Phosphorus removal and recovery from dairy manure by electrocoagulation

Abstract

Livestock manure waste, containing high level of phosphorus (P) will impair water environment if it is directly discharged to water body or improperly land applied. P recovery from animal manure not only addresses its environmental concerns, also provides a value added product as a fertilizer for better managing P. Electrocoagulation (EC) is proved to be an efficient approach that can be applied in the municipal and industrial wastewater for P removal and recovery; however, its application in the livestock manure management is less researched, especially with different selection of the electrode materials. In this study, EC process was evaluated on P removal and recovery from dairy manure. Four commonly seen electrodes, including Al, stainless steel, low carbon steel and cast iron, were compared. The results showed low carbon steel achieved the most efficient P removal (96.7%). The average particle size of manure solids was increased from 32.2 to 126.9 μm. The simulation experiment suggested that iron release and hydrolysis, which then worked as coagulants, is the major mechanism for P removal. EC by low carbon steel is an effective method for P separation from liquid phase of dairy manure to solid phase.

---

Introduction

Dairy manure contains high level of total phosphorus (P), ranging from 4.1 g kg⁻¹ to 18.3 g kg⁻¹-dry matter.¹ The majority of the P is in the form of inorganic P, e.g., ortho-phosphate ion and its insoluble minerals with small particle sizes. The remaining part is organic P dominated by phosphate monoesters, phytate, and DNA-like P.² Application of animal manure on cropland recycles a substantial portion of P, and this practice can decrease the needs for chemical fertilizers, where P is ultimately originated from unsustainable phosphate rock mining. Due to the inefficient manure transportation resulted from high moisture content (ca. 90%), the area of cropland for manure application is limited; and due to the unbalanced nutrients ratio of nitrogen to phosphorus (N/P) compared to the requirement of plant uptake, there is a possibility of P overloading and leaching when nitrogen level is used as a benchmark for manure application. The accumulated P in soil gradually moves to deep soil layer and groundwater, and becomes unavailable to plants; or it moves to surface water via agricultural drainage, potentially deteriorating aquatic water environment by eutrophication.

Nutrients separation by solid–liquid separation (SLS) will improve the manure nutrient management and reduce its environment impacts, since the liquid fraction after SLS can be used locally as a more nutrients-balanced fertilizer; while the solid fraction enriched with dry matter and other nutrients can be exported and applied to croplands where the demand on fertilizer cannot be satisfied by local manure production. Typical SLS methods include sedimentation by thickener or lagoon, centrifugation by decanter centrifuges, drainage, and pressurized filtration. Recently, some carbon-based materials like graphene were tested for phosphate adsorption in aqueous solution,³ but the pretreatment and post-treatment requirements were still unknown when those materials were used with animal manure for nutrients separation, and the suitability of applying the materials with adsorbed phosphate on cropland is not yet clear. The mass distribution of P covers a wide range of particle sizes, and P can be concentrated in either liquid phase or solid phase after SLS depending on what method is used.⁴ Chemical coagulation (CC) and flocculation (lime, iron and aluminum salts, and organic flocculants), which coagulates colloids present in liquid and increases particle sizes, is demonstrated to improve SLS as pretreatment.^5–7 However, it may also increase anions concentrations such as chloride and sulfate in waste streams, depending on the types of salts dosed, and may acidify the effluent because of pH decrease as a result of metal hydrolysis. Alternatively, electrocoagulation (EC) may achieve similar SLS performance but overcome these drawbacks of CC. EC has simple equipment requirement and is readily to be automated, potentially decreases pathogens, reduces the chemical cost, and provides gentle mixing by gas bubbling occurring on electrodes. This method has been widely tested for chemical oxygen demands (COD) or total solids (TS) separation based on different types of waters and wastewaters because of those advantages.^8,9 Behbahani et al.¹⁰ treated synthetic phosphorus solution by EC and got phosphorus removal efficiency of 100% with aluminum and 84.7% with iron electrodes. Vasudevan and his colleagues^11,12 achieved 98% of phosphorus removal efficiency from drinking water with iron and zinc electrodes. However, the physiochemical conditions in dairy manure matrices are complex, such as the highly reducing environment, various types of phosphorus compounds, high solids content, and high contaminants levels, which may consume more energy and takes longer reaction time to reach a similar removal efficiency compared with other types of waste streams. It was reported¹³ that 83% of P was removed from anaerobically digested dairy manure by stainless steel AISI 304. Another field demonstration¹⁴ showed that EC assisted with chemical pre-treatment reduced 96% of total phosphorus from dairy lagoon effluent.

Electrodes consisting of iron (cast iron, carbon steels, and stainless steels) and aluminum are common electrode materials for EC. Due to different metals composition, surface condition, reduction activity, and catalytic performance on water electrolysis, these materials may perform different EC capabilities and thus may remove P at different rates and electrical power consumptions. However, there were no reports compared these four electrodes in animal manure treatment. Meanwhile, the suitability of EC on undigested dairy manure slurry for P separation is largely uncertain at this point because of limited amount of effort input, and is further complicated by the complex matrices of the media and by the end use of the separated solids and liquid. It is necessary to empirically assess the performance of P separation in dairy manure with different electrode materials in order to select an appropriate one for scale-up experiment. This study aims to select economically competitive and effective materials among those four common materials. Simulation experiments were used to compare with EC process with the selected electrode material, so that the major mechanism for P removal during EC was clarified.

Materials and method

Manure and materials

Dairy manure slurry was taken from the dairy cattle barn of University of Minnesota-Twin Cities. One liter of manure slurry and 2 L of tap water were well mixed in a 2-gallon pail. The mixture was sieved through a stainless steel mesh sieve with 2 mm opening size, followed with another sieve of 0.295 mm opening size. The filtrate portion was termed as liquid manure, stored at 4 °C and used within two months after preparation. Before experiments, the liquid manure was diluted at 1 : 1 ratio with tap water and adjusted pH to the desirable values by 5 N H₂SO₄ or 5 N NaOH. The composition of manure is shown in Table 1. ACS reagent grade sodium hydroxide was purchased from VWR International, Inc. (Bridgeport, NJ), and concentrated sulfuric acid from Fisher Scientific (Pittsburgh, PA).

Table 1 Some chemical/physical characteristics of dairy manure

Parameters	Unit	Value
a These data are phosphorus analysis for the liquid portion of solid–liquid separation at 3000 × g; and the terms of total and reactive are defined in Hach methods 10 210 and 10 209, respectively.
pH	—	6.4
Total solids	%	2.02
Volatile solids	%	1.46
VS/TS	%	72.3
Ash	%	0.56
Total COD	g L^−1	24.8
Total VFA	mg L^−1	2686
Total phosphorus^a	mg L^−1	67.5–101.0
Reactive phosphorus^a	mg L^−1	53.0–67.6
Total nitrogen	mg L^−1	913
Total ammoniacal nitrogen	mg L^−1	123
Conductivity	S m^−1	0.32
Color	—	Brown

---

EC reactor setup

EC was performed in an electrolytic cell placed inside a 1 L plastic container with 0.8 L of liquid dairy manure. Four electrode materials (purchased from McMaster-Carr, Chicago, IL), including stainless steel (AISI type 304), aluminum (alloy 6016), low carbon steel (grade 1018) and grey cast iron, were chosen for EC. Both the anode and cathode (dimension of 7 cm × 4 cm × 0.5 cm, and effective surface area of 38 cm²) were made of the same type of materials. Before experiments, electrodes were immersed in 1 N H₂SO₄ solution for 5 min and then washed with tap water followed by sandpaper treatment to remove surface contaminants. Electrode distance was kept for 4 cm. DC power supply (Circuit Specialists, CSI5003XE) was applied for electrolytic reactions. Current was kept constant during 100 min's EC, and the medium content was agitated by a magnetic stirrer bar at 150 rpm (Fig. 1). Anode deposits were scrapped when the voltage increased around 10 V.

Fig. 1 Electrocoagulation equipment.

Simulation of EC process

There are potentially three steps that may be involved in the P removal during EC process: anodic release and formation of coagulant, coagulation process, and electrokinetic movement and attachment induced by electric field.^15,16 In order to explore the mechanisms of P removal by EC, simulation experiments were conducted to separate each of these three possible steps and assess their effects during EC in liquid dairy manure. The baseline EC process with low carbon steel was setup as reference with the following conditions: an initial pH of 7.4, current of 0.6 A, and 100 min reaction at an agitation speed at 150 rpm. The observed anode weight difference before and after EC was 1.00 g. The total Fe amount in medium will be controlled and accumulated to 17.7 mmol.

CC was carried out to test the coagulation capacity of iron ions and their hydrolysis products. Ferric chloride hexahydrate (FeCl₃·6H₂O, solid) and ferric hydroxide (or hydrated iron oxide, Fe(OH)₃) of the designated amount were used as chemical coagulants. To prepare Fe(OH)₃, 17.7 mmol of FeCl₃·6H₂O powder was transferred to 20 mL of distilled water and the pH was adjusted to around 11 through titration of 5 N NaOH. The suspension was centrifuged at 3000g for 12 min and the supernatant was discarded. After adding another 20 mL of distilled water, the mixture was vortexed for 2 min and centrifuged again. This purification step was repeated for three times to prepare Fe(OH)₃ suspension without sodium or chloride ions. Those coagulants were then added to dairy manure for CC, followed by stirring via magnetic stirrer at 150 rpm for 100 min. To test the effect of electric field on P removal (electrokinetic removal) without coagulant generation, carbon cloth (7.5 × 4 cm) instead of low carbon steel was used as both anode and cathode. The conditions applied for the electrokinetic experiment were the same as the EC method aforementioned. Finally, CC and electrokinetic processes were combined to simulate EC process. 17.7 mmol of FeCl₃·6H₂O was dissolved into 10 mL distilled water whose final volume was 12.7 mL. Carbon cloth was used as electrode to provide electric field of 0.6 A. The FeCl₃ solution was added to manure solution from the middle of two electrodes every 5 min at a rate of 127 μL min⁻¹. Another simulation experiment was conducted with carbon cloth and 17.7 mmol of Fe(OH)₃ suspension prepared by the method aforementioned.

Solids and nutrients analysis

Manure samples were taken from electrolytic cells before and during the EC process. The liquid portion of the manure samples were separated by 3000g centrifugation for 12 min, and the supernatant was tested for the total P as follows. The supernatant was dried at 100 °C overnight and incinerated at 550 °C for 30 min. One mL of H₂SO₄ was used to dissolve the solids and 1 mL of NaOH was used to neutralize the solution. Total P was tested by ascorbic acid reduction method with Hach kits (TNT 845 and TNT 843, purchased from Hach Company, Loveland, CO). The whole samples without centrifugation were used for total iron analysis as follows. One mL of sample was dried at 100 °C overnight and then incinerated at 550 °C for 30 min, followed by 2 mL of H₂SO₄ solution to dissolve the solid and 2 mL of NaOH solution to neutralize the solution. Hach kits (TNT 858 and TNT 890) were used to digest and test for the total iron through phenanthroline method. Particle size distribution of manure samples was characterized by sequential filtration: 10 mL of well mixed sample was serial filtrated by sieves with different pore sizes (45 μm, 150 μm, 250 μm, 425 μm, 850 μm and 2000 μm). After particles with different sizes were separated, these solids were dried at 100 °C overnight and the weight was measured after cooled down to room temperature. Total P and total iron of each particle size were measured with the methods aforementioned.

Electrode surface analysis

Low carbon steel electrode surface was analysed before and after EC. The electrode surface images were taken by Scanning electron microscope (SEM, Hitachi S-4700). The roughnesses (R_A)¹⁷ of the electrode were tested by Tencor P10 Profilometer.

Calculations

P removal efficiency (η, in %) is calculated as in eqn (1):where C₀ is the initial total P concentration (in mg P L⁻¹) in the supernatant; C is the total P concentration (in mg P L⁻¹) in the supernatant at t min.

Phosphorus removal kinetics was evaluated by rate equations¹⁸ as below:

where t is the reaction time (in min); C is the phosphorus concentration (in mg P L⁻¹) at time t; and r is the reaction rate.

When the reaction follows first-order reactions, the reaction rate is r = k[C] (k is the reaction rate coefficient); and when the reaction follows the second-order, r = k[C]². The integrated first-order and second-order kinetic equations are as follows, respectively:

ln[C] = −kt + ln[C₀] (3)

The energy consumption is calculated as eqn (5):

where U is the voltage (in V), I is the current (in A), t is the EC operation time (in hour), and m is the removed phosphorus (in g).

Results and discussion

Phosphorus removal

Fig. 2 shows P (total P) removal efficiencies of EC with four electrode materials accompanying the metal ions release from electrodes to manure media (Fig. 3). The P removal efficiency with aluminium electrode reached 96.4% at 70 min. The iron-containing electrode materials achieved the removal efficiencies of 87.1% with low carbon steel, 76.9% with cast iron, and 90.9% with stainless steel at 100 min. Aluminum electrode showed a better removal efficiency than other materials during 40 min and 70 min. Aluminum is more electrochemically active than those iron-containing materials and therefore is more ready to be released as metal ions (Al³⁺) to media when subjected to electrolysis. The released Al³⁺ proceeds with hydrolysis and generates coagulants (Al(OH)²⁺, Al(OH)₂⁺ and Al(OH)₃), which destabilize manure colloids. Cathodic dissolution of aluminum material may generate an additional and substantial amount of coagulants that enhance P removal.¹⁹ However, the mass transfer of Al³⁺ ions and aluminum complexes was at a low rate, and thus, aluminum electrodes, both anodes and cathodes were heavily covered by electrode deposits which substantially decrease current density and increase power consumption. In order to keep the constant current of 1 A, voltage was increased from 34.0 V to 51.6 V between aluminum electrodes. When the electrodes deposits were manually removed, the current could be kept at the set level. Similar phenomenon was also described by Tchamango, et al.²⁰ when Al was used for dairy effluents treatment. Anode deposits were scrapped when the voltage increased around 10 V. For Al electrodes, the frequency of scraping was about 1 or 2 min a time while it was about 20 min for iron electrodes. On the other hand, higher frequency of scrapping assisted ion releasing to the bulk liquid and more coagulants were generated, which is another reason that Al electrodes had higher phosphorus removal efficiency than iron electrodes. However, due to the high frequency of scraping deposits off the electrodes, aluminum is not considered proper electrode material for dairy manure in practice.

Fig. 2 Total phosphorus removal efficiency with different electrode materials (current = 1 A, initial pH = 6, initial phosphorus concentration = 100 mg L⁻¹, agitation speed = 150 rpm, retention time = 100 min).

Fig. 3 Total iron or aluminum concentration in manure during EC (current = 1 A, initial pH = 6, initial phosphorus concentration = 100 mg L⁻¹, agitation speed = 150 rpm, retention time = 100 min).

The releasing of metal ions from anode is a dominant factor for an effective EC process because those metal ions serve as the source of coagulants. As Fig. 3 shows, when applying low carbon steel electrodes, the total iron concentration in manure linearly increased throughout the testing duration, indicating that iron release rate from low carbon steel is stable over time. It has been reported that most P can be effectively removed when the molar ratio of iron to P is 1.5–3 : 1 in CC treatment in phosphate solution.^21–23 In our experiments, the Fe : P molar ratio was only 0.44 and 0.43 at 5 min when low carbon steel and cast iron were used in the current results, respectively. P removal remained at a high rate of 2.35 mg L⁻¹ min⁻¹ from the beginning to 20 min with low carbon steel, and it was 2.45 mg L⁻¹ min⁻¹ from 5 min to 20 min with cast iron. The molar ratio of Fe to P reached 1.55 for low carbon steel and 1.73 for cast iron at 20 min. Despite a small difference in ratio, these two electrode materials had a similar performance on P removal during the first 20, and 40 min as well. However, after around 55 min (shown as the cross of the two lines in Fig. 3), iron generated from low carbon steel electrode was more than that of cast iron electrode. That leads to a better P removal with low carbon steel as electrodes than that of cast iron after 55 min.

Stainless steel was used in many studies,^24–26 though, the iron content is only 53.48–74.5%.²⁷ Meanwhile, there are plenty of chromium (17.5–25%), nickel (8–15%) and other components, which can bring environmental problems. The coagulation effective metals (iron or Al) contents in low carbon steel (98.06–99.42%), grey cast iron (92.11–95.00%) or aluminum (96.00–97.35)²⁷ are higher than stainless steel. Stainless steel had the worst performance during the first 40 min among these three iron-contained electrodes, and iron released from stainless steel was less than those of the other materials at the same treatment time. Stainless steel released iron as well as the other components during EC, and catalysed oxygen evolution reaction; therefore the produced coagulants were not as many as the other electrodes. The ratio of iron to P was 0.24 at 10 min and 0.68 at 20 min. When iron concentration was higher enough (about Fe : P = 0.44) at some time between 10 and 20 min, P removal rate increased rapidly. Although the final P removal efficiency of stainless steel electrode was a little higher than those of low carbon steel and cast iron electrodes, its high contaminants content and high price discourage it as a good electrode material.

In the plot of ln C vs. t (Fig. 4), the data points for all the electrode materials show linear trends. The reaction rate coefficients for both first and second order model are summarized in Table 2. The coefficients of determination (R²) for the first-order model are above 0.92. For the second-order model, the R² of stainless steel and Al are 0.7971 while the calculated C₀ are negative values. The results show that the phosphorus removal process in this experiment was better described by first-order reaction kinetics. Low carbon steel has a larger reaction rate coefficient than that of cast iron.

Fig. 4 First-order kinetic plot (ln C vs. t) (current = 1 A, initial pH = 6, initial phosphorus concentration = 100 mg L⁻¹, agitation speed = 150 rpm, retention time = 100 min).

Table 2 Comparison between the experimental and calculated C₀ for different electrode materials in kinetic models

Electrode materials	C0 exp (mg L^−1)	First-order kinetics			Second-order kinetics
		k1	C0 cal (mg L^−1)	R^2	k2 × 10^3	C0 cal (mg L^−1)	R^2
Stainless steel	88.5	0.0236	102.2	0.9787	0.9960	−1534	0.8271
Al	101.8	0.0483	139.1	0.9611	0.3505	−38.5	0.7971
Low carbon steel	101.0	0.0202	93.0	0.9907	0.6380	162.1	0.9615
Cast iron	97.3	0.0146	84.7	0.9293	0.3420	94.0	0.9917

---

Solids analysis

Before EC, 80.0% of particles represented by total solids were smaller than 45 μm, and none particles were larger than 425 μm, due to the preliminary screening (through 295 μm) of solids from the manure slurry. Solids were initially suspended uniformly in liquid manure by visual inspection. After EC treatment, total solid content was decreased due to floating foam on manure surface and solids attachment on electrodes. Particles larger than 295 μm formed after EC. Low carbon steel achieved the results that 41.5% of the particles were larger than 45 μm, 9.62% of the particles were larger than 250 μm and the particles with size between 850 and 2000 μm were 0.10 g L⁻¹, and therefore showed the best performance on coagulation of increasing particle sizes. Al worked best on total solids removal. 58.1% of total solids were removed and 0.14 g L⁻¹ of total solids in the residual were between 850 and 2000 μm (Fig. 5).

Fig. 5 Particle size distribution before and after EC (current = 1 A, initial pH = 6, initial phosphorus concentration = 100 mg L⁻¹, agitation speed = 150 rpm, retention time = 100 min).

Fig. 6 shows the dry matter and selected chemicals distribution in raw manure and EC manure with low carbon steel as electrodes. In raw manure, 40.9% of the particles were between 25 and 45 μm while after EC treatment, most of the solids in this portion coagulated to form larger particles. After EC, the iron content in particles of different particle sizes varied between 67 and 393 mg g⁻¹ dry matter. P tended to be associated with particles smaller than 45 μm in raw manure; however, most of the P either directly reacted with iron ions or coagulated by coagulants after EC treatment. After EC, P content in particles larger than 45 μm was as high as 18.69 mg g⁻¹-dry manure matters.

Fig. 6 Particle size distribution and chemical content change before and after EC (current = 1 A, initial pH = 6.75, initial phosphorus concentration = 100 mg L⁻¹, low carbon steel, agitation speed = 150 rpm, retention time = 70 min).

Mechanisms of EC process

EC with low carbon steel as electrodes under 0.6 A performed very well, 96.7% total P was removed, and the final pH was 8.1. When dosing the same amount of iron salts or hydroxide (17.7 mmol) to manure without electric field, CC by FeCl₃·6H₂O only removed 13.1% of total P. After pH was adjusted to 8.12 at the end, total P removal efficiency reached to 96.4% that was close to what was achieved in EC. The reason for the low P removal before pH adjustment is that the hydrolysis of FeCl₃ in manure decreased the manure pH to 2.54, in which condition further hydrolysis of ferric was not allowed to proceed and phosphate salts turned to more soluble form. It was also reported in other papers^28,29 that the effective coagulating pH of ferric chloride is between 4 and 11. When the same amount of Fe(OH)₃ was used as coagulant, 54.6% of total P was removed and the final pH was 7.74 which was between 4 and 11. It indicates that during the process of coagulation, the hydrolysis of metal ion is a key step. Fe(OH)₃ has the ability of coagulating colloids which contain phosphate salts, and phosphate also can directly react with iron ions as well as complex with iron hydroxide (e.g. FeOH²⁺ and Fe(OH)₂⁺). Partial hydrolysis of Fe(OH)₃ to Fe(OH)₄⁻ may not have an positively effect on P removal since P-containing particles are considered negatively charged, too.

When using carbon cloth as electrodes, P was not removed; on the contrary, P concentration in the supernatant was increased from 57.1 mg L⁻¹ to 92.8 mg L⁻¹ due to the pH decrease from 7.40 to 6.34. During this treatment, advanced oxidation processes may occur and hydroxyl radical could oxidize the organic matter in manure into carbon intermediates or carbon dioxide.³⁰ The dissolved carbon dioxide may lead to the pH decline from 7.40 to 6.34, which released soluble phosphate from colloids of calcium phosphate. Therefore, electrokinetic process (and particles attachment to electrode) may occur when carbon cloth was used as electrodes but this process didn't remove P because of pH decrease.

Fig. 7 shows the results of simulation experiments of combining the iron salt dosing process and electrokinetic process. The combined treatment of FeCl₃ dosing and carbon cloth electrolysis was confirmed to have achieved good P removal rate, although either FeCl₃ dosing or carbon cloth electrolysis alone did not substantially remove P. It supports the assumption that electric field contributes to the release of phosphate from calcium phosphate particles, and phosphate is easier to be bound by hydrated iron oxide-hydroxide, therefore accelerating coagulation process. After about 45 min operation, pH value continued to drop to till 2.2. This pH decrease re-dissolved back part of the insoluble phosphate salts, which led to the slight increase of P concentration afterward. When adjusting the pH to 8.1 at 100 min, 98.6% total P was removed. This combination had even better performance on P removal than EC. That may be because ferric ion has a better coagulation effect than ferrous ion,²¹ while iron ions generated from low carbon steel during EC process was most considered ferrous in anaerobic condition.³¹ Simulation with Fe(OH)₃ removed P at a much slower rate and it had the worst P removal efficiency, which indicates the importance of the hydrolysis products of metal ions (e.g. FeOH²⁺ and Fe(OH)₂⁺) in the coagulation process.

Fig. 7 Phosphorus removal and simulation of EC process (current = 0.6 A, initial pH = 7.4, initial phosphorus concentration = 68 mg L⁻¹, agitation speed = 150 rpm, retention time = 100 min).

Fig. 8 shows the particle size distribution before and after treatment. Both ferric chloride and hydroxide coagulated particles, but it showed a decrease of solid size compared to that of EC. Other researchers³² also reported similar phenomenon that the CC aggregations were not as stable as EC and most of them were broken down during separation with sieves. The simulation with FeCl₃ had the best performance on coagulation, during which particles larger than 2000 μm were formed, which conforms to the fact that ferric is more effective coagulant than ferrous. The total solids of CC increased a lot due to the dosing of large amount of FeCl₃·6H₂O but EC process reduced the amount of total solids due to foam formation and floating on manure surface and solids attachment on electrodes.

Fig. 8 Particle size distribution before and after EC. “S.” indicated simulation (current = 0.6 A, initial pH = 7.4, initial phosphorus concentration = 68 mg L⁻¹, agitation speed = 150 rpm, retention time = 100 min).

Electrode surface analysis

The low carbon steel electrode surface images before and after EC are shown in Fig. 9. The inside of the electrode is defined as the surface of the anode that faced to the cathode, and the outside is defined as the other side of the anode. Iron on the surface of the anode was dissolved during EC. Most of the iron released to the bulk liquid was from the inside of the anode because there were more pits on the inside surface and they were deeper than those on the outside surface.

Fig. 9 SEM images of anode surfaces (tilted for 40°; (a) before EC; (b) after EC, inside surface; (c) after EC, outside surface).

The roughness (R_A) was 3.002 ± 0.711 μm before EC and it increased about 5 to 9 times after EC treatment. Since the iron dissolubility was different between the inside surface and the outside surface of the anode, the roughnesses after EC were different as 27.805 ± 10.844 μm and 14.197 ± 8.131 μm, respectively. The sacrificed anode needs to be replaced regularly. Using AC rather than DC or switching the anode and cathode would also help keep the electrode a longer lifetime.

Comparison to other EC and CC studies

Table 3 listed studies of EC or CC for phosphorus separation. Al and iron electrodes, such as stainless steel and mild steel are common electrode materials in EC studies and the phosphorus removal efficiency achieved more than 90% in most of the cases. However, the initial phosphorus concentration or total solid content in municipal or other wastewater are much lower than those in liquid manure. Therefore, it's more challenging and needs higher current or longer hydraulic retention time when treating with animal manure. The process designed in this study reached 96.7% P removal within 100 min, and the energy consumption was 583.3–826.5 W h g⁻¹-P removal. To decrease the energy consumption as well as increase the efficiency of EC, optimization of the operation parameters is needed.

Table 3 P removal efficiency of different wastewater treatment by EC

Effluence source	Experimental conditions	Electrode material or dosed chemical	Energy or coagulant consumption	P removal efficiency%	Ref.
a Chemical coagulation.
Phosphate solution	CD: 166.7 A m^−2	Cathode: Al	10.6 W h g^−1-P removal	98.2	10
	Initial PO4^3−: 400 mg L^−1
	Initial pH: 3	Anode: Al
	Duration time: 40 min
Synthetic wastewater	CD: 10 A m^−2	Al and iron	10.9 W h g^−1-P removal	90	34
	Initial PO4^3−: 100 mg L^−1
	Initial pH: 3
	Duration time: 12 min
Domestic wastewater	CD: 100 A m^−2	Cathode: Al	704 W h g^−1-P removal	98	35
	Initial P: 12.9 mg L^−1
	Initial pH: 7.8	Anode: Al
	Duration time: 10 min
Laundry wastewater	Voltage: 30 V	Cathode: Al	1631.6 W h g^−1-P removal	90.9	36
	Initial P: 27.6 mg L^−1
	Initial pH: 6–8	Anode: Al
	Duration time: 75 min
Mixed dairy wastewater	CD: 100 A m^−2	Cathode: iron	11.2–12.1 W h g^−1-P removal	93	37
	Initial P: 130–140 mg L^−1
	Initial pH: 6.0	Anode: Al
	Duration time: 60 min
Anaerobic digestion effluent	Current: 5.5 A	Cathode: stainless steel	—	78	13
	Initial P: 1760 mg L^−1
	Initial pH: 7.8	Anode: stainless steel
	Flow rate: 40 mL min^−1
	Duration time: 120 min
Dairy lagoon effluent	Initial P: 51.2 mg L^−1	Cathode: iron	—	96	14
	Initial pH: 7.8	Anode: iron
Wastewater	Initial P: 9.6 mg L^−1	Alum or FeCl3: 90 mg L^−1	Alum: 0.60 g Al/g-P removal	Alum: 89%	28^a
	Initial pH: 7.14
	Rapid mixing: 150 rpm, 1 min		FeCl3: 3.47 g Fe/g-P removal	FeCl3: 93%
	Slow mixing: 20 rpm, 20 min
	Settle: 30 min
Flushed dairy manure	Initial P: 255.8 mg L^−1	Alum: 216 mg Al/L	0.92 g Al/g-P removal	92	5^a
	Rapid mixing: 100 rpm, 2 min
	Slow mixing: 35 rpm, 15 min
	Settle: 30 min
Dairy manure	CD: 263 A m^−2	Cathode: low carbon steel	583.3–826.5 W h g^−1-P removal or	96.7	This study
	Initial P: 67.5–101.0 mg L^−1
	Initial pH: 7.4	Anode: low carbon steel	18.5–22.1 g Fe/g-P removal
	Duration time: 100 min

---

CC is a process similar to EC except for the source of coagulants. While achieving phosphorus removal efficiency of higher than 90%, CC needs less Fe or Al (3.47 g Fe g⁻¹-P removal, 0.60–0.92 g Al g⁻¹-P removal, respectively) than EC (18.5–22.1 g Fe g⁻¹-P removal in this study). This indicates that the utilization of iron or Al in EC is very low because lots of iron and Al dissolved from the anode is remained in the electrode deposits rather than released to the bulk liquid and generating coagulants. However, dosing chemicals like FeCl₃ or Alum may causes the pH decrease of the treated water and increases the cost by neutralizing it. On the other hand, the flocculation from the CC process is more fragile and loose than the sludge from EC. All of these can pose significant issues in the handling of the large amount of sludge. On the contrary, EC benefits for its low cost, less sludge generation, easy operation, and so on, and thus has a great potential in the animal manure management field.^29,32,33

Conclusions

This study showed that EC by low carbon steel is an effective method for P separation from liquid phase of dairy manure to solid phase. The low carbon steel was selected because it has high iron content, low contaminating metal components, and showed high and stable iron release rate. P removal can be well described by first-order kinetics and the removal efficiency reached 96.7% within 100 min. The average particle size of manure solids was increased from 32.2 μm to 126.9 μm due to EC and flocculation. P distribution analysis showed that most of the P was present as or bounded to particles larger than 45 μm after EC treatment. This particle size increase will assist solid–liquid separation of animal manure by conventional solid–liquid separation methods. Simulation experiments suggested that the release and hydrolysis of metal ions was the key and rate-limiting step in EC. Further study on operating parameters optimization based on low carbon steel is needed to implement the real applications of EC in P and fine particle separation.

Acknowledgements

The authors greatly appreciate the funding supports from the National Pork Board and Minnesota's Discovery, Research, and Innovation Economy (MnDRIVE) Bioremediation program.

References

1. Z. He, T. S. Griffin and C. W. Honeycutt, J. Environ. Qual., 2004, 33, 1528–1534.
2. P. H. Pagliari and C. A. Laboski, J. Environ. Qual., 2012, 41, 901–910.
3. S. Vasudevan and J. Lakshmi, RSC Adv., 2012, 2, 5234–5242.
4. K. Peters, M. Hjorth, L. S. Jensen and J. Magid, J. Environ. Qual., 2011, 40, 224–232.
5. K. Karthikeyan, M. Tekeste, M. Kalbasi and K. Grungor, Chemical Treatment of Dairy Manure Using Alum, Ferric Chloride and Lime, ASABE Paper, 2002.
6. D. Sievers, M. Jenner and M. Hanna, Trans. ASAE, 1994, 37, 597–601.
7. H. Lin, J. Gan, A. Rajendran, C. E. R. Reis and B. Hu, 2015.
8. D. Sharma, Membranes, 2014, 95, 19.
9. İ. A. Şengil, J. Hazard. Mater., 2006, 137, 1197–1205.
10. M. Behbahani, M. Alavi Moghaddam and M. Arami, Int. J. Environ. Res., 2011, 5, 403–412.
11. S. Vasudevan, J. Lakshmi and G. Sozhan, Desalin. Water Treat., 2009, 12, 407–414.
12. S. Vasudevan, G. Sozhan, S. Ravichandran, J. Jayaraj, J. Lakshmi and M. Sheela, Ind. Eng. Chem. Res., 2008, 47, 2018–2023.
13. W. S. University, Phosphorus and solids removal from anaerobic digestion effluent through electrochemical technology, United States Department of Agriculture, 2007.
14. S. Mukhtar, K. Wagner and L. Gregory, Field Demonstration of the Performance of an Electrocoagulation System to Reduce Phosphorus and Other Substances from Dairy Lagoon Effluent, Texas Water Resources Institute, 2009.
15. J.-H. Choi, S. Maruthamuthu, H.-G. Lee, T.-H. Ha, J.-H. Bae and A. N. Alshawabkeh, J. Appl. Electrochem., 2010, 40, 1101–1111.
16. G. Sturm, H. Weigand, C. Marb, W. Weiß and B. Huwe, J. Appl. Electrochem., 2010, 40, 1069–1078.
17. M. S. Lou, J. C. Chen and C. M. Li, Journal of Industrial Technology, 1998, 15, 1–6.
18. M. Al-Shannag, Z. Al-Qodah, K. Bani-Melhem, M. R. Qtaishat and M. Alkasrawi, Chem. Eng. J., 2015, 260, 749–756.
19. T. Picard, G. Cathalifaud-Feuillade, M. Mazet and C. Vandensteendam, J. Environ. Monit., 2000, 2, 77–80.
20. S. Tchamango, C. P. Nanseu-Njiki, E. Ngameni, D. Hadjiev and A. Darchen, Sci. Total Environ., 2010, 408, 947–952.
21. J. Barrow, H. Van Horn, D. Anderson and R. Nordstedt, Appl. Eng. Agric., 1997, 13(2), 259–267.
22. J. Sherman, H. Van Horn and R. Nordstedt, Appl. Eng. Agric., 2000, 16, 445–452.
23. J. DeBusk, J. Arogo Ogejo, K. Knowlton and N. Love, Appl. Eng. Agric., 2008, 24(4), 499–506.
24. A. A. Bukhari, Bioresour. Technol., 2008, 99, 914–921.
25. G. O. El-Sayed, M. S. Awad and Z. A. Ayad, Int. Res. J. Pure Appl. Chem., 2014, 4, 402.
26. D. Ghernaout, A. Badis, A. Kellil and B. Ghernaout, Desalination, 2008, 219, 118–125.
27. D. Hoban and L. BERG, J. Appl. Bacteriol., 1979, 47, 153–159.
28. J. M. Ebeling, P. L. Sibrell, S. R. Ogden and S. T. Summerfelt, Aquac. Eng., 2003, 29, 23–42.
29. S. Y. Lee and G. A. Gagnon, Environ. Rev., 2014, 22, 421–429.
30. A. Wang, J. Qu, L. Song and J. Ru, Acta Chim. Sin., 2006, 64, 767.
31. A. Gadd, D. Ryan, J. Kavanagh, A.-L. Beaurain, S. Luxem and G. Barton, J. Appl. Electrochem., 2010, 40, 1511–1517.
32. M. Y. A. Mollah, R. Schennach, J. R. Parga and D. L. Cocke, J. Hazard. Mater., 2001, 84, 29–41.
33. B. Zhu, D. A. Clifford and S. Chellam, Water Res., 2005, 39, 3098–3108.
34. Ş. İrdemez, N. Demircioğlu, Y. Ş. Yıldız and Z. Bingül, Sep. Purif. Technol., 2006, 52, 218–223.
35. F. Ozyonar and B. Karagozoglu, Pol. J. Environ. Stud., 2011, 20, 173.
36. F. Janpoor, A. Torabian and V. Khatibikamal, J. Chem. Technol. Biotechnol., 2011, 86, 1113–1120.
37. V. Kuokkanen, T. Kuokkanen, J. Rämö, U. Lassi and J. Roininen, Journal of Water Process Engineering, 2015, 8, e50–e57.

---