Carbon
Steel and Type 304 stainless steel samples were exposed in thirteen corrosion
tests conducted in solutions (with varying pH, oxygen content, SO2
content and chloride ion content) in a simulated sour water solution produced
in the laboratory. Corrosion rates were calculated, and U-bends were examined
for stress-corrosion cracking.
The
sour water handling system in chemical plants often show swings in pH, oxygen
content, SO2 content and chloride ion content, or different
locations. All of these variables theoretically affect the corrosivity of the
sour water, and they could be monitored and controlled to minimize corrosion in
the handling system.
The
objective of experimental study of sour water systems was to determine the
relative importance of the different possible control parameters with regard to
their effect on corrosion.
Specific
variables of interest were pH, temperature, oxygen content and inhibitor
concentration. Carbon steel, the primary material of construction in the
system, and AISI Type 304 stainless steel were to be studied. The possibility
of stress-corrosion cracking on Type 304 Stainless steel was to be addressed.
Corrosion
tests were carried out on flat coupons of AISI type 1010 carbon steel and AISI
type 304 stainless steel. Coupon dimension were 1 X 0.5 X 0.06 inches. In the
elevated-pressure tests, U-bend specimens were made of 3 X 0.75 X 0.062 inch
type 304 stainless steel sheet in the solution-annealed condition, as described
in ASTM G30. After bending over a mandrel, the arms of each U bend specimen
were pulled parallel by stressing with Hastelloy c screws and nuts.
For
each test duplicate specimens of both alloys were cleaned in acetone, dried and
weighed to the nearest tenth of a milligram. Samples for atmospheric pressure
tests were suspended in 1 liter Ehrlenmeyer flasks using Teflon tape.
Samples
for elevated-pressure tests in autoclaves were mounted on Stainless steel rods
with non-conducting Teflon washers. Duplicate samples of carbon steel and 304
stainless steel were exposed in each solution. In the autoclave tests duplicate
U-bends of 304 stainless steel were also tested.
Test
solutions and conditions are listed in Table 1. Solution
drawn from the 11T5 outlet stream was used for two tests, one at 110 F and one
at 190 F, to determine the relative effect of temperature in this location.
The
bottom stream from the 11T4 column was used for three tests: one as received at
190 F, one at 190 F with 250 ppm of IPC 2625 inhibitor added, and one at 190 F
with air sparging to simulate leakage of air into the system from the 11T13
vacuum column.
Solution
drawn from the 11E2 fin-fan condenser was used for three tests conducted in
autoclaves at 250 F. One test used solution drawn from the condenser outlet
as-received. In the second test, 11E2 inlet solution was acidified to pH 3 with
HCl, to simulate high-chloride excursions due to desalter outage. In the third
test, 250 ppm of IPC 2625 inhibitor was added to the acidified 11E2 inlet
solution described above.
Bottoms
drawn from the 11T6 desalter were used in two atmospheric pressure tests at 110
F. In one, the as-received solution was acidified to pH 4 with acetic acid. In
the other, NaOH was added to the as-received solution until the pH equaled 7.0.
Three
atmospheric pressure tests were conducted at 190 F in a simulated sour water
solution prepared by sparging SO2 through distilled water for
one-half hour. pH of these laboratory solutions was then adjusted with acetic
acid or NaOH, as needed, to produce three solutions with pH levels of 3.5, 4.0
and 4.5.
All
atmospheric pressure tests were conducted for one week. The autoclave tests
were held at temperature three weeks. After the test exposure the samples were
cleaned and re-weighed. The U-bends from the autoclave tests were cleaned after
exposure and then examined under a 20 X binocular microscope for evidence of
stress-corrosion cracking.
The
results of these corrosion tests are presented in Table 2.
Calculated corrosion rates are the average of the two exposed coupons. The
variation between the two coupons in the same environment is also listed in Table 2. Corrosion
rates on carbon steel in all these tests were significantly higher than the
rates recorded in the actual sour water handling system equipment. Films
forming on the steel surfaces apparently lead to a slow decrease in corrosion
rate with time. Longer term tests may be needed to predict field corrosion
rates. The results obtained in these tests should, however, be useful for
predicting the relative effects of the different variables.
Decreasing
the temperature of the 11T5 outlet solution from 190 F to 110 F resulted in a
28 percent decrease in corrosion rate, from 18.5 to 13.4 mils/year.
Air
sparged into the 11T4 bottoms solution increase the corrosion rate by 400
percent, from 21.8 mpy to 86.6 mpy. This is probably due to the formation of
polythionic acids from sulfides. At 250 ppm, the Ipc 2625 inhibitor was not
effective in lowering corrosion in 11T4 bottoms solution; in fact, the
calculated corrosion rate on the "inhibited" solution was slightly
higher that the rate on the as-received solution.
The
11E2 outlet condensate solution, tested at 250 F, was the most corrosive rate
of 87 mpy on carbon steel. Lowering the pH in this solution to 3.0 did not
increase the corrosion rate; in fact, the calculated rate on carbon steel dropped
fom 87 mpy in the as-received solution to 66.7 mpy in the pH adjusted solution.
The solution with 250 ppm IPC 2625 inhibitor showed significantly higher
corrosion rates than a comparable solution without inhibitor.
Raising
the pH of the 11T6 bottoms solution form 4 to 7 lowered the corrosion rate
about 25%, from 15.5 mpy to 11.9 mpy.
The
simulated sour water solutions mixed in the laboratory were far more corrosive
than any of the plant solution samples. This suggests that plant solutions are
either significantly lower in sulfur-based acids than the near-saturated
laboratory solution, or that there are organic compounds in the plant solutions
which act as natural inhibitors to some extent.
Raising
the pH of the simulated sour water solutions from 3.5 to 4.5 lowered the
corrosion rate slightly, from 360 mpy to 335 mpy.
It
was noted that the corrosion rate produced on the carbon steel samples tested
in as-received plant solutions was largely dependent on temperature rather than
the location at which the sample was drawn. Corrosion rates in the 11T6
bottoms, drawn from the desalter at the head of the system, were not
significantly different from the rates in the 11T5 outlet solution, which is
downstream of all other operations in this area. This suggests that the
concentration of corrosive species remains relatively constant from the
desalter through the 11T5 column, and that local rates of meal loss will vary
most significantly with temperature and flow rate.
No
evidence of stress-corrosion cracking was observed on the u-bends of 304
stainless steel exposed in the autoclave tests, and 304 stainless coupons
showed no significant weight loss corrosion in either the glassware or
autoclave experiments. No significant pits were detected on any of the 304
stainless steel coupons. However, crevice corrosion under the stressing-bolt's
washer occurred on one of the U-bends in the 20 psig 11E2 outlet solution
autoclave test, and on one U-bend in the autoclave test 11E2 inlet solution
acidified to pH = 3.
The
depth of attack on the U-bend immersed in the as-received 11E2 outlet
condensate indicated a crevice corrosion rate of 12.3 mpy in this solution. On
the U-bend in the acidified 11E2 inlet solution the calculated crevice
corrosion rate was 71 mpy.
The
classic defense against localized corrosion as seen on the two U-bends is, of
course, the use of molybdenum-bearing stainless steels such as type 316.
The
test results presented herein led to the following conclusions:
- The corrosivity of the sour
water itself does not change significantly between the 11T6 desalter and
the 11T5 column.
- Corrosion of carbon steel in
the sour water is relatively insensitive to pH. In the range pH = 3.5 to
pH = 7.0, corrosion rates go down with increasing pH but only about 25
percent.
- Corrosion rates of carbon steel
in the sour water are directly proportional to temperature, increasing 640
percent from 110 F to 250 F.
- Additions of 250 ppm of IPC
2625 inhibitor were ineffective at reducing corrosion rate of carbon steel
in 11T4 bottoms and 11E2 inlet solutions. Observed corrosion rates in the
"inhibited" solutions were actually greater than those from
comparable non-inhibited solutions.
- Air leaking into the 11T4
column could increase the general corrosion rate of carbon steel 400 percent
due, apparently, to increase polythionic acid formation.
- Type 304 stainless steel showed
low general corrosion rates and no stress corrosion cracking in all
solutions to which it was exposed. However, crevice corrosion was observed
on U-bends in the autoclave tests, suggesting the need for molybdenum
bearing stainless steels for these applications.
TABLE 1
Test Solutions
* as received
** 250 ppm IPC 2625 inhibitor added
+ air sparged throughout test
++ pH adjusted with HC1
# pH adjusted with HAC or NaOH, as necessary
TABLE 2
Test Results
* as received
+ air sparged throughout test
++ pH adjusted with HC1
# pH adjusted with HAc or NaOH, as necessary