Multi-bed fast cycle kinetic PSA

Abstract

Disclosed herein is a multi-bed fast cycle pressure swing adsorption (RCPSA) process for separating O2 from N2 and / or Ar wherein the process uses at least five adsorbent beds, each adsorbent bed comprising a kinetically selective adsorbent for O2 having an O2 adsorption rate (1 / s) of at least 0.20 as determined by a linear driving force model at 1 atma and 86 oF.

Description

Technical Field

[0001] This invention relates to a multi-bed rapid cyclic pressure swing adsorption (PSA) method for separating O2 from N2 and / or Ar. Background Technology

[0002] The PSA method has long been used to separate components of air. Recently, there has been considerable interest in intensifying the separation process. In cyclic processes such as PSA and TSA, reducing cycle time is a primary means of achieving greater production from a given amount of material. However, as cycle time decreases, the cyclic process typically faces challenges such as reduced working capacity per cycle for the component of interest, decreased product recovery, and increased pressure drop.

[0003] Recent developments in PSA methods involve using adsorbents with faster adsorption kinetics, such as relatively fast kinetic selectivity laminated adsorbent structures, to improve productivity. However, this increase in productivity typically comes at the cost of reduced selectivity, leading to lower product recovery. Other developments include using adsorbents with relatively slow adsorption kinetics to improve the overall product recovery of the process. However, this improvement in product recovery typically comes at the cost of reduced method productivity.

[0004] US7,645,324 discloses a rotating PSA method for gas kinetic separation using laminated adsorbents. US7,645,324 teaches that using kinetically selective laminated materials can improve productivity; however, to avoid masking the kinetic selectivity through macroporous mass transfer resistance, the macroporous structure within the adsorbent layer should be as open as possible, i.e., the macropore porosity should be relatively high. However, a problem in this regard is that high pore volume often impairs product recovery.

[0005] US9,895,646 discloses a multi-bed PSA method for generating a gas stream rich in compound X from a feed gas stream. US9,895,646 notes that introducing a pressure equalization step into the PSA process can improve product recovery, but doing so generally has an adverse effect on the specific productivity of the process. Notably, shifting the equalization from 1 to 3 allows for a 2.5% efficiency increase, but is detrimental to the 40% increase in adsorbent volume (due to the need for more adsorbent). Therefore, increasing the number of adsorption beds can improve product recovery (allowing for more pressure equalization steps), but this also leads to a decrease in the specific productivity of the method (the standard volumetric flow rate of the product divided by the total amount of adsorbent in the system).

[0006] WO2015 / 199227 discloses a multi-bed (3 or more beds) PSA method for separating methane from biogas. This method performs a pressure equalization process that transfers gas from an adsorption tower where the adsorption process has been completed and the gas is under high pressure to another adsorption tower at a lower pressure, thereby bringing the internal adsorption tower into a medium-pressure state. After completing the depressurization process, a pressure equalization process is performed, bringing the gas from the other adsorption tower to a higher pressure state, thus bringing the internal adsorption tower into a medium-pressure state. This is said to improve the energy efficiency required for pressure increases and decreases in the adsorption towers and also improve the recovery rate and purity of the gas to be purified. However, adding the pressure equalization step does not improve the specific productivity of the method. No kinetic information is provided for the adsorbent used, but the requirement of a long pressure transfer step (6 seconds) suggests the use of a slow-kinetic adsorbent.

[0007] In summary, adsorbents with relatively fast adsorption rates are known to improve method productivity, but this typically comes at the cost of lower product recovery due to larger pore volumes and / or reduced selectivity. Theoretically, introducing more adsorption beds and pressure equalization steps during the use of said adsorbents could improve product recovery; however, doing so is expected to negate the productivity gains achieved by initially using faster adsorbents. Alternatively, slower, more selective kinetic adsorbents could be used to obtain high-purity products with good product recovery, but this also comes at the cost of reduced overall method productivity.

[0008] Therefore, it is clear from the existing technology that there is a trade-off between product recovery and method productivity; method improvements that increase product recovery usually harm method productivity, and vice versa.

[0009] Therefore, there is still a need for PSA methods that offer high method productivity while maintaining high product recovery rates. Summary of the Invention

[0010] The inventors have discovered that when using a “slower” kinetic adsorbent to separate O2 from N2 or Ar via pressure swing adsorption (PSA), switching from a 2-bed PSA method to a PSA method with 5 or more adsorption beds reduces specific productivity with minimal impact on recovery. However, surprisingly, higher specific productivity can be achieved when using a “faster” kinetic adsorbent in a rapid PSA cycle, while maintaining product recovery when switching from a 2-bed to a 5 or more-bed PSA method and incorporating at least two “true” pressure equalization decompression and pressure equalization recompression steps into the process (i.e., the bed pairings involved in the first pressure equalization decompression and first pressure equalization recompression steps are different from those involved in the second pressure equalization decompression and second pressure equalization recompression steps). Therefore, the inventors have now developed a multi-bed rapid cycling PSA method for separating O2 from N2 or Ar with excellent product recovery and method productivity.

[0011] Several preferred aspects of the method according to the invention are outlined below.

[0012] Aspect 1: A multi-bed fast cyclic pressure swing adsorption (RCPSA) method for separating O2 from N2 and / or Ar, wherein the method uses at least five adsorption beds, each adsorption bed comprising components for separation of O2 from N2 and / or Ar via a linear driving force model at 1 atma and 86 atma. o The kinetic selectivity of O2 adsorbents with an O2 adsorption rate of at least 0.20 (1 / s) determined at F, wherein the RCPSA method comprises subjecting each adsorption bed to a rapid PSA cycle, including the following steps performed in sequence: i) Feeding ii) First equilibrium decompression iii) Second equilibrium decompression iv) Countercurrent decompression v) Countercurrent sweeping vi) First Equilibrium Repressurization vii) Second Equilibrium Repressurization viii) Product and / or feed repressurization When the adsorption bed undergoes the equalization depressurization step ii), it is connected to one of the other adsorption beds simultaneously undergoing the equalization repressurization step vii) and provides a repressurization flow. When the adsorption bed is undergoing the equalization depressurization step iii), it is connected to another adsorption bed that is simultaneously undergoing the equalization repressurization step vi) and provides a repressurization flow.

[0013] Aspect 2: The RCPSA method of aspect 1, wherein step iii) is a double-equilibrium decompression step and step vi) is a double-equilibrium recompression step.

[0014] Aspect 3: The RCPSA method of aspect 1 or 2, wherein step viii) is a product and feed repressurization step.

[0015] Aspect 4: The RCPSA method of any one of Aspects 1 to 3, wherein step ii) is a downstream equalization depressurization step and step vii) is a counter-current equalization repressurization step.

[0016] Aspect 5: The RCPSA method of any one of Aspects 1 to 4, wherein the method uses 5 to 18 adsorption beds.

[0017] Aspect 6: The RCPSA method of any one of Aspects 1 to 4, wherein the method uses 7 to 9 adsorption beds.

[0018] Aspect 7: The RCPSA method of any one of Aspects 1 to 4, wherein the method uses 7 or 9 adsorption beds.

[0019] Aspect 8: The RCPSA method of any one of Aspects 1 to 7, wherein the duration of the feeding step is 3 to 45 seconds.

[0020] Aspect 9: The RCPSA method of any one of Aspects 1 to 8, wherein the duration of each of the equalization decompression and equalization recompression steps is 1 to 5 seconds.

[0021] Aspect 10: The RCPSA method of any one of Aspects 1 to 9, wherein the cycle time of the fast PSA cycle is equal to or less than 100 seconds.

[0022] Aspect 11: The RCPSA method of any one of Aspects 1 to 10, wherein the feeding step is in 0 o F to 125 o Performed at a temperature of F.

[0023] Aspect 12: The RCPSA method of any one of Aspects 1 to 10, wherein the feeding step is in 20 o F to 100 o Performed at a temperature of F.

[0024] Aspect 13: The RCPSA method of any one of Aspects 1 to 10, wherein the feeding step is in 20 o F to 40 o Performed at a temperature of F.

[0025] Aspect 14: The RCPSA method of any one of Aspects 1 to 13, wherein during all or part of the feeding step, circulating gas is introduced co-currently into the bed in which the step is performed, the circulating gas including gas obtained during the counter-current depressurization step and / or from a purging step of the bed undergoing the step.

[0026] Aspect 15: The RCPSA method of any one of Aspects 1 to 14, wherein during all or part of the equalization decompression step ii), circulating gas is introduced in a co-current manner into the bed in which the step is performed, the circulating gas including gas obtained during the counter-current decompression step and / or a purging step from the bed undergoing the step.

[0027] Aspect 16: An RCPSA method of any one of Aspects 1 to 15, wherein the kinetically selective adsorbent has the properties of 1 atm and 86 atm through a linear driving force model. o At least 5% O2 / N2 kinetic selectivity was determined at F, and / or by a linear driving force model at 1 atm and 86. o At least 5% O2 / Ar kinetic selectivity was determined at F.

[0028] Aspect 17: An RCPSA method of any one of Aspects 1 to 16, wherein the kinetically selective adsorbent is a zeolite or a carbon molecular sieve.

[0029] Aspect 18: An RCPSA method for any of Aspects 1 to 17, wherein the method separates O from Ar. 2， Furthermore, the kinetic selective adsorbent is RHO zeolite with a Si / Al ratio of 3.2 to 4.5 and containing aprotic extra-framework cations, wherein each unit cell of the zeolite contains at most one proton, and wherein the size, number, and charge of the extra-framework cations present in the zeolite are such that each unit cell requires one or fewer aprotic extra-framework cations to occupy an octagonal site.

[0030] Aspect 19: An RCPSA method of any one of Aspects 1 to 17, wherein the method separates O2 from N2, and the kinetically selective adsorbent is a carbon molecular sieve (CMS) having a linear driving force model at 1 atma and 86 o The O2 / N2 kinetic selectivity was determined at F from 5 to 30.

[0031] Aspect 20: The RCPSA method of any one of Aspects 1 to 19, wherein the method is a rotating bed RCPSA method.

[0032] Aspect 21: The RCPSA method of any one of Aspects 1 to 19, wherein the method is the rotary valve RCPSA method.

[0033] Aspect 22: The RCPSA method of any one of Aspects 1 to 21, wherein the void volume of each adsorption bed is 3% to 15% of the bed volume.

[0034] Aspect 23: An RCPSA method of any one of Aspects 1 to 22, wherein the method uses seven adsorption beds, and wherein the RCPSA method includes subjecting each adsorption bed to a rapid PSA cycle, comprising the following steps performed in sequence: Feed (F); Flow-balanced pressure reduction (EQD1); Double-equilibrium decompression (DEQD2); Countercurrent decompression (CnD); Countercurrent purging (PU); Dual Equalization Repressurization (DEQR2); Countercurrent equalization and repressurization (EQR1); and Product and feed repressurization (RP / F); When the adsorption bed undergoes the co-current equilibrium depressurization (EQD1) step, it is connected to one of the other adsorption beds simultaneously undergoing the counter-current equilibrium repressurization (EQR1) step and provides the counter-current repressurization flow. When the adsorption bed performs the double equalization depressurization (DEQD2) step, it is connected to another adsorption bed that is simultaneously performing the double equalization repressurization (DEQR2) step, providing both co-current and counter-current repressurization flows.

[0035] Aspect 24: An RCPSA method of any one of Aspects 1 to 22, wherein the method uses nine adsorption beds, and wherein the RCPSA method includes subjecting each adsorption bed to a rapid PSA cycle, comprising the following steps performed in sequence: Feed (F); Flow-balanced pressure reduction (EQD1); First dual equalization decompression (DEQD2); Second Dual Equalization Decompression (DEQD3); Countercurrent decompression (CnD); Countercurrent purging (PU); First Dual Equalization Repressurization (DEQR3); Second Dual Equalization Repressurization (DEQR2); Countercurrent equalization and repressurization (EQR1); and Product and feed repressurization (RP / F); When the adsorption bed undergoes the co-current equilibrium depressurization (EQD1) step, it is connected to one of the other adsorption beds simultaneously undergoing the counter-current equilibrium repressurization (EQR1) step, providing the counter-current repressurization flow. When the adsorption bed undergoes the first double equalization depressurization (DEQD2) step, it is connected to another adsorption bed simultaneously undergoing the second double equalization repressurization (DEQR2) step, providing both co-current and counter-current repressurization flows. When the adsorption bed undergoes the second double equalization depressurization (DEQD3) step, it is connected to another adsorption bed that is simultaneously undergoing the first double equalization repressurization (DEQR3) step, providing both co-current and counter-current repressurization flows. Attached Figure Description

[0036] Figure 1 The operation of a 2-bed multi-step PSA cycle is shown.

[0037] Figure 2 The operation of a 9-bed multi-step PSA cycle is shown.

[0038] Figure 3 The operation of a 7-bed multi-step PSA cycle is shown.

[0039] Figure 4 The operation of a 4-bed multi-step PSA cycle is shown.

[0040] Figure 5 The operation of the 18-bed multi-step PSA cycle is shown.

[0041] Figure 6 The operation of a 9-bed multi-step PSA cycle is shown, in which the exhaust gas is recirculated to the PSA feed step.

[0042] Figure 7 The operation of a 9-bed multi-step PSA cycle is shown, in which the exhaust gas is recirculated to the PSA equalization step.

[0043] Figure 8 The operation of a 9-bed multi-step PSA cycle is shown, in which the exhaust gas is recirculated to the PSA equalization step.

[0044] Figure 9a , 9b Models 9c, 9d, 9e, and 9f are 2-bed models demonstrating the production of nitrogen from air using different carbon molecular sieve adsorbents (414-01, 414-02, and 414-03 adsorbents). Figure 1 The cycle shown in the diagram) and 9-bed ( Figure 2 The graph shows a comparison of the performance of the multi-step PSA cycle method (as a function of cycle time, in terms of "productivity" and "feed to product ratio").

[0045] Figure 10 It shows the use of carbon molecular sieve adsorbent 414-02 and Figure 2 The 9-bed PSA method shown in the diagram cycles at 100 o Figures showing the purging requirements for producing nitrogen products containing different amounts of oxygen at F and 7.80 atma.

[0046] Figure 11a and 11bThis shows the pore volume effect when using carbon molecular sieve adsorbent 414-02 and Figure 2 The 9-bed PSA method shown in the diagram cycles at 100 o A graph showing the effect of F and 7.80 atma on the performance of the method for producing nitrogen products containing 4500 ppm oxygen (in terms of "productivity" and "feed to product ratio").

[0047] Figure 12a and 12b This shows that the standardized performance (in terms of "productivity" and "feed to product ratio") for 414-05 is 69.8. o Figures for F and 7.80 atma of 414-06, a “fast” carbon molecular sieve adsorbent, from a 9-bed test unit (bed length = 40 inches, bed ID = 4.0 inches). Also included are the performance figures for 414-04, a “slow” carbon molecular sieve adsorbent, from a 2-bed PSA cycle (bed length = 120 inches, bed ID = 1.908 inches).

[0048] Figure 13a and 13b This indicates the use of 414-05 and 414-06 carbon molecular sieve adsorbents from a 9-bed test unit (bed length = 40 inches, bed ID = 4.0 inches), with a bed pressure of 69.8. o The effects of F and 7.80 atma on the performance of producing a nitrogen product containing 45,000 ppm oxygen (in terms of normalized "productivity" and normalized "feed to product ratio"). Also included are the performance of conventional carbon molecular sieve adsorbent 414-04 in producing a nitrogen product containing 45,000 ppm oxygen from a 2-bed PSA cycle (bed length = 120 inches, bed ID = 1.908 inches) under the same conditions, and all data were normalized. Detailed Implementation

[0049] The following detailed description provides only preferred exemplary embodiments and is not intended to limit the scope, applicability, or configuration of the invention. Rather, the subsequent detailed description of preferred exemplary embodiments will provide those skilled in the art with a feasible description of how to implement these preferred exemplary embodiments. Various changes may be made to the function and arrangement of the elements without departing from the spirit and scope of the invention as set forth in the appended claims.

[0050] When applied to any feature in the embodiments of the invention described in the specification and claims, the articles “a” and “an” as used herein mean one or more. Unless specifically stated otherwise, the use of “a” and “an” does not limit the meaning of a single feature. The article “the” preceding a singular or plural noun or noun phrase indicates a particular feature or specific characteristic and may have a singular or plural meaning depending on the context in which it is used.

[0051] As used herein, "first," "second," "third," etc., are used to distinguish multiple steps and / or features and do not indicate a total number, or relative position in time and / or space, unless explicitly stated otherwise.

[0052] As used in this article, the term "comprising" means being included by or comprising.

[0053] As used herein, the phrase “and / or” placed between the first entity and the second entity includes any of the following meanings: (1) only the first entity, (2) only the second entity, and (3) both the first entity and the second entity. The term “and / or” placed between the last two entities in a list of three or more entities means at least one entity in the list, including any particular combination of entities in the list. For example, “A, B and / or C” has the same meaning as “A and / or B and / or C” and includes the following combinations of A, B and C: (1) only A, (2) only B, (3) only C, (4) A and B but not C, (5) A and C but not B, (6) B and C but not A, and (7) A and B and C.

[0054] This invention discloses a multi-bed fast cyclic pressure swing adsorption (RCPSA) method for separating O2 from N2 and / or Ar, wherein the method utilizes at least five adsorption beds, each containing a “fast” kinetically selective adsorbent for O2, and the PSA cycle includes at least two “real” pressure equalization depressurization steps and two “real” pressure equalization repressurization steps. Such a method has been found to provide good product recovery and good method productivity.

[0055] As used herein, the term "fast" kinetically selective adsorbent for O2 refers to a kinetically selective adsorbent for O2 that has an O2 adsorption rate of at least 0.20 (1 / s), which is achieved by latma and 86 o F was determined using a linear driving force model. More preferably, the O2 adsorption rate (1 / s) of the fast CMS adsorbent 4 is at least 0.21, at least 0.22, at least 0.23, at least 0.24, at least 0.25, at least 0.26, at least 0.27, at least 0.28, at least 0.29, or at least 0.30, as determined by latma and 86. o The linear driving force model of F is determined.

[0056] As used herein, the term "slow" kinetically selective adsorbent for O2 refers to an O2 adsorbent with an O2 adsorption rate (1 / s) less than 0.2000, determined by a linear driving force model at latma and 86. o F determination.

[0057] The linear driving force (LDF) model is a well-known model for determining adsorption rates from experimental absorption curves and can be used to calculate the adsorption rate of a particular adsorbent on another adsorbent, as well as the kinetic selectivity of the adsorbent on a particular adsorbate.

[0058] More specifically, the adsorption rate of adsorbates (e.g., O2, N2, or Ar) on the adsorbent was evaluated using a standard volumetric adsorption apparatus. The experiment involved initially applying the adsorbate to a vacuum and at 303 K (86... o The adsorbent sample at F) was exposed to a measured amount of adsorbate at the same temperature at 1 atma (760 Torr / 101 kPa). Pressure changes were recorded as a function of time. Then, using the same weight of quartz beads instead of the adsorbent sample, the pressure-time data were subtracted from a similar pressure history to obtain a graph of the amount of adsorbed gas as a function of time, also known as the absorption curve. The adsorption rate of the adsorbate was then extracted using the LDF model in reverse time (1 / s) of the uptake curve. The selectivity of the adsorbent for a specific adsorbate pair can be determined individually by the ratio of the calculated adsorption rates of the two adsorbates. The analytical form of this model is given by the following equation (also in the literature Sircar, S. and Hufton, JR, “Why Does the Linear Driving Force Model for Adsorption Kinetics Work”). (Listed in Table 1 of Adsorption 2000, 6, 137-147), where f(t) is the fractional uptake, k is the LDF mass transfer coefficient, a is the correction factor for constant volume experiments, and t is time.

[0059] As stated above, the two “true” pressure equalization decompression steps and two “true” pressure equalization repressurization steps mentioned in this article mean that the bed pairings involved in the first pressure equalization decompression and the first pressure equalization repressurization steps are different from those involved in the second pressure equalization decompression and the second pressure equalization repressurization steps. For example, if the first pressure equalization decompression step in bed “A” occurs with bed “B” through the bed equalization pressure (therefore, the first pressure equalization repressurization step is in progress), then in order for the second pressure equalization decompression step to be a “true” second pressure equalization decompression step, the second pressure equalization decompression in bed “A” must be performed with a third bed, bed “C” (therefore, the second pressure equalization repressurization step is in progress) instead of bed “B”, through the bed equalization pressure.

[0060] The RCPSA method involves subjecting each adsorption bed to a rapid PSA cycle, which includes steps performed at least as follows: i) feed; ii) first equalization depressurization; iii) second equalization depressurization; iv) countercurrent depressurization; v) countercurrent purging; vi) first equalization repressurization; vii) second equalization repressurization; viiii) product and / or feed repressurization. It should be noted that the cycle may also include other steps occurring before, after, or between any of the steps listed above. When an adsorption bed undergoes the first equalization depressurization step ii), it is connected to one of the other adsorption beds and provides a repressurization flow, while the second equalization repressurization step vii) is performed. When an adsorption bed undergoes the second equalization depressurization step iii), it is connected to another adsorption bed and provides a repressurization flow, while the first equalization repressurization step vi) is performed. The reference to "another" adsorption bed here indicates that the bed undergoing the cycle, during step iii), is connected to a different bed that was previously connected to the bed during step ii). In other words, steps ii) and iii) represent two “real” pressure equalization decompression steps, and steps vi) and vii) represent two “real” pressure equalization repressurization steps.

[0061] As used herein, the term "rapid" PSA cycle means that the total duration of the feed step (i.e., step i) is preferably 45 seconds or less. The feed step (also commonly referred to as the adsorption step) is a cyclic step in which the adsorption bed is under high pressure (relative to the pressure in the bed during the process steps), introducing and passing a feed stream through the bed to adsorb one or more components from the feed stream to produce a product stream exiting the bed, which depletes the adsorbed components (relative to the composition of the feed stream). Preferably, the total duration of the feed step is at least 3 seconds. Preferably, the total duration of the feed step is 3 to 45 seconds or 3 to 16 seconds.

[0062] The fast PSA cycle preferably also has a cycle time of 100 seconds or less, where cycle time is the amount of time taken to complete a complete set of steps in the PSA cycle. More preferably, the PSA cycle has a cycle time of 60 seconds or less, 50 seconds or less, or 40 seconds or less. Preferably, the PSA cycle has a cycle time of at least 15 seconds.

[0063] It has been found that a duration of 1 to 5 seconds is preferred for each of the equalization depressurization step and the equalization repressurization step. Faster than this (i.e., <1 second) leads to reduced system efficiency, and slower than this (i.e., >5 seconds) increases cycle time and reduces overall productivity. Therefore, each of steps ii), iii), vi), and vii) preferably has a duration of 1 to 5 seconds. Furthermore, since steps ii) and vii) are connected, and steps iii) and vi) are connected (step ii) and vii) occur simultaneously between different beds where pressure equalization occurs, and similarly steps iii) and vi) occur simultaneously between different beds where pressure equalization occurs, it can be understood that steps ii) and vii) have the same duration, and steps iii) and vi) have the same duration (therefore, if step ii) has a duration of, for example, 4 seconds, then step vii) also has a duration of 4 seconds).

[0064] As used herein with respect to the PSA cycle, a “parallel flow” step refers to a step in which gas flows into and / or out of the bed in the same direction as the gas flow in the feed step. Similarly, a “counter-flow” step refers to a step in which gas flows into and / or out of the bed in the opposite direction to the gas flow in the feed step. Likewise, as used herein, the “inlet” or “inlet” end of the bed refers to the inlet or end of the bed through which the feed gas enters in the feed step, and the “outlet” or “outlet” end of the bed refers to the outlet or end of the bed through which the product gas exits in the feed step.

[0065] The term "double" indicates that gas either leaves the bed from either end or enters the bed from either end. Therefore, in a double repressurization step, gas enters the bed from both ends, and in a double depressurization step, gas leaves the bed from both ends.

[0066] Feeding step i) is usually in about 0 o F to approximately 125 o The experiment is carried out at a temperature of F, more preferably at about 20°C. o F to approximately 100 o F or approximately 20 o F to approximately 40 o The process is carried out at a temperature of F. As used herein, the temperature of the feed step refers to the temperature at which the feed gas is introduced into the adsorption bed, as measured at the inlet of the adsorption bed. Surprisingly, the inventors have discovered that in the methods described and disclosed herein, at lower temperatures (e.g., about 30°C), the process can be more efficient. oThe feed step is performed at F), which further improves the method productivity and product recovery, in contrast to known PSA systems where product recovery is improved at the expense of reduced method productivity. Performing the feed step at such a lower temperature may be particularly advantageous where the feed gas is already available at temperatures below room temperature (e.g., in the case of a method for separating oxygen and / or nitrogen from a crude argon stream derived from a distillation column), thus eliminating the need for significant cooling of the feed gas to obtain the lower temperature feed stream used in the feed step.

[0067] Any suitable feed pressure can be used during the feeding step. For example, the pressure during the feeding step, measured at the inlet of the adsorption bed, can be from about 5 to about 12 atmospheres absolute.

[0068] If the method is used to separate O2 from N2, the kinetically selective adsorbent used in the method according to the invention preferably has an O2 / N2 kinetic selectivity of at least 5, such as by latma and 86. o F is determined by a linear driving force model. Similarly, if this method is used to separate O from Ar... 2， The kinetically selective adsorbent used in the method according to the invention preferably has an O2 / Ar kinetic selectivity of at least 5, such as by latma and 86 o The dynamic selectivity is determined by the linear driving force model of F. The dynamic selectivity is determined by the LDF model as described above.

[0069] The adsorbent can be any suitable type of material. Suitable adsorbents include carbon molecular sieves (CMS); any kinetic adsorbent based on porous zeolites with a fast absorption rate, such as (but not limited to) RS-10, RHO, or chalcogenide zeolite. How to produce adsorbents with the desired "fast" kinetic adsorption performance is known in the art.

[0070] For example, methods for producing CMS adsorbents that exhibit high oxygen kinetic selectivity for nitrogen or oxygen over argon have been established. For this purpose, selective micropores are introduced into the adsorbent, where the pore diameter controls the adsorption kinetics of oxygen, nitrogen, or argon and is therefore considered a critical size for diffusion. Such micropores can exhibit very high selectivity, typically achieved at the expense of the overall adsorption rate. On the other hand, the adsorption rate can be increased at the expense of selectivity. Suitable CMS adsorbents for the RCPSA method of this invention can, for example, have O2 / N2 kinetic selectivity of 5 to 30, 10 to 25, or 15 to 20, and / or O2 / Ar kinetic selectivity of 5 to 40. CMS can have any suitable adsorption capacity. For example, it can have an adsorption capacity of 0.2-0.4 mmol / g at equilibrium, as measured by 1 atm and 86 o Standard isotherm measurement at F.

[0071] The zeolite suitable for the RCPSA method of this invention can be an RHO zeolite with a Si / Al ratio of 3.2-4.5 and containing aprotic extra-framework cations, wherein each unit cell of the zeolite contains at most one proton, and wherein the size, number, and charge of the extra-framework cations present in the zeolite are such that one or fewer aprotic extra-framework cations are required per unit cell to occupy an 8-ring site. Such RHO zeolites are disclosed in U.S. Patent Applications Nos. USSN15 / 718,467 and USSN15 / 718,620, filed September 28, 2017, the contents of which are incorporated herein by reference in their entirety.

[0072] Suitable chalcogenide adsorbents are disclosed in US9,669,349 and US9,925,514, the contents of which are incorporated herein by reference in their entirety.

[0073] The adsorption bed can be filled with any suitable form of adsorbent, but preferably contains an adsorbent in the form of random packing. The adsorbent should preferably be dense and held in place by a permeable compression system, such as a porous plate with a wire mesh facing the CMS to accommodate it and downward pressure applied, for example, by a spring. A dense load prevents additional settling, thus relieving the spring. A dense load can be achieved by methods known in the art, such as a snowfall-type load.

[0074] Preferably, each adsorption bed used in the method has a void volume of about 3% to about 15% relative to the total bed volume. Most preferably, each adsorption bed used in the method has a void volume of about 5% to about 13%, about 7% to about 12%, or about 10% relative to the total bed volume. As used herein, “void volume relative to the total bed volume” means the additional column voids divided by the total bed volume.

[0075] The process for RCPSA can be performed using any suitable apparatus. Conventional switching valves can achieve a certain level of efficiency. However, the preferred RCPSA method is a rotary bed RCPSA method using a rotary bed PSA apparatus (where the adsorption bed is mounted in a rotor located between and rotating relative to the feed and product stator assemblies, each stator assembly containing a stator plate that acts as a valve plate for switching the bed between cyclic PSA steps), or a rotary valve RCPSA method using a rotary valve PSA apparatus (in contrast, the adsorption bed is located in a fixed bed assembly and is switched between PSA steps by rotating the feed and product valves known in the art).

[0076] Now refer to Figures 1 to 8 Describe several multi-bed, multi-step PSA cycles, among which Figure 1 and Figure 4A comparative PSA cycle performed in fewer than 5 beds and not in accordance with the present invention is shown. Figure 2 , 3 Figures 5-8 illustrate exemplary PSA cycles employing five or more beds suitable for use according to the invention. In this regard, it should be noted that the term "adsorbent bed" or "bed" as used herein refers to one or more adsorbent containers, each simultaneously undergoing each step of the PSA cycle. Thus, a process using two beds has a first container or set of containers simultaneously undergoing each step of the PSA cycle (this represents one of the beds), and a second container or set of containers simultaneously undergoing each step of the PSA cycle (this represents another bed), however, the timing of the PSA cycle in the first container or set of containers is staggered from the timing of the PSA cycle in the second container or set of containers, such that each step of the PSA cycle in the first container or set of containers does not begin and end simultaneously with the start and end of the corresponding step of the PSA cycle in the second container or set of containers. All these designs are contained in tabular form in the cycle schedule. In these schedules, each row of the grid represents all the different cycle steps a given bed undergoes throughout the cycle, while a column of the grid indicates which bed runs which cycle step at a particular unit time step. The total cycle time is the sum of all the individual unit time steps in a particular row. Note that in the cyclic scheduling, there is a highlighted section, often called the "unit block," within which all steps are run by a single multi-bed (Mehrotra et al., 2011). The number of repeating unit blocks is the same as the number of beds. Typical cyclic scheduling methods can be found in the literature "Mehrotra, A.; Ebner, AD; Ritter, JA Simplified Graphical Approach for Complex PSA CycleScheduling," Adsorption It was found in “,2011,17337-345”. Figures 1 to 8 The salient features of all the loop designs presented are summarized below: 2-bed circulation ( Figure 1 ) Figure 1A two-bed PSA cycle, not according to the invention, is shown, comprising components for comparative purposes. The cycle includes a product and feed repressurization step (RP1 / F1), a feed step (F2, F3), a co-current depressurization step (CoD), a co-current and dual equilibrium depressurization step (EQD1, DEQD2, DEQD3), a counter-current depressurization step (CnD1, CnD2), a product purging step (PU1, PU2), a product repressurization step (RP2, RP3), a counter-current and dual equilibrium repressurization step (EQR1, DEQR2, DEQR3), and various idle steps (II, 12, 13). In the product and feed repressurization (F1 / RP1) step, the bed is pressurized to the highest pressure level of the cycle by introducing a primary product (e.g., enriched N2) and a feed gas (e.g., air) from the top and bottom of the bed, respectively. No product is removed during this step of the cycle. Next, in the feed step (F2 and F3 represent stages of the same feed step), feed gas continues to be introduced until the mass transfer zone of the preferentially adsorbed component (i.e., O2) reaches the outlet end of the bed without substantially disrupting it, and unadsorbed gas (e.g., purified N2) is discharged from the outlet end of the bed as the main product. The feed flow is then stopped in the co-current depressurization (CoD) step, and the bed pressure is reduced to a first intermediate level by extracting the product from the top of the bed. Then, in the co-current equilibrium depressurization (EQD1) step, the bed is connected to the bed that underwent the counter-current equilibrium repressurization (EQR1) step, and a portion of the voids and desorbed gas are transferred from the product end of the bed undergoing EQD1 to the product end of the bed undergoing EQR1, thereby partially balancing the pressure between the two beds and reducing the bed pressure in the bed undergoing EQD1 to a second intermediate level at the end of this step. Next, in the double-end equalization depressurization step (DEQD2), more co-adsorbed gas and air gas are discharged from the top and bottom of the first bed that has undergone the step, reducing the pressure in the bed to a third intermediate level. The gas extracted from the top and bottom of the bed is then placed at the top and bottom of the bed, respectively, for a double equalization repressurization step (DEQR2). Next, double-end equalization depressurization continues in step (DEQD3 / CnD1), but some of the depressurization step gas extracted in the countercurrent is used as secondary products (i.e., O2-rich gas) instead of being used to repressurize another bed, which further reduces the bed pressure to a fourth intermediate level. Then, in the countercurrent depressurization step (CnD2), the bed is countercurrently depressurized to the minimum operating pressure and purged at the cyclic minimum operating pressure in the purging steps (PU1 and PU2), producing further secondary products. Subsequently, in the countercurrent repressurization steps (RP2 and RP3), the bed is countercurrently repressurized with the primary product gas. Following the repressurization steps (RP2 and RP3), the tower is further pressurized through pressure equalization repressurization steps (EQR1, DEQR2 and DEQR3) to restore the pressure level used for startup and repeated cycles.Note that the three idling steps (I1, I2, and I3) are combined into the cycle time, during which the bed undergoing the idling step is isolated and the valve is turned off.

[0077] Although three separate equalization decompression steps and three separate equalization repressurization steps are listed above, it should be noted that all three equalization decompression steps (EQD1, DEQD2, DEQD3) and equalization repressurization steps (EQR1, DEQR2, DEQR3) in this cycle occur between the same two beds (this is unavoidable because the cycle is a 2-bed cycle, hence only two beds). Therefore, in the context of this application, the process has only one “real” equalization decompression step and one “real” equalization repressurization step, where steps EQD1, DEQD2, and DEQD3 represent stages of the same pressure equalization decompression step, and steps EQR1, DEQR2, and DEQR3 represent stages of the same equalization repressurization step.

[0078] 9-bed circulation ( Figure 2 ) Figure 2 A 9-bed PSA method applicable to one embodiment of the present invention is shown. This process includes feed (F1-F4), co-current equalization depressurization 1 (EQD1), double equalization depressurization 2 (DEQD2), double equalization depressurization 3 (DEQD3), counter-current depressurization (CnD1 to CnD3), counter-current product purging (PU1 to PU4), double equalization repressurization 3 (DEQR3), double equalization repressurization 2 (DEQR2), counter-current equalization repressurization 1 (EQR1), and product and feed repressurization (RP / F5). Each step is described in detail below.

[0079] i) Feeding Steps (F1-F4): Assuming the bed has been pre-pressurized to the highest pressure level of the cycle, the feed gas mixture is introduced into the inlet of the bed, and unadsorbed gas is discharged from the outlet of the bed. The feeding steps continue until the mass transfer zone (MTZ) of the preferentially adsorbed component reaches the outlet of the bed without significantly disrupting it.

[0080] ii) Flow Equalization Pressure Reduction 1 (EQD1): At the end of the feed step, the bed is connected to another bed in the EQR1 step (described below). A portion of the voids and the desorbed gas are transferred from the product (outlet) end of the bed in EQD1 to the product (outlet) end of the other bed in EQR1, thereby reducing the bed pressure in EQD1 to a first intermediate level.

[0081] iii) Double Equalization Decompression 2 (DEQD2): After EQD1, the bed is decompressed to a second intermediate pressure level by connecting the bed to another bed in the DEQR2 step (described below).

[0082] The effluent containing voids and desorbed gas is transferred from the bed in DEQD2 to another bed in DEQR2 through the feed (inlet) and product (outlet) ends.

[0083] iv) Double Equilibrium Pressure Reduction 3 (DEQD3): Following DEQD2, the pressure in another bed is further reduced to a third intermediate level by transferring the void and desorbed gas to another bed in the DEQR3 step (described below). Pressure is again transferred through the feed (inlet) end and the product (outlet) end of the bed.

[0084] v) Countercurrent depressurization (CnD1-CnD3): After DEQD3, the bed is depressurized by countercurrent flow to or near ambient pressure levels at the feed (inlet) end. The effluent containing more selectively adsorbed substances is removed as a second product.

[0085] vi) Countercurrent product purging (PU1-PU4): At the end of the countercurrent depressurization step, purging gas is introduced into the product (outlet) end of the bed, and the effluent is collected from the feed (inlet) end of the bed as the second product.

[0086] vii) Double Equilibrium Repressurization 3 (DEQR3): After the purging step, the bed pressure is restored from atmospheric pressure to a third intermediate level (as described in the DEQD3 step) by introducing voids and desorbed gas from another bed that has undergone DEQD3. Note that the gas is transferred through both ends of the column.

[0087] viii) Double Equilibrium Repressurization 2 (DEQR2): After DEQR3, the bed pressure is restored to a second intermediate level (as described in the DEQD2 step) by introducing voids and desorbed gas from another bed that has undergone DEQD2. This is also a two-end pressure transfer step.

[0088] ix) Countercurrent Equilibrium Repressurization 1 (EQR1): After DEQR2, the bed pressure is brought back to the first intermediate level (as described in step EQD1) by introducing voids and desorbed gas from another bed that has undergone EQD1. Note that the gas is transported through the product (outlet) end of the column.

[0089] x) Product and Feed Repressurization (RP / F5): Finally, the bed pressure is returned to the feed or maximum pressure level of the cycle to start and repeat the cycle. This step is initiated by introducing feed and a portion of the main product gas from the feed (inlet) and product (outlet) ends, respectively.

[0090] In this method, two beds are continuously fed in the cycle, so product extraction is continuous. Due to the multi-bed nature of this method, flow fluctuations in the feed and product streams can be reduced / eliminated. In the final stage of pressure equalization, during step EQR1, the beds are pressurized using a nearly pure void gas containing almost all primary products. Therefore, the pressure equalization is more efficient than... Figure 1 The conventional two-bed dynamic PSA method presented in the paper is more efficient in terms of circulation.

[0091] 7-bed circulation ( Figure 3 ) Figure 3 A 7-bed PSA method applicable to another embodiment of the present invention is shown. The process includes feed (F1 to F4), co-current equalization depressurization 1 (EQD1), double equalization depressurization 2 (DEQD2), counter-current depressurization (CnD1 and CnD2), counter-current product purging (PU1 to PU3), double equalization repressurization 2 (DEQR2), counter-current equalization repressurization 1 (EQR1), and product and feed repressurization (RP / F5). Each step is described in detail below.

[0092] i) Feeding Steps (F1 to F4): Assuming the bed has been pre-pressurized to the highest pressure level of the cycle, the feed gas mixture is introduced into the inlet of the bed, and unadsorbed gas is discharged from the outlet of the bed. The feeding steps continue until the mass transfer zone (MTZ) of the preferentially adsorbed component reaches the outlet of the bed without substantially disrupting it.

[0093] ii) Flow Equalization Pressure Reduction 1 (EQD1): At the end of the feed step, the bed is connected to another bed for step EQR1 (described below). In step EQD1, a portion of the voids and desorbed gas is transferred from the product (outlet) end of one bed to the product (outlet) end of the other bed, thereby reducing the bed pressure in EQD1 to a first intermediate level.

[0094] iii) Dual Equilibrium Decompression 2 (DEQD2): Following EQD1, the bed is decompressed to a second intermediate pressure level by connecting it to another bed that has undergone step DEQR2 (described below). The effluent containing voids and desorbed gases is transferred from the bed in DEQD2 to another bed in DEQR2 through the feed (inlet) and product (outlet) ends.

[0095] iv) Countercurrent depressurization (CnD1 and CnD2): After DEQD2, the bed is depressurized by countercurrent flow to or near ambient pressure levels at the feed (inlet) end. The effluent containing more selectively adsorbed substances is removed as a second product.

[0096] v) Countercurrent product purging (PU1 to PU3): At the end of the countercurrent depressurization step, purging gas is introduced into the product (outlet) end of the bed, and the effluent is collected from the feed (inlet) end of the bed as a second product.

[0097] vi) Double Equilibrium Repressurization 2 (DEQR2): After the purging step, the bed pressure is restored to a second intermediate level (as described in the DEQD2 step) by introducing voids and desorbed gas from another bed that has undergone DEQD2. This is again a two-end pressure transmission mode.

[0098] vii) Countercurrent Equilibrium Repressurization 1 (EQR1): After DEQR2, the bed pressure is brought back to the first intermediate level (as described in step EQD1) by introducing voids and desorbed gas from another bed that has undergone EQD1. Note that the gas is transported through the product (outlet) end of the column.

[0099] viii) Product and Feed Repressurization (RP / F5): Finally, the bed pressure is restored to the feed or maximum pressure level of the cycle to start and repeat the cycle. This step is initiated by introducing feed gas and a portion of the main product gas from the feed (inlet) and product (outlet) ends, respectively.

[0100] In this method, two beds are continuously fed in the cycle, so product extraction is continuous. Due to the multi-bed nature of this method, flow fluctuations in the feed and product streams can be reduced / eliminated. In the final stage of pressure equalization, during step EQR1, the beds are pressurized using a nearly pure void gas containing almost all primary products. Therefore, the pressure equalization is more efficient than... Figure 1 The conventional two-bed dynamic PSA method presented in the paper is more efficient in terms of circulation.

[0101] 4-bed circulation ( Figure 4 ) Figure 4 The diagram shows a 4-bed PSA method not based on the present invention, which includes components for comparative purposes. The process includes feed (F1 and F2), co-current equalization depressurization (EQD), counter-current depressurization (CnD), counter-current product purging (PU1 and PU2), counter-current equalization repressurization (EQR), and product and feed repressurization (RP / F3) steps. Each step is described in detail below.

[0102] i) Feeding steps (F1 and F2): Assuming the bed has been pre-pressurized to the highest pressure level of the cycle, the feed gas mixture is introduced into the inlet end of the bed, and unadsorbed gas is discharged from the outlet end of the bed. The feeding steps continue until the mass transfer zone (MTZ) of the preferentially adsorbed component reaches the outlet end of the bed without substantially disrupting it.

[0103] ii) Equalized Pressure Reduction (EQD): At the end of the feed step, the bed is connected to another bed in step EQR (described below). A portion of the voids and desorbed gas is transferred from the product (outlet) end of the bed in EQD to the product (outlet) end of the other bed in EQR, thereby reducing the bed pressure in EQD to an intermediate level.

[0104] iii) Countercurrent depressurization (CnD): After EQR, the bed is depressurized by countercurrent flow to or near ambient pressure levels at the feed (inlet) end. The effluent containing more selectively adsorbed substances is removed as a second product.

[0105] iv) Countercurrent product purging (PU1 and PU2): At the end of the countercurrent depressurization step, purging gas is introduced into the product (outlet) end of the bed, and the effluent is collected from the feed (inlet) end of the bed as a second product.

[0106] v) Countercurrent Equilibrium Repressurization (EQR): After the purging step, the bed pressure is restored to an intermediate level (as described in the EQD step) by introducing voids and desorbed gas from another bed that has undergone EQD. Note that the gas is transferred through the product (outlet) end of the bed.

[0107] vi) Product and Feed Repressurization (RP / F3): Finally, the bed pressure is returned to the feed or highest pressure level of the cycle to begin and repeat the above steps. This step is initiated by introducing feed gas and a portion of the main product gas from the feed (inlet) and product (outlet) ends, respectively.

[0108] 18-bed circulation ( Figure 5 ) Figure 5 An 18-bed PSA method applicable to another embodiment of the present invention is shown. This process includes feed (F1 to F4), co-current equalization depressurization 1 (EQD1), double equalization depressurization 2 (DEQD2), double equalization depressurization 3 (DEQD3), counter-current depressurization (CnD1 to CnD4), counter-current product purging (PU1 to PU3), double equalization repressurization 3 (DEQR3), double equalization repressurization 2 (DEQR2), counter-current equalization repressurization 1 (EQR1), and product and feed repressurization (RP / F5) steps. Each step is described in detail below.

[0109] i) Feeding Steps (F1 to F4): Assuming the bed has been pre-pressurized to the highest pressure level of the cycle, the feed gas mixture is introduced into the inlet of the bed, and unadsorbed gas is discharged from the outlet of the bed. The feeding steps continue until the mass transfer zone (MTZ) of the preferentially adsorbed component reaches the outlet of the bed without substantially disrupting it.

[0110] ii) Flow Equalization Pressure Reduction 1 (EQD1): At the end of the feed step, the bed is connected to another bed in the EQR1 step (described below). A portion of the voids and the desorbed gas are transferred from the product (outlet) end of the bed in EQD1 to the product (outlet) end of the other bed in EQR1, thereby reducing the bed pressure in EQD1 to a first intermediate level.

[0111] iii) Double Equilibrium Decompression 2 (DEQD2): Following EQD1, the bed is decompressed to a second intermediate pressure level by connecting it to another bed in the DEQR2 step (described below). The effluent containing voids and desorbed gases is transferred from the bed in DEQD2 to another bed in DEQR2 through the feed (inlet) and product (outlet) ends.

[0112] iv) Double Equilibrium Pressure Reduction 3 (DEQD3): Following DEQD2, the pressure in another bed is further reduced to a third intermediate level by transferring the void and desorbed gas to another bed in the DEQR3 step (described below). Pressure is again transferred through the feed (inlet) end and the product (outlet) end of the bed.

[0113] v) Countercurrent depressurization (CnD1 to CnD4): After DEQD3, the bed is depressurized by countercurrent flow to or near ambient pressure levels at the feed (inlet) end. The effluent containing more selectively adsorbed substances is removed as a second product.

[0114] vi) Countercurrent product purging (PU1 to PU3): At the end of the countercurrent depressurization step, purge gas is introduced into the product (outlet) end of the bed, and the effluent is collected from the feed (inlet) end of the bed as a second product.

[0115] vii) Double Equilibrium Repressurization 3 (DEQR3): After the purge step, the bed pressure is brought back from atmospheric pressure to the third intermediate level (as described in the DEQD3 step) by introducing voids and desorbed gas from another bed that has undergone DEQD3. Note that the gas is transferred through both ends of the column.

[0116] viii) Double Equilibrium Repressurization 2 (DEQR2): After DEQR3, the bed pressure is restored to a second intermediate level (as described in the DEQD2 step) by introducing voids and desorbed gas from another bed that has undergone DEQD2. This is also a two-end pressure transfer step.

[0117] ix) Countercurrent Equilibrium Repressurization 1 (EQR1): After DEQR2, the bed pressure is brought back to the first intermediate level (as described in step EQD1) by introducing voids and desorbed gas from another bed that has undergone EQD1. Note that the gas is transported through the product (outlet) end of the column.

[0118] x) Product and Feed Repressurization (RP / F5): Finally, the bed pressure is returned to the feed or maximum pressure level of the cycle to start and repeat the cycle. This step is initiated by introducing feed and a portion of the main product gas from the feed (inlet) and product (outlet) ends, respectively.

[0119] In this method cycle, seven beds are continuously fed, so product stopping is continuous. Due to the multi-bed nature of this method, flow fluctuations in the feed and product streams can be reduced / eliminated. In the final stage of pressure equalization, during step EQR1, the beds are pressurized using nearly pure void gas containing almost all primary products. Therefore, the pressure equalization is more efficient than... Figure 1 The conventional 2-bed dynamic PSA method shown is more efficient in its looping. Note that only one "unit block" is shown in the looping plan due to space constraints. The remaining blocks can be filled using the methods described in the literature (Mehrotra et al., 2011).

[0120] 9-bed circulation ( Figure 6 ) Figure 6 A 9-bed PSA method is shown, which has an added recycling step suitable for another embodiment of the invention. Figure 2 All features described in the 9-bed 18-step cycle shown apply to this embodiment. The only exception is that this cycle introduces a recirculation scheme. The recirculated gas is taken from the secondary product gas generated during the countercurrent depressurization (CnD1 to CnD3) and product purging (PU1 to PU4) steps. The recirculated gas is then mixed with the fresh feed. When the exhaust gas is at or near atmospheric pressure, the recirculated gas can be pressurized separately (up to the feed pressure level) and then mixed with the fresh feed (e.g., ...). Figure 6 (As shown), or the low-pressure exhaust gas can be mixed with fresh feed and the combined vapor can be pressurized (not shown) and fed to the bed as feed. If the exhaust gas is pressurized separately, or if a larger compressor is required for the combined pressurization option, a separate recirculation compressor is needed. This recirculation is designed to facilitate the recovery of primary product gases.

[0121] 9-bed circulation ( Figure 7 ) Figure 7 Another 9-bed PSA method is shown, which has an added recycling step suitable for use according to another embodiment of the invention. Again, the recycling step is included to facilitate the recovery of primary product gases. For Figure 2 The features described in the 9-bed 18-step cycle presented are applicable to this embodiment, with the following exceptions: i) Introduce a recirculating stream at the feed (inlet) end of the bed for a co-current depressurization step EQD1. The recirculating stream is the secondary product gas generated during the counter-current depressurization (CnD1 to CnD3) and product purging (PU1 to PU4) steps, which has already been pressurized by the recirculating compressor. The gas exiting from the product (outlet) end of the bed from the EQD1 / RC1 step is used to repressurize another bed through the product (outlet) end of the bed for an equalization repressurization (EQR1) step, as shown in Figure 7. This option requires a separate recirculating compressor.

[0122] ii) The double equalization decompression 2 (DEQD2) step is converted to a parallel-flow single-end equalization decompression step EQD2, and a recirculating stream is introduced at the feed (inlet) end of the bed. The recirculating stream is the secondary product gas generated during the counter-current decompression (CnD1 to CnD3) and product purging (PU1 to PU4) steps, which has been pressurized by the recirculating compressor. The product stream is introduced from the product (outlet) end of the bed from the EQD2 / RC2 step (e.g., Figure 7 The gas emitted (as shown) is used to repressurize another bed at the product (outlet) end of the bed via the countercurrent equalization repressurization step EQR2. As previously mentioned, this requires a separate recirculation compressor.

[0123] 9-bed circulation ( Figure 8 ) Figure 8 Another 9-bed PSA method is shown, which has an added recycling step suitable for use according to another embodiment of the invention, and is Figure 7 The improvement shown is a 9-bed circulation system. (Compared to...) Figure 7 The only difference is that, in this process, the gas generated from the EQD2 / RC2 step at the product end (outlet) of the bed is used for, for example... Figure 8 The diagram shows another bed repressurization step involving co-current equalization and repressurization via the feed end (inlet) of the bed. Example

[0124] Using 2 beds in multiple steps ( Figure 1 (as shown) and 9-bed multi-step ( Figure 2 The Pressure Swing Adsorption (PSA) method is cycled to evaluate the feed-to-product ratio and specific productivity using several carbon molecular sieve (CMS) adsorbents for the separation of nitrogen from air. Table 1 summarizes the oxygen, nitrogen, and argon adsorption rates and the kinetic selectivity of oxygen relative to nitrogen for each CMS adsorbent. Note that the oxygen-to-nitrogen kinetic selectivity is the same on CMS adsorbents 414-01 (“slow”) and 414-02 (“fast”), but the oxygen rate on CMS adsorbent 414-02 is twice as fast as that on CMS adsorbent 414-01. On the other hand, the oxygen rates are the same on CMS adsorbents 414-02 and 414-03 (“fast”), but the kinetic selectivity (oxygen to nitrogen) of CMS adsorbent 414-02 is twice as high as that of adsorbent 414-03. Bed geometry, operating conditions, and product oxygen concentration are also summarized in Table 1. It was also noted that the feed-to-product ratio is inversely related to the recovery rate; therefore, a lower ratio results in a better recovery rate, and vice versa. In other words, a lower feed-to-product ratio means a lower power requirement for the compressor.

[0125] The performance of the 2-bed and 9-bed PSA methods for the above CMS adsorbents is summarized in Figures 9 (9a to 9f). The evaluation conclusions are summarized below.

[0126] When using the “slow” CMS adsorbent (414-01), no improvement in process performance was observed when switching from a 2-bed to a 9-bed multi-step RCPSA process. For example, for a 9-bed process using the 414-01 CMS adsorbent, from Figure 9a It can be seen that the maximum productivity in a 40-second cycle time is 146.98 scfh60F / ft. 3 And from Figure 9b It can be seen that the corresponding feed-to-product ratio is 2.53. On the other hand, for the 2-bed multi-step method, the maximum productivity and the corresponding feed-to-product ratio are 151.38scfh60F / ft. 3 And 2.30, where the loop time is 100s (from Figure 9a and 9b (It can be seen that) For these two methods, at 100 o The performance of the PSA method for nitrogen products containing 4500 ppm oxygen was evaluated at a temperature of F and a bed pressure of 7.80 atma.

[0127] A comparative summary of the performance of 2-bed and 9-bed multi-step methods using a "fast" CMS adsorbent (414-02) with similar kinetic selectivity to "slow" (414-01) CMS is presented in [the original text]. Figure 9c and 9d In the middle. Similarly, in 100 o The performance of the method for evaluating nitrogen products containing 4500 ppm oxygen was assessed at a temperature of F and a bed pressure of 7.80 atma. The bed geometry and other parameters are summarized in Table 1.

[0128] Figure 9c and 9d It is clearly shown that when combined with a multi-bed, multi-step process, the faster-rate CMS adsorbent (414-02CMS) is significantly superior to the slower-rate CMS (414-01CMS), even though the kinetic selectivity (in this case, oxygen versus nitrogen) of the two CMS adsorbents is similar (as shown in Table 1). As mentioned above, 414-01CMS performs acceptablely when combined with a 2-bed, multi-step process (i.e., no benefit is gained by changing to a 9-bed process). Therefore, the performance of the 2-bed process using the slow-rate 414-01CMS adsorbent is used to compare the performance of the 9-bed process using the fast-rate 414-02CMS. For example, the 2-bed process using 414-01CMS achieves a productivity and feed-to-product ratio of 151.38 scfh60F / ft at a cycle time of 100 seconds. 3And 2.30. Using 414-02CMS adsorbent, a 9-bed multi-step method was implemented with a 50-second cycle time, and the productivity and feed-to-product ratio were 240.92 scfh60F / ft. 3 And 2.29. This means that when faster CMS adsorbents are combined with multi-bed methods, improvements of at least 59% in productivity can be achieved at similar feed-to-product ratios (or recoveries).

[0129] Compared to the two-bed method, the nine-bed multi-step RCPSA method maintains a higher specific productivity and offers comparable product recovery. Figure 9c It can be seen that when the cycle time is reduced from 50 seconds to 20 seconds, the specific productivity of the 9-bed method using Fast CMS increases from 240.92 to 265.35 scfh60F / ft. 3 During this period, the feed-to-product ratio increased only modestly from 2.29 to 2.93 (see [link to relevant documentation]). Figure 9d The 2-bed method using the faster CMS adsorbent (414-02) is superior to the "slow" CMS (414-01) 2-bed method, but exhibits lower specific productivity and recovery compared to the 9-bed method when operating at faster cycle times (20 s vs. 50 s). Figure 9c and 9d As shown.

[0130] Performance summary of 2-bed and 9-bed multi-step methods using a CMS adsorbent (414-03CMS) with similar rates but lower selectivity. Figure 9e and 9f Similarly, the method performance for nitrogen products containing 4500 ppm oxygen was evaluated at a temperature of 100°F and a bed pressure of 7.80 atma. The bed geometry and other parameters are summarized in Table 1.

[0131] The 9-bed method again outperforms the 2-bed multi-step method. When operating the 2-bed cycle with a faster cycle time, the losses in productivity and recovery are significant. Due to the low kinetic selectivity of oxygen to nitrogen in 414-03CMS, [the following is unclear and likely incomplete: "from..."] Figure 9d It can be seen that, compared with 414-02, the feed-to-product ratio is higher (e.g. Figure 9f (As shown). However, product recovery in 414-03CMS using the 9-bed method remains a significant improvement over the 2-bed slow CMS method, and the productivity is far superior to the 9-bed 414-03CMS method.

[0132] Another advantageous feature of combining the fast CMS adsorbent with the multi-bed, multi-step RCPSA method is that the amount of purge gas required to efficiently expel the fast desorbed component (oxygen during air separation in the CMS-based PSA method) from the bed voids is surprisingly lower than that required to produce high-purity nitrogen from air. Figure 10Overview of the use Figure 2 The 414-02CMS adsorbent and 9-bed RCPSA method shown in the figure were cycled at 100 o An overview of purge gas requirements for producing product nitrogen containing different amounts of oxygen (9 ppm, 90 ppm, and 4500 ppm) at F and 7.80 atma. A bed length of 40 inches and an inner diameter (ID) of 4 inches were used for this evaluation.

[0133] like Figure 10 As shown, when the oxygen content in the primary product gas decreases from 4500 ppm to 90 ppm, approximately 75% less purging is required. Even less purging is needed when the primary product contains 9 ppm oxygen. Due to the combination of the rapid oxygen kinetics of the fast CMS adsorbent and the efficient multi-bed, multi-step process, efficient bed regeneration was identified as the main factor for reducing purging requirements in the production of high-purity primary products. It is worth noting that all RCPSA process parameters (cycle time, all equalization valve constants, feed and product repressurization valve constants, and purging flow rates) are for... Figure 10 The three purging requirements in the report are optimized separately.

[0134] The additional column voids in conventional PSA methods using conventional valves are not limited to PSA methods employing rotary valve technology. Additional column voids include: i) any dead volume in the bed at the feed (inlet) or product (outlet) end without adsorbent; ii) in rotary valve PSA processes, the process gas transfer lines associated with the bed connecting to each valve port; and iii) any tuning or control valves or instrumentation coupled to the process gas transfer lines. Void volumes significantly degrade process performance.

[0135] The effect of void volume on method performance was investigated using a multi-bed, multi-step RCPSA method employing rotary valve technology. Figure 11a and 11b The use of 414-02 carbon molecular sieve adsorbent and Figure 2 The 9-bed multi-step RCPSA method shown in the diagram is prepared in 100 cycles. o The effects of pore volume on method performance in terms of "productivity" and "feed-to-product ratio" for nitrogen products containing 4500 ppm oxygen at F and 7.80 atma. Table 1 summarizes the adsorbent characteristics used for 414-02CMS. The bed length and inner diameter (ID) used for this evaluation were 40 inches and 4 inches, respectively. Note that the pore volume at the feed or product end increased proportionally during the study and is expressed as a percentage of one bed volume.

[0136] like Figure 11bAs shown, with increasing void volume, the feed-to-product ratio increases (or recovery decreases). However, a surprising finding is that at a certain void volume (approximately 10% of the bed volume), the void volume remains relatively constant. Figure 11a As shown, there is a maximum productivity, after which productivity begins to decline.

[0137] A 9-bed multi-step RCPSA method using rotary valve technology (cyclic sequence such as...) Figure 2 (As shown in Table 2) were used to evaluate the performance metrics of methods using different CMS adsorbents. Each bed in the RCPSA method was 40 inches high and 4 feet (ID) in inner diameter. Three CMS adsorbents (414-04, 414-05, and 414-06) with different oxygen rates, oxygen-to-nitrogen kinetic selectivity, and equilibrium capacities (as shown in Table 2) were selected for use at 69.8... o Performance evaluation of generating nitrogen (or containing different amounts of oxygen) from air at bed pressures of F and 7.80 atma. A 9-bed multi-step RCPSA process apparatus was used. Figure 2 The cyclic sequences shown in the figure) were used to evaluate CMS adsorbents 414-05 and 414-06, and the performance using these CMS adsorbents was compared with that of a conventional 2-bed multi-step method using CMS adsorbent 414-04. Figure 1 The data were compared. Note that the bed height and inner diameter (ID) of the 2-bed method used for evaluation were 120 inches and 1.908 inches, respectively.

[0138] RCPSA process parameters, such as cycle time or rotational speed, top and bottom equalization valve coefficients, feed and product repressurization valve coefficients, and purge flow rates, are optimized for each adsorbent at each purity, temperature, and bed pressure. Note that the RCPSA unit does not have feed and product tanks, as feed and product removal are continuous. The parameters for the 2-bed multi-step process are also optimized in the same manner as the RCPSA process.

[0139] Figure 12a and 12b The standardized productivity and standardized feed-to-product ratio process performance parameters for CMS adsorbents 414-04, 414-05, and 414-06 were summarized. At all product purities containing 95% to 99.9% nitrogen (or varying amounts of oxygen), the standardized productivity of the 9-bed multi-step process using fast CMS adsorbents (414-05 and 414-06) was significantly better than that of the combination of slow CMS adsorbent (414-04) and a conventional 2-bed process. For the 9-bed process using fast CMS adsorbents, the feed-to-product ratios were similar or lower at the aforementioned product purity ranges.

[0140] The standardized productivity of the CMS adsorbent considered here decreases in the order of 414-06 > 414-05 > 414-04, such as... Figure 12a As shown in Table 2, this is consistent with the uptake rate sequence (decreasing order): 414-06 > 414-05 > 414-04. A surprising feature is the similar or lower feed-to-product ratios, even though CMS adsorbents 414-05 and 414-06 have lower oxygen and nitrogen kinetic selectivity compared to CMS 414-04. Due to the higher efficiency of the multi-bed, multi-step process, equal or higher recoveries (lower feed-to-product ratios) can be achieved with CMS adsorbents having lower kinetic selectivity.

[0141] The impact of bed pressure on method performance is summarized in Figure 13a and 13b In this study, using 414-05 and 414-06 carbon molecular sieve adsorbents and the aforementioned 9-bed multi-step RCPSA unit, a standardized specific productivity of 69.8% was achieved. o The feed-to-product ratio at F produces a nitrogen product containing 45,000 ppm oxygen. The performance of a "slow" carbon molecular sieve adsorbent is also included, which produces a nitrogen product containing 45,000 ppm oxygen under the same conditions from a conventional 2-bed PSA process unit (bed length = 120 inches, bed ID = 1.908 inches). The 2-bed and 9-bed process circulation configurations are respectively... Figure 1 and 2 The information is provided in the text.

[0142] Figure 13a and 13b The results show that the specific productivity of the fast 414-06CMS is higher across all pressure ranges due to its higher oxygen uptake rate. An interesting characteristic of the 414-06CMS is that even though its oxygen selectivity is slightly lower than that of nitrogen, its feed-to-product ratio is lower than that of the 414-05CMS adsorbent. The higher equilibration capacity in this process (18.5% higher than that of 414-05CMS, as shown in Table 2) and the reduced pressure drop (due to the use of larger adsorbent particles: particle size = 0.0719 inner diameter for 414-06CMS and 0.0568 inner diameter for 414-05CMS) are considered the main reasons for the observed trend. The two-bed method using 414-04CMS performs poorly due to its slower oxygen rate and lower method efficiency.

[0143] Using rapid CMS adsorbent 414-02, respectively... Figure 2 , Figure 3 , Figure 4 and Figure 5The 9-bed, 7-bed, 4-bed, and 18-bed multi-step cycle designs shown were used to evaluate the method performance for producing nitrogen (or containing different amounts of oxygen) of varying purities under different pressures, temperatures, and bed length-to-diameter ratios (or aspect ratios). Adsorbent characteristics are summarized in Table 1. Operating conditions and bed characteristics are shown in Table 3. Note that RCPSA method parameters, such as cycle time, top and bottom equalization valve coefficients, feed and product repressurization valve coefficients, and purge flow rates, were individually optimized for each cycle design suitable for operation at different temperatures, pressures, and product oxygen purities. The evaluation summary in Table 3 is as follows.

[0144] For a 9-bed multi-step cycle producing nitrogen products containing 4500 ppm oxygen at 7.80 atma, when the temperature is increased from 100... o F becomes 30 o At step F, the feed-to-product ratio changes from 2.32 to 2.05 (Case 8, Case 2, and Case 1). The corresponding specific productivity decreases from 260.42 to 242.10 scfh60F / ft. 3 Note that the length-to-inner-diameter ratio (or aspect ratio) used for this assessment is 10.

[0145] Under the same operating conditions, the same product specifications, and the same bed width ratio, the 18-bed multi-step method is superior to the 9-bed multi-step RCPSA method, as shown in Cases 27 to 29 in Table 3.

[0146] For 7.80 atma and 100 o A 9-bed multi-step cycle produces nitrogen containing 90 ppm oxygen at F, with bed aspect ratios (or length-to-diameter ratios) varying between 2 and 18 (cases 4 to 7). For nitrogen products containing 90 ppm oxygen, the feed-to-product ratio and specific productivity are optimal at a bed aspect ratio of 18. For nitrogen products containing 4500 ppm oxygen (cases 8 to 13), the optimal bed aspect ratio is 10, as shown in case 8 in Table 3.

[0147] 7-bed multi-step circulation is superior to 9-bed multi-step circulation for use in 100 o Nitrogen containing 45,000 ppm oxygen is produced at F and 11.21 atma (cases 15 and 22). With similar feed-to-product ratios, the 7-bed method achieves a specific productivity increase of approximately 13%. Even at 100 o Nitrogen containing 4500 ppm oxygen can be produced at F and 7.80 atma. Using a 7-bed multi-step cycle, a higher productivity of approximately 14% can be achieved at a lower feed-to-product ratio (or higher recovery rate), as shown in Cases 13 and 16 in Table 3.

[0148] A 7-bed multi-step cycle is also superior to a 4-bed multi-step cycle for use in 100 oNitrogen gas containing 4500 ppm oxygen was produced at F and 7.80 atma (Case 16 and 25).

[0149] A 7-bed multi-step method was used to conduct a downsizing study using 414-02CMS adsorbent at 100 o Nitrogen containing 45,000 ppm oxygen is produced at F and 11.21 atma. When productivity requirements are lower than necessary, the process needs to be reduced. For kinetic processes, if process efficiency is low, the power requirement (or feed-to-product ratio) may be higher in the reduced mode. A summary of the adjustment evaluation studies (Case 31 to 41) is presented in Table 4 along with the best-case scenario (Case 30).

[0150] Table 4 shows that when productivity decreased to approximately 70% of the optimal level, the feed-to-product ratio increased by only about 21.9% (case 37). Adjustments were made by changing operating parameters and cycle time, as shown in Table 4.

[0151] Figure 2 , Figure 6 , Figure 7 and Figure 8 The 9-bed multi-step RCPSA cycle design shown was used to evaluate the performance of a method for producing pure argon from an argon / oxygen / nitrogen mixture using an RHO-based zeolite adsorbent. For these cycle designs, a high pressure of 7.80 atma, a low pressure of 1.05 atma, and a pressure of 100 atma were used. o The performance was evaluated using temperature F. The feed gas mixture contained 20 mol% O2, 0.05 mol% N2, and the catalyst Ar, with the final argon product containing 2 ppm O2 impurities. Table 5 summarizes the bed and adsorbent characteristics, as well as the operating conditions used to evaluate the method performance. The table also lists method performance metrics for Ar recovery and specific productivity. Note that for each cycle design, RCPSA method parameters (e.g., cycle time, top and bottom equalization valve coefficients, feed and product repressurization valve coefficients, and purge flow rates) were individually optimized.

[0152] Table 5 shows all recycling designs ( Figures 6 to 8 The cycle shown is better than none. Figure 2 The loop design for recirculation is shown below. In the recirculation options, Figure 6 The fresh feed design presented in the design makes recirculation a preferred option: i) only 44.57% of exhaust gas is recirculated, compared to Figure 7 and Figure 8Compared to other designs shown, it achieves over 75% Ar recovery with similar or better specific productivity; ii) the exhaust gas can be mixed with the low-pressure crude argon feed, and the combined feed can then be recompressed by a single compressor, thus eliminating the need for additional machinery; and iii) a simple control scheme can be implemented because the feed step operates at constant pressure, but for Figure 7 and 8 The cyclic design in the process, where the exhaust gas is recirculated back to the pressure change step, may increase some control complexity.

[0153] Although the principles of the invention have been described above in conjunction with preferred embodiments, it should be clearly understood that this description is by way of example only and is not intended to limit the scope of the invention.

[0154] Table 1. Summary of adsorbent and bed characteristics and operating conditions for carbon molecular sieve adsorbents 414-01, 414-02 and 414-03.

[0155] Table 2. Linear driving force (LDF) model based on O2 rate constant, O2-over-N2 kinetic selectivity and O2-to-414-04, 414-05 and 414-06 carbon molecular sieve adsorbents at latma and 86 o Summary of the balanced capacity of F.

[0156] Table. Use of 4-bed ( Figure 4 (shown in the image), 7 beds ( Figure 3 As shown in the image), 9 beds ( Figure 2 (shown in the image) and 18 beds ( Figure 5 (Shown) Performance summary of 414-02 carbon molecular sieve adsorbent cyclic by PSA method under different product oxygen impurities (in terms of "productivity" and "feed to product ratio").

[0157] Table 4. Production of 100g of oxygen at 45,000 ppm using 414-02 carbon molecular sieve adsorbent o F and 11.21atma 7-bed method cycle ( Figure 3 The best (Case 30) and best-adjusted (Case 31-41) performance summaries in terms of "productivity" and "feed to product ratio" are shown in the table.

[0158] Table 5. Summary of adsorbent characteristics, bed characteristics, operating conditions, and method performance based on RHO (in terms of "productivity" and "Ar recovery") for use Figure 2 , Figure 6 , Figure 7 and Figure 8 The 9-bed multi-step method shown in the diagram cyclically separates Ar from the O2 / N2 / Ar mixture.

Claims

1. A multi-bed fast cyclic pressure swing adsorption (RCPSA) method for separating O2 relative to N2 and / or Ar, wherein the method uses at least five adsorption beds, each adsorption bed comprising [a specific configuration] based on a linear driving force model at 1 atma and 86 [presumably referring to a specific pressure swing adsorption method]. o A kinetically selective adsorbent for O2 with an O2 adsorption rate of at least 0.20 (1 / s) determined at F, and wherein the RCPSA method comprises subjecting each adsorption bed to a rapid PSA cycle, including the following steps performed in sequence: i) Feeding ii) First equilibrium decompression iii) Second equilibrium decompression iv) Countercurrent decompression v) Countercurrent sweeping vi) First Equilibrium Repressurization vii) Second Equilibrium Repressurization viii) Repressurize the product and / or feed When the adsorption bed is undergoing the equalization depressurization step ii), it is connected to one of the other adsorption beds undergoing the equalization repressurization step vii) and is supplied with a repressurization flow thereon. When the adsorption bed is undergoing the equalization depressurization step iii), it is connected to another adsorption bed that is simultaneously undergoing the equalization repressurization step vi) and provides it with a repressurization flow; in, The adsorption rate of the adsorbent for O2, N2, or Ar is determined by the following steps: Make the initial state in vacuum and 86 o The adsorbent sample at F was exposed to a measured amount of adsorbate O2, N2 or Ar at the same temperature and 1 atm. The pressure changes are recorded as a function of time; Subtract the pressure versus time data from a similar pressure history of a sample using the same weight of quartz beads instead of adsorbent to obtain an absorption curve as a function of time for the amount of adsorbed gas. as well as A linear driving force model was used to extract the adsorption rate of the adsorbate in reverse time (1 / s) from the absorption curve; The cycle time of the fast PSA cycle is equal to or less than 100 seconds; The kinetic selective adsorbent described herein exhibits properties at 1 atma and 86 atma via a linear driving force model. o At least 5% O2 / N2 kinetic selectivity was determined at F, and / or by a linear driving force model at 1 atma and 86. o At least 5% O2 / Ar kinetic selectivity was determined at F.

2. The RCPSA method of claim 1, wherein step iii) is a double-equilibrium decompression step and step vi) is a double-equilibrium recompression step.

3. The RCPSA method of claim 1, wherein step viii) is a product and feed repressurization step.

4. The RCPSA method of claim 1, wherein step ii) is a downstream equalization depressurization step and step vii) is a counter-current equalization repressurization step.

5. The RCPSA method of claim 1, wherein the method uses 5 to 18 adsorption beds.

6. The RCPSA method of claim 1, wherein the method uses 7 to 9 adsorption beds.

7. The RCPSA method of claim 1, wherein the method uses 7 or 9 adsorption beds.

8. The RCPSA method of claim 1, wherein the duration of the feeding step is 3 to 45 seconds.

9. The RCPSA method of claim 1, wherein the duration of each of the equalization depressurization and equalization repressurization steps is 1 to 5 seconds.

10. The RCPSA method of claim 1, wherein the feeding step is performed at 0 o F to 125 o Performed at a temperature of F.

11. The RCPSA method of claim 1, wherein the feeding step is performed in 20... o F to 100 o Performed at a temperature of F.

12. The RCPSA method of claim 1, wherein the feeding step is performed in 20... o F to 40 o Performed at a temperature of F.

13. The RCPSA method of claim 1, wherein during all or part of the feeding step, circulating gas is introduced co-currently into the bed undergoing the feeding step, the circulating gas comprising gas obtained from the bed undergoing the counter-current depressurization step and / or purging step.

14. The RCPSA method of claim 1, wherein during all or part of the equalization decompression step ii), circulating gas is introduced co-currently into the bed undergoing the step, the circulating gas comprising gas obtained from the bed undergoing the step during the counter-current decompression step and / or purging step.

15. The RCPSA method of claim 1, wherein the kinetically selective adsorbent is a zeolite or a carbon molecular sieve.

16. The RCPSA method of claim 1, wherein the method separates O2 relative to Ar, and the kinetically selective adsorbent is RHO zeolite with a Si / Al ratio of 3.2 to 4.5 and containing aprotic extra-framework cations, wherein each unit cell of the zeolite contains at most one proton, and wherein the size, number, and charge of the extra-framework cations present in the zeolite are such that each unit cell requires one or fewer aprotic extra-framework cations to occupy an octagonal site.

17. The RCPSA method of claim 1, wherein the method separates O2 relative to N2, and the kinetically selective adsorbent has a linear driving force model at 1 atma and 86 o Carbon molecular sieves (CMS) with O2 / N2 kinetic selectivity of 5 to 30 were determined at F.

18. The RCPSA method of claim 1, wherein the method is a rotary bed RCPSA method.

19. The RCPSA method of claim 1, wherein the method is a rotary valve RCPSA method.

20. The RCPSA method of claim 1, wherein the void volume of each adsorption bed is 3% to 15% of the bed volume.

21. The RCPSA method of claim 1, wherein the method uses seven adsorption beds, and wherein the RCPSA method comprises subjecting each adsorption bed to a rapid PSA cycle, comprising the following steps performed in sequence: Feed (F); Downstream equalization pressure reduction (EQD1); Double-equalized decompression (DEQD2); Countercurrent decompression (CnD); Countercurrent purging (PU); Dual Equalization Repressurization (DEQR2); Countercurrent equalization and repressurization (EQR1); and Product and feed repressurization (RP / F); When the adsorption bed undergoes the co-current equilibrium depressurization (EQD1) step, it is connected to one of the other adsorption beds simultaneously undergoing the counter-current equilibrium repressurization (EQR1) step and is supplied with a counter-current repressurization flow. When the adsorption bed undergoes the double equalization depressurization (DEQD2) step, it is connected to another adsorption bed that is simultaneously undergoing the double equalization repressurization (DEQR2) step and is provided with co-current and counter-current repressurization flows.

22. The RCPSA method of claim 1, wherein the method uses nine adsorption beds, and wherein the RCPSA method comprises subjecting each adsorption bed to a rapid PSA cycle, comprising the following steps performed in sequence: Feed (F); Downstream equalization pressure reduction (EQD1); First dual-equalization decompression (DEQD2); Second Dual Equalization Decompression (DEQD3); Countercurrent decompression (CnD); Countercurrent purging (PU); First Dual Equalization Repressurization (DEQR3); Second Dual Equalization Repressurization (DEQR2); Countercurrent equalization and repressurization (EQR1); and Product and feed repressurization (RP / F); When the adsorption bed undergoes the co-current equilibrium depressurization (EQD1) step, it is connected to one of the other adsorption beds simultaneously undergoing the counter-current equilibrium repressurization (EQR1) step and is supplied with a counter-current repressurization flow. When the adsorption bed undergoes the first double equalization depressurization (DEQD2) step, it is connected to another adsorption bed simultaneously undergoing the second double equalization repressurization (DEQR2) step, and is provided with co-current and counter-current repressurization flows. When the adsorption bed undergoes the second double equalization depressurization (DEQD3) step, it is connected to another adsorption bed that is simultaneously undergoing the first double equalization repressurization (DEQR3) step and provides it with co-current and counter-current repressurization flows.

Images