The Effect of Sweeping Media and Temperature on Aqueous CO 2 Removal Using Hollow Fiber Membrane Contactor (HFMC): An Experimental Determination

Transport phenomena through hollow fber membrane contactors (HFMCs) indicate the exchange of a component between the two phases, inside and outside of hollow fbers. In this research, we designed and fabricated lab-made HFMCs to assess the diference between water and air as sweeping media for CO 2 exchange. Te efects of fow rates and temperature ratios on aqueous CO 2 absorption were investigated accordingly. A semiclosed circuit incorporating our fabricated HFMCs was set up to regulate the operating parameters and evaluate the aqueous CO 2 concentration using an initiative pH-based method. Te results of our experiments remarkably reveal that air tends to remove aqueous CO 2 more than water when aqueous CO 2 concentration is higher than 3.53 × 10 − 6 mlCO 2 /l. However, water would surpass air in lower concentrations. Nevertheless, tripling the fow rate of sweeping media from 500 to 1500ml/min shifts up this cutof point 50 times to around 1.66 × 10 − 4 mlCO 2 /l. Te experiments performed at three diferent temperature ratios of 22 :22, 44:12, and 22 :12 ° C (CO 2 -rich liquid: sweeping medium) demonstrated that a higher temperature gradient deteriorates the CO 2 absorption capacity of sweeping media. Nonetheless, temperature gradient becomes highly efective in aqueous CO 2 concentrations lower than 1.57 × 10 − 6 CO 2 /l. Te results of this research could be applied in performance optimization of aqueous CO 2 absorbing HFMCs, even in sophisticated medical procedures such as arterio-venous and veno-venous CO 2 removal systems where both water and air could be used as blood’s CO 2 sweeping media.

Various parameters afect the efciency of the CO 2 removal process, such as module confguration, hollow fber arrangement, type of sweeping medium, temperature, and fuid dynamics such as the fow rate, the temperature of fuids, the fow pattern, and the CO 2 concentration in both media (i.e., absorbing/sweeper and removing/carrier media) [35][36][37]. However, no specifc publication addresses the efects of the parameters mentioned previously on aqueous CO 2 removal.
In this article, the afnity of water to capture aqueous CO 2 was investigated and compared with air through identical setup designs. In addition, the efect of fuid dynamics and the temperature of the sweeper and carrier media were experimentally examined. Tis research uses the HFMC as an aqueous CO 2 removal device. Te CO 2 -containing liquid fows outside the hollow fber bundle, and the sweeping media of air or water absorbs CO 2 by fowing inside the lumens. Te results of this research bring light on the diference between water and air as a sweeping medium with respect to their ability to remove aqueous CO 2 . Tis could eventually be used in design modifcations of HFMCs in AVCO2R and VVCO2R devices.

Equipment and Materials.
In order to investigate the efect of sweeping media and temperature on the removal of the aqueous CO 2 , diferent experimental circuits had been designed and set up. Te experiments were based on the principle of CO 2 dissociation in water and the resulting aqueous pH drop. Te following equipment and materials were utilized in experimental setups to measure the pH change as a consequence of CO 2 exchange in testing HFMCs: Stöckert SIII including 2 roller pumps (Stöckert Instrumente GmbH, Germany), PT-100 temperature sensors connected to the Stöckert temperature measuring system (Stöckert Instrumente GmbH, Germany), liquid fow meter T110 ultrasonic sensor (Transonic System INC.,USA), CO 2 and compressed air (Linde AG, Germany), gas fow meters (Brooks Instrument Inc., USA), PVC and silicon tubes and connectors (Raumedic AG, Germany), Capiox ® bubble trap (Terumo Cardiovascular Group, USA), and magnet stirrer C-Mag HS7 (IKA-Werke GmbH&CO. KG, Germany). To set the intended temperatures, Omnitherm ™ heat exchangers (SciMED Life Systems Inc., USA) were employed in connection with the Stöckert cooling system (Stöckert Instrumente GmbH, Germany) for temperature reduction and HAAK water bath (Termo Fisher Scientifc, USA) for temperature increase with regard to the ambient condition. Distilled and deionized ultrapure water were prepared by employing Milli-Q Advantage A10 ultrapure water purifcation system (EMD Millipore Corp, Billerica, USA). For pH measurement, FPH-BTA Tris Compatible Flat pH sensors in combination with the compatible data acquisition board LabQuest Vernier and Logger Lite ® v.1.6.1 software (Vernier Inc., USA) were used to acquire pH data and send it to a pc by a USB port. Te acquisition frequency for pH sensor was set to 2 Hz (i.e., 2 data/s).

Fabrication of HFMCs.
A cylindrical hollow-fber membrane contactor was designed and fabricated in our laboratory, as indicated in our previous publication [38], with some modifcations. Mats of Oxyphan ® hollow fber membranes (Membrana AG, Germany) with custom-made length were ordered. Tese mats were arranged in the intermediate space of two concentric cylinders made of polycarbonate polymer (refer to Figure 1(a)). Te initial designs were performed with Autodesk Inventor ® 2012 (Autodesk Inc., USA) software, which can be used by milling machines for prototypes' construction. Ten, the designs were imported to SolidWorks ® v.2012 X64 (Dassault Systèmes SolidWorks Corp., France) software to investigate the geometric parameters of module designs in more detail. Using SolidWorks ® , the illustration of the whole design, cross sections, and fuid felds were also tested. Te basic 3D design of this HFMC created with SolidWorks ® is depicted in Figures 1(b) and 1(c).
Te hollow fbers in prototypes were then potted by polyurethane glue (Henkel AG & Co. KGaA, Germany), using a custom-built centrifuge device (Wichelhaus GmbH & Co. KG Maschinenfabrik, Germany) as shown in Figure 2 Te efective membrane surface area of this HFMC is 0.8 m 2 , while the priming volume is 0.7 × 10 −4 m 3 . Tis hollow-fber membrane contactor could be considered as either liquid-liquid or gas-liquid HFMC, provided that water or air are used as the medium fowing inside hollow fber lumens, respectively.
While fabricating the HFMC, some challenges and obstacles need to be resolved such as winding the Oxyphan ® mat homogenously around the inner shaft of the contactor. To overcome this obstacle, specifc pinches and handles were designed and constructed to properly mount the fber mat on the special lab-made winding machine. Tis machine induces a constant tension force over the fber mats; therefore, they can be winded uniformly.
Other than that, the viscosity of the polyurethane glue for potting the hollow fbers via centrifugal force plays an important role in the uniform total sealing of the hollow fbers. Te optimum viscosity of the glue was achieved by experience. After potting, both ends of the hollow fbers should be cut in a way that the ends of the fbers remain open which can be challenging and needs experience; hence, the hollow fbers were cut using a very sharp stainlesssteel blade. Figure 3, the setup was designed to measure and compare aqueous CO 2 removal at diferent temperatures using deionized (DI) water as a sweeping medium. As shown in Figure 3, on the one hand, liquid1, which is CO 2 -rich DI water, fows outside the hollow fbers of testing HFMC in a closed loop using the roller-pump1. On the other hand, liquid2, which is CO 2 -free DI water, streams counter-currently inside the hollow fber membrane lumens, in an open circuit using the roller-pump2. In this way, liquid2 removes the CO 2 content of liquid1 gradually resulting in the increase of liquid1's pH.

Liquid-Liquid HFMC. As demonstrated in
Before initiation of the experiment, in order to provide CO 2 -rich DI water, air bubbles in the closed circuit of liquid1 should be withdrawn properly with the aid of the bubble trap. Ten, pure CO 2 was directly inserted into the DI water in reservoir1. Gasifcation was maintained until the pH in reservoir1 reached a defned value. In addition, a magnet stirrer was used in order to homogenize the CO 2 distribution in the reservoir. Te method for the calculation of total aqueous CO 2 concentration based upon aqueous pH was explained in our previous publication [38].
In the frst series of liquid-liquid experiments, the fow rate of CO 2 -rich DI water was kept constant at 2800 ml/ min which is the maximum applicable fow rate of our fabricated HFMCs. Te fow rate of the CO 2 sweeping medium, i.e., liquid2, was set at 500, 750, 1000, and 1500 ml/min. Due to the limitation of the liquid fow rate inside hollow-fber membranes, the maximum applicable fow rate of liquid2 is 1500 ml/min. Te temperatures of both liquids were kept constant at 22°C utilizing both heat exchangers 1 and 2.
Te second series of experiments were designed in order to investigate the efect of temperature variation on aqueous CO 2 removal. Here, the fow rates of liquid1 and liquid2 were kept constant at 2800 and 750 ml/min, respectively. Te temperatures of liquid1 to that of liquid2 (T liq1 : T liq2 ) were set at 44 : 12 and 22 : 12°C. Te results were then compared with the corresponding investigation from the frst series of experiment where T liq1 : T liq2 was 22 : 22°C.

Gas-Liquid HFMC.
Since Oxyphan ® hollow fber membranes are highly porous, both liquid and gas can pass through and therefore are used in gas-gas, gas-liquid, and liquid-liquid HFMCs. In order to compare the afnities of water and air in removing aqueous CO 2 , in the frst approach, liquid2 should be replaced by air, as the sweeping medium. As shown in Figure 4, the air fows inside the hollow-fber lumens as a sweeping medium, while its fow rate is adjusted using the Brooks air fow meter and CO 2rich liquid fows counter currently outside the lumens in a closed loop using the roller-pump1 as well as the liquidliquid HFMC. Tis converts the HFMC into a gas-liquid contactor. In gas-liquid HFMC, the air is used to remove the aqueous CO 2 gradually which leads to the increase in pH. All other conditions are the same as liquid-liquid experiments as mentioned previously. Briefy, in the frst series of gas-liquid experiments, the liquid fow rate was kept constant at 2800 ml/min and the air fow rate was set at 500, 750, 1000, and 1500 ml/min. Temperatures of both liquids were kept constant at 22°C utilizing both heat exchangers 1 and 2.   In the second series of experiments, the fow rates of liquid and air were kept constant at 2800 and 750 ml/min, respectively. Te temperatures of CO 2 -rich liquid to that of air (T liq : T air ) were set at 44 : 12 and 22 : 12°C. Te results were then compared with the corresponding investigation from the frst series of experiments where T liq : T air was 22 : 22°C.

Efect of Sweeping Media on Aqueous CO 2 Removal.
In order to investigate the efect of sweeping media on aqueous CO 2 removal, both liquid-liquid (shown in Figure 3) and gas-liquid (shown in Figure 4)   pH variations were recorded, while the fow rate of liquid1 was kept constant at 2800 ml/min, and the fow rate of fuid2 (water and air in liquid-liquid and gas-liquid circuits, respectively) varied from 500 to 1500 ml/min. To initiate these experiments, liquid1 was gasifed with CO 2 , resulting in a pH value of around 4.2. Herein, air or liquid2 would sweep the aqueous CO 2 , and consequently, the pH level of liquid1 would increase. Figure 5 depicts the results of pH raise using water and air as the sweeping media with the fow rates of 500, 750, 1000, and 1500 ml/min. Te subsequent aqueous CO 2 removal capability could be calculated by the pH-based method [38].
To compensate the fuctuations in the recorded pH values and demonstrating smoothen curves in this work, a moving average function was applied for them. Each series of experiments was conducted 3 times. As shown, the recorded pH courses have the same trends as in Figure 5. Te intersection points of the corresponding pH courses using water or air as the CO 2 sweeping media are also shown in Figure 5 (indicated pH values for the intersection points are the mean pH of three experiment repetitions).

Efect of Temperature on Aqueous CO 2 Removal.
In order to investigate the temperature efect on aqueous CO 2 removal, experiments were conducted, while the fow ratio of the two fuids was adjusted at Q liq1 : Q liq2/air � 2800 : 750 ml/min, and the temperatures of liquid1 to that of liquid2 or air T liq1 : T liq2/air were set at 44 : 12 and 22 : 12°C. Te results were then compared with experiments using  Tables 1 and 2. Te results are reported as an average of three repetitions with their standard deviations (SD). Te outlet temperatures became steady in less than 20 seconds for the adjusted fow rates. Te curves of pH change for these experiments on liquid-liquid and gas-liquid HFMCs are depicted in Figure 6.

Discussion
Te diference between water and air in regard with their ability to remove the aqueous CO 2 by using a liquid-liquid and gas-liquid HFMC has not yet been reported. Although studies show that diferent pressures' availability of water has diferent efects on CO 2 removal from CO 2 -rich gas. At subatmospheric pressure, increasing humidity leads to less CO 2 uptake, and at high pressure, more CO 2 uptake is achieved by increasing humidity [39].
An expectable outcome according to the literature which indicates that the fow rate of the sweeping medium is in correlation with CO 2 removal [40,41] was observed. Te curves demonstrated in Figure 5 show that the higher the fow rate of fuid2, the higher the rate of pH increase in liquid1. Tis indicates that more CO 2 exchange occurs at higher fow rates of the sweeping medium. As illustrated in Figure 5, pH increases of liquid1 versus operating time, which demonstrates the CO 2 removal afnities of air and water as sweeping media, pursue nearly the same trend at the same fow rates of sweeping media. However, each fow rate has an intersection point between air and water curves. In all cases, before this point, which means lower pH or higher aqueous CO 2 concentration, the afnity of air for CO 2 removal is greater than that of water. On the contrary, after the intersection point, this phenomenon becomes vice versa.
Moreover, as shown in Figure 5, the intersection points of water and air as CO2 sweeping media appear in a lower pH by increasing their fow rates. In other words, it is observed that during aqueous CO 2 removal in higher fow rates of the sweeping medium, water becomes more efcient than air. Tis could be addressed by the efect of turbulency in enhancement of mass transfer capability of water. On the other hand, in higher CO 2 concentration, the difusivity of carbon dioxide in air is more efective than its dissolution in water. Furthermore, it is observed that the plateau section of the curves, which its related pH represents the CO 2 removal capacity of the process, in the case of using water as sweeping media is higher in relation to the cases of using air instead.

International Journal of Chemical Engineering
Tis diference can be originated from the efect of chemical dissolution in water case.
In Table 3, the intersection points of water and air, as the CO 2 sweeping media, are presented in mean ± SD considering the repetition of experiments. In addition, the corresponding total CO 2 is calculated by the pH-based method [38] and shown in Table 3.
Tese results are obtained using one-pass air or pure water as a CO 2 -extracting agent. A much higher diference in CO 2 removal capacity is expected by comparing them with the experiments implemented using diethanolamine as a sweeping or extracting agent [32].
It is noteworthy that counter-current one-pass stream of sweeping media increases the efciency of the removal process by shortening the time needed to reach the plateau section (maximum removal capacity) because it is expected to have more overall mass transfer diving force and consequently more mass transfer rate. Tis can result in smaller required retention time.
As presented in Table 3, a higher fow rate (or less retention time) results in lower pH values and more concentration of CO 2 at intersection points, leading to less CO 2 removal at the intersection point. Also, the pH changes vs. time curves in Figure 5 indicate that there is a direct    Figure 6: Efect of temperature on aqueous CO 2 removal with DI water (a) and air (b) as the sweeping medium. correlation between the fow rates and pH values, while the lower fow rates result in higher CO 2 concentration at a specifc time.
It is believed that water with high temperature has more tendency to release its dissolved CO 2 than cold water regarding endothermicity of this phenomenon [42]. Since the CO 2 absorption process is exothermic and, in contrast, gas stripping is endothermic [43][44][45], it is expected that the temperature of liquid1, as the CO 2 carrying medium, should be increased, while the temperature of fuid2, as the CO 2 absorbing medium, should be decreased in order to transfer more CO 2 from liquid1 to fuid2. However, this expectation has not been yet properly scrutinized.
According to Figure 6, the trends of pH changes versus time in three diferent temperature ratios are similar for both liquid-liquid and air-liquid HFMCs. As it can be seen, two pH ranges could be discriminated. In both types of HFMCs, for pH values between circa 5.3 and 6.2 (i.e., total CO 2 concentration between 2.17 × 10 −6 and 6.87 × 10 −5 ml CO 2 /l at 22°C), equality of temperatures of sweeping and carrying media at 22°C resulted in most efciency. After this region, for pH values above 6.3 (i.e., total CO 2 concentration lower than 1.57 × 10 −6 ml CO 2 /l at 22°C), the equality of temperatures at 22°C led to least efciency in both types of HFMCs, while the results for the two other temperature ratios were nearly identical. Our results reveal no clear conclusion for a pH range below 4.9 (i.e., total CO2 concentration greater than 3.95 × 10 −4 ml CO2/l at 22°C).
As shown in Tables 1 and 2 for both cases of using water and air as sweeping media, while there had been a temperature diference between two inlets of HFMC, and this diference became almost nonexistent at the outlet. Tis shows a high capability of selected hollow-fber membranes (here, micro-porous polypropylene) for conductive heat exchange too.

Conclusions
Tis study uses a practical HFMC to compare the removal of aqueous CO 2 between water and air as sweeping media. In addition to aqueous CO 2 concentration, the efects of fow rates and temperature ratios of the CO 2 carrier and sweeper media on the CO 2 exchange rate were examined. Te fndings show that the maximum pH diference between the water and air graphs becomes minor at larger fow rates of sweeping media. Tis has the practical efect of allowing air and water to be utilized interchangeably as sweeping media for aqueous CO 2 removal in HFMCs. Aqueous CO 2 removal from water would typically be more efective than air at lower fow rates. According to the crossing point of the water and air curves for each identical fow rate, there is a certain CO 2 concentration at which the afnities of water and air for aqueous CO 2 removal switch from positive to negative. Interestingly, when the sweeping media fow rate increases, the pH for this crossing point decreases. After this intersection point, i.e., lower aqueous CO 2 concentration, water would be a more efcient CO 2 absorbent than air.
Te results indicate that the temperature similarity of the two fuids in both types of HFMCs, i.e., liquid-liquid and gas-liquid, is the most preferred condition to remove aqueous CO 2 . However, at very low CO 2 concentrations, such as lower than 1.57 × 10 −6 ml CO 2 /l, cooling sweeping media and creating a temperature gradient between the two phases could remove more aqueous CO 2 . Te HFMC's design may be altered to employ the ideal sweeping medium, resulting in the intended CO 2 removal, depending on the intended use and operating conditions. Te accurate removal of a patient's blood's CO 2 content is critical in several challenging therapeutic applications such as AVCO2R and VVCO2R. As a result, this research could be used as a foundation for improving such treatment techniques and enhancing patient improvement.
In conclusion, this study indicates that both air and water have comparable abilities to remove aqueous CO 2 . As a sweeping medium, in higher CO 2 concentration, the affnity of air for CO 2 removal is greater than water. For further investigations, the efect of diferent sweeping media and fow regimen on aqueous CO 2 removal can be studied. Te design of the HFMC can be modifed to obtain higher contact surface and better CO 2 removal as a result. Moreover, the functionality of the proposed HFMC at low pH (lower than 4.9) is well determined.

Data Availability
All data needed for the reproducibility of this research are adequately provided within the article. Te data that support the fndings of this study can be made available from the corresponding author upon reasonable request.

Conflicts of Interest
Te authors declare that they have no conficts of interest.