The sweetening unit is a key component of the gas processing plant, and the amine regeneration tower is one of the major parts of this unit. The amine tower in the Sarkhoon gas plant in southern Iran has experienced severe corrosion [4050 mils per year (mpy)] over a 2.5-year period. The corrosion area is limited to the vapor and vapor/liquid interface spaces. The material in this area is carbon steel (A-516-GR-70) and is located on the inside of the tower shell.
A recent study investigated the cause of this corrosion, along with the corrosions tendency to move from the top trays of the tower to the bottom trays. The performance of the chosen corrosion inhibitor was also scrutinized during the investigation. Results showed that an amine solution in the unit was degraded by dissolved oxygen (O2), exposing the inside of the tower shell to corrosive amine degradation products. The injection of an O2 scavenger inhibitor into the amine solution has not corrected the problem, and design issues with the injection point and its piping have been discovered.
Corrosion sources in a gas plant
A portion of the Sarkhoon gas plants feed is sour gas. In the feed-sweetening process (Fig. 1), a solution of 70% water and 30% diethanolamine (DEA) is used to absorb H2S in the contactor tower (C-101). Absorbed H2S is released in the amine regeneration tower (C-103) in the presence of rising temperature and pressure drop.1 In comparison with other amines [e.g., monoethanolamine (MEA), methyldiethanolamine (MDEA), activated MDEA (aMDEA)], DEA is more susceptible to oxidation and degradation, and it produces organic acid bases that form heat-stable salts in an amine environment.2
| Fig. 1. Gas-sweetening unit at Sarkhoon gas plant.|
One of the major sources of corrosion on carbon steel vessels in sweetening units is heat-stable materials, which are a product of amine degradation.3,4 Oxygen plays a major role in DEA degradation. The reaction of O2 and DEA produces organic acids, such as acetic acid, formic acid and so on. O2 solubility in a solution of amine (except DEA) and water is similar to the solubility of O2 in water.5
Foaming in the amine regeneration tower results in the increased contact of corrosive components of the amine solution with the tower internals, as well as in a loosening of the protective iron sulfide (FeSX) surface layer that is formed by H2S. Heat-stable salts play an important role in foaming and corrosion in the amine regeneration tower.6,7
Organic acidssuch as formic acid, acetic acid and oxalic acidat a temperature of 200°C contacting carbon steel would result in severe corrosion. However, the corrosion activity of these acids is significantly reduced in an MDEA solution.3
The corrosiveness of a 20% amine solution on carbon steel in an environment containing either CO2 or H2S gas at a temperature of 140°F to 212°F would increase dramatically. However, environments containing either CO2 or H2S gas in this temperature range tend to experience more aggressive corrosion compared with an environment containing a CO2:H2S mixture of 1:3 or 3:1.8
One method to prevent corrosion in sweetening units is to use corrosion inhibitors. Corrosion inhibitors manufactured from heavy metals, such as arsenic and vanadium, have the ability to control corrosion in the aforementioned environments, although these heavy metals are incompatible with these environments. Furthermore, these types of corrosion inhibitors are unable to protect splash zones and vapor spaces. Film corrosion inhibitors prevent general corrosion, but are unable to control specific types of corrosion.8
To avoid certain corrosion types (e.g., sulfide stress cracking, hydrogen-induced cracking, etc.), guidelines contained in NACE MR-01-75 should be followed during the material selection of sweetening units.9 Organic amines are traditionally used in the refinery crude column overhead to neutralize and reduce pH as a means of corrosion control.10
Based on previous research on H2S concentration increases in oil, two types of scales form on contacting steel surfaces: Mackinawite and Pyrrhotite.11 Mackinawite scales have coarse grain in a loose and brittle form. Although Mackinawite scales reduce general corrosion, steel surfaces are still susceptible to aggressive, pitting corrosion. Pyrrhotite scales have fine grains and exist in continuous form, and are resistant to both general and pitting corrosion.12
There are three zones in the regeneration tower: the vapor space, the liquid/vapor interface and the liquid phase. In the vapor space, the formation of a protective layer of FeSX results in low corrosion rates for a rich solution containing H2S, compared with a lean amine solution that does not contain H2S.12
A number of field and laboratory tests have been performed to measure corrosion rates and levels at the Sarkhoon gas plant. Field tests included the following:
Weight loss (coupon) test. This test was carried out based on standard test method ASTM-G-1 via coupon installation in 10 points of amine solution cycle input. Test results are presented in Table 1.13
Iron count test. To evaluate the corrosion rate of the unit, this test was performed according to the spectrophotometric method. Samples were taken from the circulating amine solution, and the amount of iron (which is generated by corrosion on vessels) was measured. The following iron counts were observed: 16.9 ppm (Test 1), 10.6 ppm (Test 2), 35.0 ppm (Test 3) and 38.0 ppm (Test 4). Based on these data, the presence of an amine solution containing concentrated heat-stable salts results in corrosion reactions. According to the unit design, the maximum allowable amount of iron in the amine solution is 10 mpy.
Dissolved O2 in amine solution. The amount of O2 present in the amine solution, which is a major factor in DEA degradation, was measured according to test methods outlined in ASTM D-888 and ASTM D-5543. Results are represented in Table 2.14,15
Ultrasonic test. An ultrasonic thickness meter was used to measure tower wall thickness. The results of this test, which show a reduction in thickness, are represented in Tables 3 and 4.
Laboratory corrosion tests included the following:
Inlet gas analysis. This test was carried out according to ASTM D-1945 guidelines. A gas chromatography apparatus was used to pinpoint components of the feed gas entering the contactor tower.16 Results are shown in Table 5.
Polarization test. To determine the corrosion rate of the lean amine solution containing 3.71% heat-stable salts, three tests at temperatures of 25°C, 50°C and 70°C were performed according to ASTM G-3 methods. Respective corrosion rates of 2.0 mpy, 4.8 mpy and 8.8 mpy were observed at the various temperatures. These results are illustrated in Fig. 2.17
| Fig. 2. Lean amine polarization test at three |
Immersion weight loss. To simulate corrosion in the three tower regions, this test was conducted according to ASTM G-31 guidelines in static condition, with lean amine containing 3.71% heat-stable salts at a temperature of 120°C. The corrosion rates observed for the liquid phase, liquid/vapor interface, and vapor phase were 2.10 mpy, 5.02 mpy and 8.16 mpy, respectively.18
Scanning electron microscope. To evaluate corrosion products on the carbon steel section of the tower shell, a sample from the bracket of the downcomer of the ninth tray was prepared and tested by microscope. Test results are shown in Fig. 3.
| Fig. 3. SEM test result on a specimen from the |
Amine solution analysis. This test was performed to investigate lean amine solution components, and it is based on ion chromatography and National Iranian Gas Co.s test method for heat-stable salts. Results are presented in Table 6.
Corrosion inhibitor evaluation. To evaluate the O2 scavenging performance of a corrosion inhibitor injected into the amine cycle, this test was conducted according to methods outlined in ASTM D-888 and ASTM D-5543 for two samples of freshwater and a solution of 30% amine and 70% water. Test results are shown in Table 7.14,15
Corrosion test in acid/DEA environment. This test was conducted to evaluate the corrosion rate of carbon steel in a liquid phase contacting a lean amine solution. The solution contained organic acids at a temperature of 25°C to 30°C, which is the temperature of the circulated amine solution in the cycle. Eighteen separate samples from the fresh amine solution containing an organic acid (i.e., acetic, oxalic, malonic, formic, butyric, succinic or glycolic) were prepared, and corrosion coupons were installed. Test results showed a corrosion rate greater than or equal to 0.2 mpy for all of the above-named acids.
X-ray diffraction. Samples from a non-corroded portion of the tower were used in this test, as it was not possible to prepare samples from the corroded area. The formation of iron disulfide (FeS2) was observed, as shown in Fig. 4.
| Fig. 4. XRD testFeS2 scale.|
Corrosion in the gas-sweetening unit
DEA is used in the gas-sweetening unit. This unit was inspected one year after commissioning, at which time no corrosion or deterioration was detected. A second inspection was carried out 2.5 years after commissioning, and aggressive corrosion was detected in the vapor space of the tower shell, in the carbon steel section.
This tower consists of 20 trays. The tower shell is made up of 316L stainless steel from the top head through Tray 8, and the remaining 12 trays are annealed SA516 Grade 70 carbon steel. The tower wall thickness is 12 mm, with 3 mm of corrosion allowance. The chemical composition of the aforementioned alloy is presented in Table 8. Due to the existence of chloride ions in the feed, the chloride ion content of the circulating amine was higher than the allowed limit during the first three months of operation. An anionic resin bed was used to absorb the negative ions. After feed quality improved and the ion chloride content was reduced, the bed was taken out of service.
Corrosion was observed in several forms. Local or island-shaped corrosion resembled several concentric, closed curves, with each curve having an almost equal center and differentiation in the borders between it and the next curve (Fig. 5). The corrosion morphology was described as step-down edges. The most corroded areas were found at the centers of the curves. Here, the tower thickness was reduced by 3.5 mm to 4.0 mm.
| Fig. 5. General corrosion at vapor phase area.|
The valley form of corrosion occurred in down-flowing liquid in the vapor/liquid interface of the shell. The depth of the valleys was 5.5 mm to 6.0 mm (Fig. 6). The corrosion shape was described as a smooth valley and an escarpment.
| Fig. 6. Tower corrosion (valley type).|
Investigation of corroded areas for both forms showed:
1. Corrosion on the regions of the shell contacting the amine solution (liquid phase) was negligible. This was verified by coupon test results (Table 1).
2. Severe corrosion was observed on the tray vapor space and the liquid/vapor interface.
3. Corrosion severity in the aforementioned areas on trays 8 to 20 (top to bottom) increased in correlation with area and depth.
Inlet gas to the contactor contains a maximum of 40 ppm H2S. Software was used to determine the H2S concentration in each tray of the regenerator tower; this process was simulated based on rich amine content. Results are presented in Table 9.
In order to control corrosion, a corrosion inhibitor was injected into two lines with different pressures. These lines included a lean amine-to-contactor line with a pressure of 60 bar to 65 bar and a reflux stream line with a pressure of less than 2 bar. Oxygen scavengers function as the duty of this inhibitor as well as a protective layer (along with a surface scale of FeSX), based on the information provided by the inhibitor manufacturer.
Amine analysis results
The results of the circulating amine analysis showed that the amine was degraded. Oxidation and degradation products are generally organic acids that cause severe corrosion on carbon steel at proportional concentrations (Fig. 7). The analysis showed that the amine produced heat-stable salts after contacting organic acids, which can also cause corrosion on carbon steel (Fig. 8).
| Fig. 7. Carbon steel corrosion in various acids |
| Fig. 8. Carbon steel corrosion in various acids.|
Visual inspection and thickness monitoring detected no significant corrosion in the liquid phase of the amine stream on pipelines, towers or other vessels in the gas sweetening unit. An analysis of corrosion-monitoring results (Table 1) observed during an aggressive period of corrosion (as well as subsequent corrosion periods) shows a corrosion rate of fewer than 2 mpy in the liquid phase. Field corrosion-monitoring results were verified by immersion weight-loss tests in the laboratory.
The corrosion rate of the lean amine in the liquid phase was found to be within an acceptable industrial margin, according to a polarization test of the lean amine solution (Fig. 2), which contained 3.71% heat-stable salts. Organic acids were also present in the solution. Although the results of the polarization test showed rising temperature, the corrosion rate was less than 25% of the corrosion rate of the tower shell. Furthermore, the temperature increase was related to the rich amine stream in the unit. In a rich amine stream, the formation of an FeSX layer protects the surface. No H2S or FeSX layer was observed in the lean amine solution in either the polarization test or the laboratory weight-loss test.
Use of oxygen scavenger inhibitor. To omit O2, which is introduced into the solution by a demineralized water makeup, an O2 scavenger inhibitor was injected. Based on the characteristics of the inhibitor, the required temperature for starting a reaction with dissolved O2 is higher than 80°C, while the amine temperature in the tank and onstream before entering the regenerator tower (103°C) does not reach the specified temperature. Therefore, O2 cannot be entirely eliminated below 80°C.
Test results showed a partial decrease in the amount of dissolved oxygen in the amine solution after the inhibitor was added, which raised the temperature to 98°C for a period of 15 minutes. The O2 scavenging performance of this inhibitor at higher temperatures is in doubt, as it was found to be unsatisfactory during this test.
Additionally, the results of the dissolved O2 field test are evidence of the presence of O2 in the solution, which verifies the unsatisfactory performance of the inhibitor. The poor performance was due to the amine coming into contact with O2. In other words, the presence of oxygen is a factor in the production of organic acids.
The injection pump performs at a unique pressure, and it is not possible to inject the inhibitor into two points with different pressures, as setting the pressure on the reflux stream would result in lower-than-desired pressure in comparison with the contactor stream. If the pressure setting is based on contactor pressure (which is higher than reflux stream pressure), the high-pressure injection fluid would flow into the low-pressure stream (reflux); therefore, it cannot be injected into the contactor stream.
Causes of corrosion. As mentioned above, corrosion observed on the interior of the tower is located on the vapor space or on the vapor/liquid interface. According to an analysis performed on degraded amine that was circulated during a period of severe corrosion, there are several organic acid anions in amine that cause corrosion. Since tests cannot be viably performed on liquids in the reflux drum or on outlet gases from the top of the tower, process simulation software was used to evaluate the existence of acidic anions in gases flowing from the top of the tower. Simulation results (Table 9) confirmed the presence of these anions in the vapor space of the tower. Vapor pressure research also indicated the presence of these anions in the towers vapor space.
According to research and calculations, the corrosion rate is 40 mpy to 50 mpy. Taking into account the percentage of anion components found in field-sampled amine, the corrosion rate was applied to these test results, as shown in Figs. 7 and 8. The FeSX scale formation on carbon steel was found to have a smaller structure and to adhere better to metal in proportion to the increase in H2S concentration. In lower H2S concentrations, the FeSX scale has a larger structure and is less adhesive to metal. Organic acid anions overcome the sulfur element of FeSX and, in the process of destroying the scale, they form salts (e.g., iron acetate).
An X-ray diffraction analysis was performed on an FeSX scale on the vapor space of Tray 9, in an area that was not affected by corrosion. Test results (Fig. 4) show the scale type to be Pyrrhotite, which adheres suitably to metal and offers better surface protection than a Mackinawite scale.
The process simulation results presented in Table 9 show the decreasing presence of H2S in the carbon steel regions of Trays 8 through 20. Consequently, these trays have less adhesive FeSX scales.
The results of the sour gas-sweetening unit analysis show that the percentage of H2S is suitably low and meets conditions outlined in NACE MR-01-75, even though the gas is sour. On the other hand, the low concentration of H2S results in a loose and brittle FeSX scale, along with organic acid anions in the vapor phase, steam that allows for an electrochemical condition, and the inability to inject an inhibitor into the liquid phase. All of these factors create a tower environment that is susceptible to corrosion.
Furthermore, internal temperatures approach 120°C on the lower trays of the tower and 100°C on the upper trays. The higher temperature and looseness of the FeSX scale on the lower trays accelerate corrosion on the lower section of the tower.
Corrosion formed in a valley shape on the liquid/vapor interface due to FeS2 erosion from the use of an amine containing organic and inorganic contaminations. Low reforming potential and a high concentration of acidic components next to the liquid surface also contributed to the corrosion.
According to the analysis, the water used to prepare the amine solution contained an unacceptable concentration of oxygen. Also, existing O2 in the reaction with DEA resulted in DEA oxidation in the circulating amine solution, producing organic acids that entered the vapor phase of the tower. In a corrosion-susceptible environment (i.e., the existence of water, high temperature, acids, and no amine in the vapor phase), the presence of these acids contributed to corrosion on the shell interior of the tower.
The selection of a corrosion inhibitor was not correctly performed, as attention was not given to the required temperature for rationing with O2. Furthermore, the design of the corrosion inhibitor injection system was corrupt, and the injection of an inhibitor into two points is not practical. The lesson learned is that the selection of a corrosion inhibitor should be carefully performed. A revamp of the injection processone that considers simultaneous injection into two points at different pressuresis necessary. Also, the type of FeSX scale that forms on carbon steel must be considered, as this has a direct effect on the degree of surface protection achieved and the rate and extent of the corrosion that occurs.
An anionic resin bed played an unexpected and important role in corrosion prevention during the first year. In a situation where O2 cannot be eliminated, the development and use of this type of bed is recommended. Also, the implementation of a suitable lining on the tower interior can provide temporary corrosion control. HP
Research for this survey was supported by Sarkhoon & Qeshm Gas Treating Co.
1 Robert, N. and J. Morgan, Gas Conditioning and Processing, Vol. 4, Campbell Petroleum Series, Oklahoma, 1998.
2 Jenkins, H., Understanding gas treating fundamentals, Petroleum Technology Quarterly, Winter 2001.
3 Dupart, M. S., T. R. Bacon and P. C. Rooney, Effect of heat-stable salts on MDEA solution corrosivity, Hydrocarbon Processing, March 1996.
4 Rooney, C. P. and M. S. Dupart, Corrosion in Alkanolamine Plants: Causes and Minimization. NACE International, Corrosion 2000, Orlando, Fla., 2000.
5 Rooney, C. P. and D. Daniels, Oxygen solubility in various alkanolamine/water mixtures, Petroleum Technology Quarterly, Spring 1998.
6 Stewart, E. J. and R. A. Leanning, Reduce amine plant solvent losses, Hydrocarbon Processing, May 1994.
7 Tanthapanichakoon, W. and A. Veawab, Corrosion by heat-stable salts in amine-based CO2 capture unit, Second Annual Conference on Carbon Sequestration, Alexandria, Vir., 2003.
8 Dupart, M. S., T. R. Bacon and D. J. Edwards, Understanding corrosion in alkanolamine gas treating plants: Part 1, Hydrocarbon Processing, April 1993.
9 Standard Material RequirementsSulfide Stress Cracking Resistant Metallic Materials for Oilfield Equipment, NACE International, 2002.
10 Valenzuela, D. P. and A. K. Dewan, Refinery crude column overhead corrosion control, amine neutralizer electrolyte thermodynamics, thermochemical properties and phase equilibria, Fluid Phase Equilibria, Vols. 158160, June 1999.
11 Ren, C., D. Liu, Z. Bai and T. Li, Corrosion behavior of oil tube steel in stimulant solution with hydrogen sulfide carbon dioxide, Materials Chemistry and Physics, Vol. 93, 2005.
12 Fan, D., L. E. Kolp, D. S. Huett and M. A. Sargaret, Role of impurities and H2S in refinery lean DEA system corrosion, NACE International, Corrosion 2000, Orlando, Fla., 2000.
13 ASTM International, Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens, ASTM G1-03, 1990 (reapproved 2003).
14 ASTM International, Standard Test Methods for Dissolved Oxygen in Water, ASTM D888-09, 1992 (reapproved 2009).
15 ASTM International, Standard Test Methods for Low-Level Dissolved Oxygen in Water, ASTM D5543-09, 1994 (reapproved 2009).
16 ASTM International, Standard Test Methods for Analysis of Natural Gas by Gas Chromatography, ASTM D1945-03, 1996 (reapproved 2010).
17 ASTM International, Standard Test Methods for Conventions Applicable to Electrochemical Measurements in Corrosion Testing, ASTM G3-89, 1989 (reapproved 2010).
18 ASTM International, Standard Practice for Laboratory Immersion Corrosion Testing of Metals, ASTM G31-72, 1972 (reapproved 2004).
|The authors |
Ahmad Zamani Gharaghoosh is the managing director of Sarkhoon & Qeshm Gas Treating Co. He joined National Iranian Gas Co. in 1997 and has more than 15 years of experience in gas refineries as a senior corrosion and welding engineer, inspection department head and research committee member. He is an expert on static equipment inspection, risk-based inspection and corrosion investigation at gas refineries, and he has authored or co-authored five papers for ISI journals and national and international conferences.
Abolfazl Atash-Jameh heads the process engineering group at Sarkhoon & Qeshm Gas Treating Co. He joined National Iranian Gas Co. in 1999, and he has 10 years of experience in different aspects of process engineering for natural gas. He has published seven papers for national and international conferences.
Amir Reza Rashidfarokhi is a mechanical engineer and heads the technical inspection group at Sarkhoon & Qeshm Gas Treating Co. He joined National Iranian Gas Co. in 2002, and he has eight years of experience in different aspects of the technical inspection industry. He has also participated in a number of industry studies and courses.
Mahmoud Pakshir is a tenured professor at the Department of Material Science and Engineering at Shiraz University in Iran. He has published more than 70 papers on corrosion control and prevention.
M. H. Paydar is an associate professor at the Department of Material Science and Engineering at Shiraz University in Iran. He has published more than 20 papers in international magazines.