Microbubbles were applied to remove phosphorus (P) and improve environmental water conditions on the surface of the benthic sediment in a eutrophic lake. Microbubble flotation (MF) was used to remove P in a laboratory-scale experiment device from the benthic sediment and overlying water field samples. The results of P tracing observation for MF treatment, which were identified based on the mass balance, showed P accumulated at a higher concentration in the floated scum than in the deposited sludge. Furthermore, while the amount of soluble P separated was insignificant, the removal efficiency of soluble P in the floated scum was higher than total P. As an additional effect, P release was suppressed by the injection of microbubbles, which reduced anaerobicization and improved environmental conditions in the overlying water of the lake. We also confirmed that the sludge that settled after flotation separation caused capping, which restrained P release by forming a layer on the surface of the sediment. The direct P flotation separation, provision of oxygen by injected microbubbles, and capping effect of re-deposited sludge are advantages associated with MF, and our results show that field application studies are warranted.

  • MF can be applied to reduce P in the benthic sediment and overlying water of a lake.

  • Soluble P was removed more easily than T-P.

  • MF accumulated P at a higher concentration in the floated scum than in the deposited sludge.

  • MF reduced anaerobicization and improved environmental conditions in the overlying water.

  • While MF is not efficient at T-P removal, the process does effectively separate soluble P.

Graphical Abstract

Graphical Abstract
Graphical Abstract
CW

clarified water

DO

dissolved oxygen

DS

deposited sludge

FS

floated scum

MF

microbubble flotation

ORP

oxidation/reduction potential

PSD

particle size distribution

RS

raw sediment

RW

raw water

SRP

soluble reactive phosphorus

T-P

total phosphorus

US

unaffected sediment

Algal blooms die and settle at the bottom of eutrophic lakes, precipitating anaerobic conditions at the bottom and accelerating the release of nutrients from sediment. As the organic matter decays, soluble reactive phosphorus (SRP) buried in the bottom layer formed in the pore water is either adsorbed by sediment particles or bonded to iron oxides and held in the sediment (Filippelli & Delaney 1996). When the iron oxide enters a reduction condition, phosphorus (P) adsorbed at the surface is desorbed into the pore water (Thamdrup et al. 1994), and then released into the bottom water as a result of the difference in concentration between sediment and overlying water. This process, referred to as P-loading, induces circulating eutrophication of the lake. Control over internal P-loading is a necessary prerequisite to the improvement of the water quality of lakes (Wu et al. 2014).

Several means of controlling P-loading exist, including sediment dredging, aeration, and on-site chemical and biological treatment using calcium nitrate (Lin et al. 2016). None of these treatments, however, offer permanent solutions (Sadeghi et al. 2020). With dredging in particular, water quality may deteriorate due to sediment resuspension, resulting in ecosystem disturbance (Newell et al. 1998).

Flotation is a water treatment technology that uses microbubbles to separate particles from the water by spraying pressurized dissolved air at atmospheric pressure (Edzwald 1995). Flotation has long been used as a treatment strategy, and it continues to be incorporated into water treatment processes. Previous studies (Kang et al. 2016; Choi et al. 2019) reported that reducing P release in the bottom sediment through flotation has an additive effect, whereby the flotation separation using microbubbles improves aerobic conditions by directly injecting air bubbles into the bottom layer, some of the aluminum (Al) flocs that cannot float because they cannot adhere to microbubbles are deposited in the bottom layer and form a film layer on the surface of the sediment in the lake. The film that forms on the surface of the sediment induces a capping effect that inhibits P release.

Capping is being explored as a means of reducing internal P-loading (Li et al. 2017; Zhou et al. 2020). A capping treatment employing Al salt has been used continuously in more than 200 locations worldwide to restore eutrophic lakes. In some lakes, this technique has reduced P-load by up to 83%, and the effect has lasted for more than 10 years and can reduce P by 54–83% (Li et al. 2017). Capping is a low-cost means of reducing the P-load by directly addressing it in the field (Hart et al. 2003). Capping's most important functions are to protect sediment from erosion by contaminated particles and separate organisms present in sediment from the contaminated sediment (Eek et al. 2008). Pollutant flux is reduced by adsorbing pollutants in the sediment capping matrix and increasing the diffusion and transport transmission length (Go et al. 2009).

In one recent study, Al salt-based coagulants and natural minerals (limestone, zeolite, etc.) were used to remove P and cyanobacteria, and inhibit P release into the water layer after coagulation and sedimentation (Sadeghi et al. 2020). In another study, P was removed from the bottom layer sediments of a lake using microbubbles. Kim & Kwak (2019) described the mechanism of aggregation and P removal by bubbles and sediment particles through modeling contact, separation, and flotation efficiency. Another study (Kang et al. 2016) found flotation of sediment was 22.6% efficient at removing total phosphorus (T-P), and the P-loading reduction rate of P released by flotation was approximately 20–64%, confirming that the P capping effect comprised a part of the flotation process.

While additional studies have investigated other means of P removal from sediment layer and controlling P-loading using microbubbles, few studies have examined the quantitative effect of capping by Al salts and microbubbles, as well as the behavior of P in T-P and SRP sediments during flotation. In this study, we examined the separation of P from sediment, traced the behavior of P by the microbubble flotation with alum coagulation, and evaluated the additive capping effect of Al particles. As a result of the flotation process, P-containing phosphate particles floated to the water's surface, while the remaining aluminum flocs re-settled at the bottom of the lake. These settled particles formed a coating layer on the surface of the sediment and caused a capping effect that restrained the release of P from the sediment. The following two types of experiments were conducted in this study: (1) flotation experiments of phosphate particles using microbubble flotation (MF) for benthic sediments in a lake and (2) laboratory column tests of post-MF P-released settled phosphate particles on the sediment's surface. From the viewpoint of the mass balance of P, P's behavior in sediment and water as it was changed by MF was investigated based on the results of the experiments. We also compared the capping effect of MF that inhibited P release in sediments by Al particles settling after flotation against cases without MF treatment.

Samples and flotation experiments

Figure 1 is a conceptual schematic diagram of the experiment performed in this study. A hypothetical experiment was designed to remove P from the sediment of an actual contaminated lake using microbubbles. Flotation by injection of microbubbles onto the surface of the lake's sediment was implemented at a laboratory scale under conditions similar to those found at the actual site. When a bubble cloud (milky water) was injected into the lake sediment, a hydraulic disturbance occurred on the surface layer of the sediment, and P components (e.g., T-P and SRP) present in the sediment surface layer diffused into the water body along with sediment particles. The T-P that moved from the sediment to the water and the inorganic form of PO4-P that existed in a dissolved state combined with the injected chemical coagulant to form phosphate particles. These particles bound themselves to the microbubbles and floated to the surface of water.

Figure 1

Schematic diagram of the experimental apparatus for flotation and capping of P in the benthic sediment. (a) Laboratory experimental apparatus for flotation and P release test. (b) Conceptual diagram of MF device for application to the field.

Figure 1

Schematic diagram of the experimental apparatus for flotation and capping of P in the benthic sediment. (a) Laboratory experimental apparatus for flotation and P release test. (b) Conceptual diagram of MF device for application to the field.

Close modal

Sediment samples were collected from the Saemangeum Lake, an artificial lake currently under construction in southwestern Jeollabuk-do, Korea. In the Saemangeum Lake, which contains coastal brackish water in the estuary, the concentration of P has continuously increased since the sea dike was established (Jeong & Yang 2015), and an increase of algae acting as an internal load has been reported (Kim et al. 2009). The sediment samples used in this experiment were collected from four sites in which influent particles were deposited. These sites were selected as exerting the most significant effect on water quality among the possible measurement sites (Jellabukdo 2017) (Figure 2).

Figure 2

Sampling sites (Seamangeum Lake).

Figure 2

Sampling sites (Seamangeum Lake).

Close modal

Sediment was collected using a sediment corer, sealed to prevent disturbance, and transported under refrigerated condition to the laboratory. For particle size analysis, mass was measured by separating particles by size using a standard sieve. Analysis was performed using the sieve method, by which the size of the particles is determined by the distribution ratio for each size. To determine T-P of each sediment sample before and after the flotation test, each dried sediment sample was heated in an electric furnace at 550 °C for 2 h, then transferred to an extraction container where 20 mL of hydrochloric acid was added. After pretreatment by stirring for 16 h, a point at which burned sediment no longer settled, each solution was filtered through a filter paper (GF/C) and colorimetrically quantified by an ascorbic acid reduction method. For SRP, 50 mL of a 0.02 M KCl solution was added to the dried sediment sample, stirred at 150 rpm for 24 h, filtered, and colorimetrically quantified by the ascorbic acid reduction method. The T-P and SRP of sediment were analyzed according to water pollution process test standards (ARCS 1994), and the T-P and PO4-P of the sample water and of the treated water were analyzed according to procedures associated with the Standards Method (APHA 2012). SHIMADZU's UV-160A was used for the measurement.

Experiment method

Samples used in the experiment were collected using a sediment corer, and a batch-wise laboratory experiment was conducted to determine conditions at the Saemangeum Lake. Based on the coagulant dose obtained through the Jar test, the mixing intensity of coagulant in the flotation column was decreased stepwise from 120 to 30 rpm over a 10-min period. Flotation was performed by spraying saturated water (milky water) pressurized at 5.0 atm of a saturator to the flotation column of the experiment apparatus.

A 5 L column (diameter 10 cm, height 120 cm) was used as a flotation experiment device, and after 4 L of the sample was placed into the column, coagulant and saturated water were added under the same mixing conditions defined in the Jar test. In the flotation experiment, the dry weight and concentration of P in the sediment that was to be mixed with the sample water were measured prior to the experiment, and water quality analysis was performed on the sample mixed with the raw water (RW). The floated scum (FS) on the surface water and the deposited sludge (DS) on the bottom of the column were collected using a vacuum pump, filtered through filter paper, and dried in a drying oven at approximately 110 °C. T-P, SRP, and PO4-P were then measured and evaluated as shown in Figure 3.

Figure 3

Schematic diagram of P trace investigation for the stepwise experimental processes and measurements.

Figure 3

Schematic diagram of P trace investigation for the stepwise experimental processes and measurements.

Close modal

Poly-aluminum chloride (a coagulant) was injected onto the sediment to form the flocs of phosphate particles, followed by flotation separation via microbubbles. The re-deposited phosphate floc failed to adhere to the microbubbles and formed a film layer on the surface of the sediment, resulting in a kind of capping effect. The P release experiment was carried out 24 h after the flotation experiment was completed to allow sufficient time for the formation of a solid Al particle layer on the surface of the sediment, as all the fine flocs that remained in the supernatant were deposited after the flotation experiment. The P release experiment was conducted in a column of the same volume and involved comparing the flotation-separated sediment sprayed with microbubbles against a controlled sediment sample that had not been subjected to coagulation and MF.

The FS after the flotation experiment was discharged to reduce any experimental error that the FS layer may have exerted on the water quality environment, including the effect of dissolved oxygen (DO) concentration associated with the P release test in the column. The experiment was carried out for a total of 40 days, and samples were collected every day for 0–10 days, every 2 days for the subsequent 11–30 days, and every 10 days for the final 31–40 days.

After the conclusion of the flotation experiment, the weight and concentration of FS on the water surface and in the sediment deposited on the bottom were measured. The concentration of treated water was also measured to determine the overall mass balance involved in the experiment. In this manner, the change of P concentration in the overlying water that can induce a difference of the amount of P release through mass balance was calculated. DO concentration and oxidation/reduction potential (ORP) before and after the injection of microbubbles was measured to determine the effect of oxygen concentration and ORP in the bottom of the column on the P behavior of sediment. For the experiment concerning the period of P release, the difference in P concentration of the overlying water between MF column and the controlled column was considered as a quantitative capping effect that could be achieved by the application of MF in the lake.

Evaluation and calculation of P release

The amount of P released from sediment with and without microbubble flotation, including coagulant, was compared based on the amount of P in the overlying water and sediment of each column over time. In order to trace the behavior of P in the batch-type flotation experiment, the weight and concentration of P were measured from the FS on the water's surface, the remaining P of overlying water, and the DS on the bottom of sediment. Removal efficiency (R, %) was calculated using Equations (1)–(3), and the overall mass balance involved in the experiment was evaluated based on the results:
(1)
(2)
or,
(3)
where Vrw is the volume of RW taken from overlying water in the lake (L); Crw is the concentration of RW taken from overlying water in the lake (mg/L); Mrs is the mass of raw sediment (RS) taken from the benthic sediment in the lake (kg); Crs is the concentration of raw sediment taken from the benthic sediment in the lake (mg/kg); Mfs is the mass of FS after MF (kg); Cfs is the concentration of FS after MF (mg/kg); Ccw is the concentration of clarified water (CW) after MF (mg/L); Vcw is the volume of CW after MF (L); Mds is the mass of DS after MF (kg); Cds is the concentration of DS after MF (mg/kg); Mus is the mass of unaffected sediment (US) after MF (kg); Cus is the concentration of US after MF (mg/kg).

Sediment characteristics and microbubble flotation

The sediment samples were collected at sampling sites from the Saemangeum Lake and analyzed for their properties prior to the flotation experiments. Table 1 presents the average particle size found in the sediment taken from each site. At M-1 site, where the sample was influenced by the water of the Mangyeong River, 0.64% sand, 82.92% silt, and 16.44% clay, were detected. The average particle size was 29.49 (D10 = 4.97, D50 = 22.77, D90 = 64.94) μm, and the sample was composed of fine-grained sedimentary soil with high silt content. At the M-2 site, the sample was composed of silt and clay with no sand (61.49% silt and 38.51% clay) and the average particle was 12.50 (D10 = 2.23, D50 = 10.03, D90 = 26.86) μm. At the D-1 site the sediment was 61.65% silt and 38.35% clay, and average particle size was 13.94 (D10 = 2.15, D50 = 10.48, D90 = 31.75) μm. At site D-2, the sample was comprised of 0.06% sand, 75.76% silt, and 24.18% clay, and the average particle size was 25.67 (D10 = 3.15, D50 = 19.50, D90 = 57.94) μm. Across all sites, sediment was predominantly comprised of silt. The amount of P we detected in the Saemangeum Lake sediment was consistent with measurements taken by others (Kawk et al. 2018). Samples of the brackish water taken at the sampling sites showed high alkalinity (Table 2) with an average pH of 7.5, i.e., a suitable range for the coagulation of phosphate.

Table 1

Particle size distribution (PSD), T-P, and SRP concentration of sediments

SedimentsPSDT-PSRP
(μm)(mg/kg)(mg/kg)
M-1 29.49 190.433 4.1 
M-2 12.50 106.435 3.3 
D-1 13.94 121.333 3.0 
D-2 25.67 142.867 4.2 
SedimentsPSDT-PSRP
(μm)(mg/kg)(mg/kg)
M-1 29.49 190.433 4.1 
M-2 12.50 106.435 3.3 
D-1 13.94 121.333 3.0 
D-2 25.67 142.867 4.2 
Table 2

Qualities of RW used in this study

RWTempAlkalinitypHSalinityElectrical conductivityDO
(°C)(mg/L as CaCO3)(psu)(mS/cm)(mg/L)
M-1 21.0 92.4 7.20 13.6 14.09 4.40 
M-2 21.0 112.6 7.59 25.64 40.85 7.13 
D-1 18.8 65.8 7.63 2.92 5.38 8.54 
D-2 19.0 121.6 7.76 19.58 31.28 6.41 
RWTempAlkalinitypHSalinityElectrical conductivityDO
(°C)(mg/L as CaCO3)(psu)(mS/cm)(mg/L)
M-1 21.0 92.4 7.20 13.6 14.09 4.40 
M-2 21.0 112.6 7.59 25.64 40.85 7.13 
D-1 18.8 65.8 7.63 2.92 5.38 8.54 
D-2 19.0 121.6 7.76 19.58 31.28 6.41 

P concentration and turbidity were measured for each site to check for any relevant differences in coagulation conditions or water quality. The water of the Saemangeum Lake arrives from the Mangyeong and Dongjin Rivers, and its quality improves as particulate matter settles down spends more time settling in the lake (Jeong & Kwak 2021). The water samples taken of the overlying water at the four sites show that the water quality of M-2 and D-2 is improving compared to the waters of M-1 and D-1 (Figure 4).

Figure 4

Comparison of P concentration in the RW and CW treated with MF at the four sites.

Figure 4

Comparison of P concentration in the RW and CW treated with MF at the four sites.

Close modal

P concentration after MF was significantly lower than RW samples in common at all sites. Although the sediment disturbed by flotation caused the overlying water to become cloudy, flotation removed P from the water and the sediment. There were slight differences in T-P removal efficiency at each point, and the dissolved PO4-P removal efficiency was slightly higher than that of T-P. Removal efficiency was highest at M-1, where the P concentration in the initial RW was highest, with 88% of T-P and 98% of PO4-P (Figure 4). In short, the higher the initial concentration of pollutants, the more P was removed by microbubbles.

P's fate and behavior were investigated stepwise and across different spatial domains; for the FS on the surface water, the CW remained in the flotation column, the DS on the surface layer of bottom sediment, and the US by microbubble injection. The T-P concentration of FS was much higher than DS, and the concentration of DS was similar to or lower than RS, as shown in Figure 5. The reason that the FS had a higher T-P concentration than the RS and DS is considered to be because the particulate P formed by coagulation attaches easily to the injected microbubble, floats, and is concentrated with bubbles as scum on the water surface. In contrast, DS contains water on the surface of the bottom sediment, which, since it could not be consolidated like deep sedimentary soil, the concentration of P remains somewhat lower than that of FS and RS. The P itself was either dissolved or combined with particles in both water and the upper layer of sediment and was hydraulically disturbed and dispersed by the injected microbubbles before being either (1) separated to the water surface as suspended scum, (2) re-settled on the surface of the sediment, or (3) remaining as un-agglomerated dissolved P in the treated water. In light of these three alternatives, the feasibility of MF application can be grasped to remove P in the lake based on the mass balance of P.

Figure 5

Comparison of concentration of T-P before and after MF experiment for each of the four sites.

Figure 5

Comparison of concentration of T-P before and after MF experiment for each of the four sites.

Close modal

Release experiment

Using MF, bubble-floc agglomerates are separated, removed, and collected at the surface of the water, while the particles that form the agglomerates but do not adhere to the bubbles are deposited and reaccumulate on the sediment surface. Over time, these flocs accumulate on the surface of the sediment and form an aluminum film layer (Choi et al. 2019) that absorbs and blocks SRP release from the bottom of the sediment to the water layer, thereby suppressing emission into the water layer. In this study, we conducted a laboratory-scale release experiment to confirm whether the deposited aggregates had a positive effect on the control of P release through the capping of Al3+ coagulant.

The release test results showed that the amount of P released from the sediment gradually increased even when no microbubbles were injected (Figure 6). The anaerobicization is due to various environmental factors, including the biodegradation of organic matters over time in the eutrophicated lake, and thus the concentration of released P will also gradually increase. The concentration of P released was lower when MF was applied than when no coagulant or microbubble injection occurred.

Figure 6

Variation of P in overlying water released from the sediment with (W) and without (WO) MF.

Figure 6

Variation of P in overlying water released from the sediment with (W) and without (WO) MF.

Close modal

The results of the P release experiment suggest that the DS layer that contains Al3+ ion restrains P release from inside the sediment, i.e., manifests a capping effect. The difference between with and without MF in the capping effect of P release amount gradually increased at all sites after 10 days. Finally, there was a notable difference in P concentration at all sites 40 days after the completion of the P release test. At that point, the reduction rate of P release (capping effect) was 50% higher in the samples injected with microbubbles across all points than those that did not receive MF. This result confirms that the capping effect of aluminum flocs, which do not adhere to bubbles and are deposited on the bottom of the test column without bubbles, can reduce P release on the surface layer of the benthic sediment.

To accurately calculate mass balance, the amount of P released from the RW was converted into mass and expressed as ΔM, as shown in Table 3. At sites M-2, D-1, and D-2, ΔM without MF was positive, whereas MF was negative at all sites. ΔM value was decreased commonly in the cases of MF treatment. In contrast, P release was suppressed even in the absence of MF at M-1, where P was present in very high concentrations at the initial stage of the experiment. The results of the P release test suggest that when the ΔM value is positive, the release proceeds (increase of MReleased) such that the P concentration of the water increases. Conversely, when ΔM is negative, the concentration of released P decreases as it is adsorbed and removed due to the capping effect of DS.

Table 3

Mass change of P in release tests with and without MF

MRW (mg)MSat (mg)MFS (mg)MDS (mg)MReleased (mg)
Without MF 
 M-1 0.139 – – – −0.544 
 M-2 0.052 – – – 0.597 
 D-1 0.041 – – – 0.691 
 D-2 0.085 – – – 1.470 
With MF 
 M-1 0.142 0.067 0.105 0.035 −1.363 
 M-2 0.047 0.039 0.033 0.011 −0.074 
 D-1 0.044 0.024 0.032 0.011 −0.354 
 D-2 0.091 0.055 0.064 0.021 −1.040 
MRW (mg)MSat (mg)MFS (mg)MDS (mg)MReleased (mg)
Without MF 
 M-1 0.139 – – – −0.544 
 M-2 0.052 – – – 0.597 
 D-1 0.041 – – – 0.691 
 D-2 0.085 – – – 1.470 
With MF 
 M-1 0.142 0.067 0.105 0.035 −1.363 
 M-2 0.047 0.039 0.033 0.011 −0.074 
 D-1 0.044 0.024 0.032 0.011 −0.354 
 D-2 0.091 0.055 0.064 0.021 −1.040 

Note: MReleased = mass of P released for the 40 days, MSat = mass of P contained in the saturated water to produce microbubbles.

Change of DO concentration and ORP with MF

The DO concentration and ORP in the lower layer exert a significant influence on the release of P. In particular, the redox potential controls the release of P into the water layer from the sediment. As anaerobicization strengthens, DO concentration decreases, accelerating the release of P from the sediment. When P is separated in sediment through flotation, microbubbles remove particles and improve the oxygen concentration of anaerobic water bodies. In addition, the ORP of the sediment increases as the supply of oxygen increases by the presence of microbubbles. In this manner, the sediment layer enters an oxidizing condition and becomes capable of retaining released P. Increasing the DO concentration and ORP of the water body may restrain the release of T-P and SRP (Hupfer & Lewandowski 2008; Li et al. 2016).

In our experiment, the DO in the water body without MF gradually decreased as the release test proceeded until the DO reached almost 0 at most sites, as shown in Figure 7. In the water body injected with microbubbles, DO tended to temporarily rise after air bubble injection. Even after the P release test, the water body remained aerobic. Likewise, the ORP tended to be maintained to some extent by increasing after the microbubble was injected. We determined that the higher the ORP, the greater the decrease in the amount of released P. We hypothesize that P release was suppressed as soluble P in the sediment was adsorbed by aluminum oxide due to oxidation from the injection of air bubbles. Based on the results of our P release test, P in the sediment was removable not only through the DS capping effect, but also through an increase in DO concentration and ORP supplied to the water body. We attribute this to the very wide specific surface area of the microbubbles, which dissolved oxygen of air bubbles and changed the water body from anaerobic to aerobic. An important point to consider here is that DO concentration and ORP are not continuously maintained, but tend to decrease gradually over time. This shows that although the effect of decreasing the elution due to the increase of DO and ORP in this experiment is clearly shown, the adsorption of P from the sediment by the aluminum film layer has a greater effect.

Figure 7

Change of ORP and DO concentration during the P release test.

Figure 7

Change of ORP and DO concentration during the P release test.

Close modal

Behavior characteristics of P for MF

When a coagulant is added to a water body, two competitive reactions occur (Manamperuma et al. 2016). Setting aside the reaction between the condensed phosphate and organophosphate, the reaction we are concerned with involves the formation of aluminum hydroxide (aluminum hydroxide, Al(OH)3 and aluminum phosphate, AlPO4) where other compounds or ions are also involved. An Al:P stoichiometric reaction results in the formation of aluminum hydroxide phosphate crystals and precipitating particles. However, in this reaction, the Al:P molar ratio of 1:1 is difficult to achieve stoichiometrically, and the molar ratio between the aluminum ions (Al3+) injected into the treatment site and the removed P varies depending on water quality (Kwak et al. 2011).

The behavior of P transferred and migrated by flotation can be traced through the mass balance based on the measured values of this study. The sediment and water samples placed into the column, as well as phosphate ions, undergo a coagulation operation triggered by the dosed coagulant that attaches charged flocs on the surface of the microbubbles. At the same time, flotation separation is conducted through a series of steps such as collision with microbubbles, contact, and rise. Ultimately, P's fate is to either become the floated P of scum, the deposited P of sludge that never combined with air bubbles and subsequently settled, or P that remains P in treated water.

The particle layer on the initial sediment surface, in contrast, will arise an advection to the overlying water according to the hydraulic disturbance caused by the injection of microbubbles at the site. After a series of agglomeration, flotation, and sedimentation processes, it will again be deposited in the form of alum particles at the boundary film layer. The coating layer comprised of the alum particles exhibits a capping effect and may adsorb P released from the bottom sedimentary layer. Table 4 shows the amount of P distribution changed by MF in percentage.

Table 4

P distribution after MF

T-PFSDSCWUS
(%)
M-1 16.1 52.6 1.2 30.1 
M-2 16.2 76.3 2.5 5.0 
D-1 18.3 75.0 1.2 5.5 
D-2 12.5 70.5 2.1 14.9 
SRPFS (%)DSCWUS
M-1 61.0 37.1 1.6 0.3 
M-2 52.8 40.5 5.7 1.0 
D-1 51.5 43.6 2.9 2.0 
D-2 54.2 41.4 4.3 0.1 
T-PFSDSCWUS
(%)
M-1 16.1 52.6 1.2 30.1 
M-2 16.2 76.3 2.5 5.0 
D-1 18.3 75.0 1.2 5.5 
D-2 12.5 70.5 2.1 14.9 
SRPFS (%)DSCWUS
M-1 61.0 37.1 1.6 0.3 
M-2 52.8 40.5 5.7 1.0 
D-1 51.5 43.6 2.9 2.0 
D-2 54.2 41.4 4.3 0.1 

Our calculations confirmed that P was substantially removed from treated water. We also determined that some of the P contained in the sediment was re-deposited after treatment with MF. This P was never properly removed from the sediment, but in fact the re-deposited P manifested a capping effect in the sediment layer. Although the P release rate varied depending on environmental conditions specific to each site, the P release could be reduced by 49–84%, at least based on the terminated elapsed time of the release test. In sum, it proved possible to reduce the presence of nutrients in the water by removing P from the sediment through MF and by preventing the release of P from the sediment layer by capping. Table 5 and Figure 8 depict the behavior of P based on our MF experiments and P release test.

Table 5

Overall distribution and behavior of P in MF experiment

SitesT-P (g)
TotalFSDSCWUS
M-1 11,622 1876 6116 133 3496 
(water) 746 (water) 249 (water) 133 
(sediment) 1130 (sediment) 5867 (sediment) 3496 
M-2 6,723 1089 5,128 169 337 
(water) 348 (water) 115 (water) 169 
(sediment) 741 (sediment) 5013 (sediment) 337 
D-1 10,164 1274 7161 210 1519 
(water) 652 (water) 217 (water) 210 
(sediment) 622 (sediment) 6944 (sediment) 1519 
D-2 7,005 1285 5253 87 380 
(water) 367 (water) 122 (water) 87 
(sediment) 918 (sediment) 5130 (sediment) 380 
Average 8,878.5 1380.9 5914.5 149.9 1433.2 
Soluble P (g)
SitesTotalFSDSCWUS
M-1 1208.6 737.0 449.2 19.3 3.1 
(water) 723 (water) 241 (water) 19 
(sediment) 14 (sediment) 208 (sediment) 3 
M-2 759.0 401.2 307.1 44.0 6.7 
(water) 394 (water) 131 (water) 44 
(sediment) 7 (sediment) 176 (sediment) 7 
D-1 1073.7 581.9 444.4 46.8 0.6 
(water) 570 (water) 190 (water) 47 
(sediment) 12 (sediment) 254 (sediment) 1 
D-2 514.6 265.2 224.9 15.1 9.4 
(water) 255 (water) 85 (water) 15 
(sediment) 10 (sediment) 140 (sediment) 9 
Average 889.00 496.36 356.43 31.28 4.93 
SitesT-P (g)
TotalFSDSCWUS
M-1 11,622 1876 6116 133 3496 
(water) 746 (water) 249 (water) 133 
(sediment) 1130 (sediment) 5867 (sediment) 3496 
M-2 6,723 1089 5,128 169 337 
(water) 348 (water) 115 (water) 169 
(sediment) 741 (sediment) 5013 (sediment) 337 
D-1 10,164 1274 7161 210 1519 
(water) 652 (water) 217 (water) 210 
(sediment) 622 (sediment) 6944 (sediment) 1519 
D-2 7,005 1285 5253 87 380 
(water) 367 (water) 122 (water) 87 
(sediment) 918 (sediment) 5130 (sediment) 380 
Average 8,878.5 1380.9 5914.5 149.9 1433.2 
Soluble P (g)
SitesTotalFSDSCWUS
M-1 1208.6 737.0 449.2 19.3 3.1 
(water) 723 (water) 241 (water) 19 
(sediment) 14 (sediment) 208 (sediment) 3 
M-2 759.0 401.2 307.1 44.0 6.7 
(water) 394 (water) 131 (water) 44 
(sediment) 7 (sediment) 176 (sediment) 7 
D-1 1073.7 581.9 444.4 46.8 0.6 
(water) 570 (water) 190 (water) 47 
(sediment) 12 (sediment) 254 (sediment) 1 
D-2 514.6 265.2 224.9 15.1 9.4 
(water) 255 (water) 85 (water) 15 
(sediment) 10 (sediment) 140 (sediment) 9 
Average 889.00 496.36 356.43 31.28 4.93 
Figure 8

Conceptual diagram of P behavior in the stepwise experiment.

Figure 8

Conceptual diagram of P behavior in the stepwise experiment.

Close modal

Figure 9 shows the composition of the results of tracking the T-P and soluble P after treatment with MF. About 70–80% of T-P, excluding the P contained in DS that is not affected by the hydraulics of microbubble injection, was re-deposited on the bottom. It comes as no surprise, therefore, that T-P removal efficiency by MF treatment is not high. Soluble P (PO4-P and SRP), however, which is more easily absorbed by microalgae, floats to the surface and is removed at a rate of approximately 50–60%. In conclusion, although the amount of removed T-P by MF is not substantial, the removal of soluble P can nevertheless usefully suppress the proliferation of microalgae.

Figure 9

Distribution of T-P and soluble P after MF treatment.

Figure 9

Distribution of T-P and soluble P after MF treatment.

Close modal

We performed a series of flotation experiments that involved imagining a situation where MF is applied to remove P in the benthic sediment of the lake. The removal characteristics of T-P, PO4-P, and SRP were investigated for the amount of P contained in the sediment at each site, and the distribution and the behavior of P were evaluated based on the results of the flotation experiment. In addition, by the re-deposited sediment, we explored the additional capping effect that prevents P release from sediment, which occurs as a coating layer forms on top of the sediment by the coagulated flocs that were re-deposited after flotation separation.

Flotation performed on the samples of overlying water taken resulted in T-P achieving a removal efficiency of more than 73% across all sites and was removed up to 88%, while PO4-P achieved a removal efficiency over 92% across all sites. In the MF experiment of sediment, the T-P concentration of FS was much higher than DS due to P thickening to the scum layer that the particulate P attaches well to the injected microbubble, floats, and is concentrated with bubbles as scum on the water surface.

Due to the heightened DO concentration and ORP supplied by the microbubble injection, anaerobicization effectively did not occur after MF treatment. Among them, the DO concentration decreased to below 1 mg/L in the column not treated with MF at 5 days, whereas it seemed to decrease slowly in the column treated with MF. This shows that even with MF treatment, the DO concentration decreased over time, but the anaerobicization was not accelerated. The results of the P release experiment lead us to conclude that the DS layer that contains Al3+ ions prevented the P from being released from inside the sediment, i.e., P release was capped. Treatment with MF also resulted in not only flotation separation of P from water and sediment, but also water quality improvements, including DO and ORP. Over a longer timeline, the coating layer formed on the surface of the sediment by the accumulation of alum flocs may induce a secondary capping effect, as it acts as a kind of coating material that adsorbs and obstructs P release from bottom sediment.

By establishing the P change in the lake and the P contained in the sediment through the mass balance, it was possible to quantify the removed P and deposited again when microbubbles were injected. During flotation separation with microbubbles, the bubble-particle agglomerate floats to the surface of the water, where it is separated and removed. Evaluation of the mass balance revealed that while by MF treatment is not efficient at T-P removal, the process does effectively separate soluble P and remove it to the water surface as FS.

As such, in this study, P from the sediment and water can be effectively removed using MF. In addition, it was found that aluminum floc, which did not adhere to air bubbles, also precipitated at the bottom and had a capping action to adsorb P release from the sediment. If this technique is applied periodically in the actual field, it can show good effects, and more research is needed on this.

This study was supported by the National Research Foundation of Korea with grants (NRF: NRF-2019R1A2C1006441) from the Ministry of Education.

All relevant data are included in the paper or its Supplementary Information.

American Public Health Association (APHA), American Water Works Association (AWWA), & Water Environment Federation (WEF)
2012
Standard Methods for Examination of Water and Wastewater
, 22nd ed.
American Public Health Association
,
Washington D.C
Assessment and Remediation of Contaminated Sediments (ARCS)
1994
program remediation guidance document, Great lakes national program, US EPA (EPA 905-B94-003), Chicago, Illinois
.
Choi
Y. H.
,
Jeong
Y. H.
,
Yang
J. W.
,
Chae
I. S.
&
Kwak
D. H.
2019
Flotation and behavior characteristics estimation of phosphorus in sediments of lake using micro-bubble
.
J. Korean Soc. Environ. Eng.
41
(
6
),
328
336
.
Edzwald
J. K.
1995
Principles and applications of dissolved air flotation
.
Water Sci. Technol
31
(
3–4
),
1
23
.
Filippelli
G. M.
&
Delaney
M. L.
1996
Phosphorus geochemistry of equatorial Pacific sediments
.
Geochim. Cosmochim. Acta
60
(
9
),
1479
1495
.
Go
J.
,
Lampert
D.
,
Stegemann
J.
&
Reible
D.
2009
Predicting contaminant fate and transport in sediment caps: Mathematical modeling approaches
.
Applied Geochemistry
24
,
1347
1353
.
Hart
B. T.
,
Roberts
S.
,
James
R.
,
O'Donohue
M.
,
Taylor
J.
,
Donnert
D.
&
Furrer
R.
2003
Active barriers to reduce phosphorus release from sediments: effectiveness of three forms of CaCO3
.
Aust. J. Chem.
56
,
207
217
.
Hupfer
M.
&
Lewandowski
J.
2008
Oxygen control the phosphorus release from lake sediments–a long-lasting paradigm in limnology
.
Internat. Rev. Hydrobiol
93
(
4–5
),
415
432
.
Jellabukdo
2017
Research Project of Water Quality Characteristics and Management in Saemangeum Lake
, pp.
213
288
.
Kang
S. Y.
,
Kim
H. J.
,
Kim
T. I.
,
Park
H. J.
,
Na
C. K.
&
Han
M. Y.
2016
Remediation of sediments using micro-bubble
.
J. Korean Soc. Envion. Eng.
38
(
8
),
420
427
.
Kim
M.-S.
&
Kwak
D.-H.
2019
Role and applicability of bubble flotation model in sediment remediation in natural water systems
.
Separation and Purification Technology
222
,
309
320
.
Kim
S. H.
,
Kim
K.
,
Lee
M.
,
Jeong
H.-J.
,
Kim
W.-J.
,
Park
J.-G.
&
Yang
J. S.
2009
Enhanced benthic nutrient flux during monsoon periods in a coastal lake formed by tideland reclamation
.
Estuar. Coast.
32
,
1165
1175
.
Kwak
D. H.
,
Kim
S. J.
,
Jung
H. J.
,
Park
Y. K.
,
Yoo
Y. H.
&
Lee
Y. D.
2011
Particle separation and flotation efficiency by dissolved carbon dioxide flotation process
.
J. Korean Soc. Water Wastewater
25
(
4
),
471
478
.
Manamperuma
L. D.
,
Ratnaweera
H. C.
&
Martsul
A.
2016
Mechanisms during suspended solids and phosphate concentration variations in wastewater coagulation process
.
Environ. Technol.
37
(
19
),
2405
2413
.
Newell
R. C.
,
Seiderer
L. J.
&
Hitchcock
D. R.
1998
The impact of dredging works in coastal waters: a review of the sensitivity to disturbance and subsequent recovery of biological resources on the seabed
.
Oceanogr. Mar. Biol.: An Annual Rev.
36
,
127
178
.
Sadeghi
S.
,
Hua
G.
,
Min
K.
,
Johnson
T. J.
&
Gibbons
W. B.
2020
Phosphorus and cyanobacteria precipitation and sediment capping in lake water using alum and natural minerals
.
J. Env. Eng.
146
(
1
),
04019095
.
Thamdrup
B.
,
Fossing
H.
&
Jorgensen
B. B.
1994
Manganese, iron, and sulfur cycling in a coastal marine sediment, Aarhus Bay, Denmark
.
Geochim. Cosmochim. Acta
58
(
23
),
5115
5129
.
Zhou
J.
,
Li
D.
,
Zhao
Z.
,
Song
X.
,
Huang
Y.
&
Yang
J.
2020
Phosphorus immobilization by the surface sediments under the capping with new calcium peroxide material
.
J. Clean. Prod.
247
,
119135
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/licenses/by-nc-nd/4.0/).