Converting to DEA/MDEA mix ups sweetening capacity

Michael L. Spears
Union Pacific Resources
Bryan, Tex.

Kathy M. HaganUnion Pacific Resources
Ft. Worth

Jerry A. Bullin, Carl J. Michalik
Bryan Research & Engineering
Bryan, Tex.

Mixing amines can be the best method for increasing capacity or improving efficiency in an amine sweetening unit. In many cases, it may be possible simply to add a second amine to the existing solution "on the fly," or as the unit is running.

Union Pacific Resources' Bryan, Tex., gas plant provides one example. The plant was converted from diethanolamine (DEA) to a DEA/MDEA (methyl DEA) mixture after analysis by TSWEET, a process-simulation program.

After conversion, CO₂ levels in the sales gas fell to less than pipeline specifications. Data were taken for the absorber at a constant amine circulation of 120 gpm. A comparison of the performance data to the values calculated by the program proved the accuracy of TSWEET.

Expanding

In many gas-processing plants, the amine unit may limit the entire plant throughput because of, for example, increased acid-gas concentration in the feed or increased feed volume.

Increasing the unit's efficiency by converting from a single-amine solution to a mixture of amines most commonly involves use of a DEA/MDEA mixture.

The optimum solution concentration and ratio of amines depend on absorber pressure, acid-gas concentration in the feed, contactor operating temperature, and sweet-gas requirements (how much CO₂ slip is allowable).

A process simulator such as TSWEET can simplify the evaluation of adding a second amine.

Each amine has unique advantages and disadvantages.¹ For example, a primary amine (MEA) is more reactive with CO₂ than a secondary amine (DEA), which is in turn more reactive than a tertiary amine (MDEA or TEA).

In most cases, however, the maximum concentration and acid-gas loading of primary and secondary amine solutions are usually limited by corrosion problems. The most widely accepted solution concentrations are <=20 wt % for MEA and <=35 wt % for DEA.

A high concentration of MDEA in solution does not result in corrosion problems as do MEA and DEA. MDEA reacts slowly with CO₂ and is most commonly used where CO₂ slippage is desired.

Such operating parameters as longer liquid residence times on the absorber trays and higher absorber temperatures may be manipulated to accelerate CO₂ pickup by MDEA.²

MDEA has several advantages over primary and secondary amines including higher solution concentration, higher loading with fewer corrosion problems, lower heat of reaction leading to better stripping, higher resistance to degradation, and lower vapor pressure.³

Furthermore, MDEA is selective toward H_xS in the presence of CO₂. This selectivity results from the inability of tertiary amines to form a carbamate with CO₂.

MDEA does not have a hydrogen attached to the nitrogen and cannot react directly with CO₂ to form carbamate. MDEA can, however, react with H_xS by the same proton-transfer mechanism of primary and secondary amines.⁴

Using mixed amines

The concept of using amine mixtures is based on utilizing and combining the advantages of the amines in the mixture or of customizing the amine solution to a particular use. Mixing MDEA with DEA, for example, can increase the total amine solution concentration without increasing corrosion problems.

The easiest and least expensive method of switching from a single amine to a mixed amine solution is simply to add the second amine "on the fly," while the unit is operating.

If, however, the amine unit has become dirty or the amine has degraded, switching amines presents an opportunity to shut down the unit, drain the system, and clean or replace equipment.

In this case, the unit should be reloaded with the target amine solution concentrations. If the concentrations do not meet the target, the appropriate amine may be added as necessary.

Bryan plant

A recent example of a conversion to mixed amines is Union Pacific Resources' DEA-to-MDEA/DEA mixture at its Bryan, Tex., gas plant.

As shown in Fig. 1 [27456 bytes], the feed-gas stream passes through a compression station to a standard refrigeration plant where liquids are extracted. The gas is sent to an amine unit for sweetening to a product specification of 0.35 mole % CO₂. Liquids must pass the copper 1A strip test for H_xS.

The Bryan gas plant (Fig. 2 [26381 bytes]) was designed to process 35 MMscfd of gas containing 2.91% CO₂ and 0.06 gr/100 standard cu ft H_xS with 35 wt % DEA. The absorber is a 20-stage valve tray column with an amine circulation of 130 gpm.

Since the plant was built in 1980, the CO₂ content in the inlet-gas stream continually increased until the amine unit could not handle the increased CO₂ concentration and the CO₂ gas product specification was not met. The amine unit became the plant bottleneck.

The rich amine loading has increased to 0.56 mole acid gas/1.0 mole amine which could lead to corrosion problems with DEA, and the CO₂ concentration in the sweet gas is as high as 0.42%.

To increase the capacity of the amine unit, Union Pacific Resources considered designing a parallel unit, increasing the circulation rate and adding a larger reboiler, and switching to a mixture of MDEA/DEA.

Adding a parallel unit would be very expensive.

Increasing the circulation rate of the DEA solution would require replacing the stripper reboiler and pumps throughout the plant. In addition to costs, this alternative would also require extended downtime.

The third alternative, adding MDEA to the existing DEA solution, would increase the CO₂ absorption without a significant increase in corrosion. Therefore, the addition of MDEA to the DEA solution warranted further investigation.

Although MDEA is known for selectively absorbing H_xS and rejecting CO₂, an increase in CO₂ absorption was desired in the case of the Bryan gas plant. Increasing the concentration of DEA alone would absorb additional CO₂ but would increase corrosion.

Addition of MDEA to raise the total amine solution concentration would increase CO₂ absorption but would not significantly affect corrosion levels.

The feasibility study was performed by the process simulation program TSWEET which models amine sweetening units. The program was developed by Bryan Research & Engineering Inc., Bryan, Tex., and uses the flexible flowsheet and operating concept.

Beginning with the process flowsheet and operating parameters, complete material and energy balances for steady-state operating conditions can be performed.

Rigorous tray-by-tray calculations using the Ishii-Otto method⁵ design the stripper and absorber columns. Vapor pressures of H_xS and CO₂ over the amine solutions are calculated by the non random, two-liquid (NRTL) equilibrium model.

A kinetic model is used to predict the effect of residence time, temperature, solution concentration, feed pressure, and type of amine on the rate of CO₂ absorption.

The program has been reliable in calculating the performance of amine gas-sweetening plants.^{6 7}

Based on the TSWEET program results, Union Pacific Resources concluded that with the addition of MDEA to the DEA solution, the existing equipment at the Bryan gas plant had sufficient capacity to absorb enough CO₂ to meet the sweet-gas specification and that no equipment modifications were necessary.

UPR decided to convert to a DEA/MDEA mixture.

Converting, testing

There are two options for converting from DEA to a DEA/MDEA mixture. The conversion can be performed by adding the MDEA during operation without a plant shutdown, or the unit can be shut down, cleaned, and reloaded.

While shutting down the amine unit is not required, it may be a good idea if cleaning the system or replacing equipment is necessary. During amine conversion at the Bryan gas plant, Union Pacific Resources first chose to drain the system in order to replace a fouled heat exchanger.

The amine unit was taken off-line, cooled down, and drained to its low points. Next, the lean/rich exchanger was replaced and the unit was reloaded with a premixed amine solution.

The initial amine solution was 65 wt % water, 25 wt % DEA, and 10 wt % MDEA. The system was then put on-line and tested. Additional amine was gradually added until a final concentration of 50 wt % water, 35 wt % DEA, and 15 wt % MDEA was achieved.

The absorber performance at the plant was studied with a constant amine circulation rate of 120 gpm. Data were taken daily for the gas flow, composition, temperature, and pressure of the feed and exit gas, lean amine temperature, lean-amine loading, and sweet-gas composition.

Recorded stripper data included feed temperature, overhead temperature and pressure, pressure drop, and reboiler and condenser temperatures. All flow rates were measured by orifice meters and all temperatures were measured by calibrated thermocouples.

The acid-gas compositions were determined by titrimetric analysis. The remaining gas compositions were determined by gas chromatography.

Plant performance

After switching to a DEA/MDEA mixture, the plant operated without problems. The concentration of CO₂ in the sweet gas was less than 1,000 ppm and as low as 200 ppm. Table 1 [24207 bytes], shows the inlet conditions to the absorber at each test date.

Data for this study were taken during two periods: DEA-solution data were taken Dec. 5-10, 1994; DEA/MDEA mixture data were taken Feb. 2-26, 1995. Data for missing days within these periods were omitted because either no data were taken, no amine concentration was recorded, the reboiler temperature was inconsistent, or the hot oil valve was 100% open.

Table 2 [45234 bytes] shows the absorber outlet conditions along with the corresponding calculated values from TSWEET.

As can be seen in Table 2 [45234 bytes] and Fig. 3a [65597 bytes], the CO₂ concentration with the DEA solution varied 0.32-0.42%. When converted to a DEA/MDEA mixture, the CO₂ concentration in the sweet gas was often less than 0.1% and never higher than 0.2% (Table 2 [45234 bytes]; Fig. 3b [65597 bytes]).

As can be seen, the DEA/MDEA mixture easily produced H_xS and CO₂ concentrations less than the product specifications. The CO₂ concentration calculated by TSWEET matched the actual data very closely.

Table 2 [45234 bytes] and Fig. 4 [23914 bytes] also show the rich-amine loading (mole of acid gas/mole of amine) for the absorber with the DEA/MDEA amine mixture over time.

Acknowledgment

The authors acknowledge the contributions of Steve Behrens, CBS Engineering, Houston, who participated in data gathering at the conceptual stages of this project and provided technical assistance at several junctions.

References

1. Bullin, J.A., and Polasek, J.C., "Selective Absorption Using Amines," Gas Processors Convention, 1982.

2. MacKenzie, D.H., and Prambil, F.C., Daniels, C.A., and Bullin, J.A., "Design & Operation of a Selective Sweetening Plant Using MDEA." Energy Progress (March 1987), Vol. 7, No. 1, pp. 31-36

3. Blanc, C., Grail, M., and Demarais, G., "The Part Played by Degradation Products in the Corrosion of Gas Sweetening Plants Using DEA and MDEA," Gas Conditioning Conference, University of Oklahoma., 1982.

4. Jou, F.Y., Mather, A.E., and Otto, F.D., "Solubility of H_xS and CO₂ in Aqueous Methyldi ethanol amine Solution." Ind. Eng. Chem. Proc. Des. Dev., Vol. 21, p. 539.

5. Ishii, Y., and Otto, F.D., "A General Algorithm for Multistage Multicomponent Separation Calculations," Canadian Journal of Chemical Engineering, Vol. 51 (1973), p. 601.

6. Holmes, J.C., Spears, M.L., and Bullin, J.A., "Sweetening LPG's with Amines," AIChE Summer Annual Meeting, Denver, 1983.

7. Polasek J. C., Iglesias-Silva, G. A., and Bullin, J. A., "Using Mixed Amines Solutions for Gas Sweetening," Annual Gas Processors Convention, 1992.

Based on a presentation to the 75th Annual GPA Convention, Mar. 11-13, Denver.

The Authors