Risk Reduction Projects Combating

High Temperature Hydrogen Attack

This paper describes case study projects where Johnson Matthey (JM) has assisted operators in
reducing or removing the risk of high temperature hydrogen attack (HTHA) by design of upgraded
equipment or by changing the process operating conditions such that a required safety margin can be
reinstated. The following plant case studies are reviewed with learnings gained from these projects
over several years: a reformer upgrade project replacing non-PWHT carbon steel refractory lined
equipment; the replacement of carbon-½molybdenum (C½Mo) equipment found to be suffering
HTHA and; a change of catalyst allowing new process conditions reinstating the safety margin of
derated C-½Mo equipment

John Brightling, Stephen Shapcott

Johnson Matthey Plc

Introduction

M anaging the risk of damage due to

HTHA presents a risk to the synthesis
gas industries (ammonia, methanol,
hydrogen/CO). The likelihood of HTHA occur-
ring has proven difficult to predict, as witnessed
by various updates to the relevant design codes
over the years. Also, detection of HTHA by in-
spection is itself problematic.

It is well known that HTHA affects carbon and Figure 1: Failure of heat exchanger at Tesoro
low alloy steel equipment and piping. A recent (Image from U.S. Chemical Safety Board Final
catastrophic release of hydrogen which was at- Report to Tesoro Refinery fatal explosion and
tributed to HTHA was presented by the Tesoro fire).
Anacortes incident in 2010. The consequences
are illustrated by Figure 1, from the US Chemical
Safety Board’s final report, which shows the hy- HTHA – What is it?
drotreater exchanger that failed at Tesoro and re-
sulted in seven fatalities. The incident was inves- High temperature exposure of the carbon and
tigated in detail by the US Chemical Safety low-alloy steels used for piping and pressure ves-
Board (CSB) [1]. The incident also led to a revi- sels in hydrogen service leads to a special form
sion to API 941[2]. The Tesoro incident serves of degradation known as HTHA, or simply ‘hy-
as a powerful reminder of the damage potential drogen attack’. HTHA causes degradation of the
from unchecked HTHA damage. material at elevated temperatures and can result
in sudden and catastrophic brittle fracture.

2018 135 AMMONIA TECHNICAL MANUAL

 The process of HTHA requires the dissolving of Advanced Backscatter Ultrasonic Technique
atomic hydrogen into the steel. This is normal as (ABUT), and is often used as an initial screening
all ferritic steels operating at elevated tempera- method before the use of other follow-up tech-
ture and pressure will contain dissolved hydro- niques. In an ABUT inspection, a pattern-based
gen. With the amount of hydrogen penetrating backscattering technique is used as the initial
into the steel being linked to temperature, hydro- screening method. Depending on the backscatter
gen partial pressure, and time, the higher these pattern observed, one of several follow-up tech-
are the greater the amount of dissolved hydrogen. niques, including frequency dependent backscat-
ter, direction dependent backscatter, velocity ra-
Once in the steel, the hydrogen reacts with any tio, spectral analysis and spatial averaging will be
free carbon and will reduce carbides (Fe 3 C or used to determine the cause of backscattering sig-
M 3 C in low alloy steels) in the steel to form me- nal.
thane (CH 4 ). Methane is not soluble and accu-
mulates as a gas in small pockets at grain bound- However as the 8th edition update to API RP 941
aries and inclusions causing fissuring and indicates, the most likely location of HTHA at-
resulting in a reduction in mechanical properties. tack is at or near to welds. An issue being ABUT
is unsuitable for use in these locations as this
Figure 2 from API 941 shows an example of mi- straight beam technique cannot interrogate the
cro fissuring caused internally by HTHA. full volume of the weld region. Therefore angle-
beam techniques based on backscatter and spec-
tral analysis principles should be used for welds
and heat affected zones adding to the complexity
of inspection. A challenge is that this ultrasonic
inspection requires a very high degree of skill in
interpreting pulse-echo patterns on the oscillo-
scope interface.

Nelson Curves and API 941

Figure 2: Image from API 941 showing fissures Revisions
formed as a result of HTHA linked together to The history to the development of API 941 “Nel-
form a microcrack. Decarburized regions ap- son Curves” is that operating limits were deter-
pear lighter in colour (because of an absence of mined empirically by G A Nelson of Shell De-
carbon) than unaffected regions. velopment Company and presented to API in
1949. Throughout the 1950s and 1960s they con-
tinued to be revised by Nelson.
Inspection for HTHA has proved to be problem-
atic. API 941 contains discussion on the various Although the curves can yield safe operating lim-
possible inspection techniques and concludes its, they are actually curves of indicated industrial
that no one method in isolation is ideally suited failure experience and have no safety margin. It
to detection of HTHA, especially during its incu- is the responsibility of the designer/engineer/cli-
bation stage when micro fissures are just starting ent to determine and include a safety margin. In-
to develop. Ultrasonic inspection techniques cluding a safety margin is a safe practice that
have been found to have the best chance of de- must be incorporated into a proper design.
tecting HTHA, although only once fissures have
already started to develop. One of these tech- .
niques well-documented in API RP 941 is called

AMMONIA TECHNICAL MANUAL 136 2018

 .

Figure 3: API 941 curve 1st Edition 1970 – including C½Mo curve (later lowered and finally
withdrawn in 1990).

Figure 4: API 941 curve 8th Edition February 2016 – new curve for CS with no-PWHT.

2018 137 AMMONIA TECHNICAL MANUAL

 The curves were first published as API 941 operating below the Nelson curve. The equip-
(Steels for Hydrogen Services at Elevated Tem- ment had not been post weld heat treated and all
peratures and Pressures in Petroleum Refineries the observed damage was in the heat affected
and Petrochemical Plants) in July 1970. They are zone adjacent to a weld. The latest 8th Edition of
still described in API 941, and are referred to as API 941 (2016) contains relevant information
the Nelson Curves. concerning the influence of stress on HTHA risk
and contains a new lower curve for carbon steel
The curves are based on observed performance, welded without post-weld heat treatment.
and over the years they have been revised as new
empirical data has come to light. The evolution
can be seen by comparing Figure 3 which shows Engineering Reviews
the 1st Edition 1970 curve and Figure 4 which The API 941 changes mean that the original
shows the most recent 8th Edition 2016 curve. choice of metallurgy may not be appropriate.
Plants should therefore consider performing an en-
In earlier editions of API RP 941, the recom- gineering review on equipment and piping oper-
mended operational boundary of C½Mo steel in ating under conditions covered by the most up to
hydrogen service was described by a separate date Nelson curves to confirm that the metallurgy
Nelson curve, well above that for carbon steel, is considered suitable for continued operation.
see Figure 3. One of the first major decisions [4].
which impacted ammonia plants concerned
HTHA in C½Mo steels. Due to new failures hav- API RP 581, Risk-Based Inspection Technology,
ing been documented at conditions around the is a recommended practice developed and pub-
curve, in the 2nd Edition 1977 the curve for lished by API to provide quantitative risk-based
C½Mo was lowered. Also, in the 3rd Edition inspection (RBI). The 2nd Edition published in
1983 both the C½Mo and 2¼CrMo curves were 2008 attempted to assist in quantifying risk fac-
lowered, with other changes having been made, all
tors. It used a parameter, P v, which was a func-
toward a more conservative position.
tion of the hydrogen partial pressure (pH 2 ), tem-
perature and time. This P v parameter has since
Through the 1980s there were more HTHA fail-
been removed from the latest version on API 581
ures of C½Mo in the safe area of the Nelson
3rd Edition 2016.
curves, leading the API Committee to withdraw
the C½Mo curve in the 4th edition API 941, 1990
The 2016 revision adopts what is viewed as a
revision, downgrading C½Mo to the same curve
much more conservative approach, especially for
that applies to common carbon steel grades.
carbon steels and C½Mo steels for which the lat-
est version of API 581 assigns high susceptibility
This meant a previous benefit in terms of hydro-
to equipment operating above 177ºC (350ºF) and
gen resistance, allowing a higher operational
a hydrogen partial pressure exceeding 0.345Mpa
temperature for C½Mo steels versus carbon
(50psia). These parameters are considered very
steels, was removed and that C½Mo steel should
conservative with respect to the both the PWHT
be treated the same as carbon steel with respect
and non-PWHT carbon steels curves in the AP
to operating temperature and HTHA resistance.
941.
The second significant change to the Nelson
Typically, an engineering review includes an as-
curves has come following more recent incidents,
sessment of susceptibility to HTHA, risk assess-
including the Tesoro incident, where investiga-
ment and management plan.
tions concluded that the failure of equipment oc-
curred due to HTHA in a carbon steel that was

AMMONIA TECHNICAL MANUAL 138 2018

 Susceptibility review deemed high risk in the risk assessment. Solu-
tions may involve:
In API 581 3rd Edition 2016, an example guide-
line uses 27.7ºC (50ºF) increments to represent 1. Increased inspection at outages, possibly in-
relative change in susceptibility ranking for Cr- cluding destructive testing where equipment
Mo low alloy steels (Figure 5). is retired or by taking a ‘boat-sample’ from
the equipment.
2. Replacement of equipment with more re-
sistant metallurgy.
3. Reducing operating conditions, usually tem-
perature, as HTHA is closely influenced by
operating conditions.
4. Establishing Integrity Operating Window
(IOW) to be fully understood in operations
including the actions in the event of any ex-
cursions.

In many cases the safest course of action will be

to consider upgrading to more resistant metal-
Figure 5: An example of HTHA Susceptibility lurgy. The first two case studies cover equipment
Rankings for Cr-Mo Low Alloy Steels, API 581 upgrade projects to reduce risk, and the third case
3rd Edition 2016. study is an example of changing operating condi-
tions to reduce risk.

Whilst API RP581 provides this screening crite-

ria the owner-user has the responsibility to deter- Refractory Lined Systems
mine the applicability for their assets by risk as- With the latest amendment to API 941, an area of
sessment and inspections. risk to plants that needs additional management
attention is refractory lined equipment. A typical
Risk Assessment example of such a system is the transfer line from
For all equipment identified as being susceptible the primary reformer to the secondary reformer
to HTHA, a risk assessment is undertaken using which historically was often designed as a refrac-
a standard risk methodology to develop a risk tory lined (non-PWHT) carbon steel system. As
matrix considering likelihood of failure occur- originally designed, the carbon steel part of the
ring versus loss consequence of it. transfer line would operate with a shell tempera-
ture below 200ºC (392ºF). However, over time
Management Plan the refractory may degrade, crack and deteriorate
meaning that shell temperatures have increased.
Based on the level of risk, equipment is targeted Typically, such equipment is inspected by infra-
with risk management plans developed with the red thermal imaging inspection. Figure 6 shows
primary objective to reduce items of equipment an example of a transfer line operating with hot
areas.

2018 139 AMMONIA TECHNICAL MANUAL

 limit the shell temperature to less than 180ºC
(356ºF) in still air.

The extent of the equipment replacement in the

refurbishment project is shown in Figure 7.

Figure 6: Thermal Image scan showing transfer

line operating with hot areas.

In such circumstances, the steel casing is operat-

ing with areas exceeding 240ºC (464ºF) and out-
side the recommendation of the latest API 941 for
non PWHT-carbon steel. Figure 7: Case Study 1 - Equipment replacement
in the refurbishment project.
Case Study 1
For a European ammonia plant designed and Methanation Change in Vessel
built in the 1970s with a refractory lined, non-
PWHT carbon steel transfer line system which A catalyst vessel in the ammonia plant which has
had areas with refractory hotspots, a major refur- proven vulnerable to HTHA is the methanator.
bishment project for the transfer system was jus-
tified by the plant to improve plant reliability and Case Study 2
integrity. The project being scheduled to coin-
cide with a radiant section re-tube. A European ammonia plant [4] was commissioned
in 1970 and is currently operating at 1050 MTPD
To support the project JM performed an engi- (1160 STPD). As was typical for plants designed
neering design for the following items of the re- in the 1960s, the original material selected for the
methanator in hydrogen service was C½Mo, based
former - inlet pigtails, replacement reformer tube
on the Nelson curves published in API 941 at the
supports, reformer tubes, outlet pigtails, hot col-
time.
lectors (outlet manifolds) and the replacement re-
fractory lined transfer mains. The vessel had been subject to regular inspections
during its working life. Because the lifetime of the
The new transfer mains were designed to have a methanator catalyst is long, internal inspections
larger internal diameter to enable easier inspec- were infrequent.
tion and future maintenance as well as a lower
pressure drop. Some cracking was initially detected during an
internal inspection in 1992. At the time, metal-
With a new pressure shell design based on lography indicated the cracks could be original
1¼Cr½Mo material, temperature of concern with plate defects so it was unclear whether HTHA
respect to HTHA would be circa 450ºC (842ºF). was occurring. Further inspection was called for
The associated refractory system was designed to at the next inspection opportunity in 2000.

AMMONIA TECHNICAL MANUAL 140 2018

 The cracking found during the internal inspection The replacement vessel was commissioned in
in 2000 extended on either side of the 1992 2005 with the KATALCOTM 11-series methana-
cracking for over five times the length, exceeding tion catalyst charge achieving <1ppm CO+CO 2
half the circumference of the vessel (>2700 mm, slip which it has maintained over its 12 years run.
>106 inches). The height of the band had ex-
tended from 105 to 600 mm (4 to 24 inches). Iso- In addition to the vessel and catalyst replacement,
lated cracks were ground out to determine their a safety integrity level (SIL) assessment was per-
depth. formed and the instrumentation was changed
providing additional layers of protection in both
Following ultrasonic examination in February the initiating and final elements of the instru-
2003, the cracking above the band was concluded mented protection system [5]. Part of the instru-
to be present in excess of 10% of the wall thick- mentation upgrade included fast acting thermo-
ness of the vessel. It was clear that the vessel was couples to detect the very high rate of
suffering from HTHA. A complete vessel repair temperature rise in a methanator if a process up-
would require complete vessel post weld heat set results in CO 2 breakthrough from the up-
treatment, which would not be possible within stream removal section [6].
the timescale of a plant shutdown. The repair
would also retain risk so replacement was Recently in 2017, a similar project has been com-
planned for the 2005 shutdown. pleted for a North American ammonia plant. The
plant which dated from the 1960s has an existing
A replacement vessel was designed by JM to C½Mo methanation vessel. JM assisted with re-
modern design standards and several improve- placement by designing a modern vessel to
ments in the design were made: ASME Section VIII Division 1:2017 using 1¼Cr
½Mo material (ASME SA387 Gr11 Cl2), see
• Material change to 1¼Cr ½Mo to give greater
HTHA resistance.
Figure 8 using KATALCO 11-series methana-
tion catalyst and STREAMLINE low pressure
• Design temperature increase to 500ºC (932ºF) drop media.
from 427ºC (800ºF) to improve ability to han-
dle any high temperature excursions in the
event of a process upset.
• Inlet and outlet connections and the connecting
pipework were increased in diameter to give re-
duced pressure drop.
• Catalyst now supported directly on support me-
dia in the vessel base. Previously a grating sup-
ported on a welded ring inside the vessel was
used. The welded ring promoted stresses and
cracking.

Many of these improvements simplified the ves-

sel design and construction. The improvements
have a further benefit that, looking forward to the
future, the reduction of internal features and at-
tachments to the vessel shell mean that require-
ments for internal inspection of the vessel may be Figure 8: Case Study 2 – Replacement Methana-
reduced. tor vessel.

2018 141 AMMONIA TECHNICAL MANUAL

 Methanation Change in Process 570ºF) as compared to the original design tem-
Conditions perature of 427ºC (800ºF) from the withdrawn
C½Mo curve [7].
The removal of the C½Mo curve from the API
guideline placed a large number of C½Mo equip- However, the exchanger operated from 1970 to
ment items in an indeterminate state. Good prac- 2007 right on or slightly above the Nelson curve.
tice is to maintain an operating safety margin of Therefore, regular 100% ultrasonic testing (UT)
30ºC (50ºF) below the Nelson curves at any time. was performed to check for cracks on the main
welds of the pressurised shell (i.e. longitudinal,
Case Study 3 circumferential, butt welds and man hole welds).
In addition, during each turnaround, a replica test
For a 1250 MTPD (1380 STPD) European am- was performed. However, no cracks or fissures
monia plant, the inlet/outlet heat exchanger for were revealed during these investigations.
the methanator was made of C½Mo. The maxi-
mum recommended outlet temperature dictated The change to a more active catalyst capable of
by the revised Nelson curve is 290-300ºC (550- operating at a much lower inlet temperature, re-
instated a safe operating margin below the Nel-
son curve as shown in Figure 9.

Figure 9: Case study 3 - Impact of lowering methanator operating temperature shown on Nelson curve
from API 941 6th Edition, 2004.

AMMONIA TECHNICAL MANUAL 142 2018

 By taking the process operations to a lower tem- At start-up, it was proven that the CO/CO 2 slip
perature it was possible to achieve a 30ºC (54ºF) was lower than 1 ppm and the inlet temperature
safety margin below the Nelson curve, thus low was set at 220°C (428ºF). After 10 years of con-
susceptibility as per Figure 5 tinuous operation, the CO/CO 2 slip is still low
and the inlet temperature remains at 220°C with
To achieve this required lowering the inlet and a sharp and steady reaction profile, Figure 11.
exit temperatures of the methanator by 65ºC
(117ºF) by lowering the inlet temperature to
220ºC (428ºF), as shown in Figure 10.

Figure 11: Methanator exotherm profile.

In this case study due to operation with the new

catalyst charge at low temperatures, C½Mo steel
equipment is now operating with a safer margin
within the Nelson curve.

Conclusion
On-going inspection for HTHA can be a difficult
task requiring significant expertise. Depending
Figure 10: Process flow scheme around on the age of the plant many items of equipment
methanator. can be affected. Appropriate asset management
may require planned capital expenditures for
The driver behind the choice of KATALCO 11- timely renewal of vulnerable items of high risk
series as a low temperature methanation catalyst, equipment.
was reinstatement of a safe operating margin in
respect of Nelson curves. A second benefit was The case studies give some examples of how the
the improved plant efficiency due to HP steam risks inherent from HTHA have been managed as
savings on the trim heater. part of asset integrity and renewal programs at a
number of sites, with JM providing support in the
By installing KATALCO 11-series low temper- design of the new equipment constructed from
ature methanator catalyst, less duty is required inherently more resistant metallurgy.
from the trim heater fed by HP steam (100 bar,
1450 psi), as shown in Figure 10. A gain in HP An example was provided where selecting a cat-
steam consumption of 2-3 ton/h is obtained, de- alyst with a lower operating temperature range
pending on the temperature rise over the allowed a safe operation window to be estab-
methanator catalyst. This corresponds to an im- lished by moving from a high susceptibility to
provement of the energy efficiency of 0.15 low susceptibility risk and eliminating the need
GJ/ton (0.14 mmBTU/ton). to purchase new equipment.

2018 143 AMMONIA TECHNICAL MANUAL

 References: [5] M Walton, T Southerton & P Sharp. Safety
Improvements in a Methanation Reactor ,
[1] US Chemical Safety and Hazard Investiga- AIChE Ammonia technical manual Vol 48,
tion Board Report 2010-01-IWA, May 2014 2007
(www.csb.gov/tesoro-refinery-fatal-explo-
sion-and-fire/) [6] A Janssen, N Siraa, J M Blanken, Temper-
ature Runaway of a Methanator Ammonia
[2] API Recommended Practice 941. Steels for technical manual Vol 23 , 1980
Hydrogen Service at Elevated Temperatures
and Pressures in Petroleum Refineries and [7] S Van Den Broeck, M Fowles. Low temper-
Petrochemical Plants. Eighth Edition Febru- ature methanation: Operating experience and
ary 2016. something unexpected!, AIChE Ammonia
Technical manual Vol 49, 2008
[3] API Recommended Practice 941. Steels for
Hydrogen Service at Elevated Temperatures
and Pressures in Petroleum Refineries and
Petrochemical Plants. First Edition July
1977.
[4] D Keen, C Jones & C Thomas. Inspection for
high Temperature Hydrogen Attack, Proc Ni-
trogen+Syngas 2018 conference, Gothen-
burg, February 2018.

KATALCO and STREAMLINE are trademarks of the Johnson Matthey group of companies

AMMONIA TECHNICAL MANUAL 144 2018