- Email: [email protected]
Carbon dioxide removal system for closed loop atmosphere revitalization, candidate sorbents screening and test results
Acta Astronautica 86 (2013) 39–46
Contents lists available at SciVerse ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
Carbon dioxide removal system for closed loop atmosphere revitalization, candidate sorbents screening and test results$ E.M. Mattox a,n, J.C. Knox b, D.M. Bardot a a b
University of Alabama in Huntsville, United States NASA Marshall Space Flight Center, United States
a r t i c l e i n f o
abstract
Article history: Received 29 February 2012 Received in revised form 19 September 2012 Accepted 24 September 2012 Available online 5 February 2013
Due to the difficulty and expense it costs to resupply manned-spacecraft habitats, a goal is to create a closed loop atmosphere revitalization system, in which precious commodities such as oxygen, carbon dioxide, and water are continuously recycled. Our aim is to test other sorbents for their capacity for future spacecraft missions, such as on the Orion spacecraft, or possibly lunar or Mars mission habitats to see if they would be better than the zeolite sorbents on the 4-bed molecular sieve. Some of the materials being tested are currently used for other industry applications. Studying these sorbents for their specific spacecraft application is different from that for applications on earth because in space, there are certain power, mass, and volume limitations that are not as critical on Earth. In manned-spaceflight missions, the sorbents are exposed to a much lower volume fraction of CO2 (0.6% volume CO2) than on Earth. LiLSX was tested for its CO2 capacity in an atmosphere like that of the ISS. Breakthrough tests were run to establish the capacities of these materials at a partial pressure of CO2 that is seen on the ISS. This paper discusses experimental results from benchmark materials, such as results previously obtained from tests on Grade 522, and the forementioned candidate materials for the Carbon Dioxide Removal Assembly (CDRA) system. & 2012 IAA. Published by Elsevier Ltd. All rights reserved.
Keywords: Carbon dioxide removal Adsorption Commercial sorbents Zeolites
1. Introduction The purpose of this work is to evaluate sorbents being considered for carbon dioxide (CO2) removal on future longterm manned spaceflight. The experimental setup should be versatile in order to study the transient behavior of all candidate materials equally while adsorbing CO2 under the same conditions they would see in space. In the future, the sorbent will be modeled to calculate the mass transfer coefficients. When evaluating the optimum sorbent for long-term manned spaceflight purposes, the system must also be minimized in terms of mass, volume, and power.
$
‘‘This paper was presented during the 62nd IAC in cape town’’. Corresponding author. Tel.: þ1 256 348 4106. E-mail addresses: [email protected] (E.M. Mattox), [email protected] (J.C. Knox), [email protected] (D.M. Bardot).
For long-term spaceflight missions, CO2 must be removed from the air and recycled because resupply opportunities are less frequent or nonexistent. Humans require oxygen, water, and food and in return emit carbon dioxide and other waste products. An average crewmember requires approximately 0.84 kg of oxygen, and emits approximately 1 kg of carbon dioxide. [1] Optimizing which sorbent has the highest capacity is key to optimizing the capture of the carbon dioxide exhaled by the crewmembers so that it can be used for oxygen concentration downstream. Not only is carbon dioxide capture important to recycle materials, but also because carbon dioxide can be toxic to humans at high levels. 1.1. Current CO2 removal technology
n
Currently the International Space Station uses the Carbon Dioxide Removal Assembly (CDRA) to remove
0094-5765/$ - see front matter & 2012 IAA. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actaastro.2012.09.019
40
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
carbon dioxide from the environment. This assembly is a 4-Bed Molecular Sieve. The first two beds are packed with silica gel and zeolite 13X and are used to remove water that would otherwise interfere with carbon dioxide removal in the downstream zeolite 5A beds. The CDRA has been proven to be very effective because it is capable of removing 100% of metabolic CO2 generated by six crewmembers (which is about how many crew members reside on the station at any given time). [1]
swing adsorption (TSA) cycle, a sorbent can be regenerated by increasing the temperature. This is because at any given partial pressure, increasing the temperature will decrease the equilibrium loading. This cycle is paired with passing a purge gas such as nitrogen or helium through the bed to rid the bed of the desorbed components. In pressure swing adsorption (PSA), the process is swung to a low pressure to desorb the sorbent material. 1.4. Adsorption technologies
1.2. Limitations for long-term manned spaceflight Not only is it central to find an ideal sorbent, but the technology used must also be minimized in terms of mass, volume, and power. It is also imperative that the CO2 removal process minimizes use of consumables (or supplies that become waste and must be replenished). The amount of volume the system takes up must be limited because there is a restricted amount of space on any long-term manned spacecraft. The assembly should also be constrained in power usage as power is a limited resource and maximizing efficiency is important. Mass must be limited due to the limitations in launch vehicle payload capabilities. 1.3. Adsorption The CO2 removal technology exploits the process of adsorption where molecules or atoms in a fluid phase diffuse to the surface of a highly porous solid where they bond with the solid surface or are held there by weak intermolecular forces. [2] This should not be confused with absorption where a gas mixture is contacted with a liquid for the purposes of preferentially dissolving one or more components of the gas into the liquid. [2] Separation by adsorption is based on steric, equilibrium, and kinetic mechanisms. [3] Sterically, the porous solid has pores of a dimension such that it allows molecules of a smaller size to enter, while excluding those molecules which are larger. With equilibria, the solid has different abilities to accommodate different species and therefore the stronger adsorbing species are preferentially removed by the solid. In terms of kinetics, adsorption separation is based on the different rates of diffusion of different species into the pore space. [3] Adsorption is an exothermic process, and causes a release in energy. Therefore when carbon dioxide adsorbs onto sorbent, the sorbent increases in temperature; this heat is known as the heat of adsorption. Conversely, when a material is being stripped (or desorbed) it cools down because desorption is an endothermic process, which requires energy. This heat generated during adsorption creates a problem for any further adsorption. The higher temperature of the sorbent decreases the sorbent capacity, and therefore inhibits further adsorption. Similarly, during desorption the temperature drops and impedes further desorption. [4] The heat of adsorption provides a direct measure of the bonding strength of the sorbate to the surface of the sorbent. [5] In practice adsorption and desorption steps must be used for cyclic operation; examples are temperature swing adsorption and pressure swing adsorption. In a temperature
There are many different sorbents used for adsorption; some are better for carbon dioxide adsorption than others depending on the three different mechanisms of adsorption. Zeolites, silica gel, amines, hydrotalcites, activated carbon, and alumina are among many sorbents used for adsorption. Zeolite sorbents will be the primary focus of this paper. Test results are for the sorbents: GraceDavidson Grade 522 zeolite 5A (Grade 522) and UOP OXYSIV MDX (LiLSX). 1.5. Carbon dioxide breakthrough testing In order to study the working capacities of the different sorbents with computer simulation, mass transfer coefficients are required. A comparison of simulated and experimental breakthrough curves is used to obtain the mass transfer coefficients empirically. When testing the capacities of the different materials, carbon dioxide mixed with nitrogen gas is passed through the test article at a defined CO2 percentage or a constant CO2 mass flow rate. The zeolite pellets adsorb the carbon dioxide until fully saturated; this period is called the full breakthrough time. [6] Breakthrough curves can be integrated to determine the percent of CO2 upstream of the bed and the percent of CO2 downstream of the bed with respect to time. The flow upstream of the bed remains constant while the percent downstream starts at 0 and then at full breakthrough matches the percent of CO2 upstream. 1.6. Pellet size effects on adsorption Using the Ergun equation [5], the following chart (Fig. 1) was generated to demonstrate how the particle diameter of the sorbent affects the pressure drop in the bed. At smaller particle sizes, the gas flow rate has more of an impact on the pressure drop than at larger particle diameters. The pressure drop in the bed is proportional to the amount of carbon dioxide that the bed is able to load. 2. Experimental setup 2.1. Test article The test article used in this experiment was designed to be of minimal mass, ‘‘long and thin,’’ and axisymmetric. Minimizing thermal mass is desired so that when studying the heat of adsorption, results are not confounded by a large thermal mass storing and releasing heat and transferring heat axially in the tube walls. We desire to model it as infinitely long so we can neglect the thermal mass of
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
10 8
[kPa]
Predicted Pressure Drop
12
16 SLPM
6
8 SLPM
4 2 0
0
0.5
1
1.5
2
2.5
41
CO2 removal in space. Below is information on these two materials. It should be noted that the UOP OXYSIV MDX (LiLSX) material is much smaller in size than the zeolite 5A that was tested. This is because this was the only size of OXYSIV MDX that is currently available to test. If used in spacecraft applications, it will have to be sized up because the small particle size creates a much larger pressure drop in the bed (7.5 kPa as opposed to 0.5 kPa), which then requires a greater pumping requirement, which conflicts with the goal to minimize power used in space.
Particle Diameter [mm] Fig. 1. Effect of particle diameter on the pressure drop in the bed for different flow rates.
3. Methodology 3.1. Material activation To explore the different sorbents’ capacities for carbon dioxide, the sorbent must be completely purged of any water, carbon dioxide or any other trace contaminants. This is important so that the bed is not pre-loaded with water which reduces the sorbent CO2 capacity. To purge the sorbent, it must be completely desorbed. To fully regenerate the Grade 522 and the LiLSX, the sorbents were desorbed by an adapted version of a procedure developed by UOP-LLC Adsorbents and Specialties. [7] The procedure is as follows:
Fig. 2. Test article.
the endcaps. To this end, sorbent pellets are packed in the axial core of the bed and is held in place between two aluminum perforated plates. Additionally the endcaps are thermally isolated from the aluminum tube with perfluoroelastomer o-rings. The drawing is shown in Fig. 1. The static pressure of the bed is measured in the center of the upper plenum and the differential pressure from the upper to the lower plenum. This enables the absolute pressure in both ends of the bed to be calculated. The article is also axisymmetric and all thermocouples read temperatures at the same axial location (in the center of the bed). A custom thermocouple that contains 4 sensing points in one probe was used so that the thermocouples could be inserted from the ends. It is beneficial for the thermocouples to come from both ends rather than from the sides so that the mass of the metal fittings that would be required does not conflict with the heat of adsorption so that the same axial location can be read at all points in the test article. Fig. 2 2.2. Tested sorbents The baseline material used in this experiment is Grace Davidson Grade 522 zeolite 5A. A new sorbent recently developed by UOP was analyzed for its future potential for
1. Flow the purge gas (nitrogen) at a flow rate that does not cause fluidization of the pellets. 2. Set the initial oven temperature to 50 1C. Allow the sorbent to reach this temperature. Once the sorbent has reached a temperature of 50 1C, dwell for 30 min. 3. Increase the temperature setting to 100 1C and allow the zeolite to reach this temperature. Dwell at 100 1C for another 30 min. 4. Increase the temperature of the sorbent to 150 1C and then keep increasing the temperature another 50 1C until the zeolite reaches 350 1C. Be sure not to exceed a ramp rate of 3.33 1C/min. 5. Allow the zeolite to dwell at 350 1C for 3.5 h. The column used in this experiment was a thin-walled titanium tube with a 1 in. outside diameter. The ends were fitted with stainless steel Swagelok fittings so that the column could be purged and then completely sealed after regeneration. 3.1.1. Keeping the material activated After each breakthrough test, we desire to completely regenerate the sorbent before testing it again. Since it is desired that testing be done in a timely manner, the quickest regeneration method is desired. In Fig. 3, the same test was run after regenerating the bed 4 different ways: in the oven at 350 1C (note: this method requires unpacking the material because the test article is not rated to 350 1C), with a nitrogen purge for 60 h, overnight on a floor heater at 275 1C, and in-place on the test rig at 175 1C. Because each test looks about the same as the ideal ‘‘350 1C case,’’ the easiest method was chosen to regenerate the bed (regenerating in place to 175 1C).
42
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
3) Spring-load the bed and vibrate at a slightly lower level overnight. 4) Check to see if the material level has changed and if it has not, re-seal the test bed, but if it has changed add enough material to get back to the desired volume. 3.3. CO2 breakthrough testing
Fig. 3. Comparison of regeneration techniques.
Breakthrough tests ran at 5 Torr CO2 and 16 SLPM were run after using different regeneration techniques to prove it as the least time consuming technique. 3.2. Packing the sorbent canister 3.2.1. Packing grace Davidson zeolite 5A The sorbent pellets are packed according to a standard procedure so that differing results are due to differences in packing density. In general, it is the goal that the bed is packed as densely as possible. However, it is important that the fragile sorbents not be damaged in packing. In particular, the zeolite pellets can fracture under too much force. It is important to avoid creating fines because the dust is harmful to mechanical equipment downstream of the adsorption beds. The packing method was developed by starting with a Desiccant/Adsorbent Bed Packing Procedure developed by Honeywell Corporation and then adapting it for the experimental test article. The method is as follows: 1. Pour 12 cm3 of sorbent pellets into the test article. 2. Settle the particles by vibrating the test article on a vibration table for 4 min and then place a 115 g mass on top of the pellets and vibrate at a slightly higher vibration level for 1 min. 3. Repeat (1) and (2) until the packing depth of 5 in. has been reached.
3.2.2. Packing UOP OXYSIV MDX Additional packing tests were done to find how to optimally pack the test bed with the Honeywell UOP OXYSIV MDX material. The following method was used to pack the test bed because it was shown to have the best results. 1) In a drybox, put the material that has been fully desorbed and the test bed. 2) Pour material in test article at a rate of 1 mm/sec or slower until designated volume is reached, while simultaneously vibrating the material at a level just below that in which bulk movement is induced.
Sorbent capacity can be tested by running a Breakthrough test. In a Breakthrough test, nitrogen and carbon dioxide are passed through the bed and the time it takes to completely load the bed with carbon dioxide is measured. The bed temperatures are also measured to quantify the heat of adsorption that is generated. To run the test, the bed is first pre-heated with nitrogen. The nitrogen gas is pre-heated by running it through a heat exchanger. This is done to get the bed at a constant temperature that is slightly above ambient so any fluctuations in temperature that are seen are not due to fluctuations in ambient temperature. Once the bed is heated, carbon dioxide can then be mixed with the inlet nitrogen stream and heated as well. After heating the gas stream will then be sampled and a measure of its dew point and percent CO2 will be taken. The remainder of the flow will pass through the bed. At the exit of the bed, the flow will again be sampled and its dew point and a measure of its percent CO2 will be taken. The remainder of the flow will exit to ambient. The flow of CO2 can either be held at a constant percent pressure or a constant mass flow. When the flow of CO2 is held at a constant percent pressure, the mass flow is constantly fluctuating with ambient pressure fluctuations. Likewise, when the mass flow is kept constant, the percent pressure of carbon dioxide constantly fluctuates with ambient pressure fluctuations. On either case, the percent of CO2 entering the bed will stay relatively constant. Breakthrough occurs when the bed is completely saturated with carbon dioxide and can no longer take on any more. The time at which the column hits breakthrough will become apparent when the exit CO2 analyzer measures the same percent carbon dioxide as the inlet CO2 analyzer. At this point, the carbon dioxide stream can be shut-off and the bed will be purged with nitrogen. This will cause desorption and the inlet CO2 analyzer should measure 0 percent carbon dioxide instantly where the exit CO2 analyzer should measure the CO2 that is coming off of the bed. Bed temperatures will also be measured during this adsorption/desorption process. During adsorption the temperatures should rise and once breakthrough occurs, Table 1 Grade 522 information. Material:
Grade 522 8 12
Manufacturer: Lot no. Material code: Place of manufacture: Ingredients
Grace-Davidson 9 52208350201 China Sodium aluminosilicate, potassium sodium aluminosilicate or calcium sodium aluminosilicate, clay, quartz, water
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
the temperatures in the bed should fall back to steady state. Then as desorption occurs, the bed temperatures should get colder in the opposite fashion as they did during adsorption. Table 1 3.4. Calculations In a CO2 Breakthrough test, the nitrogen/ carbon dioxide gas mixture is passed through a Sable Systems CO2 Analyzer, where the partial pressure of CO2 is measured, as well as the total gas pressure. Using these measurements, the mole fraction of CO2 is calculated. It is assumed that the mole fraction of CO2 in the gas stays constant. In the entrance and exit of the bed, the total pressure is measured, so using these total pressure measurements, along with the calculated mole fraction of CO2 in the gas stream, the partial pressure of CO2 in the bed inlet and exit is calculated. Using the ideal gas law, a plot of the number of moles of CO2 being taken on by the bed at each moment in time is made. This plot is integrated to find the number of moles of CO2 the bed has taken on. Knowing the true mass of sorbent in the bed, the sorbent’s capacity for CO2 [mol CO2/ kg sorbent] is calculated. Fig. 4 is a representation of the system, showing where measurements are taken. To compare the results, the Monte Carlo Method [8] was used to calculate the uncertainty in the carbon dioxide capacity calculation. Table 2 4. Results and discussion Two pelletized sorbents are compared in this section. Grace-Davidson Grade 522 zeolite 5A (Grade 522) and Honeywell UOP OXYSIV MDX (LiLSX) were tested for their CO2 capacity in an atmosphere like that of the ISS. Breakthrough tests were run to establish the capacities of these materials. All materials were tested under four different test conditions as shown in Table 3. Grade 522 (Mesh Sizes: 8 12 and 20 50) and LiLSX (Mesh Size: 20 50) are compared. Figs. 6 and 9 compare the material adsorption breakthrough curves. Fig. 7 and 10 show desorption breakthrough curves. Fig. 8 compares the temperature profiles of the two materials. In Fig. 5, all materials are compared at 5 Torr CO2 and 16 SLPM. In this plot, it can be immediately seen that the LiLSX material has a considerably larger capacity for carbon dioxide. When looking at the Grade 522 and LiLSX, at the same mesh size, it can be seen that the mass transfer coefficients of the two materials are approximately the same, indicated by the similar slopes of the breakthrough
43
curves. However, when breakthrough curves of the Grade 522 at the two sizes are compared, it is seen that the breakthrough curve slope significantly increases when the material size is decreased. Therefore, decreasing a material’s size increases its mass transfer capabilities. The only thing that was changed when the material was crushed was the macropore size of the material. This indicates that the Grade 522 material is limited by macropore diffusion. The macropore is defined as the space in the pellet between
Table 2 UOP OXYSIV MDX information. Material:
OXYSIV MDX 20 50 (LiLSX)
Manufacturer: Lot no. Material no. Place of manufacture: Ingredients
UOP LLC, A Honeywell company 2010010918 80555-999 China Silica oxide (synthetic), aluminum oxide, lithium oxide, potassium oxide, sodium oxide, water, quartz
Table 3 Test conditions. Test condition:
1
2
3
4
Carrier gas flow rate [SLPM] Approximate CO2 partial pressure [Torr] Actual CO2 partial pressure [kPa]
16 5 0.67
16 2.5 0.31
8 5 0.67
8 2.5 0.31
Fig. 5. Material Comparison: adsorption cycle, 5 Torr (0.67 kPa) CO2, 16 SLPM.
Fig. 4. Schematic of test system with measurement locations.
44
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
Fig. 6. Comparison of bed exit temperatures. Adsorption cycle, 5 Torr (0.67 kPa) CO2, 16 SLPM.
each crystal, while the micropore is the interstitial space between each crystal where the CO2 molecule is physically adsorbs. Fig. 6 is very interesting because it gives insight into the heat of adsorption characteristics of different materials, and shows how material size can affect the heat of adsorption. Grade 522 (8 12 mesh) is shown to have a higher mass transfer resistance because the exit temperature curve is slower and rounded because the uptake is smeared over time. Its temperature rises first because it initially takes up CO2. It is noted here that the smaller materials show a sharp uptake in CO2 because the exit temperature curves rise sharply. However, the Grade 522 curve looks like a smaller version of the LiLSX curve. Also, when comparing the Grade 522 material at different sizes, it looks as though both contain the same area, although they are sloped differently. This indicates that the material size plays a large role in the mass transfer characteristics of a sorbent, but not necessarily its overall capacity. The increase in mass transfer for the smaller materials is because the macropore diffusion is increased for a smaller pelletized sorbent. This implies that a material can take on the same amount of carbon dioxide regardless of size; size just affects the rate at which it can take on that amount of carbon dioxide. In Fig. 7, the desorption curves of the three materials are compared. From this graph, it appears that the Grade 522, 8 12 mesh and the LiLSX material release the carbon dioxide at the same rate. It should be noticed that when Fig. 5 is compared to Fig. 8, there is a bigger time difference between the breakthrough time of the LiLSX material and the Grade 522, mesh 20 50 for the lower carbon dioxide partial pressure. This indicates that the LiLSX material can adsorb more CO2 than the Grade 522 material at lower CO2 partial pressures. In Fig. 9, desorption curves of the three materials are compared for 2.5 Torr CO2 and 16 SLPM. From this graph, it appears that the Grade 522 (20 50), releases the carbon dioxide the fastest and the LiLSX releases the carbon dioxide the slowest.
Fig. 7. Material comparison: desorption cycle, 5 Torr (0.67 kPa) CO2, 16 SLPM.
Fig. 8. Material comparison: adsorption cycle, 2.5 Torr (0.31 kPa) CO2, 16 SLPM.
Fig. 9. Material comparison: desorption cycle, 2.5 Torr (0.31 kPa) CO2, 16 SLPM.
Table 4 gives the results of each test. For each sorbent, the mass of sorbent used is given. The conditions as to what carbon dioxide partial pressure was used and the carrier gas (nitrogen) flow rate in standard liters per
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
Table 4 Summary of results. Sorbent
Mass [g]
Grace Grade 125 522 5A, Mesh: 8 12
Conditions [Torr]/ [SLPM] 5/16 5/8 2.5/16 2.5/8
UOP LiLSX, Mesh: 20 50
102
5/16 5/8
2.5/16
2.5/8 Grace Grade 522 5A, Mesh: 20 50
106
5/16 2.5/16
Mean capacity 95% [MCM] [mol CO2/ Confidence interval [%] kg sorbent] 1.185 1.048 0.9756 0.9871 0.5955 0.5985 0.5738 0.5569 1.9721 2.004 1.7364 1.7811 1.7795 1.5035 1.479 1.5321 1.3193 1.3246 1.3915 1.6707 0.8451 0.9806
5.74% 5.31% 3.59% 2.51% 3.76% 4.14% 3.56% 3.41% 2.59% 2.63% 2.05% 2.75% 3.61% 1.94% 1.68% 2.25% 2.70% 2.58% 4.41% 4.85% 3.74% 3.67%
minute (SLPM). The Monte Carlo Method [8] was used to calculate the mean capacity with the corresponding 95% Confidence interval. 5. Conclusion In this work, materials were compared for their potential to remove carbon dioxide from a space-habitat where resupply is difficult and under the limitations of minimizing mass, power, and volume. In the future, NASA would like to build a lunar or Martian habitat for astronauts to live. A very reliable and efficient life support system is needed to remove carbon dioxide from the habitat. This is important because at CO2 pressures above 5 Torr, astronauts start to suffer from headaches, and because the CO2 can be recycled downstream to make water or oxygen. The materials that were analyzed are Grace-Davidson Grade 522 5A (Grade 522) and UOP OXYSIV MDX (LiLSX). The Grade 522 material used was an 8 12 mesh size and the LiLSX material used was a 20 50 mesh size. In order to compare the materials without pellet size confounding the results, the Grade 522 material was crushed to a 20 50 mesh size and tested again. This also allowed us to compare the Grade 522 material at the two sizes to evaluate how pellet size affects CO2 capacity and mass transfer capabilities. The LiLSX material is a relatively new commercially available sorbent marketed for oxygen concentration. We wanted to test LiLSX for its CO2 capture potential for future spacecraft use, so it was tested along with the current standard sorbent for CO2 capture, Grade 522. A test stand was developed to uniformly compare the two materials. The test stand composed of a test article, a heat exchanger, nitrogen and carbon dioxide gases, mass flow controllers, and sample lines upstream and downstream of the test article which composed of sample pumps,
45
CO2 analyzers, and dew point analyzers. The material being tested was packed in the center of the test article. Within the test article, temperature and pressure measurements were made upstream and downstream of the packed bed. Breakthrough tests were run at two flow rates (8 SLPM and 16 SLPM) and two CO2 partial pressures (2.5 Torr and 5 Torr). During a breakthrough test, a mixture of nitrogen and carbon dioxide was passed through the bed and the amount of carbon dioxide that is adsorbed by the packed bed is calculated using the nitrogen and carbon dioxide flow rates and the pressure of carbon dioxide and temperatures at the inlet and exit of the bed. At the beginning of a test, all carbon dioxide that is being passed through the bed will be captured by the sorbent and the exit carbon dioxide analyzer will show a CO2 partial pressure of zero. Temperatures in the bed will begin to rise because adsorption is an exothermic process. Breakthrough begins when the bed can no longer hold all of the CO2 being passed through it and the exit CO2 analyzer will begin to show a positive CO2 partial pressure. The CO2 partial pressure in the exit stream will continue to rise until breakthrough is complete, at which the bed is completely saturated and cannot take on any more carbon dioxide. At this point the exit CO2 pressure will equal the inlet CO2 pressure and bed temperatures will return to steady state. Ultimately, the materials’ capacity for carbon dioxide is compared. The capacity is how much CO2 a defined amount of material can take on. This is usually expressed in units of mol CO2/ kg sorbent. The number of moles of carbon dioxide entering and exiting the bed was calculated and the difference is how much CO2 was loaded on to the bed. The capacity is calculated by dividing the number of moles loaded on to the bed by the mass of the sorbent in the packed bed. An uncertainty analysis was performed using the Monte Carlo Method [8] so that the capacity measurements for each sorbent could be accurately compared. An overall results table is shown on the previous page, in Table 5. In this table, the CO2 capacity and the results from the uncertainty analysis are given for the two sorbents under the four different sets of conditions. From examining this table, it is shown that the capacity of a material is increased slightly by decreasing its pellet diameter. Then when comparing LiLSX and Grade 522, at the same mesh sizes, it can be seen that the LiLSX material performs better at both carbon dioxide partial pressures, when tested at the higher flow rate. However, the increase in capacity is higher at the lower carbon dioxide partial pressure. The results are least consistent for the Grade 522 Mesh 20 50 tests. The uncertainty is not large enough to note these as statistically the same. These differences results from different pressure drop readings in the bed under the same conditions. The bias error is very small for the pressure drop is very small. These results indicate that the error in this reading is probably actually larger. This error was the only error that came from the NASA calibration lab. In the future, I would send this instrument back to the manufacturer for a proper inspection and manufacturer recalibration. From looking at the results, I would recommend using the LiLSX material in future spacecraft applications because it has been proven to have a higher capacity for CO2 than the standard, Grade 522. If it is used for spacecraft applications,
46
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
it will have to be scaled up to minimize the power requirements for pushing air through the bed.
Acknowledgments
[2] [3] [4]
This work was supported by the NASA Marshall Space Flight Center under Cooperative Agreements NNM05AA22A and NNM11AA01A. References
[5] [6] [7] [8]
[1] J.C. Knox, New methods for the adsorption of carbon dioxide and water vapor from manned spacecraft atmospheres: applications and
modelling, in: Proceedings of the Comsol Conference, Boston, MA, 2007, Key Note Speech. Benitez Jaime, Principles and Modern Applications of Mass Transfer Operations, 2nd ed. John Wiley & Sons, Inc., New Jersey, 2009. Duong D. Do, Adsorption Analysis: Equilibria and Kinetics, Imperial College Press, London, 1998. James C. Knox, David Howard, Clearing the Air: Life Support for Space Exploration, Comsol. News (2008) 4–6. Douglas M. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley and Sons, Inc, New York, 1984. M. Douglas LeVan, G. Carta, Adsorption and Ion Exchange, Perry’s Chemical Engineering Handbook, 8th ed, print, pp.16-1 –16-15. A Suggested Laboratory Purge Activation Procedure, uop.com, Universal Oil Products: Adsorbents & Specialties, n.d., Web, 7 September, 2010. Hugh W. Coleman, W. Glenn Steele, Experimentation, Validation, and Uncertainty Analysis for Engineers, 3rd Ed, John Wiley & Sons, Inc., 2009.
Contents lists available at SciVerse ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
Carbon dioxide removal system for closed loop atmosphere revitalization, candidate sorbents screening and test results$ E.M. Mattox a,n, J.C. Knox b, D.M. Bardot a a b
University of Alabama in Huntsville, United States NASA Marshall Space Flight Center, United States
a r t i c l e i n f o
abstract
Article history: Received 29 February 2012 Received in revised form 19 September 2012 Accepted 24 September 2012 Available online 5 February 2013
Due to the difficulty and expense it costs to resupply manned-spacecraft habitats, a goal is to create a closed loop atmosphere revitalization system, in which precious commodities such as oxygen, carbon dioxide, and water are continuously recycled. Our aim is to test other sorbents for their capacity for future spacecraft missions, such as on the Orion spacecraft, or possibly lunar or Mars mission habitats to see if they would be better than the zeolite sorbents on the 4-bed molecular sieve. Some of the materials being tested are currently used for other industry applications. Studying these sorbents for their specific spacecraft application is different from that for applications on earth because in space, there are certain power, mass, and volume limitations that are not as critical on Earth. In manned-spaceflight missions, the sorbents are exposed to a much lower volume fraction of CO2 (0.6% volume CO2) than on Earth. LiLSX was tested for its CO2 capacity in an atmosphere like that of the ISS. Breakthrough tests were run to establish the capacities of these materials at a partial pressure of CO2 that is seen on the ISS. This paper discusses experimental results from benchmark materials, such as results previously obtained from tests on Grade 522, and the forementioned candidate materials for the Carbon Dioxide Removal Assembly (CDRA) system. & 2012 IAA. Published by Elsevier Ltd. All rights reserved.
Keywords: Carbon dioxide removal Adsorption Commercial sorbents Zeolites
1. Introduction The purpose of this work is to evaluate sorbents being considered for carbon dioxide (CO2) removal on future longterm manned spaceflight. The experimental setup should be versatile in order to study the transient behavior of all candidate materials equally while adsorbing CO2 under the same conditions they would see in space. In the future, the sorbent will be modeled to calculate the mass transfer coefficients. When evaluating the optimum sorbent for long-term manned spaceflight purposes, the system must also be minimized in terms of mass, volume, and power.
$
‘‘This paper was presented during the 62nd IAC in cape town’’. Corresponding author. Tel.: þ1 256 348 4106. E-mail addresses: [email protected] (E.M. Mattox), [email protected] (J.C. Knox), [email protected] (D.M. Bardot).
For long-term spaceflight missions, CO2 must be removed from the air and recycled because resupply opportunities are less frequent or nonexistent. Humans require oxygen, water, and food and in return emit carbon dioxide and other waste products. An average crewmember requires approximately 0.84 kg of oxygen, and emits approximately 1 kg of carbon dioxide. [1] Optimizing which sorbent has the highest capacity is key to optimizing the capture of the carbon dioxide exhaled by the crewmembers so that it can be used for oxygen concentration downstream. Not only is carbon dioxide capture important to recycle materials, but also because carbon dioxide can be toxic to humans at high levels. 1.1. Current CO2 removal technology
n
Currently the International Space Station uses the Carbon Dioxide Removal Assembly (CDRA) to remove
0094-5765/$ - see front matter & 2012 IAA. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.actaastro.2012.09.019
40
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
carbon dioxide from the environment. This assembly is a 4-Bed Molecular Sieve. The first two beds are packed with silica gel and zeolite 13X and are used to remove water that would otherwise interfere with carbon dioxide removal in the downstream zeolite 5A beds. The CDRA has been proven to be very effective because it is capable of removing 100% of metabolic CO2 generated by six crewmembers (which is about how many crew members reside on the station at any given time). [1]
swing adsorption (TSA) cycle, a sorbent can be regenerated by increasing the temperature. This is because at any given partial pressure, increasing the temperature will decrease the equilibrium loading. This cycle is paired with passing a purge gas such as nitrogen or helium through the bed to rid the bed of the desorbed components. In pressure swing adsorption (PSA), the process is swung to a low pressure to desorb the sorbent material. 1.4. Adsorption technologies
1.2. Limitations for long-term manned spaceflight Not only is it central to find an ideal sorbent, but the technology used must also be minimized in terms of mass, volume, and power. It is also imperative that the CO2 removal process minimizes use of consumables (or supplies that become waste and must be replenished). The amount of volume the system takes up must be limited because there is a restricted amount of space on any long-term manned spacecraft. The assembly should also be constrained in power usage as power is a limited resource and maximizing efficiency is important. Mass must be limited due to the limitations in launch vehicle payload capabilities. 1.3. Adsorption The CO2 removal technology exploits the process of adsorption where molecules or atoms in a fluid phase diffuse to the surface of a highly porous solid where they bond with the solid surface or are held there by weak intermolecular forces. [2] This should not be confused with absorption where a gas mixture is contacted with a liquid for the purposes of preferentially dissolving one or more components of the gas into the liquid. [2] Separation by adsorption is based on steric, equilibrium, and kinetic mechanisms. [3] Sterically, the porous solid has pores of a dimension such that it allows molecules of a smaller size to enter, while excluding those molecules which are larger. With equilibria, the solid has different abilities to accommodate different species and therefore the stronger adsorbing species are preferentially removed by the solid. In terms of kinetics, adsorption separation is based on the different rates of diffusion of different species into the pore space. [3] Adsorption is an exothermic process, and causes a release in energy. Therefore when carbon dioxide adsorbs onto sorbent, the sorbent increases in temperature; this heat is known as the heat of adsorption. Conversely, when a material is being stripped (or desorbed) it cools down because desorption is an endothermic process, which requires energy. This heat generated during adsorption creates a problem for any further adsorption. The higher temperature of the sorbent decreases the sorbent capacity, and therefore inhibits further adsorption. Similarly, during desorption the temperature drops and impedes further desorption. [4] The heat of adsorption provides a direct measure of the bonding strength of the sorbate to the surface of the sorbent. [5] In practice adsorption and desorption steps must be used for cyclic operation; examples are temperature swing adsorption and pressure swing adsorption. In a temperature
There are many different sorbents used for adsorption; some are better for carbon dioxide adsorption than others depending on the three different mechanisms of adsorption. Zeolites, silica gel, amines, hydrotalcites, activated carbon, and alumina are among many sorbents used for adsorption. Zeolite sorbents will be the primary focus of this paper. Test results are for the sorbents: GraceDavidson Grade 522 zeolite 5A (Grade 522) and UOP OXYSIV MDX (LiLSX). 1.5. Carbon dioxide breakthrough testing In order to study the working capacities of the different sorbents with computer simulation, mass transfer coefficients are required. A comparison of simulated and experimental breakthrough curves is used to obtain the mass transfer coefficients empirically. When testing the capacities of the different materials, carbon dioxide mixed with nitrogen gas is passed through the test article at a defined CO2 percentage or a constant CO2 mass flow rate. The zeolite pellets adsorb the carbon dioxide until fully saturated; this period is called the full breakthrough time. [6] Breakthrough curves can be integrated to determine the percent of CO2 upstream of the bed and the percent of CO2 downstream of the bed with respect to time. The flow upstream of the bed remains constant while the percent downstream starts at 0 and then at full breakthrough matches the percent of CO2 upstream. 1.6. Pellet size effects on adsorption Using the Ergun equation [5], the following chart (Fig. 1) was generated to demonstrate how the particle diameter of the sorbent affects the pressure drop in the bed. At smaller particle sizes, the gas flow rate has more of an impact on the pressure drop than at larger particle diameters. The pressure drop in the bed is proportional to the amount of carbon dioxide that the bed is able to load. 2. Experimental setup 2.1. Test article The test article used in this experiment was designed to be of minimal mass, ‘‘long and thin,’’ and axisymmetric. Minimizing thermal mass is desired so that when studying the heat of adsorption, results are not confounded by a large thermal mass storing and releasing heat and transferring heat axially in the tube walls. We desire to model it as infinitely long so we can neglect the thermal mass of
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
10 8
[kPa]
Predicted Pressure Drop
12
16 SLPM
6
8 SLPM
4 2 0
0
0.5
1
1.5
2
2.5
41
CO2 removal in space. Below is information on these two materials. It should be noted that the UOP OXYSIV MDX (LiLSX) material is much smaller in size than the zeolite 5A that was tested. This is because this was the only size of OXYSIV MDX that is currently available to test. If used in spacecraft applications, it will have to be sized up because the small particle size creates a much larger pressure drop in the bed (7.5 kPa as opposed to 0.5 kPa), which then requires a greater pumping requirement, which conflicts with the goal to minimize power used in space.
Particle Diameter [mm] Fig. 1. Effect of particle diameter on the pressure drop in the bed for different flow rates.
3. Methodology 3.1. Material activation To explore the different sorbents’ capacities for carbon dioxide, the sorbent must be completely purged of any water, carbon dioxide or any other trace contaminants. This is important so that the bed is not pre-loaded with water which reduces the sorbent CO2 capacity. To purge the sorbent, it must be completely desorbed. To fully regenerate the Grade 522 and the LiLSX, the sorbents were desorbed by an adapted version of a procedure developed by UOP-LLC Adsorbents and Specialties. [7] The procedure is as follows:
Fig. 2. Test article.
the endcaps. To this end, sorbent pellets are packed in the axial core of the bed and is held in place between two aluminum perforated plates. Additionally the endcaps are thermally isolated from the aluminum tube with perfluoroelastomer o-rings. The drawing is shown in Fig. 1. The static pressure of the bed is measured in the center of the upper plenum and the differential pressure from the upper to the lower plenum. This enables the absolute pressure in both ends of the bed to be calculated. The article is also axisymmetric and all thermocouples read temperatures at the same axial location (in the center of the bed). A custom thermocouple that contains 4 sensing points in one probe was used so that the thermocouples could be inserted from the ends. It is beneficial for the thermocouples to come from both ends rather than from the sides so that the mass of the metal fittings that would be required does not conflict with the heat of adsorption so that the same axial location can be read at all points in the test article. Fig. 2 2.2. Tested sorbents The baseline material used in this experiment is Grace Davidson Grade 522 zeolite 5A. A new sorbent recently developed by UOP was analyzed for its future potential for
1. Flow the purge gas (nitrogen) at a flow rate that does not cause fluidization of the pellets. 2. Set the initial oven temperature to 50 1C. Allow the sorbent to reach this temperature. Once the sorbent has reached a temperature of 50 1C, dwell for 30 min. 3. Increase the temperature setting to 100 1C and allow the zeolite to reach this temperature. Dwell at 100 1C for another 30 min. 4. Increase the temperature of the sorbent to 150 1C and then keep increasing the temperature another 50 1C until the zeolite reaches 350 1C. Be sure not to exceed a ramp rate of 3.33 1C/min. 5. Allow the zeolite to dwell at 350 1C for 3.5 h. The column used in this experiment was a thin-walled titanium tube with a 1 in. outside diameter. The ends were fitted with stainless steel Swagelok fittings so that the column could be purged and then completely sealed after regeneration. 3.1.1. Keeping the material activated After each breakthrough test, we desire to completely regenerate the sorbent before testing it again. Since it is desired that testing be done in a timely manner, the quickest regeneration method is desired. In Fig. 3, the same test was run after regenerating the bed 4 different ways: in the oven at 350 1C (note: this method requires unpacking the material because the test article is not rated to 350 1C), with a nitrogen purge for 60 h, overnight on a floor heater at 275 1C, and in-place on the test rig at 175 1C. Because each test looks about the same as the ideal ‘‘350 1C case,’’ the easiest method was chosen to regenerate the bed (regenerating in place to 175 1C).
42
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
3) Spring-load the bed and vibrate at a slightly lower level overnight. 4) Check to see if the material level has changed and if it has not, re-seal the test bed, but if it has changed add enough material to get back to the desired volume. 3.3. CO2 breakthrough testing
Fig. 3. Comparison of regeneration techniques.
Breakthrough tests ran at 5 Torr CO2 and 16 SLPM were run after using different regeneration techniques to prove it as the least time consuming technique. 3.2. Packing the sorbent canister 3.2.1. Packing grace Davidson zeolite 5A The sorbent pellets are packed according to a standard procedure so that differing results are due to differences in packing density. In general, it is the goal that the bed is packed as densely as possible. However, it is important that the fragile sorbents not be damaged in packing. In particular, the zeolite pellets can fracture under too much force. It is important to avoid creating fines because the dust is harmful to mechanical equipment downstream of the adsorption beds. The packing method was developed by starting with a Desiccant/Adsorbent Bed Packing Procedure developed by Honeywell Corporation and then adapting it for the experimental test article. The method is as follows: 1. Pour 12 cm3 of sorbent pellets into the test article. 2. Settle the particles by vibrating the test article on a vibration table for 4 min and then place a 115 g mass on top of the pellets and vibrate at a slightly higher vibration level for 1 min. 3. Repeat (1) and (2) until the packing depth of 5 in. has been reached.
3.2.2. Packing UOP OXYSIV MDX Additional packing tests were done to find how to optimally pack the test bed with the Honeywell UOP OXYSIV MDX material. The following method was used to pack the test bed because it was shown to have the best results. 1) In a drybox, put the material that has been fully desorbed and the test bed. 2) Pour material in test article at a rate of 1 mm/sec or slower until designated volume is reached, while simultaneously vibrating the material at a level just below that in which bulk movement is induced.
Sorbent capacity can be tested by running a Breakthrough test. In a Breakthrough test, nitrogen and carbon dioxide are passed through the bed and the time it takes to completely load the bed with carbon dioxide is measured. The bed temperatures are also measured to quantify the heat of adsorption that is generated. To run the test, the bed is first pre-heated with nitrogen. The nitrogen gas is pre-heated by running it through a heat exchanger. This is done to get the bed at a constant temperature that is slightly above ambient so any fluctuations in temperature that are seen are not due to fluctuations in ambient temperature. Once the bed is heated, carbon dioxide can then be mixed with the inlet nitrogen stream and heated as well. After heating the gas stream will then be sampled and a measure of its dew point and percent CO2 will be taken. The remainder of the flow will pass through the bed. At the exit of the bed, the flow will again be sampled and its dew point and a measure of its percent CO2 will be taken. The remainder of the flow will exit to ambient. The flow of CO2 can either be held at a constant percent pressure or a constant mass flow. When the flow of CO2 is held at a constant percent pressure, the mass flow is constantly fluctuating with ambient pressure fluctuations. Likewise, when the mass flow is kept constant, the percent pressure of carbon dioxide constantly fluctuates with ambient pressure fluctuations. On either case, the percent of CO2 entering the bed will stay relatively constant. Breakthrough occurs when the bed is completely saturated with carbon dioxide and can no longer take on any more. The time at which the column hits breakthrough will become apparent when the exit CO2 analyzer measures the same percent carbon dioxide as the inlet CO2 analyzer. At this point, the carbon dioxide stream can be shut-off and the bed will be purged with nitrogen. This will cause desorption and the inlet CO2 analyzer should measure 0 percent carbon dioxide instantly where the exit CO2 analyzer should measure the CO2 that is coming off of the bed. Bed temperatures will also be measured during this adsorption/desorption process. During adsorption the temperatures should rise and once breakthrough occurs, Table 1 Grade 522 information. Material:
Grade 522 8 12
Manufacturer: Lot no. Material code: Place of manufacture: Ingredients
Grace-Davidson 9 52208350201 China Sodium aluminosilicate, potassium sodium aluminosilicate or calcium sodium aluminosilicate, clay, quartz, water
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
the temperatures in the bed should fall back to steady state. Then as desorption occurs, the bed temperatures should get colder in the opposite fashion as they did during adsorption. Table 1 3.4. Calculations In a CO2 Breakthrough test, the nitrogen/ carbon dioxide gas mixture is passed through a Sable Systems CO2 Analyzer, where the partial pressure of CO2 is measured, as well as the total gas pressure. Using these measurements, the mole fraction of CO2 is calculated. It is assumed that the mole fraction of CO2 in the gas stays constant. In the entrance and exit of the bed, the total pressure is measured, so using these total pressure measurements, along with the calculated mole fraction of CO2 in the gas stream, the partial pressure of CO2 in the bed inlet and exit is calculated. Using the ideal gas law, a plot of the number of moles of CO2 being taken on by the bed at each moment in time is made. This plot is integrated to find the number of moles of CO2 the bed has taken on. Knowing the true mass of sorbent in the bed, the sorbent’s capacity for CO2 [mol CO2/ kg sorbent] is calculated. Fig. 4 is a representation of the system, showing where measurements are taken. To compare the results, the Monte Carlo Method [8] was used to calculate the uncertainty in the carbon dioxide capacity calculation. Table 2 4. Results and discussion Two pelletized sorbents are compared in this section. Grace-Davidson Grade 522 zeolite 5A (Grade 522) and Honeywell UOP OXYSIV MDX (LiLSX) were tested for their CO2 capacity in an atmosphere like that of the ISS. Breakthrough tests were run to establish the capacities of these materials. All materials were tested under four different test conditions as shown in Table 3. Grade 522 (Mesh Sizes: 8 12 and 20 50) and LiLSX (Mesh Size: 20 50) are compared. Figs. 6 and 9 compare the material adsorption breakthrough curves. Fig. 7 and 10 show desorption breakthrough curves. Fig. 8 compares the temperature profiles of the two materials. In Fig. 5, all materials are compared at 5 Torr CO2 and 16 SLPM. In this plot, it can be immediately seen that the LiLSX material has a considerably larger capacity for carbon dioxide. When looking at the Grade 522 and LiLSX, at the same mesh size, it can be seen that the mass transfer coefficients of the two materials are approximately the same, indicated by the similar slopes of the breakthrough
43
curves. However, when breakthrough curves of the Grade 522 at the two sizes are compared, it is seen that the breakthrough curve slope significantly increases when the material size is decreased. Therefore, decreasing a material’s size increases its mass transfer capabilities. The only thing that was changed when the material was crushed was the macropore size of the material. This indicates that the Grade 522 material is limited by macropore diffusion. The macropore is defined as the space in the pellet between
Table 2 UOP OXYSIV MDX information. Material:
OXYSIV MDX 20 50 (LiLSX)
Manufacturer: Lot no. Material no. Place of manufacture: Ingredients
UOP LLC, A Honeywell company 2010010918 80555-999 China Silica oxide (synthetic), aluminum oxide, lithium oxide, potassium oxide, sodium oxide, water, quartz
Table 3 Test conditions. Test condition:
1
2
3
4
Carrier gas flow rate [SLPM] Approximate CO2 partial pressure [Torr] Actual CO2 partial pressure [kPa]
16 5 0.67
16 2.5 0.31
8 5 0.67
8 2.5 0.31
Fig. 5. Material Comparison: adsorption cycle, 5 Torr (0.67 kPa) CO2, 16 SLPM.
Fig. 4. Schematic of test system with measurement locations.
44
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
Fig. 6. Comparison of bed exit temperatures. Adsorption cycle, 5 Torr (0.67 kPa) CO2, 16 SLPM.
each crystal, while the micropore is the interstitial space between each crystal where the CO2 molecule is physically adsorbs. Fig. 6 is very interesting because it gives insight into the heat of adsorption characteristics of different materials, and shows how material size can affect the heat of adsorption. Grade 522 (8 12 mesh) is shown to have a higher mass transfer resistance because the exit temperature curve is slower and rounded because the uptake is smeared over time. Its temperature rises first because it initially takes up CO2. It is noted here that the smaller materials show a sharp uptake in CO2 because the exit temperature curves rise sharply. However, the Grade 522 curve looks like a smaller version of the LiLSX curve. Also, when comparing the Grade 522 material at different sizes, it looks as though both contain the same area, although they are sloped differently. This indicates that the material size plays a large role in the mass transfer characteristics of a sorbent, but not necessarily its overall capacity. The increase in mass transfer for the smaller materials is because the macropore diffusion is increased for a smaller pelletized sorbent. This implies that a material can take on the same amount of carbon dioxide regardless of size; size just affects the rate at which it can take on that amount of carbon dioxide. In Fig. 7, the desorption curves of the three materials are compared. From this graph, it appears that the Grade 522, 8 12 mesh and the LiLSX material release the carbon dioxide at the same rate. It should be noticed that when Fig. 5 is compared to Fig. 8, there is a bigger time difference between the breakthrough time of the LiLSX material and the Grade 522, mesh 20 50 for the lower carbon dioxide partial pressure. This indicates that the LiLSX material can adsorb more CO2 than the Grade 522 material at lower CO2 partial pressures. In Fig. 9, desorption curves of the three materials are compared for 2.5 Torr CO2 and 16 SLPM. From this graph, it appears that the Grade 522 (20 50), releases the carbon dioxide the fastest and the LiLSX releases the carbon dioxide the slowest.
Fig. 7. Material comparison: desorption cycle, 5 Torr (0.67 kPa) CO2, 16 SLPM.
Fig. 8. Material comparison: adsorption cycle, 2.5 Torr (0.31 kPa) CO2, 16 SLPM.
Fig. 9. Material comparison: desorption cycle, 2.5 Torr (0.31 kPa) CO2, 16 SLPM.
Table 4 gives the results of each test. For each sorbent, the mass of sorbent used is given. The conditions as to what carbon dioxide partial pressure was used and the carrier gas (nitrogen) flow rate in standard liters per
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
Table 4 Summary of results. Sorbent
Mass [g]
Grace Grade 125 522 5A, Mesh: 8 12
Conditions [Torr]/ [SLPM] 5/16 5/8 2.5/16 2.5/8
UOP LiLSX, Mesh: 20 50
102
5/16 5/8
2.5/16
2.5/8 Grace Grade 522 5A, Mesh: 20 50
106
5/16 2.5/16
Mean capacity 95% [MCM] [mol CO2/ Confidence interval [%] kg sorbent] 1.185 1.048 0.9756 0.9871 0.5955 0.5985 0.5738 0.5569 1.9721 2.004 1.7364 1.7811 1.7795 1.5035 1.479 1.5321 1.3193 1.3246 1.3915 1.6707 0.8451 0.9806
5.74% 5.31% 3.59% 2.51% 3.76% 4.14% 3.56% 3.41% 2.59% 2.63% 2.05% 2.75% 3.61% 1.94% 1.68% 2.25% 2.70% 2.58% 4.41% 4.85% 3.74% 3.67%
minute (SLPM). The Monte Carlo Method [8] was used to calculate the mean capacity with the corresponding 95% Confidence interval. 5. Conclusion In this work, materials were compared for their potential to remove carbon dioxide from a space-habitat where resupply is difficult and under the limitations of minimizing mass, power, and volume. In the future, NASA would like to build a lunar or Martian habitat for astronauts to live. A very reliable and efficient life support system is needed to remove carbon dioxide from the habitat. This is important because at CO2 pressures above 5 Torr, astronauts start to suffer from headaches, and because the CO2 can be recycled downstream to make water or oxygen. The materials that were analyzed are Grace-Davidson Grade 522 5A (Grade 522) and UOP OXYSIV MDX (LiLSX). The Grade 522 material used was an 8 12 mesh size and the LiLSX material used was a 20 50 mesh size. In order to compare the materials without pellet size confounding the results, the Grade 522 material was crushed to a 20 50 mesh size and tested again. This also allowed us to compare the Grade 522 material at the two sizes to evaluate how pellet size affects CO2 capacity and mass transfer capabilities. The LiLSX material is a relatively new commercially available sorbent marketed for oxygen concentration. We wanted to test LiLSX for its CO2 capture potential for future spacecraft use, so it was tested along with the current standard sorbent for CO2 capture, Grade 522. A test stand was developed to uniformly compare the two materials. The test stand composed of a test article, a heat exchanger, nitrogen and carbon dioxide gases, mass flow controllers, and sample lines upstream and downstream of the test article which composed of sample pumps,
45
CO2 analyzers, and dew point analyzers. The material being tested was packed in the center of the test article. Within the test article, temperature and pressure measurements were made upstream and downstream of the packed bed. Breakthrough tests were run at two flow rates (8 SLPM and 16 SLPM) and two CO2 partial pressures (2.5 Torr and 5 Torr). During a breakthrough test, a mixture of nitrogen and carbon dioxide was passed through the bed and the amount of carbon dioxide that is adsorbed by the packed bed is calculated using the nitrogen and carbon dioxide flow rates and the pressure of carbon dioxide and temperatures at the inlet and exit of the bed. At the beginning of a test, all carbon dioxide that is being passed through the bed will be captured by the sorbent and the exit carbon dioxide analyzer will show a CO2 partial pressure of zero. Temperatures in the bed will begin to rise because adsorption is an exothermic process. Breakthrough begins when the bed can no longer hold all of the CO2 being passed through it and the exit CO2 analyzer will begin to show a positive CO2 partial pressure. The CO2 partial pressure in the exit stream will continue to rise until breakthrough is complete, at which the bed is completely saturated and cannot take on any more carbon dioxide. At this point the exit CO2 pressure will equal the inlet CO2 pressure and bed temperatures will return to steady state. Ultimately, the materials’ capacity for carbon dioxide is compared. The capacity is how much CO2 a defined amount of material can take on. This is usually expressed in units of mol CO2/ kg sorbent. The number of moles of carbon dioxide entering and exiting the bed was calculated and the difference is how much CO2 was loaded on to the bed. The capacity is calculated by dividing the number of moles loaded on to the bed by the mass of the sorbent in the packed bed. An uncertainty analysis was performed using the Monte Carlo Method [8] so that the capacity measurements for each sorbent could be accurately compared. An overall results table is shown on the previous page, in Table 5. In this table, the CO2 capacity and the results from the uncertainty analysis are given for the two sorbents under the four different sets of conditions. From examining this table, it is shown that the capacity of a material is increased slightly by decreasing its pellet diameter. Then when comparing LiLSX and Grade 522, at the same mesh sizes, it can be seen that the LiLSX material performs better at both carbon dioxide partial pressures, when tested at the higher flow rate. However, the increase in capacity is higher at the lower carbon dioxide partial pressure. The results are least consistent for the Grade 522 Mesh 20 50 tests. The uncertainty is not large enough to note these as statistically the same. These differences results from different pressure drop readings in the bed under the same conditions. The bias error is very small for the pressure drop is very small. These results indicate that the error in this reading is probably actually larger. This error was the only error that came from the NASA calibration lab. In the future, I would send this instrument back to the manufacturer for a proper inspection and manufacturer recalibration. From looking at the results, I would recommend using the LiLSX material in future spacecraft applications because it has been proven to have a higher capacity for CO2 than the standard, Grade 522. If it is used for spacecraft applications,
46
E.M. Mattox et al. / Acta Astronautica 86 (2013) 39–46
it will have to be scaled up to minimize the power requirements for pushing air through the bed.
Acknowledgments
[2] [3] [4]
This work was supported by the NASA Marshall Space Flight Center under Cooperative Agreements NNM05AA22A and NNM11AA01A. References
[5] [6] [7] [8]
[1] J.C. Knox, New methods for the adsorption of carbon dioxide and water vapor from manned spacecraft atmospheres: applications and
modelling, in: Proceedings of the Comsol Conference, Boston, MA, 2007, Key Note Speech. Benitez Jaime, Principles and Modern Applications of Mass Transfer Operations, 2nd ed. John Wiley & Sons, Inc., New Jersey, 2009. Duong D. Do, Adsorption Analysis: Equilibria and Kinetics, Imperial College Press, London, 1998. James C. Knox, David Howard, Clearing the Air: Life Support for Space Exploration, Comsol. News (2008) 4–6. Douglas M. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley and Sons, Inc, New York, 1984. M. Douglas LeVan, G. Carta, Adsorption and Ion Exchange, Perry’s Chemical Engineering Handbook, 8th ed, print, pp.16-1 –16-15. A Suggested Laboratory Purge Activation Procedure, uop.com, Universal Oil Products: Adsorbents & Specialties, n.d., Web, 7 September, 2010. Hugh W. Coleman, W. Glenn Steele, Experimentation, Validation, and Uncertainty Analysis for Engineers, 3rd Ed, John Wiley & Sons, Inc., 2009.