Durable and Versatile Immobilized Carbonic Anhydrase on Textile Structured Packing for CO₂ Capture

Abstract

High-performance carbon dioxide (CO₂)-capture technologies with low environmental impact are necessary to combat the current climate change crisis. Durable and versatile “drop-in-ready” textile structured packings with covalently immobilized carbonic anhydrase (CA) were created as efficient, easy to handle catalysts for CO₂ absorption in benign solvents. The hydrophilic textile structure itself contributed high surface area and superior liquid transport properties to promote gas-liquid reactions that were further enhanced by the presence of CA, leading to excellent CO₂ absorption efficiencies in lab-scale tests. Mechanistic investigations revealed that CO₂ capture efficiency depended primarily on immobilized enzymes at or near the surface, whereas polymer entrapped enzymes were more protected from external stressors than those exposed at the surface, providing strategies to optimize performance and durability. Textile packing with covalently attached enzyme aggregates retained 100% of the initial 66.7% CO₂ capture efficiency over 71-day longevity testing and retained 85% of the initial capture efficiency after 1-year of ambient dry storage. Subsequent stable performance in a 500 h continuous liquid flow scrubber test emphasized the material robustness. Biocatalytic textile packings performed well with different desirable solvents and across wide CO₂ concentration ranges that are critical for CO₂ capture from coal and natural gas-fired power plants, from natural gas and biogas for fuel upgrading, and directly from air.

1. Introduction

Anthropogenic greenhouse gas (GHG) emissions, and carbon dioxide (CO₂) [1,2,3] in particular, are the principal drivers of global warming, provoking extreme weather and climate phenomena with negative impacts on the cryosphere, biospheres, atmosphere, and ocean [4,5]. According to the International Energy Agency (IEA), in the year 2019, 14.0 Gt of CO₂ (or 43.9% of the total) was emitted from the production of electricity and heat generation [6]. Strategies for transitioning to a net-zero emission energy system will rely heavily on improvements in the electricity sector [7], where use of existing CO₂ emitting power plants, as long-lived capital assets, will persist for the foreseeable future to meet ever increasing energy consumption needs beyond what non-emitting power generation technologies can supply. Preventing CO₂ emissions from these large point-sources is thus of paramount importance for responsible energy production from coal, natural gas, and biomass resources.

At conditions relevant for most combustion-based power plants, chemical absorption [8] (also known as reactive absorption) with aqueous monoethanolamine (MEA) solvent has been the benchmark mature technology for capturing CO₂ at atmospheric pressure [9]. However, fast MEA absorption kinetics come with a high regeneration energy cost [10] which has limited its wide adoption. Alternative solvents include sterically hindered amines that replace stable high energy carbamate bonds with less stable and lower energy carbamate bonds or with tertiary amines. Tertiary amines, such as methyldiethanolamine (MDEA), and inorganic carbonate-based solvents, such as potassium carbonate (K₂CO₃), function exclusively through a slower but lower energy bicarbonate mechanism to hydrate CO₂ based on acid-base considerations [11]. The enthalpies of reaction of MDEA (26 kJ/mol of CO₂) and K₂CO₃ (27 kJ/mol) are less than one third that of MEA (82 kJ/mol of CO₂ at 40 °C) [12,13]. However, the total regeneration heat duty also depends on both the sensible heat, to raise the solvent to the desorption temperature, and on the heat of evaporation, to produce stripping steam, all of which are highly sensitive to differences in process parameters [14]. Since lower energy solvents are also kinetically slower, they require taller absorption columns with larger amounts of solvent that can off-set cost savings from lower regeneration energy. Therefore, fast reaction rates and low regeneration energy are both important for improving process efficiency.

Biomimetic CO₂ capture using low energy aqueous solutions promoted by carbonic anhydrase (CA) is an environmentally friendly CO₂ mitigation strategy that shows promise for requiring smaller equipment size with lower operation costs [15,16,17,18,19,20]. CA enables the use of benign solvents, like non-volatile aqueous K₂CO₃, that exhibit low environmental, health and safety risks [21] compared to MEA which is corrosive [22] and generates aerosols containing hazardous nitramines and nitrosamines [23]. The hydration of CO₂ to bicarbonate (CO₂ + H₂O ↔ HCO₃⁻ + H⁺) catalyzed by CA is favored at alkaline conditions that neutralize released protons. This results in the reactive conversion of a sparingly soluble gas (CO₂) into a highly soluble ion (HCO₃⁻) to “capture” CO₂ in the absorption liquid. Representative studies using CA for enhancing CO₂ absorption rates into kinetically limited low energy solvents are summarized in Table 1. Based on modeling simulations [24], CA-enhanced alternative solvent systems have the potential to meet the CO₂ absorption efficiency of benchmark 30 wt% (~5 M) MEA [25]. To achieve this, CA must enhance CO₂ absorption rates in solvents by many-fold (sometimes called the ‘enhancement factor’) compared to comparable systems without enzyme. While capable of delivering high enhancement factors, a major obstacle to using “free” (non-immobilized) CA as the rate-enhancer results from its thermal deactivation as it flows with solvent through the heated stripper. Enzyme immobilization can overcome this issue by enabling separation and reuse, improving heat and solvent tolerance, and extending working or storage lifetime [26,27,28,29,30]. Immobilizing CA allows it to be confined to the lower temperature absorber column where it directly benefits CO₂ absorption reaction kinetics. Immobilization also helps circumvent thermal deactivation, reduces enzyme replenishment requirements, and makes it possible to operate the desorber process at the lowest energy requiring conditions without concern for damaging the enzyme, in case those conditions favor high temperature. For large industrial scale biocatalytic process, such as reactive absorption of CO₂ from point sources, immobilization could potentially be coupled with enzyme purification [31], allowing the use of crude cell extracts that could reduce cost.

Enzyme immobilization is a fabrication strategy for confining one or more enzymes in a defined space [32] with the goal of positioning enzymes relative to their substrates and reaction environment for optimal catalytic activity and to stabilize them for repeated use [33]. Materials used for enzyme immobilization come in different chemistries, physical forms, sizes and shapes [34]. Categories of physical supports commonly used for CA immobilization are diverse. They include preexisting supports, such as inorganic particles [35,36,37,38,39,40,41,42,43,44,45], polymeric hollow-fiber or flat-sheet membranes [46,47,48,49,50,51,52,53], and they include materials that are formed in situ with enzyme, such as polysaccharide beads [54,55,56,57,58], polyurethane foams [59,60], and metal organic frameworks (MOF) [61,62,63]. Alternatively, CA has been immobilized as carrier-free crosslinked enzyme aggregates (CLEAs) for use in slurry absorbers [64].

Table 1. Biomimetic CO₂ capture studies using low energy aqueous solutions promoted by carbonic anhydrase (CA).

CA Forms	Solvent and Testing Systems	Biocatalyst Performance Results	Reference
Free CA at 0–3 g/L	20 wt% K2CO3 (1.8 M); Stirred tank reactor and packed-bed column	High CA dose is able to boost CO2 absorption rate of 20 wt% K2CO3 comparable to that of 5 M (~30 wt%) MEA	Zhang et al. [24]
Free CA at 0–2.4 g/L	1–4 M MDEA; 0.3–0.6 M Na2CO3; Stirred cell contactor	Reduce the absorber size required for kinetically limited low energy solvents to the size achieved by commercial scale MEA CO2 capture systems	Penders-van Elk et al. [65]
Free CA at 0–2.5 g/L CA	23.5 wt% K2CO3 (2.2 M); Bench-scale unit packed with ceramic Raschig rings and with vacuum stripping	4.6-fold increase in CO2 capture efficiency at CA concentration of 2.5 g/L CA; a minimum of 10% daily (every 7 h run time) enzyme replenishment was required to maintain a consistent CO2 capture efficiency	Qi et al. [12,66]
Free CA at >0.3 g/L	2–3 M K2CO3; stirred cell apparatus	5–8-fold increase in CO2 absorption rate with >0.3 g/L precipitated enzyme aggregates acting as biocatalysts	Pierce et al. [15]
Immobilized CA “in vivo” anchored on cell membrane	0.5 M Na2CO3/0.5 M NaHCO3 buffer; stirred cell apparatus	Enhancement factor of 1.3–2.4 with dispersed CA-anchored cell debris acting as biocatalyst	Fabbricino et al. [67]
Immobilized CA entrapped in organosilica matrix coated on stainless steel structured packing	20 wt% K2CO3 (1.8 M)and a proprietary non-volatile alkaline salt solvent; packed column reactors from lab- to field- scale	6 to 7-fold rate enhancement for immobilized CA packing over the no-enzyme packings and a biocatalyst half-life of 539 days determined from 3500 h continuous field test	Reardon et al. [68,69]
Immobilized CA entrapped in organosilica micro-particles	30 wt% MDEA (2.5 M); counter-current packed column	Immobilized CA in the form of buoyant micro-particles that can be easily separated and reused, yielded a 6-fold enhancement in CO2 absorption	Leimbrink et al. [70,71]
Immobilized CA entrapped in chitosan matrix coated on textile structured packing	10 wt% K2CO3 (0.8 M); Lab-scale packed column	2-fold and 14.5-fold enhancement in CO2 capture efficiency over no-enzyme textile control packing and over conventional glass Raschig rings, respectively; 66% performance retention after 5 repeated test-wash-storage cycles; 76.5% performance retention after a 120 h continuous solvent flow scrubber test	Shen et al. [72]
Immobilized CA covalently attached on textile structured packing	5–30 wt% K2CO3 (0.4–2.8 M); 5–10 wt% DMG (0.4–0.8 M); 5 wt% MDEA (0.4 M); Simulated seawater (3.5 wt% solute); Lab-scale packed column	2.9-fold and 19.1-fold enhancement in CO2 capture efficiency over no-enzyme textile control packing and conventional glass Raschig rings, respectively; 100% performance retention after 10 repeated test-wash-storage cycles including a 100 h incubation in in 10 wt% K2CO3 at 45 °C; steady (~100%) performance retention over a 500 h continuous solvent flow scrubber test; 85.7% performance retention after 1 year dry storage	This study

Table 1. Biomimetic CO₂ capture studies using low energy aqueous solutions promoted by carbonic anhydrase (CA).

CA Forms	Solvent and Testing Systems	Biocatalyst Performance Results	Reference
Free CA at 0–3 g/L	20 wt% K2CO3 (1.8 M); Stirred tank reactor and packed-bed column	High CA dose is able to boost CO2 absorption rate of 20 wt% K2CO3 comparable to that of 5 M (~30 wt%) MEA	Zhang et al. [24]
Free CA at 0–2.4 g/L	1–4 M MDEA; 0.3–0.6 M Na2CO3; Stirred cell contactor	Reduce the absorber size required for kinetically limited low energy solvents to the size achieved by commercial scale MEA CO2 capture systems	Penders-van Elk et al. [65]
Free CA at 0–2.5 g/L CA	23.5 wt% K2CO3 (2.2 M); Bench-scale unit packed with ceramic Raschig rings and with vacuum stripping	4.6-fold increase in CO2 capture efficiency at CA concentration of 2.5 g/L CA; a minimum of 10% daily (every 7 h run time) enzyme replenishment was required to maintain a consistent CO2 capture efficiency	Qi et al. [12,66]
Free CA at >0.3 g/L	2–3 M K2CO3; stirred cell apparatus	5–8-fold increase in CO2 absorption rate with >0.3 g/L precipitated enzyme aggregates acting as biocatalysts	Pierce et al. [15]
Immobilized CA “in vivo” anchored on cell membrane	0.5 M Na2CO3/0.5 M NaHCO3 buffer; stirred cell apparatus	Enhancement factor of 1.3–2.4 with dispersed CA-anchored cell debris acting as biocatalyst	Fabbricino et al. [67]
Immobilized CA entrapped in organosilica matrix coated on stainless steel structured packing	20 wt% K2CO3 (1.8 M)and a proprietary non-volatile alkaline salt solvent; packed column reactors from lab- to field- scale	6 to 7-fold rate enhancement for immobilized CA packing over the no-enzyme packings and a biocatalyst half-life of 539 days determined from 3500 h continuous field test	Reardon et al. [68,69]
Immobilized CA entrapped in organosilica micro-particles	30 wt% MDEA (2.5 M); counter-current packed column	Immobilized CA in the form of buoyant micro-particles that can be easily separated and reused, yielded a 6-fold enhancement in CO2 absorption	Leimbrink et al. [70,71]
Immobilized CA entrapped in chitosan matrix coated on textile structured packing	10 wt% K2CO3 (0.8 M); Lab-scale packed column	2-fold and 14.5-fold enhancement in CO2 capture efficiency over no-enzyme textile control packing and over conventional glass Raschig rings, respectively; 66% performance retention after 5 repeated test-wash-storage cycles; 76.5% performance retention after a 120 h continuous solvent flow scrubber test	Shen et al. [72]
Immobilized CA covalently attached on textile structured packing	5–30 wt% K2CO3 (0.4–2.8 M); 5–10 wt% DMG (0.4–0.8 M); 5 wt% MDEA (0.4 M); Simulated seawater (3.5 wt% solute); Lab-scale packed column	2.9-fold and 19.1-fold enhancement in CO2 capture efficiency over no-enzyme textile control packing and conventional glass Raschig rings, respectively; 100% performance retention after 10 repeated test-wash-storage cycles including a 100 h incubation in in 10 wt% K2CO3 at 45 °C; steady (~100%) performance retention over a 500 h continuous solvent flow scrubber test; 85.7% performance retention after 1 year dry storage	This study

Table 1 includes our first generation reactive absorption “drop-in-ready” textile structured packings, wherein CAs were entrapped in thin chitosan matrices coated on cotton base-layer materials [72]. These biocatalytic textile packings exhibited excellent solvent distribution [73] and maintained uniform gas contact with the wetted solid surfaces. Average CO₂ absorption efficiencies of 52% and 82% were found for single and double-stacked CA-immobilized packings, respectively, in laboratory-scale CO₂ scrubber tests at 4 L per minute flow rate with nominal 10% CO₂ inlet gas. Textile packing retained 66% of the initial CO₂ capture performance after five cycles of wash-dry-store and re-test over a period of 66 days and retained 77% performance after a continuous 120 h recirculation longevity test. To further improve the CO₂ capture performance and extend the longevity in use, the placement of enzymes on the textile packing relative to the surface and the use of covalent bonding to affix CA at the surface were explored in the present work.

Among compounds used for biocatalyst design, glutaraldehyde (GA) is the most widely adopted and versatile tool [74]. Its chemical mechanisms of action (Figure 1b) could go beyond the conventionally acknowledged Schiff’s base formation between lysine amino groups of the enzyme and GA’s reactive aldehydes [75] to also include Michael-type additions and heterocycle species, depending on the reaction conditions [74]. Furthermore, GA has a long spacer arm that retains enzyme mobility [33] which could be beneficial for reaction catalysis at gas-liquid interfaces.

In the current study, we included GA in several ways, either as a chemical crosslinker or as a surface chemistry activating agent, to immobilize enzymes in different locations and in different physico-chemical environments relative to the textile surface (Figure 1a). The three main immobilization strategies we used were: (1) post-entrapment crosslinking, in which GA was applied to a surface already coated with enzymes entrapped in chitosan; (2) multi-step surface covalent immobilization, in which GA was applied to activate the chitosan-coated surface followed by a separate enzyme immobilization step to create a thin “mono-layer” of surface-exposed enzymes; and, (3) one-pot surface covalent immobilization, in which GA and enzymes were applied to the chitosan-coated textile surface at the same time, leading to thick coatings of enzyme at the surface, which we refer to as 3-D enzyme aggregates. Additional hybrid constructions that combined both entrapment and surface covalent immobilizations were also evaluated. Chemical and physical properties of catalytic textiles fabricated by the different immobilization approaches were characterized using spectroscopy, microscopy and liquid transport measurements. Catalytic performance of the different materials was evaluated using both a slower esterase activity assay and a faster, more application-relevant CO₂ absorption test in a laboratory gas scrubber. The combined observations from these studies revealed mechanisms behind heterogeneous catalysis robustness and longevity that guided the selection and design of prototype modular packing materials for application testing. Results from laboratory scale application tests show that biocatalytic textile structured packing modules perform well at conditions relevant for CO₂ capture from coal and natural gas-fired power plant flue gas, from natural gas and biogas for fuel upgrading, and directly from air.

2. Results and Discussions

2.1. Surface Chemistry

Enzyme immobilization at fiber surfaces was examined by Fourier-transform infrared spectroscopy (FTIR) operated in attenuated total reflectance (ATR) mode. The evanescent wave generated by ATR mode detects chemical functional groups within at least a few microns of sample surfaces and can penetrate deeper to detect underlying layers, depending on various factors [76]. Therefore, FTIR-ATR is expected to detect both the surface chemistry of a sample and the chemistry of underlying layers. Each spectrum was normalized against its maximum absorbance and the relative changes in peak intensities among different sample treatments were compared to identify differences in surface chemical composition. As shown in Figure 2, all sample spectra resemble the underlying cellulosic cheesecloth support material, which constitutes a major fraction of the total sample weight, and has a maximum absorbance at around 1058 cm⁻¹ corresponding to the C-O-C vibration of the polysaccharides ring structure [77]. The most abundant and high intensity functional group in enzymes is the amide group (Supporting Information Figure S1). Amide groups exhibit two characteristic absorbance peaks located at around 1650 cm⁻¹ (Amide I) and 1540 cm⁻¹ (Amide II), corresponding to strong carbonyl (C=O) stretching and a combination of N-H bending and C-N stretching, respectively [78]. As shown in the inset of Figure 2, neat cellulosic cheesecloth has an overlapping broad peak located near 1641 cm⁻¹ which is commonly assigned to the H-O-H bending of absorbed water [77,79].

For ease of discussion, the spectra and legends in Figure 2 are arranged according to the same order as their peak intensities at 1650 cm⁻¹. Coating with chitosan slightly increased the baseline peak intensities in this region, and introduced several overlapping peaks in the 1540–1580 cm⁻¹ region which are attributed to N-H bending and C-N stretching. Within this region, two distinct peaks at ~1540 cm⁻¹ and ~1560 cm⁻¹ were detected in all samples except in the plain cheesecloth. This is consistent with no nitrogen being present in the chemical composition of cellulose. Crosslinking the chitosan coating with glutaraldehyde (GA) made a three peak pattern at 1650, 1645, and 1633 cm⁻¹ more distinct, and all samples with GA plus enzyme exhibited more or less distinct peaks at these three wavenumbers. This triad of peaks was also previously observed when enzymes were entrapped in chitosan coating without GA crosslinking [72], albeit not as prominent as in certain (black, red and green curves of Figure 2) GA-crosslinked samples. In particular, the peak at 1650 cm⁻¹ is more distinct in GA-crosslinked samples. Furthermore, a peak at ~1600 cm⁻¹ that is routinely associated with primary amine deformation in chitosan [80,81] is missing from the chitosan coated sample, implying that most amine groups in the chitosan coating are not in the free-NH₂ form. This could be explained by formation of Schiff’s base linkages between chitosan amines and cellulose aldehydes that are either present at cellulose reducing chain ends or at the C2 and C3 positions of ring-opened glucose residues along the cellulose chain caused by oxidative damage from bleaching during cotton cheesecloth manufacturing. Support for this idea comes from close inspection of the plain cheesecloth spectrum, which reveals a peak at ~1730 cm⁻¹ that has been attributed to aldehyde C=O stretching in oxidized cellulose [79,82].

In general, peaks appearing in the 1630–1650 cm⁻¹ region after glutaraldehyde crosslinking are thought to indicate the formation of the C=N bond in Schiff’s base linkages [83,84]. However, glutaraldehyde exists in multiple forms in aqueous solution [85]. Therefore, part of the glutaraldehyde could have undergone self-polymerization to form aldol condensed oligomers [86] containing C=C bond stretching (Figure 1b), which also resides in this IR absorbance region. Furthermore, the peak at 1645 cm⁻¹ could be due to the symmetric deformation vibration of absorbed water molecules, because a peak in this location was observed to disappear when heating cotton samples to drive off water [79]. The peak patterns attributed to the combination of glutaraldehyde and chitosan alone were also observed for surface covalent mono-layer enzyme immobilization on glutaraldehyde activated chitosan coating and for crosslinked chitosan coating with enzyme pre-entrapped in the matrix. Each of these samples showed an increased baseline intensity from the foregoing ones, indicating an increase in the detectable Amide I band which indicates an increase in the amount of immobilized enzymes. Enzyme immobilization method variations shown in the top three spectra of Figure 2 (purple, yellow and dark blue) had less pronounced specific peak patterns but much higher overall peak intensities, mainly attributable to the Amide I band contribution from the higher amount of immobilized enzymes on the surface. Among all enzyme immobilization method variations, the surface covalent immobilization of enzyme 3-D aggregate on chitosan coated surfaces yielded the highest peak intensity at 1650 cm⁻¹, while the surface covalent immobilization of enzyme mono-layer on glutaraldehyde activated chitosan coating showed the lowest. The difference between the two samples lies in the sequence of adding crosslinking agent and enzyme; the 3-D aggregate method added both ingredients at the same time in a one-pot procedure while the mono-layer method was a multi-step procedure in which the chitosan coating was activated with glutaraldehyde followed by rinsing off the excess free glutaraldehyde and then adding enzyme. As a result, the 3-D aggregate method is expected to form mostly enzyme-GA-enzyme crosslinks whereas the mono-layer method should primarily form chitosan-GA-enzyme crosslinks. In the hybrid methods, with enzyme pre-entrapped in a chitosan matrix, both enzyme-GA-enzyme crosslinks and chitosan-GA-enzyme crosslinks can exist to varying degrees depending on the combination (Figure 1).

2.2. Surface Morphology

Scanning electron microscopy (SEM) was used to characterize the changes in surface morphology of the cheesecloth support materials subjected to different enzyme immobilization methods. As shown in Figure 3 and Figure S2, the cotton cheesecloth fabric had a loose plain weave construction made of single-ply yarns that were each ~200 μm in diameter containing multiple ribbon-shaped cotton fibers with individual widths ranging from a few microns across the narrow thickness of the ribbon to around 20–30 μm across the widest part. The length of individual cotton fibers (not measured) would typically be ~25–30 mm [87]. The size of the rectangular holes between yarns were around 200–500 μm in the widest dimension. As shown in Figure 3a,b, crosslinked chitosan coating no-enzyme control exhibited smooth, uniform fiber surfaces free of chitosan aggregation, which is similar to the un-crosslinked chitosan coating observed previously [72] and confirmed that the thin coating was on the single fiber level. The attachment of a mono-layer of enzyme on the activated chitosan coating (schematic shown in Figure 1) did not show any observable changes to the fiber surface as can be seen in Figure 3c,d, which is consistent with the low immobilized enzyme peaks seen by FTIR. We refer to this coating as a “mono-layer” because the two-step immobilization method expects that once the first enzymes attach to the crosslinking sites, they occupy that area and no further covalent attachment is possible for the later enzymes at the same place. As expected, the resolution limits of the SEM image do not distinguish the small diameter (about 5 nm [88]) of a typical individual carbonic anhydrase enzyme. Therefore, the “smooth uniform” surface appearance is consistent with monolayer coverage, even if the “mono-layer” has some gaps or has places where enzymes have formed slightly thicker layers or clusters that are not large enough to be detectable by SEM. In contrast, enzyme aggregates produced by the “one pot” immobilization method are clearly visible in SEM images as extra material on the fibers and at the interstitial spaces of the fabric openings and in between the cotton fibers of the surface covalently immobilized enzyme 3-D aggregate sample shown in Figure 3e,f. Some of these aggregates are large enough to span and occlude part of the empty space between yarns in a way that could restrict or alter gas and liquid flow patterns. The large amide bonds detected by FTIR for this sample indicate that the exposed extra material visible by SEM in the 3-D aggregate is due to the presence of large amounts of enzyme protein held together to form solids by substantial enzyme-to-enzyme crosslinks through glutaraldehyde. SEM images for hybrid immobilization methods with pre-entrapped enzymes are included in Supporting Information Figure S2 for comparison and are discussed below.

Whereas post-enzyme-entrapment crosslinking preserved a relatively smooth surface on individual fibers (similar appearance of Figure S2b and Figure 3b), the presence of enzyme entrapped in the chitosan coating appeared to facilitate a “thickening” of the coating when additional enzyme was applied to the surface by the “mono-layer” surface attachment method. Less underlying cotton fiber surface detail is visible in Figure S2d compared to either Figure S2d or Figure 3b. This was unexpected from the multi-step immobilization methodology perspective, because a single added mono-layer of enzymes should not be observable at this SEM resolution. However, the observed “thicker” coating after “mono-layer” enzyme was attached to a chitosan coating already containing entrapped enzyme is consistent with the more intense amide peaks observed by FTIR for this sample (purple spectrum in Figure 2) compared to the crosslinked entrapped sample (green spectrum in Figure 2). Therefore, the large increase in enzyme amide bands after “mono-layer” attachment is attributed to a thicker than expected protein coating caused by enzyme aggregation at cotton fiber surfaces from a combination of chitosan-GA-enzyme crosslinks and enzyme-GA-enzyme crosslinks. Considering the step-wise sequence of the mono-layer immobilization method, enzyme-GA-enzyme crosslinks could have formed through reaction with partially exposed enzymes previously entrapped in the chitosan coating and/or with enzymes leached to the coating surface during the glutaraldehyde surface activation process. Some amount of enzyme may also be present due to adsorption to the surface, without covalent attachment, which could be promoted by the presence of already entrapped enzymes.

The hybrid surface covalently immobilized enzyme 3-D aggregate with pre-entrapped enzymes exhibited a distinctly different fiber surface morphology compared to other samples, with high amounts of enzyme aggregate extending only between closely adjacent cotton fibers (Figure S2f). These smaller inter-fiber aggregates did not form the same larger “sheets” of inter-fiber aggregates that were observed in the sample with only chitosan coated on the fiber surface prior to enzyme immobilization by the “one pot” covalent 3-D aggregate method (Figure 3e,f). A possible explanation for this difference is that some of the crosslinking agents applied during the “one pot” stage of the hybrid treatment reacted rapidly with partially exposed enzymes that were already pre-entrapped in the chitosan coating, thus creating affinity and reaction sites to draw enzymes from the bulk immobilization solution close to the fiber surface (Figure S2f), rather than allowing time for enzyme-GA-enzyme aggregates formed in the solution to accumulate as larger “sheets” between fibers (Figure 3f) as occurred with the “one pot” surface covalent 3-D aggregate method alone. In the “one-pot” case, less GA may have initially reacted at the chitosan-only surface leading to more enzyme-GA-enzyme reaction in the solution and film (“sheet”) formation when the material dried. Both the SEM morphological observations and the ranking of amide group peak intensities in the FTIR spectra support this hypothesis.

2.3. Liquid Transport Property

Clean cotton fibers are well known to absorb and spread water quickly due to the hydrophilic nature of the cellulose surface. “Clean” means that natural cotton waxes and any hydrophobic processing aids have been removed during conventional commercial textile wet processing. Coating and enzyme immobilization alters the surface chemistry of clean cotton support materials, potentially impacting the hydrophilicity. Therefore, the wettability of the enzyme immobilized surfaces was evaluated. A previous vertical wicking experiment monitored by neutron radiography and computed tomography [73] revealed a delayed and more constrained water transport behavior within a chitosan coated cotton yarn compared to the instantaneous upwards wicking of water against gravity in a pristine cotton yarn. Similarly, static water contact angle measurements were consistent with vertical wicking experiments, indicating that chitosan coating significantly changed the wetting behavior of the cotton fabric [72]. Such observations invited further study of the dynamic liquid transport behaviors of these materials.

The Moisture Management Tester (MMT) (Supporting Information Figure S3) is an instrument that can measure liquid transport properties of textiles fabrics by monitoring the horizontal spreading of a test liquid—usually saline solution—outwards from the center of the fabric on both the top and the bottom surfaces separately, through a series of conductive sensors arranged in concentric rings. Listed in

Table 2. MMT liquid transport properties of the cheesecloth support materials with different coatings.

Sample ID	Wetting Time -Top (s)	Wetting Time -Bottom (s)	Spreading Speed -Top (mm/s)	Spreading Speed -Bottom (mm/s)
Neat cheesecloth	3.40	3.37	2.85	2.90
Chitosan coated	44.98	7.25	0.14	0.69
Chitosan coated with entrapped enzyme	19.14	8.55	0.48	0.76
Chitosan coated + surface covalent 3-D enzyme aggregate	2.95	3.04	7.04	6.98

are the top and bottom wetting times, which represent the times required for test liquid to begin wetting that surface after the test is started, and the top and bottom spreading speeds, which are the average rate that liquid traveled from the center to the maximum wetted radius. MMT results show that neat cotton cheesecloth supporting material has a short wetting time of 3.4 s and fast spreading speed of 2.9 mm/second, with both sides of the fabric behaving almost identically. However, a drastic increase in wetting time and reduction in spreading speed was observed on samples coated with chitosan. The top and bottom measurement values also differ significantly for this sample. This response is probably due to chitosan coated surfaces having a delayed interaction with (or wetting by) the liquid which gives the liquid enough time to first seep through the openings in the loosely woven textile structure to reach the bottom surface of the sample under the influence of gravity. This seeping phenomenon could overcome initial surface tension experienced on the top of the fabric, allowing the liquid to wick faster at the bottom surface. Although the very thin chitosan coating did not interfere with the liquid quickly passing through the ~300 μm openings in the fabric structure, the surface wetting and liquid transport properties of the coated fabric were significantly slowed.

When enzyme was entrapped in the chitosan matrix, the wettability and spreading speed improved somewhat, which is consistent with partially exposed entrapped enzymes presenting their hydrophilic outer surfaces for interaction with the liquid. Complete wetting recovery was observed for samples with covalently immobilized 3-D enzyme aggregates on chitosan coated cotton surfaces. The equivalent wetting times of 3.0 s and spreading speeds of 7.0 mm/s on both sides of the fabric exceeded the performance of the neat cotton cheesecloth. Since presence of enzyme was beneficial for the overall wetting time and spreading speed for the chitosan coated with entrapped enzyme sample, enzyme hydrophilicity is also attributed with helping improve the liquid transport properties of the 3-D enzyme aggregate sample. The microscopic “sheet” morphology of the 3-D aggregates (Figure 3f) may also play a role in accelerating the wetting time, but this requires further investigation. Clearly, though, entrapped or covalently bonded enzymes located on the textile support help maintain, and even contribute to further improving, the widely-recognized superior liquid transport properties of pristine cotton fabrics.

2.4. Assay Scale Longevity Tests

A primary motivation for adopting immobilized enzymes is their reusability. The working longevity of biocatalytic textiles is critical for minimizing the overall cost of such technology in a CO₂ capture system. In our prior study of a non-crosslinked chitosan entrapment method [72], the level of retained enzyme activity decreased significantly for samples that were handled and rinsed repeatedly during 31 days of testing compared to samples that were incubated in solvent for the same amount of time but were handled only once, indicating that rinsing and handling had a negative effect on immobilization longevity. Assay-scale cycling tests were therefore carried out to evaluate the effects of crosslinking on the longevity of biocatalytic textiles. In these “static” cycling tests, each textile sample remained in the well of an assay plate while p-NP enzyme activity assay reagents and wash solutions were repeatedly added and removed by pipetting. All the crosslinked samples and the entrapment-only sample remained physically intact for 14 cycles in this ambient temperature test (Supporting Information Figure S4), and all samples retained >71% activity compared to their initial activity values.

The “static” cycling test only exposed samples to mild mixing and low levels of physical stress, whereas a real application would expose materials to more stresses caused by continuous gas and liquid flows and handling during fabrication, storage and use. This means that differences between samples that were indistinguishable in the mild “static” cycling test could potentially lead to large differences in performance over the long-term. To accelerate the effects of prolonged use and handling, samples that completed the 14 cycle “static” test within 4 days were transferred to a more aggressive stress test that subjected the textile samples to continuous end-over-end mechanical agitation while immersed in a model CO₂ scrubbing solvent (10% K₂CO₃) using a rotisserie-type incubator (Supporting Information Figure S5). This level of agitation is higher than what textile packing materials would experience in a conventional gas-liquid absorber column, and hence represents a “worst case” physical stress scenario. During the test, the woven structure of the non-crosslinked samples disintegrated within a few days whereas all the crosslinked samples remained structurally intact after 31 days of mechanical agitation in the rotisserie (bottom picture in Figure S5). Additionally, as shown in Figure 4, enzyme activity retention of the non-crosslinked sample dropped well below that of the crosslinked counterparts, indicating that the non-crosslinked chitosan thin coating on cotton fibers was damaged by separation and abrasion of the fibers, causing release of enzyme into the test solution and removal during the washing steps. At the end of 25 days of rotisserie incubation at 27 °C (to mimic a “cool” absorber column condition), the non-crosslinked sample retained only 20% of its original activity in comparison to the 79–92% retained by crosslinked samples. To further stress the samples, the incubation temperature was increased to 45 °C (to mimic a “typical” absorber column condition) and maintained for an additional 6 days, overall totaling 31 days of rotisserie treatment time. The final activity retentions of crosslinked samples ranged from 49% to 59% while that of the non-crosslinked sample diminished to only 2%. Among the four crosslinking treatments, the low glutaraldehyde concentration (0.2%) with the longer crosslinking time (3 h) achieved the highest activity retention. This experiment showed that crosslinking helps preserve the fabric structure and holds enzyme on the fabric during solvent immersion with continuous mechanical agitation. These promising results with post-entrapment crosslinking led to further experiments that compared several different crosslinking strategies for securing enzyme to the textile with glutaraldehyde. The additional strategies included surface-only covalent attachment and hybrid methods that combined entrapment with surface covalent attachment (Figure 1).

Biocatalytic textiles with NZCA immobilized by surface-only or hybrid immobilization techniques were fabricated and then stress-tested by incubation in 30 wt% MDEA maintained at 45 °C with 120 RPM orbital mixing motion. Samples were taken out periodically over the course of 30 days for esterase activity assay. The results are presented in Figure 5, measured both as p-NP release rates and as percent activity retentions versus initial activities. According to p-NP release rate results (Figure 5a), cheesecloth coated with a surface-only covalent monolayer had the lowest active enzyme, followed in increasing order by surface-only covalent 3-D aggregate, cross-linked entrapped, and the two hybrid samples made with both entrapped and surface attached enzymes. Active enzyme loadings of each sample expressed as esterase activity unit per gram of cellulose were tabulated in

Table 3. Active enzyme loadings measured for different immobilization strategies by the esterase assay, compared to the maximum CO₂ absorption measured in a laboratory scrubber at L/G 30 mL/L using 10 wt% K₂CO₃ solvent.

Immobilization Preparation ID	Active Enzyme Loading (U/g of Cellulose)	CO2 Absorption (%)
Chitosan coated + surface covalent mono-layer	0.014	71.2
Chitosan coated + surface covalent 3-D aggregate	0.188	66.7
Chitosan entrapping NZCA + cross-linked	0.212	56.5
Chitosan entrapping NZCA + surface covalent mono-layer	0.280	n.a. 2
Chitosan entrapping NZCA + surface covalent 3-D aggregate	0.316	48.6
Textile control (no enzyme)	0	23.4
Textile control + dissolved NZCA (7 U/L of solvent) 1	0	84.5
Textile control + dissolved NZCA (26 U/L of solvent) 1	0	94.5

¹ Data from Ref. [72]. ² n.a. = not available.

to show these trends. The hybrid monolayer sample yielded higher p-NP release rates than one would expect from the additive sum of surface-only monolayer plus cross-linked entrapment samples, yet this relatively high activity result, implying the presence of correspondingly high amounts of enzyme, agreed with similar trends observed by FTIR protein peak intensities (Figure 2) and by the “thicker” coating on cotton fibers observed by SEM (Figure S2d). This indicates that the presence of protein in the chitosan coating provides more opportunities for enzymes applied by post-treatment to attach to the coating surface, which could be a useful strategy for optimizing enzyme loading.

Examination of the percent activity retention (Figure 5b) reveals potential enzyme deactivation mechanisms in heated alkaline scrubbing solvent. Surface-only covalent 3-D enzyme aggregate sample (red curve) experienced the steepest drop in percent activity retention over time followed by surface-only covalent monolayer (black curve). Neither of these samples had enzyme entrapped (“buried”) in their chitosan coating, therefore, the p-NP enzyme activity response of these samples relied entirely on enzymes that were physically “exposed” on surfaces. The dramatic loss of activity by these samples is consistent with a surface erosion deactivation mechanism, where enzyme molecules or aggregates detach from the surface and are washed away by the incubation solvent and assay rinsing steps prior to conducting the assay test. Even though enzymes are crosslinked to the chitosan coating or to each other by GA crosslinks, literature indicates that the reaction of GA with amino groups is reversible, except between pH 7.0 to 9.0 where only little reversibility is observed [85]. With a pH of 10.5 in the 30% MDEA test solvent, some cleavage of GA crosslinks should therefore be expected over time. Furthermore, due to its relatively long arm and low reactivity, GA is less conducive to forming highly stable multipoint covalent immobilizations and other mechanisms, such as hydrophobic adsorption and ion exchanges, can play significant roles in the overall immobilization process for molecules bonded by GA reactions [33]. Thus, the “burst” phase of activity loss observed within the first 8 days could be primarily due to leaching or desorption of non-covalently adsorbed enzymes or enzyme aggregates, followed by slower activity loss due to gradual loss of crosslinks. The slightly better activity retention of the surface-only covalent monolayer compared to the surface-only 3-D aggregate can be explained by the type of linkages they have. Surface-only covalent monolayer sample has only chitosan-to-enzyme linkages that are formed more densely and close to the surface while surface-only covalent 3-D aggregate sample has mostly enzyme-to-enzyme linkages that are loose and not anchored to the surface as tightly, making it easier for 3-D aggregate to erode. Another mode of enzyme inactivation is protein unfolding (“denaturation”) upon exposure to heat and aqueous liquids that promote molecular motion. Exposed enzymes that are not physically prevented from unfolding are more likely to denature and loose activity when the critical three-dimensional shape of their active site is compromised. Because all samples were treated in the same way in our study, other inactivation modes, like the presence of contaminants, are not implicated in explaining the different longevities observed between the different immobilization methods. Importantly, even after 30 days incubation at 45 °C in 30% MDEA, more than 40% of the initial activity remains in these samples, and this remaining activity will be at the “exposed” surface where enzymes have the greatest possibility to interact rapidly with CO₂ gas molecules when the biocatalytic textile is configured as a gas-liquid contactor.

The crosslinked enzyme entrapped sample (blue curve in Figure 5b) retained a higher percentage of activity (≥70%) than the surface-only samples (black and red curves). The hybrid covalent 3-D aggregate (purple curve) and hybrid “mono-layer” (green curve) samples yielded the highest activity retention (≥80%). Because the p-NP assay detects all active enzyme, even enzyme “buried” inside the coating, it makes sense that all samples with enzyme entrapped in chitosan both started with a higher total enzyme activity (Figure 5a and Table 3) and that the activity was better retained, because entrapped enzymes were more protected from the denaturing effects of heat and solvent and were less susceptible to leaching because they were “trapped” in the coating behind a layer of GA-crosslinked enzyme(s). This stabilization effect of the matrix and additional coating is well documented in literature [33]. However, the additional protection sometimes comes with additional mass transfer limitations. The rate of ester hydrolysis monitored by the p-NP assay is relatively slow, which is an advantage for detecting all immobilized enzyme activity, even that contributed by enzymes buried in the chitosan matrix, and is helpful for explaining enzyme location, loading and deactivation mechanisms. However, the esterase assay does not account well for molecular mass transfer limitations experienced by very fast reactions, meaning that esterase activity results can be inadequate for predicting biocatalyst performance at actual scrubber application conditions, where the catalytic pathway is through the ultra-fast CO₂ hydration reaction with water to form bicarbonate ions and where the speed of molecular mass transfer plays a dominant role. Thus, the selection of the most effective immobilization method needs to be made holistically, based both on the assay-based longevity results and after considering the actual CO₂ capture performance in simulated absorber conditions.

2.5. Lab-Scale CO₂ Gas Scrubber Test

A lab-scale CO₂ gas absorber column was set up to test the application performance of biocatalytic textile packings by simulating the physical process configuration and liquid-to-gas contact that occurs in conventional chemical absorption processes. Textile packing modules were designed to “drop-in” to the absorber column for measurement of CO₂ capture performance when exposed to CO₂-containing gas mixtures in the presence of CO₂ absorption liquids. Although the lab scale system was constrained to delivering relatively low total gas flow rates (maximum tested was 8 L per minute, LPM), the system was capable of delivering L/G (liquid-to-gas) flow ratios ranging from 3.3 to 30 mL/L, which simulate actual L/G ratios practiced in commercial CO₂ scrubbers.

As shown in Figure 6, surface-only covalent monolayer packing achieved the highest CO₂ capture efficiency of 71.2% at an L/G of 30 mL/L despite having only 7% total enzyme loading (based on esterase activity, Table 3) compared to the surface-only covalent 3-D aggregate (66.7% capture efficiency), or 4% compared to the hybrid covalent 3-D aggregate (entrapment + surface covalent 3-D aggregate, 48.6% capture efficiency). This represents about a 3-fold CO₂ capture enhancement from no-enzyme textile control packing and a 20-fold enhancement compared to conventional Raschig ring random packings. The result is consistent with Yong et al.’s finding [51] that CA immobilized through Layer-by-Layer (LbL) electrostatic adsorption on the surface of microporous polypropylene (PP) hollow fiber membranes showed a 3-fold increase in rate of CO₂ absorption into 30% K₂CO₃ solvent compared to the non-biocatalytic hollow fiber membrane contactor alone an absorption unit using 100% CO₂ gas. However, in contrast to their additive p-NP release rate results, our hybrid immobilized enzyme packing, that had additional surface covalent 3-D aggregate immobilized on top of the NZCA-entrapped chitosan matrix (NZCA:Chitosan (mL:g) = 2:1), did not exhibit an additive CO₂ capture efficiency increase. This discrepancy is due to fundamental dynamic differences between the p-NP assay (long substrate exposure; not mass transfer limited) and the CO₂ absorption test (short substrate exposure; mass transfer limited) that are influenced by microscopic physical differences between the samples. The hybrid packing with entrapped and surface covalently immobilized NZCA (Figure S2f) had a high amount of enzyme aggregate in between individual cotton fibers within the yarn structure leading to a decrease in porosity and a potential change in the liquid flow pattern that could reduce the liquid and gas contacting areas and could also conceal enzymes under a thicker liquid layer (causing localized “flooding” at the fiber level). Decreasing gas-liquid contact area and decreasing exposure to enzymes leads to lower CO₂ capture efficiency. In our previous study [72], positive control experiments using un-coated textile structured packing together with dissolved free enzyme at concentrations of 7 U/L and 26 U/L of solvent yielded high CO₂ capture efficiencies of 84.5% and 94.5% (Table 3), indicating there is still room to improve the performance of CA-immobilized packings up to these levels. Clearly, therefore, improving the packing CO₂ capture efficiency depends on positioning immobilized enzymes on or near the surface of fibers in a way that simultaneously exposes them to both liquid and gas molecules, as well as efficiently transports both gas and liquid through the system. Excess material that alters the beneficial yarn structure or decreases accessibility of gas molecules to enzyme active sites compromises the overall CO₂ capture performance of the packings. Consequently, as observed in this work, when enzymes are entrapped within an immobilization matrix, the total enzyme loading (measured by the esterase activity) does not directly correlate to CO₂ absorption efficiency (Table 3). In future studies, detailed tracking of activity balance [89] among activities of free enzyme at reference immobilization conditions, in immobilization solution supernatants, in crosslinked immobilization suspensions, and on immobilized solids will be carried out to optimize enzyme loading. In fact, the findings suggest that loading small particles comprising enzymes on the textile fiber surfaces by covalent or non-covalent means, or passing such particles across the fiber surfaces in the presence of liquid flow, could be another way to enhance CO₂ absorption efficiency. Subsequent investigations will seek to better understand the relative importance of alternative immobilization mechanisms [74], such as the possible initial ion exchange stage [75] that may occur during covalent immobilization of enzymes on GA-activated amine functionalized surfaces (i.e., chitosan coated cotton fiber), because this may affect enzyme orientation, expressed activities, and ultimately the CO₂ capture efficiency.

The effect of L/G ratio was evaluated by decreasing the liquid flow rates while maintaining a constant gas flow rate. In general, biocatalytic textile packings efficiently captured CO₂ from the gas mixture over a wide range of L/G until dropping below 6 mL/L where the amount of 10 wt% K₂CO₃ solvent delivered to the packing become so small that CO₂ saturation likely occurred. Surface-only covalent 3-D aggregate packing was selected for subsequent tests due to its acceptable initial enzyme loading (Figure 5a and Table 3), excellent CO₂ capture efficiency of up to 66.7% at high L/G and comparable capture efficiency to entrapment techniques at low L/G (Figure 6), and fast liquid transport behaviors (Table 2). Although the surface-only covalent 3-D aggregate immobilization technique resulted in a relatively high loss of total enzyme activity over time according to the p-NP assay (Figure 5b), a deactivation mechanism caused by surface erosion could result in the active surface of packing material being constantly refreshed by exposure of underlying layers, which could extend actual CO₂ capture performance beyond that predicted by the esterase assay.

Notably, the CO₂ capture efficiency of biocatalytic textile packing was rather stable across a wide L/G range of 6–30 mL/L at a fixed gas flow rate of 4 LPM. Gas and liquid flowed steadily through the textile modules without experiencing column flooding, splashing, pooling, channeling or causing liquid to run in streams along the column walls (“wall effects”) that are critical design considerations for conventional structured packing made of non-absorbent materials like stainless steel [90]. By analogy to a cotton dishcloth, the yarns of textile packings were very efficient at absorbing, spreading and directing the liquid flow throughout the spiral wound packing modules, “wicking” from side to side (Section 2.3) due to the hydrophilic properties of cotton yarns and pulled from top to bottom by the force of gravity. This excellent solvent distribution together with intrinsic high level of control over the liquid flow is a particular advantage of the textile-based packing materials and design. Liquid will self-distribute through the packing and will be constrained to the geometry of the yarns in such a way that even some tilting of the packing away from vertical alignment will not substantially disrupt the liquid flow pattern. The constrained liquid flow through textiles also opens possibilities for their use in intensified process reactor designs, such as rotating packed bed reactors [91] which exploit a high centrifugal force to overcome gravitational force in spreading liquid through the packing to enhance gas-liquid interaction and CO₂ absorption mass transfer.

2.6. Effects of Solvent and Gas Flow Rates

The extent of change in CO₂ capture efficiencies caused by lowering the solvent flow rates or increasing the gas flow rates was compared using surface-only covalent 3-D aggregate packing. As shown in

Table 4. Effects of solvent and gas flow rates on CO₂ capture efficiency of surface-only covalent 3-D aggregate enzyme packing.

Solvent Flow Rate (mL/min)	N2 Flow Rate (LPM)	CO2 Flow Rate (LPM)	Starting CO2%	Lowest CO2%	%CO2 Captured	Effect of Gas Flow Increase	Effect of Solvent Flow Decrease	Liquid to Gas Ratio (mL/L)
120	3.6	0.4	10.9	3.4	68.8%			30
120	7.2	0.8	11.1	6.3	43.2%	62.9%		15
72	7.2	0.8	11	6.8	38.2%	58.1%	88.3%	9
72	3.6	0.4	10.8	3.7	65.7%		95.5%	18

, when keeping gas flow rate constant at a total of 4 LPM (Row 1 and 4) or 8 LPM (Row 2 and 3), decreasing the solvent flow rate by 60%, from 120 mL/min to 72 mL/min, did not reduce the CO₂ capture efficiency to the same extent. Instead, over 95% and 88% of the initial CO₂ absorption performance was retained. This implies that near optimal liquid distribution throughout the packing was already achieved at the lower flow rate. The additional inventory of lean solvent flowing through the packing at the higher flow rate only provided a modest CO₂ absorption increase. This behavior implies that the benefit of fresh lean solvent was off-set by “local flooding” within yarns, resulting in that a portion of the higher solvent flow was not effective in CO₂ absorption. Higher solvent flow sends more liquid through yarns in the textile packing causing liquid to “fill” spaces between individual fibers in the yarns. Filling the “air” spaces with liquid reduces the effective contact area between gas molecules and hydrated solid interfaces and inhibits molecular mass transfer interactions with immobilized enzyme active sites. In contrast, when gas flow rate was doubled, the CO₂ capture efficiency dropped significantly, to about 63% and 58% of the original values. This is close to the approximate 50% capture efficiency reduction that would correspond to a 50% decrease in gas molecule residence time in the column. These observations indicate that textile packing performance could be improved by optimizing surface chemistries, packing materials, packing design, and process configuration (e.g., operate in co-current mode) to accelerate liquid transport, prevent yarn “flooding” and maximize gas-liquid-enzyme contacting interfaces.

For comparison, Migliardini et al. [60] found 38% (counter-current mode) and 6% (co-current mode) CO₂ conversions for their three-phase trickle bed absorption column packed with immobilized CA made by entrapment in shredded polyurethane (PU) foams. However, only a low liquid flow rate of 10 mL/min was reported for this comparison, possibly due to the flooding issue commonly seen for densely packed columns in counter-current mode at high liquid flow rates. Their highest CO₂ conversion of 45% was achieved in co-current mode at a liquid flow rate of 100 mL/min and a gas flow rate of 0.5 L/min, equating to a high L/G (mL/L) ratio of 200. Such a high L/G would require high energy demand for liquid pumping and high energy in the desorption stage to recover lean solvent. The fact that our biocatalytic textile structured packings were able to operate flood-free in counter-current mode and perform well at lower L/G ratios warrants future scale-up studies in larger scale CO₂ reactive absorption systems.

2.7. Effect of Solvent Types and Concentrations

Enzyme immobilized textile structured packing compatibility and performance was tested with several different kinetically limited CO₂ absorption solvents, including aqueous solutions of K₂CO₃ at several concentrations, N-methyldiethanolamine (MDEA) and N,N-dimethylglycine (DMG). All solvent compositions were effective in absorbing CO₂ from the gas mixture and were conducive to reaction rate enhancement by an enzyme immobilized textile packing (

Table 5. CO₂ capture efficiencies of surface-only covalent 3-D aggregate immobilized enzyme packing using various solvents.

Packing ID	10% K2CO3 pH 10.51	20% K2CO3 pH 10.51	10% DMG pH 10.90	5% K2CO3 pH 10.50	5% DMG pH 10.76	5% MDEA pH 11.15	10% K2CO3 pH 10.47	30% K2CO3 pH 10.60
Conventional Raschig ring packing	3.6%	4.5%		5.4%
No enzyme textile control packing	23.4%	20.9%		23.6%	31.3%	22.0%	24.3%
Surface-only covalent 3-D aggregate packing	68.8%		70.9%	56.3%	66.4%	62.7%	68.2%	49.1%

). In Table 5, the solvent test results are arranged in the time sequence for when each test was performed, from left (first test) to right (last test). One module of no enzyme textile control packing was repeatedly used for all the no-enzyme tests, and one module of surface-only covalent 3-D aggregate packing was repeatedly used for all the enzyme tests. These two modules were rinsed well in tap water, dried and pre-soaked in new solvent between each test with a different solvent. For the tests with conventional Raschig ring packing, the rings were randomly filled in the absorber column to the same height as the height of a textile packing module for comparison. Using the same 10 wt% K₂CO₃ solvent at the same testing conditions, the no-enzyme textile control packing and surface-only covalent 3-D aggregate immobilized enzyme packing exhibited enhancement factors of 6.5 and 19.1, respectively, over conventional Raschig ring packing. This means that the textile packing itself, even absent enzyme, provides significant CO₂ capture enhancement by promoting excellent liquid distribution and gas-liquid interaction. Since the no-enzyme textile control packing yielded very consistent capture efficiency of 23–24% using 10 wt% K₂CO₃ solvent in the first and last scrubber tests, differences in performance observed between the first and last tests can be attributed only to differences in absorption behavior by the different solvents. Among all three solvents at 5 wt% concentration without enzymes, DMG performed the best, followed by K₂CO₃ and MDEA. Notably, reducing the K₂CO₃ concentration from 10 wt% to 5 wt% did not affect the CO₂ capture efficiency of the un-catalyzed absorption process with no-enzyme textile control packing. However, for the catalyzed absorption using surface-only covalent 3-D aggregate immobilized enzyme packing, 5 wt% K₂CO₃ exhibited lower capture efficiency than 10 wt% K₂CO₃. This result is explained by that the high capture efficiency resulting from immobilized enzyme enhancement caused the 5 wt% K₂CO₃ solvent to exceed its CO₂ loading capacity which manifested in a decreased capture efficiency compared to the more concentrated 10 wt% K₂CO₃. Overall, DMG had the highest capture efficiency at both 5 wt% and 10 wt% concentrations and in both catalyzed and un-catalyzed processes, highlighting its potential as an alternative CO₂ absorption solvent when used together with CA. Another notable observation is that the higher 20 wt% and 30 wt% concentrations of K₂CO₃ reduced the capture efficiency for un-catalyzed and catalyzed absorption. Higher solution viscosities of more concentrated solvents are suspected of changing liquid flow behavior or increasing liquid film thickness leading to fewer gas-liquid-enzyme interfaces and a sacrifice of overall absorption performance. While some aspects of these fluid dynamics may be an artifact of the low gas flow rates used in the experiments, it is striking, and contrary to conventional wisdom, that textile-based packings allow low solvent concentrations to perform as well as or even better than higher solvent concentrations, which, if realized at larger scale, could result in operational cost savings and improved process safety and sustainability. Regardless, methods to minimize viscosity and surface tension effects of the solvent are expected to help textile packing system performance. Strategies to overcome viscosity issues could also include redesign of textile packing components and their porosity, use of alternative absorber configurations, surface chemistry modifications, and use of fluid dynamics simulations to optimize fluid flow.

2.8. Longevity of Enzyme Immobilized Textile Packing

Reusability and longevity are critical to achieving commercial implementation of enzyme-based CO₂ capture technologies. Preliminary selection of enzyme immobilization methods to use with textile packing materials was made based on assay-scale longevity testing (on swatches of fabrics) but the working longevity of the entire textile structured packing with immobilized enzyme must also be evaluated. A surface-only covalent 3-D aggregate immobilized enzyme packing module was tested 10 times to evaluate its reusability over a period of 71 days. When not undergoing testing, the packing was rinsed in cold tap water to remove solvent, allowed to air dry, and stored open to air at ambient lab conditions. The last 100 h were spent soaking in the scrubbing solvent (10 wt% K₂CO₃ pH~10.50) at an elevated temperature of 45 °C. Despite the repeated test, rinse, and dry cycles and the 100 h incubation in the solvent at a temperature typical for commercial absorber columns, enzyme immobilized packing exhibited 100% retention of CO₂ capture performance relative to the first cycle test (Figure 7). When the same packing was tested again after 1 year of dry storage at ambient lab conditions, over 85% of its original CO₂ capture efficiency was retained. For comparison, a biocatalytic membrane created by Xu et al. [52], using co-deposited polydopamine and poly(ethylenimine) (PEI) on PVDF hollow fiber membranes together with glutaraldehyde cross-linkers to covalently immobilize CA, generated 165% higher CO₂ absorption flux than the no-enzyme control and retained 73% of its initial esterase activity after 40 days of storage.

After the long-term repeated test-rinse-storage reusability test (Figure 7), the same packing module was later subjected to a continuous lab-scale CO₂ gas scrubber test where room temperature solvent flowed through the packing for the entire duration of the experiment and the CO₂ capture efficiency was measured at different time intervals. As shown in Figure 8, surface-only covalent 3-D aggregate immobilized enzyme packing demonstrated steady performance over a total of 500 h in this continuous solvent flow scrubber test. Throughout these tests, the module remained intact and showed no signs of wear, damage or contamination. In summary, enzyme immobilized textile structured packing is resistant to numerous application-relevant stressors such as repeated tap water wash and air drying, solvent immersion at elevated temperature, and continuous solvent flow, and it is stable for long-term dry storage at ambient conditions.

2.9. Versatile Application Conditions

CO₂ levels present in flue gas generated at coal-fired power plants are typically in the range of 10–14%, while lower percent CO₂ levels (~5%) occur in the flue gas of natural gas-fired power plants, and a higher percent CO₂ levels (~40%) occur in raw biogas sources. All these levels are important CO₂ concentration ranges for point source CO₂ capture applications. Nominal 5 vol% and 25 vol% CO₂ levels were used to evaluate and compare the effectiveness of textile packings at these CO₂ concentrations. For reference, laboratory gas scrubber tests on the same set of packing modules at a nominal 10 vol% CO₂ level yielded 23.1% and 66.7% CO₂ capture efficiencies for no-enzyme textile control packing and surface-only covalent 3-D aggregate enzyme immobilized textile structured packing, respectively. With total gas mixture flow rate (4 LPM) and solvent flow rate (120 mL/min) kept constant, reducing the CO₂ concentration in the gas mixture from the regular nominal 10 vol% to nominal 5 vol% resulted in an increase in the CO₂ capture efficiencies for both the control and enzyme packing (

Table 6. CO₂ capture efficiencies of no-enzyme textile control packing and enzyme immobilized textile structured packing at low and high CO₂ concentrations.

Packing ID	Gas Flow (LPM)	Solvent Flow (mL/min)	L/G (mL/L)	Starting CO2%	Lowest CO2%	CO2 Capture Efficiency (%)
No-enzyme control	4	120	30	5.8	4.4	24.1%
Enzyme packing	4	120	30	5.8	1.8	69.0%
No-enzyme control	4	120	30	26.2	21.6	17.6%
Enzyme packing	4	120	30	26.1	13.3	49.0%
No-enzyme control	1.6	120	75	24.6	14.9	39.4%
Enzyme packing	1.6	120	75	24.7	3.0	87.8%

). When the nominal CO₂ concentration was increased to 25 vol%, while keeping the gas and solvent flow rates constant, the CO₂ capture efficiency decreased. The capture efficiencies were then improved by slowing the gas flow rate because the slower gas flow increased gas molecule residence time in the column and also ensured sufficient absorption capacity of the solvent relative to the amount of CO₂ being delivered. Importantly, results shown here pertain to single module packing capture efficiencies. In real applications, the number of stacked packings (column height), total packing cross sectional area (by making larger modules and/or bundling several modules together), solvent and gas flow rates, as well as solvent types and concentrations would be optimized for a desired capture efficiency and cost. The lab results demonstrate the feasibility of utilizing the current packing modules across broad application conditions, including for carbon capture from coal and natural gas-fired power plant flue gas, for industrial flue gas, and for natural gas and biogas upgrading applications.

In addition to capturing CO₂ from point-sources, CO₂ removal from the atmosphere by both natural (e.g., planting trees) and technological approaches (e.g., direct air capture) [92], will be necessary to meet goals set by international agreements [5]. One of the difficulties associated with direct air capture (DAC) stems from the relative low CO₂ concentration in the atmosphere (about 415 ppm in year 2021 [93]) compared to point-source capture concentrations. The feasibility of employing enzyme immobilized textile structured packings for solvent-based DAC was explored by adapting the laboratory gas scrubber system to take in air flow supplied by a laboratory compressed air system. Simulated seawater was prepared as a solvent using generic sea salt mix (for aquarium) at a solute concentration of 3.5 wt% and the pH was adjusted to 10.0 using 1 N NaOH. The air flow rate was set at 1.5 LPM and ambient baseline CO₂ concentrations were found to stabilize between 700–800 ppm, which is higher than the 415 ppm global average, due to the composition of air in the indoor laboratory environment was influenced by human respiration and potentially other CO₂ sources. Seawater flow was turned on at a rate of 120 mL/min at the 5 min time mark after a stable baseline CO₂ level was reached. As shown in Figure 9, the red line represents CO₂ concentration of the exiting air flow when no-enzyme control textile structured packing was used; the black line marks that of the enzyme immobilized textile structured packing. As a reference, the blue line shows the result for seawater sitting statically in the reservoir of the gas absorption reactor while air was delivered to the bottom of the absorber, flowing across the top of the quiescent liquid surface. The essentially flat blue line indicates that static seawater has limited surface contact area which impedes carbon uptake and leaves the CO₂ concentration of the exiting air unchanged. Clearly, simply the process of causing seawater to flow through the packing enhanced the carbon uptake to 19.6% CO₂ capture efficiency, a value close to that achieved by no-enzyme textile control packing using the regular 10 wt% K₂CO₃ solvent and nominal 10% inlet CO₂ concentration, further evidence that the presence of textile alone fundamentally enhances gas-liquid contact. Furthermore, with enzymes immobilized on the textile packing, a CO₂ capture efficiency of 66.9% was obtained, resulting in an outlet CO₂ gas concentration (246 ppm) that is lower than the current global atmospheric average and much lower than the local indoor environment concentration at the time of the experiment. This finding is significant because it demonstrates the feasibility and effectiveness of enzyme-catalyzed textile packing modules for direct CO₂ absorption at ambient conditions from low CO₂ concentration gas mixtures found in ambient air while using low solute content solvents as the capture medium, such as naturally abundant seawater. Other capture media, like traditional CO₂ capture solvents, industrial process water, and fluid in deep saline aquifers, mine tailings or algae ponds, would also be effective and opens the door for making modular, scalable biocatalytic textile CO₂ capture contactors for diverse applications.

3. Materials and Methods

3.1. Materials, Chemicals, and Enzymes

ChitoClear^® 44020–fg95LV (chitosan, degree of deacetylation = 95%) was purchased from Primex EHF, Iceland. Cheesecloth (100% cotton) grade 90 was purchased from Testfabric Inc., West Pittston, NJ, USA. Polyester/cotton latch hook canvas (5 Mesh, IG Design Group Americas Inc., Atlanta, GA, USA) was purchased from a local crafts store. Raschig rings (8 mm × 8 mm borosilicate glass) were purchased from Chemglass Inc., Vineland, NJ, USA. Glacial acetic acid, 1N hydrochloric acid, 4-nitrophenol (p-NP), 4-nitrophenyl acetate (p-NPAc), absolute ethanol, potassium carbonate, potassium bicarbonate, tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl), and 50% glutaraldehyde aqueous solution were purchased from Fisher Scientific, USA. Trizma^® base, N-Methyldiethanolamine (MDEA), and phosphate-buffered saline (PBS) tablets were purchased from Millipore-Sigma, St. Louis, MO, USA. Instant Ocean^® Sea Salt (Spectrum Brands, Inc., Blacksburg, VA, USA) was purchased from a local pet supply store. N,N-dimethyl glycine hydrochloride (DMG-HCl, nutritional supplement grade) was purchased from PureBulk, Roseburg, OR. All chemicals were used as-received without further purification. Experimental thermostable microbial carbonic anhydrase (CA) was obtained from Novozymes A/S, Bagsvaerd, Denmark (a concentrated liquid solution with 18.8 U/mL esterase activity, referred to as NZCA in this study).

3.2. Instrumental Characterizations

A Nicolet Nexus 470 spectrometer equipped with a Nicolet OMNI germanium crystal Attenuated Total Reflection (ATR) sampling head was used to conduct Fourier-transform infrared spectroscopy (FTIR) with a total of 64 scans for each sample at a resolution of 4 cm⁻¹. An Hitachi TM 4000 desktop scanning electron microscopy (SEM) unit was operated in back-scattered electrons mode, secondary electrons mode, or a mix of the two with an accelerating voltage of 10 kV to observe surface morphology of the fibrous materials sputter-coated with approximately 10 nm of gold/palladium.

A SDL Atlas M290 Moisture Management Tester (MMT) was used to test the liquid transport properties of biocatalytic textile samples with specimen size of 8 cm × 8 cm according to American Association of Textile Chemists and Colorists (AATCC) test method 195. A TECAN Spark Microplate Reader UV-VIS spectrophotometer was used to measure the absorbance of the released product from an enzyme activity assay. A Thermoscientific MaxQ 2000 dry shaker bath was used to carry out long-term heat and solvent stress tests of the biocatalytic textiles and a J-KEM 3300 temperature controller was used to maintain the incubation temperature. A DAIGGER FINEPCR combi-D24 rotisserie-type incubator was used to apply end-to-end mechanical agitation and constant temperature control for accelerated longevity tests of the biocatalytic textiles.

3.3. Enzyme Immobilization on Textile Support Materials

Two physical forms of textile support materials were used for enzyme immobilization in this study. The first type were a flat sheets of hydrophilic cotton cheesecloth that was easily cut, coated, and sampled for evaluation in small assay-scale tests for rapid screening purposes. The second type were pre-fabricated cylinder-shaped textile packing modules that incorporated cotton cheesecloth in the design and had a diameter of 6 cm and a height of 22.5 cm. These modules were custom-made in-house following procedures that were detailed in our previous publication [72]. Solution uptake and enzyme immobilization on the packing modules mainly occurred on the flexible hydrophilic cheesecloth component while a rigid low-water-absorbing latch hook canvas acted as physical support for holding up the “jelly roll”-like spiral structure. Spacers placed at the ends of the roll aided gas and liquid flow through the packing. A 1 L graduated cylinder with sufficiently large inside diameter served as a treatment vessel for wet-processing the preformed textile packings in their final forms through dip-coating and/or solution reactions.

Enzyme entrapment using chitosan as the matrix was described in our previous study [72]. Enzyme activity loss during immobilization was minimized by using a mild pH (~5.0) chitosan solution made by re-dissolving freshly cast and air dried chitosan films. Because chitosan films made in this way are protonated and excess acid evaporates during the air drying process, adequate acidity remains to re-dissolve such chitosan films at a concentration of 1 w/v% using only deionized (DI) water and the resulting solution pH is mild enough to not harm enzyme activity. Concentrated carbonic anhydrase stock solution was added slowly into the chitosan solution with continuous stirring at a 1:1 (or 2:1 for hybrid enzyme packing) NZCA:Chitosan (mL:g) ratio. Both the base cotton cheesecloth material and the pre-formed textile packing modules were coated using the prepared solution, followed by a typical room temperature air-drying time of two days. Cheesecloth samples made using four post-entrapment crosslinking conditions, from the combination of two levels of crosslinker concentrations (0.2 vs. 2 v/v% glutaraldehyde) and two levels of reaction durations (30 min vs. 3 h), were evaluated in accelerated longevity tests. Chitosan coated cheesecloth without enzyme was crosslinked with 2 v/v% glutaraldehyde for 3 h to serve as a control for activity assay (Figure S5 bottom, Column 1). PBS buffer (pH 7.4) containing 0.2 v/v% glutaraldehyde and a reaction time of 3 h at room temperature yielded the highest activity retention and was selected for all other crosslinking reactions either on fabric-only or on preformed packings. The crosslinked samples were rinsed with DI water and soaked in Tris buffer (25 mM, pH 7.4) overnight to cap un-reacted free aldehyde groups. Finally, excess Tris buffer was removed by water rinsing and the samples were allowed to air-dry.

Surface covalent enzyme immobilization consisted of multi-step surface attachment or one-pot surface attachment methods, creating surface enzyme monolayer or surface 3-D aggregate, respectively. The multi-step surface attachment method required a first step to activate the surface, i.e., rendering the surface reactive toward enzymes, and a subsequent step to bring the activated surface and enzyme solution together. In this study, amine functionalized surfaces were obtained using chitosan dip-coating, which mimicked the procedures described for entrapment immobilization but without the added enzymes. Amine groups in the chitosan coating were then reacted with 0.2 v/v% glutaraldehyde for 3 h in PBS pH 7.4 buffer for both crosslinking and activation purposes before rinsing with PBS pH 7.4 buffer to remove unreacted glutaraldehyde. The separate enzyme immobilization step took place in PBS pH 7.4 buffer with a stock NZCA concentration of 5 µL/mL assisted by constant stirring at room temperature for 18 h followed by Tris capping and water rinses, similar to the post-entrapment crosslinking. The one-pot surface attachment method differs in that crosslinkers (0.2 v/v% glutaraldehyde) and enzymes (5 µL/mL NZCA) were introduced simultaneously into the reaction buffer (PBS pH 7.4), together with completely immersed textile support materials and constant stirring by an orbital shaker at 120 RPM for flexible cheesecloth swatches or by a magnetic stirrer at the bottom of a graduated cylinder reaction vessel for the preformed textile packings.

As the name suggests, the hybrid immobilization method combines enzyme entrapment with either one of the surface covalent immobilization protocols. The hybrid method creates biocatalytic textiles with enzymes located both inside the chitosan matrices and on the surface of the coating. Illustrations of immobilized enzymes made by different strategies are provided in Figure 1.

3.4. Assay-Scale Longevity Tests

An esterase activity assay [72] was adopted for rapid screening of different immobilization strategies. This assay can detect active immobilized CAs that are otherwise inaccessible to conventional fast diffusion-limited Wilbur−Anderson (W-A) type assays. Samples were cut into donut shapes with an outer diameter that fit snuggly in the wells of a 24-well plate and an inner diameter (hole) of 5/32 inches for a plate reader spectrophotometer signal light to pass through. Four replicate immobilized sample pieces were prepared at the same time in one batch, and each different immobilization recipe was used to make a different batch. Samples were subjected to a “static” longevity test in which the sample was placed in the well of an assay plate and was repeatedly washed and tested without leaving the assay plate. Assay solution from the previous run was replaced with a fresh buffer solution twice between each assay run. However, the “static” test did not distinguish between the longevity of crosslinked and non-crosslinked samples. To differentiate between different immobilization method variations in terms of their durability and longevity, two additional types of mechanical agitation modes were employed, end-to-end mixing in a rotisserie or swirling in an orbital shaker.

3.4.1. Stress Testing by Rotisserie Mixing

An accelerated assay-scale longevity test with continuous end-to-end mixing action was used to evaluate the importance of post-entrapment crosslinking for both textile structure integrity and enzyme activity retention. Four replicate donut shaped samples were immersed together in 20 mL nominal 10 wt% K₂CO₃ solvent (pH~10.5, K₂CO₃/KHCO₃ w/w ratio of 85/15) inside a 50 mL Oak Ridge tube and rotated at a rate of 25 RPM in a rotisserie incubator. The rotisserie incubator was non-heated for the first 25 days (27 °C) and was later heated to a constant temperature of 45 °C for the last 6 days of incubation to further differentiate performance. Over a total period of 31 days, samples were periodically taken out, filtered, and rinsed with DI water and assay buffer before an esterase activity assay was performed. Samples were rinsed with DI water and filtered before being returned back in a fresh dose of solvent to continue incubation.

3.4.2. Orbital Shaker Stress Test

A long-term heat and solvent stress test with orbital shaking agitation in a heated dry shaker bath was developed to evaluate the combined effects of temperature, solvent, and mild mechanical agitation on biocatalysts activity retention over a period of 30 days. Four replicate donut shaped samples were immersed together in 5 mL of aqueous 30 wt% MDEA solution (pH was adjusted to 10.5 by bubbling CO₂ gas into the solution) in a 20 mL glass vial. The vial was placed in a dry shaker bath maintained at 45 °C with 120 RPM orbital shaking. Samples were separated from the solvent periodically for esterase activity assay using the same filtering and rinsing protocols as described in Section 3.4.1.

3.5. Laboratory Gas Scrubber Test

CO₂ capture efficiencies of textile packings were evaluated in a laboratory CO₂ gas scrubber operated in single-pass flow-through absorption mode at near ambient temperature (20–23 °C). A schematic of the set-up is shown in Figure 10. Nominal 10 wt% K₂CO₃ solvent (K₂CO₃/KHCO₃ w/w ratio of 85/15, pH 10.5) delivered at 120 mL/min was used as the standard CO₂ scrubbing solvent condition. A nominal 10 vol% CO₂ gas mixture with a total flow rate of 4 L/min was selected as the standard gas test condition. The gas mixture was made by delivering 0.4 L/min CO₂ and 3.6 L/min N₂ from tanks, via gas regulators and dedicated mass flow controllers, to a gas mixing chamber and then to a gas humidifier before entering the bottom of the absorber column. These standard solvent and gas conditions were used for all CO₂ capture efficiency tests of textile packings for cross comparisons and repeated long-term performance tests, except where effects of solvent and gas flow rates were evaluated separately and varied conditions are specified.

The absorber column used for testing had an inner diameter 6 cm and a height of 30 cm. Lean fresh solvent was delivered to the top of the column and flowed downward through the packing installed in the column. Tests were conducted in counter-current mode. In counter-current mode, the pre-humidified gas mixture entered the column at the bottom of the absorber and flowed upwards through the packing installed in the column where it came in contact with solvent. The CO₂ concentration of the gas stream exiting from the top of the absorber column was analyzed by a CO₂ gas analyzer. Because CO₂ concentration at the inlet of the scrubber was controlled by a set of mass flow controllers, the inlet gas concentration to the absorber was approximated by the value of the stabilized CO₂ concentration measured at the absorber exit before turning on the scrubbing liquid flow. Thus, only one CO₂ analyzer was required at the exit to obtain CO₂ capture efficiency (or % CO₂ removal) [68]. The simplified formula [94] shown in Equation (1) was used in our study, which results in a conservative measurement of CO₂ capture efficiency.

$${\mathrm{CO}}_2 \mathrm{capture} \mathrm{efficiency} (\%)=\frac{{\mathrm{CO}}_{2 \mathrm{in}}-{\mathrm{CO}}_{2 \mathrm{out}}}{{\mathrm{CO}}_{2 \mathrm{in}}}\times 100\%,$$

(1)

The effect of L/G was evaluated at a constant nominal 10 vol% CO₂ flow rate of 4 LPM with nominal 10 wt% K₂CO₃ solvent flow rates varied among 120 mL/min, 72 mL/min, 24 mL/min, and 13 mL/min, corresponding to L/G ratios of 30, 18, 6, and 3.3, respectively. The relative effects of increasing the gas flow rate and reducing the solvent flow rate were evaluated at combinations of two liquid flow rates (120 mL/min or 72 mL/min) and two gas flow rates (4 LPM or 8 LPM) using the same standard solvent and gas compositions. The effects of solvent concentration and type were evaluated at the standard solvent flow rate of 120 mL/min and standard nominal 10% CO₂ gas flow rate of 4 LPM.

The longevity of the biocatalytic textile packing was evaluated over a period of 71 days, including ten repeated test-wash-dry-storage cycles followed by a final 100 h soaking in 10 wt% K₂CO₃ (pH~10.50) scrubbing solvent at an elevated temperature of 45 °C. Dry storage stability was evaluated by retesting the same packing after 1-year storage at ambient conditions. Additional longevity testing was carried out in a continuous 500 h solvent recirculation experiment. Room temperature 10 wt% K₂CO₃ flowed through the packing for the entire duration of the experiment. The standard nominal 10 vol% CO₂ gas mixture was delivered to the absorber column periodically to obtain CO₂ capture efficiency measurements at different time intervals. Fresh lean standard nominal 10 wt% K₂CO₃ solvent was used in a single-pass flow-through mode during these measurements while the recirculation was stopped.

The effects of CO₂ concentration to simulate applications other than a coal-fired power plant flue gas (nominal 10 vol% CO₂) were evaluated using nominal 5 vol% CO₂, 25 vol% CO₂, and ambient air (CO₂ conc. 600–800 ppm) at standard 120 mL/min solvent flow rate. Standard nominal 10 wt% K₂CO₃ solvent was used for the 5 vol% CO₂ and 25 vol% CO₂ conditions, while simulated seawater with a solute concentration of 3.5 wt% (pH was adjusted to 10.0 using 1N NaOH) was used for the ambient air condition at an air flow rate of 1.5 LPM supplied by a laboratory compressed air system.

4. Conclusions

Herein we reported our proof-of-concept investigations on the fabrication and testing of novel, durable and versatile “drop-in-ready” textile structured packings with immobilized carbonic anhydrase (CA) fortified by covalent linkages. The resulting biocatalytic textiles with CA immobilized on textile support materials by several different enzyme immobilization strategies were characterized by FTIR, SEM, MMT, enzyme activity assay and by CO₂ absorption in a laboratory gas scrubber.

Results from FTIR, SEM and esterase activity analyses confirmed the presence of enzyme protein and chemical crosslinks at fiber surfaces at levels that depended on and were consistent with the different immobilization strategies. The presence of covalent crosslinks between CA and chitosan-coated cotton fiber support materials significantly enhanced textile durability to mechanical stress and preserved enzyme activity on the fabric supports for long periods of time, retaining 49% to 59% of their initial activities while that of the non-crosslinked sample diminished to only 2%. Additionally, in a 30-day heat and solvent stress test in 30 wt% MDEA at 45 °C, final activity retention in the range of 42% to 82% was achieved for all samples involving crosslinker. Experimental control over the placement, and hence the location, of enzymes in the immobilization matrix, i.e., on the surface vs. buried in the chitosan matrix, together with enzyme activity retention results, suggested a surface erosion deactivation mechanism in which entrapped enzymes are more protected than those exposed at the surface, even those attached to the surface with covalent crosslinks.

In the lab-scale scrubber test, surface-only covalent mono-layer packing with CA achieved the highest capture efficiency of 71.2%, and surface-only 3-D aggregate packing achieved a similarly high performance. This corresponds to about a 3-fold enhancement from no-enzyme textile control packing and a 20-fold enhancement compared to conventional Raschig ring random packings, emphasizing the importance of “exposed” enzyme placement at the textile packing surface, where enzyme can readily contact both gas and liquid. Post-entrapment crosslinking of enzyme entrapped in chitosan coating had lower but still substantial CO₂ capture efficiency, indicating surface “exposure” of at least some of the entrapped enzymes. These insights provide strategies for further packing optimization to achieve a balance of extended durability and high CO₂ absorption performance.

Surface-only covalent 3-D aggregate packing with an excellent initial CO₂ capture efficiency of 66.7% was selected for parametric and longevity testing. After 10 repeated test, rinse, and dry cycles over a period of 71 days with the last 100 h immersed in 45 °C solvent, this single packing module retained 100% of its initial CO₂ capture performance. Furthermore, after 1 year of dry storage at lab ambient conditions, over 85% of its original capture efficiency was retained. The same packing also demonstrated stable performance during a 500 h continuous liquid flow scrubber test. One explanation for the extended longevity of this packing module could be that enzyme losses by an erosion mechanism cause underlying enzymes to become exposed such that the active enzyme is constantly refreshed over time.

Covalently immobilized CA on textile packing was resistant to stressors such as tap water wash, elevated temperature, solvent immersions, and continuous solvent flow, and was stable for long-term dry storage at ambient conditions. The versatility of enzyme immobilized textile structured packings was demonstrated by their excellent capture efficiencies using diverse and desirable CO₂ scrubbing solvents, including K₂CO₃, MDEA and DMG, as well as their applicability across wide CO₂ concentration ranges relevant for CO₂ capture in coal and natural gas-fired power plant flue gas, for natural gas and biogas upgrading, or for CO₂ removal directly from air. Remarkably, a CO₂ capture efficiency of 66.9% was obtained from ambient air (600–800 ppm) using simulated seawater.

MMT measurements showed that liquid transport properties of pristine cotton fabric was modified by the presence of chitosan coating, which decreased wicking rate, and enzyme, which increased wicking rate, in one case even surpassing the pristine cotton performance. The hydrophilic properties of these textile materials were responsible for excellent solvent distribution throughout the assembled packing modules leading to enhanced CO₂ absorption even for the control no-enzyme textile modules, compared to Raschig rings. The intrinsic high level of control over the liquid flow afforded by textile-based packing materials is a special advantage that could enable new modular packing designs suitable for unstable environments, like on ships or deep sea platforms.

The present findings with our novel biocatalytic textile structured packings that were able to operate flooding-free in counter-current mode and performed efficiently at relatively low L/G ratios warrant scale-up studies in larger scale CO₂ capture systems. Furthermore, future investigations will explore improvements in the immobilization approach, and will clarify the extent to which surface erosion or enzyme inactivation impact biocatalyst longevity. Improvements in packing module CO₂ absorption performance can be anticipated by mitigating viscosity and surface tension effects of the solvent, optimizing solvent components and concentrations, optimizing liquid distribution with novel textile materials and structural designs, employing alternative absorber configurations, adapting enzyme surfaces to improve immobilization efficiency, improving enzyme robustness—to solvents, temperature and other process variables—and by surface chemistry modifications that optimize molecular mass transfer interactions at gas-liquid-enzyme interfaces. Combined with fluid dynamics simulations and theoretical modelling, these future developments are expected to lead to even more efficient packing and process designs for CO₂ capture to help address the CO₂ mitigation challenge.

5. Patents

A patent application on the technology reported in this manuscript has been filed at the USPTO.