The Chlorella pyrenoidosa employed in this study was purchased from the Institute of Hydrobiology, Chinese Academy of Sciences (Wuhan, China). Chlorella pyrenoidosa was cultivated in the BG11 medium (Xia et al., 2018).

Figure 1 shows the setup for studying the effects of bubble cutting dynamic behaviors on microalgal growth in a bubble column photobioreactor with a novel aeration device. All trials were conducted at 25°C. The illumination on the surface of the photobioreactors was maintained at 172 μmol/m²·s using fluorescent lamps placed parallel behind the photobioreactor. The photobioreactor comprised a transparent square tube and a novel aeration device fixed at the bottom of the square tube for CO₂ gas supply. The inner cavity size of the square tube was 44 mm × 44 mm × 500 mm (length × width × height), with a working volume of 700 mL. The novel aeration device comprised a sand-core funnel, a multi-orifice plate, and cutting slices. The purchased sand core funnel was a glass funnel with a sand core pore size of 30–50 μm that allows for uniform gas flow. A multi-orifice plate was fixed on top of the sand core funnel to form bubbles with hole diameters and spacings of 1 and 6 mm, respectively. After detachment from the orifice, the bubbles rose straight before gradually transforming into a zigzag (Huang et al., 2017). Thus, the straight-rising stage was more suitable for bubble cutting.

To accurately cut bubbles without affecting bubble formation, cutting slices were installed in parallel, 10 mm above the center of the holes of the multi-orifice plate, which was placed on a fixed trestle attached to the top of the multi-orifice plate. The installation error should be kept within 1 mm. Pure CO₂ (99.99%) and dry air supplied by gas cylinders were mixed thoroughly at different gas flow rates in a gas mixer (Ding et al., 2016) to obtain a mixed gas with the required CO₂ concentrations. The mixed gas was injected from the bottom of the photobioreactor through a sand-core funnel and multi-orifice plate to form a uniform bubble flow. On contacting the cutting slices, the rising bubbles were blocked by the cutting slices due to surface tension and stretched by buoyancy on both sides of the cutting slices. After several times of deformation, bubbles were broken into small sizes, and the CO₂ in the bubbles diffused, dissolved into the suspension, and was bio-fixed by microalgae. Two materials, glass and polytetrafluoroethylene (PTFE), with static contact angles of 32°and 98°, respectively, were used for the cutting slices. The pure glass slices were marked as gg, pure PTFE slices were marked as pp, and the combined slices obtained by pasting half glass and half PTFE were marked as gp. The height and length of the cutting slices were 24 and 38 mm, respectively, and the thickness was 0.15–3 mm.

The bubble cutting dynamic behaviors in the photobioreactor were recorded using high-speed photography (FASTCAM Mini UX100 at a shooting speed of 4,000 fps) with cold light as the backlight. The acquired bubble images were processed and analyzed using matrix laboratory (MATLAB) software. The images underwent grayscale conversion, interference removal, and bubble edge extraction using the Canny operator (Zhu et al., 2014). Subsequently, the bubble equivalent diameter and rising velocity were calculated. To better reflect the bubble behavior, more than one hundred bubbles in the area of height less than 82 mm in the photobioreactor were statistically analyzed owing to their more drastic bubble behavior variation. To facilitate the analysis, the area near the cutting slices from the square tube bottom (0–82 mm height) was divided into four parts in the height direction and marked as A₁, A₂, A_3, and A₄. A₁ (0–10 mm height) reflected the bubble behavior before cutting, and the average bubble diameter d ¯ was marked at 5 mm height. At A₂ (height of 10–34 mm), the bubbles contacted the cutting slices and completed a severe deformation and cutting process. However, the equivalent bubble diameters were not calculated during cutting as they inaccurately reflected the bubble size. A₃ (34–58 mm height) and A₄ (58–82 mm height) reflected bubble-rising behaviors after cutting, and the average bubble diameter d ¯ was marked at heights of 46 and 70 mm, respectively. The bubble diameter and rising velocity listed in Table 1 adopted the average values in A₄ owing to fewer changes in bubble behavior when the height was above 82 mm.

Microalgal suspension samples were taken from the sampling connection of the photobioreactors every 12 h for testing. The hydrogen power (pH) was measured using a pH meter (Thermo Orion, United States). The microalgae concentration was represented by D (g/L dry weight of microalgae cells), calculated from the optical density measured using a spectrophotometer (TU 1901, China) when the wavelength was adjusted to 680 nm (OD _680nm ). The calculation formula for Chlorella pyrenoidosa dry weight was as follows (Ding et al., 2016): D = 0.51 O D 680 n m − 0.029 R 2 = 0.99 (1)

Microalgae biomass productivity P (g/(Ld)) represents the rapidity of microalgal growth, calculated using the following equation (Ding et al., 2016): P = D τ − D 0 τ (2) where D τ (g/L) was the dry weight of microalgae on the day of τ, D 0 (g/L) was the initial dry weight of microalgae.

All studies were performed in duplicate. All measurements were repeated five times, and the data were presented as the average value and standard deviation.

In this study, bubble dynamic behaviors were observed and compared in the photobioreactors with and without cutting units, and their effects on microalgal growth were analyzed. The cutting units used were glass slices of 1 mm thickness. The inlet CO₂ concentration in the mixed gas was 5%, and the gas flow rate was maintained at 70 mL/min. The results are shown in Figure 2. Figure 2A shows the entire bubble deformation process during bubble cutting. The bubbles accelerated vertically in the morph of ellipsoid owing to buoyancy, inertial forces, surface tension, and fluid shear stress after departure from the orifices. On touching the bottoms of the glass slices, the bubble-rising velocity decreased. They deformed into a cashew morph, and the gas-liquid interface fluctuated significantly. Furthermore, the bubbles were stretched by buoyancy on both sides of the cutting slices, and middle necking emerged. Once the middle neck of the mother bubble broke, it was divided into two daughter bubbles. Therefore, the bubble cutting behaviors reduced the bubble size, enhanced the gas-liquid interface area and fluctuation, slowed the bubble rise, and prolonged the bubble residence time, conducive to CO₂ dissolution and fixation. Figure 2B shows the bubble diameter distribution in the height direction of the bioreactor. Bubble diameters remained at approximately 6 mm in the entire bioreactor without cutting units, whereas an evident decrease was observed at 10 mm height owing to bubble cutting and were uniformly kept at approximately 4.3 mm during the subsequent rising process in the bioreactor with cutting units. Moreover, the statistical values in Table 1 reveal that adding cutting units decreased the bubble diameter by 27.97%, decreased the bubble rising velocity by 46.88%, and increased the residence time by 84.55%, all contributing to CO₂ transfer and microalgal growth.

In addition, Figures 2C, D show the variation in dry weight and pH during algal growth, validating the above analysis. As shown in Figure 2C, in the first 60 h of growth, little differences were observed between the bioreactors with and without cutting units owing to the low algal cell concentration in suspension and lower CO₂ demand. However, as microalgae grew, CO₂ demand increased. The photobioreactor with cutting units had a greater growth rate and shorter growth time, and the maximum dry weight and biomass productivity reached 1.53 g/L and 0.24 g/(Ld), respectively, which were improved by 6.99% and 33.33%, respectively, compared with the bioreactor without cutting units. This is because the photobioreactor with cutting units had higher CO₂ mass transfer and dissolution efficiency, promoting microalgal growth. Similarly, Figure 2D shows that the bioreactor with cutting units had lower pH during the entire growth process. Although bubble cutting may not achieve the equal increase of biomass productivity as existing enhancement technologies (Cheng et al., 2016; Yang et al., 2016; Xia et al., 2018; Ye et al., 2018), it causes lower energy consumption, longer bubble residence time, and bubble breakage accuracy.

To further optimize the cutting slice parameters, the effects of the glass cutting unit thickness (δ = 0.15, 1, and 3 mm) on bubble behavior and microalgal growth were investigated in the photobioreactor when the inlet CO₂ concentration in the mixed gas was 5%, and the gas flow rate was controlled at 70 mL/min. The results are shown in Figure 3. Figure 3A shows the distribution of the bubble diameters in the height direction of the bioreactor. The bubble diameters decreased at 10 mm height when the cutting unit thickness δ was less than 1 mm owing to successful bubble cutting. The bubble diameters were uniformly at a low level of 4.72 and 4.3 mm when cutting unit thicknesses δ were 0.15 and 1 mm, respectively. At 3 mm slice thickness δ, the average bubble diameter remained larger than 6 mm and increased to 6.75 mm in A₄. This is because the large thickness δ hindered the bubbles from forming a cashew morph owing to too small bubble heads on both sides of the cutting slices to provide buoyancy, causing bubbles to alternately shrink and stretch several times and slip away on one side of the slices owing to force imbalance, prolonging bubble residence time twice at A₂ compared with that of a slice thickness δ of 1 mm, and increasing bubble size owing to coalescence with subsequent small bubbles. In addition, it can be deduced from the statistical values in Table 1 that a slice thickness δ of less than 1 mm decreased bubble size and rising velocity and prolonged the residence time, facilitating CO₂ transfer and microalgal growth. At a thickness δ of 3 mm, although the residence time was prolonged by 41.82%, the bubble diameter and rising velocity increased by 13.07% and 12.50%, respectively, compared with those of the bioreactor without cutting units, which were adverse to CO₂ dissolution and fixation.

Figures 3B, C show the variation in dry weight and pH during algal growth, reaching a similar conclusion to that of the above analysis. As shown in Figure 3B, in the primary growth stage, microalgae in the bioreactor with 1 mm thickness δ had the smallest growth rate owing to the low pH (Figure 3C) inhibiting microalgal growth, while the algal growth rate gradually increased to the highest as the microalgae grew, and the largest maximum dry weight and biomass productivity were achieved owing to the increased CO₂ demand. For the photobioreactor with 3 mm thickness δ, the maximum dry weight and biomass productivity were reduced by 4.20% and 5.56%, respectively, compared with that of the bioreactor without cutting units. This is because the photobioreactor with 3 mm thickness δ had the lowest CO₂ mass transfer and dissolution efficiency, impeding microalgal growth. Additionally, Figure 3C shows that the bioreactor with 3 mm thickness δ had the highest pH, while that with 1 mm had the lowest pH during the entire growth process is conclusive.

Hydrophobic slice surfaces may prevent daughter bubble departure and prolong bubble residence time, promoting CO₂ and algal growth. To prove this hypothesis, the effects of the cutting unit hydrophilicity (gg, gp, pp) on bubble behavior and microalgal growth were investigated in a photobioreactor when the inlet CO₂ concentration in the mixed gas was 5%, and the gas flow rate was maintained at 70 mL/min. The results are shown in Figure 4, where gg indicates that the cutting units were glass slices, pp indicates that the cutting units were PTFE slices, and gp indicates that the cutting units were combined glass and PTFE slices. Figure 4A shows the bubble diameter distribution in the height direction of the bioreactor. It can be concluded that the bubbles can be successfully cut in all three cases owing to the evident decrease in the bubble diameter at 10 mm height. The most uniform bubble diameter distribution and smallest average diameter were achieved when the cutting units were gg slices. The opposite results were obtained when the cutting units used gp. This was because the glass surfaces were hydrophilic and adversely affected the bubble attachment, whereas the PTFE surfaces were hydrophobic and conducive to bubble attachment. Thus, daughter bubbles struggled to depart from the PTFE cutting surfaces owing to the high surface tension and rose along cutting slices after bubble cutting, prolonging the bubble resistance time and causing coalescence with subsequent bubbles.

Regarding gp, the bubbles were cut quickly and unevenly owing to unbalanced forces, and the daughter bubbles at the PTFE surfaces always had larger diameters owing to the higher surface tension and bubble coalescence. A non-uniform bubble diameter distribution and the largest average diameter were obtained when the cutting units were gp slices. Furthermore, the statistical values in Table 1 show that smaller bubble sizes, lower rising velocities, and longer residence times were achieved using gg slices as the cutting unit, facilitating CO₂ dissolution. In contrast, when gp was applied, the lowest decreases in bubble diameter and rising velocity were obtained at 15.58% and 25%, respectively, and the shortest residence time prolongation was 46.39% compared with those of the bioreactor without cutting units.

The growth curve and pH values shown in Figures 4B, C support the above analysis. Less distinct growth curves are shown in Figure 4B during the primary growth stage, and the bioreactor with gg slices had the lowest growth rate owing to its lower pH (Figure 4C). While evident differences occurred after 72 h of growth, the gg growth rate gradually increased to the highest and achieved the largest maximum dry weight and biomass productivity with an increased CO₂ demand, owing to the highest CO₂ mass transfer and dissolution efficiency. The maximum dry weight and biomass productivity of the gg slice were 7% and 26.32% higher than those of the gp slices, respectively, illustrating the inaccuracy of the initial assumption.

To cut bubbles into smaller units, the effects of cutting unit arrangement on bubble behavior and microalgal growth were investigated in a photobioreactor. The cutting units used were glass slices of 0.15 mm thickness. The inlet CO₂ concentration in the mixed gas was 5%, and the gas flow rate was maintained at 70 mL/min. Bubble cutting was challenging when the cutting slices were installed in crosses owing to the large three-phase contact interface resistance at the crossing. The bubbles coalesced and accumulated under the cutting units, and detached in a large size. All cut slices were arranged in parallel to improve the success rate of bubble cutting. Two cutting slices were installed above each row of holes, and the centerline of each row of holes was kept in the middle of the two cutting slices. Figure 5 shows the effects of the spacing between the two cutting slices (M = 0, 1, and 3 mm) on bubble behavior and microalgal growth. As shown in Figure 5A, the mother bubble was barely cut into three daughter bubbles. When the cutting slice spacings M were 0 and 1 mm, the bubbles could be cut into two parts, conclusive from the decreased bubble diameters at 10 mm height. When the cutting slice spacing M was increased to 3 mm, the average bubble diameter remained approximately 6 mm in region A₄. This is because when the cutting slice spacing M is larger than 0 mm, bubbles form three heads when touching the cutting slices and oscillate owing to the imbalanced force, which can promote interface transportation and prolong the bubble residence time. A small spacing (M = 1 mm) caused a small middle bubble head and larger side heads, and the bubbles were easily separated into two parts owing to the larger buoyancy provided by the two side heads, causing the middle head to shrink. When M was increased to 3 mm, the middle bubble head became larger than the side heads, and the bubbles shrank thoroughly into spaces owing to the larger buoyancy in the middle. When the spacing M was between 1 and 3 mm, the bubble behaviors varied; they may be cut, shrink into spacing, or slip away owing to severe oscillation. Moreover, the statistical values in Table 1 reveal that a spacing M less than 1 mm decreased the bubble size and rising velocity and prolonged the residence time, facilitating CO₂ transfer and microalgal growth. However, at 3 mm spacing M, the residence time was prolonged by 14.55%, and the bubble diameter and rising velocity were almost the same as those of the bioreactor without cutting units, which could not promote CO₂ dissolution and fixation.

Figures 5B, C show the variation in dry weight and pH during algal growth, consistent with the above analysis. As shown in Figure 5B, at the beginning of 24 h growth, microalgae in the bioreactor with M of 0 and 1 mm had similar growth rates, whereas the algal growth rate gradually increased to the highest when M was 1 mm, and the largest maximum dry weight and biomass productivity were achieved owing to greater interface fluctuation and longer bubble residence time, facilitating CO₂ transfer. However, the lowest maximum dry weight and biomass productivity were achieved in the photobioreactor with a 3 mm spacing M owing to the larger bubble size and shorter bubble residence time inhibiting CO₂ mass transfer and microalgae growth. In addition, it is conclusive from Figure 5C that the bioreactor with a 3 mm spacing M had the highest pH, while that of 1 mm spacing M had the lowest pH during the entire growth process.

The above studies indicated that bubble column photobioreactor with glass cutting slices (gg, δ = 1 mm, M ≤ 1 mm) parallelly arranged in the novel aeration device had the largest maximum dry weight and biomass productivity owing to smaller bubble size, lower rising velocity, and longer bubble residence time, facilitating CO₂ transfer and fixation. In addition to the structural parameters of the aeration device, the ventilation conditions, including the gas flow rate and inlet CO₂ concentration in the mixed gas, had significant effects on the bubble behavior and microalgal growth (Zhao et al., 2015), as shown in Figure 6. Figures 6A, B show the variation in the growth curve and pH with different inlet CO₂ concentrations (3%, 5%, and 10%) in the microalgae photobioreactor with the optimal aeration device (gg, δ = 1 mm, M = 0 mm) and the gas flow rate was maintained at 70 mL/min. Therefore, it can be concluded that when the inlet CO₂ concentration supplied was 5%, microalgae grew the fastest, and the maximum microalgae dry weight and productivity were achieved (Figure 6A). This was because an inlet CO₂ concentration of 10% resulted in severe acidification of the suspension and the lowest pH (Figure 6B) owing to the higher CO₂ dissolution rate, smaller bubble size, and longer bubble residence time, which completely inhibited microalgae growth (Chiu et al., 2008; Yoo et al., 2010). Moreover, a higher inlet CO₂ concentration caused severe bubble-carrying, and fewer cells were present in the middle of the suspension, impeding microalgal growth (Ding et al., 2016). In contrast, when the inlet CO₂ concentration was lower, the cells lacked CO₂ supply, the pH was maintained at a high level, and the maximum pH reached 7.36 (Figure 6B).

Furthermore, the effects of gas flow rate (30, 70, and 140 mL/min) on microalgae growth in a microalgae photobioreactor with the optimal aeration device (gg, δ = 1 mm, M = 0 mm) were investigated when the inlet CO₂ concentration supplied was 5% (Figures 6C, D). The best performance was achieved at the 70 mL/min gas flow rate. This was because a lower gas flow rate caused an insufficient CO₂ supply, restricting microalgal growth (Figure 6C), causing less gas-liquid interface area, and the weaker interface oscillation hindered CO₂ transfer. Therefore, the lowest biomass productivity of 0.15 g/(Ld) and highest pH of 7.4 was achieved at a 30 mL/min gas flow rate (Figure 6D). However, a larger gas flow rate (140 mL/min) induced more gas-liquid interface area and stronger interface oscillation, causing CO₂ excess supply, impeding algal growth owing to the lowest pH and more acidic suspension (Chiswell et al., 1999; Vogel et al., 2010). In addition, a higher gas flow rate caused a higher bubble-carrying coefficient owing to the generation of more bubbles per unit time, which was unconducive to algal growth (Ding et al., 2016).

To sum up, the optimal performance was obtained in the bubble column photobioreactor with the novel aeration device (gg, δ = 1 mm, M = 0 mm) when inlet CO₂ concentrations was 5%, gas flow rate was 70 mL/min, and the maximum dry weight and biomass productivity reached 1.53 g/L and 0.24 g/(Ld), respectively.