Cells of the mixotrophic dinoflagellate G. smaydae GSSH1005 were originally isolated from the coastal waters of Shiwha Bay, Korea; the water temperature and salinity were 19.0°C and 27.7, respectively (Kang et al. 2014). Cells of H. rotundata HRSH1201, a suitable prey of G. smaydae, were isolated from the surface waters of Shiwha Bay, Korea; the water temperature and salinity were 0.2°C and 31.0, respectively (Kang et al. 2014). Cultures of G. smaydae and H. rotundata were maintained in 10-L polycarbonate bottles (Thermo Fisher Scientific, Waltham, MA, USA) containing 0.2-μm-filtered seawater (G. smaydae) or f/2-Si medium (H. rotundata) (Guillard and Ryther 1962) at 25°C under illumination with 50 μmol photons m⁻² s⁻¹ using a light-emitting diode (LED) lamp under a 14 : 10 h light / dark cycle. These temperature and light intensity conditions are optimal for maximum growth of G. smaydae (You et al. 2020b). Cells of H. rotundata as prey were added to the G. smaydae culture every 2 or 3 days.

The semi-continuous cultivation system for mixotrophic microalgae on a 100-L scale used in this study was scaled up from a 10-L system described previously (Jeong and Lim 2020, Lim et al. 2020) (Fig. 1). This 100-L scale system consisted of three 200-L acrylic water tanks; two LED lamps for illumination per tank; two peristaltic pumps (77420-20, Masterflex; Cole Parmer, Vernon Hills, IL, USA); two devices (MultiLab 4010-3W; YSI, Yellow Springs, OH, USA), each connected to three sensors for measuring water temperature (°C), salinity, pH, and dissolved oxygen (mg L⁻¹); and silicone tubings (Masterflex). The three water tanks containing H. rotundata as prey, G. smaydae for growing, and G. smaydae for storage after growing, respectively, were connected in series through the silicone tubings (Fig. 1). Two peristaltic pumps were used to transport the prey culture from the prey culture tank to the predator culture tank and, subsequently, the predator culture from the predator culture tank to the predator storage tank. The medium was manually added to the prey culture tank after these transport processes. This 100-L scale system was executed in a temperature-controlled walk-in chamber, maintaining a temperature of 25°C under the culture conditions described above.

For experiments 1–4, G. smaydae was cultivated in the 100-L scale system using the following process: at the beginning of the procedure, 120 L of dense H. rotundata culture (>100,000 cells mL⁻¹) was added to the prey culture tank, and 10 L of dense G. smaydae cultures (approximately 50,000 cells mL⁻¹) were added to the predator culture tank. In this system, 20 or 30 L H. rotundata culture in the prey culture tank was automatically transported every day to the G. smaydae culture in the predator culture tank through a peristaltic pump (flow rate: 2 L min⁻¹). Subsequently, the prey tank was manually refilled with 20 or 30 L of f/2-Si medium. Therefore, the culture volume of 120 L in the prey tank was continuously maintained; however, the volume of G. smaydae culture in the predator culture tank gradually increased through the addition of H. rotundata culture to the tank. When the volume of G. smaydae culture in the predator culture tank reached 90 or 100 L, most of the volume (80 or 90 L) was transported to the predator storage tank. By repeating these steps, G. smaydae was cultivated semi-continuously in the predator culture tank. In the predator storage tank, 80 or 90 L of G. smaydae culture was maintained without added prey for 1 day under the same culture conditions (Lim et al. 2020), and G. smaydae cells in the predator storage tank were harvested 1 day later through centrifugation to analyze the density and fatty acid contents.

Ten-milliliter aliquots were removed from the water in the prey and predator culture tanks using an auto pipette and fixed using a 5% acidic Lugol’s solution. Subsampling from the prey or predator culture tank was performed before and after adding f/2-Si medium or prey culture to the tanks. From the predator storage tank, a 10-mL aliquot was collected before the culture was harvested. All or more than 200 cells of each species were counted in three 1-mL Sedgewick-Rafter (SR) chambers using an inverter microscope (CX21; Olympus Corporation, Shinjuku, Tokyo, Japan) at 100- or 200-fold magnification. During this cultivation, the pH and dissolved oxygen levels were also monitored.

Experiments 1 and 2 were designed to investigate the effects of continuous and no aeration on cell growth, feeding, and fatty acid composition of G. smaydae cultivated in the 100-L semi-continuous cultivation system.

In experiment 1, G. smaydae was cultivated without aeration for 9 days, as described in the section “Semi-continuous cultivation system of Gymnodinium smaydae on a 100-L scale.” The cultivation was repeated three times consecutively (periods 1–3) for 9 days. During each period, 30 L of H. rotundata culture from the prey culture tank was transported each day to the predator culture tank containing G. smaydae culture. When the volume of G. smaydae culture in the predator culture tank reached 100 L on 3 days, 90 L of G. smaydae culture was transported to the predator storage tank, which was then maintained without added prey for 1 day. One day later, triplicate 5-L aliquots from the predator storage tank were harvested through centrifugation (Labogene 1696R; Gyrozen Co., Gimpo, Korea) at 3,000 revolutions per minute (rpm) (4,315 ×g) for 10 min and then stored at −70°C in a deep freezer before measurements of fatty acids.

In experiment 2, G. smaydae was cultivated under continuous aeration for 9 days (periods 1–3) and harvested as described above. An air pump and a sparger provided continuous aeration at an airflow rate of 3 L min⁻¹ (vvm = 0.03–0.08).

Cells of G. smaydae were harvested using a continuous centrifugation system comprising an industrial continuous centrifuge (GK-2-A; Gosgo, Paju, Korea), a peristaltic pump (77420-20; Masterflex) for injecting microalgae cultures from the predator storage tank into the centrifuge, and a chiller (HD-0.5A; Hyundai ENG Co., Ltd., Seoul, Korea) for controlling water temperatures (Fig. 2). This centrifugation system facilitated the continuous harvesting of microalgae that were injected from the predator storage tank into the centrifuge by the peristaltic pump. The water from which microalgae cells had been removed was discharged in real-time through the centrifuge outlet. Consequently, only microalgae cells were collected inside the centrifuge bowl. The chiller of the centrifugation system controlled the water temperature of injected microalgae cultures; therefore, the harvesting efficiency and biomass concentration could only be affected by the rotational speed of the centrifuge and the flow rate of the injection pump.

Experiments 3 and 4 were designed to determine the effects of the rotational speed of the continuous centrifuge and the flow rate of the injection pump from the predator storage tank to the centrifuge bowl, respectively. For experiment 3, G. smaydae was cultivated for 4 days as described in the section “Semi-continuous cultivation system of Gymnodinium smaydae on a 100-L scale.” Every day, 20 L of H. rotundata culture from the prey culture tank was transported to the predator culture tank containing G. smaydae culture. When the volume of G. smaydae culture in the predator culture tank reached 90 L on day 4, 80 L of G. smaydae culture was transported to the predator storage tank and maintained without added prey for 1 day. Triplicate 5-L aliquots of G. smaydae culture from the predator storage tank were centrifuged at four different speeds: 2,000, 4,000, 8,000, and 16,000 rpm (1,611, 3,221, 6,442, and 12,884 ×g, respectively). In this experiment, the flow rate of the peristaltic pump for injecting G. smaydae cultures into the centrifuge was 0.5 L min⁻¹. The collected G. smaydae cells in the centrifuge bowl were transferred to 15-mL Falcon tubes (Corning Inc., Corning, NY, USA) and then stored at -70°C before the lipid composition was analyzed.

For experiment 4, G. smaydae was cultivated following the same method as in experiment 3. Duplicate 5-L aliquots of G. smaydae culture from the predator storage tank were centrifuged at three different flow rates of the peristaltic pump for injection: 1.0, 2.0, and 4.0 L min⁻¹. In this experiment, the rotational speed of the centrifuge was 16,000 rpm (12,884 ×g). The collected G. smaydae cells in the centrifuge bowl were transferred and stored as described above.

In experiments 3 and 4, fatty acid composition, the dry weight of total fatty acids (mg L⁻¹) and omega-3 fatty acids (mg L⁻¹ and % of total fatty acids) of G. smaydae, as well as its harvesting efficiency (%), were analyzed. To calculate the harvesting efficiency of G. smaydae cells under each experimental condition, effluent discharged during G. smaydae centrifugation was placed in one PC bottle each for experiments 3 and 4. A 10-mL aliquot was removed from each bottle and was fixed using 5% Lugol’s solution. G. smaydae cells that remained in the effluent were counted under the inverted microscope using three 1-mL SR chambers. Harvesting efficiencies for each experimental condition were calculated as follows:

, where GS_culture is the density (cells mL⁻¹) of G. smaydae in the predator storage tank before the cells were centrifuged, and GS_eff is the density (cells mL⁻¹) of G. smaydae in the effluent after the cells were centrifuged.

When the fatty acid compositions and contents of G. smaydae maintained in the predator storage tank for 1 day were analyzed, a small amount of H. rotundata cells cultivated under continuous aeration in periods 1 and 2 and a large amount of H. rotundata cells in the period 3 remained. Therefore, the analysis results were expressed as the fatty acid compositions and contents of G. smaydae + H. rotundata. According to Lim et al. (2020), the fatty acid composition and content of only G. smaydae cells were similar to those of G. smaydae cells with H. rotundata cells, when the density of H. rotundata was <1% of that of G. smaydae.

Frozen microalgal cells were lyophilized at −110°C for 24 h under vacuum using a freeze dryer (Bondiro; Ilshin, Dongducheon, Korea) to obtain dry biomass as described by Lim et al. (2020). Total fatty acid methyl esters (FAMEs) were extracted from the lyophilized cells and analyzed using the one-step hydrolysis, extraction, and methylation procedure (Garcés and Mancha 1993).

FAMEs were analyzed using gas chromatography (7890A; Agilent Technologies, Santa Clara, CA, USA). A 1-μL aliquot of the extracted FAMEs was injected into a capillary column (DB-23, Ser. No. US8897617H; 60 m × 0.25 mm × 0.25 μm film thickness) coupled with a flame ionization detector at a split ratio of 20 : 1 (Lim et al. 2020). The thermal profile consisted of the following steps: initial temperature of 50°C sustained for 1 min, increase to 130°C at 15°C min⁻¹, increase to 170°C at 8°C min⁻¹, increase to 215°C at 2°C min⁻¹, and maintenance for 10 min. The injector and detector temperatures were 250 and 280°C, respectively. FAME peaks were determined by matching the retention times of the samples with those of the reference standards (Supelco 37-component FAME mix; Supelco, Bellafonte, PA, USA). Omega-3 fatty acid contents were determined as the sums of EPA, DHA, and ALA contents.

Univariate analyses and post-hoc tests were conducted to investigate significant differences among data produced from experiments 1–4. All statistical analyses were performed using SPSS version 25 (IBM-SPSS Inc., Armonk, NY, USA). Statistical significance was considered at p < 0.05.

Before the analyses, normality and homogeneity of variance were tested using Shapiro-Wilk W and Levene’s tests. When both assumptions of normality and homogeneity of variance were satisfied, a parametric one-way analysis of variance and Tukey’s honestly significant difference post-hoc test were performed. When the assumption of normality was not satisfied, a non-parametric Kruskal-Wallis test and Mann-Whitney U comparison with a post-hoc Bonferroni correction were performed.

In the predator culture tank, the densities of G. smaydae cultivated under no aeration (experiment 1) were similar at the end of periods 1 (3 days), 2 (6 days), and 3 (9 days), i.e., 45,017, 52,333, and 45,600 cells mL⁻¹, respectively (Figs 3 & 4A). Whereas, the densities of G. smaydae in the predator culture tank cultivated under continuous aeration (experiment 2) were similar at the end of periods 1 (3 days) and 2 (6 days), i.e., 69,250 and 68,500 cells mL⁻¹, respectively; however, they were higher than that at the end of period 3 (9 days), 47,286 cells mL⁻¹ (Fig. 4B). Therefore, the densities of G. smaydae in the predator culture tank at the end of periods 1 (3 days) and 2 (6 days) cultivated under continuous aeration (experiment 2) were higher than those under no aeration (experiment 1). However, the density of G. smaydae in the predator culture tank at the end of period 3 (9 days) cultivated under continuous aeration (experiment 2) was similar to that under no aeration (experiment 1).

In the predator storage tank, the densities of G. smaydae cultivated under no aeration (experiment 1) and maintained for 1 day were 50,600, 53,000, and 47,875 cells mL⁻¹, respectively, whereas those cultivated under continuous aeration (experiment 2) and maintained for 1 day were 46,400, 58,950, and 19,417 cells mL⁻¹, respectively (Fig. 5). In the predator storage tank, the densities of G. smaydae cultivated under no aeration (experiment 1) and maintained for 1 day in periods 1 and 2 were similar, and they were also similar to those under continuous aeration (Fig. 5); however, the density of G. smaydae cultivated under no aeration (experiment 1) and maintained for 1 day in period 3 was higher than that under continuous aeration (experiment 2).

In the predator culture tank, the densities of H. rotundata at the end of periods 1 (3 days), 2 (6 days), and 3 (9 days) cultivated under no aeration (experiment 1) were 201, 1,428, and 4,034 cells mL⁻¹, respectively, whereas those cultivated under continuous aeration (experiment 2) were 0, 50,800, and 316,333 cells mL⁻¹, respectively (Figs 3, 4C & D). In the predator storage tank, the densities of H. rotundata cultivated at the end of periods 1, 2, and 3 under no aeration (experiment 1) and maintained for 1 day were 0, 82, and 52 cells mL⁻¹, respectively, whereas those cultivated at the end of periods 1, 2, and 3 under continuous aeration (experiment 2) were 0, 260, and 161,667 cells mL⁻¹, respectively (Fig. 5). Therefore, during maintenance in the predator storage tank, G. smaydae eliminated all H. rotundata cells in period 1 under both non-aeration and continuous aeration and most H. rotundata cells in period 2 under both non-aeration and continuous aeration; however, G. smaydae did not eliminate a large number of H. rotundata cells in period 3 under the continuous aeration conditions.

The fatty acid compositions of G. smaydae + H. rotundata cultivated under no aeration and maintained in the predator storage tank for 1 day in periods 1, 2, and 3 were similar (Fig. 6). Notably, the proportion of DHA in the weight of total fatty acids, 40.5–44.1%, was the highest, and that of palmitic acid (C16 0), 21.7–23.8%, was the second highest. Furthermore, the proportion of EPA in the weight of total fatty acids was 9.5–10.1% (Fig. 6). The fatty acid compositions of G. smaydae + H. rotundata cultivated under continuous aeration and maintained in the predator storage tank for 1 day in periods 1 and 2 were similar, and they were also similar to those under no aeration (Fig. 6). However, the fatty acid composition of G. smaydae + H. rotundata cultivated under continuous aeration and maintained in the predator storage tank for 1 day in period 3 was largely different from that in periods 1 and 2; the portion of DHA in the weight of total fatty acids, 29.2%, was the second highest, whereas that of palmitic acid, 31.5%, was the highest (Fig. 6).

The dry weights of total fatty acids of G. smaydae + H. rotundata cells cultivated under no aeration and harvested from 5 L in the predator storage tank in periods 1, 2, and 3 (0.52–0.62 mg L⁻¹) were similar, whereas those after cultivation under continuous aeration in the storage tank were considerably different (Fig. 7A & B). The dry weights of total fatty acids of G. smaydae + H. rotundata cells cultivated under continuous aeration in periods 1 and 2 (0.72–0.98 mg L⁻¹) were higher than that in period 3 (0.25 mg L⁻¹); however, the sum of the dry weights of total fatty acids of G. smaydae + H. rotundata cells cultivated under no aeration in periods 1–3 (1.71 mg L⁻¹) was similar to that after cultivation under continuous aeration conditions (1.95 mg L⁻¹) (Fig. 7C).

The dry weights of omega-3 fatty acids in G. smaydae + H. rotundata cells cultivated under no aeration and harvested from 5 L in the predator storage tank in periods 1, 2, and 3 (0.29–0.32 mg L⁻¹) were similar, whereas those cultivated under continuous aeration in the storage tank were considerably different (Fig. 7D & E). The dry weights of omega-3 of G. smaydae + H. rotundata cells cultivated under continuous aeration in periods 1 and 2 (0.37–0.52 mg L⁻¹) were higher than that in period 3 (0.10 mg L⁻¹); however, the sum of the dry weights of omega-3 of G. smaydae + H. rotundata cells cultivated under no aeration in periods 1–3 (0.90 mg L⁻¹) was similar to that after cultivation under continuous aeration (0.99 mg L⁻¹) (Fig. 7F).

The ratios (%) of omega-3 fatty acids relative to total fatty acids of G. smaydae + H. rotundata cells cultivated under no aeration and harvested from 5 L in the predator storage tank in periods 1, 2, and 3 (51.2–55.0%) were similar, whereas those after cultivation under continuous aeration in the storage tank were considerably different (Fig. 7G & H). The ratios (%) of omega-3 fatty acids relative to total fatty acids of G. smaydae + H. rotundata cells cultivated under continuous aeration in periods 1 and 2 (52.2–53.1%) were higher than that in period 3 (38.5%).

In experiment 3, the harvesting efficiencies (%) of G. smaydae cells cultivated for 4 days under no aeration and harvested from 5 L in the predator storage tank of the 100-L system at rotation speeds of 2,000, 4,000, 8,000, and 16,000 rpm were significantly different (Kruskal-Wallis test, H₃ = 10.421, p = 0.015) (Fig. 8A, Supplementary Figs S1 & S2). Notably, with increasing rotating speed, the harvesting efficiencies of G. smaydae cells increased from 32.8 to 93.4%. In addition, the dry weight of total fatty acids of G. smaydae + H. rotundata cells cultivated under no aeration and harvested from 5 L from the predator storage tank at rotation speeds of 2,000, 4,000, 8,000, and 16,000 rpm exhibited a pattern similar to that of the harvesting efficiency (Fig. 8B). With increasing the rotation speed, the dry weight of total fatty acids in G. smaydae + H. rotundata cells significantly increased from 0.02 to 0.50 mg L⁻¹ (Kruskal-Wallis test, H₃ = 10.009, p = 0.018). Furthermore, the dry weight of omega-3 of G. smaydae + H. rotundata cells cultivated under no aeration for 4 days and harvested from 5 L from the predator storage tank at rotation speeds of 2,000, 4,000, 8,000, and 16,000 rpm exhibited a pattern similar to that of the harvesting efficiency (Fig. 8C). With increasing the rotating speeds of the centrifuge, the dry weight of omega-3 of G. smaydae + H. rotundata cells significantly increased from 0.01 to 0.23 mg L⁻¹ (Kruskal-Wallis test, H₃ = 9.495, p = 0.023). Moreover, the ratio (%) of omega-3 relative to total fatty acids of G. smaydae + H. rotundata cells cultivated under no aeration and harvested from 5 L from the predator storage tank at rotation speeds of 2,000, 4,000, 8,000, and 16,000 rpm, 40.8–45.9%, were similar but significantly different (Kruskal-Wallis test, H₃ = 8.692, p = 0.034) (Fig. 8D). The cell morphology of G. smaydae was examined at each rotational speed (Fig. 8E), and changes from the original shape to spherical were observed immediately after centrifugation, with no differences in shape depending on the rotational speed.

The fatty acid compositions of G. smaydae + H. rotundata cells cultivated for 4 days under no aeration and harvested from 5 L from the predator storage tank of the 100-L system at rotation speeds of 2,000, 4,000, 8,000, and 16,000 rpm were similar (Fig. 9). The proportion of DHA in the weight of total fatty acids (32.4–36.7%) was the highest, and that of palmitic acid (22.0–24.4%) was the second highest. Furthermore, the proportion of EPA in the weight of total fatty acids was 7.7–8.6% (Fig. 9).

In experiment 4, the harvesting efficiencies (%) of G. smaydae cells cultivated for 4 days under no aeration and harvested from 5 L from the predator storage tank of the 100-L system at injection flow rates of 1.0, 2.0, and 4.0 L min⁻¹ (89.3–95.5%) did not differ significantly (Kruskal-Wallis test, H₂ = 3.714, p = 0.156) (Fig. 10A, Supplementary Figs S3 & S4). Similarly, the dry weight of total fatty acids (mg L⁻¹), the dry weight of omega-3 (mg L⁻¹), and the ratio (%) of omega-3 relative to total fatty acids of G. smaydae + H. rotundata cells at injection flow rates of 1.0, 2.0, and 4.0 L min⁻¹ (0.60–0.65 mg L⁻¹, 0.27–0.28 mg L⁻¹, and 43.4–44.4%, respectively) also did not differ significantly (Kruskal-Wallis test, H₂ = 1.143, p = 0.565 for the dry weight of total fatty acids; H₂ = 0.857, p = 0.651 for the dry weight of omega-3; H₂ = 3.429, p = 0.18 for the ratio of omega-3 relative to total fatty acids) (Fig. 10B–D).

The fatty acid compositions of G. smaydae + H. rotundata cells cultivated for 4 days under no aeration and harvested from 5 L from the predator storage tank of the 100-L system at flow rates of 1.0, 2.0, and 4.0 L min⁻¹ were similar (Fig. 11). The proportion of DHA in the weight of total fatty acids (34.8–35.5%) was the highest, and that of palmitic acid (22.7–23.2%) was the second highest. Furthermore, the proportion of EPA in the weight of total fatty acids was 7.9–8.2% (Fig. 11).

The results of the present study clearly demonstrated that continuous aeration at low airflow rates of 0.03–0.08 vvm affected the production of G. smaydae; however, the dry weight of total fatty acids in G. smaydae + H. rotundata cultivated under continuous aeration was greater than that under no aeration in the first two periods, indicating that aeration enhanced fatty acid production of G. smaydae + H. rotundata, probably when G. smaydae cells were healthy. In contrast, the weight of total fatty acids in G. smaydae + H. rotundata cultivated under continuous aeration was lower than that under no aeration in the last period, indicating that aeration inhibited fatty acid production by G. smaydae + H. rotundata, probably when G. smaydae cells were not healthy. Therefore, G. smaydae may be considerably affected by an increase in the aeration time (i.e., over 6 days). Subsequently, when continuous aeration is adopted, G. smaydae cells should be cultivated for less than 6 days, harvested, and replaced with a fresh culture. Meanwhile, the weights of total fatty acids in G. smaydae + H. rotundata cultivated under no aeration and continuous aeration conditions were similar after 9 days. Therefore, cultivating G. smaydae cells without aeration is a more effective approach in terms of productivity than using aeration. Furthermore, no aeration reduces the energy costs required to operate aeration devices.

Consistent with our results, the production of the haptophyte Prymnesium parvum (in mixotrophic cultivation) and chlorophyte Scenedesmus obliquus under no aeration was higher than that under aeration conditions (Table 1). Their production was lowered at low airflow rates of <0.1 vvm (Hodaifa et al. 2010, Vidyarathna et al. 2014). In contrast, the production of chlorophytes Chlorella sorokiniana and P. parvum (in autotrophic cultivation) and cryptophyte Rhodomonas salina was enhanced under low aeration (Table 1). Notably, the airflow rates of aeration by air pump devices supporting maximum production of C. sorokiniana, P. parvum, and R. salina were 0.05, 0.07, and 0.07 vvm, respectively (Vidyarathna et al. 2014, Xie et al. 2020). The production of chlorophytes Chlorella vulgaris, Chlorella sp., Haematococcus pluvialis (ZY-18), and Chlorella protothecoides was enhanced under mild aeration (Table 1). The airflow rates of aeration by air pump devices supporting maximum production of C. vulgaris, Chlorella sp., H. pluvialis (ZY-18), and C. protothecoides were 0.2, 0.2, 0.2, and 0.3 vvm, respectively (Han et al. 2015, Ding et al. 2021, Liao et al. 2023). In addition, the production of H. pluvialis (CDBB-052) and of the chlorophyte Scenedesmus quadricauda was enhanced under high aeration (Table 1). Notably, the airflow rates of aeration supporting maximum production of H. pluvialis (CDBB-052) and S. quadricauda were 0.5 and 0.6 vvm, respectively (Vega-Estrada et al. 2005, Tanh et al. 2015).

Without aeration, the dry weights (mg L⁻¹) of total fatty acids and omega-3 fatty acids of G. smaydae + H. rotundata in periods 1–3 were similar, and the ratio of omega-3 relative to total fatty acids did not change during these three periods. Therefore, the 100-L system established in the current study yields consistent fatty acid compositions and contents of G. smaydae without aeration.

In the present study, when G. smaydae cells were harvested at a rotation speed of 16,000 rpm, the harvesting efficiency was 10, 56, and 185% higher than that at 8,000, 4,000, and 2,000 rpm, respectively. Similarly, harvesting efficiencies at a high rotation speed of 7,000 rpm or 13,000 ×g for the haptophyte Pavlova lutheri, chlorophytes Chlorella sp. (UKM2), Coelastrella sp. (UKM4), Tetraselmis chui, and Chlamydomonas sp. (UKM6), ochrophyte Nannochloropsis oculata, and diatoms Thalassiosira pseudonana, Chaetoceros muelleri, Chaetoceros calcitrans, and Skeletonema costatum were 46–1,900% higher than those at a low rotation speed of 1,000 rpm or 1,300 ×g (1,000 and 7,000 rpm in Japar et al. 2017; 1,300 and 13,000 g in Heasman et al. 2000) (Table 2). However, the harvesting efficiencies of Scenedesmus sp. and thraustochytrid Aurantiochytrium sp. (KRS101), were not affected by rotational speed (Heasman et al. 2000, Kim et al. 2015, Reyes and Labra 2016) (Table 2). Therefore, the harvesting efficiency of some microalgae, such as G. smaydae, is affected by centrifugation speed, whereas that of other microalgae is not.

The energy consumption (kWh or kWh m⁻³) of centrifuges operated at high rotation speeds is typically higher than that at low speeds (Szepessy and Thorwid 2018, Abu-Shamleh and Najjar 2020, Najjar and Abu-Shamleh 2020). In general, the energy consumption of rotating centrifuges increases with increasing rotation speed, which increases energy costs. In addition, energy consumption may depend on other factors, such as power for each motor mounted on centrifuges, the density and size of microalgal cells, and the flow rates of pumps for injection of microalgae cultures into centrifuges (Abu-Shamleh and Najjar 2020). Furthermore, energy costs depend on energy consumption as well as electricity prices ($ kWh⁻¹) in each country (Krishnamurthy and Kriström 2015). Consequently, the optimal rotation speed of a centrifuge for harvesting the targeted microalgal cells should be considered in terms of both harvesting efficiency and energy costs.

The results of the present study show that the harvesting efficiency of G. smaydae cells at the same rotation speed of the centrifuge was not affected by the flow rate of the injection pump transporting them from the predator storage tank to the centrifuge bowl. However, in contrast to our results, in a previous study that harvested cells of the chlorophyte Nannochloris sp. at different injection flow rates, the highest and lowest harvesting efficiencies (94 and 17%) were obtained at the lowest and highest flow rates (0.94 and 23.2 L min⁻¹), respectively (Dassey and Theegala 2013). Therefore, whether the flow rate of an injection pump affects the harvesting efficiency of microalgae depends on the target species. Energy consumption for higher flow rates of an injection pump is likely to be greater than that at a lower flow rate at a given time; however, at a given culture volume to be harvested, the time to operate the injection pump at a higher flow rate is shorter than that at a lower flow rate (Najjar and Abu-Shamleh 2020). Therefore, the optimal injection flow rate from the storage tank to the centrifuge bowl for harvesting microalgal cells should be investigated to enhance harvesting efficiency and reduce energy costs.