Environ. Sci. Technol. 2003, 37, 5175-5180
Virus Inactivation in Aluminum and Polyaluminum Coagulation YOSHIHIKO MATSUI,* TAKU MATSUSHITA, SATORU SAKUMA, TAKAHITO GOJO, TEPPEI MAMIYA, HIROSHI SUZUOKI, AND TAKANOBU INOUE Department of Civil Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan
Inorganic aluminum salts, such as aluminum sulfate, are coagulants that cause small particles, such as bacteria and viruses as well as inorganic particles, to destabilize and combine into larger aggregates. In this investigation, batch coagulation treatments of water samples spiked with Qβ, MS2, T4, and P1 viruses were conducted with four different aluminum coagulants. The total infectious virus concentration in the suspension of floc particles that eventually formed by dosing with coagulant was measured after the floc particles were dissolved by raising the pH with an alkaline beef extract solution. The virus concentrations were extremely reduced after the water samples were dosed with aluminum coagulants. Viruses mixed with and adsorbed onto preformed aluminum hydroxide floc were, however, completely recovered after the floc dissolution. These results indicated that the aluminum coagulation process inactivates viruses. Virucidal activity was most prominent with the prehydrolyzed aluminum salt coagulant, polyaluminum chloride (PACl). Virucidal activity was lower in river water than in ultrapure waters natural organic matter in the river water depressed the virucidal activity. Mechanisms and kinetics of the virus inactivation were discussed. Our results suggest that intermediate polymers formed during hydrolysis of the aluminum coagulants sorbed strongly to viruses, either rendering them inactive or preventing infectivity.
Introduction In the treatment of drinking water, coagulation is an essential process for the removal of fine particulate matter that, owing to its small size, will not settle out from suspension by gravity. Destabilization and aggregation of the fine particulate matter into larger particulates by coagulation permit effective removal in the subsequent separation processes of sedimentation and/or filtration. When inorganic salts, such as aluminum sulfate, are used as coagulants, it is the intermediate polymers formed during the hydrolysis-precipitation reactions of the coagulant that are essential for the destabilization of the particles, i.e., bacteria and viruses, as well as inorganic particles. Viruses are also considered to be adsorbed onto the hydrolysis product (aluminum hydroxide) and, thus, may settle out during subsequent sedimentation process and/or become entrained in the filter media. This adsorption ability of aluminum hydroxide has been used to concentrate viruses found in water samples (1). Aluminum * Corresponding author phone: +81-58-293-2429; fax: +81-58239-0163; e-mail:
[email protected]. 10.1021/es0343003 CCC: $25.00 Published on Web 10/09/2003
2003 American Chemical Society
chloride or preformed aluminum hydroxide precipitate is added to the sample, and viruses are allowed to adsorb onto the aluminum hydroxide precipitate. The virus-containing precipitate is then collected by centrifugation or filtration. The viruses are eluted from the precipitate with an alkaline and/or proteinaceous solution before virus assay. Here we report that when solutions containing viruses were dosed with aluminum coagulants and mixed, the virus concentrations were extremely reduced. The results were interpreted to mean that the intermediate polymers formed during hydrolysis of the aluminum coagulant are able to sorb strongly to viruses, either rendering them inactive or preventing infectivity. Virus inactivation by disinfectants, such as chlorine, ozone, chlorine dioxide, and chloramines, has been widely studied. However, this is the first report of aluminum coagulants inactivating viruses.
Materials and Methods Viruses. To discuss the impact of aluminum coagulants on viruses, bacteriophages were selected as model viruses as they serve as ideal indicators of viral pollution and are models for enteroviruses or the hepatitis A virus in natural waters (2, 3). Seeding studies of bacteriophages Qβ (NBRC20012), MS2 (NBRC20015), P1 (NBRC20008), and T4 (NBRC20004), obtained from NITE Biological Resource Center (NBRC), were conducted. These viruses were chosen for their different compositionsswhether they had single- or double-stranded nucleic acid (DNA or RNA), and whether they were with or without an envelope. Qβ and MS2 viruses have singlestranded RNA and no envelope. They are similar in size (approximately 23 nm), shape (icosahedron), and nucleic acid (single-stranded RNA) to the hepatitis A virus and poliovirus, which have attracted attention in the treatment of drinking water. The P1 virus consists of a head (approximately 65 × 65 nm) and a tail (approximately 12 × 150 nm), whereas the T4 virus consists of a head (approximately 65 × 95 nm) and a tail (approximately 110 × 25 nm) with six tail fibers. Both of these viruses have double-stranded DNA and an envelope. Virus Solutions. Each virus was grown for 7-8 h at 37 °C in a bacterial host culture of Escherichia coli K12F+(A/λ) (obtained from Prof. S. Ogaki, University of Tokyo). The virus cultures were centrifuged at 3000×g, followed by filtration through a 0.2-µm-pore-size membrane filter (cellulose acetate; DISMIC-25cs; Toyo Roshi Kaisya, Ltd., Tokyo, Japan) to prepare the stock solutions. The concentrations of the prepared stocks of Qβ, MS2, and P1 viruses ranged from 1.5 × 1011 to 3.2 × 1012 PFU/mL. The T4 virus culture solution was further filtered through an ultrafiltration membrane (molecular weight cutoff: 20 kD; Centricon Plus-20; Millipore Co., Bedford, MA) for the concentration of the virus, and the retentate was recovered in order to prepare the T4 virus stock. The concentration of the T4 virus stock was about 2.7 × 1010 PFU/mL. Ultrapure water (Milli-Q Gradient A10 System; Millipore) and river water (Kiso River, Aichi, Japan; water quality shown in Table 1) were seeded with the stock solutions to prepare virus-spiked water samples for the experiments: the resulting virus concentrations of the virus-spiked water samples were in the range of 0.8 × 106-3.9 × 106 PFU/mL. Although the virus culture solutions contained much organic matter, such as peptone and glucose, which could affect the performance of aluminum coagulation, the high degree of dilution meant that this additional organic matter made very little difference to the total amount of organic matter in the virus-spiked water samples: the TOC contents of Qβ, MS2, and P1 virus stock solutions were about 6000 mg/L; the VOL. 37, NO. 22, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5175
TABLE 1. River Water Quality (Kiso River, Aichi, Japan) pH turbidity alkalinity TOC UV abs. (260 nm) conductivity
7.15 1.1 NTU 21.7 mg/L as CaCO3 2.2 mg/L 0.03 cm-1 8.3 mS/m
increases in TOC by spiking the water with the virus stock solutions were theoretically less than 0.06 mg/L: the TOC contents of the T4 virus stock solution were about 380 mg/L; and the increases in TOC were theoretically less than 0.01 mg/L. Besides the ultrapure water and the river water, three artificially prepared waters for virus seeding studies were prepared. The ionic constitution of the river water by ion chromatography (IC7000; Yokogawa Analytical Systems, Inc., Tokyo, Japan) was analyzed. Water containing ions of the same constitution as the river water were then prepared by adding reagent-grade cation and anion salts to ultrapure water (this water is designated “ionic water”). The two other waters were prepared after subjecting river water to ultrafiltration (UF; nominal molecular weight cutoff of 1 kDa; YM1; Amicon, Beverly, MA). The second water (designated “UF permeate”) was the filtrate resulting from the ultrafiltration. The third water (designated “UF retentate”) was that part remaining after UF, adjusted by dilution and addition of reagent-grade cation and anion salts to contain the same ionic constituents and dissolved organic matter as the UF permeate (as TOC; Model 810; Sievers Instruments, Inc., Boulder, CO). Viral Assays. The viruses were enumerated according to the PFU method with the bacterial host E. coli K12 F+(A/λ). Host cells were prepared by inoculating 15 mL of broth (polypeptone, 1.0 g; yeast extract, 0.5 g; glucose, 0.15 g; NaCl, 0.5 g; MgSO4‚7H2O, 0.02 g; MnSO4‚5H2O, 0.005 g; ultrapure water, 100 mL) with the host picked from an agar plate with a sterile platinum loop. Broth cultures were grown for 3-4 h at 37 °C (with shaking) to a concentration of 107-108 CFU/ mL. A 0.1-mL sample of the water to be assayed was poured onto a solid bottom-agar plate (polypeptone, 1.0 g; yeast extract, 0.5 g; glucose, 0.15 g; NaCl, 0.5 g; MgSO4‚7H2O, 0.02 g; MnSO4‚5H2O, 0.005 g; agar, 0.9 g; ultrapure water, 100 mL). Immediately after that, 0.3 mL of the host culture was added to 3 mL of molten (45 °C) top-agar (polypeptone, 1.0 g; yeast extract, 0.5 g; glucose, 0.15 g; NaCl, 0.5 g; MgSO4‚ 7H2O, 0.02 g; MnSO4‚5H2O, 0.005 g; CaCl2‚2H2O, 0.09 g; agar, 0.5 g; ultrapure water, 75 mL) in a test tube; the tube was then vortexed and poured onto the solid agar plate where the water sample had been poured. The agar plate was mixed and incubated for 16-24 h at 37 °C. Triplicate plating was performed for each water sample, with countable numbers of plaques ranging from 30 to 300 per plate. Coagulants and Preformed Aluminum Hydroxide Floc. The coagulants used were two laboratory-made aluminum solutions and two commercial aluminum coagulant products, the latter of which are actually used in drinking water treatment plants. The laboratory-made aluminum solutions were prepared by dissolution of reagent-grade aluminum sulfate or aluminum chloride (Wako Pure Chemical Industries, Ltd., Osaka, Japan) into ultrapure water to concentrations equivalent to 500 mg-Al/L. The commercial aluminum coagulant products were solutions of liquid alum (Al content: 4.2 wt %; Sumitomo Chemical Co., Ltd., Tokyo, Japan) and polyaluminum chloride (PACl) (Al content: 5.3 wt %, 62.5% basicity; Sumitomo Chemical). The commercial aluminum coagulant products, soon after 1:100 dilution with ultrapure water, were injected into the virus solutions. 5176
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 22, 2003
A suspension of the inactive form of aluminum (aluminum hydroxide precipitate) was prepared for control testing. A solution of reagent-grade aluminum chloride was preneutralized with 1 M sodium carbonate at pH 7.5 for 3-19 days to prepare the inactive form of aluminum hydroxide floc (preformed aluminum hydroxide floc suspension, 30 mgAl/L). The preformed aluminum hydroxide floc suspensions were dosed with the virus solutions.
Experimental Section Batch coagulation tests were conducted with 500 mL of virusspiked water in a glass beaker. The vessel contents were mixed with an impeller stirrer. Four dosages of coagulants, 0.01, 0.1, 1.0, and 10.0 mg-Al/L (0.37, 3.7, 37, and 370 µmol/L), were used. After the addition of a coagulant, the pH was adjusted and maintained at 7.0 with NaOH, HCl, or H2SO4. With the 10-mg-Al/L dosage, the solution pH started to be adjusted soon after starting the coagulant addition to avoid the momentary but intense pH drop. As a result, pH did not drop to less than 5.0 after adding the coagulant. The reaction time with the coagulants was 60 min (5 min at 139 rpm, 55 min at 34 rpm). Samples were taken from the beaker at 0, 10, 20, 40, and 60 min during the reaction. The virus concentrations in the liquid phase were measured after separating the floc particles by 5 min centrifugation at 3000×g. The concentrations of viruses in the floc mixture (viruses surviving in the liquid-phase plus the viruses in the floc matrix) were measured after dissolution of the aluminum floc particles (aluminum hydroxide), which were expected to include infectious viruses. The floc was dissolved by raising the pH to 9.5 by vortexing it with an alkaline solution (sodium hydroxide + 6% beef extract) for 5 s. Control tests were conducted to confirm the recovery of infectious viruses surviving in the floc matrix. The viruses were spiked into and mixed with the preformed aluminum hydroxide floc suspension (30 mg-Al/L). The virus concentration in the centrifuged supernatant (3000×g, 5 min) was determined after 1-6 h mixing depending on viruses, and virus adsorption onto the preformed aluminum hydroxide floc particles was confirmed. The virus concentration in the floc mixture was determined after the pH was increased to 9.5 with the 6% beef extract and sodium hydroxide solution to dissolve the floc particles and to elute any adsorbed viruses from the floc particles.
Results Recovery of Viruses Adsorbed onto the Preformed Aluminum Hydroxide Floc. Figure 1 shows the concentrations of the four viruses in the floc mixtures and in the liquid phase 1-6 h after the virus solutions were mixed with the preformed aluminum hydroxide floc. The virus concentrations in the supernatants were more than 1.5 orders of magnitude less than the virus concentrations in the initial solutions: less than 3% remained in the liquid phase and more than 97% of the incoming virus levels was removed by mixing with the preformed aluminum hydroxide floc and by centrifugal separation. However, the virus concentrations in the floc mixtures, which were measured after raising the pH to 9.5 in 6% beef extract solution, were almost the same as each initial concentration (before mixing with aluminum hydroxide floc). Recovery of Viruses in Pure Water Treated with Coagulants. Figure 2 shows the negative logarithmic values of the reduced concentrations (concentration divided by initial concentration, C/C0) of the Qβ virus after dosing with the reagent-grade aluminum chloride and the PACl. At dosages g 0.1 mg-Al/L of the reagent-grade aluminum chloride, less than 1% (-log(C/C0))2) of the Qβ virus remained in liquid phase after 10 min of coagulation (Figure 2a). The virus
FIGURE 1. Recovery of viruses adsorbed onto preformed aluminum hydroxide floc. Values represent the mean and double standard deviation. Contact times of the viruses with the preformed floc were as follows: MS2, 2 h; Qβ, 6 h; T4, 6 h; P1, 1 h.
the 0.01-mg-Al/L dosage. The variations in concentration of Qβ viruses surviving in the floc mixtures, which were quantified after dissolving the floc particles, are shown as negative logarithmic values in Figure 2b. By dissolving floc particles it was able to quantify the viruses retaining their infectious ability after being adsorbed onto floc particles as well as the viruses remaining in the liquid phase of the floc mixture. Although the concentrations of viruses in the floc mixtures were larger than the concentrations of viruses in only the liquid phase of the floc mixtures, the concentrations did not recover to the levels they were before adding the coagulant. The prehydrolyzed aluminum salt coagulant, PACl, showed a higher virus reduction ability (Figure 2c,d) than aluminum chloride. In particular, the virus counts were reduced even at the lowest dosage of 0.01 mg-Al/L. Although experiments were conducted for all combinations of the four coagulants and four viruses, the results hereafter are presented summarized in the form of figures. The time-variation trends were similar for all the experiments; these are omitted because of space limitations. Figure 3 compares the virus concentrations in floc mixtures 60 min after adding the coagulants at dosages of 0.01 mg-Al/L (Figure 3a), 0.1 mg-Al/L (Figure 3b), and 1.0 mg-Al/L (Figure 3c). Reductions in concentration were observed for all four viruses; the reductions were consistently the largest for Qβ. Virus reduction was more prominent with the commercial PACl product than with the commercial alum, reagent-grade aluminum sulfide, or reagent-grade aluminum chloride. Recovery of Viruses in River Water and Artificially Prepared Water Treated with Coagulants. Figure 4 compares virus concentrations in pure water and in river water after adding coagulants. Experiments performed with river water showed lower virus reduction than with pure water. For example, adding 1 mg-Al/L of PACl resulted in a 1- to 2-log reduction of virus concentration in the river water, as compared to the 3- to 4-log reduction in ultrapure water (Figure 4d). Adding 0.1 mg-Al/L of aluminum chloride or PACl resulted in a 2- or 3-log reduction, respectively, in ultrapure water but showed no substantial reduction in river water. Figure 5 compares virus concentrations in ionic water, UF permeate, and UF retentate as well as in the river water and the ultrapure water. The percentage of virus remaining did not increase when ionic species of the same constitution as the river water were added to pure water (ionic water). The percentage remaining was still very lowsin the order of 10-3 (0.1%)swhich was much lower than the value observed in the river water. UF permeate and UF retentate, both of which contained ionic species plus natural organic matter, however, showed much higher percentages of viruses remaining than in the pure water or in ionic water.
Discussion
FIGURE 2. Reduced concentrations of the Qβ virus in liquid phase and in floc mixture after adding aluminum coagulants. Values represent the mean and double standard deviation. Symbols and aluminum dosages are as follows: 0, 0.01 mg-Al/L; [, 0.1 mg-Al/L; 2, 1 mg-Al/L; O, 10 mg-Al/L. concentration was further decreased over the remaining 50 min, and less than 0.1% (-log(C/C0))3) of the Qβ virus remained. No substantial virus reduction was observed at
For all the viruses tested, losses were observed after adding the aluminum coagulants and dissolving the floc, in greater or lesser degrees depending on the coagulant dosages, coagulant types, viruses, and solutions. It is shown below that the virus losses were due to virus inactivation as a result of adding the aluminum coagulant, supported by the evidence that neither the alkaline condition during floc dissolution nor the acid condition at the time of adding the coagulant was related to the virus losses. It is known that products of metal hydrolysis, e.g., aluminum hydroxide, are strong adsorbents, and their adsorption ability has been used to concentrate viruses in water samples (1). Aluminum chloride or preformed aluminum hydroxide precipitate is added to the water sample, and viruses are allowed to adsorb onto the aluminum hydroxide precipitate. The virus-containing precipitate is then collected by centrifugation or filtration. The viruses are eluted VOL. 37, NO. 22, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5177
FIGURE 3. Comparison of virus concentrations in floc mixtures 60 min after adding (a) 0.01 mg-Al/L, (b) 0.1 mg-Al/L, and (c) 1.0 mg-Al/L.
FIGURE 5. Comparison of reductions in concentrations of the Qβ virus in floc mixtures between river water (9), ultrapure water (O), ionic water (3), UF permeate (2), and UF retentate ([) after adding 0.1 mg-Al/L PACl. Values represent the mean and double standard deviation.
FIGURE 4. Reduced concentrations of the Qβ virus in floc mixtures after adding aluminum coagulants. Values represent the mean and double standard deviation. O, pure water; 2, river water. from the precipitate with an alkaline or proteinaceous solution. In our experiments, losses due to adsorption onto preformed aluminum floc particles were observed for the four viruses in the control experiments. However, when the floc particles were dissolved, the concentrations of viruses recovered were the same as the initial virus concentrations. This indicated that the viruses adsorbed onto floc particles retained their infectious capability and could be entirely recovered by floc dissolution. Therefore, the reductions in virus count that were observed in the floc mixtures in the experiments involving direct addition of coagulants were not due to the alkaline conditions during floc dissolution. If the reductions in virus count were due only to the floc dissolution process, the time-variation of virus concentrations would not have been observed (Figure 2). Therefore, it appears that the viruses were inactivated by the addition of aluminum salts. Some viruses are inactivated under acid conditions. The two commercial coagulants as well as the laboratory-made 5178
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 22, 2003
aluminum chloride and aluminum sulfate solutions are acidic, decreasing the pH of the virus-spiked solutions after they were added. However, in our experiments, the minimum pH recorded after adding the coagulant was no lower than 5.0, and the viruses used are acid resistant at pH 5.0 or higher. Moreover, notable virus inactivation was observed with PACl at a dosage of 0.01 mg-Al/L where the solution pH momentarily dropped to only 5.7, while the other coagulants at the same dosage momentarily lowered the pH to the same level but showed virus reduction to a lesser extent or almost no virus reduction. Therefore, the virus inactivation did not occur because of the pH drop caused by the coagulant addition. This was also supported by the results showing that virus concentrations continued to gradually decrease with time, even after 10 min, by which time the solution pH had been adjusted to 7.0 and kept at a constant value (Figure 2). Reagent-grade aluminum sulfate and aluminum chloride solutions both showed virus inactivation, which confirmed that the inactivations observed for the commercial alum products were not caused by any impurities therein. However, preformed aluminum hydroxide did not show virus inactivation. Therefore, these results suggest that some specific chemicals formed during hydrolysis of the aluminum coagulants were related to the loss of infectivity. While aluminum salts in water hydrolyze rapidly, eventually forming nearly inactive aluminum hydroxide at the end of the hydrolysis reaction, intermediate polymeric species formed during the rapid reactions of hydrolysis and precipitation are essential for particle destabilization in the coagulation process. Of the coagulants tested, virucidal
activity was most prominent for the polyaluminum chloride (PACl), a prehydrolyzed metal salt, which contains significant amounts of polynuclear aluminum hydrolysis products (4, 5). Polymeric species formed after adding PACl are more stable than those formed during the hydrolysis reactions after adding alum (5, 6) and destabilize suspended particles faster (7). Those studies indicated that the charge-neutralization capability of the intermediate polymers is larger during the hydrolysis-precipitation reactions of PACl than during the hydrolysis-precipitation reactions of alum. In our study, PACl showed not only higher virucidal activity overall but also virucidal activity at the lowest dosage of 0.01 mg-Al/L. At this dosage, much less or almost no virus inactivation was observed with the commercial alum, reagent-grade aluminum chloride, or reagent-grade aluminum sulfate. Polymeric species would not be formed with the other three coagulants at this dosage, because the concentration was much lower than the solubility limit (0.1 mg/L at pH 7) of amorphous aluminum hydroxide (8). However, PACl already contains polynuclear aluminum hydrolysis species, which could render viruses inactive, even at a concentration of 0.01 mg/L. Therefore, the virucidal activity of aluminum salts may be related to the formation of intermediate polymers during the hydrolysis-precipitation reactions of aluminum salts, and the high virucidal activity of PACl could be attributed to the characteristics and/or the larger amounts of intermediate polymers formed by dosing with PACl. Disinfectants, such as chlorine and ozone, are known to inactivate viruses, and virus inactivation by these agents is well studied (9). However, aluminum coagulants have not previously been reported to inactivate viruses. Virus inactivation by aluminum salts is expected to have different mechanisms from these oxidants, because the aluminum hydrolysis reaction is not related to oxidation/reduction and does not produce any oxidants or radicals. Although the experimental data are not presented here, it was confirmed that there was no radical involvement with the virus inactivation by the addition of the aluminum salts. Virus inactivation by aluminum salts might be a phenomenon similar to surface inactivation: exposing viruses to air-water or solid-water interfaces leads to inactivation of the virus particles (10, 11). Trowborst et al. (12) suggest that viruses in solution approach the air-water interface via convection and diffusion and that they are adsorbed onto the air-water interface and subsequently inactivated by forces deforming the virus particle. Because of the reconfiguration at the airwater interface, virus capsids have localized polar and nonpolar regions, which could result in loss of virus infectivity. A similar mechanism has been presented for virus loss at the interface where gas, liquid, and solid phases intersect (13, 14). Currently, it has not been confirmed that the loss of the viruses by aluminum coagulation treatment is the result of virus inactivation associated with the interface between polymeric aluminum species and water. However, it is wellknown that virus particles adsorb onto aluminum hydroxide floc particles (1). Therefore, the polymeric aluminum species sorbed the virus particles so strongly that the viruses might undergo structural modifications and were subsequently inactivated by the interface forces between the polymeric aluminum species and water. On the other hand, there is also the possibility that even after the floc dissolution, polymeric aluminum species still adsorb viruses, in particular the protein in the shell of the virus that is responsible for the adsorption onto the host, thereby preventing infectivity. Aggregation of viruses is another possible explanation for the loss of infectivity (15). In this case, it can be hypothesized that some viruses remain aggregated even after the floc dissolution. Because the PFU method counts an aggregate as one infectious particle, regardless of whether the constitutive viruses are all infectious or partly infectious (some
FIGURE 6. Time-course of inactivation rate coefficient. Symbols and aluminum dosages are as follows: 0, 0.01 mg-Al/L; [, 0.1 mg-Al/L; 2, 1 mg-Al/L; O, 10 mg-Al/L. are inactive), the number of infectious particles counted by the PFU method decreases not only when viruses are inactivated but also when they are aggregated. The framework for investigating the kinetic effects of virion-virion Brownian aggregation and inactivation on viral infectivity has been presented, and the relative importance of coagulation vs inactivation has been discussed elsewhere (15). The patterns of concentration decline shown in Figures 2, 4, and 5 seem to fit the virion-virion Brownian aggregation scheme (eq 1 (16)) better than first-order inactivation kinetics
1 N ) N0 4RkTN0 t+1 3µ
(1)
where N is the total number of virus particles (singlets plus aggregates); N0 is the initial number of virus particles; R is collision efficiency (reciprocal of stability ratio); k is Boltzmann’s constant; T is the absolute temperature; µ is the viscosity of the medium; and t is time. According to eq 1, however, the time required for the number of virus particles to be reduced to 1/100 by virion-virion aggregation is 200 days or more, depending on stability ratio. The experimental results, on the other hand, showed that the time needed for the 1/100 reduction was no more than 10 min. Therefore, adhesion between virions and the polymeric aluminum particles, rather than simple virion-virion aggregation, could be responsible for the loss of infectivity. The virus reductions were also modeled by first-order kinetics with a variable inactivation rate coefficient:
N ) e-Kt N0
(2)
The rate coefficient (K) is treated as a function of time because inactivation is probably brought about by intermediate polymers formed in the hydrolysis-precipitation reactions of aluminum salts and the effectiveness of these intermediate polymers varies with time. The results shown in Figure 2 (parts b and d) were analyzed, and the inactivation rate coefficients were obtained for each reaction period (0-10, 10-20, 20-40, and 40-60 min) (Figure 6). The inactivation rate coefficients were higher with higher dosage and were higher for PACl than for AlCl3. They were also high at the beginning of the coagulation reaction and decreased with time. This also supports our view that intermediate polymers formed in the beginning of the hydrolysis-precipitation reactions of aluminum salts are responsible for the loss of virus infectivity. During the course of the hydrolysisprecipitation reactions, the polymers become less active, eventually forming the nearly inactive form of aluminum hydroxide. Aluminum hydroxide formed after more than 3 VOL. 37, NO. 22, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5179
days did not show any virus inactivation effect (control experiment, Figure 1). Loss of infectivity was observed for all four viruses, but the degrees of reduction in infectivity were dependent upon the virus; the losses of infectivity were consistently largest for Qβ. Although the electrophoretic mobility of the viruses was not measured, the degrees of reduction in infectivity may be related to the surface-charge characteristics of the viruses (17). The protein coat surrounding a virus contains weakly acidic and basic functional groups that are ionized when suspended in water resulting in a net surface charge, and viruses have different surface-charge characteristics. For example, MS2 and Norwalk viruses differ significantly with respect to their electrostatic properties (18). Viral inactivation was depressed in the river water. The result shown in Figure 5 indicates that natural organic matter (NOM) in the river water depressed viral inactivation. NOM is a major contaminant of surface waters, and most of it can be removed with aluminum coagulation. NOM removal occurs by adsorption or charge neutralization by intermediate polymers formed during the hydrolysis-precipitation reactions of the aluminum coagulant, leading to the formation of aluminum-NOM precipitates (8). Therefore, in the presence of NOM, viruses and NOM compete to interact with the intermediate polymers of aluminum species. Thus, the virucidal capability could be reduced. Aluminum-NOM precipitate still retains the ability to adsorb viruses, which is used for concentrating viruses from natural water. The adsorbability of aluminum-NOM precipitate, however, would be too weak to inactivate viruses. This report has presented evidence demonstrating that aluminum coagulation possesses a disinfection efficacy and inactivates viruses. Although the ability of aluminum coagulation to inactivate viruses was lower in river water than in ultrapure water, our findings imply that not only the treated water but also the aluminum hydroxide floc sludge is rendered virus-free or has reduced infectious virus content. The marked virucidal activity observed for PAClsa prehydrolyzed metal coagulantsreveals another practical benefit of its use as a coagulant and gives us scientific insight into the high chargeneutralization capability of prehydrolyzed metal coagulants. We are still investigating the mechanism of this virus inactivation, but we feel that the intermediate polymers formed during hydrolysis of the aluminum coagulant are able to sorb strongly to viruses, either rendering them inactive or preventing infectivity.
ma (Department of Urban Engineering, University of Tokyo) who instructed us in the methods of enumerating coliphages. This research was partly funded by a Grant-in-Aid for Science Research from the Japan Society for the Promotion of Science (Grant No. 14350284) and by NGK Insulators, Ltd. (Nagoya, Japan).
Literature Cited (1) Standard Methods for the Examination of Water and Wastewater, 20th ed.; Clesceri, L. S., Greenberg, A. E., Eaton, A. D., Eds.; American Public Health Association/American Water Works Association/Water Environment Federation: Washington, DC, 1998. (2) Wentsel, R. S.; O’Neill, P. E.; Kitchens, J. F. Appl. Environ. Microbiol. 1982, 43, 430-434. (3) Chattopadhyay, D.; Chattopadhyay, S.; Lyon, W. G.; Wilson, J. T. Environ. Sci. Technol. 2002, 36(19), 4017-4024. (4) Bottero, Y. Y.; Bersillon, J. L. In Aquatic Humic Substances: Influence on Fate and Treatment of Pollutants; Suffet, I. H., MacCarthy, P., Eds.; American Chemical Society: Washington, DC, 1989; pp 425-442. (5) Van Benschoten, J. E.; Edzwald, J. K. Water Res. 1990, 24(12), 1519. (6) Klute, R. In Chemical Water and Wastewater Treatment; Hahn, H. H., Klute, R., Eds.; Springer-Verlag: Berlin, Heidelberg, 1990; pp 33-54. (7) Matsui, Y.; Yuasa, A.; Furuya, Y.; Kamei, T. J. Am. Water Works Assoc. 1998, 90(10), 96-106. (8) Letterman, R. D.; Amirtharajah, A.; O’Melia, C. R. In Water Quality and Treatment; Letterman, R. D., Ed.; McGraw-Hill: New York, 1999; pp 6.1-6.66. (9) White, G. C. Handbook of Chlorination and Alternative Disinfectants; Wiley: New York, 1999. (10) Adams, M. H. J. Gen. Physiol. 1948, 31, 417-432. (11) Grant, S. B.; List, E. J.; Lidstrom, M. E. Water Resour. Res. 1993, 29(7), 2067-2085. (12) Trowborst, T.; Kuyper, S.; de Jong, J. C.; Plantinga, A. D. J. Gen. Virol. 1974, 24, 155-165. (13) Thompson, S. S.; Flury M.; Yates, M. V.; Jury, W. A. Appl. Environ. Microbiol. 1998, 65, 1186-1190. (14) Thompson, S. S.; Yates, M. V. Appl. Environ. Microbiol. 1999, 64, 304-309. (15) Grant, S. B. Environ. Sci. Technol. 1994, 28(5), 928-933. (16) Elimelech, M.; Gregory, J.; Jia, X.; Williams, R. A. Particle Deposition and Aggregation; Butterworth: Woburn, MA, 1995. (17) Penrod, S. L.; Olson, T. M.; Grant, S. B. J. Colloid Interface Sci. 1995, 173, 521-523. (18) Redman, J. A.; Grant, S. B.; Olson, T. M. Environ. Sci. Technol. 1997, 31(12), 3378-3383.
Acknowledgments
Received for review April 3, 2003. Revised manuscript received September 5, 2003. Accepted September 9, 2003.
We thank Masahiro Otaki (Graduate School of Humanities and Sciences, Ochanomizu University) and Hiroyuki Kataya-
ES0343003
5180
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 22, 2003
