Tailoring 3D Printed Micro‐Structured Carbons for Adsorption

The manufacture of tailored carbon‐based adsorbent structures with exceptionally low‐pressure drops and improved kinetics using stereolithographic 3D printing is presented. Adsorbent structures are printed from commercial resins with square, circular, and hexagonal cross‐sectional microchannels. These structures can reduce energy use by 50–95% compared to conventional carbon‐packed beds. The activated 3D printed carbon achieves Brunauer–Emmett–Teller surface areas over 1000 m2 g−1 and shows outstanding butane adsorption capacities, over twice the capacity of a commercial carbon and a comparable capacity to phenolic‐based carbons. The structures also show excellent uptakes of cyclohexane, up to 0.62 g g−1 in a saturated feed. The introduction of complex axial geometries including spirals and chevrons enable superior adsorption kinetics and premature breakthrough of contaminants at high gas flow rates. These results demonstrate the success of intelligent manufacturing of low‐pressure drop, high‐capacity micro‐structured adsorbents, allowing for the development of gas separation technologies for applications such as greenhouse gas removal and respiratory protection.

removal and capture of harmful noxious gases including CO 2 , CH 4 , NO x and SO x . [1,2] Porous materials such as activated carbons (ACs), zeolites, mesoporous silica, and lately metal-organic frameworks (MOFs) are often employed as adsorbents due to their high adsorption capacities, cost-effectiveness, and large surface areas. An adsorbent's selectivity for a given gas mixture can be maximized by tuning the pore geometry and properties. [3,4] ACs have attracted significant interest in gas separation with their highly tunable nature in addition to their high surface areas, low costs, and high thermal/ chemical stabilities. ACs are typically used in a packed bed form wherein small carbon granules (typically 0.8-3 mm in size) are evenly distributed within a containing module, allowing for contaminated gas flow through encouraging adsorption. [5] Although granular packed beds are widely used in industry, they have certain inherent disadvantages, including highpressure drops, channelling, and low electrical and thermal conductivities. Alternative forms for AC materials include structures such as extruded monoliths, expanded foams or woven fibers. [6][7][8] These three examples are macrostructures rather than granular carbon, allowing for a potentially more efficient pressure drop to mass transfer trade-offs. The use of straight, uniform channels allows for notably low-pressure drops and the channel size can be finely controlled to optimize mass transfer. Simple extruded monolith structures alleviate some of the issues faced by packed bed systems and have been successfully used in small-scale trials as adsorbents for removing contaminants from gas streams or as supports for catalysts. [9] Production of these monoliths requires a unique die design, high extrusion pressure and careful post-treatment, carbonization, and activation. The cell density generation, channel shape, and channel configurations during manufacture are heavily dependent on the die. The manufacturing constraints for die-extruded structures mean that primarily square, circular, or hexagonal straight channel monoliths are the more common options, and the cost and precision required to reduce the channel width further limit the size and density of these channels. [10,11] Using 3D printing techniques, it is possible to generate more complex structures such as chevron or spiral channels or to generate bespoke porosity within the walls. This means 3D printing offers significant flexibility of design. While the speed of traditional extrusion techniques is significantly higher than current

Introduction
Adsorption-based gas separation has been considered an energy-efficient alternative technology to conventional processes such as absorption and cryogenic distillation for the 3D printing speeds, the rapid development of the technology may provide cost-effective and viable ways of manufacturing complex microchannel adsorbent structures in the future.
3D printing is a rapidly evolving manufacturing technology with the potential of creating almost any geometrically complex shape or feature from a range of materials across different scales. [12,13] 3D printing technology has demonstrated its usefulness in a variety of fields, such as artificial organs, [14] biomedical applications, [15] water treatment, [16] microwave absorption, [17] construction, [18] etc., because it can provide faster, easier, and normally more economical solutions, as well as the ability to create complex geometrical structures with increased controllability targeting end-user requirements and optimized macrostructures. Furthermore, 3D printing normally uses the amount of material required and no additives (such as plasticizers and binders in extrusion) are needed to enable production. For monolith extrusion, the milling and paste preparation steps require specific particle sizes in order to control the porosity, so some off-sizes are wasted. Overall, this may mean that 3D printing has the potential to reduce the waste produced. There is a potential to revolutionize structured adsorbent manufacturing by 3D printing, imparting uniformity to the pore size distribution and geometric configuration of porous structures, at least to the micron scale. [19] While 3D printing technology has reached the nano-scale production level using techniques such as two-photon polymerization and electrohydrodynamic printing, further developments are required for nano-scale precision and cost-effectiveness. [20] Nevertheless, 3D printing technologies are expected to be the leading techniques for futuristic nanoscale fabrication. [21] Several 3D printing techniques identified with the potential to generate adsorbent structures include fused deposition modelling, stereolithography (SLA) and selective laser sintering (SLS) techniques. SLA printing involves the photocuring of a liquid resin. A point laser is used to specifically cure a layer of resin in the desired pattern, although variations such as direct light processing (DLP) use a projected image rather than a scanning laser. The cured layer attaches to the build platform which is moved up out of the resin and detaches from the bottom of the resin vat. The build platform then moves back down to a layer thickness above the bottom of the resin vat and the next layer is cured. SLS uses the thermal fusion of a fine powder in a bed via a scanning laser. The powder bed is fused in the desired pattern and a new layer of powder is deposited on top to continue the 3D printing process.
There are several options for producing structured adsorbents from 3D printing technologies, including carbons, zeolites, and MOFs, [22] using a range of printing techniques as reviewed by Blyweert (2021). [23] Earlier work in 3D printing adsorbent materials was reported in papers such as Thakkar (2016), Couck (2017) and Regufe (2019), using the direct ink writing technique to extrude structures from initially zeolite pastes and later MOFs and carbonaceous materials. [24][25][26] As outlined previously, SLA-printed activated carbon microstructures will be the focus of this work. SLA technologies are the only 3D printing methods which are established to directly form thermosets and so are promising for printing activated carbon materials. However, SLA printing of activated carbons has only tentatively been explored, with most initial work focused on resin development. [23,27] Most notably, Steldinger (2019) developed a monomer mix of a low carbon yield, high cure time acrylate with a high yield, slow curing aromatic which showed a synergistic curing behavior. [28] However, no known work has been carried out in printing microchannel structures with controllable channel geometries and properties or in developing the adsorption capacity and kinetics of light-cured printed carbons, which will be the focus of this work.
Herein, we report the use of the DLP technique (a subset of SLA) to fabricate AC-based adsorbent microchannel structures which have superior adsorption uptakes compared to commercial AC-packed beds (Figure 1). The structures are formed from photocured, commercial resins purchased from Kudo3D Inc. (USA). Three cross-sectional geometries (square, circular, hexagonal) and three axial geometries (spiral, chevron and stepped) were examined for their butane and cyclohexane uptake as volatile organic carbon (VOC) analogues. The potential of printing controlled macroporosity in the walls was also examined by using repeating unit cells to allow for optimal mass transfer into the carbon. The nominal pore size was limited to 250 µm spherical units in this work. The green polymer microchannel structures were pyrolyzed to form activated carbon structures achieving Brunauer-Emmett-Teller (BET) surface areas of  The microchannels showed much longer butane breakthrough times versus 1 mm commercial packed beds and impressive isotherm uptakes of cyclohexane of over 0.60 wt wt −1 .

Microchannel Design
As an initial step to 3D printing, a CAD model of the microchannel structure was designed using the Autodesk Inventor software. Channel dimensions were chosen to maintain the same hydraulic diameter and to be similar sizes to prior extruded work with channel width-to-wall thickness ratios (1.5:1) chosen based on Barnard et al. (2022) and to give similar free volumes to granular carbons ≈35%. [29] The resulting CAD models are presented in Figure 2. Dimensions of all structures explored in this work can be found in (Table S1, Supporting Information).
The flexibility of 3D printed design allows for almost any geometrical shape to be explored as an alternative channel crosssection. Three options were considered in this work to examine the impact of channel geometry on flow dynamics and adsorp-tion performance: square, circular, and hexagonal (Figure 2a-c). The hydraulic diameter, channel distribution and channel density were kept consistent for each design, with the voidage, thickness and shape of the inter-channel walls changing between designs. Due to maintaining the distribution of channels, the circular and hexagonal shapes are not tessellated and so will have less consistent wall thicknesses than the square pattern.
In addition to cross-sectional geometries, several unique axial channel geometries were chosen to examine: a helical pattern, a repeating chevron pattern and a stepped pattern consisting of regular cuts into the channel wall geometry, shown in Figure 2d-f. These axial designs were chosen to induce flow circulations and mixing along the channel and consequently improve the breakthrough profile. The designs are based on some common micromixer designs with high mixing efficiencies (zigzagged, convergent-divergent walls and 3D spirals). [30][31][32] The spiral and chevron patterns have the additional benefit of increasing the effective path length of the microchannel, at the expense of increased pressure drop and reduced bulk carbon in the walls. The shapes of the spiral and chevron patterns were chosen to give an average path length of twice the base microchannel length and the number of rotations of the spiral pattern tailored so that the channel length was kept consistent as the helical diameter changed. In addition to varying axial and cross-sectional channel designs, the potential for adding structured macroporosity into the channel walls was also investigated. By using repeating unit cells of varying sizes to build up the channel walls, consistent macropores could be introduced into the walls to promote diffusion from the gas phase into the bulk carbon, as shown in Figure 2g-i. The size and shape of these could be varied. For this work, a spherical-based unit cell was used, but alternative shapes such as octahedra could also work. The nominal pore size is limited by the resolution of the 3D printer to ensure each unit cell can be printed accurately. A lower limit of ten times the largest resolution is typically suitable for ensuring the accuracy of printing, which equates to 250 µm for the printer used in this work (with an XY resolution of 25 µm). These cells leave pores in the intercellular space (shown in Figure 2h) and ensure interconnectivity of the porosity and accessibility of the bulk carbon in the microchannel walls.

Carbonization, Activation and Oxidative Stabilization of 3D Printed Structures
A two-step pyrolysis process of carbonization (4 h, 500 °C) and activation (15 min, 950 °C) converts the 3D printed polymer structures into an adsorptive carbon. Carbonization of four acrylate-based commercial resins was examined: Kudo3D "UHR", "engineering hard", "engineering tough" and "cast" resins. Each resin was examined for print precision and carbonization yield -expressed as the weight percentage retained after the carbonization procedure. The "UHR" resin showed the highest carbon yield (12.9%), while the "cast" resin completely disintegrated upon pyrolysis. The "engineering hard" and "tough" resins showed comparable, but lower carbon yields to the "UHR" resin (11.2% and 10.3% respectively) and had a lower print precision. As such, the "UHR" resin was chosen for this work.
Thermogravimetric analysis (TGA) results showed a 91.1% sample mass loss after heating to 700 °C, matching the 87-91% overall mass loss typically observed during the carbonization of UHR samples ( Figure S1, Supporting Information). Low carbon yields are typical of acrylate-based polymers, which undergo de-carboxylation reactions and fragmentation above 430 °C. [33] In addition to high mass losses, structural shrinkages of 40-50% were observed in each dimension. This led to a loss of precision in the channels and surface cracking, particularly parallel to the printed plane. 3D printed models were scaled up by 65% to ensure features were the correct size after pyrolysis (shown in Table S1, Supporting Information).
A process of oxidative stabilization was introduced to improve the potential carbon yield and the integrity of the carbon structure. Oxidative stabilization involves the uptake of atmospheric oxygen by replacing hydrogen in the structure at temperatures from 200 to 300 °C. This process can be used when activating carbons to reduce the plasticity of the polymer at high temperatures and remove volatiles which may swell the material. By forming additional oxidative bonds in the structure, a reduction in fragmentation and loss of carbon during carbonization can often be observed. [34] Oxidative stabilization was found to increase the carbon yield of the 3D printed microchannels from 8-12% to 15-20% and reduced volumetric shrinkage from 40-50% to 30-35% in each dimension. This improved the integrity and precision of the activated microchannels and reduced structural cracking and channel blockage. A comparison of magnified channels can be found in ( Figure S2b vs Figure S2c, Supporting Information), with fewer, smaller channel cracks seen in the stabilized microchannels.
As a result of oxidative stabilization, the green and activated microchannel structures showed a high adherence to the CAD model (Figure 3a-c), although deviations can be seen after activation, particularly in the square design, where fine corner elements are partially distorted. However, the cross-sectional designs still remain significantly different after activation. The green printed microchannels (shown in the middle column of Figure 3a-c,g) have smooth surfaces and channels, which become rougher after pyrolysis (shown on the right column), due to cracking and increased porosity of the carbon at the surface. This should increase the surface micromixing and pressure drop of the channels, as reported by Croce (2005). [35] The accuracy of the axial geometries can be seen in Figure 3d-f, showing channel precision is reduced during the carbonization and activation steps. The patterning is still clear, but some channels are partially blocked or have narrowed. Noticeably, the stepped microchannels showed a lack of precision in the square in-cuts. Figure 3g shows the process of activation of the macroporous channel design. Porosity in the channel wall was maintained through printing and activation, although some pores were partially or fully closed at the green stage, due to exposure bleeding from the projector. Again, deformation of the corner elements was observed after activation. This process requires additional work and so the macroporous design was not examined for gas adsorption.
Apparent BET surface areas, pore volumes and pore sizes were determined for the samples using nitrogen adsorption isotherms measured at 77 K (found in Figure S3, Supporting Information). The apparent BET surface area increased from 145 m 2 g −1 when carbonized, to 882 m 2 g −1 for the activated carbon, and 1050 m 2 g −1 for the oxidatively stabilized carbon. The significant increase in apparent surface area may be due to the decreased plasticity of the carbon, reducing the collapse of the microstructure during pyrolysis, which agrees with the increased pore volume from 0.368 to 0.459 cm 3 g −1 observed after stabilization. Pore size distributions of the two activated samples were calculated using a quenched solid density functional theory (QSDFT) method [36] for slit and cylindrical-shaped carbon pores (found in Figure S4, Supporting Information) and the mean pore size was calculated as 1.3-1.4 nm for the carbons. Mercury intrusion porosimetry was used to investigate the meso-and macroporous contributions to the pore volume. In the measured mesoporous range (5-50 nm), an average pore volume of 0.054 cm 3 g −1 was found, suggesting the carbons have a low percentage of mesopores above 5 nm. The total porosity for the samples (from 5 nm to 346 µm) was found to be 1.12 cm 3 g −1 . Pore volume distributions from the mercury intrusion porosimetry can be found in ( Figure S5, Supporting Information).

Adsorption Results
The adsorption of the square-activated microstructure was compared to two baselines: an extruded phenolic-based activated carbon monolith and a 1 mm commercial granular carbon. 1000 ppm butane at a flow rate of 1 Nl min −1 was used as an analogue for other VOCs as it is non-toxic and non-polar. The oxidatively stabilized square channel model was used as the initial comparison to the baseline packed bed and extruded monolith, shown in Figure 4a. This allows the 3D printed material to be examined for its capacity as an adsorbent carbon compared to baseline carbon materials. The carbon in the phenolic-based  monolith and the square microchannels were seen to perform better than granular carbon in terms of capacity and breakthrough time (with the breakthrough point taken as 5% of initial butane concentration). The square microchannels showed a butane mass uptake of 0.095 g g −1 and breakthrough time of 22.4 min g −1 compared to 0.044 and 10.8 g g −1 respectively in the granular carbon. Breakthrough time is normalized per gram of carbon to allow for a direct comparison between different samples. The oxidatively stabilized printed carbon showed a comparable if slightly lower gravimetric capacity than the phenolic-based carbon (0.108 g g −1 ). The 3D printed square microchannels also had a shallower breakthrough curve than the phenolic monolith (with a mass transfer length of 44.7 vs 36.9 min) leading to a shorter breakthrough time (22.4 min g −1 compared to 27.9 min g −1 ). As the microchannels are slightly dense than the phenolic-based carbon but slightly less than the packed bed, they have a very similar performance when examined for volumetric uptake (0.0401 g cm −3 ) compared to the phenolic-based carbon (0.0408 g cm −3 ) and are still significantly higher capacity compared to the commercially packed bed carbon (0.0258 g cm −3 ). This suggests that the resin-based 3D-printed carbon is a suitable material for VOC adsorption. The printed carbons with different cross-sectional geometries were directly compared for their adsorption capability to determine the optimal geometry for adsorption. The square microchannels examined in Figure 4a showed a superior mass uptake of butane and a significantly longer normalized breakthrough time compared to the hexagonal microchannels (0.070 g g −1 and 11.6 min g −1 respectively) and circular microchannels (0.084 and 17.8 g g −1 respectively) (Figure 4b).
It is suggested that the lower adsorption performance of these two geometries is due to the channel distribution (shown in Figure 2) which gives inconsistent wall thicknesses for the circular and hexagonal microchannels, most prominently in the hexagonal design. This would reduce the accessibility of the wall carbon, reducing the rate of mass transfer and overall uptake. The printed carbons with different axial geometries were compared to determine whether the increased turbulence and path length had an impact on the adsorption time and breakthrough profile (Figure 4c). The spiral structure shows a shorter mass transfer zone than the baseline printed square channel structures (29.2 vs 44.7 min), suggesting the spiral design has improved the mass transfer into the carbon. However, the overall uptake of both the spiral and chevron is significantly lower than the square and stepped patterns. This may be due to a less even activation along the microchannels. The stepped structure showed a similar overall uptake and profile as the square design (0.100 vs 0.095 g g −1 ) suggesting the steps had little influence on the performance. Breakthrough runs were repeated after regeneration three times per structure, with weight uptakes having standard deviations of 4.5%, and breakthrough times having standard deviations of 2.2%.
To assess the impact of the increased channel complexity more closely, the start of the breakthrough curves at higher flow rates (2 Nl min −1 ) was examined in Figure 4d. As can be seen, the channels with the most significant channel complexity and longest channel lengths, that is, chevron and spiral patterns, had the lowest levels of instantaneous breakthrough when exposed to the higher flow rate of butane -particularly the spiral pattern. The external and internal mass transfer kinetics in monoliths are known to be strongly dependent on the channel shape and size (confirmed by the differences in Figure 4b). [37] As such, having full control over these parameters is important to prevent instantaneous breakthroughs often seen in monoliths and shown by several designs in Figure 4d. By introducing the axial changes along the design, it is clearly possible to improve the external mass transfer into the carbon walls and significantly reduce the chance of instantaneous contaminant breakthrough for a given gas flow rate, which can be an issue due to highvelocity laminar flow rates in monoliths. [38,39] The isothermal uptake of cyclohexane was examined to determine the maximum uptake capacity in saturated contaminant feeds, when allowed to reach equilibrium and as cyclohexane is the main test gas in VOC standards. [40] Extremely high cyclohexane capacities can be observed from the oxidatively stabilized 3D printed carbon, with equilibrium loadings of 62% by weight shown in Figure 4e. This is significantly above the extruded monolith baseline (32%) and a baseline packed bed carbon (28%) and much higher than the non-stabilized carbon (13%). Such high loadings suggest that the 3D-printed carbon has a very high accessible pore volume with an affinity for organic vapors. This would make the microchannels highly effective at high contaminant concentrations.

Pressure Drop
All microchannel designs tested for a breakthrough were also examined to determine the pressure drops through the designs. These were compared to a baseline commercial packed bed with a 0.8 mm particle size, along with a baseline extruded monolith. The pressure drops were examined from 0-2 Nl min −1 , comparable to the flow rates used in breakthrough testing.
The three cross-sectional channel geometries (square, circular and hexagonal) have similar pressure drops, which are less than 6% of the pressure drop of the packed bed. The square microchannel structure had the lowest pressure drop, 23% lower than the circular and hexagonal designs, and was very comparable to the extruded monolith.
By adding axial features to the channels, channel flow disturbance is increased. In the stepped channel design, the tortuosity of the channel is not changed, only a slight increase in the channel complexity, which can be seen by the increase in pressure drop over the cross-sectional geometries (which have ≈40% lower pressure drops). Adding additional channel tortuosity and effective path length of the channel can prevent premature breakthroughs in the monolith but at the cost of increasing the pressure drop. This can be seen clearly in Figure 5, where the spiral and chevron axial designs have noticeably higher pressure drops -7 and 12 times higher than the crosssectional geometries respectively but still significantly below the commercially packed bed baseline -50% and 72% lower for the chevron and the spiral channels respectively.

Conclusion
This work demonstrates that suitable adsorbent structures can be successfully produced via a new, smarter 3D-printed route from commercial resins. A wide range of channel geometries was proven to be printable and could be successfully activated into highly adsorbent, low-pressure drop systems with good adherence to the original design. When compared to the current state-of-the-art, packed bed systems, pressure drops were found to be 50-95% lower, and contaminant breakthrough times were found to be over twice longer. The three cross-sectional microchannel designs (square, circular and hexagonal) showed pressure drops comparable to conventionally extruded monoliths, whilst axial designs (spiral, chevron and stepped) showed higher pressure drops, but significantly improved mass transfer rates, preventing premature breakthrough at higher flow rates. This shows that customized structures can be designed and optimized for the application by choosing specific channel designs which balance the pressure drop, kinetics and capacity as desired. Exceptional isothermal cyclohexane uptakes were also displayed by the 3D-printed carbon, with adsorption capacities of up to 62% w/w%. These results demonstrate the success of applying 3D printing techniques to adsorption applications allowing for unique results which outperform existing start of the art technologies. These structures could be applied in safety-critical applications where complete contaminant removal is crucial or in high gas through-put applications, allowing for significant energy savings.

Experimental Section
Production of CAD Model: All the models designed were generated using Autodesk Inventor 2021/2. The designs were exported to an STL file format to be imported into a slicing software (Kudo3D print software). Supports were added at this stage, attaching the structure to the build platform at a 30° angle. The model was converted to a sliced file at a 25 µm thickness and sent directly to the printer.
Printing of Microchannels: A Titan 2 HR SLA printer from Kudo3D Inc. (USA) was used for printing all microchannel designs. The printing was setup as typical with a 38 µm XY resolution. The printing parameters used are shown in the table below ( Table 1). Four commercial resins were examined: UHR, engineering hard, engineering tough and cast. All four resins were used as received from Kudo3D Inc. (USA).
Pyrolysis of 3D Prints: After printing, the microstructures were postprocessed, by washing in water/ethanol and allowed to cure in sunlight. Where relevant, an initial oxidative stabilization step was carried out at 300 °C in the air for 4 h. All monoliths were consequently carbonized in nitrogen at 500 °C, with a heating ramp of 1 °C min −1 up to 350 °C, then a 3 h hold at 350 °C followed by a heating ramp of 0.5 °C min −1 up to 500 °C. Once carbonized, the monoliths were heated to 950 °C with a 1 °C min −1 heating ramp in a pure CO 2 atmosphere and held at the temperature for 15 min, before reducing the temperature back down to room temperature.
Characterization of Activated Microchannels: TGA was carried out using a Setaram Setsys Evolution TGA 16/18. The mass loss of the polymer samples was analyzed over the range of 30-700 °C. Nitrogen isotherms at 77 K were carried out using a Quantachrome Autosorb-iQ to determine the surface area and pore size of the carbon samples. The QSDFT method for slit/cylindrical shaped carbon pores was used to analyze the pore size and the BET method to determine the surface area. SEM images were taken using a JEOL JSM-7900F FESEM. The samples were cut to size and coated with gold via sputter deposition before imaging.
The activated carbon samples were tested for adsorption against a challenge of 1000 ppm butane in nitrogen. Samples were analyzed to produce a breakthrough curve. The compressed butane gas was passed through a mass flow controller set at 1 or 2 Nl min −1 then through the sample and into a flame ionization detector (FID). Hydrogen fuel gas and nitrogen carrier gas were also connected to the FID and left at set pressures (3.5 and 9 psi respectively). The FID (a Teledyne 4020 Total Hydrocarbon Content Detector) measured the concentration of butane in the outlet. This can be seen in Figure 6. A nitrogen generator allowed for cleaning and purging of the line to the FID during start-up and shutdown. A bypass loop around the sample then allowed for a baseline butane concentration to be established, before switching to   sample analysis. Regeneration of samples was carried out in an inert nitrogen atmosphere at 350 °C for at least 4 h. Butane breakthrough was set at 10 ppm (or 0.1%) to match standards for VOCs. Mass transfer zone lengths were calculated at 5% and 80% of the initial butane concentration to capture the main part of the curve.
Cyclohexane isotherms were measured using a Dynamic Vapor Sorption instrument (DVS Advantage 2, Surface Measurement Systems Ltd, UK). Samples of carbon, typically 20-50 mg, were loaded onto a sample pan suspended from a microbalance. The carbons were heated to 120 °C for 3 h to remove adsorbed moisture. Cyclohexane isotherms were run at 10 °C, 25 and 40 °C (each on a separate sample). At each pre-programmed relative pressure, equilibrium was defined as a mass change of less than 0.001% of the dried sample mass per minute, maintaining that rate of change for 10 min before moving to the next programmed setting Pressure drop measurements were recorded using a digital flowmeter passing air through the sample from 0-2 Nl min −1 with the air pressure measured directly before and after the monolith using a digital manometer (Figure 7).
Preparation of Baseline Samples: An extruded phenolic-based monolith and an ABEK1 granulated carbon were used as baseline samples. The monoliths were received pre-carbonized from an industrial partner and activated to the same degree as the 3D-printed samples. The ABEK1 carbon was purchased from Honeywell and was used as received.
Mercury intrusion porosimetry was carried out using a Micromeretics Autopore IV instrument, on a 200 mg activated carbon sample from 0 to 33 000 psi.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.