The invention relates generally to membranes and methods for separating a gas from a gas stream, and particularly for separating CO2 from a gas stream.
The separation of CO2 from a gas stream is a critical step in the reduction of greenhouse emissions from fossil fuel-based combustion processes. An amine process is used commonly in power plants to scrub the fuel or exhaust gas stream. However, this approach is both energy and capital intensive because the amine process involves cooling the gas stream before scrubbing. Membranes capable of separating CO2 at temperatures above 200° C. could be used in lieu of the amine process in existing plants. They could also be used in advanced integrated gasification combined cycle (IGCC) plants to improve efficiency. To be successful, a membrane must meet two sets of requirements. First, the membrane must be able to selectively separate CO2 from a gas stream. In particular, it is desirable to separate CO2 from H2 in the fuel gas stream or to separate CO2 from N2 in the exhaust gas stream. To achieve separations with a porous membrane, it is often preferable to have reverse selectivity. Reverse selectivity is selectivity in which the heavier gas is enriched relative to the level expected for Knudsen selectivity. Second, the membrane must have an operating temperature above 200° C.
For example, a high temperature membrane having reverse selectivity in separating CO2 and H2 must exhibit mechanical and functional stability up to 500° C. and CO2/H2 selectivity greater than 10. In addition, CO2 permeabilities of at least 1000 Barrer are desirable. There are no membranes currently available that meet these requirements.
Since CO2 is heavier than the other components of interest in the gas stream, Knudsen diffusion is not a viable mechanism for separation. Knudsen diffusion describes the flow of gas through a membrane in which the pore size is small compared to the mean free path of the gas. The Knudsen diffusion rate is inversely proportional to the molecular weight of the gas. A membrane relying only on Knudsen diffusion would have a CO2/H2 selectivity of 0.21. Instead, transport must occur through alternate mechanisms that enable the desired selectivity. For example, the most promising polymer membranes are based on a facilitated transport mechanism in which CO2 is selectively transported via amino groups. Those membranes exhibit selectivity of about 10 and permeability of 2000 Barrer at 180° C., but performance rapidly degrades above 180° C. due to dehydration of the membrane. Therefore, polymer membranes are not suitable at higher temperature.
Porous inorganic membranes have the capability for high temperature applications, and selectivity can be endowed through the mechanism of preferential adsorption and surface diffusion of CO2 along the pore walls. Based on this approach, CO2/N2 selectivity of ˜10 have been reported for zeolite, silica, and activated carbon membranes with permeabilities as high as ˜104 Barrer (at room temperature). Recent efforts to develop reverse selective membranes using this strategy have resulted in silica membranes having a measured selectivity of ˜5 to 7 with permeability of about 1000 Barrer at 40° C. (Moon, J. H., et al., Kor. J. Chem. Eng., 21, 477-487 (2004)). Up to this point, efforts to develop membranes with enhanced surface transport have focused on identifying a porous material which itself has suitable surface transport properties. The problem with this approach is the limited number of compositions available that satisfy both the structural requirement (well-defined pores) and the transport requirement (fast surface diffusion of CO2). Kusakabe et al. have prepared barium titanate (BTO) layers on porous alumina supports and found a CO2/N2 selectivity of 1.2 at 500° C. (J. Membrane Sci., 95, 171-177 (1994)). The expected selectivity from Knudsen diffusion is 0.8. However, the membranes contained structural defects in the form of 100 nm pinholes that limited CO2 selectivity.
Accordingly, there remains a need for membranes that can achieve CO2/H2 selectivity significantly higher than that achievable through Knudsen diffusion mechanisms at high temperatures.
It has been unexpectedly discovered that functionalized mesoporous membranes comprising a porous support layer, a well-defined porous separation layer and a surface coating that imparts the desired diffusion characteristics exhibit reverse selectivity in separation of CO2 from H2.
Briefly, in accordance with one embodiment of the present invention, a porous membrane is provided for separation of carbon dioxide from a fluid stream at a temperature higher than about 200° C. with selectivity higher than Knudsen diffusion selectivity. The porous membrane comprises a porous support layer comprising alumina, silica, zirconia or stabilized zirconia; a porous separation layer comprising alumina, silica, zirconia or stabilized zirconia, and a functional layer comprising a ceramic oxide contactable with the fluid stream to preferentially transport carbon dioxide.
In another embodiment, the present invention relates to methods for separating carbon dioxide from a fluid stream at a temperature higher than about 200° C. with selectivity higher than Knudsen diffusion selectivity. The methods comprise contacting a porous membrane according to the present invention with the fluid stream to preferentially transport carbon dioxide.
These and other features, aspects, and advantages of the present invention will become better understood when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:
The present invention relates to porous membranes for separation of carbon dioxide from a fluid stream at a temperature higher than about 200° C. with selectivity higher than Knudsen diffusion selectivity, and to methods for performing such separations using the membranes. For separation of CO2 from a H2-containing gas stream, selectivity due to Knudsen diffusion alone is 0.21. The porous membranes according to embodiments of the present invention enable reverse selectivity in such a separation, yielding selectivity greater than 2, particularly greater than 5, more particularly greater than 10 and most particularly greater than 20.
The porous membranes comprise a porous support layer comprising alumina, silica, zirconia or stabilized zirconia; a porous separation layer comprising alumina, silica, zirconia or stabilized zirconia, and a functional layer comprising a ceramic oxide contactable with the fluid stream to preferentially transport carbon dioxide.
In a particular embodiment, a membrane according to the present invention includes a porous support layer comprising Al2O3, porous separation layer comprising SiO2 and a functional layer comprising BaTiO3. In another embodiment, the membrane includes a porous support layer comprising Al2O3, a porous separation layer comprising Al2O3 and a functional layer comprising BaTiO3.
In yet another embodiment, the present invention relates to methods for separating carbon dioxide from a fluid stream at a temperature higher than about 200° C. with selectivity higher than Knudsen diffusion selectivity. The methods comprise contacting a porous membrane according to an embodiment of the present invention with the fluid stream to preferentially transport carbon dioxide.
Separation layer 120 is disposed on the surface of support layer 110 and is bonded thereto. In another embodiment (not shown), the separation layer is porous and is disposed within and at least partially fills the pores of the support layer, and is bonded thereto. In either embodiment, the separation layer provides additional mechanical stability to the membrane and ensures defect-free structures, being free of macroscopic pore defects that allow bypass flow and having high surface area, a narrow pore size distribution, and a well-ordered pore organization. The separation layer is chemically stable in a CO2/H2 (reducing) atmosphere and is thermally stable, i.e., there is no coarsening of the structure up to about 500° C.
Methods and processes for disposing a separation layer 120 on a support layer 110 are known in the art; some suitable methods are described by T. Tsuru (“Inorganic porous membranes for liquid phase separation,” Separation and Purification Methods, v30 (2), 191-220 (2001)). For example, the support layer may be fabricated using powder sintering with green bodies made from extrusion or tape casting, and the separation layer may be co-cast or co-extruded with the support layer and co-sintered. Alternately, the support layer may be sintered and the separation layer disposed on it in a subsequent processing step. Where the support layer has been prepared in a separate process, the separation layer may be fabricated using a wet chemistry process such as a sol-gel process, or a vapor deposition process such as chemical vapor deposition.
The separation layer comprises alumina, silica, zirconia or stabilized zirconia. In particular embodiments, it may be alumina or silica. A mesoporous structure may be produced using a surfactant-templated sol gel approach, on the surface of or within the pores of the support layer. See Kresge, C. T., et al., Nature, 359, 710-712 (1992), Yang, P., et al., Nature, 396, 152-155 (1998), and Ku, et al., J. Am. Chem. Soc., 127, 6934-6935 (2005).
The functional layer provides a continuous path for CO2 to move from one side of the membrane to the other, to enhance the flow of the gas through the membrane. Materials suitable as the functional layer include MgO, CaO, SrO, BaO, La2O3, CeO2, TiO2, HfO2, Y2O3, VOz, NbOz, TaOz, ATiO3, AZrO3, AAl2O4, A1FeO3, A1MnO3, A1CoO3, A1NiO3, A2HfO3, A3 CeO3, Li2ZrO3, Li2SiO3, Li2TiO3, Li2HfO3, A4N1yOz, YxN1yOz, LaxN1yOz, HfN2yOz, wherein A is Mg, La, Ca, Sr or Ba; A1 is La, Ca, Sr or Ba; A2 is Ca, Sr or Ba; A3 is Sr or Ba; A4 is Mg, Ca, Sr, Ba, Ti or Zr; N1 is V, Nb, Ta, Cr, Mo, W, Mn, Si or Ge; N2 is V, Mo, W or Si; x is 1 or 2; y is 1-3; and z is 2-7. In particular embodiments, the functional layer may be MgO, CaO, SrO, BaO, La2O3, CeO2, Y2O3,TiO2, ATiO3, AZrO3, AAl2O4, A1FeO3, A1MnO3, A1CoO3, A1NiO3, A2HfO3, A3CeO3, Li2ZrO3, Li2SiO3, Li2TiO3, and mixtures thereof. In particular embodiments, the functional layer may be MgO, CaO, La2O3, Y2O3, CeO2, TiO2, ATiO3, AZrO3, AAl2O4, A1FeO3 or a mixture thereof. More particularly, it may be MgO, CeO2, ATiO3, A1FeO3, AZrO3 or a mixture thereof. Most particularly, it may be BaTiO3, BaZrO3, LaFeO3 or a mixture thereof. In a specific embodiment, the functional layer is BaTiO3.
The functional layer may be a dense coating disposed on surfaces of pores of the separation layer as in
Methods for preparing a dense functional layer disposed on pore surfaces or a porous functional layer disposed within the pore include sol-gel techniques, wet impregnation techniques (incipient wetness; including melt infiltration), electrophoretic deposition, chemical vapor deposition, including metal organic CVD, physical vapor deposition, including evaporation and sputtering, and atomic layer deposition.
The effective average pore size of the membrane ranges from about 0.5 nm to about 60 nm, particularly from about 1 nm to about 10 nm, and more particularly from 1 nm to about 5 nm. In embodiments where a dense functional layer is disposed on the surfaces of the pores of the separation layer, ‘effective average pore size’ refers to the pore size of the layer after it has been coated by the functional layer. That is, the limiting pore size is the region defined by the functional layer after it coats the separation layer pores. In embodiments where a porous functional layer is disposed within the pores of the separation layer, ‘effective average pore size’ refers to the pore size of the functional layer. Pore size distribution is narrow to reduce dispersion in Knudsen flow. A narrow pore size distribution is defined as average pore size distribution that does not vary by more than about 100% when the average pore size is in a range from about 2 nm to about 50 nm, and an average pore size distribution that does not vary by more than about 50% when the average pore size is greater than about 50 nm.
In combination with the separation layer, the functional layer provides the overall gas selectivity of the membrane. Flow through the membrane is due to the sum of contributions from Knudsen flow through the pores and surface flow of adsorbed gas along the pore walls. For embodiments where the functional layer is disposed on surfaces of pores of the separation layer, the pore size for Knudsen flow is the size of the pores of the separation layer, reduced by the thickness of the surface functional coating. For embodiments where the functional layer is disposed within and at least partially filling pores of the separation layer, the pore size for Knudsen flow is the effective pore size.
The gas selectivity through the membrane is determined by the relative contributions of Knudsen flow and surface diffusion to the total transport of gas. To achieve reverse selectivity in a gas stream containing CO2 and H2, surface diffusion makes a significant contribution to the total CO2 transport. The rate of surface diffusion depends on the amount of CO2 adsorbed and its relative mobility.
To a first approximation, the surface diffusivity of a gas on a material can be estimated from the heat of adsorption. Since the diffusivity varies exponentially with the negative of the heat of adsorption, materials with lower heats of adsorption exhibit higher surface diffusivity. Physically, this means that materials for the functional layer have a greater affinity for CO2 than for H2, but the affinity for CO2 is not so great that the CO2 binds to the surface without transporting through the pore channel. Low heats of adsorption correspond to weakly bonded CO2 which favors high diffusivities. Accordingly, materials that are suitable for use as the functional layer are characterized by high surface coverage derivative (dθ/dp) and low heat of adsorption (ΔH). These properties can be determined from CO2 adsorption isotherms of the materials. In the context of the present invention, high surface coverage derivative and low heat of adsorption favor high surface transport fluxes when the material is formed into a membrane. Materials that are suitable for use as the functional layer possess a combination of dθ/dp and AH that falls above a line defining the minimum surface coverage derivative (dθ/dp) and maximum heat of adsorption (ΔH) values needed to obtain the desired molar selectivity under specific conditions. These values may be determined using the following analysis:
Molar selectivity can be computed from the material properties of the surface functional layer, the pore structure, and the operating conditions. Mathematically, the molar selectivity, SCO2/H2, is:
The permeability of surface diffusion, Psurt is given by:
where ε is the porosity,
The permeability of Knudsen diffusion for CO2 and H2, respectively are:
where MCO2 and MH2 are the molecular masses of CO2 and H2.
Substituting equations (1) to (3) into (4) and solving for dθ/dp gives:
The following numerical values were used to generate the curves in
The analysis assumes that all pores are straight cylinders with a uniform pore radius, and that surface diffusion is a Fickian process with Arrhenius-type transport from one site to the next. Diffusion constants of chemisorbed molecules are taken to be comparable to those for physisorbed molecules; the partial pressure difference across the membrane is treated as negligible, and the heat of adsorption is taken to be the activation energy for surface diffusion. Assuming dθ/dp to be constant across the membrane simplifies the calculation.
Benchmark surface adsorption property values for alumina, silica, lanthanum oxide and barium titanate (BTO) are shown on the graph in
While the foregoing analysis relates to separation of CO2 from H2, the surface functionalization approach described is capable of removing CO2 from other gases, such as N2, O2, and CH4. This approach is also capable of removing other gases such as H2S through enhanced surface diffusion. However, the effective surface permeability will differ for each gas due to dissimilar heats of adsorption (and consequently, surface diffusivities). Appropriate selectivity may be achieved through the identification of a material with comparable (and low) heats of adsorption for the gases to be removed.
A dense layer of barium titanate (BTO) for use as the functional layer is coated on the pores of anodic alumina using the procedure described by Hernandez, et al. (Hernandez, B. A.; Chang, K.-S.; Fisher, E. R.; Dorhout, P. K. Chem. Mater., 2002, 14, 480-482). The thickness of the coating is adjusted to achieve the desired effective pore size. BTO has dθ/dp=2.6e-6 Pa−1 (at 0.4 atm CO2 partial pressure) and ΔHSiO2=17 kJ/mol. A porous structure with 10 nm pores has selectivity ranging from 0.28 at 25° C. to 0.94 at 500° C. A porous structure with 5 nm pores has selectivity ranging from 0.48 at 25° C. to 3.1 at 500° C. A porous structure with 2 nm pores has selectivity ranging from 1.9 at 25° C. to 18.4 at 500° C.
To prepare the separation layer, a macroporous anodic aluminum oxide membrane (AAO, 200 nm diameter macropores) was immersed in an acidified ethanol-based precursor solution containing a surfactant template and an alkoxide ceramic precursor. Cetyltrimethylammonium chloride (CTAC) and nonionic block copolymers (EO20PO70EO20 [Pluronic P123] and E0106PO70EO106 [Pluronic F127]) were used as the templates and tetraethoxysilane (TEOS) as the silica precursor. The AAO macropores were filled by evaporation-driven gelation of the precursor solution. The sample was then heated to 600° C. to remove the surfactant template, shrink the ceramic deposits between growth stages, and allow deposition of additional porous regions. After heating, the membrane was recycled through the process to deposit additional material.
A macroporous anodic aluminum oxide membrane (AAO, 200 nm diameter macropores) was immersed in an acidified ethanol-based precursor solution containing a surfactant template and an alkoxide ceramic precursor. Nonionic block copolymers (EO20PO70EO20 [Pluronic P123] and E0106PO70EO106 [Pluronic F127]) were used as the templates and titanium (IV) ethoxide as the silica precursor. The AAO macropores were filled by evaporation-driven gelation of the precursor solution. The sample was then heated to 400° C. to remove the surfactant template. P123 polymer (1.0 g) was completely dissolved in 12 g of ethanol. Concurrently, 0.2 g of distilled water was added to 3.0 g concentrated HCl, followed by the addition of 4.2 g of titanium (IV) ethoxide. The solution partially gelled upon addition of the titanium ethoxide, but it re-dissolved after about 30 minutes of stirring. These solutions were combined, stirred and loaded into a Petri dish. The anodic alumina membrane was immersed in the solution for 10 hours and was subsequently removed and heated in air at 400° C. for 4 hours to remove the template. The gas permeability of the sample was measured to demonstrate a continuous path for gas flow through the mesoporous titania structure. An air permeability of 1·10−9 mol/s/Pa/m was measured at 84° F. (29° C.).
While only certain features of the invention have been illustrated and described herein, many modifications and changes will occur to those skilled in the art. It is, therefore, to be understood that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the invention.
This application is a non-provisional of and claims priority from U.S. Provisional Patent Application Ser. No. 60/721,560, filed on Sep. 28, 2005, and is related to Attorney Docket No. 165652-3, entitled “Methods And Apparatus For Hydrogen Gas Production”, filed concurrently herewith. The entire contents of both applications is incorporated herein by reference.
This invention was made with Government support under contract number DOE NETL DE-FC26-05NT42451 awarded by the U.S. Department of Energy. The Government may have certain rights in the invention.
Number | Date | Country | |
---|---|---|---|
60721560 | Sep 2005 | US |