PETROCHEMICALS & GAS PROCESSING

Predicting amine blend performance

A process simulator is able to predict the performance of an Iranian gas-sweetening plant. Various mixtures of diethanolamine (DEA) and Methyl diethanolamine (MDEA) are used to investigate the potential for an increase in plant capacity
Hamid Reza Khakdaman and Ali Taghi Zoghi National Iranian Oil Company Majid Abedinzadegan Abdi Memorial University of Newfoundland

echnology using alkanolamines, or amines, for the removal of hydrogen sulphide and carbon dioxide from natural gases has been around for decades. Since the 1960s and 1970s, several amines have come into general use, but there is little information available on which amine is best suited to a particular service. Many inefficient amine gassweetening units can be optimised by simply changing their amine solutions. The basic flow scheme for an aminesweetening unit is shown in Figure 1. In the design of the process, the primary concern is that the sweetened gas should meet the required purity specifications with respect to H2S and CO2. The secondary objective is to select the amine, which optimises equipment size and minimises plant operating costs. The following points should also be addressed in the selection of the proper amine for the design or evaluation of an existing plant: Can the amine circulation rate be reduced by selecting an amine that may be used at a higher concentration and/or at a higher acid gas loading? Could the equipment be designed more efficiently using an amine that requires a lower circulation rate and/or has lower heats of reaction with H2S and CO2? Could H2S be selectively absorbed from the sour gas while CO2 is rejected? Can the selective absorption of H2S and CO2 from the sour gas be optimised by the use of a suitable amine blend? Could corrosion and solvent loss problems be improved with an amine or mixture of amines more resistant to degradation? Between 5070% of the initial investment for an amine-sweetening unit is directly associated with the magnitude of the solvent circulation rate, and another 1020% of the initial investment depends on the regeneration energy requirement (Astarita et al, 1983). Approximately 70% of gas-

Sweet gas Lean amine cooler

Stripper O/H condenser Storage tank

Acid gas

Booster pump Sour gas

Reflux pump

Stripper

Absorber
Stm. Cond.

Reboiler Cross exchanger

Figure 1 Process flow diagram for a common sweetening plant Typical operating conditions and data for amines
Alkanolamine type MEA DEA DGA MDEA Solution strength, wt% 1520 2535 5070 2050 Acid gas loading, mole/mole 0.300.35 0.300.35 0.300.35 Unlimited Ability for selective absorption of H2S No Under limited No Under most conditions conditions

Table 1
sweetening plant operating costs, excluding labour expenses, are due to the energy required for solvent regeneration. Appropriate amine selection can significantly reduce the regeneration energy requirement and solution circulation rate. Therefore, the selection of amines best suited to the process conditions can have a dramatic impact on the overall costs associated with a sweetening unit. Several alternative flow schemes for amine-sweetening plants were discussed in detail by Polasek, Donnelly and Bullin (1983) and therefore will not be discussed here.

General considerations for amine selection

The general criteria for amine selection in sweetening plants have changed over the years. Until the 1970s, monoethanolamine (MEA) was the only amine considered for any sweetening application. In the 1970s, as exemplified in papers by Beck (1975) and Butwell and Perry (1975), a major switch from MEA to diethanolamine (DEA) occurred. In the past ten years, MDEA, DGA and mixed amines have steadily gained popularity too. In order to become accepted on an industry-wide basis, different operating

00

Heat of reaction for different types of amines

Amine Solution strength Acid gas loading, mole/mole Hr for H2S, Btu/lb Hr for CO2, Btu/lb MEA 1520 0.30.4 550 825 DEA 2535 0.30.4 511 653 DGA 4060 0.30.4 674 850 MDEA 3050 Unlimited 522 600

and Danckwerts (1979). MDEA can, however, react with H2S by the same proton transfer mechanism of primary and secondary amines (Jou et al, 1982): (2) Selective absorption of H2S can be enhanced by optimising absorber design to obtain a liquid tray residence time of 1.53.0 seconds and by increasing the temperature in the absorber. Both of these conditions favour H2S absorption with CO2 rejection.

Table 2
conditions should be tested and proven with a particular amine. Each alkanolamine solution has an accepted range of process conditions and parameters associated with it. Typical operating conditions for common alkanolamines are summarised in Table 1. concentration are typically employed in low-pressure, high-selectivity applications such as the selective removal of H2S in the Shell Claus off-gas treating (SCOT) units. Due to the considerably reduced corrosion problems, acid gas loadings as high as 0.70.8 mole/mole are considered practical in carbon steelmade equipment. Higher loadings may also be possible with a few problems. MDEA exposure to oxygen forms corrosive acids, which, if not removed from the system, can result in the buildup of iron sulphide in the system. MDEA has several distinct advantages over primary and secondary amines, which include lower vapour pressure, lower heats of reaction, higher resistance to degradation, fewer corrosion problems and selectivity toward H2S in the presence of CO2. Most of these advantages have also been reported by Blanc et al (1982). Depending on the application, some of them have special significance; for example, due to its lower heat of reaction, MDEA can be employed in pressure swing plants for bulk CO2 removal. Here, the rich amine is merely flashed at or near atmospheric pressure and little or no heat is added for stripping. The overwhelming advantage that MDEA currently possesses over the other amines is that it is readily selective toward H2S in the presence of CO2. At high CO2/H2S ratios, a major portion of the CO2 can be slipped through the absorber and into the sales gas while removing most of the H2S. The enhanced selectivity of MDEA for H2S is attributed to the inability of tertiary amines to form carbamates with CO2. MDEA does not have a hydrogen attached to the nitrogen and cannot react directly with CO2 to form carbamate. The CO2 reaction can only occur after the CO2 dissolves in water to form a bicarbonate ion, which then undergoes an acid-base reaction with the amine: (1) At least six mechanisms for the CO2MDEA reaction have been proposed by Cornelissen (1982), Barth et al (1981)

Mixed amines
Amine blends are generally mixtures of MDEA and DEA or MEA and are used to enhance CO2 removal by MDEA, as described by Polasek, Bullin and IglesiasSilva (1992). Such mixtures are referred to as MDEA-based amines, with DEA or MEA as the secondary amines. The secondary amines generally comprise less than 20% of the total amine content on a molar basis. At lower concentrations of MEA and DEA, the overall amine concentration can be as high as 55 wt% without the implementation of exotic metal equipment. MDEA-based mixtures are normally used to increase the CO2 pickup in cases where the MDEA is allowing too much CO2 to slip overhead in the absorber. Spiking the MDEA with MEA or DEA to achieve the desired CO2 pickup is often preferable to a complete amine switchout to a DEA or MEA system, because the MDEA regenerator reboiler may be undersized for a purely formulated DEA or MEA system. Operating problems associated with mixed amines influence amine mixture concentration and its maintenance. However, finding an optimum concentration for mixed amines (DEA+MDEA) strongly depends on the H2S and CO2 content of the sour gas, operating pressures and sale gas specifications. For natural gassweetening purposes, mixed amines are typically mixtures of MDEA and DEA or MEA, which enhance CO2 removal while retaining desirable characteristics of MDEA, such as reduced corrosion problems and low heats of reaction

Diethanolamine (DEA)
DEA is the most commonly used amine, within the 2535 wt% range. The total acid gas loading for DEA is limited to 0.300.35 mole/mole for carbon steel as the equipments construction material. DEA can be safely loaded up to equilibrium level (~1 mole/mole) when stainless steel is used. The degradation products of DEA are relatively less corrosive than those of MEA. Exposure to oxygen forms corrosive acids, and COS and CS2 may, to some extent, react irreversibly with DEA. DEA is not reclaimable under regenerator conditions; it decomposes below its boiling point at atmospheric pressure. Vacuum reclaimers, however, have been successfully used to reclaim DEA solutions (Meisen et al, 1996). Since DEA is a secondary alkanolamine, it has a reduced affinity for reaction with H2S and CO2. Therefore, it may not be able to produce pipeline-quality gas for some lowpressure gas streams. In general, as the gas pressure is lowered, the stripping steam must be increased or a split flow design used. In some cases, even these measures will not suffice and another solvent must be used. Under some conditions, such as low pressures and a liquid residence time on the tray of about two seconds, DEA is selective toward H2S and will permit a significant fraction of the CO2 to remain in the sales gas.

Methyldiethanolamine (MDEA)
An accepted set of operating conditions has not been firmly established for MDEA, as compared to the previously mentioned amines. This is due to the flexibility and versatility of MDEA, and the resulting wide range of applications. MDEA is a tertiary amine and commonly used in the 2050 wt% range. Solutions with lower amine

Case study
A typical Iranian gas plant is selected for this study. The gas-sweetening facility has five identical amine trains for H2S and CO2 removal. The plant management decided to consider one of the units for substituting DEA with a mixture of DEA and MDEA. Each train was composed of two absorbers and two stripper columns, which operated in parallel in the unit. The HYSYS plant simulator was used to simulate the

00

PETROCHEMICALS & GAS PROCESSING

Absorbers feed gas composition (design basis)
Flow rate, kmole/hr H2O 4.28 N2 75.94 CO2 936.3 H2S 562.36 COS 0.26 C1 12 909.22 C2 81.8 C3 13.16 i-C4 2.92 n-C4 4.38 i-C5 2.92 n-C5 2.92 C6+ 14.62 DEA Total 14 611.16 Pressure, bara 1063 Temperature, C 21 Component Mole % 0.03 0.52 6.41 3.85 17ppm 88.35 0.58 0.09 0.02 0.03 0.02 0.1 0 100

Gas-sweetening operating conditions

Parameter Amine circulation rate, m3/hr Absorber col. top/bottom pressure, bara Absorber col. top/bottom temperature, C Stripper col. top/bottom pressure, bara Stripper col. top/bottom temperature, C No. of actual tray (absorber) No. of actual tray (stripper) Typical value 935 1058/1063 55.0/77.0 22.0/27.9 52.0/120.4 20 24

Table 4 Comparison between simulation and actual operating data

Parameter Rich amine loading Lean amine loading H2S in sweet gas, ppm CO2 in sweet gas, mol% Absorber col. top/bottom temperature, C Stripper col. top/bottom temperature, C Operating data 0.450.50 0.0270.031 1.52.5 0.01 55.0/77.0 52.0/120.4 Simulation results 0.49 0.028 2.0 0.01 61.5/86.2
1

Table 3
process. The absorber feed gas composition is shown in Table 3, and operating conditions are summarised in Table 4. Current plant operating conditions were initially simulated to obtain the confidence that the simulation was performed effectively. The simulation produced a very good agreement between the HYSYS-generated results and the actual operating data. The results are listed in Table 5. The process was subsequently simulated using various mixtures of DEA and MDEA with the following constraints: Solution circulation rate was 3 considered constant at 935m /hr H2S content in sweet gas should be kept less than 2ppm CO2 content in sweet gas should be kept less than 1% Duty of each reboiler was considered constant at 1.32e+8kJ/hr (125MMBTU/hr) Condenser temperature equals 52C. DEA and MDEA concentrations in the solution were changed from 1030 and from 539 wt% respectively. The amine mixtures, which met a targeted value for the following parameters, were selected as the alternative solvent for optimum mixture concentration: Amine system

1 The condenser temperature was set to 52C and the reboiler duty of 125MMBTU/hr caused the reboiler temperature of 120.4C.

Table 5

250 MMBtu/hr
20 500 Sour gas flowrate, kmole/ hr 20 000 19 500 19 000 18 500 18 000 17 500 17 000 5 10 15 20 DEA, % 25 30 35 Total amine 35% Total amine 40% Total amine 45% Total amine 49%

Figure 2 Unit revamp at fixed reboiler duty

changed. Since the maximum MDEA concentration in an industrial application is limited to below 50%, the total composition was kept below this percentage. It can be concluded that a 49% amine blend with 2030% DEA content will be an optimised composition. A lower-end (closer to 20%) concentration for DEA will be recommended due to the need for solution corrosion and viscosity control. As can be seen, by blending DEA and MDEA mixtures, for the indicated composition, the plant capacity can be increased to 17 00020 000 kmole an rise of approximately 1637%. In order to check if the plant can handle higher gas flow rates, other pieces of equipment and plant parts,

Acid gas composition in the sweetened gas (absorber overhead) Figure 2 shows how the plant capacity can be increased for various amine blends. The throughput can be raised from the base value of approximately 14 600kmole/hr to the indicated gas flow rate shown in Figure 4 for various amine blend compositions, mol CO 2 + mol H 2S o Rich amine loading: ( ), H2S(v/v)= and will be further mol MDEA + mol DEA discussed. It should be noted that the H2S Vol .Flow CO 2 Vol .Flow , CO2 (v/v)= reboiler duty and A min e Vol .Flow A min e Vol .Flow other parameters previously indicated o Lean amine loading: ( mol CO 2 + mol H 2S ) , H2S(v/v), CO2 (v/v) were fixed and that mol MDEA + mol DEA only the gas throughputs were

00

PETROCHEMICALS & GAS PROCESSING

MDEA capabilities
0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 5 10 15 20 DEA, % 25 30 35 Total amine 35% Total amine 40% Total amine 45% Total amine 49% CO2, %

Due to its lower corrosion tendency and heat of reactions with acid gases compared to other amines, MDEA is a favourable option. Using the HYSYS plant simulation, different mixtures of DEA and MDEA were investigated. Since mixed amines have a higher capacity for acid gas removal at constant amine circulation rates compared to DEA solvents, the gas-processing capacity of gas-sweetening units can be increased. These results show that the gas flow rate capacity for a typical unit can easily be increased by up to 20%.
References 1 Beck J E, Diethanolamine an energy conserver, Laurence Reid Gas Conditioning Conference, Norman, Oklahoma, 1975. 2 Donnelly S T, Henderson D R, A new approach to amine selection, Proceedings of the AIChE Spring National Meeting New York, 1982. 3 Holmes J W, Spears M L, Bullin J A, Sweetening LPGs with amines, Chemical Engineering Progress, 80, 5, 47, 1984. 4 Jou F Y, Lal D, Mather A E, Otto F P, Solubility of H2S and CO2 in aqueous methyldiethanolamine solutions, Ind. Eng. Chem. Proc. Des. Dev., 21, 539, 1982. 5 Kohl A and Riesenfeld F, Gas purification, Gulf Publishing Company, Houston, Texas, 1979. 6 Meisen A, Abedinzadegan Abdi M, Abry R, Millard M, Degraded amine solutions: nature, problems, distillative reclamation, Proceedings of the 45th Annual Laurence Reid Gas Conditioning Conference, Norman, Oklahoma, 168190, 1996. 7 Polasek J C, Bullin J A, Selective absorption using amines, Proceedings of the 61st Annual Gas Processors Convention, Tulsa, USA, 1992. 8 Polasek J C, Bullin J A, Iglesias-Silva G A, Using mixed amine solutions for gas sweetening, presented at the 71st Annual Gas Processors Association 2528 February, Norman, Oklahoma, 2001.

Figure 3 CO2 concentration in sweet gas for different amine mixtures (DEA+MDEA)

14.0 12.0 10.0 H2S, ppm 8.0 6.0 4.0 2.0 0.0 5 10 15 20 DEA, % 25

Total amine 35% Total amine 40% Total amine 45% Total amine 49%

30

35

Figure 4 H2S concentration in sweet gas for different amine mixtures (DEA+MDEA)
including heat exchangers, pumps, pipe sizes and towers, should be examined. It was observed that the diameter of the existing absorption tower columns could handle an increase in gas throughput of around 22%. Using this figure, the performance of the plant was then evaluated using various amine blends. The lean loading increases with DEA content, as a higher heating rate is required to release the acid gas from the amine. The rich amine loadings remain relatively constant for a particular amine blend, but since a solvent blend with a higher concentration of total amine can naturally absorb more acid gas a reduction of acid gas loading will be expected when the total amine content increases. Figure 3 and the previously mentioned Figure 4 show a variation in the acid gas (CO2 and H2S) content of the sweetened gas for the enhanced capacity scenario a 22% gas flow rate increase when the amine flow rate and reboiler duties were kept constant. It is evident that unless the total amine concentration was increased beyond the 35 wt% mark, the acid gas specifications could not be met. Beyond this total amine composition, the acid gas content of the sweetened gas remains nearly constant for varying DEA content.

Since mixed amines have a higher capacity for acid gas removal at constant amine circulation rates compared to DEA solvents, the gas-processing capacity of gas-sweetening units can be increased

Hamid Reza Khakdaman is head of the modelling and simulation department at the das division of the Research Institute of Petroleum Industry (RIPI) of the National Iranian Oil Company (NIOC), Tehran, Iran. He is involved in natural gas sweetening and gas-to-liquids processes as a research engineer. He holds a BSc from the Iranian University of Science and Technology (IUST) and MSc from Tehran University, both in chemical engineering. Email: khakdamanhr@ripi.ir Ali Taghi Zoghi is a senior project engineer at NIOC-RIPI, Tehran, Iran where he is involved in gas-processing projects. He holds a BSc from Tehran University and MSc from IUST, both in chemical engineering. Email: zoghiat@ripi.ir Majid Abedinzadegan Abdi is the former director of the gas division at NIOC-RIPI, Tehran, Iran and is currently a faculty member at the Memorial University of Newfoundland and process engineering lead at the centre for marine CNG in St Johns, NL, Canada. Email: abdi@engr.mun.ca

00