Performance of MEA and amine-blends in the CSIRO PCC Pilot Plant at Loy Yang Power in Australia

Transcription

1 NOTICE: this is the author s version of a work that was accepted for publication in Fuel. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Fuel, vol. 101 (Nov. 2012). Performance of MEA and amine-blends in the CSIRO PCC Pilot Plant at Loy Yang Power in Australia Yuli Artanto a,1, James Jansen a, Pauline Pearson a, Thong Do c, Aaron Cottrell b, Erik Meuleman a and Paul Feron b a CSIRO Energy Technology, Clayton, VIC, Australia, b CSIRO Energy Technology, Newcastle, NSW, Australia c CSIRO Energy Technology, North Ryde, NSW, Australia ABSTRACT Chemical reactive liquid absorption in post-combustion carbon capture (PCC) has sparked an interest of many researchers for improvement in creating high CO 2 absorption capacity and minimizing the reboiler heat duty for regeneration. This paper discusses results from performance trials on different solvents: mono-ethanolamine (MEAbaseline), a mixture of MEA with 2-amino-2-methyl-1-propanol/AMP (Blended Amine 1) and Blended Amine 2 (a propietary solvent developed by Research Institute of Innovative Technology for the Earth (RITE), Japan). The trials were carried out in CSIRO s PCC pilot plant at Loy Yang Power. The pilot plant receives flue gas from this 2,2 GW brown coal-fired power station in the Latrobe Valley, Victoria. The study has shown the benefits of using blended solvents for CO 2 capture in terms of CO 2 recovery and reboiler duty for solvent regeneration. The correlation of CO 2 recovery and process parameters, i.e., liquid and gas (L/G) ratio and lean loading, has been established for the range of solvents tested. The results show that, in general, an increase of L/G ratio increases the CO 2 recovery. It is also noted that a solvent at higher lean loading requires a greater solvent (recycle) flow rate, or vice versa, in order to capture a given amount of CO 2. The effect of reducing the stripper s bottom temperature from 115 to 112 C reduced the CO 2 recovery of both MEA and Blended Amine 1 as their solvent lean loadings are raised. The same temperature change, however, did not significantly decrease CO 2 recovery of Blended Amine 2, as it also shown by insignificant changes in solvent lean loadings. The results also indicate that the reboiler heat duty is dependent upon L/G ratio, solvent lean loading and blending solvent type. For the MEA-115 C, a minimum reboiler heat-duty as a function of L/G ratio is clearly observed and the Blended Amine C has also shown a similar pattern. In constrast, Blended Amine C showed excessive energy duty, which indicates unfavourable condition for this solvent. This is not the case for Blended Amine 2. It is observed that the reboiler heat duty decreases as stripper bottom temperature decreases, which in turn raises the lean loading for both MEA and Blended Amine 1. On the other hand, the lean loading of Blended Amine 2 in the temperature range examined did not affect the reboiler heat duty changes. The investigation also found that the magnitude of reboiler duty is decreasing following the order MEA > Blended Amine 1 > Blended Amine 2. In order to obtain CO 2 recovery of 85%, the Blended Amine C (L/G = 3.60) has similar reboiler duty relative to that of MEA-115 C with L/G ratio of For the Blended Amine 1 (L/G = 3.04), however, the reboiler heat duty increased by 51% at a similar temperature. The temperature reduction to 112 C decreased the reboiler heat duty of both Blended Amines 1 and 2 by 11% (L/G = 4.01) and 14% (L/G = 3.60) respectively. Further raising the CO 2 recovery to 90-95%, in comparison to that of MEA at 115 C with L/G ratio of 4.20, the reboiler heat duty of Blended Amine 1 has risen by 49% (L/G = 3.95) and of Blended Amine 2 dropped by 5% (L/G = 3.52) under a similar temperature. At 112 C, the reboiler heat duty of Blended Amine 1 decreased by 18% (L/G = 5.02). Blended Amine 2 may give a similar reboiler heat duty if run at a larger L/G ratio. Overall, the result of the MEA baseline and blending solvents has shown different behaviours of CO 2 stripping. The distribution of the components of reboiler heat duty may explain this difference. For MEA, the condenser heat, which equals the heat needed for water evaporation, is a major contributor to the 1 Corresponding author: yuli.artanto@csiro.au

2 reboiler duty. For both Blended Amine solvents, the magnitude of sensible heat and the heat of CO 2 desorption are more pronounced than the condenser heat. At the minimum reboiler duty reached by MEA and Blended Amine 1, the condenser heat share is lower than the other reboiler duty components. Therefore, further works involve optimisation of the PCC process with blended Amines and other blends. Keywords: PCC pilot plant, CO 2 recovery, MEA, AMP, Blended Amines 1 and 2, temperature, reboiler heat duty and L/G ratio 1. Introduction The brown coal-fired power generation is the source of about half of Victoria s current greenhouse gas emissions [1]. It is therefore clear that CO 2 -emission reduction strategies aimed at the existing power stations are urgently needed to provide a path towards environmental sustainability of the Victorian brown coal industry. The use of carbon capture and sequestration (CCS) is essential for this path and Victoria is well placed due to the vicinity of vast storage capacity in the Gippsland basin, which equals to more than 100 years of CO 2 produced in Victoria at current rates [2]. Post-combustion capture (PCC) is an important first part of the CCS chain. The implementation of PCC in the Victorian case requires specific focus towards its technological development in regards of three following issues: 1. Brown coal is not sold into a world market due to its high moisture content in contrast to black coal, oil or natural gas. Therefore, it is expected that brown coal prices will remain at low price levels thus continuing to provide the basis for low cost electricity for Victoria. The capture of CO 2 will result in a large increase in the cost of electricity generation, which needs to be addressed; 2. Brown coal flue gases are available at high temperature, have high water content and contain alkaline ash. This provides a challenging environment for chemical absorption processes; 3. The combined process of coal mining, power generation and PCC should use less water than current power generation as ground water level retreats to an unsustainable situation. A wide range of CO 2 separation techniques such as absorption into a liquid, adsorption onto a solid and membrane permeation processes [3] are in principle available to capture CO 2 from flue gases. Due to a high volume flow rate of gas stream from a coal-fired power plant and low CO 2 partial pressure in the gas stream [4], absorption into a liquid is currently the most suitable option for capturing CO 2 [3, 5, 6]. The use of mono-ethanolamine (MEA) as a post-combustion capture (PCC) based solvent has found practical application in CO 2 separation from flue gases [3]. This is due to its low cost, its ability to capture CO 2 from low pressure flue gas and its fast reaction kinetics with CO 2. The PCC unit can also be retrofitted to an existing and/or easily integrated to a new power station [7-9]. A large amount of energy is required for solvent regeneration resulting in a large drop in power station efficiency [6, 8, 9]. Degradation and solvent losses are also identified as an important environmental and cost factor, particularly when it comes to commercial scale. Furthermore, in a financial cost analysis of an MEA-based PCC plant, Veawab et al. [10] concluded that absorption and desorption (solvent regeneration) sections are the major contributors to capital cost investment. They also specified that up to 70% of the total operating costs will be needed to provide the heat duty for the solvent regeneration (desorption section) [10]. The solvent scrubbing technique is considered to be the most advanced post-combustion capture technology [11]. A strategy of optimizing absorber conditions, in order to maintain a higher masstransfer rate by controlling minimum reboiler heat duty required, is a way to introduce a trade-off approach in order to make the PCC plant feasible. In this paper, discussion on initial focus to look at MEA based results as a baseline and an attempt to reduce regeneration energy (reboiler heat duty) 1

3 without impeding the CO 2 capture performance are presented. This may be achieved by utilizing MEA-based blends and/or a new mixture solvent as an alternative solvent. CSIRO has devised a transportable pilot plant based on MEA. For the Latrobe Valley Post-combustion Capture project CSIRO operates the pilot plant, based on amine technology, which is connected to flue gases from Victorian brown coal-fired power station at Loy Yang Power (2.2 GW). The objective of the PCC pilot plant trial program is to set a baseline based on 30 wt-% MEA and benchmark other solvents against this baseline. Up to date, our work has examined two different solvents, i.e., Blended Amine 1, which is a mixture of MEA and 2-amino-2-methyl-1-propanol (AMP) and Blended Amine 2, which is a proprietary solvent. The result for these two solvents are compared to the MEA baseline results in terms of CO 2 recovery and heat duty required. 2. Experimental 2.1. Chemicals and feed gas composition Concentrated MEA and NaOH 32 wt-% are obtained from Water Treatment Services (Aus) Pty Ltd. 20 wt-% MEA and 10 wt-% AMP are blended for making Blended Amine 1. The concentrated AMP is also supplied by the Water Treatment Services (Aus) Pty Ltd. Blended Amine 2 is supplied by RITE, Japan. These solvents (except NaOH 32 wt-%) were diluted with mains water prior to use. Table 1 shows important flue gas constituents and its concentration in the flue gas from Loy Yang Power used during the operation. Furthermore, the flue gas contains significant amounts of SO 2 and NO x. The SO 2 is washed off in the pre-treatment column with 32 wt-% NaOH Description and operation of the PCC pilot plant at Loy Yang Power The transportable PCC pilot plant at Loy Yang Power was designed to capture CO 2 from real flue gases using 30 wt-% MEA. Due to availability of the MEA-based result, it was decided to use the results for this solvent as the baseline case. The PCC pilot plant is able to receive 150 kg/h real flue gas from unit 2. The pilot plant consists of one flue gas pre-treatment unit, two absorber columns to capture the CO 2, a stripper column for stripping off the CO 2 from the solvent and a 120 kw electricboiler unit to generate steam for heating the media in the stripper s reboiler which uses a plate reboiler. The pilot plant has been operated with two absorbers in series (cf. Fig. 1). Two short absorbers were designed to provide a compact plant that is easily transportable. A 200DN stainless steel (211 mm ID) column is used for each absorber. The column height of each absorber column is 940 cm. Each column consists of two packed beds of pall rings, the height of a single bed being 135 cm. The height of the stripper column is 690 cm and it is made from 150DN stainless steel (161 mm ID). The stripper column contains a packed bed of pall rings with a height of 390 cm. During the operation, the level of liquid in each column is kept constant to ensure a constant liquid flow-rate. Water and solvent vapour released through the top of the stripper column are recovered in a 60 kw condenser. The vapour and condensed-liquid are separated in a flash drum situated near the solventfeed tank. The liquid phase is returned to the solvent feed tank after prior mixing with hot lean solvent from the stripper s bottom product. The vapour phase, containing about 98 vol-% CO 2, is delivered back to the flue gas duct. The mass flow of concentrated-co 2 gas is measured with a coriolis meter. 2

4 Fig. 1 gives a simplified flow diagram of the pilot plant. A portion of flue gases is taken-off by a blower before reaching in chimney. The flue gas is cooled and then flows into the knock-out drum where condensates and particulates are separated. The pilot plant consists of a flue gas pre-treatment unit, which use a caustic solution (32% NaOH), before the flue gas enters the absorber column 2. The average SO 2 concentration in the flue gas from the power station is around 200 ppmv and up to 98% is removed in the pre-treatment column. In series operation mode, the pre-treated flue gas from the pre-treatment column initially flows to the absorber column 2, where it is counter-currently contacted with a CO 2 -rich solvent from the bottom of absorber 1 at around C under atmospheric pressure. Then, the gas of low CO 2 content goes out from the top of absorber 2 to the bottom section of absorber 1 where it also counter-currently contacts with fresh CO 2 -lean solvent pumped from the solvent feed-tank. Absorption and chemical reactions between CO 2, a weak acid, and aqueous CO 2 -contained solvent, a weak base, occurs in these two absorber columns. The gas of low CO 2 content, so-called gas out or lean gas, is emitted from the top of absorber column 1 and is subsequently returned to the power plant s stack. The cold CO 2 -rich solvent stream emerging from the bottom of the absorber column 2 is pre-heated to C (the exact temperature depending on the T approach to be used) in a cross-solvent heat exchanger, utilizing the heat from the hot lean solvent (typically at C) flowing out of the bottom stripper column. This pre-heated solvent will be referred as hot rich solvent. Temperature of the hot lean solvent drops after releasing the heat into the cold rich solvent and is then further reduced with cooling water until it reaches the desired temperature before the lean solvent flows into the solvent-feed tank. From the solvent-feed tank, the solvent is then pumped into the absorber column 1. In the stripper column, water vapour and CO 2 separation from the solvent (through endothermic reaction) have occurred due to heat supplied from the reboiler (typically C). The reboiler heat is obtained from latent heat of the steam (at a steam pressure range of kpa). Typical stripper bottom pressures at given stripper bottom temperatures are presented in Table 2 below. Gas sampling points (indicated by G1 to G5) and liquid sampling points (indicated by L1 to L4) are installed at several points in the pilot plant as shown in Fig. 1 in order to determine species composition in gas streams and CO 2 loading in liquid streams respectively. Flue gas flow-rate at the blower inlet, solvent flow-rate and ph of the rich and lean solvent are monitored Experimental Program Three experimental campaigns with different solvents were carried out at the pilot plant. The aim of the first campaign was to evaluate the performance of 30 wt-% MEA, as a baseline result for our pilot plant. In this baseline operation, it should be noted that adequate CO 2 balance throughout the operation was achieved with typical fluctuations of ± 10% around the zero-point [12]. In the second campaign MEA was blended with AMP that has been reported [13, 14] to be useful in enhancing the loading capacity and lowering the binding energy, thereby decreasing the energy duty. This blended solvent was named Blended Amine 1. In the third campaign a novel proprietary amine blend was trialed; so-called Blended Amine 2, which has been developed by RITE (Japan). The two Blended Amines were examined in the pilot plant aiming at similar percentages of CO 2 recovery as compared to the baseline case with 30 wt-% MEA. Table 3 illustrates a range of processing conditions applied in the pilot plant in order to trial different solvents for CO 2 recovery. The performance of solvent is evaluated based on its ability in capturing CO 2 and the lowest reboiler heat duty necessary for solvent regeneration. In this study, a temperature 3

5 of 115 C was selected as maximum stripper bottom temperature in order to minimize thermal degradation and corrosion which will become excessively high at higher temperature. From Table 4, it can also be seen that there is an upper limit to the pressure in the stripper column because the temperature difference between top packing and bottom cannot be too small. Therefore, using the current design could not do better than that. To increase the pressure further, the packing height in the stripper column would have to be increased in order to improve CO 2 stripping and hence reduce the reboiler heat duty. In our pilot plant, even though stripper bottom temperature reduction is small, however, it can give upper packing temperature significantly different depending on the amount of water vapour flows. This, therefore, gives different amount of reboiler heat duty. In this paper, it also shows that operating the stripper column bottom at a temperature below 112 C gave a little of benefit as the CO 2 recovery will be smaller due to the increased lean loading even though it can give lower reboiler heat duty Gas and liquid analysis Gas analysis is conducted on-line by using a GASMET CEMS. The instrument incorporates a Fourier Transform Infrared spectrometer, a temperature controlled sample cell, and signal processing electronics. The gas analyser is designed for continuous emission monitoring (CEM). Gas components detected are H 2 O, CO 2, CO, N 2 O, NO, NO 2, SO 2, NH 3, HCl, CH 4, C 2 H 4, C 2 H 6, C 3 H 8, C 6 H 14, Formaldehyde, Acetaldehyde, Ethanol, Ethanolamine, HF and O 2. A liquid sample is usually collected within the time period when the absorption process is considered to have reached equilibrium, before the trial is stopped. Liquid samples, which represent lean (L1 and L4) and rich loadings (L2 and L3), are collected into heat and shock resistant glass bottles. The liquid sample is then delivered overnight to CSIRO Clayton laboratories for further analysis to determine CO 2 loading, total/free amine and CO 2 concentrations. An agilent 6850 series gas chromatograph with an automatic liquid sampler (8 samples) is used to determine CO 2 concentration and total/free amine by a volumetric titration method. The instrument is also equipped with an FID detector. The analysis conditions are 1 µl injection, 200 ml/min flowrate, 100:1 split; oven 110 C for 2 min, ramp at 20 o C/min to 250 C then hold for 5 min. The procedure is described below: CO 2 determination: A sample of absorber solution (~2.5 g) is weighed accurately into a vial. The ph of a methanol solution (~50 ml) is adjusted to ph 11 using NaOH (0.5 M). The MEA sample is added to the methanol solution and the mixture is titrated back to ph 11 with NaOH (0.5 M). The second volume of NaOH is used to calculate the amount of CO 2. Amine determination: Free MEA: A sample of MEA solution (~0.25 g) is weighed accurately and diluted with water to ~50 ml. The diluted sample is titrated against HCl (0.1 M) until the equivalence point is reached at ~ ph 4-5. Total MEA: A sample of MEA solution (~0.25 g) is weighed accurately and diluted with water to ~50 ml. A known excess of HCl (20 ml, 0.1 M) is added to the mixture. The mixture is titrated against NaOH 4

6 (0.5 M) until the equivalence point is reached. The volume of NaOH used gives the amount of HCl in excess, hence the amount of HCl that reacted with the MEA can be calculated to determine the total MEA in solution. The errors in the CO 2 loadings, CO 2 and amine concentrations were based on duplicate or triplicate determinations CO 2 recovery and the reboiler duty for solvent regeneration CO 2 recovery is determined based on analysing the gas at the inlet to absorber column 2 (G2), outlet of absorber column 1 (G4) and CO 2 product at the top of the stripper column (G5). When there were large moment to moment fluctuations in the CO 2 product at the top of stripper column, the average CO 2 recovery calculated based on gas analysis (G2 and G4) and liquid analysis was used. Reboiler heat duty for the regeneration of CO 2 -loaded solvent was determined by making an energy balance around the reboiler and/or the stripper column. Two approaches can be used to determine reboiler heat duty in these pilot plant tests. Firstly, actual reboiler heat duty was determined by a manual measurement of steam condensed flow and the steam pressure supplied to the reboiler. This actual reboiler heat duty does incorporate heat loss to the ambient. The actual reboiler heat duty is expressed per amount of CO 2 absorbed. Secondly, the energy supplied by the reboiler can be devided into three contributors [8, 12, 15-17]: the heat required for evaporating the water, the sensible heat to heat up the solvent to reboiler temperature (Q sh ) and the heat of CO 2 desorption (Q des ). The heat required to evaporate the water is actually equivalent to the latent heat of water condensation which can be measured directly around the condenser (Q cond ). The equation can be represented as follows; Q reboiler = = = Q condenser Q cond m H w + Q + Q solvent heat desorption + Q + Q sh des + m c (T w s p - T bottom top ) - m H CO CO 2 2 where Q reboiler is the reboiler heat duty; m w is the amount of water flowing into the condenser; H w is the latent heat of water condensation; ms is the solvent flow rate; c p is the heat capacity of the solvent; T is the temperature of hot lean solvent going out from the bottom of stripper column; bottom T top is the temperature of hot rich solvent entering the top of stripper column; mco is the amount of CO 2 2 produced from the stripper column and H CO is the enthalpy of CO 2 2 desorption. Due to large variations in the ambient temperature and to the fact that the pipelines were not insulated, it is not easy to obtain representative values of actual reboiler heat duty. Therefore, in this paper we express the reboiler heat duty using the second method excluding the heat loss to the ambient Error determination The absolute error or proportional error of each instrumentation/equipment is listed in Table 5. The systematic error for specific gas components is listed in Table 6 below. Errors for CO 2 - lean loading and CO 2 -rich loading were obtained from triplicate measurements. The error for CO 2 - lean loading and CO 2 -rich loading is 0.01 and 0.02 mol CO 2 /mol Amine, respectively. 5

7 The error for CO 2 recovery is varied between 1% to 2%. The error of reboiler heat duty is varied between 0.3 to 0.4 MJ/kg CO 2. The errors for Q cond, Q sh and Q des are 0.1, 0.2 and 0.3 MJ/kg CO Results and discussion 3.1. MEA baseline L/G ratio and lean loading affect CO 2 recovery The effect of varying the L/G ratio on CO 2 recovery at different stripper bottom temperatures, i.e. 112 and 115 C, is illustrated in Fig. 2. The graph implies that for both temperatures, an increase in L/G results in an increase of available solvent to capture CO 2 from the same amount of flue gas. Fig. 2 also shows the effect of the stripper bottom temperature over a range of L/G ratios on CO 2 recovery. Increasing the stripper bottom temperature from 112 C to 115 C elevates the CO 2 recovery over the range of L/G ratios tested. This indicates increasing the bottom stripper temperature led to increasing CO 2 strip-off from the stripper column and hence, lowering of the solvent loading. This suggests that raising the temperatures increases the driving force, which promotes CO 2 transfer from flue gas to the solvent. This finding is supported by Fig. 3. The figure indicates a combination effect between L/G ratio and solvent-lean loading. The figure shows that to obtain higher CO 2 recovery, the plant can be operated in three modes; (i) at a constant L/G ratio, circulate lower solvent-lean loading, (ii) at a constat solvent-lean loading rate, increase L/G ratio and (iii) circulate high solvent-lean loading at L/G ratio as high as possible. It should be noted, however, that to operate the plant at higher L/G ratio will elevate the energy requirement for the pumps and increase plant dimensions. In order to make the solvent leaner, there will have to be an increase in energy required for solvent regeneration in the stripper column (reboiler heat duty). Therefore, there should be a trade-off in order to decide optimum plant operation MEA baseline L/G ratio and lean loading affect reboiler heat duty The dependency of CO 2 recovery upon L/G ratio and lean loading should also translate into a relationship between reboiler heat duty on the one hand and the L/G ratio and the lean loading on the other. Fig. 4 shows the influence of L/G ratio on the reboiler heat duty of solvent regeneration. The reboiler heat duty of MEA baseline examined at a stripper bottom temperature of 115 C shows a parabolic trend as a function of L/G ratio. This finding was also reported by Mangalapally et al. [16] and Cifre et al. [8], where they presented reboiler heat duty as a function of solvent mass flow rate. Fig. 5 shows the effect of L/G ratio associated with the lean loading on the reboiler heat duty. At stripper bottom temperature of 115 C, the lean loading slightly changes across the L/G ratio ( ) tested from 0.17 to Similar changes in the lean loading (0.21 to 0.27) at different L/G ratio ( ) also noticed at 112 C. It is observed that at similar L/G ratio (3.22 and 3.19 at 115 C and 112 C respectively), the reboiler heat duty did not significantly change within the limits of error as their lean loading also did not change within the limits of error. A similar situation is also noticed for another similar L/G ratio (4.00 and 3.97 at 115 C and 112 C respectively). It shows, however, that the reboiler heat duty significantly increased even though at similar L/G ratio (2.32 and 2.34 at 115 C and 112 C respectively) and similar lean loading (0.19 and 0.21 at 115 C and 112 C respectively) within the limits of error. It is also noticed that the reboiler heat duty of L/G ratio at 4.2 increased as it has a low lean loading of The increase in the reboiler heat duty of both at lower L/G ratios (2.30 and 2.32) and higher L/G ratios (4.2) for 115 C can be explained in considering the three components of reboiler heat duty separately (see sub section 2.5 and Fig. 6A). The increase of reboiler heat duty at L/G ratio of 2.30 and 2.32 may be attributed to the extra heat removed by the cooling water in the condenser (Q cond ) as more water is 6

8 evaporated for stripping-off the CO 2, which in turn makes the solvent leaner (see Fig. 5). In contrast to what is observed at 115 C, Q cond for 112 C is very low compared to Q sh and Q des (see Fig. 6 Fig. 6B), indicating less water was evaporated than that at 115 C. The values of Q sh and Q des do not change within the limits of error from 112 C and 115 C. The major increase of Q cond for the 115 C experiment is associated with the reduction of lean loading to below 0.20 (top packing temperature around 110 C, see Fig. 7). The Q cond, however, does not change significantly at lean loading between 0.20 and The higher Q cond for 115 C (top packing temperature of 105 C) compared to that for 112 C (top packing temperature of 100 C), at a similar lean loading of 0.21, may be attributed to the increase in top packing temperature (cf. Fig. 7) as a consequence of the increase in the steam flow to the top of the stripper column due to the stripper bottom temperature increase. The raise in reboiler heat duty for the L/G above 3.8 is due to an increase in the solvent flow, hence increasing the sensible heat (Q sh ). For L/G at 4.2, not only Q sh, but also Q cond contributes to the increase in reboiler heat duty. The guide-line in Fig. 4 shows that the reboiler heat duty is a minimum at L/G around 3.2 for 115 C stripper bottom temperature. As seen in Fig. 6 Fig. 6, the Q cond contribution at this L/G value is small compared to the Q sh associated with heat up of the solvent and the heat of desorption (Q des ). The amount of Q des varied little with L/G, because the Q des for CO 2 -loadings below 0.5 is relatively constant even with temperature changes [16]. Fig. 4 also indicates that, for the stripper bottom temperature of 112 C, reboiler heat duty is not significantly affected by L/G ratio (2.34 to 3.97) and lean loading (0.21 to 0.27) (cf. Fig. 5). Fig. 7 also indicates that the changes of lean loading (0.21 to 0.27) did not give significant enhancement in Q cond within the limits of error, which in turn reflects a small amount of steam leaving the stripper column A comparison between MEA alone and Blended Amine solvent CO 2 -lean loading and CO 2 - rich loading Table 7 shows lean and rich CO 2 -loadings for MEA alone, Blended Amines 1 and 2 at different stripper bottom temperatures. In general, CO 2 absorption into solvent occurred mainly in absorber column 1 rather than in column 2. This is because the CO 2 -solvent loading into column 1 is lower than that going into column 2 and hence, the driving force of CO 2 -mass transfer in column 1 is higher than that in column 2. For all solvents tested in the pilot plant, the CO 2 -lean loading also reduces as the stripper bottom temperature increases. It can be seen from the Table that at 115 C, MEA has higher lean loading amounts than those of blended amines. The difference of lean loading between MEA and Blended Amine 1, however, was not significantly large at 112 C. Blended Amine 2 has lean loading much lower than those of MEA and Blended Amine 1 at both stripper bottom temperatures. At 112 C, the lean loading values for Blended Amine 1 are also comparable to those for MEA at 115 C. For Blended Amine 2, however, the lean loading did not change as the stripper bottom temperature was reduced from 115 C to 112 C, but, the lean loading did increase as the temperature dropped further to 107 C. There are two factors that can explain the difference among the solvents. The first factor is heat of absorption. It can be assumed that the amount of heat required for reaction with CO 2 is similar to the heat required for breaking the CO 2 -Amine bonds. Among the three alkanolamines tested in the pilot plant, MEA has the highest heat of reaction with CO 2. It requires an energy of about 1920 kj/kg CO 2 [3]. It is noted that Blended Amine 2 properties data cannot be shown as it is a proprietary solvent and readers are referred to papers published by RITE. Our estimation reveals that Blended Amine 1 needs 7

9 kj/kg CO 2, indicating the difference in heat of reaction between MEA and Blended Amine 1 is not large, but this suggests that in the blended amines amine-co 2 bonds break more easily than in the MEA at a similar amount of heat supplied. The second factor is the heat required to generate steam. The steam is needed to create a driving force for CO 2 stripping in the stripper column. Using a model developed in our group [18], we can also estimate an equilibrium CO 2 partial pressure of MEA, AMP and Blended Amine 1 as described in Fig. 8. The Figure indicates that MEA requires a lower operating CO 2 partial pressure than Blended Amine 1 and AMP, hence MEA needs the largest heat consumption to create steam. At a given CO2 loading the order of equilibrium CO 2 partial pressures is Blended Amine 2 > Blended Amine 1 > MEA. In term of rich solvent and cyclic capacity (the difference between rich solvent from column 2 minus lean solvent), Blended Amine 1 has higher capacity than those of MEA and Blended Amine A comparison between MEA alone and Blended Amine solvent L/G ratio and lean loading affect CO 2 recovery and reboiler heat duty It would be desirable to seek a remedy for the MEA solvent s drawback in terms of energy used for regeneration. An examination of a mixture solvent, which was intended to show benefits in terms of adequate CO 2 absorption capacity and lower reboiler heat duty, might be one of the counter-measures to this shortcoming. Pilot plant results from previous workers [5, 19] have shown that a blended solvent has a lower heat of desorption than a single solvent, which can contribute to reduce the reboiler heat duty. In this study, the behavior of Blended Amines 1 and 2 was examined as a function of L/G ratio and of stripper bottom temperature. The effect of changing the stripper bottom temperature on solvent lean loading was also examined. The performance of blended solvents is expressed by the CO 2 recovery and the reboiler duty. They are compared with the MEA baseline. The amine concentration of the blended solvent was maintained constant throughout the campaign. Fig. 9A and 9B show the variation of CO 2 recovery with L/G ratio at different stripper bottom temperatures for both Blended Amine solvents. Fig. 9A indicates that, for both Blended Amines tested under the highest stripper s bottom temperature (115 C) the magnitudes of CO 2 recovery are comparable to that of the MEA baseline at 115 C. Fig. 9B shows that for Blended Amine solvent 1 operating at L/G below 4.0, a stripper bottom temperature reduction from 115 C to 112 C and 109 C significantly decreased the CO 2 recovery compared to those of MEA and Blended Amine 2 at 115 C. A comparable CO 2 recovery to that of MEA at 115 C, however, can be obtained when Blended Amine 1 is circulated at higher L/G ratio. It is noted that for Blended Amine 2, the CO 2 recovery is not greatly lowered when the stripper bottom temperature lowered to112 C.. The effect of lowering the stripper bottom temperature on the CO 2 recovery reductions of Blended Amine 1 may be associated with increasing the solvent lean loading. Fig. 10 shows that a reduction of stripper bottom temperature raises the lean loading. For all temperatures examined, at a relatively constant lean loading, increasing the L/G ratio improves the CO 2 recovery. The figure also shows that operating the stripper bottom temperature at 115 C for Blended Amine 1 resulted in an increase of reboiler heat duty, which in turn produced much leaner solvent compared to that of MEA at 115 C (cf. Table 7). This suggests that 115 C is an unfavourable operating condition for Blended Amine 1. The graph also indicates that operating the stripper bottom temperature at 112 C for this solvent gave a comparable lean loading to that of MEA at 115 C (cf. Table 7) and also significantly lowered the reboiler heat duty compared to that of operating at 115 C. 8

10 For 115 C, as L/G ratio increased, the reboiler heat duty decreased significantly. The reboiler heat duty might be expected increased if the L/G ratio increases. Nevertheless, the magnitude of reboiler heat duty shows large difference across the L/G ratio compared to that of MEA at 115 C. The reboiler heat duty as a function of L/G ratio for Blended Amine 1 at 112 C did not change significantly within the limits of error. The reboiler heat duty for Blended Amine 1 at 112 C is comparable to that for MEA at the same temperature. Further reduction of temperature to 109 C did not lower the reboiler heat duty significantly even though the CO 2 recovery was significantly reduced. It is also observed that the effect of L/G changes is not sensitive to the changes of reboiler heat duty at 109 C, as all L/G values generated similar lean loadings within the limit of error. Fig. 11 also shows that the reduction of stripper bottom temperature for Blended Amine 2 led to a decrease in the reboiler heat duty regardless of L/G ratio value even though the change in temperatures had little effect on the lean loading within the limits of error (except for lean loading at 112 C and L/G ratio 3.56, (cf. Table 7). Both temperatures show similar trends in the increase of reboiler heat duty as a function of increasing the L/G ratio. Further trials could be conducted at lower L/G ratio in order to investigate whether the reboiler heat duty at low L/G ratio increases as in the case of MEA at 115 C and Blended Amine 1 at 112 C. Operating at the L/G ratio lower than 2.00, however, will not benefit CO 2 recovery so that such trials would not have any practical value. It can also be seen in Fig. 12 Fig. 12 that to achieve 85% CO 2 recovery, MEA trialled at 115 C with L/G of 4.12 requires the reboiler heat duty of 5.1 MJ/kg CO 2. Both MEA at 112 C and Blended Amine 1 at 109 C, however, require L/G ratios higher than 6.0 (cf. Fig. 9B) in order to attain similar CO 2 recovery, which may not favourable from a practical point of view. If Blended Amine 1 is operated at 115 C, the reboiler heat duty, unfortunately, rises to 10.5 MJ/kg CO 2, which is also not feasible. However, the reboiler heat duty of Blended Amine 2 at the same temperature with L/G ratio of 3.60 has a similar value to that of MEA. It is also notable that the reduction of temperature to 112 C decreases the reboiler heat duty of both Blended Amines 1 (L/G = 4.01) and 2 (L/G = 3.61) by 11% and 14%, respectively. In order to obtain 90%-95% CO 2 recovery, the reboiler heat duty of MEA at 115 C with L/G = 4.20 increases by 17% compared to that of required to get 85% CO 2 recovery. This increase is mainly due to the increase of condenser duty (cf. Fig. 6A Fig. 6) as indicated earlier. In comparison with MEA at 115 C, the reboiler heat duty of Blended Amines 1 (at L/G = 3.95) is higher by 2.0 MJ/kg CO 2 (increase of 49%) and the reboiler heat duty of Blended Amine 2 (at L/G = 3.52) is a little lower by 0.6 MJ/kg CO 2 (reduction of 5%). If Blended Amine 1 is operated at 112 C with L/G ratio of 5.02, it has a reboiler heat duty 18% lower than that of MEA at 115 C. Blended Amine 2 at 112 C may possibly give a similar reboiler heat duty to that of Blended Amine 1 at 112 C if the L/G ratio is raised, to a higher value than that of MEA at 115 C (L/G of 4.12) with the CO 2 recovery of 85%. 4. The behaviour of component reboiler heat duty distributions for MEA, Blended Amines 1 and 2 It is found (section 3) that the reboiler heat duty trend demonstrates different patterns with regard to changes in L/G ratios and lean loadings for the three solvents. The three energy contributions to the reboiler heat duty (section 2.5) can be used to elucidate the differences in behaviour of reboiler heat duty. 9

11 Fig. 13 shows the three different components of heat duty for MEA, Blended Amines 1 and 2 as a function of L/G ratio at different stripper bottom temperatures and CO 2 recovery. The Q con for Blended Amine 1 at 115 C is higher than any other component energy. Q con for Blended Amine 1 at 115 C possibly slightly increases at L/G ratios higher than This trend is also noticed for the reboiler heat duty for MEA at 115 C and Blended Amine 1 at 112 C, suggesting that the heat of water vaporization for CO 2 stripping, which is equivalent to Q con, is a major factor in increasing the reboiler heat duty. Fig. 13 also shows that the Q con for Blended Amine 1 at all L/G ratios are higher than others when they are trialled at 115 C. MEA at 115 C and Blended Amine 1 at 112 C also reached minimum reboiler heat duty at L/G ratios around Thus the Q con for MEA at 115 C and Blended Amine 1 at 112 C has a similar pattern to their total reboiler heat duty. The Blended Amine 1 at 112 C has slightly lower Q con than MEA at 115 C, suggesting that relatively less heat of water vaporization is required for Blended Amine 1 to obtain similar lean loading to that of MEA at 115 C (cf. Table 7). This is also indicated in Fig. 14. The chart reveals that MEA at 115 C generated more steam compared to Blended Amine 1 at 112 C in order to achieve similar CO 2 lean loading (0.18~0.19). The greater amount of steam generated results in the raising of Q con. As the consequence of increasing the steam flows to the top of the stripper column, the temperature at the top packing in the stripper column also rises. Thus, at the same CO 2 -lean loading (0.18~0.19), MEA at 115 C gave higher Q con than Blended Amine 1 at 112 C as can be noted from Fig. 15. Both Q con for MEA at 112 C and Blended Amine 1 at 109 C are relatively constant at different L/G ratios. The magnitude of the Q con for both of them are also lower compared to others. Nevertheless, for Blended Amine 1 across the operating L/G ratio range up to 4.10, the proportions of condenser heat in the reboiler duty is generally less than those of the individual sensible heat and desorption heat (cf. Fig. 12). Yet, for the highest L/G ratio, the three component energies are similar. For Blended Amine 2 at 112 C and 115 C, in general they have similar pattern of Q con to their reboiler heat duty. The Q con for both temperatures are initially constant and then steadily increase at a higher L/G ratio. It is noticed that Q con at 112 C is lower than that of at 115 C, as expected. The Q con of Blended Amine 2 at 115 is comparable to the value of Q con from MEA at 115 C (at lower L/G ratio) and Blended Amine 1 at 112 C at lower and higher L/G ratios. The magnitude of Q con for Blended Amine 2 at 112 C is also similar within the limits of error compared to that of Blended Amine 1 at 112 C. For Blended Amine 2 at 112 C and 115 C, it is noticed that the sensible heat and the heat of desorption are nearly constant from the lowest L/G ratio up to an L/G ratio of 2.6. In general, the heat of water vaporization is much lower than the other components of reboiler heat duty. The Q con does steadily increase as the L/G ratio increases. For Blended Amine 2, the sensible heat also rises after L/G ratio 0.26, which is attributed to the increasing of solvent flow rate. The heat of desorption is apparently levells-off regardless of L/G ratio values. Blended Amines 1 and 2 are similar in that the magnitude of desorption heat is slightly higher than that of sensible heat. The Figure also shows that for Blended Amine 2, the three energy components make similar contributions to the reboiler heat duty. The sensible heat (Q sh ) for all solvents at different conditions is similar within the limits of error, except for Blended Amine 1 at 115 C, where the Q sh steadily decreases as L/G ratio increases and then starts to level-off above a L/G ratio of 3.0 (within the limits of error). This may be attributed to the temperature difference between the cold-rich solvent entering the stripper column and the hot-lean solvent leaving the stripper bottom not being constant. 10

12 Fig. 13 also shows that MEA at both temperatures has the highest heat of CO 2 desorption (Q des ) compared to the other solvents, suggesting that the Q des has contributed to increase reboiler heat duty for the MEA at both 112 C and 115 C. In this study, the magnitude of desorption heat decreases in the following order: MEA > Blended Amine 1 > Blended Amine 2. For Blended Amine 2, the Q des has more contribution than other component energy to reduce the reboiler heat duty. 5. Conclusion Performance tests of mono-ethanolamine (MEA) and Blended Amine solvents were carried out in CSIRO s transportable PCC pilot plant. The plant is fed with real flue gas from the brown coal-fired power station of Loy Yang Power in the Latrobe Valley, Victoria, Australia. Several campaigns have been devised to determine the solvent performance against the CO 2 absorption capacity (CO 2 recovery) and the requirement of reboiler duty in stripper column. The effect of increasing the stripper bottom temperature is to decrease the lean loading of the solvent. A reduction of lean solvent loading to below 0.20 for all solvents leads to an increase in the reboiler heat duty. This suggests that increasing the lean loading reduced the reboiler heat duty. A value of lean solvent loading above 0.20 for all solvents has less effect on the change of reboiler heat duty. In general, the magnitude of CO 2 recovery for all sets of solvents tested in the pilot plant is dependent on the L/G ratio and solvent lean loading. It is shown that a solvent with a higher lean loading requires a greater solvent (recycle) flow rate, or vice versa, in order to capture a given amount of CO 2. For Blended Amine 1, the operating temperature of stripper bottom at 112 C at L/G ratio between 4.0 to 5.0 is favourable because it can reach high CO 2 recovery at the low reboiler heat duty compared to MEA at 115 C. Even this reboiler heat duty is not higher than that when operating at 109 C. Because Blended Amine 2 has similar recovery for operating at 112 Cand 115 C, a lower temperature will result in reduction of the reboiler heat duty. Blended Amine 1 has higher CO 2 capacity and richer solvent loading than those of MEA and Blended Amine 2. The reboiler duty is also significantly dependent upon the L/G ratio and solvent lean loading. For all solvents at higher temperature, the contribution of Q cond is large compared to that of Q sh and Q des. For MEA and Blended Amine 1, the relationship between Q cond and L/G ratio shows a parabolic trend resembling to that of their total reboiler heat duty. MEA needs the highest reboiler duty followed by Blended Amine 1 and then Blended Amine 2. To achieve a similar CO 2 recovery of 84-85%, the Blended Amines 1 and 2 require 11% and 14% less reboiler duty than that of MEA, respectively ACKNOWLEDGEMENTS The authors gratefully acknowledge the PCC team from RITE and Chiyoda Corporation for their invaluable co-operation during campaign 4 Blended Amine 2, which has been developed by RITE within the COCS project supported by Ministry of Economy, Trade and Industry in Japan. Also thanks to Loy Yang Power for hosting the CSIRO s PCC pilot plant on their site and for having fruitful discussions. CSIRO s contribution in the LVPCC project was performed within CSIRO's Advanced Coal Technology and it is supported by the Victorian State Government through the Energy Technology Innovation Subsidy (ETIS) and Loy Yang Power Management Pty Ltd. 11

13 REFERENCES [1] Energy for Victoria: A Statement by the Minister for Energy and Resources final report submitted to Department of Natural Resources and Environment. 2002; quoted in Energy for Victoria: A Statement by the Minister for Energy and Resources [2] Bradshae J. GEODISC Project 1: Regional Analysis Final Report. APCRC Report, final report submitted to CO2CRC - CRC for Greenhouse gas Technologies. 2003; quoted in GEODISC Project 1: Regional Analysis Final Report. APCRC Report, [3] Kohl AL, Nielsen RB. Gas Purification, 5th. Houston, Texas. Gulf Publishing Company; 1997 [4] Aroonwilas A, Veawab A. Energy Procedia 2009; 1: [5] Idem R, Wilson M, Tontiwachwuthikul P, Chakma A, Veawab A, Aroonwilas A, Gelowitz D. Ind. Eng. Chem. Res. 2006; 45: [6] Chakma A, Mehrotra AK, Nielsen B. Heat Recovery Systems and CHP 1995; 15: 231. [7] Cottrell AJ, McGregor JM, Jansen J, Artanto Y, Dave N, Morgan S, Pearson P, Attalla MI, Wardhaugh L, Yu H, Allport A, Feron PHM. Energy Procedia 2009; 1: [8] Cifre PG, Brechtel K, Unterberger S, Scheffknecht G. "Experimental studies on CO 2 desorption from amine solutions" 2009, Clean Coal Technologies 2009 (CCT2009), Dresden, Germany. [9] Feron PHM. Energy Procedia 2009; 1: [10] Veawab A, Tontiwachwuthikul P, Aroonwilas A, Chakma A, Gale J, Kaya Y. Performance and Cost Analysis for CO 2 Capture from Flue Gas Streams: Absorption and Regeneration Aspects. Greenhouse Gas Control Technologies - 6th International Conference. Oxford: Pergamon; 2003, p.127. [11] Davison J. Energy 2007; 32: [12] Meuleman EEB, Artanto Y, Jansen J, Osborn M, Pearson P, Cottrell AJ, Feron PHM. In: Proceedings of the 35th International Technical Conference on Clean Coal and Fuel Systems. Sheraton Sand Key Clearwater, Florida, USA: June , [13] Xiao J, Li C-W, Li M-H. Chemical Engineering Science 2000; 55: 161. [14] Dey A, Aroonwilas A. Energy Procedia 2009; 1: 211. [15] Goto K, Okabe H, Shimizu S, Onoda M, Fujioka Y. Energy Procedia 2009; 1: [16] Mangalapally HP, Notz R, Hoch S, Asprion N, Sieder G, Garcia H, Hasse H. Energy Procedia 2009; 1: 963. [17] Kvamsdal HM, Jakobsen JP, Hoff KA. Chemical Engineering and Processing: Process Intensification 2009; 48: 135. [18] Puxty G, Rowland R. Environmental Science & Technology 2011; 45: [19] Sakwattanapong R, Aroonwilas A, Veawab A. Ind. Eng. Chem. Res. 2005; 44: [20] Dugas RE. Pilot Plant Study of Carbon Dioxide Capture by Aqueous Monoethanolamine, M.S.E, The University of Texas at Austin,

14 Fig. 1. Typical Flow Diagram of PCC Pilot Plant at Loy Yang Power 13

15 100% 90% 115 o C CO 2 recovery 80% 70% 60% 112 o C 50% 40% L/G (L/Nm 3 ) Fig. 2. Correlation between L/G and CO 2 recovery for two different stripper s bottom temperatures 100% 90% 115 o C CO 2 recovery 80% 70% 60% 112 o C 50% 40% Lean loading (mol CO 2 /mol MEA) L/G = L/G = L/G = Fig. 3. Correlation between lean loading MEA solvent at similar L/G on CO 2 recovery 14

16 6.5 Heat duty (MJ/kg CO 2 ) o C 112 o C L/G (L/Nm 3 ) Fig. 4. Correlation between L/G and reboiler heat duty for two different stripper bottom temperatures Heat duty (MJ/kg CO 2 ) L/G (L/Nm 3 ) Lean loading (mol CO 2 /mol Amine) Duty C Duty C Lean loading C Lean loading C Fig. 5. The effect of L/G ratio and solvent-lean loading on reboiler heat duty 15

17 Component of heat duty (MJ/kg CO 2 ) Component of heat duty (MJ/kg CO 2 ) 6.0 A B L/G (L/Nm 3 ) Total heat duty Q condenser Q solvent heat Q desorption Fig. 6. Reboiler duty distribution as a function of L/G ratio for MEA baseline (A: 115 C, B: 112 C) 16