Page 237 - Gas Purification 5E

Mechanical Design and Operation of Alkanolamine Plants 223

In 1984, an amine LPG treater ruptured, causing an explosion and fire that killed 17 people
(McHenry et al., 1986; 1987). Although the disaster was apparently caused by a combination
of HIC and SOHIC, this disaster and reports of extensive cracking in other amine systems
(Gutzeit and Johnson, 1986; Anon, 1985) initiated investigations of carbon steel cracking in
amine units. These investigations, which revealed many instances of ASCC, were undertaken
by NACE Group Committee T-8 on Refining Corrosion in cooperation with the API.
Results of an industry survey by this group were reported in 1989 by Richert et al. This
survey included 294 amine units, 272 of them in refineries, and most of them MEA and DEA
units. This survey indicated that cracking occurred primarily in MEA service. Nearly every
crack was associated with a weld that had not been PWHT. Cracks occurred in every type of
equipment and at temperatures as low as ambient. However, no conclusive correlation was
found between cracking and most process variables, including source of gas, amine concen-
tration or acid gas loading, use of filters or reclaimers, use of corrosion inhibitors, type of
carbon steel, or addition of caustic to neutralize heat-stable salts. ASCC was found to be
generally independent of the HzS/COz ratio. However, all amine plants with less than 1
mol% HzS in their feed experienced cracking. This suggests that at least 1 mol% H2S in a
predominantly COz stream has some tendency to inhibit ASCC. This conclusion was con-
firmed by later laboratory investigations (Schutt, 1988; Parkins and Foroulis, 1988; Lyle,
1988). The survey data could not be used to demonstrate that ASCC was temperature depen-
dent because most of the equipment operating at higher temperatures was PWHT. However,
the experience documented by Garwood (1953) suggests that ASCC increases with increas-
ing temperature when equipment and piping are not PWHT. Since 98% of the cracks report-
ed in the survey occurred in carbon steel welds that had not been PWHT or at nozzles where
PWHT is difficult, PWHT of all carbon steel piping and equipment in amine service was
judged to be the single most effective measure to prevent ASCC. Later, additional data on
DEA and DIPA gas treating units, which included data on ASCC below 66°C (150°F) for
both DEA and DIPA, confirmed that PWHT is required to prevent ASCC in plants using
these amines (Bagdasarian et al., 1991).
A report by the Southwest Institute (SwIU, 1989) and a related article by Lyle (1988)
point out that the NACE survey includes data from only 16 natural gas treating units out of a
total of 294 units of all kinds. These references state that a more comprehensive survey was
made for natural gas treating units, but gave the results of the survey only in brief summary,
and did not state how many units had been surveyed. Both reports did, however, give
detailed results of a laboratory study, which concluded that ASCC of carbon steel was inhib-
ited by the presence of HzS. Also, according to both reports, ASCC in refineries occurs pre-
dominantly in lean amine solutions; whereas, ASCC cracking in natural gas plants occurs
primarily in rich amine solutions.

Corrosion Inhibitors

Corrosion inhibitors are often classified as cathodic inhibitors, which inhibit reaction 3-2,
anodic inhibitors, which inhibit reaction 3-1, and oxidizing passivators, which are discussed
in the next paragraph. Cathodic and anodic inhibitors are often adsorbed on the corroding
metal, like filming amines, or plated out on it, like arsenic and antimony. They have been
recommended and patented at various times, but none has had much commercial success in
amine units (API, 1990).
Oxidizing passivators are considerably stronger oxidizing agents than hydrogen ion, and they
operate by shifting the potential of the steel to a more positive value where reactions 3-1 and 3-2