AU2005290082B2 - Removal of carbon dioxide from air - Google Patents

Description

WO 2006/036396 PCT/US2005/029979 REMOVAL OF CARBON DIOXIDE FROM AIR The present invention in one aspect relates to removal of selected gases from air. The invention has particular utility for the extraction of carbon dioxide (CO 2 ) from air and will be described in connection with such utilities, although other utilities 5 are contemplated. Extracting carbon dioxide (CO 2 ) from ambient air would make it possible to use carbon-based fuels and deal with the associated greenhouse gas emissions after the fact. Since CO 2 is neither poisonous nor harmful in parts per million quantities but creates environmental problems simply by accumulating in the atmosphere, it is 10 possible to remove CO 2 from air in order to compensate for equally sized emissions elsewhere and at different times. The overall scheme of air capture is well known. The production of CO 2 occurs in a variety of industrial applications such as the generation of electricity power plants from coal and in the use of hydrocarbons that are typically the main components of fuels that are combusted in combustion 15 devices, such as engines. Exhaust gas discharged from such combustion devices contains CO 2 gas, which at present is simply released to the atmosphere. However, as greenhouse gas concerns mount, CO 2 emissions from all sources will have to be curtailed. For mobile sources the best option is likely to be the collection of CO 2 directly from the air rather than from the mobile combustion device in a car or an 20 airplane. The advantage of removing CO 2 from air is that it eliminates the need for storing CO 2 on the mobile device. Various methods and apparatus have been developed for removing CO 2 from air. In one of these, air is washed with an alkaline solution in tanks filled with what are referred to as Raschig rings. For the elimination of small amounts of C0 2 , gel 25 absorbers have also been used. Although these methods are efficient in removing

CO

2 , they have a serious disadvantage in that for them to efficiently remove carbon dioxide from the air, the air must be driven by the sorbent at a fairly high pressure, because relatively high pressure losses occur during the washing process. Furthermore, in order to obtain the increased pressure, compressing means of some 30 nature are required and these means use up a certain amount of energy. This additional energy used in compressing the air can have a particularly unfavorable effect with regard to the overall carbon dioxide balance of the process, as the energy WO 2006/036396 PCT/US2005/029979 required for increasing the air pressure would produce its own CO 2 that would have to be captured and disposed of. Thus, the prior art methods result in the inefficient capture of CO 2 from air because these processes heat or cool the air, or change the pressure of the air by 5 substantial amounts, i.e., the net loss in CO 2 is negligible as the cleaning process introduces CO 2 into the atmosphere as a byproduct of the generation of electricity used to power the process. Furthermore, while scrubber designs for separating CO 2 from air already exist, generally they are limited to packed bed type implementations whose goal is typically 10 to remove all traces of an impurity from another gas. One such device, described in U.S. Patent 4,047,894, contains absorption elements comprising porous sintered plates made of polyvinylchloride (PVC) or carbon foam assembled spaced from one another in a housing. Prior to the plates being assembled in the housing, potassium hydroxide is impregnated in the porous plates. Such a device has the disadvantage that the 15 sorbent material used to separate CO 2 from air cannot be replenished without disassembling the device housing In another aspect the present invention relates generally to methods and apparatus for separating carbon dioxide (C0 2 ) bound in a solvent. The invention has particular utility in connection with processing hydroxide solvents containing CO 2 20 captured from air (or other alkaline sorbents that are used to collect CO 2 ) and will be described in connection with such utilities, although other utilities are contemplated. Processes that collect CO 2 from the air typically rely on solvents that either physically or chemically bind CO 2 from the air. A class of practical CO 2 solvents include strongly alkaline hydroxide solutions like, for example, sodium and potassium 25 hydroxide. Hydroxide solutions in excess of 0.1 molarity can readily remove CO 2 from air where it is bound, e.g., as a carbonate. Higher hydroxide concentrations are desirable and an efficient air contactor will use hydroxide solutions in excess of 1 molar. Sodium hydroxide is a particular convenient choice, but other solvents such as organic amines may be used. Yet another choice of sorbents include weaker alkaline 30 brines like sodium or potassium carbonate brines. The following discussion applies to all solvents that store CO 2 at least in part in an ionic carbonate or bicarbonate form. 2 WO 2006/036396 PCT/US2005/029979 The design of air contactor systems that ain to contact the air for CO 2 is dealt with in other patents and in the literature [1,2,3]. This aspect of the present invention relates to the recovery of the sorbent, wherein the C02 laden sorbent is rejuvenated and the CO 2 is separated from the liquid. We are describing a set of electrochemical 5 processes that can be combined with an air capture -unit to refresh the hydroxide solution and collect the CO 2 in a separate and in sorne cases pressurized stream. All processes have in common that they separate sodium hydroxide from the carbonate or another salt by electrochemical means - While there are some electrolytical processes that involve only a pair of electrodes, most processes involve 10 separation schemes that use bipolar membranes and/or at least one type of cationic or anionic membranes. In addition some of these pro-cesses involve conventional calcination and/or acid base reactions that lead to the evolution of gaseous CO 2 . Several such processes are claimed in this invention and have been group into seven distinct classes as will be discussed below. 15 Thus, a purpose of this invention is to improve and streamline process designs for capture of carbon dioxide from air, which is an important tool in allowing the use of hydrocarbon fuels in a carbon constrained world. Many of these processes could also find use in other applications in which CO 2 bound into a hydroxide solvent has to be completely or partially removed from the solvent. 20 The disadvantages in the art are addressed and overcome by the CO 2 separation membranes and methods of use thereof as embraced by the present invention. The purpose of the removal of CO 2 fron1 the air is to balance out the CO 2 emission resulting from, for example, the operation of vehicle or a power plant. 25 While the most obvious source of CO 2 emissions that could be remedied by this invention are those for which it would be difficult or impossible to capture the CO 2 at the point of emission, the invention is not restricted to such sources but could compensate for any other source as well. Indeed this approach of CO 2 mitigation could be used to lower the atmospheric concentration of C0 2 , if at some future time 30 society deems the anthropogenic carbon dioxide concentration in the air too high. While the goal of this invention is to capture carbon dioxide from air for purposes of managing the overall carbon dioxide budget of the atmosphere, the 3 4 concepts would apply equally well if the reason for carbon dioxide capture from a gas with low concentrations of C02 is a different one. Examples include, capture for the purpose of the sale of C02 in the food industry or the oil industry, capture of carbon dioxide or other acid gases from dilute streams 5 as they would occur in indoor air, in tunnels or other closed environments. In an aspect of the invention, there is provided a method for capturing C02 from air, which comprises exposing sorbent covered surfaces to air streams where the air streams have a flow that is kept laminar, or close to a laminar regime. 10 In another aspect of the invention, there is provided a method for removing CO 2 from the air, wherein CO 2 is absorbed in a carbonate brine, the carbonate is separated through thermal swing precipitation of the carbonate from the brine; the separated carbonate is electrochemically treated to recover sodium bicarbonate solution in a first electrochemical process step; the 15 bicarbonate is mixed with an acid to release C02 in a more concentrated form; and, the acid is recovered from its salt in a second electrochemical process step. In another aspect of the invention, there is provided a device for separating an alkaline carbonate brine into a cation and bicarbonate, said 20 device including an anode and a cathode to which power is delivered whereupon the cation is moved across the cationic membrane whereby to convert the brine to bicarbonate. In another aspect of the invention, there is provided a device for separating C02 from a bicarbonate brine containing C02, which device 25 comprises: a reservoir having acidic cells and basic cells separated by anionic membranes alternating with bipolar membranes for producing, in a stream, bicarbonate ions which is mixed with acid in the acidic cells which produces C02, and leaving behind in the basic cells a residual brine enriched in carbonate ions, and optionally including a pressurizing device for the C02.

4A The air scrubber according to an embodiment of the invention operates at a minimal air pressure drop and is effective in removing a large fraction of the C02 from the air that is flowing through the air scrubber. We refer to the scrubber design as a lamella design for reasons that become clear below. 5 The lamella based air scrubber unit could become a module in a larger superstructure for funneling the air that can be modified to suit the particular design. The air can be driven by natural wind, by thermal convection or by fans. In another embodiment of the invention, a method and apparatus is 10 proposed to recover the carbon dioxide that has been captured in the scrubber device. In nearly all air capture designs, the overall process of C02 capture from air requires an air contractor that removes C02 from the air by binding the C02 into a solvent or sorbent. The spent sorbent is then processed to recover all or part of the C02, preferably in a concentrated, 15 pressurized stream. The rejuvenated solvent is recycled to the CO 2 collector. This application lays out several processes for recovering an hydroxide based sorbent by means of electrochemical processes that can separate acids from base. Such processes exist and have been demonstrated for a variety of acids. Here we take these processes and combine them in such a 20 way as to built a functional and efficient C02 recovery unit. Embodiments of the invention are also concerned with several novel designs of unit processes that are specifically adapted to the application considered here. The advantages of this invention are several: First the process greatly 25 streamlines the overall flow sheet of carbon dioxide capture from air, by avoiding the intermediate step of transferring the carbonate ion to calcium carbonate which is then calcined to free the C02. The mass handling of such a transfer process is WO 2006/036396 PCT/US2005/029979 complicated. Secondly, the more direct electrochemical process provides also a way of reducing the overall energy consumption. Thirdly it greatly reduces the need for complex, moving equipment to manage solid material streams, as would be necessary in a conventional calcium carbonate driven recovery unit. 5 Finally, we note that implementations of this type could also be used in systems that need to separate carbonate and hydroxide solutions that result from processes other than air extraction. Further features and advantages of the present invention will be seen from the following detailed description taken in connection with the accompanying drawings 10 wherein like numerals depict like parts, and wherein: Fig. 1 is a perspective view of a air scrubber unit made in accordance with one preferred embodiment of the present invention; Fig. 2 is a top plan view of the air scrubber unit of Fig. 1; Fig. 3 is a front, i.e., air inlet view of the air scrubbing unit of Fig. 1; 15 Fig. 4 is a side elevational view of the air scrubber unit of Fig. 1; Fig. 5 is a diagrammatic view of an apparatus for separating carbonate and hydroxide solutions in accordance with another aspect of the invention; and Figs. 6-13 are flow diagrams of various processes and process systems for separating carbonate and hydroxide solutions in accordance with the present 20 invention. Referring first to Figs. 1-4, an air scrubber unit according to one aspect of the present invention removes CO 2 from an airflow that is maintained by a low-pressure gradient. The air scrubber units consist of a wind collector 10 having lamella, which are two sheets or plates 5 covered in downward flowing sorbent bounding a thin air 25 space, and a liquid sump 12. The two sheets forming the lamella preferably are separated by spacers 4 laced between the sheets on thru-rods 2 supported by a rigid frame 1 although the lamella may be supported in spaced relation by other means. In general, the sorbent material flows down the lamella sheets, while the airflow passes between the thin airspace between the sheets. The contact between the 30 air and the sorbent material causes a chemical reaction that removes CO 2 . However, the air scrubber units could also capture other gases present in the air. Sorbent is applied to the lamella sheets according to established state of the art 5 WO 2006/036396 PCT/US2005/029979 approaches, e.g. spray nozzles or liquid extrusion, for example from corrugated tubing 3 fed from a header 6. Also, designs could wet vertical surfaces near the top and let gravity run the fluid over the surface until the entire area is covered. Alternatively, the surfaces could be shaped as flat disks which are wetted as they rotate through a sump. 5 The motion would distribute the liquid along these surfaces. Typical pressure gradients for moving the airflow across the lamella are such that they could be generated by natural airflows, e.g. wind, or thermal gradients. Pressure drops across the unit range from nearly zero to a few hundreds of Pascal, a preferred range is from 1 to 30 Pa and an optimal range may be from 3 to 20 Pa. 10 However, fans either with or without ductwork to guide the air and convection could also be used to move the airflow. The Lamella The purpose of the wind collector is to bring the airflow into close contact with sorbent coated surfaces of the scrubber or wind collector. The basic unit of the 15 wind collector is a single lamella which is a thin air space bounded by two sorbent covered sheets. In the most simple design the sheets are flat, but it is possible that the sheets are curved as long as the air passing over them can move in a straight line, i.e. the sheets curve in the direction normal to the wind flow. Each air scrubber device includes a means of distributing the sorbent on the sheets of the lamella and 20 recapturing the spent sorbent. The following is a list of exemplary designs for the wind chamber lamella: 1) Flat rectangular sheets or plates that are aligned parallel to each other. 2) Corrugated sheets that are lined up parallel to each other, with surfaces straight in the direction of air flow. 25 3) Flat disks rotating around a center axis with with the air flowing at right angle to the axis of rotation. Sorbent could be applied by the wheels dipping into fluid near the bottom of the circular motion. The standing sorbent may only cover the, outer rim of the disks or reach all the way to the axle. Alternatively sorbent may be injected onto the rim by liquid wetting near the axle and flowing around the disk due to gravity 30 and rotary motion. 4) Concentric tubes or similar shapes where air would be blowing along the tube axis. Such tubes could be arranged vertically for counterflow designs with 6 WO 2006/036396 PCT/US2005/029979 wetting initiated at the upper rim or nearly horizontally with sorbent entering at one end and one point and getting distributed through a slow rotating motion of the tubes. Airflows across the lamella may be natural wind flows, or they may be obtained by other means, for example through engineered thermal updrafts. However, 5 high wind speeds would be counterproductive as higher speeds lead to higher rates of energy dissipation. Slow airflow speeds maximize air contact time with the sorbent material on the lamella while minimizing the loss of kinetic energy in the system. Thus, airflow velocities through the scrubber unit may range from virtually stagnant to a few tens of meters per second. A preferred range would be from 0.5 to 15 m/s an 10 optimal range for wind driven systems ranges from 1 to 6m/sec. Practically, the flow speed of the airflow through the wind collector needs to be a substantial fraction of the typical wind speed. The choice of better geometries may reduce the flow speed somewhat, but those enhancements will be factors of two not orders of magnitude. 15 In an exemplary embodiment of the invention, an airflow speed of 2m/s is assumed, but airflow speeds may range from 0.5m/s to about 4m/s. At the nominal flow speed of 2m/s, the flux of CO 2 per unit of wind area is 30 mmol/m 3 /s. The flux into the sodium hydroxide solution is limited to about 0.06 mmol/m 2 /s of hydroxide surface and the air side transport coefficient dominates for boundary layer thicknesses 20 in excess of about 2 to 4mm. In this embodiment, the capture system is as compact as possible, and the size constraints determine the geometry of the apparatus. Placing flat absorbing sheets approximately 0.5cm apart provides nearly 500m 2 of sheet surface area inside a cubic meter. The actual length is approximately 1.2m plus an allowance for the finite 25 thickness of the sheets. It is possible to obtain a slightly larger sheet surface area if the sheets are folded or shaped into tubes. However, since the liquid sorbent flows from the top of the wind collector to the bottom, flat vertical sheets are a natural choice because sheets that have breaks or folds would deflect the passing air to add turbulence to the system and reduce the boundary layer thickness. Since the flat plates 30 already operate at the optimal boundary layer thickness, such turbulence would not improve the CO 2 uptake performance, but it would increase the energy dissipated in the device. 7 WO 2006/036396 PCT/US2005/029979 However, a very large system is also contemplated that, for structural reasons, might have a wind collector design with a depth greater than 1 meter. Such a device would still be optimized to 500m 2 of sheet surface per square meter of frontal opening. In this embodiment, the natural spacing between plates would exceed the 5 optimal boundary layer thickness and thus the introduction of shapes that cause turbulence would be necessary. The turbulence would drive the boundary layer thickness back to the desired value of 2 to 4 mm. For example a 20 m deep filter system, would require about 25m 2 of packing per cubic meter. And, the typical spacing for the sheets would be about 8 cm, far too large for an optimal boundary 10 layer. Creating eddies at the centimeter scale would in effect reduce the boundary layer thickness and thus provide the necessary airside CO 2 flux. The flow through a 1 m deep unit with sheets or plates 0.6 cm apart would be laminar up to relatively high flow velocities. If the Reynolds number is defined as Re= pd v )1 (1) 15 the laminar regime extents to about 1400. In other words, for plates spaced at a distance of 0.6cm apart, the flow remains laminar to about 4m/s. In the following, the resistance to the airflow in such a stack of plates was calculated. 20 The pressure drop per unit length is given by: a P ex (2) If the pressure across a plane normal to the two side walls is assumed to be constant, then the force on a parcel of air of with width Ay, height h, and depth Ax is given according to the literature by 8 WO 2006/036396 PCT/US2005/029979 AYh! Ax=-rn ±Ey hd x(3 or 01V~ =-I P (4) 57 Ox or V(y)=- y'+Cjy+C2 (5) The two integration constants follow from the two boundary conditions, namely v(0) = v(d)=O (6) 5 From that we obtain v(y)- 6y(d-y) 21r8 x (7) The peak velocity between the plates is therefore, 10 d 2 Ap max (8) where L is the length of the plate and AP the pressure drop across this distance. The average flow velocity is given by d2 AP(9 VJ v(y)dy= (9) d~i L or 12A P , (10) d' or L= APd (11) 12rly 15 If we want to determine L such that 9 WO 2006/036396 PCT/US2005/029979 P =A, P 2 (12) then L=pvd 2 Red we find thatat v= 2m/s, L= 0.2m 24 r 24 (13) More generally, a design rule would be that 5 d/L ~ 24/Re (14) where Re is the Reynolds number of the flow. 10 And, the flow between plates is affected by fluctuations in the distance between the plates. Note that the mass flow per unit width of the system is given by Q=pd- d' dp 12r7 ax (15) If we assume that Q is a constant and d is a function of x, then we find 127 Q f L dx A P=- j p 0 x) 15 (16) In a simplified case, the width on one half the lamella is di and on the other half it is d 2 . We furthermore assume that d2d 2d__ dl+d 2 =2d and d= or d, - and d2= d ( dnd d 1

+

1 /0 (17) 20 With that we find that 12I7L d 1 1 2IL ((1+tc) 3 +(1+IC) 3 ) A P -+- 1= pd 3 2 d dI pd 3 16 (18) The correction factor is 1 for a=1, but it rises to 1.89 for u-=2 . In a fully three dimensional system where the air can flow around a narrow spot the total constriction 10 WO 2006/036396 PCT/US2005/029979 is actually smaller. Note, the equations derived above only apply to fully developed laminar flow between the plates. However there is a section at the onset of the plates where the flow is not fully developed. In that region the pressure drop is best characterized by 5 the drag on the separate plates. As the boundary layer thickness increases and the boundary layers from adjacent plates start to overlap the flow develops into the steady flow pattern observed between two plates. The drag per unit width on one side of an infinitely thin plate is given by 10 F2ragC,}xv (19) where x is the distance from the beginning of the plate. The drag coefficient Cd is given according to the literature by Cd=1.308Re =1.308q12P-2v-12 - (20) The pressure drop across a set of parallel plates short enough that they do not yet 15 interfere with each other would be given by P(0)-P(x)=C Pv2l1 2 d (21) where d is the spacing between the infinitesimally thin plates or 20 P(0)-P(x)=1.308riu 2 p 1 2 v 3 2 2 d~' (22) In the onset the airflow looks different as the boundary layers affected by the onset are smaller. Furthermore, in one particular design one might have an ambient wind velocity vo. In front of the lamellae the air stagnates and slows down to a velocity v. 25 The pressure for driving the air through the lamellae is given by 11 WO 2006/036396 PCT/US2005/029979 p '2 212YLv V6 -V )=(23) 2 d' or + v-av =0 (24) 2 d 2O V0 f 12r\ 12)? L (25) * pd pd Of the two solutions to the quadratic equation, we chose the only physical solution, i.e., the one that is positive. 5 The Sorbent The rate of uptake of CO 2 into a strong hydroxide solution has been well studied. The air scrubber of the instant invention is a device that will pull CO 2 , or other gas, directly out of a natural wind flow, or out of a flow subject to a similar driving force, e.g., a thermally induced convection. 10 CO 2 uptake into a strong hydroxide solution involves a chemical reaction that greatly accelerates the dissolution process. The net reaction is C0 2 (dissolved) +2OH~ -+COT~+H 2 0 (26) There are several distinct pathways by which this reaction can occur. The two steps that 15 are relevant at high pH are C02 (dissolved)+OH~->HCQs (27) followed by HCO+0H-+CO 3 ~+H,0 (28) 20 12 WO 2006/036396 PCT/US2005/029979 The latter reaction is known to be very fast; the first reaction on the other hand proceeds at a relatively slow rate. The reaction kinetics for reaction (2) is described by d -[CO2]=K[OH][CO 2 ] (29) 5 Hence the time constant describing the reaction kinetics is . K[OH ] (30) The rate constant K has been measured at 20*C and infinite dilution, 10 =5000litermol-s1 =5m 3 mol-'s (31) The ionic strength correction is given by K=K., 1 0 0.13 (32) At high concentration of CO 2 in the gas, the rate of reaction (2) limits the rate of 15 uptake, even though the time constant for a one molar solution at 0.14 ms is quite short. Following standard chemical engineering models, e.g. Dankwert or Astarita, one can describe the transfer process in which a gas component is dissolved or chemically absorbed into a sorbent with a standard model that combines a gas-side 20 flow transfer coefficient and a liquid side transfer coefficient to describe the net flow through the interface. The total flux is given by F=KG(p(x=-o) -p(x=0)) =Ki(P '(x=0)-p '(X =c)) where p and p' are the molar concentrations of CO 2 in the gas and in the solution 25 respectively. The parameter x characterizes the distance from the interface. Distances into the gas are counted negative. At the boundary Henry's law applies, hence 13 WO 2006/036396 PCT/US2005/029979 p '(0)=KH P (0) (34) Expressed as a dimensionless factor, KH = 0.7. For the gas side the transfer constant can be estimated as 5

K

0 = D (35) where A is the thickness of the laminar sublayer that forms on the surface of the interface. The thickness of this layer will depend on the geometry of the flow and on the turbulence in the gas flow. Assuming the geometry of the flow and the turbulence 10 in the gas flow is given, then the optimal choice for A must be determined. For a fluid package, the standard approach to estimating the transfer coefficient assumes a residence time TD for the parcel on the surface of the fluid. This time results from the flow characteristic of the sorbent and it include surface creation and surface destruction as well as turbulent liquid mixing near the surface. 15 A = D TD (36) Since diffusion in the time TD can mix the dissolved CO 2 into a layer of thickness the flux from the surface is given by a p' 20 P a (37) where DL is the diffusion constant of CO 2 and p' the liquid side concentration of CO 2 . The gradient is evaluated at the surface. The transfer coefficient of the liquid is defined from the equation 25 F=KL(p'(x=0)-p'(x=oo)) (38) Approximating the gradient by 1 Note that typically, Henry's constant has dimensions, as concentrations on the gas side are measured as partial pressures, i.e., in units of Pascal or units of atmospheres (atm), whereas the liquid side concentrations are typically measured as moles per liter. Thus a typical unit would be liter/mol/atm. 14 WO 2006/036396 PCT/US2005/029979 8p _ p'(0)-p'(o) ax A (39) shows that for a diffusion driven absorption process DL 5 (40) Here DL is the diffusion rate of CO 2 in the sorbent. In the presence of a fast chemical reaction where the reaction time rR «TD, the layer that absorbs CO 2 is characterized by this shorter time, hence the transfer coefficient is given by 10 DL KL=

FL

TR (41) In the presence of a chemical reaction the transfer coefficient is thus increased therefore by a factor 15 TR (42) However, this enhancement can only be maintained if the supply of reactant in the sorbent is not limited. In the case of CO 2 neutralizing a hydroxide solution, it is possible to deplete the hydroxide in the boundary layer. The layer thickness X contains an area density of hydroxide ions of poHX and the rate of depletion is 2 KL p' CO 2 : 20 Thus for the fast reaction limit (eqn. 41) to apply POH- TR 1 2p 'co, TD (43) In our case 15 WO 2006/036396 PCT/US2005/029979 1 POH- TR: K (44) Hence the condition can be rewritten as 2p 'co2 KTD > 1 (45) 5 The critical time for transitioning from fast reaction kinetics to instantaneous reaction kinetics is approximately 10 see for ambient air. The transition does not dependent on the hydroxide concentration in the solution. However, once past the transition, the rate of uptake is limited by the rate at which hydroxide ions can flux to the surface. It is therefore lower than in the fast limit, and the CO 2 flux is given by 10 1 DoH F=- PoH 2 TD (46) In the instantaneous regime the flux is independent of the CO 2 concentration in the boundary layer. 15 The flux can be characterized by an effective transfer coefficient, which can be written as F=Kej(peo,-p 'co, KH) (47) Here the molar concentrations are for the asymptotic values in the far away gas and 20 far away liquid. In the case of hydroxide solutions, the latter is zero. Hence, F=Kg ePco, (48) and 11 Kefg= -+ 25 KG KL Ku (49) An optimal design is close to the border between gas side limitation and liquid side limitation. Therefore, we establish a design value for the air side boundary thickness 16 WO 2006/036396 PCT/US2005/029979 As* DG ;DL/TR (50) This is approximately 4 mm for air based extraction of CO 2 . These constraints together very much limit a practical design. For a 1 molar 5 solution, the total solution flow has been measured as 6x10~5 mol m~ 2 s-, which translates into an effective value of 0.4 cm/s which is close to the theoretical value. As for types of sorbents that absorb CO 2 , there are a wide variety of options that can be used. In one embodiment, aqueous hydroxide solutions are used as the sorbent material. These would tend to be strong hydroxide solutions above 0.1 molar 10 and up to the maximum possible level (around 20 molar). The hydroxides used as a sorbent could be of a variety of cations. Sodium hydroxide and potassium hydroxides are the most obvious, but others including organic sorbents like MEA, DEA etc. are viable possibilities. Furthermore, the hydroxides need not be pure, they could contain admixtures of other materials that are 15 added to change or modify various properties of the sorbent. For example, additives may improve on the reaction kinetics of the hydroxide with the CO 2 from the air. Such catalysts could be surfactants or molecules dissolved in tlie liquid. Additions of organic compounds like MEA are just one example. Other additives may help in reducing water losses by making the solution more hygroscopic. Yet other additives 20 may be used to improve the flow or wettability characteristic of the fluid or help protect the surfaces from the corrosive effects of the hydroxide solution. In addition, any sorbent used in the invention must wet the surfaces of the lamella sheets. To this end, there are various means known in the art. These include surface treatments that increase hydrophilicity, surfactants in the sorbent and other means. 25 The invention includes the following important design features: 1) Lamella sheets are substantially smooth in the direction of the airflow on a size scale consistent with the size sheet separation. (However, incidental or engineered structures on a much finer scale may be used to improve the CO 2 transport coefficient.) Variations in shape that are at right angles to the airflow, are of relatively 30 little concern, as long as they do not interfere with the efficient wetting of the plates, sheets or surfaces; 17 WO 2006/036396 PCT/US2005/029979 2) The sheets are held in place sufficiently tightly or rigidly such that their flexing or flapping does not significantly reduce pressure variations between the lamellae. 3) Airflow through openings in the surfaces is inhibited so that it cannot 5 significantly influence pressure variations between the lamellae. 4) The spacing between the lamellae is chosen such that the system does riot transition out of the laminar flow or at least does not deviate much from that regime 5) The depth of the membrane units is kept short enough to avoid nearly complete depletion of the air in the front part of the unit. 10 6) For utilization of both sides of the sheets it is preferable to arrange the lamella vertically. However, deviations from such a design could be considered for other flow optimizations. 7) The height of the lamella is chosen to optimize wetting properties of the surfaces and to minimize the need for reprocessing the fluid multiple times. 15 The Building Blocks of the CO2 Recovery System. In another aspect of this invention, the following electrochemical processes may be utilized in the CO 2 capture systems described in this invention, or in any other device that has collected CO 2 . These electrochemical processes are all based on the separation of a salt into its acid and base, where the acid and the base stay in solution, 20 by means of electrodialysis with bipolar membranes. Examples include the formation of sodium hydroxide and hydrochloric acid from sodium chloride, and the formation. of sodium hydroxide and acetic acid from sodium acetate. Other combinations of acid and base have also been demonstrated in the literature, in the patent literature and in industrial practice. In the context of this invention, units of this type will be used to 25 separate a hydroxide and carbonate solution, as well as units that separate the salt of a weak acid into the corresponding acid and base. In the following we describe a number of processing steps which become the basic building blocks of the processes we consider. 1. The separation of a mixture of sodium hydroxide and sodium carbonate 30 electrochemically into sodium hydroxide and sodium carbonate. For this process step we can rely on existing building blocks or use specifically designed units using electro-dialysis for the separation. These techniques also can be extended to other 18 WO 2006/036396 PCT/US2005/029979 cations than sodium, such as, but not limited to potassium and ammonia, and the cations of organic amines, such as monoethanolamine (MEA), diethanolamine (DEA) and the like. The basic reaction in all cases is the separation of a mixture of R-OH and R 2 C0 3 through a membrane process into separate solutions of R-OH and RHCO 3 . 5 2. The electrochemical separation of a metal bicarbonate into the metal carbonate and CO 2 . This process preferably uses electrodialysis involving bipolar membranes, but other electrolytic processes have been described in the literature and may be used. 3. The separation of the metal bicarbonate into the metal hydroxide and CO 2 . Again this process preferably relies on electrodialysis with bipolar membranes, but it 10 also could be accomplished by electrolysis of metal bicarbonate producing hydrogen that is reused in a hydrogen electrode producing CO 2 . 4. Units that combine two or more of the above building blocks 2 and 3 or 4 into a single unit. For example, processes that take a mixture of carbonate and hydroxide all the way to a hydroxide solution and CO 2 gas. 15 The following are additional building blocks that do not involve electrochemistry: 1. A membrane process that uses concentration gradients to separate cations such as sodium from the solvent to reduce or eliminate the hydroxide in the input solvent. In some cases this unit could partially transform the solvent from carbonates into 20 bicarbonates. 2. Temperature swing processes to separate sodium carbonate from a mixture of sodium carbonate and sodium hydroxide via precipitation. . 3. Processes that take bicarbonate solutions to carbonate solutions by thermal or pressure swing. Such processes are conventionally deployed in certain C0 2 25 scrubbing systems that operate at CO 2 pressures sufficiently high for the reaction between sodium or potassium carbonate and CO 2 to form bicarbonates. 4. Processes that take bicarbonate solutions and use evaporation or thermal swings to precipitate bicarbonate from solution. 5. Processes for the calcination of bicarbonate to carbonate. Specifically of 30 interest here are sodium or potassium bicarbonates. 6. A process that mixes an acid with hydroxide-carbonate mixture to neutralize the mixture and to fonn solid precipitates of these salts. The process can stop either 19 WO 2006/036396 PCT/US2005/029979 at pure carbonate or move on to form carbonate/bicarbonate mixtures or move all the way to bicarbonate. 7. A process that uses an acid to drive all CO 2 out of the bicarbonate, or carbonate or hydroxide mixture. This process can be performed at elevated pressure 5 in order to deliver the CO 2 at pipeline pressure. An outline of the overall process schemes All processes begin with the extraction of carbon dioxide from air in a unit that here is not further specified. A specific implementation has been dealt with in another aspect of this invention, The details of this unit are not of interest here, other 10 than to note that this unit will consume a hydroxide based solvent that is fully or partially converted into a carbonate. It may be possible to convert the solvent partially into a bicarbonate. In this latter case on may also consider the use of carbonate as the starting solvent. The input solvent may contain other chemicals than just the hydroxide. For example it could contain certain additives that improve the 15 process performance, but in particular it could contain residual carbonate from previous process cycles. The purpose of this section of the invention is to outline processes and methods for recycling the solvent and a partial or complete recovery of the CO 2 into a concentrated stream preferably at a pressure suitable for subsequent processing steps. 20 In the following discussion for the sake of clarity we will refer to specific hydroxides and specific acids. However, we emphasize that the process is not limited to these specific chemicals but can easily be generalized to encompass other ionic species. In the following example the air contactor unit uses a sodium hydroxide solution whose concentration is in excess of one mole per liter of sodium hydroxide. 25 Some remnant carbonate may still be in the solvent from the previous process cycle but as the solvent is exposed to air, hydroxide is converted into carbonate and the carbonate concentration of the solution starts rising until further conversion would not be desirable. There are several reasons for stopping the absorption process. In particular the process may be stopped because the hydroxide is exhausted, or the 30 carbonate concentration reaches saturation levels. For most capture designs precipitation of carbonate in the absorber would be undesirable. The resulting carbonate solution is then returned from the capture unit for further processing. 20 WO 2006/036396 PCT/US2005/029979 Conceptually one can consider three steps in the recovery process as follows: 1. Separation of unconverted hydroxide from the carbonate; 2. Decomposition of sodium carbonate into sodium hydroxide and sodium 5 bicarbonate, which is an acid base decomposition; and 3. Decomposition of sodium bicarbonate into sodium hydroxide or sodium carbonate and carbonic acid. In some implementations these steps could be combined together into two process steps or even a single process step. 10 Alternatively, one can accomplish each of these steps by neutralizing the base, (here sodium) with a weak acid. If the sodium salt of the acid precipitates, then the process can be stopped at any point because it is straightforward to separate the acid anion in its precipitated form from the liquid; otherwise the neutralization process has to run to completion in which case the result is gaseous CO 2 and the salt of the base. 15 If the air capture uses sodium hydroxide and the acid is acetic acid, the result would be sodium acetate. The resulting sodium acetate would be separated into sodium hydroxide and acetic acid. Both of them are recycled. The decomposition of sodium acetate is best accomplished with electrodialysis units encompassing bipolar membranes. If a high pressure CO 2 is required an acid stronger than acetic acid is 20 required. Process 1: Referring to Figs. 5-7, process 1 breaks the upgrading of the solvent into three distinct steps. First it separates a large fraction of the carbonate from the brine. Then it uses an electrochemical step to in effect withdraw sodium ions from the brine 25 leading to sodium hydroxide and sodium bi-carbonate. Finally the resulting sodium bicarbonate releases its CO 2 under addition of an acid, which again is recycled in an electrochemical step. The advantage of this process implementation is that it combines high energy efficiency, with the ability to produce pressurized CO 2 . It s an advantage of the electrochemical separation that carbon dioxide can be delivered at 30 elevated pressure. 21 WO 2006/036396 PCT/US2005/029979 Step 1.1 Extract sodium carbonate from the spent solvent by a temperature swing. Sodium carbonate solubility is far smaller than that of sodium hydroxide. (Similar reasoning applies to some of the other hydroxides, but this implementation is limited 5 to those for which the solubility ranges match). Consequently, for concentrated hydroxide solutions the maximum amount of sodium carbonate that can converted to sodium carbonate by CO 2 absorption is limited. One disadvantage of operating at high sodium hydroxide concentrations is that the spent solvent is still dominated by sodium hydroxide, which should not be processed through a number of expensive 10 stages. The temperature swing method overcomes this problem, because it allows one to separate the carbonate without having to pass all sodium hydroxide through membrane systems. If the spent solution is nearly saturated in sodium carbonate, one can extract a fraction of the carbonate through precipitation. Solubility of sodium carbonate changes by more than a factor of three between 0*C and 25"C. Thus it is 15 possible to refresh the sodium hydroxide solution through a temperature swing, with heat exchange between the incoming fluid and outgoing fluid. This approach could utilize ambient heat in warm dry climates where the maximum temperature swing is large. The refreshed hydroxide solution is sent back to the air contactor unit. This approach also is more advantageously deployed in dry climates where high NaOH 20 concentrations would help to reduce the concurrent water losses. Step 1.2 The sodium carbonate precipitate is dissolved in water at maximum concentration. The sodium carbonate is processed further in an electrochemical unit for acid/base separation that can separate sodium carbonate into sodium hydroxide 25 (the base) and sodium bicarbonate (the acid). There are several different designs possible for this electrochemical separation. Some are conventional and state of the art generic separators for acid and base that use bipolar membranes. Others involve hydrogen electrodes. Below we describe a particular unit specifically designed for sodium carbonate disassociation. 30 Step 1.3 The bicarbonate solution resulting from Step 1.2 is injected into a pressure vessel where it mixes with a weak acid. Preferred acids include citric, formic and 22 WO 2006/036396 PCT/US2005/029979 acetic acid. However, the invention is not limited to any specific acid. The acid base reaction drives carbonic acid out of the salt. The carbonic acid then decomposes into CO 2 and water. CO 2 at first dissolves into the brine but soon reaches a pressure that exceeds the container pressure, leading to the release of a pressurized CO 2 stream. 5 The design constraints on this unit put some limits on the choice of an acid. Most importantly, the acid needs to be strong enough to drive CO 2 out of the solution, even at the design pressure. For a further discussion of this unit see below. The advantage of such a system is that it allows the release of concentrated CO 2 at pipeline pressure without having to put a large electrochemical unit into a pressure vessel. Left behind 10 is a brine of the salt of the weak acid. This could be sodium acetate, sodium citrate or any other salt of a weak acid. Step 1.4 The salt of the weak acid and the base used in the capture is decomposed in an electro-dialysis unit utilizing cationic, anionic and bipolar membranes to recover 15 sodium hydroxide and the weak acid. There are several variations of this unit that could be used. With the conclusion of Step 1.4 the CO 2 is recovered, and the residual sodium hydroxide is returned to the overall cycle. In choosing among various design options, it is advantageous to use a unit that removes sodium ions from the solution rather than removing the anion from the solution, as it would generally be undesirable 20 to send residual anions into the air contactor. This also makes it possible to control the concentration of the sodium hydroxide brine. Depending on the detailed conditions of the implementation, this last unit can therefore be used to adjust the water content of the sodium hydroxide to match what is desired in the air contactor. While we refer here generally to a weak acid, because the electrodialysis process 25 requires less energy in recovering a weak acid, we note that the process in principle also works with a strong acid. In some special cases strong acids may have other advantages that overcome the inherently higher electrochemical potential. For example some membranes can sustain larger currents on simple ions of strong acids, then on larger organic acids. 30 Process 2: Referring to Fig. 8, this process is very similar to Process 1, but it replaces the first step with a membrane separation system. This will create a relatively dilute 23 WO 2006/036396 PCT/US2005/029979 -NaOH solution that in turn needs to be concentrated. It could be used in subsequent steps as the starting brine on the hydroxide side of the membrane. Process 2 works particularly well, if the air extraction step has led to evaporative water losses from the solvent and thus additional water needs to be added to the solvent in any case. 5 Step 2.1: Use a periodic system of cells with dilute NaOH solutions alternating with concentrated NaOH/Na 2

CO

3 brine. On the one side the cells are separated by a cationic membrane and on the other by a bipolar membrane. The last cell is connected to the first cell making the system periodic. A design could be reduced to a 10 simple pair of cells, but geometrical constraints generally favor a multiple cell system. As the sodium diffuses through the cationic membrane, charge neutrality of the cells demands that the bipolar membrane provide an H*-OH~ pair. The H* neutralizes the left behind OH~; the OH~ forms a base with the withdrawn sodium in the other chamber. To a first approximation, the sodium concentration in the two chambers 15 will balance out, suggesting that this separation can be performed without electric power input if at least half of the NaOH in the spent solvent has been converted into sodium carbonate. If this is not the case, it is still possible to use this system to partially reduce the NaOH concentration, or if one is willing to increase the water content of the solution, one can transfer a larger fraction of the sodium ions into the 20 new hydroxide chamber which needs to maintain a sodium ion concentration that is lower than the remaining sodium ion concentration in the carbonate side of the system. Diluting the brine at this point may actually be desirable, as many air contactor designs will have lost some of the water that was originally in the solution. However, process step 2.2 which is the direct analog of process step 1.2 can also 25 proceed if the extraction of NaOH was not entirely complete. By taking a number of these cell arrangements (without closure at the end) and incorporating them into a stack that is used in step 2.2 to generate sodium bicarbonate, one can harness the power of the concentration driven cells to partially provide the driving expression for the second step in the conversion (Fig. 8). 24 WO 2006/036396 PCT/US2005/029979 Step 2.2 This process is very similar as Step 1.2 above. The difference is that the sodium carbonate is delivered in dissolved form, and it is likely that there is residual sodium hydroxide left in the input brine. 5 Step 2.3 and Step 2.4 The same as Steps 1.3 and 1.4. Process 3 Referring to Fig. 9, for the sake of process simplicity we eliminate the step of electrochemically separating sodium carbonate into sodium hydroxide and sodium 10 bicarbonate. Instead we use the weak acid directly to produce CO 2 . This implementation is included for its simplicity, and because it allows to take advantage of the future state of the art, that may have reached extremely efficient implementations for acid/base separation in some specific acid/base pair. It is of course possible to also generate a hybrid process where steps 1.1 and 2.1 may be 15 pushed further than just to the carbonate boundary. As another alternative one could use the electrochemical separation in 1.2 and 2.2 but stop short of the full formation of sodium bicarbonate. Step 3.1 This step separates sodium carbonate from the sodium hydroxide in the input 20 brine. This step could either be accomplished as in Step 1.1 or as in Step 2.1. It could also completely be eliminated by introducing a hydroxide carbonate mixture into step 3.2. Step 3.2 This step is the analog to Steps 1.3 and 2.3 but it requires twice as much acid. 25 The advantage of such an implementation is a substantial streamlining of the flow sheet. Step 3.3 The step is the analog to Steps 1.4 and 2.4, but it produces twice as much acid. Process 4: 25 WO 2006/036396 PCT/US2005/029979 Referring to Fig. 10, process 4 starts out like processes 1 and 2, but then replaces the acid decomposition with a bipolar membrane process that drives the CO 2 out of solution. Step 4.1 5 This step is the same as Step 1.1 or Step 2.1 Step 4.2 This step is the same as Step 1.2 or Step 2.2 Step 4.3 Electrochemical separation of NaHCO 3 into CO 2 and NaOH. This is based on 10 electrodialysis with bipolar membranes. In order to obtain high pressure CO 2 the electrodialysis unit should be put into a pressure vessel, which maintains the desired

CO

2 pressure over the cell. For this reason it would be desirable not to combine steps 4.2 and 4.3 as this would increase the size of the unit that needs to be maintained at pressure. It is however possible to combine the two units into one. The advantage of 15 such a design would be a reduction in process steps. It would even be possible to combine all three units into one. Other implementations would use other electrochemical nieans, as for example an electrolysis system that on the cathode generates hydrogen and for the anode uses a hydrogen electrode that consumes the hydrogen produced at the cathode. 20 Process 5: Process 5 and 6 extract CO 2 from the bicarbonate brine producing at least in part sodium carbonate and thus introduces a new recirculation loop between the final steps and the upstream steps. Process 5 precipitates out sodium bicarbonate whereas process 6 implements an aqueous version of the process. As a result these processes 25 are well suited for implementations that only produce carbonate and use this carbonate as a fresh sorbent for CO 2 capture. Refer to Fig. 11: Step 5.1 This step is the same as in Step 1.1 or Step 2.1 Step 5.2 26 WO 2006/036396 PCT/US2005/029979 This step is the same as in Step 1.2 or Step 2.2. However, the input to this unit is in part derived from process 5.1 and in part from recycled sodium carbonate derived from Step 5.5 Step 5.3 5 Increase the concentration of bicarbonate through water removal. This is best accomplished by letting water pass through water permeable membranes into concentrated brines. There are two possible sources for these brines (1) the concentrated brines that leave the air contactor; this is particularly useful if Step 5.1 follows 2.1; and (2) the concentrated brines that are derived from Step 5.1 if it is 10 analogous to 1.1 and results in solid sodium carbonate precipitate. The result is a concentrated brine of sodium bicarbonate. It needs to be contained in an air tight container so as to contain the higher than ambient CO 2 partial pressure over the solution. Another option for dewatering the brine is to run a conventional 15 electrodialysis unit (without bipolar membranes) in reverse. Rather than using the pure water, which will be reused elsewhere in the cycle (the total system loses water), the concentrated brine on the other side of the membrane will be collected for further use. The advantage of this approach is that it requires smaller volumes to pass through membranes, but it requires an electromotive force to succeed. 20 Step 5.4 Temperature swing to precipitate sodium bicarbonate from the brine. The temperature swing is not as efficient as the temperature swing for the precipitation of Na 2

CO

3 . However, operating between 25 and 0*C would allow one to remove roughly 1/3 of the bicarbonate. Heat exchange between input and output minimizes 25 heat losses in the system. The reinaining brine is sent back to Step 5.3 for further dewatering. Step 5.5 Calcination of solid sodium bicarbonate to form sodium carbonate and pressurized CO 2 . In order to pressurize the CO 2 , the calciner is contained in a 30 pressure vessel. Such a system could utilize various sources of waste heat, e.g. from a refinery or from a power plant. Another alternative might be solar energy which has 27 WO 2006/036396 PCT/US2005/029979 the advantage of being carbon neutral. If fossil carborn is used the heat source should use oxygen rather than air and collect the CO 2 that results from its combustion. Hydrogen and oxygen produced in the upstream electrodialysis units would provide another CO 2 free source of energy. Alternatively, a srnall fraction of the sodium 5 carbonate produced could be used in part to adsorb the CO 2 from the combustion process. This sodium bicarbonate brine is returned to 5.3 in order to be dewatered again. The remaining sodium carbonate is sent back to Step 5.2 The CO 2 stream leaves from this unit. The advantage of this implementation is that it reduces the electricity demand 10 and replaces it in part with low grade heat. This method is therefore particularly useful in regions where electricity is expensive, or very CO 2 intensive. Methods 1-4 are advantageous in regions with low cost, low carbon electricity. E.g., hydroelectricity or excess wind power from a large wind mill farm. Process 6: 15 Process 6 is similar to Process 5, but it replaces the precipitation/calcination with a thermal decomposition of sodium bicarbonate directly in solution. The advantage of Process 5 is that it easily can achieve high pressure in the CO 2 stream, whereas Process 6 is easier to implement and it follows conventional processing streams. Referring to Fig. 12: 20 Step 6.1 This step is the same as Step 5.1 Step 6.2 This step is the same as Step 5.2 Step 6.3 25 This step is the same as Step 5.3, but concentrations can be kept lower than in 5.3 and in some implementations it could be omitted. Step 6.4 Temperature swing to heat the solution to remove CO 2 from the brine and return a brine enriched in sodium carbonate back to Step 6.2. Heat exchangers are 30 used to minimize energy demand. Water condensation can be managed inside the unit. See discussion below. Potential heat sources are similar to those listed in Step 28 WO 2006/036396 PCT/US2005/029979 5.5. A fraction of the brine produced in 6.2 can be used to absorb CO 2 produced in the heat generation. The resulting sodium carbonate rich brine is returned to Step 6.2. Process 7: Process 7 is similar to 5 and 6 in that it operates the CO 2 generating anit 5 strictly between bicarbonate and carbonate and that it makes no attempt to drive the electrodialysis of the CO 2 generator past this point. It may indeed stop slightly before that so as to avoid creating high pH solutions. Refer to Fig. 13: Step 7.1 This step is the same as in Step 6.1 10 Step 7.2 This step is the same as in Step 6.2 Step 7.3 This step is the same as in Step 6.3 Step 7.4 15 A cell alternating anionic and bipolar membranes with the basic brine starting out as bicarbonate solution and the acidic brine as pure water, where the applied voltage drives the bicarbonate ions and carbonate ions across the anionic membrane to create carbonic acid on the acid side, which will release CO 2 . With the removal of carbonic acid anions, the brine on the basic side gradually rises in pH. The process 20 must stop when OH~ concentrations start to compete with dissolved inorganic carbon. This would allow the transformation of the bicarbonate brine to a carbonate brine. The remaining carbonate brine is sent back to the previous unit, so that after some dewatering it can be reconverted into a bicarbonate brine. Discussion of the Processes 25 The processes outlined above represent different optimizations for different situations and different goals. Which one will prove optimal will depend on the typical temperatures at which the units operate, on the local cost and carbon intensity of electricity, on the progress of various electrochemical schemes to generate acid and base. As this field is still young and in flux, it is possible that over time the advantage 30 will move more and more to the fully electrochemical designs. 29 WO 2006/036396 PCT/US2005/029979 Process 1 through 4 which all rely on a second acid to complete the transformation of the spent solvent into CO 2 and fresh solvent make it possible to independently optimize acid/base separation and pressurization of CO 2 . The advantage of these methods is that they completely eliminate the need of compressors 5 for driving CO 2 up to pipeline pressure. The same is true for Process 5, but for Process 6 the maximum pressure that can be achieved is limited by the temperature to which one is willing to drive the carbonate/bicarbonate brine. One advantage of Process 6 is that Step 6.4 has been implemented in the past on large scales and thus reduces cost uncertainties associated with the scale up of new processes. 10 Other process units may be integrated into the overall stream to deal for example with impurities. For example, the carbonate brine arriving from the air contactor should be filtered to remove dust accumulation. While we discuss below in some detail more specialized implementations of unit processes that are optimized for our design, one can use standard 15 implementations for all process units. Implementation of the Separation of Carbonate into Bicarbonate and Hydroxide In principle any implementation of an established electrochemical process for separating acid and base can be adapted for this process unit. Not all of them rely on bipolar membranes but many of them do. One we have developed for this purpose 20 combines a series of cationic and bipolar membranes. The system ends in two standard electrodes producing hydrogen and oxygen. These will be responsible for a few percent of the total energy consumption. They can either be integrated into the process via a fuel cell or - in Processes 5 and 6, which require heat -- they can be combusted to produce heat without CO 2 emission. 25 Sodium ions follow either a concentration gradient or an electric gradient from the mixture into the next cell which is accumulating sodium hydroxide. Different sections of the cell may be working on different concentrations in order to minimize potential differences in the system. In particular, as mentioned before it is possible to include the upstream separation of hydroxide from carbonate which can be driven by 30 concentration gradients alone. Since none of the units reach acidic pH, the proton concentration is everywhere small enough to avoid the need for compartments separated by anionic membranes. The system is therefore simpler than a conventional 30 WO 2006/036396 PCT/US2005/029979 bipolar membrane system that needs to control proton currents. In these cells the negative ions do not leave the cell they started in. The advantage of extracting sodium carbonate from the solvent brine prior to this step is that it reduces the amount of sodium that has to pass through these membranes. However, a simplified version 5 of the process can eliminate the first step. Implementation of the Acid Driven CO 9 Generator Mixing an acid with sodium carbonate or bicarbonate leads to the vigorous production of CO 2 . If the acid is strong enough, the entire process can generate high pressures of CO 2 if the reaction is contained in a vessel that is held at the desired 10 pressure. One possible use for such a system would be to generate CO 2 at pressures that are above pipeline pressure, eliminating the need for subsequent compression. One possible implementation of such a system envisions three small reservoirs, one filled with acid, one filled with bicarbonate and the third filled with the salt (e.g., sodium salt) of the acid. The bicarbonate and acid are injected from their 15 respective reservoirs into a flow channel shaped to enhance mixing of the two fluids. If the acid is weak and the reaction therefore slow, it is also possible to introduce a container vessel that is actively stirred. In the fast reactor, the mixing channel rises to a high point where the gas is separated from the liquid flow which then is channeled downward again to enter the salt solution reservoir. The injectors into the acid and 20 base reservoir are mechanically coupled to the salt exhaust reservoir. The mechanical energy harnessed at the exit is nearly sufficient to drive the injection pumps. A direct mechanical coupling could be based on piston displacement pumps which are mechanically connected. Small turbines could similarly be coupled together. There are many state of the art approaches that allow for the mechanical coupling. 25 Small systems may instead operate in a batch operation where the input tanks and output tanks are separated for example by a diaphragm. When the pressure is released filling the empty input tanks forces the draining of the full output tank. Then the system is pressure isolated from its environment and CO 2 is produced as the two fluids are pumped from the input tank into the output tank. Once the output tank is 30 full, the CO 2 line is valved off, and the cycle repeats itself. Another implementation could use pistons, which in effect replace the moving diaphragm. 31 WO 2006/036396 PCT/US2005/029979 It is of course also possible to provide electric coupling, by converting the output energy of the salt stream and CO 2 stream into electric power. A small mismatch in volumes could be made up by withdrawing some pressure energy from the CO 2 output line. In principle, this could be a substantial source of mechanical 5 energy satisfying a large number of pumping needs within the overall system. One can use this ability to adjust the mismatch in strength between the carbonic acid and the acid used to drive the system. In this way the acid production becomes a convenient way of providing mechanical energy which is removed from the exhaust carbon dioxide. 10 Prior to injection of the carbon dioxide into the output stream, it needs to be cleaned and dried so that it meets whatever requirements are put on it in the particular application or particular means of disposal Water Management in A Thermal Swing C0 Generator In heating a bicarbonate solution, the CO 2 will carry with it water vapor that 15 needs to be condensed out. The CO 2 which will leave the solution at some pressure and will flow out of the reservoir mixed with water vapor. In the next stage it is used to preheat the incoming solution and in the process it condenses out the water vapor. The collected water is best kept out of the bicarbonate solution as increasing the brine concentration raises the CO 2 partial pressure over the solution. 20 The water can be used in providing input feed for creating fresh sodium bicarbonate in the electrochemical acid/base separations in Step 6.2. As the opportunities for the use of CO 2 in the oil industries become exhausted, work will be underway to put in place regulatory allowances for the CO 2 "credits" earned through sequestration. These "credits" then will have a market value used in a 25 number of ways. One possibility will be for local regulatory agencies to offer a "credit certificate" to an auto manufacturer or purchaser as a means to boost fleet mileage while allowing the continued use of popular vehicle designs that may not perform to desired levels. It is not unreasonable to for see the time when an automobile or truck may be 30 driven with conventional internal combustion technology (or advanced propulsion systems relying on hydrocarbon fuels) while at the same time making the claim as a 32 33 zero emission automobile since sufficient CO 2 had already been removed from the atmosphere through this process. This might be arranged as an accessory certificate attached to the purchase price of the automobile or truck, or as a regulatory demand placed upon the transportation industry or some other arrangement yet to be defined. 5 Or a socially conscious person may "buy-out" carbon upfront, i.e., at the time of purchase of an automobile. In the specification and the claims the term "comprising" shall be understood to have a broad meaning similar to the term "including" and will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the 10 exclusion of any other integer or step or group of integers or steps. This definition also applies to variations on the term "comprising" such as "comprise" and "comprises". The reference to any prior art in this specification is not, and should not be taken as an acknowledgement or any form of suggestion that the referenced prior art 15 forms part of the common general knowledge in Australia. While the invention has particular utility in extracting C02 from the air, the air scrubber of the invention may be used for removing other gases from the air by employing a different sorbent material.

Claims (11)

1. A method of reducing greenhouse gases in the atmosphere which comprises extracting carbon dioxide from ambient air at a location adjacent or remote from where the carbon dioxide is generated, by exposing sorbent covered surfaces to air streams where the air streams have a flow that is kept laminar, or close to a laminar regime, collecting and storing the carbon dioxide in a concentrated form, wherein the amount of CO 2 generated in collecting, storing and concentrating the CO 2 is less than the amount of CO 2 removed from the atmosphere, generating a carbon credit based on the net amount of carbon dioxide collected and stored, and selling, trading or transferring the resulting carbon credits to a third party.

2. The method of claim 1, comprising one or more of the following features: (a) wherein the surfaces comprise smooth parallel plates; (b) wherein the surfaces are not entirely flat, and follow straight parallel lines in the direction of air flow; (c) wherein the surfaces comprise corrugations, pipes, tubes, angular shapes akin to harmonica covers, or any combination thereof, but with the air flow still following a straight line; (d) wherein the surfaces are roughened with grooves, dimples, bumps or other small structures that are smaller than the surface spacing, and wherein these surface structures remain well within the laminar boundary of the air flow; (e) wherein the surfaces are roughened with grooves, dimples, bumps or other small structures, and the Reynolds number of the flow around these grooves, dimples,. bumps or other small structures is small, optimally between 0 and 100; and (f) wherein the surface is roughened through sand blasting, etching or other similar means.

3. The method of claim 1, comprising one or more of the following features: (a) wherein the direction of the air flow is substantially horizontal; (b) wherein the surfaces - or the line of symmetry of the surfaces - is substantially vertical; (c) wherein the sorbent is a liquid which is flowed at about an approximate right angle to the airflow direction; 35 (d) wherein the surfaces are spaced apart from about 0.3 cm to about 3 cm; (e) wherein the surfaces are at approximate right angles to air flow direction, and are from about 0.30 m to about 10 m in length; (f) wherein air flow speed is from about 0.1m/s to about 10 m/s; (g) wherein the air flow between the surfaces is from about 0.10 m to about 2 m; (h) wherein the sorbent is a liquid which is applied by spraying a flow onto upper edges of the surfaces; (i) wherein the sorbent is a liquid which is applied to both sides of the surfaces; (j) wherein the sorbent is a liquid which is applied in a pulsed manner; (k) wherein the sorbent is a liquid which is flowed on the surfaces or plates and collected at the bottom of the surfaces or plates; (1) wherein the sorbent is a liquid which is flowed on the surfaces or plates and collected at the bottom of the surfaces or plates, and the collected fluid is immediately passed on to a recovery unit; (in) wherein the sorbent is a liquid which is flowed on the surfaces or plates and collected at the bottom of the surfaces or plates, and the collected fluid is recycled to a scrubbing unit for additional CO 2 collection; (n) wherein the method is conducted using apparatus that further comprises and is equipped with air flow straighteners to minimize losses from misalignment between the surfaces and an instantaneous wind field; and (o) wherein the method is conducted using apparatus that further comprises and is equipped with mechanisms that either passively or actively steer the surfaces so that they point into the wind.

4. The method of claim 1, wherein the method is conducted using surfaces that are rotating disks where wetting by sorbent is achieved at least in part by the rotary motion of the disks, and the air is moving substantially at a right angle to the axis.

5. The method of claim 4, comprising one or more of the following features: (a) wherein the axis of rotation is approximately horizontal and the disks are dipped into the sorbent at their rims, whereby said rotary motion promotes distribution of the sorbent on the disks; 36 (b) wherein sorbent is sprayed onto the disk as it is moved by a radially aligned injector; and (c) wherein sorbent is extruded onto the disk adjacent its axis of rotation.

6. The method of claim 1, wherein the surfaces are concentric tubes of circular or other cross-section shape and the air is flowed in a direction of tube axis, and optionally including one or more of the following features: (a) wherein the tubes rotate around a central axis; (b) wherein the tubes have axis oriented approximately vertically and sorbent is applied to flow downward on surfaces of the tubes; and (c) wherein the tubes have axis oriented at an angle to the vertical and the sorbent is inserted at an upper opening and flows downward in a spiral motion covering the entire surface.

7. The method of claim 1, wherein CO 2 is absorbed in a carbonate brine, the carbonate is separated through thermal swing precipitation of the carbonate from the brine; the separated carbonate is electrochemically treated to recover sodium bicarbonate solution in a first electrochemical process step; the bicarbonate is mixed with an acid to release CO 2 in a more concentrated form; and, the acid is recovered from its salt in a second electrochemical process step.

8. The method of claim 7, comprising one or more of the following features: (a) wherein the second electrochemical process comprises electrodialysis with bipolar membranes; (b) wherein at least some of the carbonate is separated in the second electrochemical process step; (c) wherein the brine is neutralized by acid injection; (d) wherein the brine is neutralized in a first low pressure system to adjust the brine to a pH level that supports the formation of bicarbonate, followed by further acidification in a second high pressure system in which CO 2 is generated; (e) wherein CO 2 is released by an electrochemical process in a pressure vessel to provide CO 2 under pressure; (f) wherein the CO 2 is released in an electrochemical process which optionally comprises electrodialysis with bipolar membranes; 37 (g) wherein the CO 2 is released in an electrochemical process which generates hydrogen on the cathodes and uses it again in a hydrogen anode; (h) wherein the sodium bicarbonate is subjected to thermal decomposition into sodium carbonate and CO 2 followed by recycling of the sodium carbonate to an earlier stage of the process; (i) wherein the bicarbonate solution is reduced in water content through membrane separation by concentration gradients or electrochemical gradients (reverse electrodialysis), bicarbonate is extracted from the concentrated brine in a thermal swing precipitation followed by a thermal calcination of the bicarbonate to CO 2 and carbonate, and a resulting dilute bicarbonate output stream is recycled to another dewatering of the bicarbonate solution; (j) wherein the bicarbonate solution is heated until CO 2 is released resulting in a carbonate/bicarbonate brine which is electrochemically reprocessed to bicarbonate; (k) wherein the bicarbonate solution evolves CO 2 inside a pressure vessel; (1) including a heat exchange between inputs and outputs of the thermal steps to minimize energy consumption; (m) wherein dilute water streams generated are treated as off-water; and (n) wherein dilute water streams are used as make-up water in the input in an air contactor unit.

9. The method of any of claims 1-8, wherein a carbon credit is sold, traded or transferred with the sale or lease of an automobile or truck or with fuel for the automobile or truck, or by a producer of a hydrocarbon fuel.

10. The method of any of claims 1-9, wherein the carbon dioxide (C0 2 ) is permanently removed from the atmosphere.

11. A method of creating tradable carbon credits substantially as herein described and claimed.