Paper No.

392

AMMONIUM SALT CORROSION IN

HYDROTREATING UNIT STRIPPER COLUMN OVERHEAD SYSTEMS

C.A. Shargay, G.E. Jacobs

Fluor Daniel, Inc. -
3353 Michelson Dr.
Irvine, CA 92698

M.D. Price
Los Angeles Refinery Company
Equilon Enterprises LLC
P.O. BOX 817
Wilmington, CA 90748

ABSTRACT

In the past, the presence of ammonium bisulfide (NHqHS) or ammonium chloride (NHdCl) in hydrotreater stripper
overheads was relatively rare. However, designs of newer units and revamps to older units, are resulting in corrosive levels
of NHdHS or NHdCl in overhead systems. This is primarily due to the addition of a hot high pressure separator (HHPS) with
the bottoms going directly (or through another separator) to the stripper. This paper presents data on one corrosion case
history in detail, a summary of another case history reported in Retln-Cor 3.0 and the results of a brief survey targeted to
units with HHPS designs. Some correlations of the process conditions to the severity of corrosion are made based on the
survey results, and recommended materials selection and other corrosion control methods to minimize the risk of this
problem are given.

Keywords: hydrotreaters, hydrodesulfurizers, ammonium bisulfide, ammonium chloride, corrosion, material selection

HISTORY OF CORROSION PROBLEMS

Ammonium bisulfide (NH4HS) corrosion is a well-known problem in Hydrotreater and Hydrocracker reactor effluent
air cooler (REAC) areas. There have been numerous case histories of severe corrosion problems reported, and this corrosion
has been the cause of several major f~es and costly unscheduled shutdowns. Numerous articles detailing the root causes of
the corrosion and the recommended prevention methods have been published.’-’ Results of a NACE survey on the topic were
published in 1976.’

In contrast, this paper focuses on NH4HS and ammonium chloride (NH~Cl) corrosion in other areas of the plant. These
areas have been considered lower “risk” based on most risk rating systems, because they are significantly lower pressure.

Copyright
W 999 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole must be made in writing to NACE
International, Conferences Division, P.O. Box 218340, Houston, Texas 77218-8340. The material presented and the views expressed in this
paper are solely those of the author(s) and are not necessarily endorsed by the Association. Printed in the U.S.A.
However, the corrosion can be rapid and localized, which can make it difficult to locate by inspection programs. Also, the
streams contain HZSand can include relatively light products, hence the impact of leaks can still be significant.

The prima~ areas of the Hydrotreater and Hydrocracker units covered by this paper are:

+ Hot Low Pressure Separator vapor streams, and

+ Product Stripper (or Fractionator) overheads.

One of the frost documented cases of a major corrosion problem in these areas was reported in 1991. In the NACE
International CORROSION/9 1 minutes of the T-8 Refining Industry Corrosion Group Meeting, one refiner reported
corrosion of the stripper overhead air cooler, piping and tower top section. This is published in the Refm-Cor 3.0 database
available from NACE, as item 91C9.7-O1.

From informal surveys, numerous other case histories were uncovered. Additional information on these cases will be
discussed later in this report.

REAC CORROSION MECHANISM AND PREVENTIVE MEASURES

Since the details on the NH4HS corrosion mechanism are very similar between the REAC and the other plant locations,
a detailed review of that mechrmism is justified.

Ammonium Bisulfide Corrosion Mechanism

Ammonium bisulfide is the product of ammonia and hydrogen sulfide gases. As the reactor effluent stream cools
down, solid NHqHS can crystallize out of the vapor phase and may quickly plug up exchanger tubes and cause underdeposit
corrosion. However, ammonium bisulfide readily dissolves in water, so a common practice is to inject water ahead of
effluent air coolers to prevent ammonium bisulfide salt deposition and to form an aqueous bisulfide solution that is dilute
enough to be only mildly corrosive. Concentrated 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 chlorides in the process fluid
tends to fiu-ther accelerate corrosion.

Industry studies of reactor eflluent air cooler and piping corrosion has led to a general understanding of the following
important factors:

● Corrosion is fimdamentally caused by ammonium bisulfide and corrosion rates increase in relation to its concentration.
The NHOHSconcentration is typically controlled in operating units by adjusting the water injection rate. The appropriate
NHqHS limit for a given unit seems to depend on many variables, and typical limits for the cold high pressure separator
water range from 2°/0to 10°/0(by weight). Some rough guidelines are: QO/Ohas a negligible risk of corrosion, at 2°/0to
about 80A,CS can be used with added corrosion prevention steps, and systems with >8°A may need to be alloyed to be
reliable.
● NHqHS corrosion is velocity-sensitive. At high velocities and with susceptible concentrations of NHqHS, corrosion-
erosion of carbon steel tube ends or piping elbows is likely to occur. Hence, maximum velocity limits are typically
applied during design of new units.
● In air cooler or heat exchanger tubing, there is also a minimum velocity to avoid deposit formation, which can then lead
to underdeposit pitting.
● In the materials selection 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 @ stream entering the reactor eflluent coolers to
determine whether alloy is needed for air coolers. The additional expense for a materials upgrade maybe partially
justified by the difficulty and cost of the frequent extensive inspections that are normally required for carbon steel.
 {Kp is one prediction tool for REAC’S which is probably not applicable to the other plant areas. The criteria for
predicting corrosion levels were based on empirical data from REAC’S (generally in units with no HHPS). Since many
other process variables are different in the other plant areas and since some Kp in these areas in operating plants are
extremely high with no major corrosion problems on CS, Kp does not seem to be representative of the corrosivity of these
services. }

. For operating units, the NH,HS concentration is generally considered a better indicator of the REAC corrosivity than Kp.
This is because Kp is determined on a dry basis, and would not reflect the effect of changes in the water injection rate on
NH~HS concentration and hence, on corrosion rates

. When corrosive conditions are present, the corrosion locations can include:
REAC,
REAC inlet piping from the water injection point to the air cooler,
REAC outlet piping, and
Cold High Pressure Separator (CHPS).
At extremely high NHqHS concentrations, corrosion has also been reported in the vapor line (i.e. recycle hydrogen line)
and liquid hydrocarbon line from the CHPS, and in downstream separators and their associated piping. This occurs when
there is water carryover into these lines. The water has the same NHqHS concentration as the water draw from the CHPS.

Corrosion Prevention Measures With NH4HS >2%

● Proper water injection rates ahead of the cooler are key to controlling corrosion. Water injection at less-than-
recommended rates can lead to accelerated corrosion. The water rate should be high enough that i) >25°/0 remains
unvaporized, ii) the NHdHS content in the separator water meets the desired limit, and iii) the process velocity in the
tubes is within the specified limits (however, velocities are usually not significantly influenced by the water rate).
● Injection water must be ilee of oxygen (less than 15-50 ~. The presence of oxygen leads to rapid corrosion and
fouling, even at lower NH,HS concentrations. 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 maybe
“outside of desired limits”, but not significantly affect corrosion, if all the other variables are acceptable. However,
oxygen contamination is one factor which, alone, can immediately lead to severe corrosion.
● Air coolers with carbon steel tubes and carbon steel piping are typically limited to 6 m/s (20 Ms) maximum. Stainless
steel (SS) ferrules may be used at both inlet and outlet ends of the tubes that handle streams with high ammonia and H2S
content. For alloy tubing and piping, the upper limit for velocity can be raised to 9 m/s (30 tVs).
● To prevent corrosion, good flow distribution of vapor, liquid hydrocarbon, and water phases is essential. This is
accomplished by using a balanced (symmetrical) design for inlet and outlet headers and by keeping fluid velocities high
enough to maintain an acceptable flow regime. Phase separation problems, and risks of deposits and underdeposit
corrosion can occur when velocity is <3 m/s (<10 fth) in the air cooler, hence this is the typical minimum velocity limit.
● 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.
However, some refiners consider the reliability of the multiple injection systems (with many small valves, flow meters,
etc.) so poor, that single injection points are sometimes preferred.
● Air-cooled exchangers designed for this service should have header boxes at both ends. U-bends should be avoided
because they are se&itive to-erosion-corrosion. Header boxes should be designed to permit access for tube cleaning.

This type of corrosion is still the subject of research programs and surveys. The effect of flow regimes, chlorides,
excess HZS and various other variables are still being studied. Guidelines for achieving good water distribution are
continually being improved. Some of the recent causes of unexpected corrosion in REAC’S have been attributed at least
partially to:
+ oxygen contamination of injection water due to loss of nitrogen blanketing on the water storage tank,
+ uneven shutdown of fans which led to poor distribution of both hydrocarbons and water,
+ air cooler designs with multiple tube rows in the inlet pass (water goes preferentially to bottom rows), and
+ high chloride levels and NH~Cl deposits and pitting.

There is less published literature and guidelines on the NHqCl corrosion problems in these units, but water washing to
prevent deposits helps avoid corrosion.

VARIATIONS IN UNIT DESIGNS

There are numerous variations of the designs of Hydrotreaters in both the effluent separation sections and the
stripper/fractionator areas. Units can have between one to five separators, and for the purposes of this paper, the following
names and acronyms will be used:

Hot High Pressure Separator (HHPS)

Cold High Pressure Separator (CHPS)
Hot Low Pressure Separator (HLPS)
Cold Low Pressure Separator (CLPS), and
Cold Flash Drum (CFD)

We will be discussing three of the more common separator schemes, namely:

1. Effluent directly to REAC to CHPS (no HHPS) - Figure 1

2. Effluent to HHPS; HHPS vapor to REAC, HHPS liquid to Stripper - Figure 2
3. Effluent to HHPS; HHPS vapor to REAC, HHPS liquid to HLPS;
HLPS vapor to condenser then to CLPS - Figure 3

Figure 1 shows the historically most common separator scheme in hydroprocessing reactor loops. After cooling via
heat integration, the reactor effluent flows through the REAC and into the CHPS. Wash water, injected upstream of the
REAC, removes the ammonia horn the reactor effluent in the form of ammonium bisulfide. The sour water is then decanted
in the CHPS. Most of the ammonia formed in the reactor is removed in the sour water. Hence, with this process scheme,
there are generally no areas of NHqHS corrosion concern other than the REAC area (and CHPS along with its outlet piping if
the NHqHS concentration and/or velocities are high). The stripper overhead water phase would contain H2S, but no
significant amounts of ammonia.

With this scheme, the CHPS operating temperature is typically as low as possible (32° to 49”C, {90”F to 120°F}),
maximizing the condensation of light ends and, therefore, the hydrogen purity of the recycle gas. The viscosity and density
of the cold reactor effluent influence the efllciency of the oil-water separation, and may necessitate somewhat higher
operating temperatures. The inability to efficiently perform this oil-water separation was one of the primary driving forces
(aside from improved energy efficiency) for the incorporation of the HHPS into the reaction loop, as shown in Figure 2.

In this second scheme, the HHPS is located just upstream of the REAC. The HHPS liquid, which is the highest boiling
portion of the reactor effluent, is routed hot to the fractionation section. The HHPS vapor is water-washed and cooled in the
REAC, prior to flowing to the CHPS. The HHPS operating temperature (typically 204°- 288°C {400”F - 550”F}) is selected
based on several considerations, including: 1) the heat requirements of the fractionation section; 2) its impact on recycle gas
purity; 3) the deposition temperature of ammonium salts, especially ammonium chloride; and 4) its impact on the metallurgy
of the HHPS.

Because the HHPS liquid exits the reaction loop upstream of the water wash, this liquid can contain dissolved gases
such as ammoni~ hydrogen sulfide, and hydrogen chloride. In one unit with a relatively cold HHPS, about 21% of the
ammonia was dissolved in the liquid, and 79’%went with the overhead vapors to the REAC. A unit with a hotter HHPS had
6% of the ammonia dissolved in the liquid. The gases in the HHPS liquid will concentrate in the wet overhead vapor from
the stripper/fractionator column. The overhead system may now be subject to the same NHIHS/NH~Cl corrosion mechanism
and design considerations as the REAC.
 One way of limiting the quantities of dissolved gases in the column overhead system is to utilize a HLPS, as shown in
Figure 3. The gases in the HHPS liquid tend to flash into the vapor phase when let down across the level control valve. The
HLPS vapor can then be water-washed and air-cooled, prior to routing to the CLPS where there is a sour water drawoff. The
HLPS vapor air cooler is typically much smaller than the stripper column overhead condenser, thereby reducing the cost of
designing to the same criteria as the REAC. The stripper column overhead in this design will have low ammonia levels
(NH,HS <2%). The operating pressure of the HLPS should be set at a practical minimum, to achieve maximum degassing of
the HLPS liquid.

One variation on this scheme does not inject water into the HLPS overhead, and maintains the air cooler outlet
temperature above the dew point. Although this avoids corrosion of the HLPS overhead air cooler, corrosion may still occur
downstream as all the ammonia may not be removed in the CLPS sour water draw and some may go into the CLPS vapor.

Compelling process reasons often exist for including a HLPS in the design of a unit, but they extend beyond the scope
of this paper. From a metallurgical viewpoint, the benefits of including a HLPS in the design of a unit are driven by several
factors, including: 1) the concentration of ammonia in the reactor effluent; 2) the operating pressure of the reaction loop,
specifically the HHPS; and 3) to a lesser extent, the operating temperature of the HHPS. The quantity of dissolved ammonia
in the HHPS liquid is directly proportional to the vapor phase ammonia concentration and the operating pressure. The
ammonia volubility is non-linear with temperature, and is predicted to go through a minimum value somewhere between
204°C and 260°C (400°F and 500”F), because the Henry’s Law constant is at a maximums

Most stripping columns use steam for stripping, which goes overhead and condenses out. This provides water for
“dilution” of the NHIHS and in some units, may decrease the concentration to acceptable levels. However, columns that are
reboiled and reaction loops with higher NHJ concentrations would need additional water, injected into the overhead, to
achieve the desired NHqHS concentrations. Most designs limit the volume of stripping steam to stay 110 to 17°C (20° to
30”F) above the water dew point at the top of the tower (for both corrosion and process performance reasons).

CORROSION CASE HISTORIES

Summary of the Corrosion Cases Reported During an Informal Suwey (Table 1):

In 1991, a major refiner reported at NACE on a recent failure in their naphtha HDS feed stripper fractionator. They
found a band of localized corrosion in the 122° to 280”C (250”F to 280”F) temperature range, that was wasting away a
1-year-old Type 309 (UNS S30980) SS overlay over carbon steel, and penetrated the tower with a pinhole leak. They
suspected that ammonium bisulfide deposition on the tray caused the tower shell thinning. They previously had aggressive
corrosion throughout the CS tower top, which was why it was overlayed with 309 stainless. During this survey, they
reported that the initial design had a CS tower and overhead air cooler which reportedly corroded out in about one year. The
air cooler and piping were replaced with Duplex stainless steel 2205 (UNS S31803). After the 309 SS corrosion in the
tower, they put a new top on, clad with Alloy 2205 (based on corrosion coupons that had been in the tower). Type 300-series
coupons did not show good resistance. A rough estimate of the bulk NHqHS concentration was 7- 10°/0,however, it may have
been higher in localized areas. Ammonium chloride may also have played a role in this attack.

A second refiner bought the design of the above unit. They modified the process and installed alloys as part of the
original design. It reportedly has not experienced any corrosion or materials problems.

Another location had approximately 4.5% NH~HS in the stripper overhead, which resulted in corrosion in the CS air
cooler and outlet piping. The top rows of air cooler tubes were severely pitted due to wet NH~Cl deposits and the lower rows
or tubes were severely thinned by NHqHS corrosion. There was also localized thinning of the elbows in the air cooler outlet
piping due to NH~HS corrosion. Another refinery with a HHPS has 2-4% NH~HS in the stripper overhead system which is
all CS. They rely on a corrosion inhibitor, which was shown to be beneficial by corrosion probes on the inlet and outlet of
the air cooler.

One unit using a HHPS and HLPS was designed with a water wash and an Alloy 825 air cooler in the HLPS overhead
system. At the time of the survey, they reported that no corrosion problems had occurred in this area or in the CS stripper
overhead system. ~estiipper overhead accuulator water waslast tested shotilyafierst@ph 1995 andshowed O.4%
NHIHS. A few months after completing the survey, the refinery found severe corrosion in the stripper tower, the overhead
line and the overhead air cooler. The tower shell was overlayed with SS in the areas of corrosion. The line was replaced
with thicker-walled CS, and a new air cooler is currently being designed and purchased with a predicted installation within
six months. The causes of the corrosion are cun-ently under investigation, but some theories are that the tower top
temperature was lowered to below dew point, and the concentrations of NHiHS and/or NHdCl in the overhead were increased
by operational changes.

Detailed Corrosion History Case Study:

The case study unit was originally designed as a two cold separator system in FCC Feed Hydrotreating service.
Original design conditions were not in the region that indicated a HHPS would be justified. After several years of operation,
it became apparent that a hot separator would be necessary to eliminate a severe emulsionlfoaming problem in the CHPS,
and a revamp was performed.

The optimum heat recovery for the revamp project was to feed the liquid from the HHPS directly to the product stripper
on the top tray. The function of the product stripper was to remove the light ends only; no final liquid products were taken
from the column. The stripped product was then fractionated into naphtha, diesel, and DGO product streams in the next
tower. Figure 4 shows the applicable parts of the process flow diagram, along with the relevant temperatures and pressures.

While it was recognized that some ammonia would be dissolved in the hydrocarbon leaving the liquid in the hot
separator, there was a question as to how much. The presence of ammonia is probably the most significant factor in defining
what corrosion mechanisms would be present, and of what magnitude the problems might be.

The expected corrosion mechanisms were wet NHQHSand underdeposit corrosion in the top portion of the stripper, wet
H# corrosion, and general NHQHScorrosion in the ov&head system. Initial tests on the gas oils and hydrogen entering the
unit did not indicate detectable ievels of chloride.

The original product stripper, air cooler, and overhead system were carbon steel. The concentrations of NH4HS were
not expected to be very high (<30/0),and hence no upgrades were made when the HHPS was added. Nonetheless, provisions
were made to allow fiture injection of corrosion inhibitor and wash water to handle higher levels of NHqHS if the nitrogen in
the feed to the unit was increased.

Operational problems in the effluent separation system all but disappeared after the HHPS was commissioned.
However, ongoing monitoring (both with lab tests and inspection procedures) of the stripper overhead system indicated that
higher than expected corrosion rates were occurring.

The lab data showed NHqHS concentrations in the range of 4 to 6 wt% in the sour water from the product stripper
accumulator water boot. In addition, higher than “background” levels of chlorides were observed. Neither of these items
were expected.

It was important to understand exactly what was happening to properly mitigate the problem. Through the use of
different equations of state and Henry’s Law constants, the levels of ammonia leaving the HHPS entering the product stripper
were simulated. In addition, unit-wide chloride testing began on internal streams, as well as feed streams, using different
methods, and confirmed that chlorides were entering the unit with the gas oil feed. Closing a chloride balance on the unit
ended up being an impossible task, due to not knowing how much chloride was being deposited on the catalyst, and how
much was being converted to HC1 in the reactors. Nevertheless, determining where the chlorides distributed to was the most
important factor, since even ppb levels of chloride can cause salt formation at the ammonia concentrations present.

Within about 12-15 months of operation after the revamp, the on-line inspection available (radiography {RT} and
ultrasonics {UT}) indicated some isolated deadleg corrosion in the vapor line leaving the tower. Scans of the tower itself
indicated no corrosion of the top head and the first two trays. Salt build-up (probably NHdCl) in the pressure safety relief
valve (PSV) inlet line was discovered. Similar dead leg corrosion was also reported by the unit identified as Case 3 in Table
1. Inspection of the splash zone in the accumulator showed some corrosion activity, but was not conclusive. The inspection
indications were thought to be possibly blistering, dirty steel, or pitting. The vapor line leaving the product stripper overhead
accumulator showed thinning with conventional UT techniques, and part of it was confirmed to be blistering with RT. No
corrosion activity was observed in the reflux line. These observations were not all immediately explainable.

Knowing the inspection history, and exactly what feeds were entering the product stripper, plans were made to begin
design modifications to the product stripper overhead system. The decision was made to treat the overhead system with the
same design considerations as used in the design of the REAC. In the meantime, the corrosion inhibitor injection into the
vapor line leaving the tower was commissioned, as was an on-line water wash in the same line. Corrosion inhibitor injection
was also commissioned in the vapor line leaving the accumulator due to the level of blistering and the observed
corrosion/erosion in two elbows. While waiting for the next turnaround to implement the modifications, a detailed inspection
plan was developed and carried out.

Planned modifications included velocity reduction in the vapor line leaving the tower, elimination of dead legs, and
changeout of metallurgy from carbon steel to Alloy 825 (UNS N08825), including the existing carbon steel overhead
condenser. Alloy 825 was used in lieu of SS to avoid any risk of chloride stress corrosion cracking. The accumulator was
replaced in kind, with additional SS cladding in the splash zone. A mist eliminator/coalescer was installed in the accumulator
to improve phase separation.

The vapor line leaving the accumulator was a concern. Blisters on piping, in locations where testing eliminated the
possibility of cyanides, was somewhat of a mystery. Studying the pattern and location of the blisters (increasing with
distance from the accumulator) led back to NH,HS corrosion and the associated hydrogen activity. Gradual cooling of the
vapor was causing small droplets of aqueous NHqHS to condense out, which appeared at the far end of the piping run in the
elbows.

Once again, the decision was made to treat the stripper overhead system with the same design parameters used in
REAC design, including velocity reduction and materials upgrade. The size of the replacement air cooler was restricted to
the same dimensions due to air coolers in other services on either side. Velocities in the tubes needed to be reduced by more
than 40% to get below the target upper limit. One alternative considered was going to additional tube rows per pass, but flow
distribution concerns arose when considering more than two tube rows per pass. Two phase flow regimes were also
evaluated, but manipulation of process variables to achieve mist flow regime that might make additional tubes rows
acceptable was not successful. The desired velocity was finally achieved through the use of 1-!4 in. diameter tubes.

During the turnaround, visual inspections were performed to confirm on-line inspection results. Visual inspection
confirmed that the top of the tower had no significant corrosion (near the HHPS liquid inlet or the splash zone at the reflux
return). This confirmed the original assessment that feeding the top tray directly from the HHPS, at a temperature more than
120°F above the calculated water dew point, mitigated the possibility of corrosion in the tower.

The corrosion activity in the accumulator splash zone was confirmed to be blisters.

The carbon steel air cooler removed from service had been in service only 10 months. The bundle was tested and
underwent selective internal rotary ultrasonic (IRIS) inspection. The air cooler design was a two-pass exchanger, with two
rows per pass. Tubes were 1 in. diameter. Wall losses in the tubes are shown in table below.

Inspection Results on 10-Month Old

Striuuer Overhead Air Cooler Tubing

Wall Loss 0/0of total tubes

o 0
1-20’% 37.2
21-40% 48.8
41-60?40 7.9
>600/0 0.6 (1 tube)
Not tested 5.5 (9 tubes)
 The wall loss was all on the ID and consisted of sca~ered pitting with length, made up of both groove-like pits as well
as wide shallow pits. This wall loss is considered significant as the bundle was only in service ten months. The wall loss (pit
depth and fi-equency) was far more severe in the first pass than the second pass.

The areas of piping elbow corrosion detected by the on-stream inspection program were confined to be present, and
had a dramatic, localized appearance, and no additional areas were detected by the visual inspection.

CONCLUSIONS AND RECOMMENDATIONS

1. NH,HS corrosion in hydrotreaters is well-known in the REAC areas, but in units with HHPS’S, NHqHS and/or NH~Cl
corrosion can also occur in the HZSstripper column overhead and/or the HLPS overhead system. Corrosion problems
can be severe. When a HHPS is part of a new unit design or is added to an existing unit as a revamp, the NHqHS and
NHiCl contents predicted for these areas should be reviewed.

2. The concentrations of NHqHS which can cause corrosion in stripper or HLPS overheads, are apparently similar to the
REAC. Some rough guidelines are: <2°/0has a negligible risk of corrosion, at 2°/0to about 8°/0,CS can be used with
added corrosion prevention steps, and systems with >8V0may need to be alloyed to be reliable.

3. The corrosion prevention steps needed to minimize the corrosion risk with CS are also similar to those used on
REAC’S. A water wash with high quality water, proper rates and proper distribution are key corrosion control steps.
Maintaining proper flow regimes and/or velocities are important, and inhibitors or other additives are used in some
units. Alloys used for upgrading,”are also similar to REAC’s.

REFERENCES

1. Piehl, R.L., “Survey of Corrosion in Hydrocracker Effluent Air Coolers,” Materials Performance, January 1976, p. 15.
2. Alvarez, A.M. and Robertson, C.A., “Materials and Design Considerations in High Pressure HDS Eflluent Coolers,”
Materials Performance, May 1973, p. 16.
3. Ehmke, E., “Corrosion Correlations with Ammonia and Hydrogen Sulfide in Air Coolers,” Materials Performance, July
1975, p. 20.
4. Damin, D.G. and McCoy, J.D., “Prevention of Corrosion in Hydrodesulfurizer Air Coolers and Condensers,” Materials
Performance, Dec. 1978, p. 23.
5. Scherrer, C., Durrieu, M. and Jamo, G., “Distillate and Resid Hydroprocessing: Coping with Corrosion with High
Concentrations of Ammonium Bistdfide in the Process Water,” Materials Performance, Nov. 1980, p. 25.
6. Shargay, C.A., Bagdasarian, A.J., Coombs, J.W. and Jenkins, W.K., “Corrosion in Hydroprocessing Units,” NACE
International Symposium on Corrosion in the Oil Refining Industry, Houston TX, Sep. 1996.
7. Turner, J., “Design of Hydroprocessing Effluent Wash Water Systems,” NACE International CORROSION/98, Paper
98593, San Diego CA, Mar. 1998.
8. Tremper, Kevin K. and Prausnitz, John M., “Volubility of Inorganic Gases in High-Boiling Hydrocarbon Solvents,”
Journal of Chemical and Engineering Data, Vol.21, No. 3, 1976, p. 295.
 TABLE 1

SUMMARY TABLE ON CORROSION CASE HISTORIES

PART 1 – PROCESS DATA

CASE UNIT SIZE TYPE START FEED FEED FEED HHPS HHPS CHPS STRIPPER
(BPSD) (Note l) UP API SULFUR NITROGEN TEMP. PRESS NH,HS C)HNHqHS
YEAR GRAV. (wt. %) (wppm) (c) (MPa)
1 Cat Feed Hydrotreater 30,000 2 1990 18 2.6 6800 224 10.7 6-8 -lo%
2 Cat Feed Gas Oil 48,000 3 1995 21 2.2 3700 212 11.8 0.4%
Hydrotreater
3 Cat Feed Hydrodesulfurizer 70,000 2 1986 21 1.3 1800 204 10.2 8-1 ()~o 2.5?4.
4 Vacuum Resid Desulfiu-izer 70,000* 3** 1976 18 2.6 5400 360 13.1 12-20%
5 Naphtha HDS 7-10?40
6 Naphtha HDS 7-1o%
7 Cat Feed Hydrotreater 4.5%

NOTES:

1. The numbers refer to Figures 2 and 3, and basically indicate how many separators are present in the effluent separation section of the unit. Figure 2 has two
separators and Figure 3 has four separators.

* This unit was originally designed for 24,000 BPD, and has been extensively debottlenecked.

** This unit deviates from Figure 3 in that there are no feed/effluent exchangers upstream of the HHPS, hence the HHPS and HLPS are at approximately the
reactor outlet temperature.
 TABLE 1

PART 2 – CORROSION HISTORY

ZASE PREDICTED CARBON STEEL CORROSION CORROSION

CORROSION CONCERN AREAS HISTORY CONTROL METHODS
1 - REAC and piping - Various problems, including leaks, in the REAC piping. REAC and piping have been upgraded
- Stripper OH cond. and piping - Severe corr in Stripper OH cond. and piping to Alloy 825, with water wash, velocity
Corrosion also occurred due to NH4CI limits, etc.
Initially, Stripper OH corrosion was
reduced by adding a water wash and
corrosion inhibitor,
Now, the condenser and sections of
piping have been upgraded to Alloy 825.
2 - REAC and piping None - REAC, HLPS Vapor Cond. and piping
HLPS Vapor Condenser and were designed as Alloy 825, with water
piping wash (at both locations), velocity limits, etc.
3 - REAC and piping Corrosion in Stripper OH piping on branch connection to - REAC and inlet piping were designed as
Stripper OH cond. and piping Psv. Alloy 800. Inlet piping has been upgraded
Corrosion activity (short term) detected on corrosion probes to Alloy 825, after SCC on Alloy 800.
on outlets of REAC when water injection rates were Stripper OH utilitzes corrosion inhibitor.
temporarily reduced and NHqHS concentration increased.
Corrosion activity also shown on probes in the Stripper OH
when inhibitor was removed.
4 - REAC and piping Corrosion and leaks in piping elbows in vapor product line - REAC is Alloy 825.
. Stripper OH cond. and piping leaving the Stripper OH accumulator. 80-240 mpy. Corrosion - Stripper OH cond is CS and has lasted
began after 1993 debottleneck modifications. 22 years to date.
Moderate corrosion in reflux line.
5 - Stripper OH cond. and piping - Initial design had CS for the Stripper OH cond. and piping, - Upgrading to 2205.
- Probably REAC and piping, but which corroded out in about one year. It was extensively
not enough information was replaced with Duplex SS 2205. 300 SS cladding in the tower
supplied to comment top also corroded and was replaced with 2205 clad plate.
6 - This unit is based on the design of - None - Alloy was used in original design.
the unit given in Case 5.
7 - Stripper OH cond. and piping - Stripper OH Cond. and outlet piping showed corrosion. It was - Many areas were upgraded to Alloy 20.
- Probably REAC and piping, but due to both NHqCl and NH,HS.
not enough information was - CLPS vapor outlet piping showed severe corrosion after unit
supplied to comment revamp resulted in higher ammonia levels.
 WASH
WATER

REACTOR RECYCLE
b
EFFLUENT GAS
4
REAC
FEED/EFFLUENT
EXCHANGERS

EFFLUENT LIQUID TO
SOUR ➤ FRACTIONATION
WATER SECTION

FIGURE 1- Effluent system with no HHPS. Usually in these units, the only area with
NH4HS corrosion concerns is the REAC system.

WASH
WATER

I &
➤ RECYCLE GAS
REAC

HHPS
REACTOR
EFFLUENT

FEEDI
EFFLUENT EFFLUENT LIQUID TO
EXCHANGER v
➤ FRACTIONATION
I
1 SOUR SECTION

Ii WATER

B
I
1
.------

I HOT EFFLUENT
➤ FRACTIONATION
LIQUID TO
SECTION

FIGURE 2- Effluent system with two separators; a HHPS and a CHPS. This design often leads to potential
NH4HS corrosion in the stripper column overhead (in the fractionation section).
 WASH
WATER

REAC
I t I
➤ RECYCLE GAS

CHPS
HHPS

=~~~-%EE
I r I I
I

FEEDI
H17 h
I
I
I

EFFLUENT
EXCHANGER

r+
1

l-’ I
I
I
I
I WASH
SOUR
WATER I
I
------- . WATER

SOUR GAS
, 1 q~,
➤ TO AMINE
TREATING
REAC t v
f

HLPS

I
I
1
I
I
------ -

3 L
c1 I
I
I
I
+ L yI&NT
LIQUID TO
I SOUR

-------
k) : I
I
I
1
WATER FRACT.
SECTION

LPlwiR”ENT FRACT.SECTION

FIGURE 3 -Effluent system with four separators. With this design, both the HLPS vapor and the stripper
overhead systems should be reviewed for potential NH4HS corrosion concerns.
 ~. RECYCLE GAS
WASH

1
WATER

CHPS
20C
10,6

lJ7
MPa
HHPS cl
REACTOR
EFFLUENT 224C
10.7
MPa
FEEDI

3,
I

P
I
EFFLUENT c1 -.-----

EXCHANGER

i
1
(
I

0 ------ - t
STRIPPEROH CLPS
20C 0.17MPa

213C I
I
0.07 MPa I
-----.
1)

=+L--
SOUR 4
WATER

H
SOUR $
WATER

u
———

STRIPPER + 216C
STEAM

FRACTIONATOR
FEED

FIGURE 4- Process Flow Diagram of the unit described in the detailed case history.