- Email: [email protected]
ELPI+ measurements of aerosol growth in an amine absorption column
International Journal of Greenhouse Gas Control 23 (2014) 44–50
Contents lists available at ScienceDirect
International Journal of Greenhouse Gas Control journal homepage: www.elsevier.com/locate/ijggc
ELPI+ measurements of aerosol growth in an amine absorption column Jan Mertens a,∗ , L. Brachert b , D. Desagher a , M.L. Thielens a , P. Khakharia c , E. Goetheer c , K. Schaber b a b c
Laborelec, Rodestraat 125, 1630 Linkebeek, Belgium Karlsruhe Institute of Technology (KIT), Institut für Technische Thermodynamik und Kältetechnik, Engler-Bunte-Ring 21, 76131 Karlsruhe, Germany TNO, Leeghwaterstraat 46, Delft 2628CA, The Netherlands
a r t i c l e
i n f o
Article history: Received 1 October 2013 Received in revised form 30 January 2014 Accepted 1 February 2014 Available online 6 March 2014 Keywords: Monoethanolamine ELPI+ Emission measurement Aerosols Mist Post combustion carbon capture
a b s t r a c t Recently, studies have appeared pointing out that aerosols can dominate the total amine emission from amine based PCCC pilot plant scale installations. For the design of countermeasure types (upstream or downstream of the PCCC installation), it is crucial to have an idea of the aerosol size distribution and numbers entering or leaving the absorber. This study is the first to present this kind of data and should serve future installations when designing aerosol emission countermeasures. H2 SO4 aerosols entering the absorber are observed to be extremely small (i.e. <0.2 m) with number concentrations exceeding 1E8 cm−3 . The aerosols grow in size as they travel through the absorber through the taking up of water and amine to sizes close to but staying below 1 m. However, despite the fact that most of the aerosols (expressed in number concentrations) are well below 1 m, most of the water (and thus amine) is found in the aerosol sizes between 0.5 and 2 m. Therefore, if one aims at designing efficient countermeasures, eliminating this size fraction is crucial. This amine emission stream is therefore very difficult to remove using water washes as aerosols are known to travel through water wash sections. Moreover, also classical demisters show very little efficiency for these small aerosol sizes and are therefore believed not to be suitable for the removal of aerosols. This information will therefore serve future installations when designing aerosol emission countermeasures. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Post combustion carbon capture (PCCC) is based on the removal of CO2 after the combustion of a fossil fuel. Reactive absorption is one of the most important techniques for the removal of CO2 from flue gas. This reactive absorption process makes use of the reversible nature of the chemical reaction of an aqueous alkaline solvent (usually an amine) with an acid gas (CO2 ). The amine based solvent can be lost during the process due to: degradation, vaporization, mechanical losses and aerosol (mist) formation. Studies on emission processes such as vaporization and mist formation exist at laboratory conditions (Chi and Rochelle, 2002; Nguyen et al., 2010; Freeman et al., 2010). Only recently, studies have appeared pointing out that aerosols can dominate the total
Abbreviations: PCCC, post combustion carbon capture; ELPI+ , electrostatic low pressure impactor; MEA, monoethanolamine; PM, particulate matter; FTIR, Fourier transformed infra red; ELV, emission limit value. ∗ Corresponding author at: Sustainable Process Technology, Laborelec, Rodestraat 125, 1630 Linkebeek, Belgium. Tel.: +32 2 382 05 71; fax: +32 2 382 02 41. E-mail address: [email protected] (J. Mertens). http://dx.doi.org/10.1016/j.ijggc.2014.02.002 1750-5836/© 2014 Elsevier Ltd. All rights reserved.
amine emission at pilot plant scale (Kamijo, 2011; Kolderup et al., 2012; Knudsen and Bade, 2012; Carter, 2012; Mertens et al., 2012, 2013; Khakharia et al., 2013). Mertens et al. (2013) discussed the origin and driving factors of the NH3 and MEA emissions including mist formation phenomena as measured at different pilot plants using a Fourier transformed infra red (FTIR) analyser. Mist precursors can be ultrafine liquid or solid particles of sulfuric acid, salts or any form of particulate matter in the flue gas entering the absorber from the coal fired power plant (Kamijo, 2011; Khakharia et al., 2013). Kamijo (2011) describes that submicron (<1 m) particulate matter or H2 SO4 aerosols may serve as nuclei for the formation of amine aerosols. Khakharia et al. (2013) show that particles in the form of soot (106 number of particles per cm3 ) can cause MEA emissions in the order of 200 mg Nm−3 . H2 SO4 aerosols with a particle number concentration in the order of 108 per cm3 can lead to MEA emissions in the range of 600–1100 mg Nm−3 . Schaber (1995) and Kolderup et al. (2011) describe how in gas–liquid contact devices like absorbers, aerosols can be formed. Two nucleation mechanisms are described: (i) homogeneous nucleation: when nuclei are formed only by molecules of condensable components and
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
(ii) heterogeneous nucleation: when fine particles are present. Heterogeneous nucleation is the dominant mechanism of aerosol formation in most industrial gas cleaning processes, because there is a large number of nanoparticles present in these flue gases (Schaber, 1995). The aerosol droplets grow by condensation until the supersaturation of the gas phase disappears. Only recently, this phenomenon has been described in amine based CO2 absorbers. Whether and to what extent mist is formed depends to a large extent on the flue gas composition upstream the CO2 capture plant and on the pilot plant’s design and operating conditions. Knudsen and Bade (2012) similarly describe that mist formation may be a significant mechanism of amine emission and that amine mist, once formed, may be difficult to remove with conventional washing stages and demisters. If formed, aerosols may pass through water wash and demisters because of their assumed small particle size. Two types of countermeasures can be thought of: (i) removal of the aerosol precursors (SO3 , fine PM, etc.) in the flue gas upstream the absorber or (ii) removal of the MEA containing aerosols leaving the absorber. Future full scale amine scrubber installations will be imposed emission limit values (ELV) for a number of components including NH3 and the amine itself. Most likely these ELV will be expressed as maximum concentrations tolerated in the CO2 poor flue gas leaving the stack. The exact value of the ELV will depend on the local authorities but it is clear that the high amine concentrations measured during mist formation phenomena will not be tolerated. Therefore, countermeasures need to be implemented. For the design of both countermeasures types (upstream or downstream), it is crucial to have an idea of the aerosol size distribution and numbers entering or leaving the absorber. However, until present, no relation between measurements of the aerosol sizes and numbers entering and leaving the absorber has been published. This study is the first to present this kind of data and should serve future installations when designing aerosol emission countermeasures.
2. Materials and methods
45
Fig. 1. Pilot plant layout at the Karlsruhe Institute of Technology. SO2 is converted into SO3 inside a micro-reactor over a Pt based catalyst. This SO3 stream is added to the flue gas of a natural gas combustion transforming it into H2 SO4 aerosols inside a quench column. Part of this H2 SO4 aerosol containing flue gas is consequently send into TNO’s mini carbon capture plant.
which is several orders of magnitude lower than the sulfuric acid aerosol concentration (Brachert et al., 2013). The pilot plant has been used to simulate a FGD and understand H2 SO4 aerosol formation. As indicated in Fig. 1, at the outlet of the quench, a split stream is directed to TNO’s CO2 mini-capture plant described below. A T-connection enables the simultaneous measurement of the aerosol number concentration and size distribution by means of a condensation particle counter (UFCPC, PALAS) and the ELPI+ analyser (DEKATI).
2.1. Pilot plant installation: H2 SO4 aerosol generation The experiments are performed at the pilot plant (Fig. 1) at the Institut für Technische Thermodynamik und Kältetechnik (ITTK) in Karlsruhe Institute of Technology (KIT) where a flue gas of ∼180 m3 (STP)/h is generated by burning natural gas (Brachert et al., 2013). This pilot plant has been extensively described in the literature before (Gretscher and Schaber, 1999). The constant dosing of sulfur trioxide (SO3 ) is done via a microreactor where SO2 is oxidized to SO3 at 500 ◦ C. The catalyst of the microreactor is a Pt-catalyst on TiO2 on microstructured titanium foils (Pfeifer et al., 2011). In the flue gas the sulfur trioxide combines with the water vapour from the combustion to sulfuric acid. FTIR has been used to measure the conversion factor from SO2 to SO3 in the reactor by gas composition measurements directly behind the inlet of the flow from the microreactor into the flue gas stream. The possibility of inhomogeneity in the flue gas stream and the low absolute concentrations of SO2 makes these measurements challenging. The measurements give ∼20% conversion factor at 50 sccm (standard cubic centimeters per minute) SO2 which corresponds to ≈15 mg of sulfuric acid per standard cubic meter in the flue gas. The aerosol formation takes place in a quench cooler (packed column absorber). In this absorber, the hot flue gas is cooled down fast from ∼200 ◦ C to the adiabatic saturation temperature of approximately 46 ◦ C by circulating water. The high supersaturation leads to homogeneous nucleation of sulfuric acid and water (Wix et al., 2010). The soot concentration in the pilot plant, which is a possible source for heterogeneous nucleation, is in the order of magnitude of 104 cm−3
2.2. Pilot plant installation: mini-capture plant TNO’s mobile carbon capture mini-plant is a conventional absorption–desorption system with a capacity of 4 m3 h−1 of flue gas. The absorber section is a packed bed (Sulzer Mellapak 2X) with a total height of 3.5 m composed of 4 sections each of 0.5 m height. The absorber section neither has a water wash section nor a demister fitted to the top. The stripper section has the same dimension as the absorber. The total liquid inventory of the system is about 20 l. Additional CO2 was added to the flue gas to ensure a CO2 concentration of about 12 vol% to mimic coal fired power plant flue gas conditions. More details on this plant and its functioning can be found in Khakharia et al. (2013). 2.3. FTIR A FTIR analyser (GASMET CX 4000) was used to analyze the gas phase. The gas leaving the absorber column is heated to 180 ◦ C using a trace heated transfer line. At this temperature the aerosol phase is vaporized implying that the concentration of MEA measured by FTIR is the sum of vapour phase MEA and MEA that was present in the aerosols (Mertens et al., 2012, 2013). The FTIR analyzer was specifically calibrated for MEA. The uncertainty of the components measured by the FTIR depends on the chemical composition (and thus possible interferences) of the matrix in which it is measured. Therefore, stating detection and quantification limits is
46
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
difficult; however they are in the low mg Nm−3 range at the applied conditions (for more information, Mertens et al., 2012, 2013). 2.4. Electrical low pressure impactor (ELPI+ ) The main measurement device used in this work is the electrical low pressure impactor ELPI+ . The particles are charged by corona charging and subsequently separated in a low pressure cascade impactor with 14 electrically insulated collection stages. The measured current signals are proportional to the number concentration and size. By using kernel functions in order to account for the charging efficiency dependency on diameter and for the collection efficiencies of the different stages, the number concentration in every channel can be calculated. A more precise description of the original ELPI can be found in Keskinen et al. (1992) and Marjamäki et al. (2000). The ELPI+ features an additional impactor stage which enlarges the measurement range covered by impactor stages from a cut-off of 30 nm down to a cut-off of 16.7 nm (Yli-Ojanperä et al., 2010). Additionally, a filter stage has been added which collects all the particles that are not trapped in one of the impactor stages. The charging efficiency decreases strongly with the particle size, so that below 6 nm no charging is expected anymore. Thereby, particles down to 6 nm can be measured. The maximum number concentrations that are detectable in every stage as indicated in the manual are not reached during the experiments. Nevertheless, at least a dilution of 10 is used in this work in order to avoid condensation on the impactor plates since the aerosol stream sampled from the pilot plant is saturated with water. The dilution cascade DC 10000 (PALAS GmbH) has been used for this purpose. With this device dilution factors of 10, 102 , 103 and 104 are possible. A last important remark concerns the non-isokinetic sampling of the aerosols and the effect this may have on the observed aerosol sizes. This effect is only important for aerosols that are larger than a few m (Angelo, 2008; Mertens et al., 2013). Mertens et al. (2013) showed the absence of the effect of iso-kinetic sampling on the measured MEA concentrations using the FTIR. This was explained by the fact that the aerosols should be smaller than 1 m which is confirmed in this study as shown in detail below. Therefore, it was judged not necessary to iso-kinetically sample the aerosols for both the FTIR and ELPI+ measurements presented in this study.
Fig. 2. The measured aerosol sizes as measured by Brachert et al. (2013) in front of the absorber are larger than in the current layout because the aerosols shrink as they pass through a fan before entering the absorber.
(ii) An overestimation of the number concentrations measured by the filter stage (i.e. smaller than 10 nm) With respect to the dilution effect, Brachert et al. (2014) show that the measured size distributions for low H2 SO4 concentrations (up to 7.6 sccm) were not affected by the dilution since they are already close to their dry droplet diameter. This term conveys that water cannot be evaporated from these aerosols in the dilution step as the high sulfuric acid concentration in the droplet results in a negligible low vapour pressure. A shrinking of the aerosols at 50 sccm was however noted because of their much larger water content (i.e. increasing H2 SO4 concentration only increases the size and not the number of aerosols) and thus potential for water evaporation and shrinking.
3. Results and discussion Brachert et al. (2014) used the same H2 SO4 generation pilot plant set-up for their comparison of two different measurement principles (Condensation Particle Counter; PALAS, UFCPC with sensor 200 and ELPI+ ) for the measurement of the number concentration of volatile sulfuric acid aerosols. They conclude that both devices confirm the high number concentrations of >108 cm−3 which also corresponds to the simulations carried out by Wix et al. (2010). With the ELPI+ , we were for the first time able to obtain information on sizes by measurement. The prediction from simulations was confirmed: i.e. the number concentrations are always in the same order of magnitude upon increase of the sulfuric acid concentration. However, higher sulfuric acid concentrations lead to an increase in droplet sizes. The ELPI+ size distribution confirmed the simulation results with additional coagulation calculations and revealed droplet sizes well below 100–200 nm which makes them very difficult to precipitate. Brachert et al. (2014) learns that ELPI+ measurements provide a good estimation of the true size distribution but two important things must be taken into account when interpreting the measurements: (i) The effect that the dilution may have on the shrinking of aerosols that contain a lot of water
3.1. Number concentration and size of H2 SO4 aerosols entering the absorber and the effect of dilution A main and important difference in the set-up between the Brachert et al. (2014) set-up and the set-up used in this study is the fact that in the current set-up a fan is used to blow the flue gas into the mini-capture plant. In contrast, in Brachert et al. (2014), a vacuum pump was installed behind the mini-absorber locally creating a strong vacuum. This strong vacuum, however, makes ELPI+ measurements impossible behind the absorber which is one of the two locations of interest in this study. This sampling set-up including the fan, leads to a significant decrease in size as a result of the passing through the fan as clearly visible from Fig. 2. This size decrease is associated to a number increase by a factor 3, from 2.1 E8 to 6.5 E8 cm−3 . This increase is most likely mainly due to the overestimation of the number concentrations when most of the particles are found in the filter stage (Brachert et al., 2014). As shown in Fig. 3, shrinking is no longer observed behind the fan since inside the fan, a similar shrinking process takes place as in the dilution step. This process of water evaporation reduces the aerosols in sizes close to their dry droplet diameter. Similarly to Brachert et al. (2014), we conclude that H2 SO4 aerosol sizes entering the absorber are extremely small and well below 200 nm. It
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
47
Fig. 3. Applying different dilution ratios has no effect on the measured aerosol size distributions in front of the absorber (behind the fan).
is important to stress that at this location, the aerosols are only composed of H2 SO4 and no amine is yet inside. 3.2. Number concentration and size of aerosols leaving the absorber and the effect of dilution Fig. 4 compares the measured aerosol size distributions into the absorber to the measured size distribution leaving the absorber. It is clear that the aerosols grow inside the absorber and their number is reduced by a factor of almost 5 from 6.5 E8 to 1.3 E8 cm−3 . The reduction in number concentration can be explained on the one hand by the reduced overestimation in the ELPI+ due to larger droplets and on the other hand by coagulation in the column.
The size increase inside the absorber is attributed to the uptake of water and MEA as the aerosols travel through the absorber. This water uptake is confirmed by Fig. 5 which shows that behind the absorber, there is now a clear dilution effect, in contrast to Fig. 3 in front of the absorber where no distinct dilution effect has been observed. This indicates that there is a significant amount of water surrounding the aerosols enabling the shrinking effect observed when diluting with dry clear air. Fig. 5 suggests that at a dilution factor of 100, we are close to the dry droplet diameter since dilution 1000 times hardly effects the measured size distribution. Of course the question remains what the true size distribution would be without dilution? This question is treated below where an evaluation of the measured number and sizes of the aerosols is correlated to the
Fig. 4. Aerosols grow as they move through the absorber.
48
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
Fig. 5. Applying different dilution ratios has a clear effect on the measured aerosol size distributions behind the absorber because water is evaporated in the dilution step.
measured MEA emissions from the absorber as measured by the FTIR. 3.3. MEA emissions from the absorber Fig. 6 presents the emissions from the top of the absorber of the time period during which ELPI+ measurements were made behind the absorber. First of all, it is important to repeat that neither water wash nor demister is installed at the top of the mini-absorber and thus no emission control whatsoever. NH3 emissions are observed to be at around 20 mg Nm−3 , oxygen remains at around 10 vol% and CO2 is around 1.5 vol% which corresponds to close to 90% capture efficiency since the CO2 concentration send into the absorber is artificially kept constant at 12 vol%. Between 3000 and 3500 mg Nm−3 of MEA is leaving the absorber and a rising trend in time is observed
Fig. 6. NH3 , MEA, CO2 and O2 concentrations as measured by the FTIR system in the flue gas leaving the absorber during the period in which the ELPI+ measurements were carried out behind the absorber.
most likely due to some instability (e.g. temperature variations) in the mini-pilot plant. It is important to mention that the FTIR is not calibrated to measure these high numbers of MEA. Earlier research has proven that the FTIR overestimates the true value at high concentrations by at least 20%. Therefore, a value of 3000 mg Nm−3 is used for the evaluation presented below although the FTIR readings suggest values up to 3500 mg Nm−3 . It is worth mentioning that MEA emissions from the top of the absorber in the absence of H2 SO4 aerosols are observed around 45 mg Nm−3 as presented in Khakharia et al. (2013). 3.4. Theoretical evaluation of the measured aerosol sizes and numbers and corresponding MEA emissions In this paragraph, a theoretical calculation is carried out to evaluate whether the measured aerosol numbers and corresponding size distribution can contain the high MEA concentration of around 3000 mg Nm−3 as measured by the FTIR. First assumption made here is that the MEA leaving the absorber is all in the form of aerosols. This is not entirely true since Nguyen et al. (2010) report a volatility of 147 mg Nm−3 for 30 wt% aqueous MEA at 40 ◦ C and a typical lean loading of 0.25 mol CO2 mol−1 amine. Temperatures at the top of the absorber in this study were around 30 ◦ C which would reduce the volatile MEA contribution to even lower values so the assumption that all MEA is aerosol MEA is justified. However, one very important assumption to be made here is the MEA concentration inside the aerosols. Condensate measurement of flue gas leaving the absorber carried out on the Maasvlakte pilot plant in 2012, a MEA concentration of 15 wt% was found (Kolderup et al., 2012). The question remains whether this number is valid for the tests carried out here, but since we have no better, it is used in the calculation. This implies that the flue gas needs to contain 20 ml of water per Nm−3 to contain 3000 mg Nm−3 of MEA. Assuming that the aerosols are made up of 100% water, we can calculate the ml of water the aerosols contain based on their size and numbers. Fig. 7 presents this simple calculation by plotting the ml of water per m3 of gas that the aerosols can contain as a function of their assumed size for different aerosol number concentrations. With the help of this plot, the needed droplet sizes
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
49
Comparing the 2.5 ml Nm−3 based on ELPI+ measurements with the required 20 ml Nm−3 based on the calculation reveals a discrepancy which can be explained by: 1. Going from a dilution factor of 10 to no dilution would reveal larger aerosol sizes; similarly as we observed going from a factor 100 to a factor of 10. Since going from 100 times dilution to 10 times dilution increased the water content by a factor of 10, a factor of 10 discrepancy in water content between no dilution and 10 times dilution seems plausible. 2. There is a large uncertainty associated to the 15 wt% assumed in the study. If the concentration would be higher, then again less than 20 ml of water per Nm3 of flue gas would be required to contain 3000 mg Nm−3 of MEA.
Fig. 7. Theoretical relation between aerosol size and water content as a function of the number concentration of the aerosols.
for a certain required water content in the aerosol can be identified. To come to the 20 ml assumed above, the figure reveals that we would need an aerosol number concentration between 1E7 and 1E8 cm−3 assuming that all aerosols have a diameter of 1 m. Fig. 8 presents the calculation of the aerosol water content (plotted as cumulative distribution) for each of the size fractions based on the ELPI+ measurements of the flue gas behind the absorber both at dilution factors 10 and 100 as presented in Fig. 5. It is clear that most aerosols are smaller than 0.1 m and contain hardly any water due to their small size. As also discussed above, the dilution has an important effect on the measured sizes and even more so on the total water content of the aerosols. In case of a 100 times dilution, a total water content of only 0.25 ml Nm−3 is calculated whilst for a ten times dilution 2.5 ml Nm−3 is obtained.
We conclude that the measured number concentrations and corresponding size distribution shown as measured by the ELPI+ behind the absorber and presented above in Fig. 4 are close to reality. It is true that it is likely that in reality, the aerosol size distribution will shift a little to the right (i.e. to larger sizes) although only a slight change is sufficient to match the calculations performed here. We therefore conclude that most of the aerosols are significantly smaller than 1 m and thus very hard to eliminate from the flue gas leaving the absorber using traditional demisters which show very low removal efficiency of submicron particles. However, despite the fact that most of the aerosols (expressed in number concentrations) are well below 1 m, most of the water (and thus amine) is found in the aerosol sizes between 0.5 and 2 m. Therefore, if one aims at designing efficient countermeasures, eliminating this size fraction is crucial. Eliminating the aerosols smaller than 500 nm will likely no longer reduce the amine concentrations leaving the absorber to a large extend. 4. Conclusions The study is the first to present what happens to H2 SO4 aerosol sizes and numbers as they travel through an amine PCCC absorber. H2 SO4 aerosols entering the absorber are observed to be extremely small (i.e. <0.1 m) with number concentrations exceeding 1E8 cm−3 . The aerosols grow as they travel through the absorber through the taking up of water and amine to sizes close to but staying below 1 m. Their number concentration is observed to be slightly reduced most probably due to coagulation processes taking place inside the absorber. However, despite the fact that most of the aerosols (expressed in number concentrations) are well below 1 m, most of the water (and thus amine) is found in the aerosol sizes between 0.5 and 2 m. Therefore, if one aims at designing efficient countermeasures, eliminating this size fraction is crucial. Eliminating the aerosols smaller than 500 nm will likely no longer reduce the amine concentrations leaving the absorber to a large extend. The main conclusion of this study is the fact that aerosols entering and leaving the absorber are small (<2 m) although they can contain lots of amine. This amine emission stream is therefore very difficult to remove using water washes as aerosols are known to travel through water wash sections. Moreover, also classical demisters show very little efficiency for these small aerosol sizes and are therefore believed not to be suitable for the removal of aerosols. This information will therefore serve future installations when designing aerosol emission countermeasures. Acknowledgements
Fig. 8. Measured aerosol size distributions behind the absorber for dilution ratios of 10 and 100 and their corresponding calculated water content (presented as a cumulative distribution).
This research was carried out within the framework of the Dutch CATO-2 program. We are very grateful to TNO and the University
50
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
of Karlsruhe for the collaboration during this experimental work. This research was feasible thanks to the GDF SUEZ CCS research program which partly funded this research. References Angelo, M., 2008. Automatic Particle Monitoring Technology. Presentation for the SICK Company. Brachert, L., Kochenburger, T., Schaber, K., 2013. Facing the sulfuric acid aerosol problem in flue gas cleaning: pilot plant experiments and simulation. Aerosol Science and Technology 10, 1083–1091. Brachert, L., Mertens, J., Khakharia, P., Schaber, K., 2014. The challenge of measuring sulfuric acid aerosols: number concentration and size evaluation using a condensation particle counter (CPC) and an electrical low pressure impactor (ELPI+ ). Journal of Aerosol Science 61, 21–27. Carter, 2012. National Carbon Capture Center: Post-Combustion. In: 2012 NETL CO2 Capture Technology Meeting, July 10 http://www.netl.doe.gov/publications/ proceedings/12/co2capture/presentations/2-Tuesday/T%20Carter-NCCC-Postcombustion.pdf Chi, S., Rochelle, G.T., 2002. Oxidative degradation of monoethanolamine. Industrial & Engineering Chemistry Research 41, 4178–4186. Freeman, S.A., Davis, J., Rochelle, G.T., 2010. Degradation of aqueous piperazine in carbon dioxide capture. International Journal of Greenhouse Gas Control 5, 756–761. Gretscher, H., Schaber, K., 1999. Aerosol formation by heterogeneous nucleation in wet scrubbing processes. Chemical Engineering and Processing 38, 541–548. Kamijo, T., 2011. Amine Emission Control Technology of KM CDR ProcessTM . In: Presentation at the EPRI meeting: amines for post combustion capture, 16 August. Keskinen, J.K., Pietarinen, K., Lehtimäki, M., 1992. Electrical low pressure impactor. Journal of Aerosol Science 23, 353–360. Khakharia, P., Brachert, L., Mertens, J., Huizinga, A., Schallert, B., Schaber, K., Vlugt, T., Goetheer, E., 2013. Investigation of aerosol based emission of MEA due to sulphuric acid aerosol and soot in a post combustion CO2 capture process. International Journal of Greenhouse Gas Control 19, 138–144.
Knudsen, J., Bade, O., 2012. Emission reduction technologies. In: CLIMIT Workshop, 5–6 December 2011, Oslo, Norway, Available at: www.akercleancarbon.com Kolderup, H., da Silva, E., Mejdell, T., Tobiesen, A., Haugen, G., Hoff, K.A., Josefsen, K., Strøm, T., Furuseth, O., Hanssen, K.F., Wirsching, H., Myhrvold, T., Johnsen, K., 2011. Emission Reducing Technologies. Report H&ETQP Amine 6 for Gassnova NF, Norway. http://www.gassnova.no/gassnova2/frontend/files/CONTENT/ Rapporter/EmissionquantificationandreductionWP1and3 Sintef.pdf Kolderup, H., Hjarbo, K., Huizinga, A., Tuinman, I., Zahlsen, K., Vernstad, K., Hyldbakk, A., Holten, T., Kvamsdal, H., van Os, P., da Silva, E., Goetheer, E., Khakharia, P., 2012. Emission Quantification and Reduction (No. WP 1 and 3 in the Project: CCM TQP Amine 6 for Gassnova SF, Norway). Marjamäki, M., Keskinen, J., Chen, D., Pui, D.Y.H., 2000. Performance evaluation of the electrical low-pressure impactor (ELPI). Journal of Aerosol Science 31, 249–261. Mertens, J., Lepaumier, H., Desagher, D., Thielens, M.L., 2013. Understanding ethanolamine (MEA) and ammonia emissions from amine based post combustion carbon capture: lessons learned from field tests. International Journal of Greenhouse Gas Control 13, 72–77. Mertens, J., Thielens, M.L., Knudsen, J., Andersen, J., 2012. On-line monitoring and controlling emissions in amine post combustion carbon capture: a field test. International Journal of Greenhouse Gas Control 6, 2–11. Nguyen, T., Hilliard, M., Rochelle, G.T., 2010. Amine volatility in CO2 capture. International Journal of Greenhouse Gas Control 4, 707–715. Pfeifer, P., Zscherpe, T., Haas-Santo, K., Dittmeyer, R., 2011. Investigations on a Pt/TiO2 catalyst coating for oxidation of SO2 in a microstructured reactor for operation with forced decreasing temperature profile. Applied Catalysis A 391, 289–296. Schaber, K., 1995. Aerosol formation in absorption processes. Chemical Engineering Science 50, 1347–1360. Wix, A., Brachert, L., Sinanis, S., Schaber, K., 2010. A simulation tool for aerosol formation during sulphuric acid absorption in a gas cleaning process. Journal of Aerosol Science 41, 1066–1079. Yli-Ojanperä, J., Kannosto, J., Marjamäki, M., Keskinen, J., 2010. Improving the nanoparticle resolution of the ELPI. Aerosol and Air Quality Research 10, 360–366.
Contents lists available at ScienceDirect
International Journal of Greenhouse Gas Control journal homepage: www.elsevier.com/locate/ijggc
ELPI+ measurements of aerosol growth in an amine absorption column Jan Mertens a,∗ , L. Brachert b , D. Desagher a , M.L. Thielens a , P. Khakharia c , E. Goetheer c , K. Schaber b a b c
Laborelec, Rodestraat 125, 1630 Linkebeek, Belgium Karlsruhe Institute of Technology (KIT), Institut für Technische Thermodynamik und Kältetechnik, Engler-Bunte-Ring 21, 76131 Karlsruhe, Germany TNO, Leeghwaterstraat 46, Delft 2628CA, The Netherlands
a r t i c l e
i n f o
Article history: Received 1 October 2013 Received in revised form 30 January 2014 Accepted 1 February 2014 Available online 6 March 2014 Keywords: Monoethanolamine ELPI+ Emission measurement Aerosols Mist Post combustion carbon capture
a b s t r a c t Recently, studies have appeared pointing out that aerosols can dominate the total amine emission from amine based PCCC pilot plant scale installations. For the design of countermeasure types (upstream or downstream of the PCCC installation), it is crucial to have an idea of the aerosol size distribution and numbers entering or leaving the absorber. This study is the first to present this kind of data and should serve future installations when designing aerosol emission countermeasures. H2 SO4 aerosols entering the absorber are observed to be extremely small (i.e. <0.2 m) with number concentrations exceeding 1E8 cm−3 . The aerosols grow in size as they travel through the absorber through the taking up of water and amine to sizes close to but staying below 1 m. However, despite the fact that most of the aerosols (expressed in number concentrations) are well below 1 m, most of the water (and thus amine) is found in the aerosol sizes between 0.5 and 2 m. Therefore, if one aims at designing efficient countermeasures, eliminating this size fraction is crucial. This amine emission stream is therefore very difficult to remove using water washes as aerosols are known to travel through water wash sections. Moreover, also classical demisters show very little efficiency for these small aerosol sizes and are therefore believed not to be suitable for the removal of aerosols. This information will therefore serve future installations when designing aerosol emission countermeasures. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction Post combustion carbon capture (PCCC) is based on the removal of CO2 after the combustion of a fossil fuel. Reactive absorption is one of the most important techniques for the removal of CO2 from flue gas. This reactive absorption process makes use of the reversible nature of the chemical reaction of an aqueous alkaline solvent (usually an amine) with an acid gas (CO2 ). The amine based solvent can be lost during the process due to: degradation, vaporization, mechanical losses and aerosol (mist) formation. Studies on emission processes such as vaporization and mist formation exist at laboratory conditions (Chi and Rochelle, 2002; Nguyen et al., 2010; Freeman et al., 2010). Only recently, studies have appeared pointing out that aerosols can dominate the total
Abbreviations: PCCC, post combustion carbon capture; ELPI+ , electrostatic low pressure impactor; MEA, monoethanolamine; PM, particulate matter; FTIR, Fourier transformed infra red; ELV, emission limit value. ∗ Corresponding author at: Sustainable Process Technology, Laborelec, Rodestraat 125, 1630 Linkebeek, Belgium. Tel.: +32 2 382 05 71; fax: +32 2 382 02 41. E-mail address: [email protected] (J. Mertens). http://dx.doi.org/10.1016/j.ijggc.2014.02.002 1750-5836/© 2014 Elsevier Ltd. All rights reserved.
amine emission at pilot plant scale (Kamijo, 2011; Kolderup et al., 2012; Knudsen and Bade, 2012; Carter, 2012; Mertens et al., 2012, 2013; Khakharia et al., 2013). Mertens et al. (2013) discussed the origin and driving factors of the NH3 and MEA emissions including mist formation phenomena as measured at different pilot plants using a Fourier transformed infra red (FTIR) analyser. Mist precursors can be ultrafine liquid or solid particles of sulfuric acid, salts or any form of particulate matter in the flue gas entering the absorber from the coal fired power plant (Kamijo, 2011; Khakharia et al., 2013). Kamijo (2011) describes that submicron (<1 m) particulate matter or H2 SO4 aerosols may serve as nuclei for the formation of amine aerosols. Khakharia et al. (2013) show that particles in the form of soot (106 number of particles per cm3 ) can cause MEA emissions in the order of 200 mg Nm−3 . H2 SO4 aerosols with a particle number concentration in the order of 108 per cm3 can lead to MEA emissions in the range of 600–1100 mg Nm−3 . Schaber (1995) and Kolderup et al. (2011) describe how in gas–liquid contact devices like absorbers, aerosols can be formed. Two nucleation mechanisms are described: (i) homogeneous nucleation: when nuclei are formed only by molecules of condensable components and
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
(ii) heterogeneous nucleation: when fine particles are present. Heterogeneous nucleation is the dominant mechanism of aerosol formation in most industrial gas cleaning processes, because there is a large number of nanoparticles present in these flue gases (Schaber, 1995). The aerosol droplets grow by condensation until the supersaturation of the gas phase disappears. Only recently, this phenomenon has been described in amine based CO2 absorbers. Whether and to what extent mist is formed depends to a large extent on the flue gas composition upstream the CO2 capture plant and on the pilot plant’s design and operating conditions. Knudsen and Bade (2012) similarly describe that mist formation may be a significant mechanism of amine emission and that amine mist, once formed, may be difficult to remove with conventional washing stages and demisters. If formed, aerosols may pass through water wash and demisters because of their assumed small particle size. Two types of countermeasures can be thought of: (i) removal of the aerosol precursors (SO3 , fine PM, etc.) in the flue gas upstream the absorber or (ii) removal of the MEA containing aerosols leaving the absorber. Future full scale amine scrubber installations will be imposed emission limit values (ELV) for a number of components including NH3 and the amine itself. Most likely these ELV will be expressed as maximum concentrations tolerated in the CO2 poor flue gas leaving the stack. The exact value of the ELV will depend on the local authorities but it is clear that the high amine concentrations measured during mist formation phenomena will not be tolerated. Therefore, countermeasures need to be implemented. For the design of both countermeasures types (upstream or downstream), it is crucial to have an idea of the aerosol size distribution and numbers entering or leaving the absorber. However, until present, no relation between measurements of the aerosol sizes and numbers entering and leaving the absorber has been published. This study is the first to present this kind of data and should serve future installations when designing aerosol emission countermeasures.
2. Materials and methods
45
Fig. 1. Pilot plant layout at the Karlsruhe Institute of Technology. SO2 is converted into SO3 inside a micro-reactor over a Pt based catalyst. This SO3 stream is added to the flue gas of a natural gas combustion transforming it into H2 SO4 aerosols inside a quench column. Part of this H2 SO4 aerosol containing flue gas is consequently send into TNO’s mini carbon capture plant.
which is several orders of magnitude lower than the sulfuric acid aerosol concentration (Brachert et al., 2013). The pilot plant has been used to simulate a FGD and understand H2 SO4 aerosol formation. As indicated in Fig. 1, at the outlet of the quench, a split stream is directed to TNO’s CO2 mini-capture plant described below. A T-connection enables the simultaneous measurement of the aerosol number concentration and size distribution by means of a condensation particle counter (UFCPC, PALAS) and the ELPI+ analyser (DEKATI).
2.1. Pilot plant installation: H2 SO4 aerosol generation The experiments are performed at the pilot plant (Fig. 1) at the Institut für Technische Thermodynamik und Kältetechnik (ITTK) in Karlsruhe Institute of Technology (KIT) where a flue gas of ∼180 m3 (STP)/h is generated by burning natural gas (Brachert et al., 2013). This pilot plant has been extensively described in the literature before (Gretscher and Schaber, 1999). The constant dosing of sulfur trioxide (SO3 ) is done via a microreactor where SO2 is oxidized to SO3 at 500 ◦ C. The catalyst of the microreactor is a Pt-catalyst on TiO2 on microstructured titanium foils (Pfeifer et al., 2011). In the flue gas the sulfur trioxide combines with the water vapour from the combustion to sulfuric acid. FTIR has been used to measure the conversion factor from SO2 to SO3 in the reactor by gas composition measurements directly behind the inlet of the flow from the microreactor into the flue gas stream. The possibility of inhomogeneity in the flue gas stream and the low absolute concentrations of SO2 makes these measurements challenging. The measurements give ∼20% conversion factor at 50 sccm (standard cubic centimeters per minute) SO2 which corresponds to ≈15 mg of sulfuric acid per standard cubic meter in the flue gas. The aerosol formation takes place in a quench cooler (packed column absorber). In this absorber, the hot flue gas is cooled down fast from ∼200 ◦ C to the adiabatic saturation temperature of approximately 46 ◦ C by circulating water. The high supersaturation leads to homogeneous nucleation of sulfuric acid and water (Wix et al., 2010). The soot concentration in the pilot plant, which is a possible source for heterogeneous nucleation, is in the order of magnitude of 104 cm−3
2.2. Pilot plant installation: mini-capture plant TNO’s mobile carbon capture mini-plant is a conventional absorption–desorption system with a capacity of 4 m3 h−1 of flue gas. The absorber section is a packed bed (Sulzer Mellapak 2X) with a total height of 3.5 m composed of 4 sections each of 0.5 m height. The absorber section neither has a water wash section nor a demister fitted to the top. The stripper section has the same dimension as the absorber. The total liquid inventory of the system is about 20 l. Additional CO2 was added to the flue gas to ensure a CO2 concentration of about 12 vol% to mimic coal fired power plant flue gas conditions. More details on this plant and its functioning can be found in Khakharia et al. (2013). 2.3. FTIR A FTIR analyser (GASMET CX 4000) was used to analyze the gas phase. The gas leaving the absorber column is heated to 180 ◦ C using a trace heated transfer line. At this temperature the aerosol phase is vaporized implying that the concentration of MEA measured by FTIR is the sum of vapour phase MEA and MEA that was present in the aerosols (Mertens et al., 2012, 2013). The FTIR analyzer was specifically calibrated for MEA. The uncertainty of the components measured by the FTIR depends on the chemical composition (and thus possible interferences) of the matrix in which it is measured. Therefore, stating detection and quantification limits is
46
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
difficult; however they are in the low mg Nm−3 range at the applied conditions (for more information, Mertens et al., 2012, 2013). 2.4. Electrical low pressure impactor (ELPI+ ) The main measurement device used in this work is the electrical low pressure impactor ELPI+ . The particles are charged by corona charging and subsequently separated in a low pressure cascade impactor with 14 electrically insulated collection stages. The measured current signals are proportional to the number concentration and size. By using kernel functions in order to account for the charging efficiency dependency on diameter and for the collection efficiencies of the different stages, the number concentration in every channel can be calculated. A more precise description of the original ELPI can be found in Keskinen et al. (1992) and Marjamäki et al. (2000). The ELPI+ features an additional impactor stage which enlarges the measurement range covered by impactor stages from a cut-off of 30 nm down to a cut-off of 16.7 nm (Yli-Ojanperä et al., 2010). Additionally, a filter stage has been added which collects all the particles that are not trapped in one of the impactor stages. The charging efficiency decreases strongly with the particle size, so that below 6 nm no charging is expected anymore. Thereby, particles down to 6 nm can be measured. The maximum number concentrations that are detectable in every stage as indicated in the manual are not reached during the experiments. Nevertheless, at least a dilution of 10 is used in this work in order to avoid condensation on the impactor plates since the aerosol stream sampled from the pilot plant is saturated with water. The dilution cascade DC 10000 (PALAS GmbH) has been used for this purpose. With this device dilution factors of 10, 102 , 103 and 104 are possible. A last important remark concerns the non-isokinetic sampling of the aerosols and the effect this may have on the observed aerosol sizes. This effect is only important for aerosols that are larger than a few m (Angelo, 2008; Mertens et al., 2013). Mertens et al. (2013) showed the absence of the effect of iso-kinetic sampling on the measured MEA concentrations using the FTIR. This was explained by the fact that the aerosols should be smaller than 1 m which is confirmed in this study as shown in detail below. Therefore, it was judged not necessary to iso-kinetically sample the aerosols for both the FTIR and ELPI+ measurements presented in this study.
Fig. 2. The measured aerosol sizes as measured by Brachert et al. (2013) in front of the absorber are larger than in the current layout because the aerosols shrink as they pass through a fan before entering the absorber.
(ii) An overestimation of the number concentrations measured by the filter stage (i.e. smaller than 10 nm) With respect to the dilution effect, Brachert et al. (2014) show that the measured size distributions for low H2 SO4 concentrations (up to 7.6 sccm) were not affected by the dilution since they are already close to their dry droplet diameter. This term conveys that water cannot be evaporated from these aerosols in the dilution step as the high sulfuric acid concentration in the droplet results in a negligible low vapour pressure. A shrinking of the aerosols at 50 sccm was however noted because of their much larger water content (i.e. increasing H2 SO4 concentration only increases the size and not the number of aerosols) and thus potential for water evaporation and shrinking.
3. Results and discussion Brachert et al. (2014) used the same H2 SO4 generation pilot plant set-up for their comparison of two different measurement principles (Condensation Particle Counter; PALAS, UFCPC with sensor 200 and ELPI+ ) for the measurement of the number concentration of volatile sulfuric acid aerosols. They conclude that both devices confirm the high number concentrations of >108 cm−3 which also corresponds to the simulations carried out by Wix et al. (2010). With the ELPI+ , we were for the first time able to obtain information on sizes by measurement. The prediction from simulations was confirmed: i.e. the number concentrations are always in the same order of magnitude upon increase of the sulfuric acid concentration. However, higher sulfuric acid concentrations lead to an increase in droplet sizes. The ELPI+ size distribution confirmed the simulation results with additional coagulation calculations and revealed droplet sizes well below 100–200 nm which makes them very difficult to precipitate. Brachert et al. (2014) learns that ELPI+ measurements provide a good estimation of the true size distribution but two important things must be taken into account when interpreting the measurements: (i) The effect that the dilution may have on the shrinking of aerosols that contain a lot of water
3.1. Number concentration and size of H2 SO4 aerosols entering the absorber and the effect of dilution A main and important difference in the set-up between the Brachert et al. (2014) set-up and the set-up used in this study is the fact that in the current set-up a fan is used to blow the flue gas into the mini-capture plant. In contrast, in Brachert et al. (2014), a vacuum pump was installed behind the mini-absorber locally creating a strong vacuum. This strong vacuum, however, makes ELPI+ measurements impossible behind the absorber which is one of the two locations of interest in this study. This sampling set-up including the fan, leads to a significant decrease in size as a result of the passing through the fan as clearly visible from Fig. 2. This size decrease is associated to a number increase by a factor 3, from 2.1 E8 to 6.5 E8 cm−3 . This increase is most likely mainly due to the overestimation of the number concentrations when most of the particles are found in the filter stage (Brachert et al., 2014). As shown in Fig. 3, shrinking is no longer observed behind the fan since inside the fan, a similar shrinking process takes place as in the dilution step. This process of water evaporation reduces the aerosols in sizes close to their dry droplet diameter. Similarly to Brachert et al. (2014), we conclude that H2 SO4 aerosol sizes entering the absorber are extremely small and well below 200 nm. It
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
47
Fig. 3. Applying different dilution ratios has no effect on the measured aerosol size distributions in front of the absorber (behind the fan).
is important to stress that at this location, the aerosols are only composed of H2 SO4 and no amine is yet inside. 3.2. Number concentration and size of aerosols leaving the absorber and the effect of dilution Fig. 4 compares the measured aerosol size distributions into the absorber to the measured size distribution leaving the absorber. It is clear that the aerosols grow inside the absorber and their number is reduced by a factor of almost 5 from 6.5 E8 to 1.3 E8 cm−3 . The reduction in number concentration can be explained on the one hand by the reduced overestimation in the ELPI+ due to larger droplets and on the other hand by coagulation in the column.
The size increase inside the absorber is attributed to the uptake of water and MEA as the aerosols travel through the absorber. This water uptake is confirmed by Fig. 5 which shows that behind the absorber, there is now a clear dilution effect, in contrast to Fig. 3 in front of the absorber where no distinct dilution effect has been observed. This indicates that there is a significant amount of water surrounding the aerosols enabling the shrinking effect observed when diluting with dry clear air. Fig. 5 suggests that at a dilution factor of 100, we are close to the dry droplet diameter since dilution 1000 times hardly effects the measured size distribution. Of course the question remains what the true size distribution would be without dilution? This question is treated below where an evaluation of the measured number and sizes of the aerosols is correlated to the
Fig. 4. Aerosols grow as they move through the absorber.
48
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
Fig. 5. Applying different dilution ratios has a clear effect on the measured aerosol size distributions behind the absorber because water is evaporated in the dilution step.
measured MEA emissions from the absorber as measured by the FTIR. 3.3. MEA emissions from the absorber Fig. 6 presents the emissions from the top of the absorber of the time period during which ELPI+ measurements were made behind the absorber. First of all, it is important to repeat that neither water wash nor demister is installed at the top of the mini-absorber and thus no emission control whatsoever. NH3 emissions are observed to be at around 20 mg Nm−3 , oxygen remains at around 10 vol% and CO2 is around 1.5 vol% which corresponds to close to 90% capture efficiency since the CO2 concentration send into the absorber is artificially kept constant at 12 vol%. Between 3000 and 3500 mg Nm−3 of MEA is leaving the absorber and a rising trend in time is observed
Fig. 6. NH3 , MEA, CO2 and O2 concentrations as measured by the FTIR system in the flue gas leaving the absorber during the period in which the ELPI+ measurements were carried out behind the absorber.
most likely due to some instability (e.g. temperature variations) in the mini-pilot plant. It is important to mention that the FTIR is not calibrated to measure these high numbers of MEA. Earlier research has proven that the FTIR overestimates the true value at high concentrations by at least 20%. Therefore, a value of 3000 mg Nm−3 is used for the evaluation presented below although the FTIR readings suggest values up to 3500 mg Nm−3 . It is worth mentioning that MEA emissions from the top of the absorber in the absence of H2 SO4 aerosols are observed around 45 mg Nm−3 as presented in Khakharia et al. (2013). 3.4. Theoretical evaluation of the measured aerosol sizes and numbers and corresponding MEA emissions In this paragraph, a theoretical calculation is carried out to evaluate whether the measured aerosol numbers and corresponding size distribution can contain the high MEA concentration of around 3000 mg Nm−3 as measured by the FTIR. First assumption made here is that the MEA leaving the absorber is all in the form of aerosols. This is not entirely true since Nguyen et al. (2010) report a volatility of 147 mg Nm−3 for 30 wt% aqueous MEA at 40 ◦ C and a typical lean loading of 0.25 mol CO2 mol−1 amine. Temperatures at the top of the absorber in this study were around 30 ◦ C which would reduce the volatile MEA contribution to even lower values so the assumption that all MEA is aerosol MEA is justified. However, one very important assumption to be made here is the MEA concentration inside the aerosols. Condensate measurement of flue gas leaving the absorber carried out on the Maasvlakte pilot plant in 2012, a MEA concentration of 15 wt% was found (Kolderup et al., 2012). The question remains whether this number is valid for the tests carried out here, but since we have no better, it is used in the calculation. This implies that the flue gas needs to contain 20 ml of water per Nm−3 to contain 3000 mg Nm−3 of MEA. Assuming that the aerosols are made up of 100% water, we can calculate the ml of water the aerosols contain based on their size and numbers. Fig. 7 presents this simple calculation by plotting the ml of water per m3 of gas that the aerosols can contain as a function of their assumed size for different aerosol number concentrations. With the help of this plot, the needed droplet sizes
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
49
Comparing the 2.5 ml Nm−3 based on ELPI+ measurements with the required 20 ml Nm−3 based on the calculation reveals a discrepancy which can be explained by: 1. Going from a dilution factor of 10 to no dilution would reveal larger aerosol sizes; similarly as we observed going from a factor 100 to a factor of 10. Since going from 100 times dilution to 10 times dilution increased the water content by a factor of 10, a factor of 10 discrepancy in water content between no dilution and 10 times dilution seems plausible. 2. There is a large uncertainty associated to the 15 wt% assumed in the study. If the concentration would be higher, then again less than 20 ml of water per Nm3 of flue gas would be required to contain 3000 mg Nm−3 of MEA.
Fig. 7. Theoretical relation between aerosol size and water content as a function of the number concentration of the aerosols.
for a certain required water content in the aerosol can be identified. To come to the 20 ml assumed above, the figure reveals that we would need an aerosol number concentration between 1E7 and 1E8 cm−3 assuming that all aerosols have a diameter of 1 m. Fig. 8 presents the calculation of the aerosol water content (plotted as cumulative distribution) for each of the size fractions based on the ELPI+ measurements of the flue gas behind the absorber both at dilution factors 10 and 100 as presented in Fig. 5. It is clear that most aerosols are smaller than 0.1 m and contain hardly any water due to their small size. As also discussed above, the dilution has an important effect on the measured sizes and even more so on the total water content of the aerosols. In case of a 100 times dilution, a total water content of only 0.25 ml Nm−3 is calculated whilst for a ten times dilution 2.5 ml Nm−3 is obtained.
We conclude that the measured number concentrations and corresponding size distribution shown as measured by the ELPI+ behind the absorber and presented above in Fig. 4 are close to reality. It is true that it is likely that in reality, the aerosol size distribution will shift a little to the right (i.e. to larger sizes) although only a slight change is sufficient to match the calculations performed here. We therefore conclude that most of the aerosols are significantly smaller than 1 m and thus very hard to eliminate from the flue gas leaving the absorber using traditional demisters which show very low removal efficiency of submicron particles. However, despite the fact that most of the aerosols (expressed in number concentrations) are well below 1 m, most of the water (and thus amine) is found in the aerosol sizes between 0.5 and 2 m. Therefore, if one aims at designing efficient countermeasures, eliminating this size fraction is crucial. Eliminating the aerosols smaller than 500 nm will likely no longer reduce the amine concentrations leaving the absorber to a large extend. 4. Conclusions The study is the first to present what happens to H2 SO4 aerosol sizes and numbers as they travel through an amine PCCC absorber. H2 SO4 aerosols entering the absorber are observed to be extremely small (i.e. <0.1 m) with number concentrations exceeding 1E8 cm−3 . The aerosols grow as they travel through the absorber through the taking up of water and amine to sizes close to but staying below 1 m. Their number concentration is observed to be slightly reduced most probably due to coagulation processes taking place inside the absorber. However, despite the fact that most of the aerosols (expressed in number concentrations) are well below 1 m, most of the water (and thus amine) is found in the aerosol sizes between 0.5 and 2 m. Therefore, if one aims at designing efficient countermeasures, eliminating this size fraction is crucial. Eliminating the aerosols smaller than 500 nm will likely no longer reduce the amine concentrations leaving the absorber to a large extend. The main conclusion of this study is the fact that aerosols entering and leaving the absorber are small (<2 m) although they can contain lots of amine. This amine emission stream is therefore very difficult to remove using water washes as aerosols are known to travel through water wash sections. Moreover, also classical demisters show very little efficiency for these small aerosol sizes and are therefore believed not to be suitable for the removal of aerosols. This information will therefore serve future installations when designing aerosol emission countermeasures. Acknowledgements
Fig. 8. Measured aerosol size distributions behind the absorber for dilution ratios of 10 and 100 and their corresponding calculated water content (presented as a cumulative distribution).
This research was carried out within the framework of the Dutch CATO-2 program. We are very grateful to TNO and the University
50
J. Mertens et al. / International Journal of Greenhouse Gas Control 23 (2014) 44–50
of Karlsruhe for the collaboration during this experimental work. This research was feasible thanks to the GDF SUEZ CCS research program which partly funded this research. References Angelo, M., 2008. Automatic Particle Monitoring Technology. Presentation for the SICK Company. Brachert, L., Kochenburger, T., Schaber, K., 2013. Facing the sulfuric acid aerosol problem in flue gas cleaning: pilot plant experiments and simulation. Aerosol Science and Technology 10, 1083–1091. Brachert, L., Mertens, J., Khakharia, P., Schaber, K., 2014. The challenge of measuring sulfuric acid aerosols: number concentration and size evaluation using a condensation particle counter (CPC) and an electrical low pressure impactor (ELPI+ ). Journal of Aerosol Science 61, 21–27. Carter, 2012. National Carbon Capture Center: Post-Combustion. In: 2012 NETL CO2 Capture Technology Meeting, July 10 http://www.netl.doe.gov/publications/ proceedings/12/co2capture/presentations/2-Tuesday/T%20Carter-NCCC-Postcombustion.pdf Chi, S., Rochelle, G.T., 2002. Oxidative degradation of monoethanolamine. Industrial & Engineering Chemistry Research 41, 4178–4186. Freeman, S.A., Davis, J., Rochelle, G.T., 2010. Degradation of aqueous piperazine in carbon dioxide capture. International Journal of Greenhouse Gas Control 5, 756–761. Gretscher, H., Schaber, K., 1999. Aerosol formation by heterogeneous nucleation in wet scrubbing processes. Chemical Engineering and Processing 38, 541–548. Kamijo, T., 2011. Amine Emission Control Technology of KM CDR ProcessTM . In: Presentation at the EPRI meeting: amines for post combustion capture, 16 August. Keskinen, J.K., Pietarinen, K., Lehtimäki, M., 1992. Electrical low pressure impactor. Journal of Aerosol Science 23, 353–360. Khakharia, P., Brachert, L., Mertens, J., Huizinga, A., Schallert, B., Schaber, K., Vlugt, T., Goetheer, E., 2013. Investigation of aerosol based emission of MEA due to sulphuric acid aerosol and soot in a post combustion CO2 capture process. International Journal of Greenhouse Gas Control 19, 138–144.
Knudsen, J., Bade, O., 2012. Emission reduction technologies. In: CLIMIT Workshop, 5–6 December 2011, Oslo, Norway, Available at: www.akercleancarbon.com Kolderup, H., da Silva, E., Mejdell, T., Tobiesen, A., Haugen, G., Hoff, K.A., Josefsen, K., Strøm, T., Furuseth, O., Hanssen, K.F., Wirsching, H., Myhrvold, T., Johnsen, K., 2011. Emission Reducing Technologies. Report H&ETQP Amine 6 for Gassnova NF, Norway. http://www.gassnova.no/gassnova2/frontend/files/CONTENT/ Rapporter/EmissionquantificationandreductionWP1and3 Sintef.pdf Kolderup, H., Hjarbo, K., Huizinga, A., Tuinman, I., Zahlsen, K., Vernstad, K., Hyldbakk, A., Holten, T., Kvamsdal, H., van Os, P., da Silva, E., Goetheer, E., Khakharia, P., 2012. Emission Quantification and Reduction (No. WP 1 and 3 in the Project: CCM TQP Amine 6 for Gassnova SF, Norway). Marjamäki, M., Keskinen, J., Chen, D., Pui, D.Y.H., 2000. Performance evaluation of the electrical low-pressure impactor (ELPI). Journal of Aerosol Science 31, 249–261. Mertens, J., Lepaumier, H., Desagher, D., Thielens, M.L., 2013. Understanding ethanolamine (MEA) and ammonia emissions from amine based post combustion carbon capture: lessons learned from field tests. International Journal of Greenhouse Gas Control 13, 72–77. Mertens, J., Thielens, M.L., Knudsen, J., Andersen, J., 2012. On-line monitoring and controlling emissions in amine post combustion carbon capture: a field test. International Journal of Greenhouse Gas Control 6, 2–11. Nguyen, T., Hilliard, M., Rochelle, G.T., 2010. Amine volatility in CO2 capture. International Journal of Greenhouse Gas Control 4, 707–715. Pfeifer, P., Zscherpe, T., Haas-Santo, K., Dittmeyer, R., 2011. Investigations on a Pt/TiO2 catalyst coating for oxidation of SO2 in a microstructured reactor for operation with forced decreasing temperature profile. Applied Catalysis A 391, 289–296. Schaber, K., 1995. Aerosol formation in absorption processes. Chemical Engineering Science 50, 1347–1360. Wix, A., Brachert, L., Sinanis, S., Schaber, K., 2010. A simulation tool for aerosol formation during sulphuric acid absorption in a gas cleaning process. Journal of Aerosol Science 41, 1066–1079. Yli-Ojanperä, J., Kannosto, J., Marjamäki, M., Keskinen, J., 2010. Improving the nanoparticle resolution of the ELPI. Aerosol and Air Quality Research 10, 360–366.