Subscriber access provided by Stockholm University Library
Better Absorbents for Ammonia Separation Mahdi Malmali, Giang Le, Jennifer Hendrickson, Joshua Prince, Alon V. McCormick, and Edward L. Cussler ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.7b04684 • Publication Date (Web): 30 Mar 2018 Downloaded from http://pubs.acs.org on March 31, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Sustainable Chemistry & Engineering
Better Absorbents for Ammonia Separation Mahdi Malmali, Giang Le, Jennifer Hendrickson, Joshua Prince, Alon V. McCormick, and E. L. Cussler University of Minnesota, Chemical Engineering and Materials Science, 421 Washington Ave SE, Minneapolis, MN 55455 corresponding author:
[email protected] ABSTRACT: Making ammonia from renewable wind energy at a competitive price may be possible if the conventional ammonia condenser is replaced with an ammonia absorber. Such a process change, shown in the TOC figure, requires an ammonia selective absorbent. Supported metal halide sorbents for this separation display outstanding dynamic capacity close to their equilibrium thermodynamic limits. Alkaline earth chlorides and bromides supported on silica and zeolite Y are the most promising. MgCl2-Si with 40% loading on silica shows the highest observed capacity, 160 mgNH3/gsalt at 150 °C and 4 bar. Overall, cations with smaller atomic numbers show more affinity to ammonia; bromides hold ammonia more strongly than chlorides. Different solvents and metal halide mixtures do not show significant changes in the absorption capacity. These absorbents can be incorporated to ammonia reaction-absorption syntheses to achieve faster production rates. KEYWORDS: Ammonia, Absorption, Metal Halides, Wind Energy
Introduction Ammonia is one of the most important chemicals in the world, the key to the “Green Revolution.”1,2 Without synthetic ammonia, about two billion of the world’s population would starve.3 This ammonia is almost entirely produced by the Haber-Bosch process, carefully optimized by a century of effort. This catalytic process produces a mixture of hydrogen, nitrogen, and carbon dioxide by burning fossil fuels with air and water. After the carbon dioxide is removed, the hydrogen and nitrogen are reacted at above 400 °C and 150 bar.4 The high temperature is needed to break the strong nitrogen bond, and the high pressure is used to drive the reaction towards higher conversion. Conventionally, the ammonia is separated by condensation, and the unreacted hydrogen and nitrogen are reheated and recycled to the reactor. The process works well and is a mainstay of commodity chemical manufacture.4 However, the process does have disadvantages. It requires fossil fuels and is, by itself, responsible for 2-3% of global carbon dioxide emissions.5–7 The high pressure and high temperature required mean capital expenses are large, which can limit production in developing countries. The catalyst works well but still requires a temperature which, in theory, is higher than necessary for a reaction that has a negative free energy difference and so should run
ACS Paragon Plus Environment
ACS Sustainable Chemistry & Engineering 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
spontaneously. These disadvantages to the current process are serious enough to sustain continuing research, especially for new catalyst recipes.8 In this work, we seek to improve ammonia synthesis not by finding better catalysts but by discovering better separations. We are especially interested in an ammonia synthesis which does not require fossil fuels but uses wind-generated electricity to make hydrogen from electrolysis of water and uses nitrogen by pressure swing absorption of air.9–12 Experiments with an existing, small scale pilot plant show that these sources of hydrogen and nitrogen are feasible in a windbased process. However, the new separation described here can work either in a conventional or in a wind-based process. Both processes can be idealized as shown in the TOC figure.12 The conventional one, shown on the left, consists of a catalytic reactor to make ammonia, a condenser to remove the ammonia, and a pump to return unreacted hydrogen and nitrogen to the reactor. The alternative, shown on the right, also has a reactor and a pump, but replaces the condenser with a bed of solid absorbent. The use of an absorbent that captures the ammonia in a solid crystal is superior to an adsorbent in two ways: it captures the ammonia at higher temperatures, and it is more selective.13,14 However, the absorption may take longer to achieve its full capacity.15,16 In this, we are referring to the take up of ammonia by salts as “absorption.” By this, we assert that the ammonia molecules are going into the salt crystal, and not just being retained on the crystal surface. This incorporation into the solid is responsible for the selectivity of the absorption; ammonia is absorbed, but nitrogen and hydrogen gases are not. The absorption can take place even at temperatures over 300oC, and can exceed four moles of ammonia per mole of salt at lower temperatures. The process is much more like hydration, where ammonia takes the place of water to form “ammoniates.” The amount and the selectivity in these absorptions are much greater than those expected for adsorption, where the gas being separated adheres to the crystal surface, and does not penetrate significantly into the solid.
At the same time, we should stress that these gas-solid absorptions can be significantly different from gas-liquid absorptions. In the gas-liquid case, the gas phase and the liquid phase are usually each well mixed. In the gas-solid case, the gas phase is again well mixed, but the solid phase may not be. This means that the kinetics of absorption is also important: its rate can be proportional to the area of crystals even though the majority of the ammonia is located not at the surface of the crystal but in the bulk. In this paper, however, we limit the majority of the discussion to the amount of uptake, and defer the investigation of kinetics to another paper.
This paper seeks the best absorbent materials for ammonia separation at high temperature. The materials developed in the past range widely. Ammonia sorption on activated carbon (2-20 mgNH3/gsorbent)17, metal organic frameworks (MOFs—0.27-105 mgNH3/gsorbent)18,19, covalent
ACS Paragon Plus Environment
Page 2 of 27
Page 3 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Sustainable Chemistry & Engineering
organic frameworks (272 mgNH3/gsorbent)20, MCM-41 (136 mgNH3/gsorbent)20, alumina (34 mgNH3/gsorbent)21, silica gel (85 mgNH3/gsorbent) 21, and zeolites (130 mgNH3/gsorbent)21 have all been studied. (Capacities listed for these materials are at room temperature and 1 bar partial pressure of ammonia.) These materials are excellent for removal of toxic gases from gaseous streams at room temperature; however, they start to release adsorbed ammonia at temperatures as low as 40 °C. Additionally, cyclic ammonia uptake and release from most of these materials can result in decreased ammonia capacity, a consequence of changing meso- or micro-porous geometry.20 Almost none of the above mentioned materials are capable of separating ammonia at temperature above 200 °C. Hence, they are not attractive options for ammonia separation (high temperature) process. Separation and storage of ammonia on different inorganic materials are also well studied. Metal halides13,22,23, metal borohydride,24 and metal fullerides25 are some of the examples. These materials have excellent capacity for ammonia storage and have been proposed to be ideal for on-board hydrogen storage, since it is safer to transport absorbed ammonia in the solid form than liquid hydrogen. While borohydride and fulleride data are incomplete, these materials may not completely separate ammonia and hydrogen. Additionally, the ammonia capacity at high temperatures (> 200 °C) is low, and some of these materials may undergo a phase change upon absorption.22 Among these inorganic sorbents, metal halides are the most promising, though bulk halides are not always stable, and achieving equilibrium capacity is difficult.16,26 One strategy to improve the performance of metal halides for ammonia separation is to disperse a thin layer of these materials on porous supports with large surface areas. Now, diffusion of the ammonia into the dispersed metal halide clusters is faster and may not limit the uptake of ammonia.16,27–30 Barpaga and LeVan30 functionalized carbon silica composites to improve ammonia and sulfur dioxide separations. Aristov showed supporting metal halides on different supports can improve separations.27,31 Wagner et al.16 reported that 2% loaded MgCl2 on alumina is stable and outperforms bulk MgCl2. Following these leads, we evaluate in this paper the performance of different supported metal halides with high salt loadings for ammonia separation as described below.
Experimental Sources of Materials Calcium bromide hydrate (98% purity) was purchased from Acros Organics (Geel, Belgium). Magnesium bromide hexahydrate (99% purity) was obtained from ChemImpex International (Wood Dale, IL). Strontium bromide hexahydrate (95% purity), anhydrous strontium chloride (95% purity), and zeolite Y (5.1:1 mole ratio SiO2:Al2O3) were purchased from Alfa Aesar (Ward Hill, MA). Anhydrous calcium chloride (97% purity), anhydrous magnesium chloride (98% purity), aluminum oxide (98% purity), diatomaceous earth (CAS No. 68855-54-9), kaolinite (CAS No. 1318-74-7), hydrophilic bentonite (CAS No. 1302-78-9), silica gel technical grade (pore size 60 Å—CAS No. 112926-00-8) were purchased from Sigma Aldrich (St. Louis, MO). Ultrahigh purity nitrogen and ammonia gas cylinders were obtained from Matheson (Eagan, MN). Two hundred proof ethanol (ACS grade) and 99.8% HPLC grade
ACS Paragon Plus Environment
ACS Sustainable Chemistry & Engineering 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
methanol were purchased from Pharmco-Aaper (Shelbyville, KY) and Sigma Aldrich (St. Louis, MO), respectively.
Materials Characterization Past studies of ammines are occasionally compromised by using not pure salts but partial hydrates. Thermogravimetric analysis (TGA) is a proven tool for studying water release from hydrates. Because ammonia has some characteristics similar to those of water, TGA is also helpful in understanding desorption of ammonia. Results for the dehydration of bromides are shown in Figure 1. High resolution TGA was conducted on a TA Instruments Q500. Samples were heated from 22 °C to 550 °C at a dynamic rate of 3 °C min-1, with 100 SCCM nitrogen flow as the purge gas. MgBr2 and SrBr2 were received as hexahydrates; while CaBr2, nominally received as a monohydrate, also behaved as a hexahydrate. Magnesium bromide shows three peaks at 140, 220, and 375 °C, which indicate three stages of water release (hexa-, di-, and mono-hydrates).14 Calcium bromide shows a similar trend, with peaks at 50, 95, 130 °C. Lower temperature desorption of water was anticipated for calcium bromide hydrate: according to the literature, we should expect at least four peaks (octa- to hexa-, hexa- to di-, dito mono-, and mono- to an-hydrous) for calcium bromide hydrate.14 However, we did not observe the octa-hydrate, probably because 1 bar partial pressure of water is required to form the mono-amminated calcium bromide at 25 °C.32 Compared to magnesium bromide and strontium bromide, we did see a relatively small peak at 50 °C, indicating that not all calcium bromide crystals are hexa-hydrated. We observe a similar trend for SrBr2 but do not detect the hexahydrate structure because it is unstable at room temperature and humidity. With 20% reduction in the weight, CaBr2 has the least water in its hydrated form, while MgBr2 loses 70% of its initial weight. X-ray powder diffraction (XRD) results shown in Figure 2 indicate dried CaBr2 and SrBr2 are obtained after drying, consistent with the weight loss observed. However, for MgBr2, we observe more than 80% decrease in the sample weight. Dehydration of six molecules of water cannot be responsible for such change. XRD spectrum for MgBr2 shows the final product of drying magnesium bromide hexa-hydrate is not pure magnesium bromide. Magnesium bromide may partially oxidize to form magnesium oxide, which is consistent with the weight loss observed in TGA results. Supported salts Making supported salts can involve pretreating both the salts and the supports. Chloride salts were used as received. Bromides were dried in a tube furnace, under vacuum at 25 ± 2 in Hg, with a nitrogen sweep flowing at approximately 50 standard cubic centimeter per minute (SCCM). The oven was initially heated at a rate of 5 °C min-1 to 210 °C, and then at 0.5 °C min-1 to 400 °C. To pretreat the supports, the oven was heated at 15 °C min-1 to 400 °C, and then 5 °C min-1 to 450 °C. The drying was continued for 2 hours at this temperature. Supported samples were prepared in a two-neck round bottom flask with a condenser. Unless otherwise noted, 10 g of dried metal halide was mixed with 200 mL of 50:50 ethanol/methanol in the flask. The mixture was then heated with an oil bath to the boiling point of methanol-ethanol mixture. The solution was stirred for at least an hour under total reflux, with nitrogen bubbling within the mixture. Then 15 g of porous support was added to the hot flask while hot. After the
ACS Paragon Plus Environment
Page 4 of 27
Page 5 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Sustainable Chemistry & Engineering
suspension was stirred continuously under the same conditions for another three hours, the solvent was vacuum-evaporated. The solid impregnated material was transferred to a petri dish and stored in a vacuum oven at 150 °C and 30 ± 2 in Hg vacuum overnight. It was then dried in a tube furnace, using the same conditions mentioned above. Absorption The absorption apparatus is illustrated in Figure 3. N2 and NH3 volumetric flow rates were controlled with Brooks Instrument mass flow controllers. Unless otherwise noted, 10.0 g of sorbent was packed into a six-inch piece of half-inch Swagelok stainless steel tubing. Two grades of stainless steel wool were inserted in the ends of the column to confine the packing materials, and a thermocouple was positioned in the center of the packed bed. The column was wrapped with heating tape and insulation, and pressure gauges were placed upstream and downstream. An inline needle valve was used to control the system’s pressure. Nitrogen flow at 50 SCCM and ammonia flow at 10 SCCM were pumped through the bed, and the absorber exhaust was directed toward a gas chromatograph equipped with a TCD detector (Agilent Technologies, Santa Clara, CA). For desorption tests, the exhaust was connected to a vacuum pump or fume hood. As illustrated in Figure 4, supported metal halides display sharp and reproducible breakthrough curves, while bulk unsupported metal halides can show dispersed and irreproducible breakthroughs. Since all breakthrough curves were similar, we report the apparent capacity of absorbent as only two parameters, the coordination number and the apparent sorbent capacity. The coordination number is calculated from the breakthrough time BT and the amount of the salt loaded into the support: :
× ,"## %&'() * +(), ×-# , $ 0 ./*
(1)
The sorbent capacity is calculated based on the cumulative weight of support and salt: × ×56
,"## 1 234: 789:; ?9@; @=9AB ×5C B
(2)
where the BT is defined as the time at which 5% of the inlet concentration exits.
Results Here, we report our efforts to find better absorbents for ammonia at conditions of lower pressure but closer to temperatures used in the existing ammonia synthesis. These experiments explore when absorption may be a superior separation to ammonia condensation now used in conventional, fossil-fuel fired, ammonia synthesis. They may also aid evaluation of smaller scale ammonia manufacture, both to supply fertilizer locally to farms and to capture stranded wind energy as a liquid carrier for hydrogen (later to be used in fuel cells).12 The important results are those for alternative supports for absorption, for varied ratios of support to salt mass, and for alternative absorption chemistries. Other results include the effects
ACS Paragon Plus Environment
ACS Sustainable Chemistry & Engineering 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 27
of feed flow, different absorbent preparations, and mixtures of absorbents. Details of results follow.
Bulk metal halides Metal halides have extraordinary capacity for ammonia storage.14,15,33–35 However, achieving the equilibrium capacity is not straightforward.13,16 Results summarized in Table 1 indicate metal halide salts in their bulk form are not suitable ammonia absorbent materials for our application. All these experiments used feeds of N2:NH3 = 5:1 and a total pressure of 4 bar. The absorption capacity does not exceed 6 mgNH3/gsalt. This dynamic capacity is 1-5% of the equilibrium capacity.12,16 Moreover, under absorption conditions, these metal halides are unstable.16,26 After each loading/unloading cycle, the breakthrough of the subsequent cycle decreases. Table 1. Ammonia Absorption Using Pure Metal Halides. All experiments used feeds of N2:NH3 = 5:1 and 4 bar, with a total flow of 60 SCCM. The amounts absorbed are much less than those expected at equilibrium. Sorbents
T (°C)
MgCl2
150 200 300 150 200 300 150 200 300 150 200 200 150 200 300 150 200 300
CaCl2
SrCl2
MgBr2
CaBr2
SrBr2
BT time (min) 5.1 2.3 1.6 1.4 0.5 0.0 0.21 0.10 0.00 0.00 7.58 5.40 0.46 0.49 0.27 3.11 1.03 0.44
Coordination Number (molNH3/molsalt)
Sorbent Capacity (mgNH3/gsalt)
0.03 0.01 0.01 0.01 0.00 0.00 0.01 0.00 0.00 0.00 0.06 0.04
7.2 1.7 1.2 1.1 0.4 0.0 0.0 0.0 0.0 7.4 6.1 4.5
0.06 0.04 0.02 0.01 0.00 0.00
5.2 3.7 2.8 0.0 0.0 0.0
Different Supports We next investigated the effect of different porous supports on the performance of composite sorbents. For these studies, we used CaBr2 as our model metal halide. Figure 5 summarizes breakthrough studies performed at 4 bar (0.66 bar partial pressure of ammonia) on different supports; these same results are detailed in Table 2. The data are at three different temperatures. The third column shows the breakthrough times; the fourth reports the
ACS Paragon Plus Environment
Page 7 of 27 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Sustainable Chemistry & Engineering
coordination number inferred from these times; and the last column shows the sorbent capacity. This capacity is the total amount absorbed minus that absorbed by the pure support, which is always less than 10% of the total. There are two main conclusions to be drawn from this table. First, the results support the contentions of Sharanov,27,28 van Hassel,29 and Wagner16 that supporting metal halides on porous supports improves their apparent capacity and the structural stability. More specifically, the amounts absorbed are much greater than those we have measured for the pure halides, summarized in Table 1. Second, alumina (Al), kaolinite (Ka), and diatomaceous earth (DE) are not effective for supporting metal halides: the apparent capacity for ammonia never exceeds 9 mgNH3/gsalt for these materials. Of course, the equilibrium capacity of CaBr2 must be the same in all these experiments, independent of the support used. The apparent capacity changes because of the particular preparation and exposure time used. The reasons are unclear. The differences will not be due to changes in the diffusion coefficient, because diffusion in all cases is of ammonia into calcium bromide. They are not the result of an altered partition coefficient between the gas and the salt. The most likely explanation is that the surface area per mass of salt is altered by how the salt spreads out onto the support. This is supported by BET areas measured for each sample, as summarized in the last column of Table 2. Both samples supported on Si and Ze give large BET surface areas. Originally, Si and Ze have surface areas equal to 540 and 985 cm2/g, respectively; impregnating salts into these supports leads to formation of materials with relatively large surface areas (399 and 541 cm2/g, respectively). While this implies that supports with larger surface areas will form more absorbent structures, results for alumina contradict this notion. Alumina’s BET area is 184 cm2/g, but the area for alumina impregnated with salt is only 1 cm2/g. Other factors, like surface tension, must also be involved. This is indirectly supported by the observation that the pure salt goes from a free-flowing powder to a concrete mass after exposure to ammonia. The salt is apparently softened by the ammonia and seems to fuse. Supports like silica seem to reduce this fusion. Some evidence for this comes from the micrographs shown in Figure 6. In previous work,16 we suggested some supports may be able to disperse the salt as nanocrystals inside small cracks appearing throughout the support. Figure 6 shows a similar phenomenon may be involved here, but it does not prove this. The CaBr2 on Si and Ze supports achieves a high coordination number, even for temperatures as high as 300 °C, close to the theoretical capacity of CaBr2 at these temperatures. At 150 °C, the sorption capacity for CaBr2-Ze is less than CaBr2-Si, while at 200 and 300 °C, it is higher. Still, because silica is cheap and showed a more consistent absorption capacity at the temperatures studied, we decided to use it as our base case for studying different salts.
Table 2. Ammonia Absorption for Calcium Bromide Supported on Different Substrates. Silica and zeolite Y work best. The capacity shown is that observed minus that for the support alone. All experiments used feeds of N2:NH3 = 5:1 and 4 bar.
ACS Paragon Plus Environment
ACS Sustainable Chemistry & Engineering 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Sorbents
T (°C)
CaBr2-Al
150 200 300 150 200 300 150 200 300 150 200 300 150
CaBr2-DE
CaBr2-Ka
CaBr2-Si
CaBr2-Ze
BT time (min)
200 300
Coordination Number (molNH3/molsalt)
3 1 0 2 4 2 7 3 1 80 54 21 48 50 59 57 30
0.02 0.00 0.00 0.01 0.06 0.03 0.11 0.05 0.01 1.66 0.85 0.98 0.95 0.99 1.27 1.23 0.65
Sorbent Capacity (mgNH3/gsalt)
1.89 0.36 0.39 0.96 5.25 2.55 9.65 4.24 0.80 141.67 72.66 83.39 81.26 84.60 107.95 104.63 55.59
Page 8 of 27
BET Surface Area (cm2 /g) (Support’s Actual BET area) 1 (184) 2 (85)