Environmental and Economic Evaluation of Downﬂow Hanging Sponge Reactors for Treating High-Strength Organic Wastewater

: This study evaluated the performance of a downﬂow hanging sponge (DHS) in reducing the concentrations of chemical oxygen demand (COD), ammonia (NH 3 ), total suspended solids (TSS), and total dissolved solids (TDS) in high-strength organic wastewater (HSOW). The DHS unit was composed of three segments connected vertically and operated under different organic loading rates (OLRs) between 3.01 and 12.33 kg COD/m 3sponge /d at a constant hydraulic retention time (HRT) of 3.6 h. The results demonstrated that the DHS system achieved COD, NH 3 , TSS, and TDS removal efﬁciencies of 88.34 ± 6.53%, 64.38 ± 4.37%, 88.13 ± 5.42%, and 20.83 ± 1.78% at an OLR of 3.01 kg COD/m 3sponge /d, respectively. These removal efﬁciencies signiﬁcantly ( p < 0.05) dropped to 76.39 ± 6.58%, 36.59 ± 2.91%, 80.87 ± 5.71%, and 14.20 ± 1.07%, respectively, by increasing the OLR to 12.33 kg COD/m 3sponge /d. The variation in COD experimental data was well described by the ﬁrst-order ( R 2 = 0.927) and modiﬁed Stover–Kincannon models ( R 2 = 0.999), providing an organics removal constant (K 1 ) = 27.39 1/d, a saturation value constant (K B ) = 83.81 g/L/d, and a maximum utilization rate constant (U max ) = 76.92 g/L/d. Adding another DHS reactor in a secondary phase improved the ﬁnal efﬂuent quality, complying with most environmental regulation criteria except those related to TDS concentrations. Treating HSOW with two sequential DHS reactors was economically feasible, with total energy consumption of 0.14 kWh/m 3 and an operating cost of about 7.07 USD/m 3 . Accordingly, using dual DHS/DHS units to remove organics and nitrogen pollutants from HSOW would be a promising and cost-efﬁcient strategy. However, a tertiary treatment phase could be required to reduce the TDS concentrations.


Introduction
Huge amounts of high-strength organic wastewater (HSOW), with a chemical oxygen demand (COD) concentration greater than 1000 mg/L [1], have recently been produced owing to the considerable expansion of residential activities.These human activities include bathing, toileting, showering, and washing/grooming [2].In developing countries suffering from limited water resources, decreasing the amounts of flushing water utilized by conventional toilets tends to increase the black water quantities, characterized by high COD concentrations [3].Disposing of HSOW in the aquatic environment, without proper treatment, could result in dissolved oxygen depletion due to organic matter decomposition caused by microorganisms.Decreasing the dissolved oxygen content in water bodies below 2 mg/L negatively affects marine fish farming, water quality, and aquatic life [4].Treating HSOW becomes challenging because most conventional domestic wastewater treatment systems have been designed based on influent COD levels below 1000 mg/L [5].Furthermore, HSOW requires additional units to improve the performance of the wastewater treatment system so it can remove organic pollution and nutrients [6].Using a series of bioreactors to treat HSOW is costly, especially in developing countries needing a higher budget, more experience, and/or more trained operators [7].Hence, essential studies are required to find a cheap and proper HSOW treatment technique, providing a final effluent complying with continually expanding environmental regulations.
Aerobic biological systems, e.g., activated sludge processes, have been widely used to treat HSOW, where microorganisms utilize dissolved oxygen to convert organic wastes into CO 2 , H 2 O, and new cells [4].Although this biological-based aerobic approach maintains a high HSOW treatment performance, its operation generates large quantities of sludge [8].Moreover, these systems suffer from the consumption of excessive amounts of energy as an oxygen supply, limiting their practical application in low-income countries [9].On the other hand, anaerobic biological systems have also been implemented to treat HSOW [10].This biological-assisted process tends to degrade a large portion of organics into methane, CO 2, and H 2 O in the absence of oxygen [11].Moreover, the amount of sludge generated from the anaerobic-related systems is lower than that from the aerated units [4].Although anaerobic biological pathways are also used for energy production, they encounter multiple obstacles during real applications.These challenges include a slow bacterial growth rate, instability under fluctuating environmental conditions, and inefficiency at varying hydraulic and organic loading rates [12].In addition, anaerobic systems require a posttreatment process, especially if the effluent contains high concentrations of ammonium ion (NH 4 + ) and hydrogen sulfide (HS − ) [4].Moreover, anaerobic processes cannot fully stabilize HSOW, requiring a further treatment step to produce a final effluent complying with disposal standards [13].
The downflow hanging sponge (DHS) reactor has recently been employed to avoid most of the drawbacks of conventional biological wastewater treatment systems.The DHS technology was established based on the conventional trickling filters idea, where wastewater is uniformly distributed over a packing medium that holds active microorganisms [14].A sponge is used as the DHS medium, providing porosity greater than traditional rocks, gravel, or plastic pieces [15].The sponge material can carry a high mass of bacteria either inside the voids or on the surface, prolonging the solids retention time (SRT) operational factor [16].Polyurethane foam, with a porosity of about 90% and a specific surface area of 2400 m 2 /m 3 , is one of the sponge media used to retain large amounts of microorganisms [17].Because the sponge carrier is subjected to the atmosphere through natural ventilation openings, oxygen is spontaneously transferred into the influent wastewater.Accordingly, the DHS does not usually require external or artificial aeration facilities.Moreover, the DHS can remove a high portion of ammonium via autotrophic nitrifying bacteria grown under aerobic conditions [8].A DHS unit can also be constructed in an area relatively smaller than the physical footprint of other conventional wastewater treatment methods [18].Furthermore, due to the wide microbial diversity occupying the aerobic and anoxic zones of the sponge, higher biomass concentrations and small amounts of generated sludge can be monitored during the DHS's operation.In this context, the DHS system is simple to operate, has cheap treatment costs, and has little sludge discharge, making it a proper treatment method for low-and medium-income countries.
The organic loading rate (OLR) represents the daily organic mass, expressed by COD, introduced to a unit volume (m 3 ) of a bioreactor [19].Increasing the value of this operational factor reduces the bioreactor's performance in decomposing organic matter, oxidizing ammonia, and eliminating pathogens [20].Moreover, at higher OLRs, the heterotrophs and nitrifiers tend to compete for the available dissolved oxygen (electron acceptor) in the bioreactor, affecting the final effluent quality [21].Sequential wastewater treatment units could be implemented to withstand repeated organic load shocks.Hence, monitoring the performance of a DHS subjected to higher OLR levels would be an interesting research point.
First-order kinetic and modified Stover-Kincannon models have recently been employed to estimate the substrate bio-removal rate of wastewater treatments [10].Due to their simplicity of application, first-order kinetic models have been broadly used to evaluate the ability of most biological wastewater treatment processes to remove COD and nitrogen [22].The modified Stover-Kincannon model was established by Kincannon and Stover to determine the correlation between the substrate consumption rate and the OLR of a biofilm reactor without taking biomass growth into consideration [23].The kinetic constants of this model are beneficial in estimating the bioreactor's performance in removing carbonaceous organic matter and nutrients during wastewater treatment [24].Moreover, this model can predict effluent COD concentrations in continuous and semicontinuous bioreactors.Various nonlinear regression techniques have been employed to obtain the substrate utilization constant by fitting the kinetic models to the experimental data [10].The estimated kinetic rates are suitable for designing a bioreactor and predicting its treatment performance under both anaerobic and aerobic conditions [22].For a moving bed biofilm reactor treating synthetic starch wastewater, the kinetic parameters estimated from the modified Stover-Kincannon model could be used to predict the reactor's behavior on various scales [25].Kinetic models have also been studied extensively for a wide range of wastewater types, such as slaughterhouse wastewater [26], sugar-industry wastewater [27], fruit [24], and synthetics [22,28].Hence, the derived kinetic constants could also be used to evaluate a DHS's performance when the treatment system is operated under different OLRs.
Based on a comprehensive literature search of original articles published over the previous decade, a DHS system was used to treat domestic wastewater with COD concentrations below 1000 mg/L [29][30][31].Hence, assessing the DHS's performance in treating domestic wastewater with higher COD levels, equivalent to OLRs greater than 3 kg COD/m 3 sponge /d, would be an interesting topic to investigate.The current study objectives are fivefold: (1) evaluating the DHS performance for COD, NH 3 , and solids removal when the imposed OLRs reach 12.33 kg COD/m 3 sponge /d; (2) investigating the applicability of using two sequential DHS reactors to treat HSOW; (3) understanding the organic degradation performance by estimating the substrate utilization kinetics; (4) monitoring biomass attachment on the sponge using surface morphology and structural characterization; and (5) determining the economic feasibility of treating HSOW using the DHS technology.

Wastewater Characteristics
Synthetic wastewater was prepared using CH 3 COONa•3H 2 O (260 mg/L), MgSO   [32].Trace elements [32] were added, and the solutions were thickened to prepare different phases, varying COD concentrations from 450 to 1850 mg/L.Six phases (Table 1) were prepared to mimic a real-time domestic wastewater composition with different organic strengths.The highest COD concentration at phase 6 was used to investigate the applicability of the biological system in treating HSOW and determine the effluent water quality compared with the environmental standard.Moreover, the DHS bioreactor was operated under a wide nutrient condition (C:N ratio of 3.5:1 to 14.5:1).The NH 3 concentration was maintained so it was approximately constant to provide a reliable comparison between the 6 phases based on the decomposition of carbonaceous organic matters (i.e., the effect of varying the nutrient concentrations on the DHS's performance is outside the scope of this study).

Downflow Hanging Sponge (DHS) Reactor Configuration
Figure 1 shows a graphical layout of the pilot-scale DHS unit used for HSOW treatment.The DHS reactor was composed of three segments connected vertically, where wastewater flows downward from segment-1 to segment-3.It had a cylindrical shape with a 1.5 m height and a 0.15 m diameter and was constructed from polyvinyl chloride (PVC).Each segment had a 0.4 m height, and the space between every two segments was 0.1 m.This gap was used to facilitate the diffusion of natural air from the surrounding atmosphere to the DHS interior zone.About 270 sponge pieces (sponge dimensions: 3.3 cm diameter and 3.3 cm height) were randomly placed in the DHS unit, providing a sponge volume of 6.95 L. The sponge element was manufactured from polyurethane and covered by a spherical net made of polypropylene plastic.The sponge material had a density of 30 kg/m 3 , a specific surface area of 256 m 2 /m 3 , 90% porosity, and a pore size of 0.63 mm.A peristaltic pump (Masterflex, Cole-Parmer, Vernon Hills, IL, USA) with a variable discharge flow was used to feed the reactor.The influent wastewater was distributed homogenously using a spray dispenser with a perforated disc fixed to the DHS's top.A settling tank (0.15 m in height × 0.2 m in diameter) was installed at the DHS's bottom to allow for solid/liquid separation.

Downflow Hanging Sponge (DHS) Reactor Configuration
Figure 1 shows a graphical layout of the pilot-scale DHS unit used for HSOW treatment.The DHS reactor was composed of three segments connected vertically, where wastewater flows downward from segment-1 to segment-3.It had a cylindrical shape with a 1.5 m height and a 0.15 m diameter and was constructed from polyvinyl chloride (PVC).Each segment had a 0.4 m height, and the space between every two segments was 0.1 m.This gap was used to facilitate the diffusion of natural air from the surrounding atmosphere to the DHS interior zone.About 270 sponge pieces (sponge dimensions: 3.3 cm diameter and 3.3 cm height) were randomly placed in the DHS unit, providing a sponge volume of 6.95 L. The sponge element was manufactured from polyurethane and covered by a spherical net made of polypropylene plastic.The sponge material had a density of 30 kg/m 3 , a specific surface area of 256 m 2 /m 3 , 90% porosity, and a pore size of 0.63 mm.A peristaltic pump (Masterflex, Cole-Parmer, Vernon Hills, IL, USA) with a variable discharge flow was used to feed the reactor.The influent wastewater was distributed homogenously using a spray dispenser with a perforated disc fixed to the DHS's top.A settling tank (0.15 m in height × 0.2 m in diameter) was installed at the DHS's bottom to allow for solid/liquid separation.

Experimental Setup
Initially, the sponge carriers were soaked in a tank containing activated sludge collected from a nearby wastewater treatment plant.This step was used to reduce the time required for the bacteria culture to acclimatize to the HSOW [2].The DHS system was operated for one month (adaptation period) using the prepared synthetic wastewater.The water quality of the effluent was monitored to determine the steady-state condition.At the end of the adaptation phase, stable COD and NH 3 concentrations (within 6% variation for the triplicate readings) were observed.During the 150-day experimental period, the applied OLRs increased gradually over six operational phases (3.01, 4.22, 6.44, 8.36, 10.28, and 12.33 kg COD/m 3 sponge /d) at a constant hydraulic retention time (HRT) of 3.6 h.Synthetic wastewater was fed to the DHS's top at a flow rate of 46.34 L/d, controlled by a peristaltic pump.The influent wastewater was then trickled through the reactor via gravity flow, and the DHS unit was operated at 24-34 • C.

Analytical Analysis
The influent and effluent concentrations of COD, TDS, TSS, and NH 3 were determined based on the "Standard Method for Examination of Water and Wastewater" by the American Public Health Association [33].The wastewater sample was passed through a glass fiber filter (0.45 µm) to measure TSS.All concentrations were measured in triplicate (n = 3) and reported as average ± standard deviation (SD).The amount of sludge occupying the DHS unit was obtained by compressing and squeezing the sponge, followed by oven drying for 24 h [33].The crystallinity index (CrI) of the sponge before and after treatment was determined using the X-ray diffractometer technique (XRD-6100, Shimadzu, Japan), following Segal et al. [34].The XRD procedure was conducted at 40 kV and 30 mV, and the samples were scanned across a 2θ range equaling 0-80 • with a 0.02 • step size.Fourier transform infrared (FTIR) spectroscopy (Bruker Optics, ALPHA, Ettlingen, Germany) was conducted using the KBr pellet method [35] to calculate the fluctuation of surface functional groups in a frequency range wavelength of 4000 to 400 cm −1 .A scanning electron microscope (JCM-6000PLUS NeoScope Benchtop SEM, Tokyo, Japan) was applied to specimens mounted on stubs and sputter-coated with gold-palladium to detect the change in the surface morphologies of the samples after treatment.

Kinetic Models
The first-order and Stover-Kincannon kinetic models were used to determine the substrate utilization parameters, describing COD removal via DHS.

First-Order Substrate Removal Model
Organic matter removal using aerobic biological systems can be expressed by Equation ( 1) [36].
There is a negligible variation in −ds/dt in the pseudo-steady-state situation; therefore, Equation ( 2) can be derived.
where (ds/dt) represents the substrate removal rate (g/L/d); S i and S e denote the feed and effluent substrate concentrations (g/L), respectively; HRT is hydraulic retention time (d); K 1 is the speed constant of the first-order model for removing organic matter (1/d); Q expresses the flow rate to the DHS (L/d); and V is the DHS effective volume (L).

Stover-Kincannon Model
Equations ( 3)-( 5) were used to estimate the kinetic parameters of the modified Stover-Kincannon model, considering the substance removal rate in a steady-state condition [22]: where U max represents the maximum utilization rate constant (g/L/d), and K B represents the saturation value constant (g/L/d).
The linearization form, Equation ( 6), was used to determine the U max and K B kinetics: The effluent substrate concentration can be predicted using Equation ( 7):

Effect of Organic Loading Rate on COD Removal
Figure 2a shows the variation in the influent and effluent COD throughout the experiment.The average COD concentration in the effluent increased from 52.6 ± 9.8 to 436.7 ± 38.6 mg/L due to elevating the OLR from 3.01 to 12.33 kg COD/m 3 sponge /d.The sponge medium in the DHS was occupied by high-density microbial biomass, utilizing the readily biodegradable fraction of COD.Under higher OLRs, the attached biomass could not convert most organic compounds into CO 2 , H 2 O, NH 3 , energy, or other end products.In a similar study, El-Kamah et al. [37] found that increasing OLR from 2.2 ± 0.7 to 5.1 ± 1.8 kg COD/m 3 sponge /d reduced the COD removal efficiency from 87.6% to 70%.Their study [37] found that, at higher OLRs, the imposed COD mass could be greater than the amount required for decomposition by the available biomass in the sponge pieces.For a DHS treating reactive dye wastewater [38], elevating the OLR from 2.8 to 12.44 kg COD/m 3 sponge /d was associated with influent and effluent COD removal efficiencies of 66.5 ± 7.07% and 14.8 ± 2.9%, respectively.Feeding the DHS reactor with higher OLR levels could also be associated with the separation of the attached biomass from the media, resulting in biosolid escape [39].This biomass is located either on the surface or inside the sponge pieces.It is suggested that an additional treatment unit should be used to polish the DHS's effluent before final disposal.

Effect of Organic Loading Rate on NH 3 Removal
The NH 3 content evolution in the DHS effluent for each OLR is seen in Figure 2b.This NH 3 concentration elevated from 45.95 ± 3.41 to 81.23 ± 6.93 mg/L with an OLR increase of 3.01 to 12.33 kg COD/m 3 sponge /d.The results indicate that the nitrification rate was highly dependent on the OLR imposed on the DHS, following a similar pattern to that of COD removal (see Figure 2a).The deterioration of nitrification activity under high loads of biodegradable COD could be due to competition between heterotrophs and autotrophs for the available substrate (COD and NH 3 ).In addition, heterotrophic bacteria and autotrophic nitrifiers compete for oxygen as a common electron acceptor [40].The interaction between heterotrophic and autotrophic bacteria as a function of organic substrate loading has also been demonstrated [41].This finding could validate the drop in the NH 3 removal efficiency within the investigated OLR range.These results are comparable to Tandukar et al. [17], who noticed that increasing the OLR of the DHS system from 2.03 to 3.15 kg COD/m 3 sponge /d caused an increase in the residual ammonia concentrations, from 10 to 20 mg/L.A lower OLR favored the extension of the solids retention time (SRT), providing appropriate circumstances for slow-growing nitrifying bacteria [41].This long SRT (at a low OLR) is important for nitrifiers to successfully complete the ammonia oxidation process.Watari et al. [12] also found that higher OLRs (i.e., shorter SRTs) might lead to the effective washing out and dilution of the nitrifier biomass responsible for NH 3 removal.

Effect of Organic Loading Rate on NH3 Removal
The NH3 content evolution in the DHS effluent for each OLR is seen in Figure 2b.This NH3 concentration elevated from 45.95 ± 3.41 to 81.23 ± 6.93 mg/L with an OLR increase of 3.01 to 12.33 kg COD/m 3 sponge/d.The results indicate that the nitrification rate was highly dependent on the OLR imposed on the DHS, following a similar pattern to that of COD removal (see Figure 2a).The deterioration of nitrification activity under high loads of biodegradable COD could be due to competition between heterotrophs and autotrophs for the available substrate (COD and NH3).In addition, heterotrophic bacteria and autotrophic nitrifiers compete for oxygen as a common electron acceptor [40].The interaction between heterotrophic and autotrophic bacteria as a function of organic substrate loading has also been demonstrated [41].This finding could validate the drop in the NH3 removal efficiency within the investigated OLR range.These results are comparable to Tandukar et al. [17], who noticed that increasing the OLR of the DHS system from 2.03 to 3.15 kg COD/m 3 sponge/d caused an increase in the residual ammonia concentrations, from 10 to 20 mg/L.A lower OLR favored the extension of the solids retention time (SRT), providing appropriate circumstances for slow-growing nitrifying bacteria [41].This long SRT (at a low OLR) is important for nitrifiers to successfully complete the ammonia oxidation process.Watari et al. [12] also found that higher OLRs (i.e., shorter SRTs) might lead to the effective washing out and dilution of the nitrifier biomass responsible for NH3 removal.

Effect of Organic Loading Rate on TSS Removal
The data shown in Figure 2c illustrate the change in TSS removal with respect to OLR variation.The DHS reactor achieved high TSS removal efficiencies of 88.13 ± 5.42% and 80.87 ± 5.71% at OLRs of 3.01 and 12.33 kg COD/m 3 sponge /d, respectively.The corresponding residual concentrations of TSS in the final effluents were 3.23 ± 1.94 and 48.67 ± 9.59 mg/L, respectively.Elevating the OLR values was associated with an increase in the amount of sludge (biofilm growth) attached to the sponge pieces.This pattern was continuously monitored until the sponge could not hold the thick biofilm, leading to the biomass sloughing phenomenon.Subsequently, the TSS concentrations in the final effluent increased.Raising the OLR is connected to a higher food-to-microorganism (F/M) ratio of 0.36 kg COD/kg VS/d, decreasing the possibility of sludge degradation (i.e., bacterial autolysis and protozoan and metazoan predation) [42].As such, lower F/M ratios tend to provide suitable conditions for bacterial cell lysis (i.e., cell wall breakdown and sludge reduction) [43].However, the presence of a well-designed settler connected to the DHS reactor (see Figure 1) could mitigate the impacts of OLR and reduce effluent TSS concentrations under the permissible limit (100 mg/L).

Effect of Organic Loading Rate on TDS Removal
The effluent TDS concentrations increased gradually with an increasing OLR in the 3.01-12.33kg COD/m 3 sponge /d range (Figure 2d).The lowest TDS removal efficiency of 14.20 ± 1.07% was noticed at the largest OLR of 12.33 kg COD/m 3 sponge /d.Insufficient TDS reduction via DHS could be illustrated due to several reasons, including (i) the secretion of soluble microbial products as a defense mechanism against harsh environments and (ii) bacteria decay due to substrate nonavailability [44].

DHS Profile for COD Removal
The average COD removal efficiencies for the three segments were 38.06%, 28.24%, and 18.04% of the overall DHS performance (Figure 3a).As such, the majority of COD was removed in the first and second segments of the DHS system.However, no noticeable improvement in COD removal was achieved by the third (last) segment.Furthermore, Figure 3a demonstrates that the COD removal efficiencies in the segments declined with increasing OLR levels, especially in segment-1 and segment-2.Most of the coarse and soluble organic matter was deposited and decomposed in the first segments.Heterotrophic bacteria situated in the packing materials tended to utilize the available substrate (as an electron donor) and natural oxygen (as an electron acceptor) from the air for aerobic growth (Equation ( 8)).
C 6 H 12 O 6 + 6O 2 → 6CO 2 + 6H 2 O + energy (8) 0.36 kg COD/kg VS/d, decreasing the possibility of sludge degradation (i.e., bacterial autolysis and protozoan and metazoan predation) [42].As such, lower F/M ratios tend to provide suitable conditions for bacterial cell lysis (i.e., cell wall breakdown and sludge reduction) [43].However, the presence of a well-designed settler connected to the DHS reactor (see Figure 1) could mitigate the impacts of OLR and reduce effluent TSS concentrations under the permissible limit (100 mg/L).

Effect of Organic Loading Rate on TDS Removal
The effluent TDS concentrations increased gradually with an increasing OLR in the 3.01-12.33kg COD/m 3 sponge/d range (Figure 2d).The lowest TDS removal efficiency of 14.20 ± 1.07% was noticed at the largest OLR of 12.33 kg COD/m 3 sponge/d.Insufficient TDS reduction via DHS could be illustrated due to several reasons, including (i) the secretion of soluble microbial products as a defense mechanism against harsh environments and (ii) bacteria decay due to substrate nonavailability [44].

DHS Profile for COD Removal
The average COD removal efficiencies for the three segments were 38.06%, 28.24%, and 18.04% of the overall DHS performance (Figure 3a).As such, the majority of COD was removed in the first and second segments of the DHS system.However, no noticeable improvement in COD removal was achieved by the third (last) segment.Furthermore, Figure 3a demonstrates that the COD removal efficiencies in the segments declined with increasing OLR levels, especially in segment-1 and segment-2.Most of the coarse and soluble organic matter was deposited and decomposed in the first segments.Heterotrophic bacteria situated in the packing materials tended to utilize the available substrate (as an electron donor) and natural oxygen (as an electron acceptor) from the air for aerobic growth (Equation ( 8)).

DHS Profile for NH 3 Removal
According to Figure 3b, the nitrification efficiency was relatively limited in the DHS's first segment, providing an average NH 3 removal rate of about 8.69%.Because the first segment was subjected to the highest OLR, the existing ammonium oxidizers could not compete with heterotrophs for space and oxygen.However, the nitrification rates improved in the DHS's second and third segments, attaining removal efficiencies of approximately 16.05% and 26.26%, respectively.These results show that the aerobic growth of autotrophs remained dominant in the DHS's lower portions, with decreased OLRs.In another study, it was reported that the sponges in the lower DHS compartment were occupied mainly by nitrite-oxidizing (Nitrospira and Nitrospina) and ammonia-oxidizing (Nitrosomonas) bacteria, supporting the nitrogen removal capability [40].Fleifle et al. [14] also found that N was highly removed in the third compartment of the DHS because the upper segments were subjected to greater OLRs that encouraged heterotrophic biomass growth.The operating condition of these top segments was also responsible for suppressing the growth and activity of the nitrifying biomass.

DHS Profile for TSS Removal
Figure 3c demonstrates that a major amount of TSS was eliminated in the first segment, with an average removal efficiency of 48.37% of the DHS's total performance.This value reduced to about 23.54% in the second segment.Segment-1 played a prominent role in reducing the particulate organic shock load subjected to the DHS unit.The sponge voids could be partially occupied by microbial communities during the initial operating phase, making a suitable condition to entrap most of the wastewater particulates.Accordingly, the physical retention mechanism in the DHS's upper segment could remove most of the coarse suspended solids.This finding suggests that bio-filtration could be another route for degrading organic matter via the attached biomass in the DHS.However, biosolids in segment-1 might escape from the sponge carriers at higher OLRs, reducing the TSS removal performance.

DHS Profile for TDS Removal
Although the DHS was insufficient in removing the dissolved contaminants, the first segment reduced the largest portion of TDS (~6.47%) (Figure 3d).A large amount of biomass (sludge) retained in the first segment's sponge could adsorb the dissolved organic carbon and ions.However, this sludge had a small quantity in the DHS's final segment, probably due to the insufficient substrate required for bacterial growth.Bacteria decay and lysis could be associated with releasing soluble products, elevating the TDS in the third compartment.

Predicting Effluent COD Using the Kinetic Coefficients of Substrate Removal Models
The effect of varying the OLRs on the substrate consumption rates was explored by estimating the kinetic coefficients.The kinetic constants (K 1 , U max , and K B ) of the first-order and modified Stover-Kincannon models were derived by fitting them to the COD data (Table 2).These coefficients were also used to predict the concentration of the COD effluent and the corresponding removal efficiencies.The results attained by operating the DHS system in a steady-state condition with varying influent COD concentrations were utilized to estimate the kinetic coefficients of the first-order model.A straight line between (S i − S e )/HRT and S e was used to determine the kinetic K 1 constant (see Supplementary Figure S1).The K 1 and R 2 values related to the entire DHS performance were calculated as 27.39 1/d and 0.927, respectively.According to Table 2, the K 1 values associated with the influent and effluent COD concentrations of each segment were 3.77, 4.91, and 6.00 1/d.The variation in the rate of substrate degradation could be because segment-1 received the highest COD concentration compared with other segments.As such, the microorganisms of segment-1 were subjected to a high quantity of biodegradable food ingredients, requiring a longer HRT than other compartments.A lowvalue coefficient of determination, especially in the third segment (R 2 = 0.91), demonstrates that the first-order substrate consumption model can imprecisely predict the COD removal efficiencies of the studied DHS reactor.The effluent COD concentration could be expected based on the determined K 1 value using Equation (9).The first-order model constant achieved in this study was lower than a K 1 of 37.4 1/d, estimated by removing diethyl phthalate (DEP) using a biofilm reactor [45].This variation in K 1 values could be ascribed to the differences in the biodegradable fraction of COD, the microbial community distribution, and/or the operating condition.

Stover-Kincannon Model Kinetic Parameters
A plot of V/(Q x (S i −S e )) against V/Q × S i was used to determine the K B and U max of the Stover-Kincannon model (see Supplementary Figure S2).A high R 2 value of 0.99 (see Table 2) indicates good agreement between the prediction and the experimental COD data for the DHS reactor and its segments.It is supposed that the modified Stover-Kincannon model can describe the COD removal performance in this study, providing more precise results than the first-order kinetic reaction.Table 2 lists the U max and K B values for different segments of the DHS unit.The U max value was 24.27 g/L/d in segment-1, declining to 11.25 g/L/d in segment-2.The final segment had the lowest U max of 9.02 g/L/d.As such, the maximum speed of COD consumption steadily reduced from the beginning to the end of the DHS process.This finding agrees with the data shown in Figure 3a, where greater COD removal occurred in the first segment.In the DHS reactor, K B and U max were derived as 83.81 and 76.92 g/L/d, respectively.Therefore, the effluent COD concentrations can be predicted by Equation (10), following the Stover-Kincannon equation.

Model Testing
Both the first-order (Equation ( 9)) and Stover-Kincannon (Equation ( 10)) equations were applied to predict the COD concentrations of the DHS effluent.The observed effluent COD and the corresponding predicted values were compared to validate the proposed kinetic models (see Supplementary Table S1).This step was conducted using unseen data not considered in the model's calibration.The COD results predicted by the Stover-Kincannon model were more precise than the first-order kinetic reaction.For example, in phase 1, the measured effluent COD was 52.60 mg/L compared with 88.30 mg/L and 51.41 mg/L, as estimated by the first-order model and Stover-Kincannon model (see Supplementary Figure S3).Accordingly, the estimated K B and U max could be further used to express the substrate consumption rate as a function of the OLR and design a DHS system operated on a large scale in real-time.To obtain more accurate predictions of the effluent COD values at each phase, nonlinear methods, genetic algorithms, and artificial intelligence techniques could be used [46].4 depicts the changes in the sponge's surface morphology after running the DHS system for 150 days.The hexagonal-shaped voids in the unloaded sponge pieces were almost empty, with many clear pores (Figure 4a).After a long-term operation, the gaps became partially occupied by agglomerated biomass, demonstrating the entrapment of microorganisms and particulate contaminants (Figure 4b).Tawfik et al. [38] also used the SEM technique to show biomass deposition on the sponge surface and inside the holes, suggesting the entrapment of both biodegradable and inert organic fractions because of the carriers.Moreover, the SEM images also demonstrated that the sponge elements could be responsible for capturing particulate pollutants and bacterial cells [41]; however, the voids were partially clogged due to the balance between the microorganisms' growth and decay patterns.This result suggests that the sponge pieces could retain solids, making the sponge an attractive alternative to other traditional media such as gravel and plastic.The immobilization of biomass within and outside the sponge material could result in a longer SRT, equivalent to 86 days for the current DHS unit.Accordingly, a smaller amount of sludge was generated, making a good scenario for reducing the costs of sludge disposal.Furthermore, there were no issues related to sponge clogging, probably because the attached biofilm could be self-degraded [42].The expenses related to cleaning, backwashing, or replacing the packing sponge material could also be minimized due to a lack of deformation or destruction in the sponge pieces.

XRD
Figure 4c represents the XRD patterns of (i) the sponge loaded with sludge before the beginning of the experiment (i.e., the sponges were initially soaked in sludge to reduce the lag-phase adaptation) and (ii) the sponge at the end of operations.The detection of multiple broad peaks between a 2θ of 13° and 24° before treatment could be ascribed to the crystallographic (002) plane.This XRD pattern corresponded to CrI = 41.29%.After wastewater treatment, a sharp peak became visible at 2θ around 30.5°.This new peak, in addition to other peaks, could justify increasing the CrI to 45.76%.For instance, the accumulation of organic molecules and the associated bacterial growth/decay increased the XRD crystalline peak over the amorphous region.The interaction in sponge structure (i.e., fibers) between organic compounds and bacterial cell debris could be one of the main reasons for the formation of larger crystal sizes and highly organized structures (more regular morphology) [12].However, the slight variation between the two XRD profiles could be ascribed to the microorganism's self-degradation and the balance between biomass

XRD
Figure 4c represents the XRD patterns of (i) the sponge loaded with sludge before the beginning of the experiment (i.e., the sponges were initially soaked in sludge to reduce the lag-phase adaptation) and (ii) the sponge at the end of operations.The detection of multiple broad peaks between a 2θ of 13 • and 24 • before treatment could be ascribed to the crystallographic (002) plane.This XRD pattern corresponded to CrI = 41.29%.After wastewater treatment, a sharp peak became visible at 2θ around 30.5 • .This new peak, in addition to other peaks, could justify increasing the CrI to 45.76%.For instance, the accumulation of organic molecules and the associated bacterial growth/decay increased the XRD crystalline peak over the amorphous region.The interaction in sponge structure (i.e., fibers) between organic compounds and bacterial cell debris could be one of the main reasons for the formation of larger crystal sizes and highly organized structures (more regular morphology) [12].However, the slight variation between the two XRD profiles could be ascribed to the microorganism's self-degradation and the balance between biomass growth and decay [45] (as previously suggested in SEM; Figure 4a,b).

FTIR
The FTIR spectroscopy data were collected to determine the existing functional groups and bonds in the sponge samples (Figure 4d).For both samples, it was found that the functional groups and assigned corresponding peaks were comparable but with slightly different peak intensities.This finding indicates that the changes in the chemical structure of both sponge samples (sponge before treatment soaked in sludge and sponge at the end of the experiment) were minor.The absorption band at 3450 cm −1 could be assigned to the O-H bond in hydroxyl functional groups [47].This band could emphasize the existence of hydrogen bonds in the amines and alcohols (or phenols) in the sponge samples.Observing an absorption peak at 1635 cm −1 could be related to C=O stretching and amide I (C-N) [48].Moreover, an absorption peak at 1384 cm −1 (fingerprint region) could indicate nitrate accumulation, owing to the nitrification pathway (ammonia conversion to nitrate).The characteristic band in the 1200-1500 cm −1 region was previously used to quantify nitrate contents [49].The FTIR spectrum also showed a peak at 556 cm −1 , probably attributed to the existence of (i) aromatic compounds in humic-like substances, and/or (ii) the Si-O bond present in the raw sludge.Overall, the results of the FTIR analysis revealed the presence of organic macro-molecules made of proteins, polysaccharides, and lipids in the sponge samples.The sample after treatment had slightly higher peak intensities due to the retention and deposition of organic molecules caused by the sponge medium accompanied by dead-end products and debris sloughing.

Suggested Pollutant-Removal Mechanisms via DHS Based on Experimental Results and Literature Survey
Based on the current results and observations obtained from literature studies [2,37], organics and solids pollution reduction via DHS could follow several hypotheses (Figure 5).The sponge carrier acquires a large amount of void space (see SEM image in Figure 4a), allowing for biomass growth and acclimatization to high OLRs.Heterotrophic microorganisms occupying the surface (aerobic) and interior (anoxic) zones of the sponge could be responsible for COD (substrate) reduction.As such, the organic compounds are converted into CO 2 , H 2 O, and new cells [50].The DHS's upper segments are responsible for removing the most biodegradable organic matter from the wastewater.The presence of anoxic conditions within the sponge media effectively facilitates the accomplishment of the denitrification process simultaneously [41].The bottom compartments, with lower organic concentrations, provide a suitable condition for autotrophs to utilize the NH 3 content and form the nitrification pathway (see Figure 3b).This mechanism was supported by FTIR characterization, where an absorption peak was found at 1384 cm −1 , probably due to nitrate formation (see Figure 4d).Autotrophic ammonium-oxidizing bacteria also consume large quantities of dissolved oxygen; hence, they dominate the sponge exterior surface [40].In addition to the nitrification and denitrification processes, converting organically bound nitrogen to ammonia has also been noticed by previous researchers [8,51].The sponge medium also acts as a filter to retain suspended particles, biomass, and sludge.Prolonging the sludge age would provide a positive situation for the domination of nitrifying bacteria [42,52], improving the NH 3 removal performance (see Figure 2b).
to nitrate formation (see Figure 4d).Autotrophic ammonium-oxidizing bacteria also consume large quantities of dissolved oxygen; hence, they dominate the sponge exterior surface [40].In addition to the nitrification and denitrification processes, converting organically bound nitrogen to ammonia has also been noticed by previous researchers [8,51].The sponge medium also acts as a filter to retain suspended particles, biomass, and sludge.Prolonging the sludge age would provide a positive situation for the domination of nitrifying bacteria [42,52], improving the NH3 removal performance (see Figure 2b).

Performance of Sequential DHS System for Wastewater Treatment
Two sequential DHS reactors were applied for the HSOW treatment, supporting the concept of unconventional wastewater treatment applications in small and decentralized communities (see Supplementary Figure S4).Table 3 lists the results obtained by running the DHS system under different OLR phases.The effluent characteristics of the combined system demonstrated that a promising performance was achieved.At an OLR of 3.01 kg COD/m 3 sponge /d, the COD, NH 3 , TSS, and TDS concentrations in the effluent of the first DHS were 52.60 ± 18.6, 45.95 ± 17.19, 3.21 ± 1.07, and 855.0 ± 53 mg/L, respectively.These values were further reduced after passing the first effluent through the second DHS, yielding final concentrations of 27.0 ± 9.61, 15.37 ± 8.22, 2.99 ± 1.13, and 730.0 ± 43.19 mg/L, respectively.A similar DHS/DHS performance was noticed when imposing other OLRs on the treatment system.The average removal efficiencies for COD, NH 3 , TSS, and TDS using the first DHS unit were 84.33%, 51.0%, 84.57%, and 17.05%, respectively.These observations were significantly (p < 0.05) improved after the second DHS step to 94.72%, 75.83%, 88.94%, and 26.30%, respectively.The results in Table 3 indicate that most of the organic compounds (expressed by influent COD) could be adequately removed by the first DHS phase.However, the second DHS unit could be appropriately employed to oxidize a large portion of ammonium nitrogen.In particular, the competition between heterotrophs and nitrifiers for available oxygen was reduced in the second unit, which was associated with a low F/M ratio of 0.24 Kg COD/kg VS/d.Moreover, prolonging the entire HRT to 7.20 h provided a suitable condition to increase the contact between biodegradable organics and microbes.Moreover, the second DHS had an SRT of 171 days, which was greater than the SRT of DHS-1 by about twofold.Due to the endogenous decay, this long SRT is essential to reduce the amount of sludge generated by the second DHS.Moreover, this condition revealed the presence of slow-growing nitrifiers in DHS-2.The integrated DHS/DHS also showed an efficient performance in removing various pollutants compared with other decentralized wastewater treatment systems (Table 3).This DHS technology has been established by different researchers in Vietnam [29,31], Japan [12,42], Thailand [53], the United Kingdom [30], and Indonesia [18].Moreover, this technology has been combined with other aerobic processes [30,53] and anaerobic reactors [20,39].In addition to domestic sewage, onion wastewater [37], agricultural drainage water [14], and textile industrial effluents [12] have also been properly treated using DHS-based technology.Increasing the OLR from 0.2 to 0.4 kg COD/m 3 /d for domestic wastewater decreased the COD and NH 3 removal efficiencies from 86.6% to 84.2% and 98.6% to 81.3%, respectively [30].

Economic Consideration for DHS Implementation
Artificial aeration is a common energy-consuming process in aerobic-based wastewater treatment systems, accounting for 40-60% of the total operating costs.This cost is followed by the expenses caused by energy consumption in sludge treatment (15-25%) and pumping the effluent to the secondary clarifier (15%) [56].The current study used a DHS system to treat HSOW without purchasing either external or mechanical aerators.Moreover, the DHS unit produced a low quantity of sludge (due to extended SRT), reducing the expenses of sludge dewatering and drying.The sponge pieces were synthesized and delivered locally, costing about 300 USD/m 3 .The sponge elements were not replaced during the experimental period (150 d).Furthermore, Watari et al. [12] demonstrated that the sponge modules do not need to be substituted over a long-term operation due to their stability and durability.The initial investment of the DHS reactor (construction work and mechanical and electrical facilities) was estimated to be USD 40 to treat 1 m 3 /d of sewage.This price was calculated based on a land requirement of approximately 0.015 m 2 /PE (square meters per person equivalent) [57].Hence, the DHS's initial investment was 80 USD/m 3 /d because it could be simply constructed in situations with low space availability.The amount of energy input to the DHS was assigned to the pump's electrical power consumption.This item represents the pumps used for DHS feeding and effluent disposal.The total energy consumption was estimated to be 0.14 kWh/m 3 of sewage load (i.e., 0.07 kWh/m 3 for DHS-1 and DHS-2; power cost calculated at a rate of 0.07 USD/kWh).The sludge drainage pump consumed less than 0.001 kWh energy/m 3 of sewage load, owing to the low amount of sludge produced [11].Other running costs, including operation and maintenance, manpower, and chemical utilization, accounted for around 2.59 USD/m 3 /d for each reactor.This value is equivalent to 3.53 USD/m 3 (i.e., 7.07 USD/m 3 for a combined DHS/DHS system), assuming an interest rate of 4% annually with a 15-year project lifetime.The estimated running cost was considerably lower than that of other conventional aeration systems, such as 47 USD/m 3 /d for the activated sludge process [58], because the air diffuses naturally into the sponges' DHS (there is no need for external aeration).

Conclusions
A downflow hanging sponge (DHS) system was successfully employed to treat highstrength organic wastewater (HSOW), accomplished by varying the organic loading rate (OLR) from 3.01 to 12.33 kg COD/m 3 sponge /d.By examining the sponges' surface morphologies and structures, the DHS was able to entrap the particulate matter, reduce COD in the upper segments, and perform nitrification in the middle and lower compartments.Moreover, the DHS's performance in removing COD from the HSOW was adequately accomplished using the kinetic coefficients of the modified Stover-Kincannon model (R 2 0.99), which was able to predict the effluent COD using new data sources.To obtain allowable wastewater disposal criteria (COD < 80 mg/L), a single DHS unit could be operated at an OLR below 4.22 kg COD/m 3 sponge /d; however, a combined DHS/DHS system could withstand a higher OLRs until 10.28 kg COD/m 3 sponge /d.As such, additional treatment units could be essential when OLR levels reach 12.33 kg COD/m 3 sponge /d.The economic feasibility study demonstrated that the combined DHS system could be a good alternative to conventional treatment systems, especially in decentralized and low-income communities.

Supplementary Materials:
The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/su15076038/s1.Author Contributions: A.Z.: methodology, formal analysis, writing-original draft; M.N.: supervision, visualization, writing-review and editing; M.F.: supervision, visualization, writing-review; M.G.I.: supervision, conceptualization, writing-review and editing.All authors have read and agreed to the published version of the manuscript reported.
Funding: Funds were partially granted by Egypt-Japan University of Science and Technology (E-JUST) and the Japan International Cooperation Agency (JICA).

Figure 1 .
Figure 1.Diagram of a downflow hanging sponge (DHS) unit used to treat high-strength organic wastewater.

Figure 2 .
Figure 2. Effect of OLR variation on DHS performance regarding the removal efficiencies of (a) COD, (b) NH3, (c) TSS, and (d) TDS.Error bars are estimated as the difference between the maximum and minimum value of each result (n = 3).

Figure 2 .
Figure 2. Effect of OLR variation on DHS performance regarding the removal efficiencies of (a) COD, (b) NH 3 , (c) TSS, and (d) TDS.Error bars are estimated as the difference between the maximum and minimum value of each result (n = 3).

Figure 3 .
Figure 3. Monitoring the removal efficiencies of (a) COD, (b) NH3, (c) TSS, and (d) TDS at each DHS segment.Error bars are estimated as the difference between the maximum and minimum value of each result (n = 3).

Figure 3 .
Figure 3. Monitoring the removal efficiencies of (a) COD, (b) NH 3 , (c) TSS, and (d) TDS at each DHS segment.Error bars are estimated as the difference between the maximum and minimum value of each result (n = 3).

3. 4 .
Figure4depicts the changes in the sponge's surface morphology after running the DHS system for 150 days.The hexagonal-shaped voids in the unloaded sponge pieces were almost empty, with many clear pores (Figure4a).After a long-term operation, the gaps became partially occupied by agglomerated biomass, demonstrating the entrapment of microorganisms and particulate contaminants (Figure4b).Tawfik et al.[38] also used the SEM technique to show biomass deposition on the sponge surface and inside the holes, suggesting the entrapment of both biodegradable and inert organic fractions because of the carriers.Moreover, the SEM images also demonstrated that the sponge elements could be responsible for capturing particulate pollutants and bacterial cells[41]; however, the voids were partially clogged due to the balance between the microorganisms' growth and decay patterns.This result suggests that the sponge pieces could retain solids, making the sponge an attractive alternative to other traditional media such as gravel and plastic.The immobilization of biomass within and outside the sponge material could result in a longer SRT, equivalent to 86 days for the current DHS unit.Accordingly, a smaller amount of sludge was generated, making a good scenario for reducing the costs of sludge disposal.Furthermore, there were no issues related to sponge clogging, probably because the attached biofilm could be self-degraded[42].The expenses related to cleaning, backwashing, or replacing the packing sponge material could also be minimized due to a lack of deformation or destruction in the sponge pieces.

Figure 4 .
Figure 4. Characterization of sponge media by (a) SEM of raw sponge unloaded with biomass, (b) SEM of sponge occupied by biomass after experimentation, (c) XRD of sponge before treatment (soaked in sludge) and at the end of the experiment, and (d) FTIR for sponge before treatment (immersed in sludge) and at the end of the experiment.

Figure 4 .
Figure 4. Characterization of sponge media by (a) SEM of raw sponge unloaded with biomass, (b) SEM of sponge occupied by biomass after experimentation, (c) XRD of sponge before treatment (soaked in sludge) and at the end of the experiment, and (d) FTIR for sponge before treatment (immersed in sludge) and at the end of the experiment.

Table 1 .
Characteristics of high-strength organic wastewater during the operational period.

Table 1 .
Characteristics of high-strength organic wastewater during the operational period.

Table 2 .
Comparison of the kinetic constants according to first-order and Stover-Kincannon models for the entire DHS and its segments.

Table 3 .
Comparison of the sequential DHS system with previous studies regarding the removal of COD, N, and TSS.