Using current technology to capture the carbon in flue gas at fossil fuel power plants would increase the cost of electricity they produce by between 35 per cent and 85 per cent, but a simulation shows that using supported amine sorbents (SASs) instead could be a less expensive solution.
SASs would reduce these facilities’ capital expenditure on a capture facility because the adsorber which contains these materials would be about 60 per cent smaller in volume than the absorber that would be required if the monoethanolamine (MEA) solutions of today’s technology were used to perform the same task. This is because the SAS system avoids the slow mass transfer associated with dissolution and diffusion in the liquid in the MEA system. And these cost savings come on top of energy savings.
Today capture and compression costs contribute to around 80 per cent of the total cost of carbon capture and storage (CCS). But excluding compression, thermal energy consumption and other operational costs dominate capture costs, as shown in Figure 1, and the capital expenditure, 28 per cent, is mainly made up of the cost of the absorption unit, covering almost 50 per cent of the total CAPEX.
Figure 2 breaks down the energy requirement fraction (44 per cent of the total cost of CO2 capture) into contributions for reaction or desorption heat, sensible heat and solvent evaporation.
A large part of the energy requirement of an MEA process arises in the heating of the amine solution from the absorption temperature to the desorption temperature, and in the evaporation of solvent in the desorber column, so replacing the water phase with an SAS greatly reduces the energy required for regeneration. An adsorption-based capture process could reduce the net energy requirement for CO2 capture from roughly 3.3 GJ/tonneCO2 (GJ/tCO2) to below 2 GJ/tCO2, assuming that 75 per cent of the heat contained by the regenerated sorbent is recovered for heating the sorbent before it enters the desorber.
If co-adsorption of water can be prevented and no water has to be evaporated in the desorber, the process energy requirement could even be 1.7 GJ/tCO2.
Another advantage of using SAS would be the reduction of the emissions of toxic degradation compounds such as nitrosamines and nitramines.
Our simulation employed DNV KEMA Energy & Sustainability’s Spence® flowsheeting tool, which determines power plant performance accurately. Data about the sorbent has come from SAS that the University of Twente in the Netherlands has produced through the physical impregnation of polymethylmethacrylate, (PMMA) in this case Diaion™ HP-2MG, with tetraethylenepentamine (TEPA) and polyethyleneIimine (PEI) of different molecular weights and chain lengths.
Our analysis calculated the efficiency of a typical natural gas-fired combined-cycle (NGCC) plant of 446 MWe and a ultra supercritical pulverised coal (PC) plant of 1069 MWe, each equipped with either an MEA-based capture facility or an SAS-based one, providing four cases. The analysis compared the amount of energy consumed per kg of CO2 captured in each case.
SAS consist of a support material of high surface area with amine functional groups on their surface. The university’s sorbent material had a CO2 capacity of 3.8 mol/kg sorbent .
Figure 3 shows adsorption isobars for a typical developed sorbent. The sample was heated from 40ºC to 140ºC at a rate of 0.1 K/min in fractions by volume of 1 per cent, 5 per cent, 10 per cent and 80 per cent CO2 atmosphere (balance N2) at a total flow rate of 100 ml/min. From the selected adsorption and regeneration conditions the sorbent operating capacities were calculated for CO2 capture at NGCC and PC plants.
|Figure 3: Adsorption isobars for 1 per cent, 5 per cent, 10 per cent and 80 per cent of CO2 by volume (Ptotal = 100 kPa). The sorbent was 38 per cent by weight TEPA on PMMA|
Flue gas from an NGCC plant typically contains 4-7 per cent of CO2 by volume, whereas from a PC plant it contains 10 to 15 per cent.
The sorbent working capacity is the difference between the CO2 loading under adsorption conditions and the loading under desorption conditions. During capture, when the flue gas contains 10 per cent of CO2 by volume and desorbing is at 130ºC, working capacities of around 3.1 mol/kg can be achieved with the developed sorbent. For capture from flue gas containing 5 per cent CO2, working capacities of about 2.6 mol/kg can be achieved under these conditions.
The Spence flowsheeting tool calculated how much thermal energy each case of power plant would require for regeneration of the solvent or sorbent and what the resulting drop in power from the plant would be. This thermal energy would come from low-pressure steam at 4.6 bara (460 kPa) for the NGCC plant and 3.5 bara for the PC fired plant.
Table 1 shows general information about the carbon capture facility. The boundary of interest of the system excludes CO2 transport and storage systems but includes compression.
The modelled MEA-based capture facility was a standard regenerative absorption-desorption system, with a net thermal energy input of 3 GJ/CO2. Values for the heat requirement of leading absorption technologies are between 2.7-3.3 GJ/tCO2. Operational assumptions for the absorber and desorber column were 110 kPa and 40ºC, and 170 kPa and 118ºC, respectively. The flue gas side pressure drop was 8.15 kPa for the PC system and 4 kPa for the NGCC system.
The electrical energy requirement of the capture facility came from three types of equipment: the flue gas blower, the pumps to circulate the absorption liquid and the CO2 product compression equipment.
The SAS model was also a regenerative system in which sorbent material circulated between the adsorber and desorber columns, the operating temperatures in which were 60ºC and 130ºC, respectively, and the pressure atmospheric in both cases.
Calculation of the thermal energy input of this facility relied on the work of Li et al and sums the desorption heat, equal to 1.5 GJ/tCO2, and the sensible heat required to raise the sorbent from the adsorption temperature to the desorption temperature, which depends on the heat capacity of the sorbent (1.5 KJ/kg/K), the temperature difference between the adsorption and desorber columns (70 K) and the cyclic CO2 capacity of the sorbent.
Figure 4 shows the net thermal energy requirement of the process as a function of sorbent working capacity. For the selected adsorption and regeneration conditions, this is about 1.7 GJ/tCO2 at sorbent working capacities of 2.6-3.1 mol/kg when heat integration is applied, or 40 per cent lower than in MEA-based systems with advanced stripper configurations – typically 3 GJ/tCO2.
|Figure 4: Thermal energy input for an SAS facility in a process simulation. Heat integration yields 75 per cent of the energy needed to raise the sorbent temperature from that of adsorption to that of desorption.|
At working capacities higher than 2 mol/kg the reaction heat dominates the thermal energy demand. Furthermore, increasing the sorbent working capacity beyond 4 mol/kg does not improve the overall performance.
Heat integration (lean/rich heat exchanger) is more important in aqueous solvent based processes, as figure 5 shows. In the calculations underlying Figure 4 the assumption was made that 75 per cent of the sensible heat required for heating the sorbent can be recovered. Possible co-adsorption of water in the adsorber column was not taken into account.
|Figure 5: Importance of capture plant heat integration for MEA and SAS. For MEA it is assumed that the pinch in the rich-lean heat exchanger is 10ºC|
The calculation of the electrical energy required summed the energy consumed by the flue gas blower, sorbent circulation and CO2 compression. Generally compression of captured carbon accounts for 20-25 per cent of the total electrical energy requirement. The experimental work at the University of Twente shows sorbent regeneration at elevated pressure of up to 1000 kPa can further improve the energy efficiency of an SAS-based capture process in a CCS setting. Releasing CO2 at 500 kPa, for example, would reduce the electrical energy required for compression significantly and eliminate one compression stage.
The flue gas side pressure drop in the absorber depends strongly on the type of contactor chosen. Fixed bed operation leads to very high pressure drops and is not a realistic option. Moving bed and fluidisation technologies are more suitable options.
From an ongoing process optimisation study by Veneman et al the most promising option appears to be a countercurrent gas–solid (G-S) trickle bed contactor. The pressure drop for this contactor is estimated to be around 6.4 kPa. Our energy calculations used a value of 7.5 kPa. On the regenerator side the pressure drop is less critical and a (staged) fluid bed configuration might be a more cost-effective option.
The installation of a capture facility at a power plant results in a decrease in electrical power output because steam is extracted from the plant for the regeneration of the solvent or sorbent. Table 2 shows the effect.
However, the lower thermal energy requirement of the SAS-based process means the loss in gross power is less than in the case of a solvent-based process.
A PC power plant emits 2.7 times more CO2 per MWh than an NGCC plant, so application of carbon capture at a PC plant has a higher impact on the power output.
Figure 6 plots the energy demand in GJ/tCO2 of the capture facility for each case. The columns labelled ‘Thermo min. work’ show the thermodynamic minimum amount of work required for CO2 separation and compression to 11,000 kPa.
|Figure 6: Electricity consumption for capture facilities studied|
The desorption energy fraction is related to the decrease in the electrical energy output of the power plant caused by the extraction of low-pressure steam for the regeneration of sorbent/solvent. The electrical energy fraction includes energy for the circulation of sorbent or solvent and the power demand of the flue gas blower. The ‘compression energy’ fraction is related to the power consumption of the CO2 compressor.
Figure 6 shows that an SAS-based capture facility at an NGCC plant would be 32 per cent more efficient than an MEA-based one. For a PC plant an SAS-based plant is 35 per cent more efficient.
Additional savings could be reached by further reduction of the flue gas side pressure drop and possibly via high-pressure sorbent regeneration, which could save up to 10 per cent of the CO2 compression energy. A value of 65 kJ/molCO2 for the reaction heat was assumed here, typical for secondary amines. The values reported by Gray et al of 47.5 and 51.6 kJ/molCO2 for immobilised PEI would reduce the reaction heat contribution by another 0.3-0.4 GJ/tCO2 to below 1.5 GJ/tCO2 and reduce the overall energy consumption by 0.1 GJ/tCO2.
Process design & CAPEX
Figure 7 shows the breakdown of purchased equipment costs for an MEA-based capture unit at a 500 MWe PC plant (excluding the CO2 compressor). Almost 30 per cent of the costs is associated with the absorber. The size of the absorber is predominantly determined by the residence time of the flue gas.
|Figure 7: Breakdown of the costs of purchased equipment of an MEA-based capture facility (excluding the compressor) at a 500 MWe coal fired power plant|
The application of SAS could potentially reduce the size of the absorber column because G-S systems have a higher overall mass transfer rate.
For conventional MEA scrubbers, overall mass transfer rates depend on the type of packing but are 10-2 to 10-1 (kLa/s). Typical values for G-S systems are 101 kGa/s . So CO2 uptake rates are expected to be higher than for the MEA benchmark technology, even though CO2 absorption is significantly enhanced by the chemical reaction between CO2 and the dissolved MEA.
The enhancement factor EA can be as high as 102. This means an SAS-based facility could be equipped with an adsorption column smaller than the absorption column of an MEA-based facility.
When applying SAS, a G-S trickle flow reactor could perform as well as a CO2 adsorber. It allows for counter-current G-S contacting and can be operated at gas velocities equal to or higher than those in an MEA scrubber, without resulting in unacceptable pressure drops. Preliminary sizing of a G-S trickle flow adsorber was performed using the hydrodynamic model for trickle flow reactors developed by Dudukovic et al.
Table 3 compares SAS and MEA systems for productivity, pressure drop in view of electricity consumption for the flue gas blower, and operating velocity in view of the footprint of the absorber/adsorber. Productivity was defined as the amount of CO2 captured per second per m3 of installed reactor volume, which is a measure of compactness and, therefore, of capital expenditure.
Expectations are that a gas-solid trickle bed adsorber would outperform an MEA scrubber column in productivity (1.20 mol/m3/s versus 0.56 mol/m3/s). The total installed reactor volume is expected to be two to three times smaller primarily because of the higher mass transfer rates in trickle bed reactors.
Table 4 compares a sorbent-based capture process with state-of-the-art capture technology that employs aqueous amine solvents.
Co-adsorption of water by a sorbent results in extra energy consumption during capture and complicates sorbent regeneration because CO2-water separation would be needed. However, first measurements performed by the University of Twente have shown that CO2 capacity is not negatively affected by the presence of water.
A second recommendation is to determine possible sorbent degradation by traces of NOx and SOx in the flue gas, as well as the potential presence of nitrosamines and nitramines in the treated flue gas and, with this, the need for a washing section.
The mechanical stability of the sorbent should also be examined to determine the maximum number of cycles that can be attained with a single batch.
Sorbent regeneration under elevated pressure needs further examination too. Pressurised regeneration would enable further energy savings as CO2 compression would consume less energy and one compression stage can be omitted.
A final recommendation is to prepare a conceptual design of a full-scale sorbent-based carbon capture unit to provide a more detailed estimate of what the savings will be in energy and capital expenditure.
1.Brunetti A, Scura F, Barbieri G, Drioli E, Membrane technologies for CO2 separation, J. of Membrane Science, 359 (2010) 115-125.
2. Abu-Zahra M R M, Schneiders L H J, Niederer J P M, Feron P H M, Versteeg G F, CO2 capture from power plants: Part I. A parametric study of the technical performance based on monoethanolamine, Int. Journal of Greenhouse Gas Control, 1 (2007), 37-46
3. Abu-Zahra M R M, Niederer J P M, Feron P H M, Versteeg G F, CO2 capture from power plants. Part II. A parametric study of the economical performance based on mono-ethanolamine, Int. Journal of Greenhouse Gas Control, 1 (2007), 135-142
4. Mathieu P, The IPCC special report on carbon dioxide capture and storage, ECOS 2006: Proceedings of the 19th International Conference on Efficiency, Cost, Optimisation, Simulation and Environmental Impact of Energy Systems, Vols 1-3, (2006) 1611-1617.
5. Ebner A D, Gray M L, Chisholm N G, Black Q T, Mumford D D, Nicholson MA, Ritter JA, Suitability of a solid amine sorbent for CO2 capture by pressure swing adsorption, Industrial and Engineering Chemistry Research, 50 (2011) 5634-5641.
6. Veneman R, Li Z S, Hogendoorn J A, Kersten S R A, Brilman D W F, Continuous CO2 capture in a circulating fluidised bed using supported amine sorbents, Chemical Engineering Journal, (2012).
7. Li Z S, Cai N S, Croiset E, Process analysis of CO2 capture from flue gas using carbonation/calcination cycles, AIChE J, 54 (2008) 1912-1925.
8. Fauth D J, Gray M L, Pennline H W, Krutka H M, Sjostrom S, Ault A M, Investigation of porous silica supported mixed-amine sorbents for post-combustion CO2 capture, Energy and Fuels, 26 (2012) 2483-2496.
9. Rochelle G, Chen E, Freeman S, Van Wagener D, Xu Q, Voice A, Aqueous piperazine as the new standard for CO2 capture technology, Chemical Engineering Journal, (2011).
10. Fisher K S, Beitler C, Rueter C, Searcy K, Rochelle D G, Jassim D M, Integrating MEA regeneration with CO2 compression and peaking to reduce CO2 capture costs, DOE/NETL report, Contract No. DE-FG02-04ER84111, (2005).
11. Jassim M S, Rochelle G, Eimer D, Ramshaw C, Carbon dioxide absorption and desorption in aqueous MEA solutions in a rotating packed bed, Industrial and Engineering Chemistry Research, 46 (2007) 2823-2833
12. Kiel J H A, Prins W, van Swaaij W P M, Mass Transfer between gas and particles in a gas solid trickle flow reactor, Chemical Engineering Science, 48 (1993), 117-125
13. Dudukovic A P, Nikacevic N M, Petrovic D L, Predojevic Z J, Solids holdup and pressure drop in gas flowing solids-fixed bed contactors, Industrial & Engineering Chemistry Research, 42 (2003), 2530-2535.
A. B. M. Heesink and G. Magneschi are from DNV KEMA, based in the Netherlands, and R. Veneman and D. W. F. Brilman are researcher at the Dutch University of Twente.
Power Engineering International Archives
View Power Generation Articles on PennEnergy.com