the distillation section. Typically, reactor pressures are between 106 to 211 kg/cm2 (1500
to 3000 psi) and temperatures are from 343oC to 454oC (650oF to 850oF).
There are several common variations in flow scheme. In some vacuum resid desulfurizers, the
reactor effluent enters the separator vessels directly from the reactor with only a small amount of
prior cooling. Hot vapor and liquid streams are separated and individually cooled. Another
common process variant is that some units use a high pressure amine absorber to remove H2S
from the recycle hydrogen stream while others units do not. Distillation systems can vary
substantially in flow scheme. Some units have an H2S stripper column before the fractionator.
The essential differences in configuration that influence corrosion will be discussed later.
8.4
MATERIALS SELECTION
SELECTION
8.4.1
Reactor System
The materials of construction used in the reactor system of a hydrotreater, single stage
hydrocracker or the first stage of a two stage unit must be resistant to the following forms of
corrosion damage:
High temperature hydrogen attack

High temperature H2-H2S corrosion
Aqueous corrosion by ammonium bisulfide
Stress corrosion cracking by chorides, sulfur acids or sulfides
Naphthenic acid corrosion (if feed has a high neutralization number)
Page 3 - 76
8.4.2
October 1999
Reactor Feed System

Up to the point of recycle hydrogen addition, the hydrocarbon feed to the plant is generally noncorrosive to carbon steel, except when the feed contains H2S at >260oC (500oF) or naphthenic
acid at >232oC (450oF). In those cases where the plant feed is corrosive because of high
temperature and dissolved H2S, corrosion can be minimized by using alloys containing 5%
chrome or better. Naphthenic acids may necessitate the use of Type 316 or 316L.
After the point of recycle hydrogen addition, progressively higher alloys are required as the stock
is heated. This is to resist both hydrogen attack and high temperature H2/H2S corrosion. The
threshold temperature for H2/H2S corrosion depends on the amount of H2S introduced with the
recycle gas, but in most plants it is on the order of 232oC (500oF).
Above this temperature, austenitic stainless steels are typically used for piping and exchangers to
provide corrosion resistance. Hot piping is commonly constructed of Type 321 as Type 347
piping is more costly, generally less available, and somewhat more difficult to weld than Type
321. Exchanger bundles are typically Type 321 and the shells and channel sections are clad with
Type 321 or Type 347. However, Type 347 is normally used for cladding or weld overlay on
thick-walled components.
Hydrogen attack becomes a materials consideration in the reactor feed system when above about
232oC (450oF). Above this temperature threshold, carbon steel cannot be used, even as base
metal on stainless steel clad components. Although stainless steels are immune to hydrogen
attack under plant conditions, hydrogen can diffuse through stainless cladding to attack the base
metal.
Based on these considerations, the reactor feed system is normally built of carbon steel where
temperatures are below about 232oC (450oF). Above this temperature, piping will generally be
Type 321 steel to prevent H2/H2S corrosion, and exchanger channel sections and shells will be
stainless clad with 1 1/4Cr-1/2Mo or 2 1/4Cr-1Mo base metal used as needed for protection
from hydrogen attack.
The role of corrosion products in plugging catalyst and reducing run lengths may also provide
economic justification for upgrading alloys in feed exchangers, piping and furnace tubes. Even
acceptable corrosion rates can generate large volumes of corrosion products. Corrosion products
can be a cause of plugging in downflow reactors, but rarely cause problems with radial flow
reactors.
Heater tubes and return bends are commonly constructed of Type 347 stainless steel, although
Type 321 has also been used. Return bends should be wrought rather than cast, both to obtain
superior quality and because castings tend to develop sigma embrittlement above 538oC
(1000oF).
8.4.3
Reactors
Reactors are constructed of low alloy steel for hydrogen attack reasons and are protected against
H2/H2S corrosion by austenitic stainless steel roll-bond cladding or weld overlays. The most
common base metal for reactors is 2 1/4Cr-1Mo although 3Cr-1Mo has also been used. Alloys
lower than 2 1/4Cr-1Mo are occasionally used when temperature and hydrogen partial pressure
permit.
Reactor internals are constructed of an austenitic stainless steel, typically Type 321 or 347.
Aluminizing or aluminum diffusion coatings are sometimes specified for catalyst support screens
Page 3 - 77
October 1999
to help prevent corrosion which could result in plugging from scales. Aluminum is essentially
immune to H2S corrosion.
8.4.4
Reactor Effluent System

In the reactor effluent system, from reactor to the reactor effluent/stripper feed exchanger,
materials selection is based on the same criteria as in the reactor feed system. Stainless steels
should be used for corrosion protection until the stock is cooled below the threshold for high
temperature H2/H2S corrosion [about 232oC (450oF)]. Alloys resistant to hydrogen attack must be
used down to about the same temperature. The exact threshold temperatures for H2/H2S
corrosion and hydrogen attack vary somewhat as a function of the partial pressures of H2S and
hydrogen. On surfaces such as exchanger tubes and tube sheets which are exposed to two-side
attack, the conditions existing on both sides must be considered.
From the reactor outlet temperature down to about 232oC (450oF) (the H2/H2S corrosion
threshold), piping and exchanger bundles are generally Type 321 and exchanger shells are Type
321 or Type 347 clad. Base metals used for exchanger shells may again be 2 1/4Cr-1Mo or 1-1/4
Cr-1/2Mo, depending on the alloy content required to provide resistance to hydrogen attack.
Below the 232oC (450oF) hydrogen attack threshold, carbon steel is generally used.
8.4.5
Reactor Effluent - Distillation Feed Exchangers

Many, but not all plants, use an exchanger that cools the reactor effluent stream by exchanging
it with the separator liquid on its way to the first distillation column after the reactor system. Such
exchangers pose special corrosion problems. One problem is entrainment of small quantities of
salt-containing water in the separator liquid. As this stock is heated, the water evaporates, leaving
salt deposits on the tubes. Carbon steel tubes may corrode in the presence of these deposits.
Tube life is highly variable, depending primarily on temperature and the amount of salt entrained
into the exchanger. Chrome-moly steels perform no better than carbon steel in this instance.
Austenitic stainless steels are likely to fail by chloride stress corrosion cracking or underdeposit
ammonium chloride pitting. In general, austenitic stainless steel tubes should not be used in this
exchanger except in existing plants where good performance has been proven.
This results in the materials selection for these tubes being a choice between carbon steel and
very expensive alloys such as Alloy 825, AL6XN or Alloy 625. For new plant construction, carbon
steel is generally specified. The exception is when reactor effluent-side temperatures are so high
that a better alloy is required to resist high temperature H2S corrosion. Under these conditions
and for replacement of existing exchanger bundles where carbon steel shows inadequate life, an
alloy with very high resistance to chloride corrosion and SCC should be used.
8.4.6
Effluent Air Coolers

This is probably the piece of equipment most vulnerable to ammonium bisulfide corrosion. Most
plants initially install carbon steel tubes for effluent air coolers, however, some units with high Kp
values have installed duplex stainless steel or Alloy 800 or 825. In other cases, carbon steel has
experienced corrosion failures due to excessive velocities, oxygen in the injection water,
maldistribution of flow or other causes. Where such problems have occurred and tube materials
have been upgraded, Monel, Alloy 800 or 825 have been used for tube replacement.
The Alloy 800 series provide resistance to high concentrations of ammonium bisulfide. However,
in recent years, there have been cases of polythionic SCC of Alloy 800 piping and equipment.
Page 3 - 78
October 1999
Alloy 825 is stabilized and contains molybdenum and, hence, has superior resistance to
polythionic and chloride SCC, as well as NH4HS corrosion. It is generally the preferred alloy over
Alloy 800. However, if due to cost or availability, Alloy 800 is selected, care should be exercised
to avoid purchasing high carbon, coarse grain material to minimize sensitization during
fabrication and welding. Current experience indicates that fine grained material may be
impossible to obtain.
Duplex stainless steels such as Type 2205 are increasing in use for tubes and header boxes.
Numerous special requirements should be imposed on the materials and fabrication and welding
practices. Welds or heat affected zones that do not have the proper ratio of austenite/ferrite in
their microstructure can be susceptible to hydrogen embrittlement and SSC. Both positive and
negative experiences, with regards to obtaining acceptable fabrications have been reported.
Although they should have good NH4HS corrosion resistance, austenitic stainless steel tubes have
seldom been used in this service due to the risk of chloride stress corrosion cracking. In the past,
several companies have used Type 410 or Type 430 stainless steel tubes in effluent air coolers but
failures have occurred by isolated pitting. Alloy 400 (Monel) has been used successfully for a few
air coolers in the past, but may not be suitable for units with high levels of NH4HS.
In some plants, effluent air coolers have been constructed with stainless steel ferrules at both inlet
and outlet ends of the steel tubes. This provides increased protection against tube end erosioncorrosion. Ferrules have also been installed with good results in existing steel air coolers where
tube end attack had occurred. When ferrules are used, the ends of the ferrules must be tapered
to provide a smooth flow transition.
Carbon steel header boxes may also experience corrosion if velocities or turbulence are excessive.
Industry experience indicates that the majority of those effluent air coolers experiencing tube
corrosion will also suffer attack on header boxes. For this reason, alloy header boxes should be
used with alloy tubes.
Although ammonium bisulfide corrosion is the major concern, failures can also occur from NH4Cl
corrosion. Units with two air coolers in series, with the water injection downstream of the first air
cooler, may experience NH4Cl deposition and pitting at the outlet ends of the first air cooler. This
generally occurs when the air cooler's outlet temperature is relatively low. NH4Cl sublimes at
temperatures above the NH4HS sublimation temperature. The deposits are hygroscopic and often
provide enough cooling of the metal to result in water condensation and acid formation beneath
the deposit. No practical materials upgrade will resist this problem, hence, it is usually avoided by
raising the process temperature.
8.4.7
Effluent Air Cooler Inlet

Inlet and Outlet Piping
The piping upstream (from the water injection point) and downstream of the effluent air cooler is
often subject to the same NH4HS erosion-corrosion problem as the air cooler. Corrosion is
typified by highly localized metal loss at bends, tees and other points of local turbulence. Such
corrosion is most likely to occur when the process fluid is high in NH4HS concentration and
where fluid velocities are high. Carbon steel piping should be designed with a 6 m/s (20 ft/s)
maximum limit. Special attention must be given to areas downstream of any injection lines or
even small branch connections in this service.
When new units are predicted to be extremely corrosive, when high reliability is desired, when
periodic rigorous inspection is considered difficult or uneconomical, or when corrosion occurs in
existing plants, alloy piping is often installed. Alloy 800, Alloy 825, Type 316L [for application
below 60oC (140oF)], Alloy 20 and Alloy 2205 have been used. SCC failures have occurred on
Page 3 - 79
October 1999
Alloy 800 when it was supplied with high carbon contents and large grain sizes as discussed in
8.5.6. The upper velocity limit for alloy piping can be raised to 9 m/s (30 ft/s).
When carbon steel is used, it generally has a high corrosion allowance of 6.4 mm (1/4 in).
Balanced inlet and outlet piping is also typically specified.
NH4HS corrosion can also occur downstream of the separators in lines handling wet
hydrocarbons or foul water and in other piping where the process fluid contains appreciable
quantities of H2S and ammonia, and any quantity of liquid water. For example, lines handling
separator vapor could corrode if the vapor were further cooled to condense out additional water.
In many but not all plants, the wet, sour process fluids which are present downstream of the
effluent air cooler are capable of causing rapid SSC of high strength steels. Valve stems and trim
are the components most likely to suffer damage. In new plant construction, Type 316 trim is
typically specified for services where SSC appears likely to occur.
8.4.8
Separator Vessels
Except for units operating with greater than 10% NH4HS, separator vessels normally have very
low corrosion rates. The major concern is that the incoming process fluid may impinge on the
shell or heads, causing localized NH4HS corrosion at that point. Typically, installing a stainless
steel impingement baffle or wear plate of adequate size to shield the entire impingement area
avoids the problem.
The only major exception is the hot separator in a hydrotreater designs where the first separator
operates at or near the full reactor outlet temperature. Accordingly, stainless-clad construction is
used to provide resistance to high temperature H2/H2S corrosion. The base metal is chosen to
resist hydrogen attack at operating temperature.
Cold separators containing sour water may be subject to severe HIC and SOHIC. On recent units,
these vessels have been built of HIC resistant steel or entirely clad with a 300 series stainless steel.
8.4.9
Recycle Hydrogen System

Significant corrosion is seldom encountered in this part of the system. The only potentially serious
materials problem is SSC of the recycle gas compressor, as it typically contains materials like 4330
or 4140 steel, which can be susceptible to SSC if too strong or too hard. To avoid this problem, it
is common practice to limit the strength and hardness of compressor materials. However, for
additional protection against SSC, precautions should be taken to keep the compressor dry by
methods such as maintaining the mist eliminator in the compressor knock out drum in good
condition and by steam-tracing the compressor suction line from the knock out drum. Keeping
the system dry also protects against NH4HS erosion-corrosion. Valve trim is also typically 316.
8.4.10 Distillation Section

Construction materials used in the distillation section are chosen on the basis of the need to resist
high temperature H2S corrosion. Where H2S is present at temperatures above 232oC to 316oC
(500oF to 600oF) (depending on H2S concentration), alloy is required. Where H2S is absent, or
where temperatures are below 232oC (500oF), carbon steel is generally adequate. Where the
temperature exceeds 316oC (600oF), corrosion may occur at H2S levels as low as 1 ppm.
Page 3 - 80
October 1999
If the transfer line temperature were as high as to 316oC (600oF), carbon steel would be expected
to corrode rapidly in the transfer line and column flash zone. A new plant designed for such
conditions would use 5 Cr for the transfer line and 12 Cr cladding in the flash zone of the
column. If the transfer line temperature were well over 316oC (600oF), furnace tube corrosion
could occur even on 5 Cr. It would then be necessary to use 9 Cr or Type 321 tubes, depending
on the H2S content of the stock, the stock temperature and fumace design. 12 Cr tubes are not
recommended because they cannot be readily welded and have embrittlement problems when
used as pressure parts.
In the bottom half of many fractionator columns, corrosion of carbon steel is often minimal
despite the high temperature because H2S has been stripped out of the hydrocarbon. The same is
true for the fractionator column reboiler. The key factors are the H2S content of the bottoms and
effectiveness of H2S stripping. At temperatures over 316oC (600oF), corrosion of carbon steel can
occur if the H2S content exceeds about 1 ppm. Under these conditions, corrosion of carbon steel
could occur in the bottom of the fractionator column, the reboiler, the bottoms line to the
splitter and the flash zone of the splitter column. Corrosion would not be expected in the bottom
of the splitter column or in its reboiler because H2S should be stripped out.
Aside from the possibility of high temperature H2S corrosion, the only other corrosion concern is
in the overhead of the distillation column. Overhead condensers and drums exposed to both
water and H2S may experience moderate corrosion, but this is rarely a serious problem and may
be further controlled by injection of a filming amine inhibitor. In the overhead systems, many
refiners may apply materials and fabrication controls to minimize wet H2S cracking, however, HIC
steels are not typically used in this location.
8.5
CORROSION PHENOMENA
PHENOMENA IN HYDROPROCESSING UNITS
8.5.1
High Temperature Hydrogen Attack

Hydrogen at >232oC (>450oF) and partial pressures >7 kg/cm2 (>100 psi) can cause hydrogen
attack of carbon and low alloy steels. It results in decarburization which weakens the metal. In
addition, methane forms at the interstices and builds up internal pressure which induces
fissuring, blistering and possible failure. Alloying with chromium and molybdenum reduces the
degree of attack because of their strong carbide-forming characteristics.
All hydroprocessing units involve the use of hot, high pressure hydrogen in the reactor systems.
Therefore, it is essential to use construction materials which are resistant to hydrogen attack at
the operating conditions. Hydrogen attack is not a consideration in the distillation systems, since
hydrogen partial pressures are very low.
API 941, "Steels for Hydrogen Service at Elevated Temperatures and Pressures in Petroleum
Refineries and Petrochemical Plants", contains operating limits for steels in hydrogen service. The
API 941 curves are also referred to as the "Nelson Curves" as they were originally developed by
G.A. Nelson. The axes are hydrogen partial pressures and operating temperature and the area
below a given material's curve is considered acceptable operating conditions for that material.
The common upgrades for cases when carbon steel is not acceptable, are 1 1/4 Cr-1/2 Mo and 2
1/4 Cr-l Mo alloys. Historically, many oil companies apply a 28oC (50oF) safety factor when using
these curves to select materials (except for reactors, for which a 14oC (25oF) safety factor is
typically used).
Page 3 - 81
October 1999
C-1/2 Mo was commonly used for equipment and piping in the past, but due to failures at
conditions where it had been predicted to be acceptable, it is no longer used for new
construction in hydrogen service. Its design curve has been removed from the Fourth Edition of
API 941 which recommends rigorous inspection of existing C-1/2Mo equipment. A new Fifth
Edition of API 941 was published in January, 1997.
Hydrogen can diffuse through overlays and attack the underlying base material, so regardless of
the type of overlay, the base material should satisfy the API 941 requirements for the operating
conditions. (Refer to Section 2 5.4)
8.5.2
High Temperature Hydrogen Sulfide / Hydrogen Corrosion

Hydrotreater feedstocks typically contain sulfur compounds, such as mercaptans, sulfides,
disulfides, thiophenes and others. Under the reactor conditions, most of these compounds are
converted to hydrogen sulfide (H2S). H2S reacts with metals at high temperatures >232oC (450oF)
by direct sulfidation. The presence of hydrogen typically increases H2S corrosion rates on carbon
steel and low alloy steels. Areas of hydroprocessing units susceptible to H2-H2S corrosion are the
reactor feed downstream of the hydrogen mix point, the reactor, the reactor effluent and the
recycle hydrogen gas including the exchangers, heaters, separators, piping, etc. in these systems.
A reasonably good estimate of corrosion rates in H2-H2S systems can be made from data
commonly known as the "Couper-Gorman Curves."
Most refiners have found that under H2-H2S corrosive conditions, alloys with up to five percent
chromium offer no significant improvement over carbon steel, and nine percent chromium alloys
provide only marginal improvements and are thus considered ineffective in resisting corrosion.
While some 12% Cr stainless steels are resistant to most H2S ranges, corrosion may occur on
materials with chromium contents on the low end of the acceptable range or where service
conditions are severe. Also, the 12% Cr alloys are not a commonly used material for this service
due to fabrication difficulties and possible embrittlement problems. Even in cases where Cr-Mo or
12%Cr alloys would have acceptable lives, if they are predicted to have moderate corrosion rates,
there may be economic justification to upgrade to reduce fouling or plugging from corrosion
products. Austenitic stainless steels (18% Cr) are generally required to meet the corrosion and
plugging resistance requirements. (Refer to Section 2 5.1.5)
8.5.3
High Temperature Hydrogen Sulfide Corrosion

Corrosion in Areas with Negligible Hydrogen
In the some feed systems upstream of the hydrogen injection point, and in some fractionator
sections after the majority of hydrogen has been separated out, another form of H2S corrosion
can occur in areas >288oC (550oF). In these cases, the hydrogen partial pressures are typically <4
kg/cm2 (50 psi). Note that the fractionation schemes shown in Figure 3-8 pages 1 and 2 do not
show high temperature areas, but there are other designs which are up to 382oC (720oF). Some
process variations are discussed later in 8.5.
One primary difference between this corrosion mechanism and the mechanism discussed above
is that an alloy's corrosion resistance is directly proportional to its chromium content, so alloys
with intermediate chromium contents provide an improvement over carbon steel. The typical
upgrades used when above 288oC (550oF) are 5 Cr, 9 Cr, 12 Cr or 300 Series stainless steels. The
12 Cr alloys (such as 405 and 410S stainless steels) are typically used for only nonpressure thin
components such as cladding, trays and tubing, as thick sections are relatively difficult to
fabricate.
Page 3 - 82
October 1999
The available published corrosion rate data is inadequate to address the H2S corrosion
encountered in hydroprocessing unit fractionation sections. The most commonly used published
literature for the feed areas upstream of the hydrogen injection point are known as the
"McConomy Curves". These curves were developed from crude unit and hydroprocessing unit
feed furnace tube field data (upstream of the hydrogen injection point) and laboratory tests.
Since their publication, these curves have come under criticism of being too conservative for
some applications. A revised set of McConomy Curves reduced the predicted corrosion rates by
about 2.5. The revised curves are commonly used for crude, coker and FCC units and
hydroprocessing feed systems upstream of the hydrogen injection point.
While the revised curves are still somewhat conservative for the applications listed above, they
have been shown to be inaccurate in a non-conservative direction for predicting -corrosion rates
for low concentration, high temperature H2S corrosion found in some hydroprocessing unit
fractionation sections. One proposed reason is that in the data sources for these curves, the total
sulfur content included a wide range of sulfur species, some of which were corrosive and some
were not. In contrast, the total sulfur content in the hydroprocessing unit fractionator feed
stream is almost entirely corrosive. The selection of materials for these areas are primarily based
on field experience of similar units. (Refer Section 2 5.1.5)
8.5.4
Naphthenic Acid Corrosion

Naphthenic acid corrosion may be a problem in hot feed piping and equipment if the feed
contains a high concentration of naphthenic acids. The naphthenic acid concentration in the
feed is indicated by the total acid neutralization number (TAN or neutralization number) as
determined by ASTM test methods D664 or D974. One caution when using these methods to
test the feed to a hydroprocessing unit is that the H2S and mercaptans should be removed prior
to the naphthenic acid determination.
At high acid concentrations (approximately neut. no. >1.5) and at >232oC (>450oF), carbon
steel, chrome moly steels and even some 300 Series stainless steels can suffer accelerated
corrosion. Upgrading to Type 316L or other high molybdenum alloys (>2-3% Mo) is the most
common prevention method used to avoid this corrosion. Modern 316L is often more susceptible
to corrosion than older 316L, as in the past its average molybdenum content was 2.5%, while
today it is 2.0%.
The experience to date has been that components upstream of the hydrogen mix point,
operating at 232-288oC (450-550oF) have shown accelerated corrosion. Downstream of the
hydrogen mix point, in reactor feed piping, heater tubes and exchanger tubes, no problems due
to naphthenic acid corrosion have been reported to date. This is somewhat surprising as these
components are typically Type 321 or 347, which contain no molybdenum and are susceptible
to this type of corrosion. Type 316L is generally not used as it would sensitize at >371-399oC
(>700-750oF) and could suffer intergranular polythionic acid Stress Corrosion Cracking (SCC)
during shutdowns (Refer to Section 2 5.2.12) Pilot plant data has shown that a significant
portion of the naphthenic acid content is destroyed in the first reactor, hence no special material
considerations need to be taken for this corrosion problem downstream of the reactor.
8.5.5
Ammonium Bisulfide
Bisulfide Corrosion
Ammonium bisulfide (NH4HS) is the product of ammonia and hydrogen sulfide gases. As the
reactor effluent stream cools down, solid NH4HS can crystallize out of the vapor phase and may
quickly plug up exchanger tubes and cause underdeposit corrosion. However, ammonium
bisulfide dissolves readily in water, so the common practice is to inject water ahead of effluent air
Page 3 - 83
October 1999
coolers to prevent bisulfide deposition. Unfortunately, ammonium bisulfide solutions may also be
highly corrosive, leading to rapid attack of carbon steel tubes. The presence of small quantities of
cyanides, oxygen or ammonium chloride in the process fluid tends to further accelerate
corrosion.
Industry studies of effluent air cooler and piping corrosion have led to a general understanding of
the following important factors:
Corrosion is fundamentally caused by ammonium bisulfide and corrosion rates increase in

relation to its concentration. The appropriate NH4HS limit for a given unit seems to depend
on many variables and typical limits for the separator water range from two to ten percent.
The NH4HS concentration can be somewhat controlled in operating units by adjusting the
water injection rate.
In the materials selection phase for a new unit, some refiners use the Kp factor (developed by
R.L. Piehl), which is the mole percent of ammonia times the mole percent of hydrogen
sulfide in the dry stream entering the reactor effluent coolers to determine the need to alloy
air coolers. The additional expense for a materials upgrade may be partially justified by the
difficulty, unreliability and cost of the frequent inspections that are normally required for a
carbon steel air cooler in corrosive service.
NH4HS corrosion is velocity-sensitive. At tube velocities in excess of 6 m/s (20 ft/s), severe
corrosion erosion of carbon steel tube ends is likely to occur unless the process fluid is very
low in ammonia and H2S. Hence, air coolers with carbon steel tubes should be limited to 6
m/s (20 ft/s) maximum. Stainless steel ferrules may be used at both inlet and outlet ends of
the tubes that handle stocks with high ammonia and H2S content. For alloy tubing, the
upper limit for velocity can be raised to 9 m/s (30 ft/s).
Proper water injection rates are key to controlling corrosion. Water injection at less-thanrecommended rates can lead to accelerated corrosion. The water rate should be high enough
that >25% remains unvaporized, the NH4HS content in the separator water meets the desired
limit, and the process velocity in the tubes is within the specified limits.
Injection water must be free of oxygen (less than 15-50 ppb). The presence of oxygen leads
to rapid corrosion and fouling. An oxygen contamination problem, including intermittent
spikes of high oxygen can be detected by regular testing of the injection water. This problem
can also be indicated by analysis of corrosion products in the exchanger. Many of the other
factors affecting corrosion are interrelated, and one factor may be up to desired ideal if all the
others are acceptable. However, oxygen contamination is one factor which alone can
immediately lead to severe corrosion.
To prevent corrosion, good flow distribution of vapor, liquid hydrocarbon and water phases
is essential. This is accomplished by using a balanced design for inlet and outlet headers and
by keeping fluid velocities high enough to minimize phase separation. The phase separation
problem and risks of deposits and underdeposit corrosion occur when velocity is <3 m/s (<10
ft/s) in the air cooler.
Accumulation of fouling deposits in the cooler tends to cause flow maldistribution and can
lead to corrosion. Good water distribution is essential to minimize fouling. Hence, many units
have injection points on the inlets to each bank.
Page 3 - 84
8.5.6
October 1999
Air-cooled exchangers designed for this service should have header boxes at both ends. Ubends should be avoided because they are sensitive to erosion-corrosion. Header boxes
should be designed to permit access for tube cleaning. (Refer to Section 2 5.2.3)
Chloride Stress Corrosion Cracking

Cracking
In general, Chloride Stress Corrosion Cracking (SCC) of austenitic stainless steels can occur when
the following conditions are present:
liquid water at >60oC (140oF)

chlorides, especially with a concentrating mechanism
tensile stresses
Chlorides (generally in ppm levels) may be present in the plant feed and/or in the makeup
hydrogen when the hydrogen source is a catalytic reformer unit. In either case, the result can be
ammonium chloride deposition in heat exchangers downstream of the reactor. Austenitic 300
Series stainless steels are acceptable for feed/effluent exchangers as they are too hot for
significant amounts of salt formation and any salts that do accumulate are dry. However, in
cooler exchangers, the 300 Series stainless steels are generally not used due to the risk of chloride
SCC.
Numerous refineries have experienced chloride SCC in austenitic stainless steel drain lines
branching off from the piping in the reactor feed and effluent systems. The failures generally
occur shortly after startup following a turnaround. They are due to water with chlorides
accumulating in the lines during the shutting down washing steps and/or when the piping is
open to the atmosphere. During the startup, when the temperature in the branches exceeds
60oC (140oF), SCC begins to occur. As the trapped water begins to boil off, the chlorides will
become concentrated and hot. Cracks and leaks have generally occurred at non stress-relieved
welds which would have high residual stresses. The solutions being used by various refiners are to
stress relieve these drain lines or to upgrade them to Alloy 20 or Alloy 825.
There have been isolated cases of other chloride SCC problems in the austenitic stainless steel
feed and effluent piping. One reported site for chloride cracking initation is in the ring grooves of
flange faces.
A risk of chloride cracking also exists in reactor effluent/stripper feed exchangers when austenitic
stainless steel is used for the tubing. Hydrocarbon liquid from the separator (i.e. stripper feed)
may contain a small amount of dissolved water. This water, in turn, contains some dissolved
chloride salts. When the hydrocarbon is reheated, the water vaporizes, leaving salt deposits.
Under these conditions, austenitic as well as duplex stainless steels may suffer rapid chloride
cracking or pitting.
Feed heaters, especially those with vertical austenitic stainless steel tubes, may occasionally be
exposed to shutdown conditions capable of causing chloride cracking. If pools of water are
allowed to remain in the tubes after a water wash, chlorides in this water will concentrate when
the furnace is fired and chloride cracking of the tubes could occur. To minimize this possibility,
the water must be removed from the tubes before startup. This can be accomplished by airblowing or, on vertically-tubed furnaces, by oil circulation.
Stainless-clad reactors can experience chloride cracking if they are water-flooded during the
catalyst removal process. To minimize chloride cracking problems during catalyst removal, the
water used for flooding should be low in chloride. Condensate, or fresh water with chloride
content below about 50 ppm can be used. Some refiners use a soda ash solution to flood the
reactors during catalyst removal. (Refer to Section 2 5.3.1)
Page 3 - 85
8.5.7
October 1999
Polythionic Acid Stress Corrosion Cracking (PASCC)

This form of SCC occurs only on austenitic stainless steels and a few related austenitic alloys such
as Alloy 800, and only when these alloys have become sensitized by welding, postweld heat
treatment, or service over 371-454oC (700-850oF). Polythionic acids form from the reaction of
iron sulfide scales with oxygen and moisture. Hence, this stress corrosion cracking occurs during
shutdowns when the equipment is exposed to air and moisture, and not during normal
operation on hydroprocessing units.
Commonly-used stainless steels and their limits to avoid sensitization are:
Type 304 or 316 - Used only for parts which are not welded or heat treated, and have
operating and regeneration temperatures less than 371o-399oC (700o-750oF).
Type 304L or 316L - Acceptable for welded or heat-treated parts with maximum
operating temperatures up to 371o-399oC (700o-750oF).
Type 321 - Acceptable for welded or heat-treated parts with maximum operating
temperatures up to 418oC (780oF) depending on whether the material was given a mill
stabilization heat treatment.
Type 347 - Acceptable for welded or heat-treated parts with maximum operating
temperatures up to 454oC (850oF) depending on whether the material was given a mill
stabilization heat treatment.
HF Modified - In the past, this cast material was considered immune to PTA SCC and it
was used for piping with maximum operating temperatures up to 454oC (850oF).
However, in recent years, there have been reports of a few incidences of PTA cracking at
operating conditions below 371oC (700oF).
Other methods to lessen the risk of polythionic acid stress corrosion cracking involve minimizing
the formation of the polythionic acids. One method is to nitrogen blanket the equipment and
piping during turnarounds to prevent exposure to air and moisture. Another is a wash with soda
ash solutions, per NACE RP0170-93, "Protection of Austenitic Stainless Steels and Other Austenitic
Alloys from Polythionic Acid Stress Corrosion Cracking during Shutdown of Refinery Equipment".
Soda ash washing leaves a residual alkaline film on the metal to neutralize any acids as they are
formed. Many refiners apply these practices to all austenitic stainless steel components (with the
possible exception of clad reactors), however, others use them only on austenitic stainless steel
suspected to have some degree of sensitization. (Refer to Section 2 5.3.4)
8.5.8
Wet H2S Cracking

Wet H2S cracking within hydroprocessing units can be Sulfide Stress Cracking (SSC), Hydrogen
Induced Cracking (HIC) or Stress Oriented Hydrogen Induced Cracking (SOHIC). These types of
cracking are possible where steel is exposed to liquid water with about 50 ppm H2S or greater.
(i) Sulfide Stress Cracking

Sulfide Stress Cracking (SSC) is primarily a problem on high strength ferritic or martensitic steels
and generally occurs only on steels above about 621 N/mm2 (90,000 psi in yield strength or
above Rockwell C) Rc 20-22 in hardness. It occurs only when liquid water and hydrogen sulfide
are present, but even a thin film of condensed moisture provides enough water for cracking to
occur.
Page 3 - 86
October 1999
One of the most commonly encountered sulfide cracking problems in hydrotreaters is the failure
of valve stems and valve trim. The commonly-used trim material of 400 Series stainless steels (1213% Cr) are particularly prone to failure. Sulfide cracking of valve trim occurs most frequently in
the reactor effluent system after it has cooled sufficiently for liquid water to be present. This
includes the area of the effluent air coolers and separators. SSC also has occurred in the recycle
hydrogen system and some distillation overhead systems.
The common solution to sulfide cracking of valve trim is to use an 18 Cr-8 Ni stainless steel, since
austenitic stainless steels are essentially immune to sulfide cracking. Another option is to use 12
Cr trim which has been certified as meeting NACE MR0175, "Sulfide Stress Cracking Resistant
Metallic Materials for Oilfield Equipment" . However, the 18-8 trim is usually cheaper than the 12
Cr with NACE MR0175 requirements. Cobalt-based, stellite valve trim is also immune to sulfide
cracking and is used for pressure let-down valves in wet H2S services for resistance to various
possible problems including SSC. ASTM A 638 Grade 660 precipitation hardened stainless steel is
also used where high strength is required for valve stems. Grade 660 should have a 35 Rc
maximum hardness and conform to NACE MR0175.
Centrifugal recycle gas compressors may be subject to SSC. Depending on strength and
hardness, commonly-used rotor materials such as AISI 4330 and 4140 steel can be highly
susceptible. Normal practice for hydrotreater recycle hydrogen service is to specify that
compressors be constructed of alloys with <621 N/mm2 (90,000 psi) yield strength and 22 Rc,
hardness plus be heat treated such that there is no untempered martensite present in the
microstructure.
Hard weldments will also be susceptible to SCC, and should be avoided by using good
fabrication practices. NACE RP0472, "Methods and Controls to Prevent In-Service Environmental
Cracking of Carbon Steel Weldments in Corrosive Petroleum Refining Environments" should be
followed to limit hardness of both weld deposits and heat affected zones. The procedures used
in welding of carbon steel should ensure that the 200 HB weld hardness limit is not exceeded.
On low alloy steels, most refiners use a maximum allowable weld hardness of 225 HB.
(ii) Hydrogen Induced Cracking and Stress Oriented Hydrogen Induced Cracking
There is not yet a single well-defined method for preventing HIC and SOHIC. They can occur on
vessels which have received postweld heat treatment (PWHT) but the likelihood of SOHIC is
significantly reduced. For this risk reduction, and since PWHT nearly eliminates the risk of SSC,
most refiners specify PWHT for equipment in wet H2S service. Requiring 100% wet fluorescent
magnetic particle testing (WFMT) after PWHT is also common. Seamless grade B piping has not
experienced SOHIC to date (except under extremely severe conditions), so most refiners do not
require PWHT of this piping.
Low impurity HIC-resistant steels (clean steels) have recently become available, normally at a
considerable premium over standard carbon steel. However, their performance has not always
been successful in avoiding hydrogen-related problems. Most refiners currently use these steels
for only critical pieces of equipment in the most severe services. In hydroprocessing units, these
steels have been used for the headers in effluent air coolers and for separators.
Coatings are being tried with mixed success throughout the industry. Cladding with stainless
steel (such as 304L or 316L) is the most reliable way of avoiding wet H2S cracking, but it is
expensive. The base metal under the cladding typically has no added requirements, and PWHT is
not needed for a fully clad vessel unless required by Code. (Refer to Section 2 5.3.6)
Page 3 - 87
8.5.9
October 1999
Material Property Degradation Mechanisms
(i) Temper Embrittlement

Temper embrittlement of chrome-moly steels, particularly 2 1/4Cr-1Mo, occurs when the steel is
heated for a long time in the 360o to 566oC (680o to 1050oF) range. It is generally believed that
embrittlement is due to the gradual accumulation of tramp elements such as antimony, tin,
arsenic and phosphorus in the grain boundaries of the steel. The embrittlement results in a
significant rise in the steel's ductile-to-brittle transition temperature (i.e. the temperature below
which cracks may propagate in an instantaneous, catastrophic manner and above which cracks
will be arrested by the inherent toughness of the material).
As delivered, a typical older vintage reactor would have had a minimum impact strength of 54 J
(40 ft-lb) at 10oC (50oF). But after long-time, high-temperature exposure, the 54 J impact
strength could rise in extreme cases to as high as 149oC (300oF). In other words, the steel could
behave in a brittle manner at temperatures below 149oC (300oF). The extent to which the
transition temperature rises is a function of steel chemistry and microstructure, as well as the
duration of exposure to service temperatures above about 360oC (680oF). Many hydrotreating
reactors that operate above 360oC (680oF) end-of-run temperatures may suffer temper
embrittlement in service. Below 360oC (680oF), temper embrittlement occurs so slowly that it is
generally of little practical consequence.
To avoid the possibility of sudden brittle fracture of reactors operating in the embrittling
temperature range, the normal procedure is to maintain low pressure in the reactor when below
the ductile-to-brittle transition temperature. Low pressure is generally defined as ~25%-40% of
the reactor's design pressure. It is generally accepted that the probability of brittle fracture is
minimal if the stress level is less than 20% of the material's yield strength.
Each reactor must be considered individually in determining these temperature and pressure
restrictions. Current practice in the construction of new reactors is to require ultra clean steels
and to carefully screen the materials being used to minimize temper embrittlement and ease the
required pressurizing restrictions. Adhering to the pressure restrictions and Minimum Pressuring
Temperatures (MPT) are essential elements of safe operation of hydroprocessing units. [Refer to
Section 2 5.5 (vii)]
(ii) Hydrogen Embrittlement
Hydrogen embrittlement is also a concern in the reactors of hydroprocessing units. These
reactors operate at high enough temperatures and hydrogen partial pressures to result in a
significant concentration of dissolved hydrogen within the walls at operating temperatures. If the
reactor walls are thick enough and are cooled rapidly when shutting down, the dissolved
hydrogen will not be able to escape from the metal during cooling. If significant amounts of
hydrogen remain dissolved in the steel after cooling, mechanical properties will be temporarily
affected. This degradation of mechanical properties, called 'hydrogen embrittlement', exists only
while the hydrogen remains in the steel, and the steel will regain its original properties as the
hydrogen is allowed to escape. Even though hydrogen is present in the metal, properties are
affected only at temperatures below about 149oC (300oF).
One concern is that pre-existing fabrication flaws or other defects will grow due to hydrogen
embrittlement. To avoid this risk, special cooling procedures are usually required when removing
reactors from service to let a significant amount of the hydrogen diffuse out of the metal before
the reactor is cooled below 149oC (300oF). This typically involve cooling no faster than a specified
maximum rate. In most cases, cooling rates of 28oC/hr (50oF/hr) to 56oC/hr (100oF/hr) provide
enough time for degassing.
Page 3 - 88
October 1999
Heavy-walled reactors in hydrogen service should be inspected with particular care after initial
construction and during plant turnarounds to guard against existing defects which might enlarge
due to hydrogen embrittlement. (Refer to Section 2 5.4.)
Page 3 - 89
October 1999
Page 3 - 90
CW
TO GAS RECOVERY
HYDROGEN
MAKE-UP
CW
LIGHT NAPHTHA
FURNACE
FRACTIONATOR
750 Fo
LEAN
DEA
ABSORBER
HEAVY NAPHTHA
LIGHT GAS OIL
REACTOR
QUENCH
REACTOR
EFFLUENT
600 F
RICH DEA
TO GAS
RECOVERY
100 Fo
750 F
100 Fo
CW
HEAVY GAS OIL
HIGH-PRESSURE
SEPARATOR
LOW-PRESSURE
SEPARATOR
180 Fo
o
350 F
FRACTIONATOR
FEED
HYDRODESULFURIZERS AND
HYDROFINERS
Figure 3-8
Page 1
START
WATER
October 1999
Page 3 - 92
LPG
100 Fo
LT GASOLINE
STABILIZER
700 Fo
RECYCLE GAS
COMPRESSOR
REACTOR
SPLITTER
FEED
HEATER
FLASH
GAS
HY GASOLINE
630 Fo
FROM
HYDROGEN
PLANT
FEED
HP
SEPARATOR
NAPHTHA
100 Fo
MAKE-UP GAS
COMPRESSOR
750 Fo
KEROSENE
GAS OIL
WATER
START
WATER
LP
SEPARATOR
590 Fo
HYDROCRACKING
Figure 3-8
Page 2
330 oF
AIR COOLER
170 Fo
SOUR
WATER
FRACTIONATOR
BOTTOMS
October 1999
Page 3 - 94
9.0
October 1999
ALKYLATION UNIT
SULFURIC ACID ALKYLATION
9.1
ABSTRACT
This section reviews fundamental corrosion and materials issues in sulfuric acid alkylation units.
Corrosion in these units is primarily driven by sulfuric acid and by esters formed in the process.
Materials of construction in these units are generally carbon steel although alloys are used
selectively to resist acid corrosion. Corrosion control is usually achieved by limiting acid strength
and by caustic washing product streams. Because the process involves refrigeration, external
corrosion under insulation is also a concern.
9.2
PROCESS DESCRIPTION
DESCRIPTION
Refinery cracking operations, primarily fluid catalytic cracking, yield large quantities of light gases
which may be converted to gasoline blending components through an alkylation reaction. The
sulfuric acid alkylation process chemically combines a low octane olefin (usually butylene,
propylene or amylene) and an isoparaffin (usually isobutane) in the presence of sulfuric acid
catalyst to yield higher octane alkylate (primarily isooctane or isoheptane).
The basic types of sulfuric acid alkylation units are shown in Figures 3-9 and 3-10. Both types of
units have four major sections as follows:
Reaction
Treating
Fractionation
Refrigeration
The principal differences between the two are based on the reactor designs and the method in
which refrigeration is accomplished.
9.2.1
Reaction Section
In the reaction section, olefin feed is brought into intimate contact using mixing, with
concentrated (93 to 98 wt%) sulfuric acid at temperatures between about 5o and 15oC (40o and
60oF). The olefin feed may be pretreated prior to entering the reactor. For example. caustic
washing may be done to remove sulfur, coalescers may remove water thus reducing detrimental
acid dilution, or filters may be used to eliminate solids which could cause plugging and fouling
problems.
In the auto-refrigerated stirred reactor design (Figure 3-9), alkylation occurs in multiple
compartment reactors which are agitated using relatively low speed paddle-type mixers in each
of the compartments. The isobutane and acid are mixed and added to the reactors separately
from the olefin feed. Heat of reaction is removed by allowing a portion of the light hydrocarbon
to vaporize and auto-refrigerate.
In the effluent refrigerated contactor design (Figure 3-10), the acid-feed emulsion is mixed in a
large reactor vessel. The acid and olefin-isobutane mixture are added separately at the eye of a
Page 3 - 95
October 1999
mixing impeller which maintains the emulsion and moves it along the reactor. Refrigerant is
circulated through tubes in the reactor to remove the heat of reaction.
Following the reactor, the acid-hydrocarbon emulsion is separated in a settler. A majority of the
spent acid is returned to the reaction stage with fresh, concentrated make-up acid. A small
portion of the acid is purged to maintain the acid concentration. Alkylation reduces the
concentration of the sulfuric acid, thus creating lower concentration spent acid of 88 to 90 wt%
concentration.
9.2.2
Treating Section
In the treating section, residual acid catalyst and acidic by-products are removed from the reactor
effluent by one or more of several consecutive treating steps, including neutralization with dilute
caustic (NaOH) and water washing. A typical system is shown in Figure 3-11. Mixing is often
accomplished using in-line mixers followed by drums to separate the alkylate from the caustic
and water. Some plants have installed acid wash facilities to extract esters from the reactor
product thus minimizing the impact on downstream fractionation facilities. Poor mixing or
caustic wash operation can result in caustic carryover into downstream fractionation equipment.
Concentrating the caustic in the column reboiling system may result in caustic cracking problems
in the column and reboiler.
9.2.3
Fractionation Section
In the fractionation section (Figure 3-12) alkylate is separated from butane and excess isobutane.
Following the settler and treating section, the reactor products are usually sent in succession to a
deisobutanizer and a debutanizer where isobutane, normal butane and alkylate product are
separated. In some locations, the alkylate is further fractionated to provide flexibility in product
use. In addition to use as a fuel blending stock, alkylate may be used as a feed to solvent
production units.
9.2.4
Refrigeration Section
Because the alkylation process is exothermic, refrigeration is used to limit temperatures to a
favorable operating range, generally in the 5o to 15oC (40o to 60oF) range. Refrigeration is
accomplished in one of two ways depending on the reactor type. Where stirred auto-refrigerated
reactors are used, cooling is accomplished by controlled vaporization of a portion of the light
hydrocarbon contained in the reactor as illustrated in Figure 3-9. This approach is known as autorefrigeration. In the contactor design, isobutane becomes the refrigerant for the cooling coils of
the contactors as shown in Figure 3-10. In both types of systems, flashed vapors are
recompressed and propane is removed before recirculating the remaining stream to the reactor.
The depropanizer feed is often caustic and water washed to remove acid contamination in a
treating system similar to the one on the reactor product.
Page 3 - 96
TYPICAL AUTO-REFRIGERATION
ALKYLATION PLANT WITH STIRRED
REACTORS
Figure 3-9
Propane
Depropanizer
Condenser
Compressor
Reactor
Accumulator
Settler
Feed
Chiller
Olefin
Feed
Feed/
Effluent
Exchanger
Treating
Section
Feed/
Effluent
Exchanger
Recycle to Acid
Fractionation
Section
Butane
Alkylate
Recycle Isobutane
Spent Acid
Fresh Acid
TYPICAL EFFLUENT REFRIGERATION

ALKYLATION PLANT
WITH CONTACTOR-TYPE REACTOR
Figure 3-10
Propane
Condenser
Depropanizer
Treating
Section
Compressor
Fractionation
Section
Accumulator
Butane
Refrigerant
Recycle
Isobutane
Contactor
Olefin
Feed
Feed/
Effluent
Exchanger
Alkylate
Settler
Vapor-Liquid
Separator
Recycle
Acid
Spent Acid
Fresh Acid
Recycle Isobutane
TYPICAL CAUSTIC and

WATER WASH FACILITY
Figure 3-11
Feed to
Deisobutanizer
In-Line Mixer
Reactor
Product
Fresh
NaOH
Fresh
H20
Spent
NaOH
H20
Purge
Caustic Wash
Water Wash
Butane
Isobutane to Reactor
Deisobutanizer
Debutanizer
Reboiler
Reboiler
From
Treating
Alkylate
TYPICAL FRACTIONATION
FACILITY
Figure 3-12
October 1999
Page 3 - 100
9.3
October 1999
CONSTRUCTION
Sulfuric acid alkylation units are built primarily of carbon steel, typically with a 1/4 in (6 mm)
corrosion allowance in the areas where sulfuric acid is present. In the fractionation section, a 1/8
in (3 mm) corrosion allowance may be used for towers.
Carbon steel welds likely to be contacted by sulfuric acid may in some cases be postweld heat
treated to minimize preferential corrosion of welds and weld heat-affected zones. Weld root
beads are often made using Gas Tungsten Arc Welding (GTAW) to provide high quality welds
with limited slag and weld deposit penetration into the line to minimize turbulence which can
increase corrosion rates. Cold-worked metal (usually bends) are often stress relieved.
In high concentration sulfuric acid, carbon steel depends on a film of iron sulfate for corrosion
resistance and if this film is destroyed by high flow velocities and turbulence, corrosion can be
quite severe. For this reason, flow velocities of any streams containing significant amounts of
concentrated sulfuric acid are usually limited to 0.6 to 0.9 m/s (2 to 3 ft/sec). Special alloys are
used for valves, pump internals and injection and mixing nozzles. Piping just upstream and
downstream of the caustic and wash-water injection points in the treating section often requires
selective alloying.
Where excessive corrosion of carbon steel is encountered, alloy 20, an austenitic alloy especially
designed to resist corrosion by sulfuric acid, Ni-Mo alloy B-2, Ni-Cr-Mo alloys C-4 or C-276, highsilicon cast iron, or high-nickel cast iron are usually suitable alternatives. It should be noted that
alloy B-2 is susceptible to higher corrosion rates in the presence of oxidizing agents in the acid or
by air contamination.
In practice, valves and pumps in concentrated and spent sulfuric acid service often are made
from solid alloy 20. For hydrocarbon streams containing traces of concentrated or dilute sulfuric
acid, steel-body valves with Type 316 stainless steel or alloy 20 trim can be used. In this service,
steel pump casings, sometimes weld overlaid with aluminum bronze, have been used
successfully. High silicon iron pump impellers are often used.
Piping for hydrocarbon/acid mixing lines ahead of the reactors may require alloy 20 because
water contamination of feed stocks can cause severe corrosion of carbon steel. Ni-Cu alloy 400
has been found useful for reactor effluent lines around the point of caustic injection. Alloy 400
and titanium Grade 2 have been used as a replacement for carbon steel or Admiralty brass for
tubes in overhead condensers.
In general, most organic coatings are not resistant to concentrated sulfuric acid. Fluoropolymers
such as PTFE have excellent resistance, however, and are used extensively for gaskets, pump and
valve packing and for mixing nozzles.
9.4
MATERIALS AND CORROSION

CORROSION PROBLEMS
Under ideal operating conditions, few corrosion problems occur. Many streams, however,
contain potentially troublesome compounds and any meaningful corrosion control program
must be aimed at controlling these compounds through suitable process changes. Except where
uncontrolled acid dilution occurs, most corrosion problems occur in the fractionation section.
Much of the damage will be found in reactor-effluent lines, overhead systems and reboilers.
Damage can be especially severe in deisobutanizer (DIB) overhead systems. Compounds which
Page 3 - 101
October 1999
are responsible for corrosion are either contaminants in feed stocks or reaction by-products
contained in the reactor effluent.
9.4.1
Sulfuric Acid Corrosion
(i) Acid Concentration

The principal corrosive in alkylation units is concentrated sulfuric acid which is used as a catalyst
in the alkylation reaction. Within the typical ranges of H2S04 concentration, temperature and
velocity, carbon steel can be used satisfactorily for a large part of the unit. Mixtures of sulfuric
acid and hydrocarbon are generally less corrosive than fresh acid. However, sulfuric acid
corrosivity is not linear with concentration. Diluted acid is much more aggressive than
concentrated acid. For example, carbon steel is satisfactory above 85 wt% concentration, but
usually cannot be used below 85 wt%.
Acid corrosion can be particularly troublesome during unit shutdowns and preparation for entry.
Without proper operating precautions, diluted acid can be present and result in very high
corrosion rates on carbon steel. To compound this problem, dilution of acid is exothermic and
the temperature increase can further accelerate corrosion.
(ii) Acid Temperature and Velocity
Temperature and velocity also have a direct relationship to corrosion rates and the rates can
increase substantially if they become excessive. As a result, control of the velocity within specific
limits is a normal practice and the temperature is usually kept as low as possible. Typical velocity
limits are 0.6 to 0.9 m/s (2 to 3 ft/s) in sulfuric acid. Higher velocities may be permissible in the
low temperature sections of the plant. Alloy materials will be required at some high velocity areas
and in some spent acid systems, especially those where the acid may be heated.
Particularly troublesome locations for erosion corrosion, even when velocities are well controlled,
are mixing tees, throttling valves, restriction orifices, check valves and low point bleeders along
with pump suction and discharge where high turbulence is likely to occur. Under such
circumstances, it is common to specify alloy 20, alloys B-2, C-4 or C-276 or PTFE components in
place of carbon steel.
It was reported at a NACE International T-8 corrosion information exchange that high nickel iron
impellers of contactor reactor impellers can experience accelerated tip erosion. Two plants
reported similar excessive erosion of the impellers. The solution developed for this type of erosion
was to hardface the impeller tips with a material like hard facing alloy 21.
(iii) Acid Dilution
Control of acid concentration is very important not only from the process aspect but also to
minimize excessive corrosion. For example, when the acid strength drops substantially below 88
wt.%, a phenomenon called 'acid runaway' can take place. During acid runaway, reactions other
than alkylation take place. Polymerization forms large quantities of acid soluble materials. Such
reactions drastically raise the acid dilutents. Under such extreme conditions, it is not possible to
maintain the acid strength, even with adding fresh acid at the maximum rate possible. Esters are
also formed which cause high corrosion rates in equipment downstream of the settlers.
Page 3 - 102
October 1999
(iv) Hydrogen Grooving

A special problem associated with sulfuric acid corrosion is called 'hydrogen grooving'. Hydrogen
grooving is a form of localized, accelerated corrosion which can occur in mixed phase acid
piping, manways of vessels and especially above some kinds of nozzles in acid storage tanks.
Hydrogen generated by sulfuric acid corrosion rises along the vessel, tank or pipe wall causing
accelerated turbulence which helps to remove protective sulfate scales. Usually a pattern of
parallel grooves is observed, thus the name hydrogen grooving.
Corrosion due to hydrogen grooving occurs at a rate faster than H2S04 corrosion. In one
publicized case, accelerated corrosion and hydrogen grooving caused an acid tank to split
vertically in a line coincident with the position of a flush-mounted, side-entry acid inlet nozzle.
Such a failure could also have occurred with a top entry nozzle that discharges near the tank wall.
In piping systems, hydrogen grooving has been observed at elbows at the top of a vertical piping
run and along the pipe wall leading up to that point.
(v) Feed Contaminants
Typical feedstock contaminants include water, mercaptans and diolefins. When present in
sufficient quantity, contaminants tend to stabilize the acid emulsion, consume acid by H20 or
polymer dilution, or react with the acid to produce troublesome by-products. These can
ultimately result in corrosion problems downstream of the reactor if improper levels of
contaminants are allowed to enter the system.
Water enters the reactor with butane-butylene feed. It may also be present in recycle isobutane
because reactor effluent is caustic treated and water washed. The amount of dissolved water in
feed stocks depends on the temperature, on average doubling for a 16oC (30oF) increase in
temperature. The amount of entrained water can be controlled to some extent by modifying the
feed treating operations.
Mercaptans and other sulfur compounds in the feed normally are removed by caustic treating
and water washing, but residual amounts may still remain.
Diolefins originate from catalytic-cracking operations and cannot normally be removed from
alkylation feed streams without special processing. Butadiene polymerizes forming acid soluble
compounds. Formation of these compounds reduces acid concentration and increases make-up
acid requirements. Very high levels of diolefins make the plant more susceptible to upsets
resulting in acid runaway.
9.4.2
Acid and Neutral Esters

Water, mercaptans and diolefins not only dilute and consume sulfuric acid catalyst, but also
increase the amount of undesirable byproducts in the reactor effluent. These include entrained
sulfuric acid catalyst as well as acidic alkyl sulfates and neutral dialkyl sulfates from secondary
alkylation reactions. Both acid and neutral esters of hydrocarbon and sulfuric acid are produced
in alkylation reactors. High space velocity reactors and high temperatures resulting from poor
operating conditions favor ester formation.
(i) Acid Esters

Reactor effluent normally contains 30 to 100 ppm of alkyl acid sulfates, largely dissolved in the
hydrocarbons. However, they may be greater due to entrained sulfuric acid. The acid esters will
be corrosive to equipment upstream of the caustic wash facilities. Typically, the reactor product
Page 3 - 103
October 1999
pump and the reactor product exchanger are affected by these acid esters. The caustic wash will
eliminate corrosion of downstream equipment by neutralizing the acid esters as well as small
quantities of acid carried over from the settler. Alkyl acid sulfates are not corrosive after
neutralization with caustic in the treating section. If not effectively removed in the caustic wash,
they may revert to acid at the high temperatures encountered in reboilers and cause corrosion
problems in towers and overhead systems, as well as fouling problems in reboilers. Upsets in
reactor operations increase alkyl acid sulfate formation. Upsets in the treating section increase
carryover of water and neutralized esters.
(ii) Neutral Esters
Reactor effluent also contains 10 to 150 ppm of dialkyl sulfates. These neutral esters are largely
dissolved in the hydrocarbon phase and ordinarily cannot be neutralized or removed by caustic
and water washing. Breakdown of these esters can occur, starting at about 75oC (170oF), when
they are heated in the deisobutanizer preheater or reboiler. This decomposition can then cause
corrosion and fouling problems similar to those caused by alkyl sulfates. Dialkyl sulfates appear to
be especially troublesome in the alkylation of propylene feed.
Decomposition of these esters forms H2S04 which will combine with water and cause acid
corrosion in feed preheaters. Corrosion can also occur in the reboilers and in the deisobutanizer
overhead condenser. The residues that remain in the reboiler from the coking of hydrocarbon on
the hot tubes of the reboiler can cause fouling and overheating, leading to failure of the tubes.
The polymeric compounds are responsible for fouling of the reboilers.
When alkyl and dialkyl sulfates decompose in fractionation section reboilers, sulfur dioxide and
polymeric compounds are formed. The S02 formed from the reduction of the H2S04 rises in the
tower and can cause problems in the overhead system due to the low pH water. Sulfur dioxide
combines with water and forms sulfurous acid in the top tower sections, the overhead
condensers and reflux drums. Although the pH of the overhead water is usually 6 to 7, lower pH
could be a symptom of the presence of esters.
9.4.3
Acid Carry-over
Downstream of the acid/hydrocarbon separator, the reactor effluent normally contains 50 to 500
ppm of acid catalyst. Under normal circumstances, and with proper contacting, the caustic and
water wash systems of the treating section remove the trace acid. During upsets, however, large
slugs of acid may pass through the treating section essentially un-neutralized. Acid slugs can
cause considerable corrosion damage quickly because the high boiling-point [165oC to over
315oC (330oF to over 600oF)] and the high specific gravity tend to concentrate sulfuric acid in
tower bottoms and reboilers of the fractionation towers.
Acid carry-over is also a concern in the refrigeration systems of sulfuric acid alkylation units. While
the corrosion rate of the sulfuric acid is generally tolerable, accumulated corrosion over many
years can go undetected since acid may accumulate locally unnoticed. Designing vapor systems
in particular to minimize low points can help reduce acid collection.
9.4.4
Corrosion Under Insulation

Insulation
Equipment in the unit which operates below the ambient air temperature may be vulnerable to
external corrosion. When breaches of the insulating system and vapor barrier occur, the low
metal temperature will cause moisture to condense and this can cause localized Corrosion Under
Insulation (CUI) where water may collect. It is important to recognize the point of the insulation
breach is not often the same location where water will collect and cause corrosion.
Page 3 - 104
9.4.5
October 1999
Fouling Problems
Fouling problems often accompany corrosion in the fractionation section, primarily in reboilers of
various towers. High vaporization rates make reboilers particularly vulnerable to fouling. As
discussed previously, fouling is caused by alkyl and dialkyl sulfates in the reactor effluent which
decompose and polymerize at temperatures above 121oC (250oF). The deposits vary from
varnish-like coatings on relatively cool surfaces to asphalt and coke-like deposits on hot surfaces.
Insoluble corrosion products and inorganic salts often accumulate in reboilers and add to the
bulk of deposits with polymers acting as binders.
Fouling occurs to a lesser extent in the lower section of certain towers, particularly the rerun and
DIB towers. Antifoulants may be injected into the tower feed streams to minimize this problem.
Reboilers are protected, at the same time, because antifoulants tend to stay with the heavier
hydrocarbon fractions.
9.5
CORROSION CONTROL
CONTROL MEASURES
Good process control plays an important part in avoiding corrosion problems by good feed
preparation and in maintaining acid concentration. Feed preparation includes the removal of
contaminants by caustic treating, water washing, coalescing and filtering. Alkylation plants that
operate with healthy conditions do not need the use of any special corrosion control additives.
However, as esters increase with increasing throughput and more severe alkylation conditions,
the resultant increased esters can cause corrosion control to be more of an issue.
9.5.1
Reactor Section Corrosion

Corrosion
Corrosion of reactor effluent lines downstream of the effluent treating section is caused by
excessive amounts of entrained acid catalyst or the presence of esters and can usually be
corrected by suitable changes in the caustic treating and water washing operations. Otherwise,
partially replacing the carbon steel effluent lines with type 316L or alloy 20 may be required as a
long-term solution to the problem.
9.5.2
Tower Overhead Corrosion

To control excessive corrosion in fractionation section overhead systems, the use of caustic and
water wash of the reactor product is of primary importance to remove acidic contaminants.
However, tower overhead corrosion may still occur as a result of acid carry-over and ester
decomposition. In tower overhead systems, neutralizing and filming-amine corrosion inhibitors
have sometimes been used. However, the application of treating chemicals to the deisobutanizer
overhead system must be approached cautiously as recirculation of the amines with the
isobutane may contribute to problems in the reaction section. Such problems are typically not
associated with chemical treatment of the debutanizer overhead system.
Because a neutralizer has the benefit of helping to control corrosion in downstream equipment
by chemically reacting with the corrosive, they may more commonly be added to the
deisobutanizer feed to tie-up acid released during decomposition of esters. This approach to
neutralizer usage avoids isobutane contamination by the amine. In this operating scenario,
Page 3 - 105
October 1999
fouling by neutralizer salts of the reboiler may occur, but these may be removed by water
washing the tower.
If their use is determined to be acceptable from an operations standpoint, filming amine
inhibitors are typically injected into the overhead line at rates ranging from 5-10 ppm by volume,
based on the amount of hydrocarbon condensed in the reflux drum. In order to reduce corrosion
rates and inhibitor consumption as much as possible, overhead-water condensate can be
neutralized to a pH value between 6 and 7. Both neutralizing amines and ammonia have been
used, but in some cases ammonia has caused stress-corrosion cracking of Admiralty brass tubes in
overhead condensers.
Corrosion control in the depropanizer overhead is most effectively achieved by keeping the
system dry. Wet systems vulnerable to acid carry-over have sometimes been treated before the
depropanizer with caustic wash to remove acid contamination and water wash to remove
caustic.
9.5.3
Reboiler Corrosion and Fouling Control

Reboiler corrosion is most easily avoided by keeping reboilers free of fouling deposits and
minimizing the presence of esters. In severe cases, injection of filming-amine corrosion inhibitors
into the reboiler feed line has proven to be beneficial.
Process changes designed to control corrosion problems will also minimize fouling problems. In
addition, periodic blowdown and water jetting are beneficial in resolving the fouling problems of
reboilers. In more severe cases, antifoulants can extend operating runs by keeping potential
foulants dispersed in the hydrocarbon phase. Maximum benefits are realized by continuous
injection into the tower-bottoms line ahead of the reboiler at rates ranging from 5-20 ppm by
volume, based on the amount of liquid hydrocarbon which enters the reboiler. Antifoulants are
usually effective in preventing the formation of hard, baked-on deposits on heat exchanger
surfaces. As a result their use, reboiler bundles may be cleaned more easily by steaming or water
washing.
9.5.4
Acid Tanks
NACE International Recommended Practice RP0294-94 is a valuable reference on H2S04 acid
storage tank design to minimize corrosion. This RP includes the following recommended design
details:
Tanks should be fitted with a top inlet nozzle.

Inlet nozzle should be placed away from the side wall and be fitted with a section of pipe
protruding at least 150 mm (6 in) into the tank or with a dip tube terminating no less than
600 mm (24 in) from the tank floor.
Splash plates should be provided beneath side outlet nozzle piping and inlet nozzle dip
tubes.
Manways and side nozzles should be lined or weld overlaid with alloy to prevent hydrogen
grooving.
Valves, inlet and outlet pipes, vents and wear or splash plates should be a corrosion resistant
alloy such as alloy 20.
The internal upper shell, roof and roof structure exposed to acid vapors may be coated.
Components exposed to acid vapors are more vulnerable to corrosion than those immersed in
Page 3 - 106
October 1999
acid. Coating manufacturers should be consulted on appropriate coatings for this application.
Where feasible, the roof supporting structure may be external to the tank.
9.5.5
Corrosion Control During Unit Shutdowns

Corrosion can be particularly troublesome during unit shutdown as equipment is rinsed. Acid
dilution increases the corrosion rates on carbon steel through the effect of the reduced
concentration as well as through increasing temperature during acid dilution. Consequently,
proper draining and flushing procedures are needed. Dilute acid can form which is corrosive to
carbon steel. To gas free the unit, equipment is often flooded with water until the pH of the
drained water is greater than 6.0 and drained as soon as possible.
Flushing of the reactor, settler, tank and other equipment with low strength caustic is often done
during turnarounds before water washing. Since much of this equipment is not postweld heat
treated, control of caustic strength is important. At a recent NACE meeting, one company
reported foaming the neutralizing solution to reduce the required volume to 10% of their normal
usage. They also added 0.1 vol.% of an acid indicator to serve as a visual indicator of the
neutralization process.
9.5.6
Corrosion Under Insulation

Corrosion Under Insulation (CUI) is usually controlled by the application and maintenance of
coatings on equipment subject to CUI. A large number of papers have been published on the
subject of CUI prevention. NACE International is developing concensus guidelines on coatings for
CUI prevention. On cold-service equipment which normally operates below ambient
temperatures, the insulation system should be designed with an effective vapor barrier.
Protrusions through the insulation need to be well sealed, and the sealant maintained to avoid
moisture ingress. Insulation damage should be repaired promptly.
9.5.7
To monitor for corrosion and to warn operators of corrosive conditions, corrosion probes may be
used. Some common locations for probes are listed in Table 3-6 and 3-7. Unusual corrosion
activity should be carefully evaluated and corrective measures taken.
Page 3 - 107
TABLE 3-6
October 1999
COMMON CORROSION PROBE LOCATIONS IN SULFURIC ACID ALKYLATION UNITS
To monitor for corrosion and to warn operators of corrosive conditions, corrosion probes may be used.
Unusual corrosion activity should be carefully evaluated and corrective measures taken.
CORROSION PROBE LOCATION
PURPOSE
Effluent piping from Settler to Deisobutanizer caustic wash
Determines whether excessive acid is being carried over
Feed piping to deisobutanizer
Helps determine if caustic wash is controlling corrosion that

could be caused by acid entrainment.
Effluent piping from deisobutanizer overhead condenser
Determines whether caustic wash is controlling acidic species in

tower feed and/or neutral esters may be decomposing to SO2
Effluent piping from debutanizer reboiler
Helps determine if caustic wash is controlling corrosion or if

caustic carry-over is occurring
Effluent piping from depropanizer overhead condenser
Determines whether caustic wash is controlling SO2 esters in

refrigerant purge steam and/or neutral esters may be
decomposing to SO2.
In addition to corrosion probe monitoring, specific process steams are normally sampled for chemical analyses as a further check
on corrosion control. Typical samples are listed below in Table 3 -7.
TABLE 3 - 7
COMMON STREAM ANALYSES FOR H2SO4 ALKYLATION
SAMPLE
ANALYSIS
PURPOSE
Spent acid from acid settler
% free acid
Warns when acid concentration is low
Water draw-off from deisobutanizer

overhead drum
pH
Values under 5.5 to 6.o indicate corrosive

conditions.
Water draw-off from depropanizer

overhead drum (for systems with a
treating system on the feed to the
depropanizer.)
pH, if water present
Values under 5.5 to 6.o indicate corrosive

conditions
(i) Inspection Reaction Section

In the reactor feed system, only routine inspection of equipment is required because of the
relatively mild conditions. The reactor feed chiller and downstream piping and equipment
operated at low temperature are subject to corrosion under insulation (CUI). Regular inspections
for the insulation integrity and CUI are warranted.
In the reactors themselves, the inlet and outlet nozzles are of particular concern. The reactor and
settler are also vulnerable to CUI. Internals of reactors, particularly wear plates, feed nozzles,
mixers and impellers, and the vessel shell near mixers/impellers are of special interest. Velocity in
the settler is generally low so turbulence-related corrosion should be minimal.
Acid-containing equipment most vulnerable to corrosion includes the following:
Pump discharge piping

Hot spent acid piping
Turbulent areas such as elbows (include high points where hydrogen grooving has been
observed), tees and reducers
Page 3 - 108
October 1999
Valves, particularly control valves, check valves ,valves used for throttling and small
connections (low points)
Dead legs in piping systems
Uninsulated/uncoated cold steel
Insulation integrity
The spent acid system generally requires a comprehensive inspection program, particularly of
carbon steel equipment because it handles ambient temperature, reduced strength acid without
hydrocarbon. Although hot spent acid is much more corrosive than cold, cold systems need also
to be thoroughly inspected, particularly at turbulent areas.
Periodic onstream monitoring can be carried out using Non-Destructive Evaluation (NDE)
methods such as radiography. Larger acid-filled lines may be a limitation to obtaining wall
thickness information using conventional radiography and higher energy sources may be needed.
Cold service equipment and piping do not lend themselves well to frequent Ultrasonic Thickness
(UT) measurements because such testing would require a break in the insulation vapor barrier
which could permit water ingress, condensation and corrosion under insulation. However, colder
systems are generally less vulnerable to internal corrosion so less frequent measurements are
suitable. Often, insulation removal and visual and UT inspection are considered the most reliable
methods to inspect for CUI. However, radiographic techniques are being used that can provide
the type of wide area inspections that are needed for this form of corrosion.
(ii) Treating Section
Mixing points have been a source of accelerated corrosion in a variety of refinery services.
Inspection of the caustic and water wash mixing points for accelerated corrosion is important in
the alkylation unit. The mixing points in the deisobutanizer feed system are particularly
important. If excessive acidity is present, the mix points can experience low pH and wide pH
variability, and even materials like alloy 20 may not be suitable.
(iii) Fractionation Section
Hot equipment downstream of the caustic and water wash facilities which has not been postweld
heat treated may be vulnerable to caustic stress corrosion cracking due to caustic carry-over.
While the inspection of piping welds for such cracking is difficult, vessels such as the
deisobutanizer feed heater, deisobutanizer tower and the tower reboiler should be inspected if
caustic carry-over can occur. The DIB tower, reboiler and overhead condenser are also vulnerable
to accelerated corrosion which may be localized, due to acid carry-over or the break-down of
esters.
(iv) Refrigeration Equipment
Refrigeration equipment is naturally subject to CUI. Inspection for CUI includes radiographic
methods but the most reliable method is still often considered to be insulation removal and visual
inspection. Visual inspections should focus on insulation integrity.
Equipment should be inspected for signs of corrosion due to acid carry-over. Specific locations of
concerns include low points, the refrigeration compressor knockout drum and the refrigeration
condensers.
Page 3 - 109
October 1999
(v) Acid Tank

Inspection of acid tanks should be in accordance with 'API 653'. However, corrosive conditions in
the tank may warrant more frequent inspections than 'API 653' intervals might suggest. Key
locations for inspection are the following.
UT measure tank shell and nozzles/manways in the vicinity of and above acid inlet nozzles
Inspect vapor space for corrosion and lining integrity
Measure tank floor in the vicinity of acid outlet nozzle
Page 3 - 110
10.0
AMINE TREATING UNIT
10.1
ABSTRACT
October 1999
This section reviews the fundamental corrosion issues encountered in the alkanolamine based
process for removing acid gases from refinery gas and liquid streams. It is intended to be a
concise description of the process, including major equipment typically found in amine units, the
types of corrosion reported and the locations where corrosion may occur, as well as corrosion
monitoring techniques.
10.2
INTRODUCTION
Amine units are used throughout modern, integrated refineries to remove hydrogen sulfide,
mercaptans and carbon dioxide from hydrocarbon process streams. The amine unit typically
processes gas streams from the crude, coker, FCC and hydrotreating process units and liquid
hydrocarbon streams such as mixed C3 and C4 liquid hydrocarbons. The wide variety of feed
streams processed by the amine treater has resulted in a variety of alkanolamine based processes.
Acid gases such as hydrogen sulfide and carbon dioxide are absorbed into the amine solution by
chemical reaction resulting in a dissolved salt. This reaction is easily reversible for weak acids like
hydrogen sulfide and carbon dioxide by increasing the temperature and lowering the partial
pressure of the acid gases.
Amine units also remove many other compounds that contribute to corrosion, fouling and
reduced operating efficiency. The reaction of stronger acids like formic or thiosulfurous acid with
the amine solution is much more difficult to reverse. The resulting amine salts are called heat
stable because the acids are not removed under the normal operating conditions of the process.
10.3
TYPES OF AMINES USED

The most common amines utilized in refinery amine treaters include monoethanolamine (MEA),
diethanolamine (DEA), diisopropanolamine (DIPA), methyldiethanolamine (MDEA) and
Diglycolamine Agent (DGA). Each amine has significant advantages that may make it the
appropriate choice for a specific amine treating application. Table 3-8 presents data on the
amines discussed below.
10.3.1 Monoethanolamine (MEA)

MEA, a primary amine, removes hydrogen sulfide, carbon dioxide and mercaptans with good
efficiency. MEA is reported to degrade more rapidly than other amines such as DEA when used in
carbon dioxide service. MEA forms irreversible degradation products with CO2, CS2 and COS
which requires continuous atmospheric reclaiming. MEA does remove mercaptans such as
carbonyl sulfide and carbon disulfide measurably better than most other amines including DEA
and MDEA.
Page 3 - 111
October 1999
10.3.2 Diethanolamine (DEA)

DEA, a secondary amine, is probably the most widely employed amine and is frequently used in
refinery amine systems upstream to the sulfur plant. DEA is resistant to degradation resulting
from reactions COS, CS2 and somewhat by C02. DEA removes hydrogen sulfide, mercaptans and
carbon dioxide. DEA has the lowest hydrocarbon solubility at comparable molar concentrations
of the common alkanolamines. DEA is not reclaimed by atmospheric distillation.
10.3.3 Methyldiethanolamine (MDEA)

MDEA is a tertiary amine used most often in sulfur plant tail gas amine units that require selective
removal of hydrogen sulfide. MDEA reacts with hydrogen sulfide like any amine but reacts much
more slowly with carbon dioxide. This difference in reaction rates allows MDEA to selectively
remove more hydrogen sulfide than carbon dioxide if the unit is designed and operated properly.
MDEA is a larger amine molecule and thus a higher weight concentration must be used to
achieve similar treating capacity. MDEA has poor absorption of carbonyl sulfide and carbon
disulfide and significant hydrocarbon solubility.
10.3.4 Diisopropanolamine (DIPA)

DIPA is a secondary amine which was the first to be utilized commercially to selectively remove
hydrogen sulfide. Many of the amine treaters using DIPA have switched to MDEA. DIPA is used
primarily in sulfur plant tail gas units. DIPA is reported to degrade rapidly enough in carbon
dioxide service to require frequent reclaiming by vacuum distillation.
10.3.5 Diglycolamine Agent (DGA)

DGA is a primary amine that is utilized when carbonyl sulfide and carbon disulfide, in addition to
hydrogen sulfide and carbon dioxide must be removed to low concentrations. DGA is reclaimed
by atmospheric distillation when in CO2, COS or CS2 service. DGA is similar to MEA in many ways
but is typically used at higher molar concentrations due to a lower vapor pressure.
10.3.6 Specialty Amines

Specialty amines are increasingly popular because they can often provide increased performance
or meet unique needs compared to a single amine. Specialty amines are usually blends of MDEA
and other amines often with additives to enhance performance characteristics.
All of the alkanolamines discussed react directly with hydrogen sulfide, as listed in Table 3-8,
(Equations 1 and 4). The reaction between the alkanolamines and carbon dioxide differ in that all
of the amines will react with carbonic acid (Equations 2 and 5) but MDEA and other tertiary
amines will not react by the carbamate mechanism (Equation 3). The additional time required for
carbon dioxide to dissociate to carbonic acid and then react is the principle reason MDEA can be
used to selectively remove hydrogen sulfide while absorbing less carbon dioxide. The selective
absorption of hydrogen sulfide is often measured as CO2 Slip, the percentage of carbon dioxide
that is not absorbed.
Caution is required when comparing corrosion rate data of the different amines due to the large
difference in molecular weights. Laboratory results comparing MEA at 20 wt% with MDEA at 30
wt% would not be a reasonable comparison unless the selectivity of MDEA warranted the use of
Page 3 - 112
October 1999
a 25% less active solution. Ideally, testing should be conducted at equal molar, equal acid gas
removal or actual plant concentrations.
10.4
REFINERY AMINE UNIT PROCESS DESCRIPTION

Figure 3-13 presents a generalized flow diagram for a refinery primary amine system. This system
contains a fuel gas absorber, a liquid treater, a hydrogen recycle absorber, one amine stripper or
regenerator, a flash drum or more precisely, a three phase separator, in addition to the required
pumps, heat exchangers and other associated equipment.
Gas absorbers typically have 20 trays or an equivalent amount of packing unless the absorber is
designed specifically for H2S selectivity. The feed gas enters the absorber through a distributor at
the bottom of the vessel. The preferred feed gas temperature range is 27-38oC (80o to 100oF) but
temperatures as high as 54oC (130oF) are common. The gas moves upward through the
absorber. The amine enters the absorber vessel near the top as well as through a distributor. The
amine flows downward (counter current) across the trays and comes in contact with the gas
stream, absorbing the acid gases species in the process. The amine becomes enriched with acid
gas and is usually called 'Rich Amine', although other names like 'Fat Amine' are also used.
Liquid treaters may have trays or packing but usually have fewer equivalent stages. Liquid treaters
must balance the need for good mixing, and thus good acid absorption, with the need for good
phase separation. The tower liquid level is usually maintained near the top with the bulk phase
amine but the reverse is also practiced. Liquid treaters are operated at temperatures in the 49o60oC (120o to 140oF) range. The hydrocarbon feed, treated hydrocarbon product, amine feed
and rich amine enter and exit the treater as described for gas absorbers.
The rich amine from several absorbers is commingled together in the flash drum. The flash drum
pressure is reduced to flash soluble light hydrocarbons and aid in the removal of entrained or
condensed hydrocarbons. The flash gas is usually returned to the refinery fuel gas system. A flash
drum with liquid separation internals is typically called a three phase separator. Entrained
hydrocarbons can be skimmed from the rich amine to reduce hydrocarbon build up in the
stripper or amine filters.
The pressure of the flash drum determines if a rich amine pump is required to move the rich
amine to the stripper. The rich amine solution is pressured or pumped through the Lean/Rich
cross heat exchanger(s). The rich amine is tube side to reduce local pressure changes that may
cause flashing of the amine. The rich amine is typically heated to 82o to 99oC (180o to 210oF).
After the cross exchanger, the hot rich amine is flashed to the top section of the amine stripper.
The reduction in pressure and the application of heat in the regenerator strips the acid gas from
the amine as the solution flows down the regenerator. Increasing the temperature slows and then
reverses the absorption reactions. Acid gas is liberated as the reaction equilibrium is shifted from
the salt to amine and acid gas. The steam stripping reduces the vapor pressure of the acid gas in
the vapor phase further shifting the reaction equilibrium to amine and acid gas.
The stripper overhead stream consists of steam and acid gas at a temperature of about 99o
to113oC (210o to 235oF). Most designs utilize a conventional condenser and accumulator. The
steam is condensed and returned to the stripper as reflux, typically at 43o to 60oC (110o to
140oF). The amine solution flows down the stripper to the reboiler which operates at 110o to
127oC (230o to 260oF). The majority of the hydrogen sulfide and carbon dioxide should be
removed from the amine prior to reaching the reboiler. The reboiler heating medium should be
Page 3 - 113
October 1999
limited to 149oC (300oF) to reduce amine thermal degradation, 65 psia saturated steam is
typically specified.
The amine exiting the bottom of the stripper tower is called 'Lean Amine' due to the low
concentration of acid gas remaining. The hot lean amine is cooled in the Lean/Rich cross
exchanger(s) to about 17o to 88oC (170o to 190oF). Additional cooling is provided in the lean
amine coolers prior to entry into the absorber(s). The lean amine temperature entering a gas
treater/absorber should be 2.7o to 6oC (5o to 10o F) warmer than the feed gas to reduce
hydrocarbon condensation. The lean amine feed to liquid treaters is often about 54oC (130oF).
The amine is filtered by both particulate and carbon unless the amine reclaimer is used routinely
(for MEA, DIPA and DGA). Older designs utilize filtration principally on the lean amine perhaps
due to the concern over potential overexposure to H2S. Newer amine designs utilize rich amine
filtration which appears to be more effective. Some designs utilize filtration on both lean and rich
amine streams. The particulate filters remove corrosion products and other solids. Carbon filters
remove surfactants and hydrocarbons. Neither filter is particularly effective at removing water
soluble organic acids.
Knockout pots located prior to the absorber are included in many designs to remove entrained
water and liquid hydrocarbons from feed or product streams. Knockout pots on the feed gas can
be useful in removing water soluble organic and inorganic acids.
Tail Gas amine Units (TGU) are the last opportunity to remove hydrogen sulfide and other sulfur
species from the sulfur plant off gas before it enters the atmosphere. Figure 3-14 is a generalized
process flow diagram for a TGU. The most significant differences in the primary amine system
and a TGU are:
1) the very low pressure of the TGU absorber,
2) the composition of the feed gas and
3) the need to achieve a very low hydrogen sulfide content in the treated gas.
The TGU absorber operates typically at 5 psig or less as compared to primary amine unit
absorbers which range from about 50 psig up to over 500 psig (7.25-72.5 kPa). The lower
operating pressure reduces the amount of acid gas that will be absorbed per mole of amine, thus
the rich loading is lower for a TGU than most other systems.
The TGU feed gas contains primarily hydrogen, nitrogen, carbon dioxide, hydrogen sulfide and
traces of other species. Because the acid gas off the TGU stripper is recycled to the sulfur plant,
absorbing the carbon dioxide is not desired. This means that the TGU is the only amine system
that universally will benefit from a selective amine solvent such as MDEA.
Page 3 - 114
REFINERY PRIMARY AMINE SYSTEM

with MULTIPLE ABSORBERS
Figure 3-13
QUENCH TOWER and TAIL

GAS UNIT
Acid Gas
Recycle to
Claus Reactor
Figure 3-14
Sour Tail Gas
Treated Tail Gas

to Incinerator
Quench Water
pH 7.5 - 8.5
Quench
Tower
Lean
Cooler
Absorber/
Contactor
Stripper
Lean
Filter
Sulfur Plant
Tail Gas
Reboiler
Lean/Rich
Exchanger
Rich Amine
Water Blow
Down
Water Make-Up & Caustic or

Ammonia for pH Control
Rich Filler
Lean Amine
10.5
October 1999
CORROSION PHENOMENA
PHENOMENA
Most of the variables that influence corrosion rates in other process systems are important in
amine systems as well. These variables include temperature, velocity and concentration of
corrosive species. These three variables contribute to corrosion phenomena described in amine
literature as acid gas flashing, heat stable salts and amine degradation. Flashing acid gas is a
reoccurring theme in amine systems because this allows for the local metal surface to have a
significantly lower pH than the bulk solution and because flashing can occur at pressures and
temperatures lower than the boiling point of the solution. Flashing is caused by temperature
increases or pressure reductions that upset the acid gas/amine reaction equilibrium.
TABLE 3-8
ACID GAS ABSORPTION REACTIONS
Primary and Secondary Amines (DEA shown)

(HOCH2CH2) 2NH +H2S <=> (HOCH2CH2) 2NH2+ + HS
1
-
(HOCH2CH2) 2NH +H2S + CO2 <=> (HOCH2CH2) 2NH2+ + HCO3S
2(HOCH2CH2) 2NH + CO2 <=> (HOCH2CH2) 2NCO2+ + (HOCH2CH2) 2NH2
Tertiary Amines (MDEA shown)

-
(HOCH2CH2) 2NCH3 +H2S <=> ( (HOCH2CH2) 2NCH3) H+ + HS
(HOCH2CH2) 2NCH3 +H2O + CO2 <=> ((HOCH2CH2) 2NCH3) H+ + HCO3S
The high pH of the amine solvent creates a relatively non-corrosive environment for carbon steel
in most areas in the amine unit. The most severe corrosion is acidic in nature caused by localized
areas of depressed pH. The addition of hydrogen sulfide, carbon dioxide and stronger acids to
the amine solution reduces the pH.
Corrosion can be aggressive in the high temperature areas of the unit. As the rich amine is
heated in the Lean/Rich cross exchanger, the chemical equilibrium between the amine, acid gas
and the salt is shifted away from the salt as described earlier. Acid gas flashing is more likely to
occur with high rich amine loadings and a reduction in pressure. Flashing acid gas produces a
vapor phase containing little amine to prevent a low pH situation at the location of
recondensation.
Velocity has both a direct and indirect impact on corrosion. Increasing velocity directly increases
corrosion by physically damaging protective iron sulfide scales and increasing the effective
concentration of corrosive specie. Physical damage to protective scales increases with the
presence of solids or two phase flow in the amine solution.
Page 3 - 117
October 1999
Indirectly, increasing velocity creates areas of higher and lower local pressure caused by flow.
Local changes in pressure, differential pressure, create the environment for acids to flash from the
bulk aqueous amine solution. The collapse of the vapor phase in piping, pumps and exchangers
contributes to the physical damage caused by flow in the form of cavitation or impingement.
The areas of highest corrosion potential are the reboiler, hot lean amine piping, Lean/Rich cross
exchanger, hot rich amine piping, stripper and stripper overhead. Acid gas flashing is possible in
all of these areas due to the high temperature and potential for fluctuating differential pressure
caused by flow.
10.6
CORROSIVE SPECIES
SPECIES
Most corrosion in an amine system is acidic in nature. Acids that enter the system include carbon
dioxide, hydrogen sulfide and a variety of stronger acids. A listing of the most common corrosive
specie are listed in Table 3-9. Carbon dioxide and hydrogen sulfide are the corrosive components
in the highest concentration. Carbon dioxide has a pKa only slightly lower than H2S, but
experience has proven it to be much more corrosive. Table 3 -10 presents the corrosion reactions
discussed below.
TABLE 3 - 9
CHEMICAL DATA ON SELECTED SUBSTANCES

Mole Wt.
pKa
(25oC)
Formula
Hydrogen Chloride
36.46
-6.1
HCl
Sulfuric Acid
96.06
-3
H2SO4
Thiocyanic Acid
59.09
-1.85
HSCN
Thiosulfurous Acid
114.14
0.60
H2S2O3
Oxalic Acid
90.02
1.27
HOOCCOOH
Sulfur Dioxide
64.06
1.89
SO2 + H2O
Formic Acid
46.02
3.75
HCOOH
Glycollic Acid
76.03
3.83
CH2OHOOH
Acetic Acid
60.03
4.76
CH3COOH
Butyric Acid
88.11
4.82
CH3CH2CH2COOH
Propionic Acid
74.05
4.87
CH3CH2COOH
Carbon Dioxide
44.01
6.36
CO2
Hydrogen Sulfide
34.08
6.97
H2 S
Methyl diethanolamine (MDEA)
119.16
8.56
(HOCH2CH2) 2NCH3
Diethanolamine (DEA)
105.14
8.90
(HOCH2CH2) 2NH
Diisopropanolamine (DIPA)
133.19
8.97
(HOCH2CH2CH2) 2NH
Hydrogen Cyanide
27.02
9.21
HCN
Ammonia Ion
18.04
9.24
NH4 +1
Diglycolamine (DGA)
105.14
9.50
H(OCH2CH2) 2NH2
Monoethanolamine (MEA)
61.08
9.52
HOCH2CH2NH2
Calcium Ion
40.08
12.8
Ca +2
Sodium Ion
22.99
14.8
Na +1
Page 3 - 118
TABLE 3 - 10
October 1999
CORROSION REACTIONS IN AMINE SYSTEMS
CARBON DIOXIDE
C02 + H20 2(HO)~C=O H+ + HCO3
Fe + 2H2C03 Fe(HCO3)2 + H2
Fe(HCO3)2 + H20 Fe(OH)2 + C02
Fe + (C02 + H20) FeCO3
(6)
(7)
(8)
(9)
HYDROGEN SULFIDE
Fe Fe+2 + 2e
H2S H+ + HS
Fe+2 + HS- (FeSH)+
(FeSH)+ + HS- (HS-Fe-SH)
(HS-Fe-SH) + (FeSH)+ (HS-Fe-S-Fe-SH)
(HS-Fe-S-Fe-SH) + HCI (HS-Fe-S-Fe-CI) + H2S
Fe + H2S FeS + H2
OXYGEN (17)
3Fe + 4H2O Fe3O4+4H2
4Fe +302 + 6H20 4Fe(OH)3
2Fe(OH)3 Fe2O3 + 3H20
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(magnetite) (17)
(ferric hydrozide) (18)
(ferric oxide) (19)
ORGANIC ACIDS
Fe(OH)3 + 2(CH3COOH) Fe(C2H302)20H + 2H20
Fe + 2(CH3COOH) Fe(CH3COO)2 + H2
INORGANIC ACIDS
2Fe(OH)3 + 3(H+ + HSO4) Fe2(S04)3 + 6H20
Fe + H2SO4 FeSO4 + H2
FeS + 6HCN Fe(CN)6 -4 + H2S + 4H+
Fe + 2HSCN Fe(SCN)2 + H2
2Fe + 6HSCN Fe2(SCN)6 + 3H2
(ferric acetate) (20)

(ferrous acetate) (21)
(ferric sulfate) (22)

(ferrous sulfate) (23)
(24)
(ferrous thiocyanate) (25)
(ferric thiocyanate) (26)
HEAT STABLE SALTS (HSS)

HCOO + H+ + (HOCH2CH2)2NH (HOCH2CH2)2NH2+ + HCOO
(27)
(HOCH2CH2)2NH + H+ + HS04- (HOCH2CH2)2NH2+ + HS04-
(28)
STRONG BASE CONTAMINATION

4NaOH + Fe3O4 Na2FeO2 + 2NaFeO2 + 2H20
2NaOH + FeS Fe(OH)2 + Na2S
(29)
(30)
AMMONIA
NH3 + H2S NH4+ + HS
(ammonia bisulfide) (31)
Page 3 - 119
October 1999
Carbon dioxide dissolves into the amine solution and forms carbonic acid. Carbonic acid reacts
with iron to form iron carbonate which usually does not protect the metal from continued
corrosion in these systems. Systems processing little or no hydrogen sulfide will produce
corrosion products composed of iron carbonates, iron oxides and iron hydroxides.
Hydrogen sulfide reacts with iron to form a scale that can be protective against continued
corrosion in many systems. The reaction of iron and hydrogen sulfide is thought to proceed by
the mechanism proposed by R.H. Hausler and N.D. Coble. The mechanism proposes that an iron
sulfide "polymer" forms with a bisulfide ion terminating each end. The inclusion of another ion,
such as chloride, in the structure terminates its growth and if present in high enough
concentrations, destroy the scale. This mechanism provides a reasonable explanation of the
impact of chloride and other ions over a wide range of pH values.
Magnetite scales have been reported in literature. Magnetite can form a protective scale as in
Equation 17. Magnetite is naturally occurring in scales and deposits in systems processing carbon
dioxide but not hydrogen sulfide. Protective magnetite scales are formed at temperatures above
about 110o to116oC (230o to 240oF). Literature suggests that magnetite is always present in case
studies of intergranular stress corrosion cracking and would be necessary for a film rupture
mechanism of cracking.
Oxygen reacts with all of the alkanolamines resulting in the formation of formic acid, oxalic acid
and to a lesser extent, acetic acid . High concentrations of oxygen in the amine may also lead to
the formation of iron oxides and iron hydroxides. Amine sumps and makeup water are a
traditional means of oxygen entry into the amine system. Many feed gases contain low
concentrations of oxygen from the upstream process such as the FCC. The elimination of oxygen
from amine unit feeds, make up water and amine storage will reduce amine degradation and
corrosion issues due to oxygen.
The contamination of the amine solution with acids stronger than hydrogen sulfide or carbon
dioxide changes the corrosion picture entirely. Stronger acid salts are not efficiently removed in
the regenerator and reach the stripper bottoms, reboiler and hot lean piping. These acids form
Heat Stable Amine Salts (HSAS) with the amine because they are essentially not removed by
normal unit operation. These acids include the organic acids of formate, acetate, oxalate and the
inorganic acids of chloride, sulfate, cyanide, sulfur dioxide, thiosulfate and thiocyanate.
The organic acids are reported in literature to form from reactions of the amine and oxygen, from
reactions between carbon dioxide, carbon monoxide, oxygen and light hydrocarbons. These
acids are also found in crude oils and many crude oil fractions including amine unit feed streams.
The reaction of these acids with iron creates a water soluble corrosion product.
Inorganic acids may also be in the feed streams or be formed in the amine solution. Most liquid
and gas feeds to the amine system contain small concentrations of chloride, sulfate, thiosulfate,
cyanides and thiocyanides. Feeds from the crude, vacuum, coker and fluid units are have been
identified as containing these species. These acids react with iron and iron corrosion products like
iron sulfide forming both water soluble and insoluble iron salts.
HSAS represent a loss of gas treating capacity because the amine cannot absorb acid gas if
reacted with another stronger acid. The term 'Heat Stable' is not entirely true because a
significant quantity of these salts dissociate in the stripper and the acids can be found in the
stripper overhead. The acids dissolve in the condensing steam in the overhead and return to the
regenerator tower in the reflux. Systems that have a top pump around rather than the traditional
overhead condenser will accumulate large quantities of strong acids in the circulating water.
Page 3 - 120
October 1999
Several rules of thumb are used for the maximum allowable concentration of HSAS before
corrective action is required. The oldest rule sets 2% HSAS (as amine) as the safe limit before the
onset of problems. Later, a second rule allowed 10% of the amine concentration to be the safe
limit. This second rule appears to be a modification of the first and was developed for DEA and
other amines as they became popular commercially. Experience indicates that DEA can usually
operate with HSAS up to 10% of the amine concentration (28% DEA*O.1 = 2.8 wt% HSS as DEA)
but MDEA should be limited to a lower concentration.
HSAS concentration is most often controlled by normal amine solution losses, amine reclamation
or intentional purging of amine solution or reflux. HSAS are selectively removed from the amine
solution by atmospheric distillation, vacuum distillation, ion exchange and proprietary methods.
All of these practices are beneficial from a corrosion point of view.
Neutralization of the HSAS insitu is a common industry practice to restore acid gas removal
capacity. The impact of neutralized HSS on corrosion is a current industry concern. Adding a
strong base like sodium hydroxide will restore acid gas removal capacity but the salt remains in
the solution.
Excessive amounts of a strong base can damage or destroy the iron sulfide scale protecting the
metal as shown in Equations 29 and 30. This will result in a high solids content in the solution
but usually this requires a period of several weeks to months. Most literature sources indicate that
neutralization of HSAS with a strong base reduced corrosion but at least one example of
increased corrosive attack of stainless steel is reported.
HSAS concentrations can be measured by a titration method or by measuring each acid directly
utilizing ion chromatography. The titration method does not report the amount of HSAS
neutralized by a strong base, ion chromatography measures the concentration of each anion
which allows for improved trouble shooting should the rate of HSAS formation increase
dramatically. Ion chromatography reports the neutralized acids and HSAS.
Excessive stripping of the acid gas can result in high corrosion rates. Excessive stripping removes
the acid gas to very low levels. The hot lean amine may contain too little hydrogen sulfide to
keep the protective iron sulfide scale intact. The exposed metal can be aggressively attacked by
other acids in the solution. Excessive stripping and corrosion in the hot lean amine piping is more
common with MDEA than other amines.
Corrosion in stripper overhead systems is caused by carbon dioxide, hydrogen sulfide and other
acids dissolved in the condensing steam. Ammonia and a low concentration of amine is normally
present in the overhead, keeping the pH from dropping too low. Large concentrations of
ammonium bisulfide can accumulate in the overhead condensate producing many of the
problems found in hydrotreater reactor effluent. Purging the reflux is a common practice to
remove ammonia and some of the acidic species.
10.7
AMINE DEGRADATION
DEGRADATION
Degradation of alkanolamines occurs as the result of exposure to high temperatures and
compounds such as oxygen and carbon monoxide. Oxygen formed degradation products
include organic acids like formic and oxalic acid.
MEA, DIPA, DGA and DEA also form chemical degradation products in systems containing
carbonyl sulfide (COS), carbon disulfide (CS2), and carbon dioxide. C02 Degradation of DEA is
well documented and will not be repeated here. Literature sources indicate that many of these
Page 3 - 121
October 1999
degradation products are chelation agents that may be strong enough to chemically remove iron
from iron oxide or sulfide scales and perhaps directly from the base metal. Most degradation
products are basic and have minimal impact on corrosion.
10.8
CRACKING PHENOMENA
PHENOMENA
Cracking in amine systems is due to two primary phenomena, carbonate stress corrosion cracking
and hydrogen induced cracking. Variations on these two mechanisms to cover different concerns
about metal hardness, stress loading, purity have yielded a number of names like SSC, HIC and
SOHIC.
Detailed discussion of the causes and variations of cracking in amine systems is beyond the scope
of this section. Hydrogen blistering is caused by the accumulation of molecular hydrogen in the
steel. Atomic hydrogen from corrosion reactions penetrates the steel and combines at inclusions
or irregularities. Increasing quantities of molecular hydrogen create pressure in the alloy until a
blister is formed. Hydrogen Induced Cracking (HIC) is a type of cracking in which small blisters
link up. Hydrogen Embrittlement involves the same process but occurs in high strength steels or
in areas of high localized stress. The term Hydrogen Embrittlement has fallen out of favor for
more specific terminology for different cracking mechanisms. Sulfide Stress Cracking (SSC) is a
form of hydrogen embrittlement in steels of high hardness or strength. Stress Oriented Hydrogen
Induced Cracking (SOHIC) in which the blisters align perpendicular to the stress.
Reducing corrosion activity and using clean steels may be the most effective methods of reducing
hydrogen blistering and HIC. Post weld heat treatment (PWHT) is recommended to reduce
hydrogen embrittlement, SSC and SOHIC. The use of clean steel and more extensive use of post
weld heat treatment have become the most common practices to reduce the incidence of
cracking. Stress corrosion cracking (SCC) is controlled by PWHT (carbonate cracking) and
avoiding exposure of inappropriate alloys to a specific environment.
10.9
CORROSION INHIBITORS
INHIBITORS
Corrosion inhibitors for amine systems consist of two basic types, film forming or passivating.
Film forming inhibitors are organic compounds or mixture of compounds that attach themselves
to the metal surfaces of the system, thus forming a protective barrier. Passivation type inhibitors
react with the metal and local environment to form a protective scale.
Filming inhibitors used in amine systems must be compatible with the amine solution and
process. Amine systems are unique when compared to most other processing units. The pH of
the system ranges from near neutral (7) in the reboiler to the 11 - 12.5 range at the top of the
absorber. Amine systems also will cause a corrosion inhibitor to cycle up as little opportunity
exists for the product to leave the system. Foaming can be a problem if the product is not
correctly formulated. Any inhibitor considered for amine unit service should be tested at the
suppliers recommended dosage and at several times the recommended dosage.
Organic filming inhibitors were first used over 20 years ago and some have been patented for
their special ability to protect amine systems from corrosion without causing adverse effects such
as foaming.
Page 3 - 122
October 1999
Research has shown that some inhibitors are very effective in producing more protective iron
sulfide scales. Organic filming inhibitors may also act as dispersants to existing deposits in the
system. A rapid increase in solution solids content may occur after initial injection of one of these
products.
Inhibitor injection location will vary from unit to unit but injection into the regenerator overhead
with a slip stream of reflux is common. Filming inhibitors have low volatility and will not protect
the overhead if not injected at that location. Inhibitor feed rate and concentration should be
determined by monitoring corrosion activity.
Passivation inhibitors rely on the formation of a protective scale formed insitu on the metal
surface. Older formulations included sodium metavanadate and compounds of arsenic, tin and
bismuth. These inhibitors should be used in systems processing only H2S or C02. The
environmental problems of these inhibitors has nearly eliminated their use.
Newer formulations utilize oxygen scavengers to form magnetite (Fe304). Magnetite scales can be
very protective if formed under the proper conditions. Under other conditions, such as at too low
a temperature, the magnetite formed is not protective. Repeated formation and spalling of the
scale results in increased corrosion rates.
Oxygen scavengers, and passivators in general, do not form protective scales at temperatures
lower than 110o to116oC (225-240oF). The magnetite scale also must compete with the
formation of iron sulfide and carbonate. Galvanic type attack has been reported at the
boundaries between the different type of scale. Oxygen scavengers are occasionally used to
remove oxygen that can react to form organic acids, thiosulfates or sulfates.
10.10 MATERIALS OF CONSTRUCTION

CONSTRUCTION
Carbon steel is the most prevalent alloy used in construction of amine systems. Carbon steel has
provided good service in many units and failed quickly in others. Carbon steel remains the
material of choice for exchangers shells, vessels and most exchanger bundles and piping. Tower
trays, packing and fasteners are usually made of 410, 304 or 316 stainless steel but occasionally
polypropylene or ceramic packing is used. Pumps utilizing cast iron housings and 316 or 317
stainless steel impellers are most commonly specified. Flow control equipment is usually in
carbon steel bodies with internals of 316 or 17-4PH.
Reboiler bundles of carbon steel are common but 316 or 304 are increasingly prevalent. If
replacing a carbon steel bundle with stainless steel, be careful to ensure that the lower
conductivity of stainless steel does not require a higher heat flux than recommended. Lean/Rich
cross exchangers are usually 304 or 316 but carbon steel was more prevalent just a few years
ago. Lean amine coolers are usually carbon steel. Stripper overhead condensers are usually
carbon steel but stainless steel and titanium have also been used.
10.11 CORROSION MONITORING

MONITORING
Corrosion has been successfully monitored with corrosion coupons, electrical resistance probes,
linear polarization probes, hydrogen finger and patch probes and many other methods.
Corrosion coupons should be used to insure that electrical resistance or linear polarization probes
are working properly.
Page 3 - 123
October 1999
Corrosion coupons or insertion type probes are typically used to monitor corrosion in the hot
areas of the unit reboiler feed line, hot lean amine piping, hot rich amine piping and the stripper
overhead condenser.
Amine solution analysis has been used to monitor metals concentrations but is no more reliable.
Even a minor unit upset can dramatically increase or decrease the amount of soluble or insoluble
iron present in the amine solution on any given day. The amine solution should be completely
acidified to insure a good iron analysis is achieved. Sampling technique and frequency may limit
the usefulness of amine analysis as a corrosion monitoring tool.
10.12 CORROSION CONTROL

CONTROL MEASURES
The proper operation of the amine system is the most effective method to control corrosion. The
key variables which contribute to corrosion include: temperature, differential pressure, acid gas
loading, HSAS concentration, amine concentration, solution velocity and heat flux.
1. Adequate feed preparation is one of the most effective methods to reduce amine
contamination and corrosion. An inlet separator can be used to remove entrained water from
the inlet gas. Injecting water upstream to the separator allows for many of the HSS forming
acids to be removed from the feed. Liquid feeds often contain more acidic species than gas
flows. A water wash injected prior to the inlet separator should use the best quality water
available. Adding caustic to the water wash will increase the removal of strong acids. The use
of sour water as the water source may increase the acid concentration in the feed gas.
Stripped sour water is often not of acceptable quality.
2. The temperature of the amine should not exceed 127oC (260oF). Reboiler bulk temperatures
in excess of 129oC (265oF) can result in aggressive corrosion and thermal degradation of the
amine. Steam pressures in excess of 65 psia, saturated can result in high metal surface
temperatures and corrosion. The steam control valve should be on the steam inlet to allow
the condensate to continuously drain from the tubes. Temperatures above about 98o
to104oC (210o to 220oF) in the rich amine can result in acid gas flashing and severe localized
corrosion.
3. The reboiler steam rate should be operated to maintain a regenerator top temperature
adjusted for variable overhead pressure. The use of excessive steam rates accelerates weak
acid corrosion in the reboiler and lean piping and also increases velocities in the stripping
tower. Insufficient steam rates also increase corrosion activity because higher acid gas
concentrations reach the stripper bottoms and reboiler. Many control systems target a
stripper top temperature without correction for changing pressure. Manual adjustments can
be made by using a reflux ratio graph.
4. Local pressure changes must be minimized by design and operation. The net positive suction
head to pumps must include a large margin of safety to prevent acid gas flashing. Piping
should be designed to minimize turbulence. Areas that must be exposed to large pressure
changes should be made of stainless steel, preferably Type 316L or better.
5. The optimal combination of amine concentration, circulation rate and acid gas loadings
should used to minimize corrosion. Corrosion will generally be lower if amine concentration
is increased prior to increasing rich loading or circulation rate. Rich amine acid gas loadings
should be controlled to make product specification and to control corrosion. Increasing
circulation rate increases velocity, local pressure changes and steam requirement. Major
Page 3 - 124
October 1999
problems occur when the unit is at maximum amine concentration, circulation rate and rich
loading. Corrosion inhibitors may allow continued operation in these situations with
acceptable corrosion rates.
6. Solids should be removed from the system by particulate filtration. Filter pore sizes of 10 to
20 microns are generally specified but sizes ranging from 75 to 5 microns have been
effective. Total flow filtration is required on some units to maintain the solution in good
condition. Rich filtration has the potential to be much more effective than lean amine
filtration. Filters should be changed based on differential pressure but the system should be
designed to last about 2 weeks. Systems requiring filter changes more frequently may benefit
from corrosion inhibitors.
7. Carbon filtration removes hydrocarbons from the amine solution. Hydrocarbons cause
foaming that may contribute to fouling and foaming. Carbon filtration is ineffective in
removing HSAS.
8. Oxygen entry into the system should be eliminated. Routine testing of the make up water
should confirm the absence of oxygen. Amine storage and sump tanks should be protected
from oxygen entry. The entrance of oxygen into upstream processes may need to be
investigated and corrected.
9. Heat Stable Amine Salts should be controlled to a concentration that is economically
acceptable, balancing removal costs with equipment cost. Corrosion rates and acid gas
removal performance should be monitored in determining the acceptable concentration of
HSAS.
10. MEA, DIPA and DGA systems should utilize the reclaiming equipment on a continual basis.
Other amines can be reclaimed or purged by a number of methods. Caustic addition into the
amine solution will restore the acid gas treating capacity but increase the salt and ash
content.
11. Corrosion inhibitors can be used to reduce many corrosion and fouling problems if used and
monitored correctly. Inhibitor selection should include screening for compatibility with the
amine solution.
12. Fabrication methods should include PWHT in all locations. Materials selection should stress
clean steels and good welding practices. Stainless steels should be considered for areas of
high heat transfer, velocity and pressure changes. Trays or packing should be made of
stainless steel.
13. Routinely review the operation of upstream equipment to minimize the introduction of
strong acids to the system. Changes in crude overhead water washes can increase or
decrease the amount of organic acids in the fuel gas feed. Changes in FCC catalyst circulation
can alter the amount of carbon dioxide and other acids in the FCC off gas.
Page 3 - 125
11.0
October 1999
MEROX TREATING UNITS

The Merox process is a catalytic process developed for the chemical treatment of petroleum
distillates for removal of sulfur present as mercaptans (Merox Extraction) of conversion of
mercaptan sulfur to another much less objectionable form (Merox Sweetening). Depending upon
the application, extraction and sweetening can be used either separately or in combination. The
process uses a catalyst in a alkaline environment to promote the oxidation of mercaptans to
disulfides using air as the source of oxygen. The reaction proceeds at a practical rate at normal
temperatures for refinery rundown streams. For light stocks, operating pressure is controlled
slightly above the bubble point to assure liquid phase operation. For heavier stocks, the reaction
section operating pressure is normally set to keep air dissolved in the hydrocarbon phase.
There are essentially four different versions of the Merox process:
1. Merox Extraction (Figure 3-15)
2. Merox Liquid/Liquid Sweetening (Figure 3-16)
3. Merox Conventional Fixed Bed Sweetening (Figure 3-17)
4. Merox Minalk Fixed Bed Sweetening (Figure 3-18)
Some units require pretreatment for removal of contaminants such as H2S or post treatment to
eliminate trace carryover of caustic or corrosion inhibitors to tankage.
11.1
MEROX EXTRACTION
EXTRACTION
In mercaptan extraction units, LPG or light gasoline fraction is charged to an extractor column in
which mercaptans are countercurrently extracted by a caustic stream containing Merox catalyst.
The treated material passes overhead to a settler in which any entrained caustic solution is
separated and returned to the circulation system.
The rich caustic solution from the bottom of the extractor column flows to the regeneration
section. Air is injected into this stream and the mixture flows upward through the oxidizer where
the mercaptans are catalytically converted to disulfides by the dissolved Merox catalyst. The
oxidizer effluent flows into the disulfide separator where spent air, disulfide oil and the caustic
solution are separated. Spent air and disulfide oil is sent to disposal. The regenerated caustic
stream is returned to the extractor column.
11.2
MEROX LIQUID/LIQUID
LIQUID/LIQUID SWEETENING
In sweetening units, hydrocarbon, air and aqueous caustic soda containing Merox catalyst are
simultaneously contacted in a mixer, where mercaptans are converted to disulfides. Mixer
effluent is directed to a settler from which the treated hydrocarbon stream is separated from the
solutions. The disulfides and spent air remain in the product.
11.2.1 Merox Solid Bed Sweetening

Fixed bed sweetening is employed with charge stocks having endpoints above 135o to150oC
(275o to 300oF). With these stocks, the heavier mercaptan types are only partially soluble in
caustic solution. Mercaptans are converted to disulfides in a fixed bed reactor system. The
reactor contains specially selected activated charcoal impregnated with Merox catalyst and the
Page 3 - 126
October 1999
bed is periodically wetted with caustic solution. Air is injected into the feed hydrocarbon stream
ahead of the reactor to convert the mercaptans to disulfides. Spent air leaves with the treated
product. The reactor is followed by a settler which also serves as a caustic reservoir. A caustic
circulation pump is provided to circulate caustic from the settler to wet the catalyst bed.
11.2.2 Merox Minalk Sweetening

This version of the Merox process began in the early 1970s. Sweetening is accomplished with
charge stock end points below 135 o to 150 oC (275o to 302oF). The Merox Minalk process relies
on a small controlled continous injection of alkali solution, rather than gross alkali saturation of
the catalyst bed as in Merox fixed bed sweetening. The basic flow scheme is similar to that of the
conventional fixed bed unit, but involves a minimum of equipment.
11.2.3 Pretreatment
A caustic scrubber is employed with some units to remove H2S which will cause excessive
consumption of caustic and Merox catalyst. The caustic scrubber also removes naphthenic acids
that would form soaps which will coat the catalyst and block the pores of solid bed units.
Contaminants which are not removed by the caustic scrubber but have an effect on Merox
catalyst activity are corrosion inhibitors and antifoulants injected in processes upstream of the
Merox treater.
Laboratory studies have shown that at concentrations above a few ppm, corrosion inhibitors will
deactivate the sweetening catalyst. While some inhibitors are worse than others, the overall effect
is the reduction of catalytic activity of the sweetening catalyst as shown in the table below. The
feed RSH was 500 ppm for each test.
TABLE 3-11
Corrosion Inhibitor
ACTIVITY OF SWEETENING CATALYST

Inhibitor Concentration
none
RSH in Product
28 wt ppm
10 wt ppm
106 wt ppm
50 wt ppm
225 wt ppm
5 wt ppm
100 wt ppm
10 wt ppm
45 wt ppm
10 wt ppm
53 wt ppm
As can be seen, the product mercaptan content for these tests was drastically increased at only 510 wt ppm and at 50 ppm the catalyst was 50% deactivated. Thus, whenever a corrosion
inhibitor must be used upstream of a Merox treater, it should be tested to measure its effect on
the sweetening catalyst and should be used only at minimum levels.
Page 3 - 127
October 1999
11.2.4 Post Treatment

Post treatment varies with the product specifications and end use but normally consists only of a
sand filter to coalesce any caustic haze that might be present. Extensive post treating is needed
for jet fuels. This consists of a water wash to remove entrained caustic solution and water soluble
surfactants, including sodium naphthenates and corrosion inhibitors added upstream of the
Merox treater. The additives will prevent the jet fuel from passing the water reaction or WSIM
test.
11.3
MATERIAL OF CONSTRUCTION
CONSTRUCTION
Merox units are constructed of carbon steel. Because the unit is normally operated in the range
of 21o-43 oC (70 o -110 oF), in most refineries, the carbon steel is not stress relieved. If heat tracing
is used on pipe or vessels in cold climates, a standoff is used to prevent the hot tracing
equipment from direct contact with the metal surface under insulation. Copper and its alloys are
not recommended as materials of construction since mercaptans will corrode these metals.
Copper is also a well known catalyst to degrade fuel properties.
11.4
TYPES OF CORROSION
CORROSION
Hydrocarbon feed to a Merox Unit contains H2S which is removed in the caustic scrubber and
mercaptans which are either extracted or converted to disulfides. Corrosion reactions may be
caused by impurities contained in the feed. Impurities will be defined as nonhydrocarbon
constituents, although they may be and quite often are, attached to a hydrocarbon molecule.
The word impurity is also a misnomer because this word implies something undesirable. Not
everything that one may want to remove, or modify, can be considered undesirable. For
example, certain phenolic, naphthenic acid, or sulfur bearing compounds are now being credited
with giving jet fuel the lubricity needed by the fuel pumps, and their removal from the fuel is
being questioned.
Elemental sulfur may come into the unit in fuels as a result of oxidation of hydrogen sulfide.
Elemental sulfur will, therefore, be found in those fractions that contain H2S as produced and
which have come in contact with oxygen prior to the removal of H2S.
If the spent air is routed to the refinery flare line, elemental sulfur can form in the line if hydrogen
sulfide is present. The presence of elemental sulfur in a petroleum product makes it corrosive to
iron and will cause the product to fail the copper strip test.
Carbonyl sulfide (COS) is a compound that appears only in the lower boiling fractions, such as
the overhead from a depropanizer or debutanizer. It is known to be found in some natural gas
fields and COS also appears to be a decomposition product of sulfur compounds in thermal and
catalytic cracking operations. If the Merox treated product is intended as a raw material for
chemical syntheses, COS will hydrolyze under certain conditions to produce H2S plus CO2. Metal
corrosion may be a problem when the COS hydrolyzes.
The disulfide oil produced in a Merox extraction is separated from the aqueous regenerated
caustic in the main body of the disulfide separator vessel. Upon separation, the disulfide oil is
periodically withdrawn to disposal as an upper oil phase.
Page 3 - 128
11.5
October 1999
CORROSION CONTROL
CONTROL MEASURES
Merox units normally experience very little, if any, corrosion in normal operations. Contaminants
entrained in the feed stream should be examined whenever corrosion is experienced. The
contaminants entering the unit with the hydrocarbon feed will many times be first observed by
unit operators. Problems in meeting product specifications is often a sign that upstream feed
composition has changed.
Merox Extractions units can experience some corrosion in the piping located between the
oxidizer vessel and the settling drum. Hydrogen sulfide or reentry disulfide oils can enhance the
corrosion rate on this line. The pipe is normally constructed of schedule 80 carbon steel.
Replacement of this piping should be considered at shutdown time if the line has experienced
high corrosion rates.
11.5.1 Corrosion Monitoring

On-stream corrosion monitoring is not used due to the extremely low corrosion rates
experienced. Shutdown unit inspection is adequate to insure safe corrosion allowances for the
next planned operating schedule.
11.6
INSPECTION
Merox units are normally constructed of carbon steel. Operating conditions do not require stress
relief of the carbon steel to protect against caustic stress corrosion cracking. When a unit is
shutdown and entry into vessels is planned, it is common practice in many refineries to steam out
the vessels. Steam temperatures may result in caustic stress corrosion cracking. The unit should
be flooded with fresh water to dilute and remove caustic from the equipment. Low temperature
steam may not dilute and remove all the caustic from low pockets. Water flooding of the vessels
and piping will normally prevent cracking problems when used before a steamout.
Under certain conditions, activated charcoal can become an excellent oxygen scavenger,
depleting the oxygen content of the atmospheric air contained in Merox fixed bed reactors.
Therefore, precautions should be taken anytime personnel find it necessary to enter any vessel
which contains activated charcoal.
Page 3 - 129
October 1999
Page 3 - 130
MEROX MERCAPTAN
EXTRACTION UNIT
Extracted
Product
Figure 3-15
Excess
Air
E
Disulfide
H2S Free
Feed
DS
Merox-Caustic
Solution
Rich
Merox
Caustic
LEGEND
E =Extractor
O = Oxidizer
DS = Disulfide Separator
= Intermittent Use
Merox WS
Catalyst
Injection
LIQUID - LIQUID MEROX

SWEETENING UNIT
SOLID BED CONTACTOR
Figure 3-16
H2S
Free Feed
Air
Caustic
Recycle
LEGEND
C =Contacting Bed
B = Disengaging Basket
= Intermittent Use
Merox WS
Catalyst
Injection
Sweet
Product
CONVENTIONAL FIXED - BED

MEROX SWEETENING UNIT
Figure 3-17
Air
2
H S Free
Feed
Merox No.8 Catalyst

or
Merox No. 10 Catalyst
Sweet
Product
Activator
CS
LEGEND
R = Reactor
CS = Caustic Settler
= Intermittent Use
Caustic
Circulation
MINALK FIXED - BED MEROX

SWEETENING UNIT
Figure 3-18
Air
H2S Free Gasoline
Activator
Continuous
Caustic
Injection
LEGEND
R = Reactor
D = Drain Interface Pot
Sweetened
Gasoline
D
Spent Alkali
and Water
12.0
October 1999
SOUR WATER
WATER STRIPPING UNIT
A Sour Water Stripper is a unit designed to remove primarily ammonia and hydrogen sulfide from
refinery process water streams. These streams carry contaminants such as ammonia, hydrogen
sulfide, carbon monoxide, carbon dioxide, cyanides, thiocyanates, phenols, salts, organic and
inorganic acids, inhibitors, and some hydrocarbons. Although the amounts of the contaminants
vary depending on the process and types of crude being run, the primary contaminants,
ammonia and hydrogen sulfide are typically in the 1000 to 10000 ppm range. The stripper
removes the ammonia and hydrogen sulfide and sends it to a Sulfur Recovery Unit.
12.1
PROCESS
There are several different types of sour water strippers including acidified, nonacidified
condensing and noncondensing units. The most common strippers are the condensing and
noncondensing units. The process objective is to remove ammonia and hydrogen sulfide from
the process water. Sour water originates from the accumulator, reflux and knockout drums at the
various refinery units. A typical single tower stripper flow diagram is shown in Figure 3 -21. In a
typical stripper, the sour water is charged to a feed drum, through the stripper tower
feed/bottoms exchanger to heat the feed and cool the stripped water. The heated water is fed to
the stripper where the ammonia and hydrogen sulfide are steam stripped and go overhead. The
overhead is partially condensed and fed back into the tower as reflux to reduce the water content
of the offgas. The offgas containing concentrated quantities of ammonia and hydrogen sulfide
are sent to the Sulfur Recovery Unit.
12.2
CORROSION PROBLEMS
PROBLEMS
The major area of corrosion in a typical nonacidified condensing stripper is the overhead
condensing system. The acidic condensate that forms in the overhead system includes ammonia,
hydrogen sulfide and chlorides. At temperatures below 60oC (140oF) ammonium bisulfide is
formed and becomes very corrosive at concentrations over about 2%. The overhead system is
very sensitive to velocity effects along with the corrosion problem.
12.3
CORROSION CONTROLS
CONTROLS
A major corrosion control is to keep the overhead temperature above about 82oC (180oF) to
minimize ammonium bisulfide formation. If hydrogen blistering has been a problem, some
inhibitors have been tried. Generally, alloy exchange equipment and piping in the reflux system
along with a 316 reflux pump have been fairly successful.
12.4
MATERIALS SELECTION
SELECTION
Carbon steel is used in the strippers where it does not suffer too high a corrosion rate. The feed
surge drum , feed piping to the feed/bottoms exchangers and the stripper are usually carbon
steel. Often the stripper tower is either acid resistant concrete lined or clad with 304L or 316L.
Page 3 - 133
October 1999
The feed/bottoms exchangers are usually carbon steel up to about 82oC (180oF). At temperatures
above this the duplex stainless steels like 3RE60 or 2205 are used. The bottoms reboiler is usually
carbon steel but sometimes the shell requires an austenitic stainless steel liner and duplex
stainless steel tubes. There has been some 304 and 316 used but many refiners are concerned
with chloride stress corrosion cracking with the austenitic stainless steels. The overhead system
has been aluminum, titanium, duplex stainless steel and austenitic stainless steels. Typically, the
condenser has had duplex tubes. The accumulator drum has been 6mm corrosion allowance
carbon steel and the piping between the condenser and the accumulator and reflux back to the
tower has been austenitic stainless steel. The reflux pump has been 316 or Alloy 20. The valves
have typically used 316 trim instead of 12%Cr which could have a cracking problem. NACE
valves with low hardness 12%Cr trim would be an option but are generally more expensive than
316 trim.
Page 3 - 134
OVERHEAD
CONDENSER
Figure 3-19
H2S & NH3

to Disposal
SOUR WATER
STRIPPERS
(Acidified)
Feed
Surge
Tank
Reflux
Drum
Sulfuric
Acid
Equipment for
Acidified Strippers
STM
STRIPPER
REBOILER
Treated Water
To Disposal
OVERHEAD
CONDENSER
Figure 3-20
H2S & NH3

to Disposal
SOUR WATER
STRIPPERS
(Non-Acidified)
Feed
Surge
Tank
Reflux
Drum
STRIPPER
STM
REBOILER
Treated Water
To Disposal
OVERHEAD CONDENSER:
- If water-cooled, use Ti Tubes with Ti Tubesheet, baffles and tie rods
- If air-cooled, use Ai-finned Ti tubes with 304LSS header boxes
NH3, H2S to sulfur

plant or
incineration
WET H2S CRACKING:
- Use 18-8 Valve Trim
- Limit hardness of CS welds 200
BHN
- Use HIC resistant steels and
PWHT for top head of column,
0H exchanger shell and reflux
drum
STRIPPER COLUMN:
Reflux
Drum
REFLUX DRUM CS
with 1/4 c.a.
REFLUX PUMP & CONTROL VALVE:

Polysulfide to resist
NH4HS corrosion
- Hastelloy C for NH4HS corrosion resistance
CS PIPING:
- Maintain velocities 20 fps to avoid
severe NH4HS corrosion, upgrade
is T304LSS
- CS shell with 1/4 c.a. for top head

and upper shell (NH4HS
corrosion)
- 3/16 c.o. for remaining shell
- 304LSS top five trays
REBOILER:
- CS tubes - duplex SS or Ti upgrade
for NH4Ci pitting resistance
Feed Water
(NH3 , H2S, CN)
Feed Bottoms Exchanger

typical CS
Stripped Water
TYPICAL CONVENTIONAL
REFLUXED SOUR WATER STRIPPER
Figure 3-21
13.0
SULFUR RECOVERY UNITS
13.1
PROCESS
October 1999
The Claus sulfur plants are typical of the units that convert hydrogen sulfide to elemental sulfur.
Since environmental regulations have required lower sulfur products and restrict the sulfur levels
of refinery fuels, the sulfur units have become important in refining operations. Environmental
regulations determine whether the sulfur plants will require a Tail Gas Unit to assist in converting
100% of the hydrogen sulfide to sulfur. Typically, a three converter sulfur unit will convert 9498% of the hydrogen sulfide to sulfur. Figure 3-22 shows a typical Claus unit with the major
equipment and materials. The major equipment includes the reaction furnace and waste heat
boiler, the reheaters(auxillary burners), catalytic converters, and the sulfur condensers. In the
reaction furnace about 1/3rd of the hydrogen sulfide is burned to produce SO2 and heat.
2H2S+3O2
2H2O+2SO2 + Heat
The SO2 is reacted with the remaining H2S to produce sulfur and water.
2H2S +SO2
2H2O + 3S + Heat
Other reactions produce some sulfur in the reaction furnace. Total conversion in the furnace is
about 50-65%. The gas from the reaction furnace is passed through the waste heat boiler which
is connected to the furnace. Some sulfur is condensed out in the waste heat boiler. The
remaining gas is trim heated with the reheater burners and is passed into the catalytic converters
over an alumina catalyst. In going through multiple converters in series up to 98% of the
hydrogen sulfide is converted to sulfur depending on the impurities in the hydrogen sulfide
stream.
The sulfur is condensed out in the waste heat boiler and by heat exchange in the sulfur
condensers after the converters. The condensed sulfur flows to a sulfur pit or holding tank for
storage and shipping as liquid sulfur. There are several variations that include shipping solid sulfur
depending on the expected end use.
Two of the major impurities in the hydrogen sulfide streams are ammonia and hydrocarbons. The
ammonia must be properly burned or iron ammonium sulfates are formed which can plug
equipment downstream of the reaction furnace with deposits. Hydrocarbons and amines tend to
polymerize and foul the converter catalyst and heat exchange equipment.
13.2
MATERIALS
Carbon steel is the basic material for sulfur plants. Various grades of firebrick and castable
refractories are used to line the furnace, waste heat boiler, converters and ductwork to minimize
the heat loss and prevent rapid sulfidation of the carbon steel pressure envelope.
Ceramic ferrules are used in the waste heat boiler tube inlets to protect them from the high inlet
temperature. Austenitic stainless steels are usually used where very severe conditions prevent the
use of carbon steel. Locations such a burner tips, demister pads, catalyst screens and refractory
anchors are generally austenitic stainless steel. Carpenter 20 is usually used for sulfur pit steam
coil risers.
Most material problems are caused by:
Page 3 - 137
October 1999
High temperature in the reaction furnace and reheaters have caused refractory loss and
corrosion of the burner tips and furnace shell.
Hot H2S/SO2 corrosion of the tubesheet and tubes of the waste heat boiler because of
refractory problems.
H2SO3/H2SO4 condensation causes severe rapid corrosion in the system when process
temperatures are too low or when the unit is shutdown improperly. Low temperatures
are generally due to low feed rates.
Sulfur or iron sulfide deposits when moist can cause severe corrosion. It is generally
difficult to properly shutdown a sulfur unit.
Boiler Feed Water corrosion is not unusual and can cause tube failures if the waste heat
boiler tubes are not properly cooled.
Page 3 - 138
Castable refractory-lined below catalyst bed. 300

Series SS catalyst screens and alonized carbon
steel gratings and supports.
Steam
CATALYTIC
CONVERTERS
2300 Fo
BFW
470 Fo
Reheaters are
refractory lined.
600 Fo
Refractory-lined carbon steel for

high temperature resistance.
Waste
Heat
Boiler
Stage 2
600 Fo
Reaction
Furnace
Stage 1
420 Fo
430 Fo
Air
H2S
#1
Stainless Steel Demister pads. SO

+HO
2 condensation leads to
corrosion.
BFW or
steam
270 Fo
#1
SULFUR PLANT
Burner tips are 316

SS. Upgrade to
alonized 310 SS or
alonized Incoloy 800.
Recently alonized
Haynes 188 and 556
installed.
#2
BFW or
steam
BFW
#2
#3
Reheaters
#1, #2
(often
auxiliary
burners)
Figure 3-22
BFW
BFW
r
Su
lfu
Su
Su
lfu
r
lfu
BFW
Tail
Gas
Type 1 Portland cement preferred for new

construction. Consider external moisture barrier
(membrane) if water table is high. Repair with acid
resistant potassium silicate cement.
Sulfur Pit
o
280 F
Alloy 20 Cb-3 stream/condensate risers for heating
coils, 300 Series SS is common.
Aluminum Alloy 6061-T6 is commonly used for manway

covers, pit roofs and structural components.
Sulfur
Condensers
#1, #2, #3
October 1999
Page 3 - 140
October 1999
14.0
SELECTION OF CORROSION
CORROSION MONITORING METHODS FOR REFINERY APPLICATIONS
14.1
ABSTRACT
Corrosion monitoring is one of the basic functions required for safe and efficient refinery
operations. This section presents an overview of the various corrosion measurement methods
available for monitoring refinery systems. They range from non-direct/non-intrusive to
direct/intrusive techniques and include everything from simple corrosion coupons to
sophisticated electrochemical techniques. These systems should be selected to meet the needs of
particular plant to minimize downtime and equipment failures. Major sources of corrosion in
refinery processes include dew point corrosion in overhead systems where an acidic, chloride
containing condensate precipitates from the hydrocarbon phase. Other conditions that require
monitoring include high temperature non-aqueous corrosion (i.e. naphthenic acid), aqueous
sulfide corrosion that can produce hydrogen charging and cracking of steels, exposures to strong
acids used as catalysts, and amine solutions used in gas sweetening equipment. Each process
poses a different set of conditions that must be analyzed. This section also describes several
refinery processes and indicates typical monitoring methods.
14.2
INTRODUCTION
The are many purposes for corrosion monitoring in refinery systems. These purposes generally
include one or more of the following:
Diagnosis of corrosion problems

Monitoring of corrosion control methods (i.e. inhibition, pH control, etc.)
Advanced warning of system upsets leading to corrosion damage
Invoke process controls
Determination of inspection and/or maintenance schedules
Estimation of useful service life of equipment
There are also several basic aspects of corrosion monitoring that are often overlooked. First,
corrosion assessment can be complex since refining operations provide a wide variety of
environments and service conditions. Second, no single method will work or provide optimum
results in all applications. Third in some cases, multiple measurement technologies may be
needed in combination to provide accurate and reliable corrosion monitoring information.
The purpose of this section is to introduce several of the more common techniques utilized for
monitoring corrosion in refinery operations and to define their advantages and limitations.
Additionally, examples will be presented on how these techniques are employed for in-plant
corrosion monitoring. This discussion will include selection of the monitoring technique(s) and
placement of probes and coupons for various types of refinery processes.
14.3
TYPES OF CORROSION
CORROSION MONITORING TECHNIQUES
There are four basic categories of corrosion monitoring techniques. The first distinction is based
on whether or not the technique is direct or indirect. That is, whether or not the monitoring
Page 3 - 141
October 1999
technique measures a direct result of corrosion (i.e. mass loss, corrosion current, etc.). Examples
of direct techniques are the three most common corrosion monitoring methods:
Corrosion Coupons - Coupons can be weighed before and after exposure to determine the
extent of mass loss corrosion. In some cases, valuable information on the nature and
distribution of corrosion can also be assessed.
Electric Resistance (ER) technique - ER probes utilize the change in resistance produced by the
change in cross sectional area resulting from corrosion over an interval of exposure.
Linear Polarization Resistance (LPR) technique - LPR probes utilize the relationship of Potential
vs. Current on a corroding electrode.
Indirect corrosion monitoring techniques measure an outcome of the corrosion process. Two of
the most common indirect techniques are ultrasonic testing and radiography. Both techniques
can be used to determine the remaining wall thickness of a pipe, vessel or other equipment
affected by corrosion.
The other two classifications of' corrosion monitoring techniques are intrusive or non-intrusive.
An intrusive technique requires entry into the process stream. Corrosion coupons, ER and LPR
probes and on-line pH or water analysis are intrusive since they require access to the inside of the
equipment to be monitored. Examples of non-intrusive techniques are external hydrogen flux
probes and analysis of water samples obtained through an existing valve.
14.3.1 Corrosion Coupons

The evaluation of corrosion coupons is one of the most basic methods of corrosion monitoring.
Consequently, coupon testing is often overlooked as an archaic procedure. However, it should be
realized that coupons provide the most reliable physical evidence possible. They yield information
on average mass loss corrosion rate as well as on the extent and distribution of localized
corrosion. They can also provide information on the nature of corrosion through analysis of their
corrosion products. The biggest drawbacks that coupons have are that they usually require
significant time in terms of labor and they provide time averaged data that can not be utilized for
real time or online corrosion monitoring This not withstanding, corrosion coupons should be
utilized to provide periodic information and their data should be considered a basis from which
all other methods should be compared.
Procedures for in-plant corrosion coupon testing are given in ASTM G4. This standard give
methods for placement, installation and exposure of in-plant corrosion coupons. ASTM G1 gives
useful guidelines for preparation, cleaning, weighing of corrosion coupons. To evaluate the
general mass loss corrosion rate (in millimeters per year) from coupon tests, the following
equation can be used:
Corrosion Rate (in mmpy) = (8.76 x 104) M/ADT
Where:
M = mass loss resulting from the difference initial and final specimen weights (g),
A = coupon surface area (cm2),
D = material density (g/cm3) and
T = time of exposure (hours).
The corrosion rate in mmpy can then be converted to mils per year (mpy) by multiplying by 40.
Page 3 - 142
October 1999
Average corrosion rate determined with the above equation also assumes that the corrosion of
the metal is uniform. This is why it is important to evaluate coupons visually. ASTM G46 gives
procedures for analysis of localized corrosion. It is important to document the density (#/area),
size (diameter) and depth of localized attack per procedures given in ASTM G46.This information
can assist in the evaluation of the mode of corrosion and potential problems in inspection
associated with localized corrosion.
Useful parameters for localized corrosion are:
1. the ratio of maximum localized attack rate to the general corrosion rate determined by mass
loss and;
2. the percent of corrosion affected area on the coupon.
In many systems the ratio of pitting rate to general corrosion rate will be low indicating that the
general corrosion rate is an accurate predictor of corrosion performance. In other cases, such as
where oxygen ingress, pitting of stainless alloys and velocity accelerated corrosion (i.e.
naphthenic acid corrosion) occur, the local attack rate can be over ten times the general
corrosion rate. Such differences are important when trying to assess the relevance of inspection
techniques such as ultrasonic tests of remaining section thickness.
14.3.2 Electrical Resistance Monitoring

The ER probe is comprised of a sensing element that is basically a loop of material made from a
wire, strip or tube which is used to conduct an electrical signal. When exposed to a corrosive
environment, the cross-section of the loop is reduced over time which increases the resistance of
the sensing element thus producing a change in the output on the ER meter according to the
following formula:
R=s(L)A
Where: R = is the resistance
s = is the resistivity of the metal
L = is the length of the sensing element
A = is the cross sectional area.
Relevant information on this technique can be found in "ASTM G96" for in-plant corrosion
monitoring.
The main benefit of the ER technique is that it can be utilized in on-line process monitoring.
Multiple probes can be used to access various locations in the process stream. Telemetry can be
used to send this information back to a central location so that corrosion rates and the effects of
process changes can be identified. Probes of varying materials and sensitivities are commonly
available. One of the best aspects of the ER technique is that it does not require a continuous
electrolyte current path to make measurements. Therefore, these probes work in multiphase
environments that contain distinct aqueous and hydrocarbon phases and they can be utilized for
monitoring corrosion in non-aqueous environments (e.g. sulfidic and naphthenic acid corrosion).
The limitations of the ER technique are that they only provide representative data for general
corrosion. They do not have the ability to accurately detect localized attack. ER probes while
available in varying sensitivity, typically require several days to determine a reliable corrosion rate
trend. Therefore, if the process is prone to rapid changes in corrosivity, ER probes typically may
Page 3 - 143
October 1999
not provide accurate and reliable corrosion rate data. In some cases, namely where H2S is
present, they can be prone to error due to the presence of conductive sulfide corrosion products
on the sensing element which may lead to non-conservative results. The results of ER probes
should be compared to those obtained from coupon exposures during the same time period.
While, in some cases, ER data may not give reliable indications of the absolute corrosion rate,
they can yield useful indications of trends and changes in plant corrosion activity.
14.3.4 Electrochemical Methods

Electrochemical corrosion monitoring is based on the premise that corrosion is basically an
electrochemical process that can be monitored through the measurement of potential and
current that characterize the corrosion process. In simple terms. corrosion processes can be
described as an electrochemical potential which indicates thermodynamic driving force and a
current that indicates the reaction rate or kinetics of the process. This corrosion current (icorr) can
be converted to a corrosion rate by employing Faraday's Law:
Corrosion rate=K(icorr)(EW)/D
Where:
K = is a constant
EW = is the equivalent weight
D = is the density of the metal.
Electrochemical methods depend strongly on the ability to measure current flow through the
solution. Therefore, it has limitations in many multiphase or non-aqueous environments.
The main benefits of electrochemical methods are that they can provide faster and more dynamic
information coming very close to being able to measure an instantaneous corrosion rate in the
system. Therefore, it can identify rapid changes in process corrosivity. Some electrochemical
techniques can be used to obtain more mechanistic information through controlled polarization
of the material in the environment. This information includes identification of transitions between
active and passive behavior, shifts in polarization by impurities or additives, and the mechanisms
of inhibition.
As indicated above, a limitation in using electrochemical techniques for corrosion monitoring is in
environments which have very high resistivity or where a non-conductive oil phase is present
which may isolate the electrodes used for these techniques. Also, standard analyses usually
assume that the measured corrosion rate is the result of general corrosion. In actuality, many
cases are observed where only a small portion of the surface of the specimen is actually corroding
thus leading to corrosion rate errors. In some environments, non-corrosion electrochemical
reactions can lead to the measurement of current that does not contribute to corrosion. H2S or
other soluble sulfur species may complicate electrochemical measurements since sulfur can be
easily oxidized and reduced which produces a measurable non-corrosion current. As in the case
with ER probes, electrochemicallv derived corrosion rates should be compared to corrosion
coupon data. Again, the trends may be meaningful, but the absolute corrosion rate values may
be different when compared to coupon data.
14.3.5 Linear Polarization Resistance Method

One of the most popular electrochemical techniques utilized for corrosion monitoring is the LPR
technique. As described in ASTM G59, LPR utilizes a measurement of the slope of the potential
Page 3 - 144
October 1999
versus current plot +20 mV around the corrosion potential to define a parameter defined as the
polarization resistance (Rp) according to the following formula:
icorr=B/Rp
This relationship indicates that the polarization resistance is inversely proportional to the
corrosion current density where B is a combination of the anodic (ba) and cathodic (bc) Tafel
slopes:
B=(ba x bc)/[2.303(ba + bc)]
Very often automated corrosion monitoring equipment utilize a constant (0.12V/decade) of
current for both the anodic and cathodic polarizations of steel. For this condition, the above
equation reduces to simply:
dE/diapp=0.026/icorr=Rp
Due to the inverse relationship between the corrosion rate and the polarization resistance, it can
be seen that high values of measured polarization resistance yield low corrosion rates.
An LPR (Linear Polarization Resistance) plots "E" versus "i" of a corroding metal electrode where
the resultant slope is the polarization resistance. Such determinations can typically be made by
producing only small perturbations (+20 mV) on the working electrode with measurements
being made either manually or they can be automated to produce data at a frequency from 30
minutes to 24 hours. This yields the ability to determine the corrosion rate versus time by taking
multiple measurements over short or extended periods. The data can also be transmitted to a
central location using telemetry. Further guidelines for on-line in-plant corrosion monitoring
using electrochemical and electronic techniques are given in ASTM G96.
14.3.6 Other Types Of EC Monitoring

(i) Potential Monitoring
In some systems, it is important to know how the potential of a material changes. Two examples
of these situations are where either cathodic protection or anodic protection are being utilized to
mitigate corrosion. Variations in the electrochemical potential can indicate if the proper levels of
protection are being maintained or if local changes in corrosion behavior are occurring. For
stainless alloys, such measurements may also indicate the influences of process changes or
additives on the corrosion potential relative to the pitting potential for the material. Potential
monitoring provides a useful way to differentiate conditions of active corrosion from passive
conditions. However, this an indirect method of corrosion monitoring and assesses risk of
corrosion and not the actual corrosion rate. It also requires the use of a stable reference electrode
that can be used in a field or plant setting. For the evaluation of cathodic protection of
Page 3 - 145
October 1999
underground steel pipelines, tank bottoms and buried vessels compensation for the IR drop
associated with soils of relatively low conductivity must be also utilized.
(ii) Zero Resistance Ammetry (ZRA)

ZRA is a relatively simple technique used to monitor the flow of current between two electrodes
in a corrosive environment. The most common case, is to utilize ZRA for the evaluation of
galvanic corrosion between dissimilar materials. In this case, the current flow between two
electrodes of the two materials are connected through a zero resistance ammeter. The resultant
galvanic current is a measure of the galvanic acceleration of the corrosion rate and the value in
terms of mmpy or mpy can be obtained through application of Faraday's Law as discussed
previously. Methods for applying ZRA for the evaluation of galvanic corrosion is given in ASTM
G71. This technique can also be utilized with two electrodes from the same material to monitor
electrochemical current noise as discussed in a following section.
(iii) Electrical Impedance Spectroscopy
With the increased attention toward accurate corrosion monitoring and methods that are
compatible with modern data acquisition technology, several new techniques are currently being
utilized. One of these newer techniques is electrochemical impedance spectroscopy (EIS) which
utilizes an AC signal to excite or perturb a corroding specimen. In fact, one convenient way to
visualize the EIS technique is to picture the low frequency limit of EIS as essentially the same as
the LPR method. In a mechanistic sense, EIS monitors the electric response of the
metal/environment interface to the applied AC signal over a frequency spectrum usually in the
range of 5 to 10kHz to 50uHz. In these cases, a full EIS scan may take up to two hours. In some
cases, however, selected frequencies or a more limited range of frequencies can be utilized to
reduce the time required for corrosion monitoring with EIS.
One of the problems with EIS data is that the analysis is relatively complex compared to the
commonly used ER or LPR techniques. The Nyquest curve plots the imaginary and real
components of the impedance while the Bode curve gives a plot of impedance versus the phase
angle. The benefit of these data representations is that they allow separation of the various
components of the system resistance that are assumed to be the polarization resistance. This
assumption can lead to errors in more coventional LPR corrosion rate determinations.
In environments of low conductivity the solution resistance can be separated from the actual
polarization resistance using EIS. EIS techniques can also be utilized to examine coated or
inhibited materials much more effectively than with LPR techniques. EIS data can be used to
determine the fundamental properties of the surface layers such as pore resistance and film
capacitance. Other areas of application for EIS are in the evaluation of corrosion of steel in
concrete structures and in the evaluation of cathodic protection since both applications require
significant compensation for resistive losses. The main limitations of these techniques are that the
analysis of the data is complex and its interpretation is not fully developed for all applications.
They require application of a theoretical equivalent circuit to analyze and interpret the data.
Quite often these techniques still require benchmarking with other more common corrosion
monitoring techniques such as corrosion coupons before meaningful data can be obtained.
14.3.7 Hydrogen Flux Monitoring

A common byproduct of corrosion of steel is atomic hydrogen (H+). Atomic hydrogen can either
recombine to form molecular hydrogen (H2) and bubble off the corroding metal surface or it can
remain in the atomic form and may diffuse into the steel. The presence of hydrogen sulfide
promotes hydrogen absorption, which in turn can lead to problems associated with the various
Page 3 - 146
October 1999
forms of wet H2S cracking: blistering, hydrogen induced cracking, stress oriented hydrogen
induced cracking and sulfide stress cracking. There have been several intrusive and non-intrusive
devices designed to monitor hydrogen absorption called hydrogen probes.
Intrusive probes(often referred to as finger probes) are actually inserted into a vessel or pipe
section. They consist of a steel sensing element which has a hollow space inside connected to a
pressure sensing devise which monitors the build up of hydrogen pressure. Atomic hydrogen is
formed on the external surface of the probe as a result of sulfide corrosion and diffuses through
the steel element. Once inside of the hollow space, it will recombine to form molecular hydrogen
which can not escape. The rate of hydrogen pressure buildup (or vacuum loss) is proportional to
the severity of hydrogen absorption and should be at least qualitatively proportional to the
potential for wet H2S cracking. Non-intrusive hydrogen probes have also been designed that
utilize an externally applied cell or patch to monitor the rate of hydrogen egress from the outer
surface of the steel pipe or vessel wall, In this case, the wall of the equipment is actually the
sensing element. These non-intrusive devices can utilize relatively simple sensing elements such
as a patch and gauge on the outer surface which traps hydrogen. Conversely, the more
sophisticated probes utilize an electrochemical cell which reacts with the hydrogen as it exists the
outer surface of the steel to produce a current signal.
Data from hydrogen flux probes commonly utilize a hydrogen pressure increase (or vacuum loss)
per unit time. The data is presented in terms of an output current versus time which is generally
compatible with modern telemetry and data acquisition systems. The current versus time
information can be converted to a hydrogen pressure build-up rate, if desired.
The limitations of hydrogen probe monitoring are numerous, but still, such techniques can be of
benefit by providing qualitative information which can be used to monitor the severity of
hydrogen charging in wet H2S operations. Most importantly, hydrogen probe data does not
correlate with weight loss corrosion in wet H2S service since this is only one of many factors
involved in defining the severity of hydrogen charging. Secondly, the diffusivity of hydrogen in
steel changes rapidly with temperature which can produce large, periodic swings in the data with
operating conditions. Finally non-intrusive probes and in some cases even intrusive probes can be
subject to error or delays in reading process transients and can experience decreased
performance if internal blisters or cracks form in the steel. The response time is particularly a
problem when non-intrusive techniques are used to monitor equipment since wet H2S refinery
processes can be prone to rapid changes in hydrogen charging severity with plant operating
conditions.
The following are some practical hydrogen probe comments:
Assign ownership for monitoring

Helium pressure test the gauges before field installation
Maintain 1-2 psig on pressure gauge during operation to ensure the gauge is not leaking
Measure and record hydrogen probe pressures weekly, accounting for temperature
fluctuations
Look for any increase in pressure; Increases in pressure mean increased corrosion activity
and blistering
Patch probes are not considered accurate
Rules of thumb
1. <5 psi increase per month
Minimal damage is occurring
2. 5-30 psi per month
Some activity
3. >30 psi per month
Damage possible
Page 3 - 147
14.4
October 1999
CORROSION MONITORING
MONITORING
Corrosion monitoring is a vital part of any plant corrosion control program. It should be
integrated with other programs designed optimize the process conditions, chemical injection and
inspection to recognize the full potential to successfully managed plant operations.
14.4.1 Chemical Injection

A typical refinery may have dozens of process additives including antifoulants, corrosion
inhibitors, neutralizers and emulsion breakers. Due to the number of additives and the
complexity and dynamic nature of hydrocarbon process environments it is very difficult to
properly simulate these situations to optimize performance or to identify potential interactions.
Therefore, in-plant corrosion monitoring is essential in most processes so that protection and
long term serviceability of equipment can be assured.
Suitable corrosion monitoring devices need to be utilized which are compatible with the nature
of the process media. An evaluation of the process and the intended corrosion monitoring
requirements needs to be conducted. For example, the locations of chemical injection and
distribution of the aqueous phase and the injected chemicals through the system are important
conditions. This will usually dictate the proper location and type of equipment required for
optimum results. Corrosion monitoring locations should be located in sites where water will
condense, pool or impinge since corrosion will usually be more severe in these locations. Other
factors included in the evaluation of monitoring locations are variations in temperature and
concentration of corrosive species through the process equipment. Coupons should generally be
installed for relatively long intervals (i.e. on overlapping 30 to 90 day intervals) to provide
baseline corrosion rates at multiple locations. However, in locations where high corrosion rates
are expected exposure intervals maybe as short as a few days to a week.
As indicated previously, coupon data should provide a basis for interpretation of data from
corrosion probes and other on-line corrosion monitoring devises. Coupons should be installed on
retractable probes which allows for installation and periodic removal while the system is under
operation. The decision on whether to use LPR or ER will often depend on whether or not a
continuous aqueous phase is likely to be present which is a requirement for LPR or other
electrochemical methods. Probes or their sensing elements should generally be replaced when
they reach 80 to 90% of their nominal life. However, they may have to be replaced after as little
as 50% of their nominal life if crevice or underdeposit corrosion or other forms of localized attack
have occurred. In some cases, the elements may even need to be cleaned or changed more
frequently if the data produced appears questionable.
Process water samples can also be a valuable source of indirect monitoring data that can be
obtained non-intrusively and possibly correlated with corrosion rates. Information such as pH
(closed and open cup), dissolved metal content, chloride, hydrogen sulfide and ammonia
content. Systems downstream from catalytic cracking units should also be analyzed for cyanide.
This data can be used to assess the corrosivity of the system and the potential severity of
hydrogen charging and wet H2S cracking and the need for chemical treatment, water washing
and/or other process controls. Hydrogen probe data can then be used to verify the severity of
hydrogen charging, the effectiveness of these remedial actions and the need for periodic
inspection and non-destructive testing.
Page 3 - 148
October 1999
14.4.2 Dewpoint
Overhead systems are a particular problem area due to the potential for precipitation of water
and the influence of acidic chloride containing condensates. Effective corrosion monitoring can
take many forms and in most cases involves multiple sources of data. These usually start with
inspections during a unit turnaround. Visual observations of corrosive damage is valuable
information since it sets the location of highest corrosivity and shows the effectiveness of injected
chemicals used for neutralization and corrosion inhibition. This information may be useful in
verifying that the present corrosion monitoring sites are adequate or for the identification of
other more appropriate corrosion monitoring sites. Samples of corrosion deposits can also be
analyzed to determine the nature of the corrosion and the presence of reactive species that cause
the attack.
Retractable corrosion coupons and ER probes are valuable only if placed in a region of potential
corrosion. One complication is that the location and severity of corrosion may vary with
feedstock and plant operating conditions since these can influence the location and amount of
water precipitated and the amounts of corrodents. This often places some degree of uncertainty
in dewpoint corrosion monitoring. Therefore, corrosion data at fixed sites is usually
supplemented by ultrasonic inspection, radiography and visual examination, where and when
possible.
One way to reduce the need for multiple corrosion monitoring sites is to utilize modern corrosion
monitoring technology. A more sophisticated corrosion monitoring probe can be employed
which has multiple sensor types where the temperature of the sensing element can be controlled
while making corrosion measurements. The readings from the various sensors are viewed in
combination along with the probe temperature data to identify conditions where corrosion is
most severe.
14.5
SUMMARY OF CORROSION
CORROSION MONITORING CONSIDERATIONS IN REFINERY UNITS
The principals of corrosion monitoring are typically to first identify the hot spots for corrosion
based on the location of the following situations:
Precipitation of water from the hydrocarbon phase. Typically, the most corrosive areas
tend to be sites of condensation having the highest temperatures.
Tower overheads where air cooled or water condensers are required.
The accumulator, water draw-off boot and piping.
Effluent coolers.
Locations where corrosive contaminants may concentrate or where potentially corrosive
process chemicals are injected.
Typically, the most important consideration in the process of in-plant corrosion monitoring is to
understand the specific needs for corrosion monitoring and the likely locations for corrosion.
These considerations may vary from plant-to-plant and with varying feedstock and process
conditions.
14.5.1 Atmospheric Distillation Unit

Concerns for corrosion in the atmospheric distillation unit will vary with the crude being
processed. Units processing low sulfur crudes (<0.6% sulfur) typically experience minimum
corrosion. Concerns for corrosion increase as the feedstock goes increasingly sour. These include
Page 3 - 149
October 1999
concerns in both high temperature (non-aqueous) systems due to sulfide corrosion and in lower
temperature systems exposed to wet H2S environments. Critical locations for corrosion
monitoring include crude oil preheat exchangers and piping, the tower, tower overhead,
overhead piping and condensers. These latter sites are locations of water condensation
complicated by acid chlorides which have not been completely neutralized and/or inhibited. If
the crude oil feedstock contains high levels of naphthenic acids or acidic chloride condensates,
corrosion problems may extend into the downstream system from the atmospheric distillation
unit.
14.5.2 Vacuum Distillation Unit

Vacuum distillation units are designed to volatilize additional products coming from the crude
atmospheric bottom These include naphthenic and other organic acids, sodium chloride and
sometime organic chloride salts produced by incomplete hydrolysis in the atmospheric section.
Critical sites for corrosion monitoring are again related to sites of condensation of acidic chloride
environments in the distillate drum boot and the ejector intercondenser collector drum.
If the naphthenic acid content of the crude is significant (Whole crude TAN >0.3; Side cut TAN >
1.5). then concerns for high temperature process corrosion must also be considered. Likely sites
for naphthenic acid corrosion are typically regions of high temperature and high velocity such as
found in the vacuum furnace and transfer lines and in regions of naphthenic acid condensation
as found inside of the column. The problems associated with these conditions are typically
twofold. The first is to determine these locations and to select equipment and sensors that can
handle the high process temperatures. An alternative is to utilize external inspection techniques
such as ultrasonic and radiography to augment corrosion monitoring.
14.5.3 Fluid Catalytic Cracking Unit

Excessive corrosion and hydrogen activity in the light ends section of the plant may be
experienced due to H2S ammonia and cyanides formed in the FCCU reactor from nitrogen
compounds in the feed. Corrosion control may involve water washing of wet gas streams to
reduce the concentration of cyanides, water removal and/or inhibition. These systems can be
optimized when they are combined with effective corrosion monitoring and hydrogen flux
monitoring where concerns for wet H2S cracking persist. Corrosion monitoring sites are found in
the fractionator overhead system, effluent piping of the compressor after coolers and debutanizer
overhead system.
14.5.4 Amine Treating Unit

These units utilize amine solvents to selectively remove H2S and/or CO2 from the process stream.
The most reactive is monoethanolamine (MEA) which is typically the most corrosive in terms of
both weight loss corrosion and the potential for stress corrosion cracking of non-stress relieved
steel equipment. In most cases, either ER or LPR probes can be utilized for corrosion monitoring
and for early identification of corrosive conditions and removal of impurities from the system.
However, the experience with wire element ER probes in this service has indicated questionable
reliability particularly where high velocity or impingement has been encountered. Sites for
corrosion monitoring in amine units include stripper tower and bottoms piping where particle
erosion and amine degradation can add to the severity of corrosion. Other sites are:
Stripper reboilers exposed to rich amine and flashing.

Stripper overhead condenser/receiver and piping.
Page 3 - 150
October 1999
Amine reclaimer where corrosives may concentrate

Lean/Rich MEA exchanger due to erosion effects.
14.5.5 Sour Water Stripping Units

Waste water condensates from various refining processes contain a variety of corrosive impurities
such as hydrogen sulfide, ammonia, chlorides, cyanides and phenols. They are sent to stripping
units for decontamination with steam. A major area for corrosion is in the overhead condenser of
the stripper column. Corrosion monitoring is recommended for optimizing of the chemical
treatment of such streams to protect the overhead condensers. In some cases, replacement of
steel with corrosion resistant alloys has eliminated the need for corrosion monitoring and difficult
to apply chemical treatment.
14.5.6 Sulfuric Acid Alkylation Unit

Sulfuric acid is a catalyst in the alkylation process and is a potentially one of the most corrosive
streams in refining. Typically, it is the equipment in contact with sulfuric acid concentrations of
less than 90% that will experience the highest rates of corrosion although spent acid is less
corrosive than fresh acid. It is these dilute streams, particularly at elevated temperatures and at
velocities greater than 1.2 m/sec (47 in/s), that require the highest attention in terms of corrosion
monitoring. Specific sites for corrosion monitoring include the following:
14.6
Effluent piping from the deisobutanizer where caustic and inhibitor treatment needs to be
monitored.
Effluent piping from the deisobutanizer reboiler to the debutanizer.
Effluent piping from the debutanizer reboiler to the rerun circuit.
Effluent piping from the depropanizer overhead condenser where S02 is formed in the tower
reboiler which will readily combine with water in the overhead system.
AUTOMATED ON-LINE
ON-LINE MONITORING AND DATA ANALYSIS
The combination of increased emphasis on minimizing unscheduled downtime and costly
equipment failures with a decrease in overall plant staffing, has increased emphasis in recent
years on online monitoring. These techniques are being applied so that corrosion monitoring can
become an integral part of process monitoring so that conditions that can jeopardize the long
term serviceability of plant equipment can be minimized or avoided. However, there are typically
numerous potential problem areas associated with automated online monitoring, many of which
also exist in more conventional manually monitored systems. These problem areas include:
Selection of representative corrosion monitoring sites and monitoring technique(s).

Specification and selection of automated corrosion monitoring equipment .
Troubleshooting problems associated with equipment reliability, telemetry (data links)
and signal noise.
Verification of data (comparison with data from other sources such as inspection and
coupons).
Selection of hardware and database software.
Training of support personnel.
Analysis of data, definition of trends, correlation with process variables/upsets and
integration with process controls.
Page 3 - 151
October 1999
Use of data to make process changes.
The advances in PC-based systems and software, network connectivity and computer compatible
corrosion monitoring equipment has simplified but not eliminated these problem areas. For
example, equipment and probe reliability still can be a problem in many cases. Furthermore, it is
important not to underestimate the magnitude of problems in system integration,
troubleshooting, and training of operation personnel. These areas often require major efforts in
order to make online systems work.
14.7
NDT AND CORROSION

CORROSION MONITORING TECHNIQUES
A matrix of available NDT AND Corrosion Monitoring techniques are shown in Table 3-12 and
Table 3-13.
14.8
CONCLUSIONS
Based on this summary of refinery corrosion monitoring the following conclusions have been
drawn:
1. There are a number of corrosion monitoring techniques available for use in refinery systems.
2. No one technique will be appropriate for all systems. Where possible, corrosion data from
more than one technique should be compared and analyzed.
3. Corrosion coupons should be the basis for any corrosion monitoring program. This
information can be augmented by more sophisticated techniques such as ER and LPR
particularly where on-line monitoring capabilities are required.
4. Electrochemical methods such as LPR require a continuous, conductive electrolyte path to
work properly, This may be a problem in condensing or multiphase systems where
nonconductive oil films may complicate accurate readings.
5. ER methods do not require a continuous electrolyte path and can be used effectively in many
multiphase and high temperature (non-aqueous) systems. However, conductive sulfide
corrosion products and fouling of electrodes can be sources of misleading data.
6. In plant operations it is important to develop a corrosion monitoring plan that is based on
knowledge of the specific operating conditions in the plant. These considerations may also
vary with changing feedstocks.
7. Many refinery plant corrosion problems result from precipitation of water from hydrocarbon
process streams. Therefore, major emphasis should be placed on monitoring of corrosion
under dewpoint conditions, particularly where conditions exist for maximum temperature or
the most concentrated corrosive ions exist.
8. Selection of corrosion monitoring sites needs to be based on identification of locations where
water and corrosive species coexist and where other factors such as flow, impingement,
flashing and particle erosion may lead to locally accelerated corrosion.
9. Corrosion monitoring should be integrated with periodic visual inspection during
turnarounds and with nondestructive testing using ultrasonic testing, radiography and other
techniques.
Page 3 - 152
October 1999
TABLE 3-12 MATRIX OF AVAILABLE CORROSION MONITORING TECHNIQUES

General Info:
Technology
Cost
On-Line
Service:
Aqueous
Low Conductivity
Non-aqueous
Sweet
Sour
Max. Temperature
Corrosion Detected:
General
Pitting
Cracking
Trend Only
Coupons
ER
Galvanic
LPR
PD/CP
H2 Patch*
AE*
EIS
ECN
FSM*
Mature
Low
Mature
Medium
X
Mature
Medium
X
Mature
Medium
X
Mature
Medium
Mature/New
Low
Mature
High
X
New
High
X
New
High
X
New
High
X
X
X
X
X
X
None
X
X
X
X
X
300oF
X
X
X
X
X
X
X
X
300oF
X
X
300oF
X
X
300oF
X
X
X
X
X
?
X
X
300oF
X
X
500oF
X
X
X
X
X
>500oF
X
X
X
X
X
X
X
X
X
X
X
* Non-intrusive
X
300oF
Coupons = Corrosion Coupons

Galvanic = Galvanic Probes
LPR = Linear Polarization Resistance
ER = Electrical Resistance
PD/CP = Potentiodynamic/Cyclic Polarization
Page 3 - 153
H2 Patch = Hydrogen Patch Probes/Foil

EIS = Electrochemical Impedance Spectroscopy
AE = Acoustic Emission
ECN = Electrochemical Noise
FSM = Field Signature Method
October 1999
TABLE 3-13
NDT Technique
Visual
Mechanical, Gauging
Penetrant
Magnetic Particle
Longitudinal UT
Shear Wave UT
Radiography
Eddy Current
Acoustic Emission5
Hydrostatic
APPLICATION OF NDT TECHNIQUES TO DETECT AND MONITOR CORROSION
General
Corrosion
Thickness
Localized
Corrosion,
Pitting,
Erosion, Etc.
C
B
A
A
High Temperature
H2 Attach
Advanced,
Gross
Early
Stages
B
B1
C
B
B
A
B3
A
B
C3
C3
Detection
Sizing
B = Often Suitable
H2
Blistering
Leaks
A4
A4
A
A
C
B
C
A = Preferred Method
SCC on
Accessible
Surface
C
SCC on Inaccessible
Surfaces
B
C
C
A
C = Questionable
Notes: This table deals only with the application of NDT techniques to detect and monitor corrosion phenomena. Many of the techniques are also useful for detecting original
fabrication defects and other flaws that affect the integrity of equipment.
1. For metal thickness <0.25 in.
2. If cracks present at accessible surface.
3. Method under development based on attenuation and sound velocity.
4. Magnetic particle is preferred to penetrant but it can be performed on ferromagnetic materials.
5. Insufficient Company experience to highly recommend. Useful if used in conjunction with more proven techniques.
Page 3 - 154

Caltex Process Plant Corrosion

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Caltex Process Plant Corrosion

Загружено:

Авторское право:

Доступные форматы

SECTION 3

REFINERY MATERIALS MANUAL

SAFETY, THE ENVIRONMENT, AND RELIABILITY............................................................................2

CRUDE UNIT CORROSION AND CORROSION CONTROL ...........................................................13

FLUID CATALYTIC CRACKING UNIT ................................................................

REFINERY MATERIALS MANUAL

5.4.4 Reaction Mix Line, Main Fractionator and Bottoms Piping...................................................45

CATALYTIC LIGHT ENDS RECOVERY UNIT ................................................................

HYDROPROCESSING UNITS ................................................................

REFINERY MATERIALS MANUAL

CORROSION PHENOMENA IN HYDROPROCESSING UNITS........................................................81

ALKYLATION UNIT ................................................................

10.0 AMINE TREATING UNIT ................................................................

REFINERY MATERIALS MANUAL

11.0 MEROX TREATING UNITS ................................................................

12.0 SOUR WATER STRIPPING

13.0 SULFUR RECOVERY UNITS................................

14.0 SELECTION OF CORROSION

Related Standards and Regulations ........................................................................................2

REFINERY MATERIALS MANUAL

Fractionator Damage Mechanisms (1) ...................................................................................57

Simplified Refinery Flow Diagram.. 9

REFINERY MATERIALS MANUAL

REFINERY MATERIALS MANUAL

REFINERY MATERIALS MANUAL

SAFETY, THE ENVIRONMENT,

RELATED STANDARDS AND REGULATIONS

Mechanical integrity programs

Inspection, repair and re-rating of pressure vessels

Inspection and repair of piping

API 650, 651, 652, 653

Design, inspection, repair, cathodic protection of tankage

Design of fired heaters

National Board code covering inspection and repair of pressure vessels

ASME Boiler and Pressure Vessel

New pressure vessel design, fabrication and inspection

ASME Piping Code

New piping design, fabrication and inspection

Prevention of polythionic acid stress corrosion cracking

Prevention of cracking of CS welds

Inspection, fabrication and repair of equipment in wet H2S service

South African Standards

New vessel design, fabrication and inspection

New vessel design, fabrication and inspection

Japanese high Pressure Gas Law

New vessel design, fabrication and inspection

Korean High pressure Gas Law

New vessel design, fabrication and inspection

British Standard 5500

New vessel design, fabrication and inspection

REFINERY MATERIALS MANUAL

REFINING PROCESS OVERVIEW

REFINERY MATERIALS MANUAL

TABLE 3-2 COMMON REFINING PROCESSES

Light Product Upgrading

Heavy Oil Upgrading

Oil Movement and Storage

REFINERY MATERIALS MANUAL

TABLE 3-3 TYPICAL REFINERY PROCESS SELECTION

Crude Light Ends Separation

Crude Light Ends Separation

Sour Water Stripping

Fluid Catalytic Cracking

Oil Movement and Storage