Experimental Oxygen Mass Transfer Study of Micro-Perforated Diffusers

: We studied new micro-perforated diffuser concepts for the aeration process in wastewater treatment plants and evaluated their aeration efﬁciency. These are micro-perforated plate diffusers with oriﬁce diameters of 30 µ m, 50 µ m and 70 µ m and a micro-perforated tube diffuser with an oriﬁce diameter of 50 µ m. The oxygen transfer of the diffuser concepts is tested in clean water, and it is compared with commercial aerators from the literature. The micro-perforated tube diffuser and micro-perforated plate diffusers outperform the commercial membrane diffusers by up to 44% and 20%, respectively, with regard to the oxygen transfer efﬁciency. The most relevant reason for the improved oxygen transfer is the ﬁne bubble aeration with bubble sizes as small as 1.8 mm. Furthermore, the more homogenous cross-sectional bubble distribution of the micro-perforated tube diffuser has a beneﬁcial effect on the gas mass transfer due to less bubble coalescence. However, the pressure drop of micro-perforated diffusers seems to be the limiting factor for their standard aeration efﬁciencies due to the size and the number of oriﬁces. Nevertheless, this study shows the potential for better aeration efﬁciency through the studied conceptual micro-perforated diffusers.


Introduction
Gas dispersion in liquids is crucial for a large number of multiphase reactions in chemical and biochemical processes. In general, this is a highly energy-intensive process. One example is the gas dispersion in wastewater treatment, where high amounts of air are dispersed into large tanks for activated sludge aeration or ozonation for contamination removal. This work focuses on the air dispersion, which is needed in the biological nitrification process of wastewater treatment. In the activated sludge process, the air is used for the microbial degradation of nitrogen compounds, especially ammonia. There, the air is compressed by blowers and injected through several diffusers from the bottom of the basin of typically 4 m-6 m depth. Moreover, the aeration promotes suspension of the activated sludge and mixing for improved contacting of microorganisms with organic matter. This is the most energy-intensive process step in the activated sludge wastewater treatment. According to the International Energy Agency (IEA), wastewater treatment plants (WWTPs) are responsible for about 1% of the total global electricity consumption in the water sector and IEA 2016. The biological wastewater treatment is accountable for the majority of the energy consumption [1]. Aeration is responsible for more than 50% of the overall electric energy consumption of a treatment plant operating with activated sludge [2]. Because of the high electricity consumption of sludge aeration, improvements are needed to reduce the energy consumption of the aeration process without diminishing contaminant removal during the aeration process. Producing small

Bubble Size Measurement
The bubble size investigation was conducted in an acrylic glass column with a crosssection of 400 mm × 400 mm and 300 mm height filled with deionized water. Bubble size from various diffusers was measured using videometry with a backlight technique. To avoid overlapping of bubbles, one row of orifices per diffuser was operated in isolation. A high-speed camera from Vision Research, Inc. model VEO 710L was used together with a microscope lens, Model K2 DistaMaxTM from Infinity Photo-Optical Company. The optical system has a spatial resolution as small as 2 µm/pix and a temporal resolution as small as 25 µs, and the exposure time was set to 8 µs. The backlight was supplied by a 200 W pulsed LED light source from Veritas light model constellation 120E. The final bubble volume was determined using a proprietary image processing algorithm developed by Ziegenhein [16]. A detailed explanation of each step is provided in Mohseni et al. [15]. Eventually, the Sauter mean diameter d 32 , also known as the surface-volume mean, was reported as the representative mean value of the bubble diameter.

Experimental Setup for Oxygen Transfer
Oxygen absorption experiments for various diffuser types were performed in a pilotscale test facility. The facility comprises two instrumented activated sludge bubble columns that have a maximum aeration depth of 4 m and an inner diameter of 900 mm (see Figure 1). Deionized water (σ < 10 µS/cm) was used as a reference in the absorption experiments. The airflow was controlled using mass flow controllers (up to 250 slpm, standard conditions: 1.01325 bar at 25 • C) from Omega Engineering Inc. The range of the volumetric air flow rate was kept between 1 m 3 ·h −1 and 7 m 3 ·h −1 for clean water. For the investigation of the oxygen absorption, the columns were equipped with dissolved oxygen sensors. A submersible relative pressure sensor measured the pressure drop in the gas feed line upstream of the diffusers.

Bubble Size Measurement
The bubble size investigation was conducted in an acrylic glass column with a crosssection of 400 mm × 400 mm and 300 mm height filled with deionized water. Bubble size from various diffusers was measured using videometry with a backlight technique. To avoid overlapping of bubbles, one row of orifices per diffuser was operated in isolation. A high-speed camera from Vision Research, Inc. model VEO 710L was used together with a microscope lens, Model K2 DistaMaxTM from Infinity Photo-Optical Company. The optical system has a spatial resolution as small as 2 µ m/pix and a temporal resolution as small as 25 µ s, and the exposure time was set to 8 μs. The backlight was supplied by a 200 W pulsed LED light source from Veritas light model constellation 120E. The final bubble volume was determined using a proprietary image processing algorithm developed by Ziegenhein [16]. A detailed explanation of each step is provided in Mohseni et al. [15]. Eventually, the Sauter mean diameter d 32 , also known as the surface-volume mean, was reported as the representative mean value of the bubble diameter.

Experimental Setup for Oxygen Transfer
Oxygen absorption experiments for various diffuser types were performed in a pilotscale test facility. The facility comprises two instrumented activated sludge bubble columns that have a maximum aeration depth of 4 m and an inner diameter of 900 mm (see Figure 1). Deionized water (σ < 10 µ S/cm) was used as a reference in the absorption experiments. The airflow was controlled using mass flow controllers (up to 250 slpm, standard conditions: 1.01325 bar at 25 °C) from Omega Engineering Inc. The range of the volumetric air flow rate was kept between 1 m 3 ·h −1 and 7 m 3 ·h −1 for clean water. For the investigation of the oxygen absorption, the columns were equipped with dissolved oxygen sensors. A submersible relative pressure sensor measured the pressure drop in the gas feed line upstream of the diffusers.  The aeration efficiency of the micro-perforated plate and tube diffusers were compared with commercially available diffusers in wastewater treatment plants with disc, tube and plate shapes. The diffusers were tested under the same conditions while placed at the  Table 1 provides the characteristics of the studied diffusers. The micro-perforated plate diffusers were manufactured using a laser drilling technique. A single micro-perforated plate diffuser had an active aeration area of 0.01 m 2 . To achieve an active aeration area comparable to the one from commercial diffusers, a set of four micro-perforated plate diffusers with the same orifice diameters were installed together in the columns (see Figure 2). The aeration efficiency of the micro-perforated plate and tube diffusers were compared with commercially available diffusers in wastewater treatment plants with disc, tube and plate shapes. The diffusers were tested under the same conditions while placed at the bottom of the DN 900 columns. Table 1 provides the characteristics of the studied diffusers. The micro-perforated plate diffusers were manufactured using a laser drilling technique. A single micro-perforated plate diffuser had an active aeration area of 0.01 m 2 . To achieve an active aeration area comparable to the one from commercial diffusers, a set of four micro-perforated plate diffusers with the same orifice diameters were installed together in the columns (see Figure 2).  The micro-perforated tube diffuser had an orifice diameter of 50 µ m. The tubes were made out of polyamide. A laser drilling technique was used for manufacturing the orifices of the tubes. Figure 3 shows a scheme and a close-up picture of this tube diffuser concept. Moreover, the tube diffuser provides a more homogenous cross-sectional distribution of the gas phase in favor of limited bubble coalescence when compared to other diffusers while maintaining a high orifice density. The tubes were installed within a frame for the tests in the experimental facility (see Figure 3). In the latter, each tube was sealed from one side while there was a gas feed from the other side. The lateral surface of the cylindricalshaped tubes embodied the active aeration area for one micro-perforated tube. Including The micro-perforated tube diffuser had an orifice diameter of 50 µm. The tubes were made out of polyamide. A laser drilling technique was used for manufacturing the orifices of the tubes. Figure 3 shows a scheme and a close-up picture of this tube diffuser concept. Moreover, the tube diffuser provides a more homogenous cross-sectional distribution of the gas phase in favor of limited bubble coalescence when compared to other diffusers while maintaining a high orifice density. The tubes were installed within a frame for the tests in the experimental facility (see Figure 3). In the latter, each tube was sealed from one side while there was a gas feed from the other side. The lateral surface of the cylindrical-shaped tubes embodied the active aeration area for one micro-perforated tube. Including twelve twelve micro-perforated tubes, the diffuser had an active aeration area of 0.104 m 2 , which is similar to commercial diffusers.

Oxygen Mass Transfer Measurement
Oxygen transfer was measured in clean water using an absorption measurement technique, and the specific parameters were calculated with Equations (1)-(5) according to DWA-M 209 [17]. The measurements were conducted as batch experiments based on the non-steady-state mode. First, nitrogen gas was supplied for the removal of oxygen from the liquid phase. Subsequently, re-oxygenation with air at a certain flow rate was provided until the oxygen saturation concentration in the liquid phase (CS,p*,T) was reached. The dissolved oxygen concentration was monitored with an oxygen sensor during the whole process.
The temporal oxygen concentration Ct can be expressed with with the volumetric mass transfer coefficient kLaT, the dissolved oxygen saturation concentration CS,p*,T for the experimental conditions and the initial dissolved oxygen concentration C0. Based on the measured values of C0 and CS,p*,T, kLaT was determined using (1) as a nonlinear regression model. The determined kLaT and CS,p*,T values had to be adjusted to standard conditions in the following way: in order to enable a comparison of different experimental conditions. SOTR was calculated from kLa20, and the tank volume (V) was calculated according to SOTR was then used to compare the oxygen transfer of various diffusers over a wide range of volumetric gas flow rates. SAE was used for comparison in terms of power efficiency since it shows how much power input is necessary to dissolve 1 kg of oxygen in water. This measurement considers the required electrical power and was calculated with

Oxygen Mass Transfer Measurement
Oxygen transfer was measured in clean water using an absorption measurement technique, and the specific parameters were calculated with Equations (1)-(5) according to DWA-M 209 [17]. The measurements were conducted as batch experiments based on the non-steady-state mode. First, nitrogen gas was supplied for the removal of oxygen from the liquid phase. Subsequently, re-oxygenation with air at a certain flow rate was provided until the oxygen saturation concentration in the liquid phase (C S,p*,T ) was reached. The dissolved oxygen concentration was monitored with an oxygen sensor during the whole process.
The temporal oxygen concentration C t can be expressed with with the volumetric mass transfer coefficient k L a T , the dissolved oxygen saturation concentration C S,p*,T for the experimental conditions and the initial dissolved oxygen concentration C 0 . Based on the measured values of C 0 and C S,p*,T , k L a T was determined using (1) as a nonlinear regression model. The determined k L a T and C S,p*,T values had to be adjusted to standard conditions in the following way: in order to enable a comparison of different experimental conditions. SOTR was calculated from k La20 , and the tank volume (V) was calculated according to SOTR was then used to compare the oxygen transfer of various diffusers over a wide range of volumetric gas flow rates. SAE was used for comparison in terms of power efficiency since it shows how much power input is necessary to dissolve 1 kg of oxygen in water. This measurement considers the required electrical power and was calculated with The experimental setup was connected to the in-house pressurized air supply. Therefore, it was necessary to calculate the equivalent power (P) that would be needed for the air compression at the specific flow rates.
In this case, P was acquired using from Pöpel and Wagner and Pöpel et al. [18,19]. The equations were derived from manufactures' data for three different blower and compressor types. For the experimental conditions in this study, we used the positive displacement blower. Hence, a specific energy E 0 of 4.3 (Wh·m −3 ·m) and an exponent Y = 1 were used. Figure 4 illustrates the trend of d 32 with regard to the volumetric gas flow rate per opening (q) from the diffusers in this study and the available data from the literature. Among our own data, MT1 generates the largest bubbles with a slight increase in d 32 with q. This is followed by MP3-MP1, which show a similar trend to that of MT1. Although the orifice diameter of MP3 is bigger than that of MT1, the latter generates larger bubbles. A plausible reason behind this observation may be the difference in the surface wettability of polyamide pipes and stainless steel diffusers, orifice orientation and the deformation in the geometry of the orifices during the laser drilling process. Among the data from the literature, the diffusers used by Amaral et al. [10] and Behnisch et al. (D2) [11] generate the smallest bubbles. It should be noted that the range of volumetric gas flow rate per orifice q used by Amaral et al. is below the smallest one in this study [10]. Moreover, Behnisch et al. only measured the bubbles from the side of the spargers to avoid the overlapping of bubbles [11]. The values reported by Hasanen et al. were measured using a suction probe at several points across the diffuser [9]. The experimental setup was connected to the in-house pressurized air supply. Therefore, it was necessary to calculate the equivalent power (P) that would be needed for the air compression at the specific flow rates.

Bubble Size Measurements
In this case, P was acquired using from Pöpel and Wagner and Pöpel et al. [18,19]. The equations were derived from manufactures' data for three different blower and compressor types. For the experimental conditions in this study, we used the positive displacement blower. Hence, a specific energy E0 of 4.3 (Wh·m −3 ·m) and an exponent = 1 were used. Figure 4 illustrates the trend of d32 with regard to the volumetric gas flow rate per opening (q) from the diffusers in this study and the available data from the literature. Among our own data, MT1 generates the largest bubbles with a slight increase in d32 with q. This is followed by MP3-MP1, which show a similar trend to that of MT1. Although the orifice diameter of MP3 is bigger than that of MT1, the latter generates larger bubbles. A plausible reason behind this observation may be the difference in the surface wettability of polyamide pipes and stainless steel diffusers, orifice orientation and the deformation in the geometry of the orifices during the laser drilling process. Among the data from the literature, the diffusers used by Amaral et al. [10] and Behnisch et al. (D2) [11] generate the smallest bubbles. It should be noted that the range of volumetric gas flow rate per orifice used by Amaral et al. is below the smallest one in this study [10]. Moreover, Behnisch et al. only measured the bubbles from the side of the spargers to avoid the overlapping of bubbles [11]. The values reported by Hasanen et al. were measured using a suction probe at several points across the diffuser [9].  [11]). Figure 5 shows the results of SSOTE as a function of air flow rate per diffuser area of disc diffusers (QA,DA) as well as the correlation for the disc diffusers derived by Jolly et al. [13]. According to DWA, this study covers the typical QA,DA operational ranges of 35 Sm 3 ·h −1 ·m −2 -150 Sm 3 ·h −1 ·m −2 for disc diffusers and 5 Sm 3 ·h −1 ·m −2 -60 Sm 3 ·h −1 ·m −2 for plate  Figure 5 shows the results of SSOTE as a function of air flow rate per diffuser area of disc diffusers (Q A,DA ) as well as the correlation for the disc diffusers derived by Jolly et al. [13]. According to DWA, this study covers the typical Q A,DA operational ranges of 35 Sm 3 ·h −1 ·m −2 -150 Sm 3 ·h −1 ·m −2 for disc diffusers and 5 Sm 3 ·h −1 ·m −2 -60 Sm 3 ·h −1 ·m −2 for plate diffusers [20]. The results for the standard disc diffusers from this study are in good agreement with those of Jolly et al. [13]. However, the micro-perforated tube diffuser (MT1) accomplishes more than 14%·m −1 SSOTE at the lowest Q A,DA and 9%·m −1 at the diffusers [20]. The results for the standard disc diffusers from this study are in good agreement with those of Jolly et al. [13]. However, the micro-perforated tube diffuser (MT1) accomplishes more than 14%·m −1 SSOTE at the lowest QA,DA and 9%·m −1 at the highest QA,DA. For the best SSOTE of the disc diffuser DD1 with 7% m −1 for the lowest QA,DA and 5%·m −1 for the highest QA,DA (DD4), the MT1 outperforms these values by 50% and 44%, respectively. Taking into account the literature findings of Jolly et al., this represents an enhancement in the SSOTE of 20% at the highest QA,DA and an improvement of 33% for the lowest QA,DA of MT1 [13]. According to Figure 5, MT1 exceeds the SSOTE of the membrane tube diffusers for all measured flow rates. For the lowest flow rate, MT1 exceeds the SSOTE of 6.18%·m −1 from TD1 by more than 55%. For the highest QA,DA, MT1 shows an improvement in the SSOTE of more than 46% compared to the 4.85%·m −1 of TD1. Regarding SSOTE, the diffusers MP1-3 are in good agreement with the values reported by Behnisch et al. and Jolly et al. [13,14]. The SSOTE values of MP1-3 are slightly above those of PD1. With 7.40%·m −1 at the lowest QA,DA, the SSOTE of MP1 is 20% better compared to the SSOTE of PD1.

Oxygen Transfer Efficiency
MT1 shows improvements for all QA,DA compared to all other diffusers. This is believed to be due to the generated bubble size, which results in a better mass transfer and, therefore, better oxygen transfer efficiency. With a reduction in bubble size, the partial pressure of the dissolved gas component increases, and the gas dissolves easier. This partial pressure is the driving force for the gas dissolution. Moreover, smaller bubbles have a lower bubble rising velocity and, therefore, a higher residence time. Among the studied diffusers, only MT1 exceeds the favorable SSOTE range of 8.5%·m −1 -9.8%·m −1 , especially for QA,DA < 35 Sm 3 ·h −1 ·m −2 [14]. According to Figure 5, MT1 exceeds the SSOTE of the membrane tube diffusers for all measured flow rates. For the lowest flow rate, MT1 exceeds the SSOTE of 6.18%·m −1 from TD1 by more than 55%. For the highest Q A,DA , MT1 shows an improvement in the SSOTE of more than 46% compared to the 4.85%·m −1 of TD1. Regarding SSOTE, the diffusers MP1-3 are in good agreement with the values reported by Behnisch et al. and Jolly et al. [13,14]. The SSOTE values of MP1-3 are slightly above those of PD1. With 7.40%·m −1 at the lowest Q A,DA , the SSOTE of MP1 is 20% better compared to the SSOTE of PD1.
MT1 shows improvements for all Q A,DA compared to all other diffusers. This is believed to be due to the generated bubble size, which results in a better mass transfer and, therefore, better oxygen transfer efficiency. With a reduction in bubble size, the partial pressure of the dissolved gas component increases, and the gas dissolves easier. This partial pressure is the driving force for the gas dissolution. Moreover, smaller bubbles have a lower bubble rising velocity and, therefore, a higher residence time. Among the studied diffusers, only MT1 exceeds the favorable SSOTE range of 8.5%·m −1 -9.8%·m −1 , especially for Q A,DA < 35 Sm 3 ·h −1 ·m −2 [14].
It should be noted that this evaluation does not include the effect of diffuser density. A higher diffuser density enhances SSOTE, especially for low gas flow rates [14]. Thus, the  [14].
MP1-3 achieve SSOTE values comparable to those of the tested plate diffuser PD1, although no enhancement in mass transfer is observed. Nevertheless, the SSOTE of the micro-perforated plate diffuser is better than that of PD with less than one-half of the diffuser surface area and a lower number of orifices and orifice density, respectively. A lower orifice density means a wider distance between the orifices and, thereby, a more homogenous cross-sectional bubble distribution [7]. Therefore, bubbles will remain smaller because of less bubble coalescence and, consequently, preserve their small size. Moreover, a wider bubble distribution provides a more homogenous gas concentration gradient across the column. This avoids local degradation in the concentration gradients due to the creation of bubble plume. Because of the low orifice density and due to the novel shape, MT1 has a more homogenous spatial bubble distribution compared to other diffusers, including the micro-perforated plate diffusers. We observed that MT1 reaches oxygen saturation in the liquid phase quicker than all other tested diffusers. This is important in terms of gas transfer efficiency with respect to the same gas flow rates for all tested diffusers. Figure 6 depicts the variation in SOTR per aerated tank volume, SOTR ATV, as a function of air flow rate per aerated tank volume, Q A,ATV , for the various diffusers. The investigated disc diffusers achieved similar SOTR ATV values, and the results show a good linear correlation as well as an agreement with Behnisch et al. [14]. MT1 achieved SOTR ATV ≈ 130 gO 2 ·m −3 ·h −1 with a Q A,ATV =1.1 Sm 3 ·m tv −3 ·h −1 , whereas all tested disc diffusers needed more than 2.20 Sm 3 m tv −3 h −1 to reach a comparable SOTR ATV . For the given experimental facility, this would represent an improvement in the Q A,ATV of about 50%. According to Behnisch et al., a Q A,ATV of 1.65 Sm 3 ·m tv −3 ·h −1 would be necessary to have an SOTR ATV of 130 gO 2 ·m −3 ·h −1 [14]. Compared to the value of MT1, this would still be an improvement of 33% for the required Q A;ATV . Moreover, to reach a favorable SOTR ATV of 120 gO 2 m −3 h −1 over peak loads, a Q A,ATV of 0.97 Sm 3 ·m tv −3 ·h −1 is needed for MT1. This number is 35% below the suggested value of 1.50 Sm 3 ·m tv −3 ·h −1 [14].

Oxygen Transfer Rate
The tube diffusers have a comparable SOTR ATV , and the results are in agreement with the literature and illustrate a linear correlation. For an SOTR ATV of about 130 gO 2 ·m −3 ·h −1 , TD1 and MT1 need Q A,ATV = 2.65 Sm 3 ·m TV −3 ·h −1 and Q A,ATV = 1.10 Sm 3 ·m TV −3 ·h −1 , respectively. This is an enhancement of nearly 60% for MT1 compared to TD1 in terms of Q A,ATV . Moreover, the SOTR ATV values of MP1-3 do not have a wide distribution. PD1 and MP1-3 would reach an SOTR ATV of 120 gO 2 ·m −3 ·h −1 over peak loads with Q A,ATV ≈ 1.95 Sm 3 ·m TV −3 ·h −1 . The latter is 50% more than the Q A,ATV for the same SOTR ATV of MT1. For SOTR ATV = 130 gO 2 ·m −3 ·h −1 , PD1 and MP1-3 require Q A,ATV ≈ 2.10 Sm 3 m tv −3 h −1 , where MT1 achieves a similar SOTR ATV with Q A,ATV ≈ 1.008 Sm 3 ·m tv −3 ·h −1 .
Based on the observations of SOTR ATV , the calculation of SSOTE is possible. In this case, the micro-perforated diffusers accomplish better results compared to the membrane diffusers. Moreover, the micro-perforated diffusers could surpass the recommended SOTR ATV of 120 gO 2 ·m −3 ·h −1 [14]. The reason is a homogenous cross-sectional bubble distribution and, consequently, less bubble coalescence as well as the generation of small bubbles. Accordingly, k L a improves and, as a result, SOTR improves with it.  Figure 7 shows the pressure drop Δp of different types of diffusers. Since the Δp of diffusers from the same category is similar, one diffuser from each diffuser group is presented in Figure 7. Clearly, there is a difference in Δp values between the chosen disc, tube and the plate diffusers compared to MT1 and MP1. For MT1 and MP1, Δp at QA = 1 Sm 3 ·h −1 are 43 hPa and 65 hPa, respectively. However, PD1 and TD1 have a Δp of 28 hPa and 46 hPa at QA = 7 Sm 3 h −1 . The maximum Δp among all diffusers is recorded for MT1 at QA = 4.2 Sm 3 ·h −1 with a value of 286 hPa. Regarding MP3 and DD3, both diffusers show a comparable Δp. From Figure 7, it can be concluded that the orifice size and total orifice number have a significant effect on Δp. For similar QA, less orifices have to cope with the same gas volume. Therefore, the volumetric gas flow rate from these orifices has to increase, which causes more flow resistance and eventually an increase in Δp. MP1 has the smallest orifices among the other diffusers but, in total, more orifices than MP2 and MP3. However, the total orifice surface area of MP1, i.e., AT,O = 8.2 mm 2 , is less than the total orifice surface area of MP2, i.e., AT,O = 11.8 mm 2 , and MP3, i.e., AT,O = 20 mm 2 . Consequently, Δp decreases with the change from MP1 to MP3. Moreover, this would explain the greater increase in the Δp of DD3 due to the smaller total surface area of the orifices compared to the other tested disc diffusers. This is also true when comparing the Δp of DD3 to the other plate and tube diffusers. Similarly, the high Δp of MT1 is related to the low AT,O. MT1 has the lowest AT,O = 7.2 mm 2 among all the investigated diffusers.  Figure 7 shows the pressure drop ∆p of different types of diffusers. Since the ∆p of diffusers from the same category is similar, one diffuser from each diffuser group is presented in Figure 7. Clearly, there is a difference in ∆p values between the chosen disc, tube and the plate diffusers compared to MT1 and MP1. For MT1 and MP1, ∆p at Q A = 1 Sm 3 ·h −1 are 43 hPa and 65 hPa, respectively. However, PD1 and TD1 have a ∆p of 28 hPa and 46 hPa at Q A = 7 Sm 3 h −1 . The maximum ∆p among all diffusers is recorded for MT1 at Q A = 4.2 Sm 3 ·h −1 with a value of 286 hPa. Regarding MP3 and DD3, both diffusers show a comparable ∆p. From Figure 7, it can be concluded that the orifice size and total orifice number have a significant effect on ∆p. For similar Q A , less orifices have to cope with the same gas volume. Therefore, the volumetric gas flow rate from these orifices has to increase, which causes more flow resistance and eventually an increase in ∆p. MP1 has the smallest orifices among the other diffusers but, in total, more orifices than MP2 and MP3. However, the total orifice surface area of MP1, i.e., A T,O = 8.2 mm 2 , is less than the total orifice surface area of MP2, i.e., A T,O = 11.8 mm 2 , and MP3, i.e., A T,O = 20 mm 2 . Consequently, ∆p decreases with the change from MP1 to MP3. Moreover, this would explain the greater increase in the ∆p of DD3 due to the smaller total surface area of the orifices compared to the other tested disc diffusers. This is also true when comparing the ∆p of DD3 to the other plate and tube diffusers. Similarly, the high ∆p of MT1 is related to the low A T,O . MT1 has the lowest A T,O = 7.2 mm 2 among all the investigated diffusers.  In the case of tube diffusers, MT1 has a higher SAE value compared to the fusers for QA ≤ 3 Sm 3 ·h −1 . Moreover, MT1 reaches a better SAE for QA ≤ 2.4 Sm 3 pared to PD1 and the MP(s). PD1 and MP3 reach a similar SAE of 4.35 kgO2·kW QA of 1.0 Sm 3 ·h −1 . In general, MT1 shows the best results for SAE in the air flow ra of 0.6 Sm 3 ·h −1 to 2.4 Sm 3 ·h −1 . MP1-3 attain a similar SAE compared to the investiga ventional plate diffuser.

Pressure Loss and Aeration Efficiency
All investigated membrane diffusers in Figure 8 have a decreasing trend for S an increase in QA. As previously explained, the micro-orifices lead to a higher Δp increasing QA compared to the membrane diffusers. Since SAE is calculated as th SOTRATV to P and Δp, the decrease in the SAE of MT1, MP1 and MP2 is mostly a to the Δp. Moreover, since SOTR does not decrease with QA, Δp remains as th influencing parameter for the decrease in SAE. One way to address this issue co increment the number of micro-perforated tubes and plates per m 2 coverage a would increase the number of orifices, resulting in a lower flow rate per orifice t to a lower Δp due to an enhanced AT,O.  In the case of tube diffusers, MT1 has a higher SAE value compared to the tube diffusers for Q A ≤ 3 Sm 3 ·h −1 . Moreover, MT1 reaches a better SAE for Q A ≤ 2.4 Sm 3 ·h −1 compared to PD1 and the MP(s). PD1 and MP3 reach a similar SAE of 4.35 kgO 2 ·kWh −1 for a Q A of 1.0 Sm 3 ·h −1 . In general, MT1 shows the best results for SAE in the air flow rate range of 0.6 Sm 3 ·h −1 to 2.4 Sm 3 ·h −1 . MP1-3 attain a similar SAE compared to the investigated conventional plate diffuser.
All investigated membrane diffusers in Figure 8 have a decreasing trend for SAE with an increase in Q A . As previously explained, the micro-orifices lead to a higher ∆p over an increasing Q A compared to the membrane diffusers. Since SAE is calculated as the ratio of SOTR ATV to P and ∆p, the decrease in the SAE of MT1, MP1 and MP2 is mostly attributed to the ∆p. Moreover, since SOTR does not decrease with Q A , ∆p remains as the major influencing parameter for the decrease in SAE. One way to address this issue could be to increment the number of micro-perforated tubes and plates per m 2 coverage area. This would increase the number of orifices, resulting in a lower flow rate per orifice that leads to a lower ∆p due to an enhanced A T,O . Energies 2021, 14, x FOR PEER REVIEW 11 of 14

Conclusions and Outlook
We studied the performance of micro-perforated diffusers with regard to oxygen transfer and compared them with commercially available membrane diffusers. Our assessments included various parameters, such as kLa, SOTR and SAE, in clean water with an aeration depth of 4 m in an instrumented pilot-scale test facility. Furthermore, we compared our results with the literature. The following conclusions can be drawn from our investigation:  Micro-perforated diffusers are able to deliver bubbles with 32 ≤ 2. In this case, the micro-perforated plate diffusers are able to generate smaller bubbles than those of the micro-perforated tube diffuser.  Regarding the results of the SSOTE, the micro-perforated tube diffuser outperforms the membrane diffusers with up to 50% and 44% higher SSOTE for the lowest flow rate and the highest flow rates, respectively. Additionally, the micro-perforated small tube diffuser surpasses the referenced trend lines by at least 20% and reaches the recommended SSOTE range of 8.5%·m −1 -9.8%·m −1 .  In the case of SOTRATV, the micro-perforated tube diffuser achieves 48% diminution in QA,ATV compared to the best accomplished results of the membrane aeration elements. In addition, the novel tube system results in a reduction of the QA,ATV of 33% and 35% compared to the literature trend line.  The reasons behind the good performance of the micro-perforated tube diffuser are believed to be the fine bubble aeration, with bubble sizes as small as 1.8 mm, and the homogenous spatial bubble distribution.

Conclusions and Outlook
We studied the performance of micro-perforated diffusers with regard to oxygen transfer and compared them with commercially available membrane diffusers. Our assessments included various parameters, such as k L a, SOTR and SAE, in clean water with an aeration depth of 4 m in an instrumented pilot-scale test facility. Furthermore, we compared our results with the literature. The following conclusions can be drawn from our investigation: • Micro-perforated diffusers are able to deliver bubbles with d 32 ≤ 2. In this case, the micro-perforated plate diffusers are able to generate smaller bubbles than those of the micro-perforated tube diffuser. Regarding the results of the SSOTE, the microperforated tube diffuser outperforms the membrane diffusers with up to 50% and 44% higher SSOTE for the lowest flow rate and the highest flow rates, respectively. Additionally, the micro-perforated small tube diffuser surpasses the referenced trend lines by at least 20% and reaches the recommended SSOTE range of 8.5%·m −1 -9.8%·m −1 .

•
In the case of SOTR ATV , the micro-perforated tube diffuser achieves 48% diminution in Q A,ATV compared to the best accomplished results of the membrane aeration elements. In addition, the novel tube system results in a reduction of the Q A,ATV of 33% and 35% compared to the literature trend line.

•
The reasons behind the good performance of the micro-perforated tube diffuser are believed to be the fine bubble aeration, with bubble sizes as small as 1.8 mm, and the homogenous spatial bubble distribution. • The novel tube system results in an improved SAE of up to 20% for Q A ≤ 2.4 Sm 3 ·h −1 . However, increasing Q A results in a decrease in SAE. The latter is shown to be mainly due to the enhanced ∆p as Q A increases. The resultant SAE of the micro-perforated plate diffusers with 50 µm and 70 µm orifices corresponds to the outcomes of the membrane plate diffuser, although ∆p for the micro-perforated plate diffusers is higher than the ∆p of the membrane plate diffuser.
Based on the current findings, we suggest Q A ≤ 3.0 Sm 3 ·h −1 as a preferred working range for micro-perforated tube and plate diffusers. Within this range, micro-perforated diffusers are able to achieve the required oxygen demand of 120 gO 2 ·m −3 ·h −1 with a lower or equal flow rate in comparison with state-of-the-art membrane diffusers. Moreover, we propose a more homogenous cross-sectional bubble distribution instead of a centered aeration surface. A better aeration efficiency due to a small bubble production and a wider bubble distribution would decrease the air consumption in the aeration process and, therefore, reduce the required power supply of the air blowers because of an improved gas transfer. Nevertheless, novel diffuser concepts with micro-perforated orifices should be investigated in experiments with wastewater to assess the effect of biofouling, the resilience of the material and the impact on the aeration efficiency. Further, different diffuser densities and, for the novel tube system, more tubes should be examined to investigate the effects on pressure loss.

Nomenclature
A Perforated Diffuser Area m 2 C 0 Dissolved oxygen concentration at t = 0 kg/m 3 C S, 20 Oxygen saturation concentration at T = 20 • C and p = 1013 hPa kg/m 3 C S,St, 20 Oxygen saturation concentration from EN 25814, C S,St,20 = 9.09 mg/L kg/m 3 C S,p*,T Oxygen saturation concentration at T and p* kg/m 3