Jun Ma*, Wei Liu
School of Municipal and Environmental Engineering, Harbin Institute of Technology, PO Box 2627, 202 Haihe Road,Harbin 150090, People’s Republic of China
Received 17 October 2000; received in revised form 1 June 2001; accepted 18 June 2001

Abstract
Jar tests were conducted to evaluate the effectiveness of potassium ferrate preoxidation on algae removal by coagulation. Laboratory studies demonstrated that pretreatment with potassium ferrate obviously enhanced the algae removal by coagulation with alum [Al2(SO4)3
18H2O]. Algae removal efficiency increased remarkably when the water was pretreated with ferrate. A very short time of preoxidation was enough to achieve substantial algae removal efficiency, and the effectiveness was further increased at a prolonged pretreatment time. Pretreatment with ferrate resulted in a reduction of alum dosage required to cause an efficient coagulation for algae removal. The obvious impact of cell architecture by potassium ferrate was found through scanning electron microscopy. Upon oxidation with ferrate, the cells were inactivated and some intracellular and extracelluar components were released into the water, which may be helpful to the coagulation by their bridging effect. Efficient removal of algae by potassium ferrate preoxidation is believed to be a consequence of several process mechanisms. Ferrate preoxidation inactivated algae, induced the formation of coagulant aid, which are the cellular components secreted by algal cells. The coagulation was also improved by increasing particle concentration in water, because of the formation of the intermediate forms of precipitant iron species during preoxidation. In addition, it was also observed that ferrate preoxidation caused algae agglomerate formation before the addition of coagulant, the subsequent application of alum resulted in further coagulation. r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Potassium ferrate; Algae; Preoxidation; Coagulation; Enhanced coagulation; Oxidation
1. Introduction
The eutrophication of surface water is a worldwide problem, which is increasing in significance. Eutrophication is caused by excessive inputs of nutrients, especially phosphorus, that stimulate nuisance growth of algae. The omnipresence of algae caused by the eutrophication
of surface water is the current and growing problem in the production of drinking water. To control the massive growth of algae in lakes and reservoirs, the impact of some chemicals such as copper sulphate, potassium permanganate, on algae were studied. Copper sulphate [CuSO4
5H2O] has been used to control nuisance algae in lakes and reservoirs for more than 80 years, and is considered to be an effective algicide available [1]. Potassium permanganate was also studied on the specific use as an algicide for reservoirs [2,3].
In drinking water treatment, conventional coagulation is still the main treatment process for algae removal. Whereas other treatment processes, for example, dissolved air flotation [4], sand filtration [5], direct filtration [6], which are aimed at algae removal, have also been researched. Massive growths of algae have caused many problems. Some algae cause uncomfortable tastes and odors, some algae cause filter clogging, some algae can penetrate the filter, leading to the deterioration of drinking water quality. Algae is also a precursor of
disinfection by-products. Algae removal from water treatment process is difficult because of their small size and the low specific gravity. It was reported that pre-treatment with oxidants may enhance the coagulation process and specifically enhance the removal of algae and other particulate matters in subsequent treatment steps [7–9]. The effects of chlorine, ozone and chlorine dioxide on Scenedesmus
sp. cultures were studied [10]. Algal cells activity and chlorophyll concentration decreased, and the concentration
of dissolved organic substances increased with increasing applied oxidant concentration. It was found that pretreatment with chlorine dioxide (1, 3 or 5 mg/l) or ozone (2.6, 4.6 or 8.1 mg/l) on algal cultures enhanced algal coagulation with aluminium sulphate, while prechlorination with 10 or 20 mg/l chlorine increased the required dosage of alum by 15%. However, the negative effect of using chlorine and chlorine dioxide resulting from the formation of by-products are limiting the use of these chemicals as pre-oxidants. In addition, it was recently recognized that the ozonation of waters containing bromide may lead to the formation of bromate at a level suspected of being hazardous to health, which is a negative aspect for using ozone as a preoxidant. Potassium permanganate has been investigated as an alternative preoxidant for the direct filtration of impounded surface water. The experiments of modified jar test apparatus and pilot plants showed that permanganate pre-treatment followed by coagulation with dual coagulants (ferric sulphate and cationic polymer) distinctly improved the particle and algae removal commonly achieved in direct filtration [6]. It is suggested that the common mechanism of algae removal by oxidant is the destruction of the algae architecture to various extent through different ways of oxidation.
Potassium ferrate [K2FeO4] is another strong oxidizing agent, which has a strong redox potential through the entire pH range, ranging from 2.2V in acid to 0.7V in base [11]. Several investigations have been conducted in applying potassium ferrate as a favorable alternative disinfection to chlorine for the disinfection of water and wastewater [12,13]. It is found that ferrate(VI) ion appeared to be an effective antifoulant [14], as only short contact times were required for ferrate concentration of 105M to control the biofilm growth. In addition, recent study found that ferrate treatment did not produce any mutagenic by-products during the treatment process [15]. The effects of pre-treatment with ferrate(VI) ion on algae coagulation by aluminium sulphate was investigated in this paper. Also, the effects of the pre-oxidation by ferrate on algae cell surface architecture and dissolved organic materials (DOM) of samples were studied by scanning electron microscopy (SEM) and UV spectrophotometer, respectively.
2. Materials and methods
2.1. Raw lake water
Raw water of shallow lake located in northeast part of China, which is in deep green color indicating high algal concentration, was selected in this study. Observation results by microscopy shows that the lake water principally contains green algae, such as Chlorella,
Spirogyra, Chlorococoum, Scenedesmus etc. The raw water quality was listed as follows: Turbidity 10–30 NTU; pH 7.5–7.7; temperature 15–181C; algal concentration 8106–2107 cells/l, CODMn(permanganate index) 10.5 mg/l (measured after filtration with 0.45 mm membrane).
2.2. Culture conditions
A solution containing cultured green algae species was also used in this study in order to overcome the influence of other materials except algae in natural surface water on experimental results. Algae species chosen in cultured process are Chlorococoum and Scenedesmus, because
they are commonly found in natural waters and are typical green algaes. It is also because that they are easily available and are easily cultured in the laboratory. The algae seeds were cultured in the plastic culture tank containing total 280 l modified inorganic nutrient solution (the concentration of inorganic salts that contained in the solution is listed in Table 1), in which 200 ml soil exudation liquid was mixed. The starting cell concentration in the water was 4106 cells/l at a constant temperature of 15711C, pH 7.3. Continuous light was provided by incandescent lamp and daylight lamp. A gas mixture (1% CO2 in air) was bubbled into the medium for a period of 15 min every other day. After grown for 25 days, algal concentration in nutrient

solution achieved 3.5–4.2108 cells/l, pH 9.1, turbidity20–40 NTU, CODMn(permanganate index) 5.2 mg/l(measured after filtration with 0.45 mm).
2.3. Coagulation tests
Standard jar tests were conducted in a mixer equipped with six-paddle jar test apparatus. The effects of various dosages of aluminium sulphate [Al2(SO4)3
18H2O] on cell coagulation with and without preoxidation was tested in six 0.5 l beakers. The pH of each 0.5 l cultured sample was adjusted to 7.1 with 1N HCl. Potassium ferrate [K2FeO4] was prepared by the modification of the method of reaction between OCl– and Fe(OH)3 (gel) in strongly basic media and isolated from saturated KOH solution [16]. The chemical solution was transferred
into six test tubes fixed on one stick and injected into the individual sample simultaneously by rotating the stick in order to minimize the systematic errors resulting from differences in the time of addition. A carefully calculated amount of potassium ferrate solution
was injected into beakers a certain time before the addition of alum solution. The freshly prepared alum solution using analytical reagent was predetermined (10 mg/l aluminium sulphate). During ferrate and alum addition, samples were stirred at 200 rpm for 1 min and then at 45 rpm for 10 min. Afterwards, samples were allowed to settle quiescently for 20 min. Thereafter, the upper 100 ml of the water sample was siphoned 1 cm below the water surface and taken for determination of residual algae cell concentration. In the case of lake water, the settled samples were further filtered with filter paper (1–2 mm pore size), and the residual algal concentration after filtration was also determined. Residual algal concentration after coagulation test was determined by microscope counting of cells.
2.4. Scanning electron microscopy
A blob of treated algae solution by potassium ferrate oxidation and the control algae solution without ferrate oxidation were dried for 2 h in the drying table, and they were gold-coated to a calculated coating thickness of 150nm by Eiko IB-3 ion emitting apparatus. Then, they
were examined in a Hitachi S-520 scanning electron microscope operated at 15 KV.
2.5. UVA at 254 nm and UVA scanning
The scanning of ultraviolet absorbance was performed at the wavelength ranging from 200nm to 320nm of untreated and treated cultured samples after filtered with a 0.45 mm cellulose acetate membrane filter. The absorbance of ultraviolet absorbance at 254nm was also determined. The absorbance of ultraviolet at 254nm by natural waters is a semi-quantitative indicator of the concentration of natural organic materials (NOM) in water. In water treatment practice, the use of absorbance at 254nm has been found to be useful for monitoring the concentration of DOC [17]. UV absorbance was also used to characterise NOM by the degree of its aromaticity. UVA at 254nm and UVA scanning were used in this study to indicate the variation of dissolved organic materials (DOM) in cultured solutions during different treatment processes.
3. Results and discussion
3.1. Effects of ferrate preoxidation on coagulation
Fig. 1 shows the effects of increasing ferrate concentration on algae removal by coagulation with alum. The removal efficiency is expressed as the ratio of the algal concentration before addition of chemicals to the algal concentration measured at the end of the coagulationsedimentation test. It can be seen that alum coagulation partially removed the algal cells in lake water, 20–30% of algae removal was achieved at the low alum dosage of 20–50 mg/l; 50% of algae removal was observed at higher alum dosage (e.g. 80 mg/l, see Fig. 1a). While in
the case of cultured solution, substantial algae removal was observed with alum coagulation alone in whole range of alum dosages (Fig. 1b). Thus the difference of algae removal between the case of coagulation alone and that with ferrate preoxidation was not very obvious for
cultured water. It is worth noting that there was a sharp increase of algae removal efficiency when the alum dosages were between 50–60 mg/l in lake water and 40–50 mg/l in cultured solution at the case with alum only. This indicates an optimum dosage range of alum for
effective coagulation of algal cells, due to the conceivable isoelectric point between the alum and algal cells. The figure shows that ferrate preoxidation has obvious effect on the coagulation of algae in either lake water or cultured solution. At any coagulant dosages
adopted in the tests the algae removal of settled samples pretreated with ferrate is higher than that without ferrate pretreatment. Even if ferrate dosage is only 1 mg/l, an obvious effect of algae reduction can be observed. Meanwhile, the removal efficiency increased gradually with the increase of alum dosage when pretreated by ferrate without an obvious isoelectric point especially in cultured solution, and this lead to a relative wider optimum coagulant dosage range for the removal of algae. Pretreatment with ferrate remarkably enhanced the algae removal, so that the alum dosage required for achieving a certain algae removal efficiency can be reduced. With the continuing increase of ferrate dosages, the residual algae removal increased further. Algal biocolloids carry negative surface charges at most pH levels [18], and the basic mechanism of iron- or


aluminium-hydroxide coagulation consists in mutual attraction and neutralization of the charge by the positively-charged hydroxide coagulant. While, it has been previously demonstrated that the NOM have a very strong influence on coagulation effectiveness. In
the presence of NOM, the coagulant reacts first with the free natural organic acids, e.g. humic acids and fulvic acids in waters, and only when the coagulant dosages are high enough to neutralize the surface charges of the organic materials can the coagulant take part in the electro-neutralization and bridging process [22]. Those observations lead to the conclusion that the NOM contained in surface water result in different algae removal between lake water and cultured solution which is lack of natural organic materials, when treated with alum alone (see Fig. 1). Meanwhile, the effects of ferrate preoxidation on algae removal in the case of lake water is more significant than that in cultured solution. It is believed that these effects are attributed to ferrate preoxidation, in which ferrate act as an aid to
coagulation processes [19].
The effect of various pretreatment times on algae removal by coagulation was tested (see Fig. 2). Algae removal efficiency increased obviously even in a short preoxidation time (as 1 min), and the removal efficiency further increases with the continuing extension of the
contacting time. Due to the slight increase of removal efficiency with the extension of pretreatment times (longer than 1 min), it was suggested that pretreatment with ferrate can influence the surface characteristics of algal cells in a very short time of oxidation and thus cause an enhanced coagulation.
In addition, the residual algae removal rate preoxidized with ferrate is further increased after filtered with 1–2 mm pore size filter papers, as shown in Fig. 3, indicating that the filtration process accentuated the effects of ferrate preoxidation on algae removal. The
algal cells after ferrate preoxidation and alum coagulation are easily intercepted by the filter, which otherwise did not precipitated within the sedimentation stage.
3.2. Effects of ferrate preoxidation on algal cell surface architecture
The scanning electron microscopy (SEM) was used to provide additional information that was essential to better describe the underlying process mechanisms. Typical appearance of Scenedesmus (Fig. 4a) under our cultivation condition showed four cells coenobia with two spines arising from the ends of each terminal cells. All four cells have approximately the same dimensions. The cells were enclosed by a sheath (here referred to as the reticulate layer) [20]. Randomly scattered warts on the cell surface were clearly seen, which is the typical
surface characteristic of Scenedesmus. Two striations as the longitudinal axis can be seen on the surface of each of the inner cells. Opposite spines regularly distribute around the spherical cell of living algae Chlorococoum (Fig. 4b). All two species of the above algal cells appeared in the pictures shrink a little compared to the living cells, because of the drying process during pretreatment needed for SEM tests. Results of the comparison of SEM micrographs between algal cells before and after pretreatment with ferrate demonstrated that the ferrate preoxidation induced a number of clearly discernible effects on algal behavior and cell architecture. The treatment caused the
release of intracellular component into the surrounding


medium (Figs. 5a and 6a). This phenomenon possibly caused by ferrate stimulation on algal cell or cleaved sheath by ferrate oxidation. Algae may release organic compounds into water, that are species-specific and growth phase-specific. It is reported that extracellular
organic matters (EOM) from cultures of green and bluegreen algae and diatoms behave like anionic and nonionic polyelectrolytes [21]. Therefore, it is suggested that algal biopolymers secreted in response to ferrate oxidation behave as a coagulant aid.
Another result of ferrate’s effect on algal cells was the intense sheath convolutions. SEM micrographs (Figs. 5b and 6a) showed that the cell surface architecture was eminent damaged, row organization of warts of Scenedesmus were not remained and the spines fell off
from the Chlorococoum.
What is special in potassium ferrate preoxidation is the formation of ferric hydroxide [(Fe(OH)3] colloids after it is decomposed. Fig. 5c shows that Fe(OH)3 possibly precipitates on the algae surface. These precipitates can obviously change algal surface properties. Once it is attached to the algal surface, the weights


of the algal cells is increased and the algae settling character is improved, which have been observed in the process of coagulation. In addition, the Fe(OH)3 colloids increased the concentration of particles in water, which is too low to cause effective coagulation.
Conglomeration of algal cells was also observed after pretreatment with ferrate (Fig. 6b). It was suggested that

preoxidation with ferrate enhanced the coagulation (Fig. 1) through the modification of algae envelope and their behavior as well, thus reducing the stability of algae colloids.
3.3. Effects of ferrate preoxidation on UVA
Ultraviolet absorbance was used to indicate the variation of concentration and the chemical changes of DOM before and after treatment, because DOM is one of the most important factors of water quality that affects coagulation. Fig. 7 shows the comparison of UVA at 254nm with and without alum coagulation in various preoxidation time. UVA at 254nm of the samples increased after ferrate oxidation at a very short contacting time (as 1 min, similar to Fig. 2), and varied smoothly with the continuing extension of oxidation time. However, UV absorbance decreased after following coagulation-sedimentation process, and it also varied slightly with the extension of oxidation time. It means that ferrate preoxidation possibly increased the dissolved organic concentration of cultured solution or changed the chemical structure of DOM. This result also suggests that the increment can be removed by following alum coagulation. Residual DOM concentration


preoxidized with ferrate followed by alum coagulation is lower than that without ferrate pretreatment. Corresponding to Fig. 7, scanning of UVA from 200nm to 320nm shows the same information (see Fig. 8). Figs. 7 and 8 together with Figs. 5a and 6a possibly support that algal biopolymers secreted in response to ferrate preoxidation, which may behave as a coagulant aid.
Fig. 9 shows the effects of ferrate preoxidation on algae removal and the variation of UVA at 254nm of filtered samples. It can be seen that the large change

of three curves occurred during the first 1 min of preoxidation time corresponding to the phenomenon that ferrate decomposed while its characteristic violet disappeared soon after injection. During the prolonged preoxidation time (e.g. longer than 1 min), algae removal
increased gradually comparing to the smoothly variation of UVA at 254 nm. It demonstrated that ferrate preoxidation inactivated the algal cells in a very short contacting time and then the algae architecture was destroyed (Fig. 5b), in consequence, the cellar components
were released to act as coagulant aid (Figs. 5a and 6a), which largely enhanced the following coagulation. Induced coagulant aid and the ferric hydroxide [Fe(OH)3] colloids derived from the decomposition of ferrate caused the conglomeration of algal cells (Fig. 6b) in prolonged preoxidation time, leading to the primary coagulation of algae, which also enhanced the following algae coagulation by alum (Fig. 2).
4. Conclusions
Laboratory studies using algae-bearing lake water and cultured algae solution demonstrated that pretreatment with potassium ferrate obviously enhanced the algae removal by coagulation-sedimentation process with alum [Al2(SO4)3
18H2O]. Algae removal efficiency
increased remarkably even at a short period of preoxidation time, and the efficiency was further increased at a prolonged contact time. To achieve a certain extent of algae removal, pretreatment with ferrate can reduce the dosage of alum required to cause
an efficient coagulation and filtration.
Efficient removal of algae caused by potassium ferrate preoxidation and coagulation with alum is suggested to be a consequence of several process mechanisms. Ferrate preoxidation inactivated algae, and also induced coagulant aid secreted by algal cells. Meanwhile, ferric
hydroxide derived from the decomposition of ferrate improved the coagulation condition by increasing particle concentration in water. In addition, ferrate preoxidation caused algae agglomerate formation before the addition of coagulant, and the subsequent application of alum resulted in further coagulation.
Acknowledgements
This work was supported by the National Natural Science Foundation of China under the scheme of National Science Fund for Distinguished Young Scholars (Project number 59825106 ).
References
[1] Elder JF, Horne AJ. Copper cycles and copper sulphate algicidal capacity in two Californian lakes. Envir Mgmt 1978;2:17–30.
[2] Fitzgerald GP. Laboratory evaluation of potassium permanganate as an algicide for water reservoirs. J South West Water Works Assn 1964;45:16–25.
[3] Kemp HT, Fuller RG, Davidson RS. Potassium permanganate as an algicide. J Am Water Works Assoc 1966;58:255–63.
[4] Bare WFR, Jones NB, Middleebrooka EJ. Algae removal using dissolved air flotation. J Water Pollut Control Fed 1975;47:153–69.
[5] Borchardt JA, O’melia CR. Sand filtration of algae suspension. J Am Water Works Assoc 1961;53:1493–508.
[6] Fetrusevski B, Van Breemen AN, Alaerts G. Effect of permanganate pre-treatment and coagulation with dual coagulants on algae removal in direct filtration. J Water SRT-Aqua 1996;45:316–26.
[7] Ma J, Graham N, Li GB. Effectiveness of permanganate preoxidation in enhancing the coagulation of surface waters-Laboratory case studies. J Water SRT-Aqua 1997; 46:1–11.
[8] Ma J, Li GB. Laboratory and full-scale plant studies of permanganate oxidation as an aid to the coagulation.Water Sci Technol 1993;27:47–54.
[9] Ma J, Li G, Chen ZL, Xu XG, Cai GQ. Enhanced coagulation of surface waters with high organic content by permanganate preoxidation. IAWQ/IWSA Joint Specialized Conference. Particle Removal from Dams &Reservoirs. 15–18, February, 2000, Durban, South Africa,2000.
[10] Sukenik A, Teltch B, Wachs AW, Shelef G, Nir I, Levanon D. Effect of oxidants on microalgal coagulation. Water Res 1987;21:533–9.
[11] Wood RH. The heat, free energy and entropy of the ferrate(VI) ion. J Am Chem Soc 1958;80:2038–41.
[12] Gilbert MB, Waite TD, Hare C. An investigation of the applicability of ferrate ion for disinfection. J Am Water Works Assoc 1976;68:495–7.
[13] Schink T, Waite TD. Inactivation of f2 virus with ferrate(VI). Water Res 1980;14:1705–17.
[14] Fagan J, Waite TD. Biofouling control with ferrate(VI). Environ Sci Technol 1983;7:23–125.
[15] Decula SJ, Chao AC, Smallwood C. Ames test of ferrate treated water. J Environ Eng 1983;109:1159–67.
[16] Goff H, Murmann RK. Studies on the mechanism of isotopic oxygen exchange and education of ferrate(VI) ion (FeO4 2). J Am Chem Soc 1971;93:6058–65.
[17] Owen DM, Amy GL, Chowdhury ZK, Paode R, McCoy G, Viscosil K. NOM characterization and treatability. J Am Water Works Assoc 1995;87:46–63.
[18] Ives KJ. Electrokinetic phenomena of planktonic algae. Proc Soc Water Treat Exam 1956;5:41–53.
[19] Ma J, Liu W. Enhanced coagulation of low temperature and low turbidity water by ferrate composite chemicals. Water Wastewater Eng 1997;23:9–11 (in Chinese).
[20] Staehlin LA, Pickett-Heaps JD. The ultra-structure of Scenedesmus (chlorophydeae). I. Species with the “reticulate” or “wary” type of ornamented layer. J Phycol 1975;11:163–85.
[21] Bernhart H, Clasen J. Flocculation of micro-organisms. J Water SRT-Aqua 1991;40:76–87.
[22] Narkis N, Rebhun M. The mechanism of flocculation processes in the presence of humic substances. J Am Water Works Assoc 1975;67:101–8.