Article pubs.acs.org/est
Carbon Fiber as an Excellent Support Material for Wastewater Treatment Biofilms Shinya Matsumoto,† Akihito Ohtaki,‡ and Katsutoshi Hori†,* †
Department of Biotechnology, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan Teijin Ltd., WPT Project, Kasumigaseki Common Gate West Tower 2-1, Kasumigaseki 3-chome, Chiyoda-ku, Tokyo 100-8585, Japan
‡
S Supporting Information *
ABSTRACT: Fibrous materials made of carbon fiber (CF), aromatic polyamide (AP), preoxidized polyacrylonitrile (PAN), and polyethylene (PE), which are widely used in the textile industry, were evaluated as biofilm supports for wastewater treatment. We found that CF has a high capacity for adsorbing nitrifying bacterial sludge. The adhesion rate of four pure strainsCytophaga hutchinsonii, Alcaligenes faecalis, Bacillus subtilis, and Escherichia coliwas highest to CF. The ζ-potentials of the fibrous supports, and the cell surface potentials of these bacteria on the basis of the soft particle theory, were experimentally determined. Bacterial cell adhesion to the fibrous supports could be explained by Derjaguin− Landau−Verwey−Overbeek theory. Interaction energy profiles based on this theory indicated the disappearance of the energy barrier in bacterial cell adhesion to the CF support, whereas an insurmountable energy barrier was observed in the adhesion to the other fibrous supports. This result was attributed to the less negative ζ-potential of CF and the relatively large Hamaker constant for the CF/bacterium interaction in water; through simulations, the latter factor was found to make a greater contribution to lowering the energy barrier. In practice and theory, CF is an excellent material as a microbial and biofilm support for wastewater treatment.
■
organisms.5 In fact, fibrous supports have been used in bioprocesses for wastewater treatment, for example, in a biofilm reactor for ammonia removal from inorganic wastewater,5 in an attached-growth pond system to enhance organic carbon removal,6 and in a fiber-based biofilm reactor for denitrification of nitrate-contaminated groundwater.7 However, when choosing fiber materials in process design, little is known about their appropriateness as biofilm supports for wastewater treatment. The present study was aimed at identifying, based on a theoretical index for material selection, fiber materials that are appropriate biofilm supports for wastewater treatment. Four types of fiber materials that are widely used in the textile industry were evaluated in terms of their capacity for supporting microorganisms: carbon fiber (CF), aromatic polyamide fiber (AP), preoxidized polyacrylonitrile fiber (PAN), and polyethylene fiber (PE). PAN has been used as fibers in hot gas filtration systems, outdoor awnings, sails for yachts, and even fiber-reinforced concrete. CF, which is derived from PAN, has a high Young’s modulus and high thermal conductivity. Therefore, CF is used as a frame material in aerospace vehicles, airplanes, and automobiles, and has recently
INTRODUCTION Microorganisms tend to adhere to surfaces, to form a microcolony on them, and to develop a biofilm, which is an accumulated biomass of microorganisms and extracellular materials on a solid surface.1 Biofilms can, on the one hand, be detrimental to both human health and industrial processes, for example, causing infection, pathogen contamination, and slime formation, while on the other hand, be beneficial in environmental technologies and bioprocesses.2 In wastewater treatment systems, bioreactors implemented with biofilms are greatly advantageous in removal efficiency because they can retain a larger amount of microorganisms than conventional activated sludge systems can. Biofilms are used in several different reactor systems such as trickling filters, moving bed reactors, and rotating contactors.3 Although granules formed by microbial aggregates are often considered to be a type of biofilms, supports to which microorganisms can adhere have important roles in effective biofilm formation. Supports with a large surface area, such as porous and fibrous materials, have frequently been used for this purpose. On a porous support, however, biofilms grow excessively in the pores, clog them, and inhibit mass transfer of substrates and oxygen to the inner parts of the support, resulting in inactivation of microorganisms.4 On the other hand, flexible fiber textiles sway in water flow and thick biofilms that excessively grow on the fibrous support are likely to detach under high shear force, maintaining an appropriate thickness of the biofilms with active micro© 2012 American Chemical Society
Received: Revised: Accepted: Published: 10175
May 23, 2012 August 17, 2012 August 20, 2012 August 20, 2012 dx.doi.org/10.1021/es3020502 | Environ. Sci. Technol. 2012, 46, 10175−10181
Environmental Science & Technology
Article
microscopy confirmed that all the fibrous supports have no surface pores which may affect bacterial adsorption (SI Figure S1).
been used in implanted medical devices. AP is heat-resistant and is used in aerospace applications, in tires, and as a substitute of asbestos. PE is a thermoplastic polymer and is used in clothes or fishing nets. These materials have also been used as biofilm supports in bioprocesses including wastewater treatment. For example, in Japan, CF has been empirically used for water purification in many aquatic fields and model treatment systems8−10 since Kojima and co-workers found the high capacity of CF for adsorbing activated sludge,11 although the adsorbing mechanism remains unclear. CF has also been used in methane fermentation for degrading organic solid materials within anaerobic bioreactors12 or hydrogen fermentation in upflow anaerobic sludge blanket.13 PE and AP have been used in aerobic biofilter systems for high-loaded treatment of chemical enterprise sewage and of oil-contaminated recirculating water from nuclear power plants.14 In the present study, we demonstrated that CF is a superior material which exhibits an extremely high capacity for immobilizing activated sludge and nitrifying bacterial sludge and revealed the mechanism of adsorption of microbial cells onto CF on the basis of physicochemical interactions.
Table 1. Details of Fibrous Supports Used in This Study material carbon fiber (CF) aromatic polyamide (AP) polyacrylonitrile (PAN) polyethylene (PE) a
diameter (μm)
surface area per unit weight (m2/g)
zeta potentiala (mV)
7 12
0.33 0.22
−10.7 −21.3
12
0.22
−24.9
23
0.13
−34.5
pH 7.0.
The ζ-potentials of AP, PAN, and PE were determined by electroosmosis measurement on an electrophoretic lightscattering spectrophotometer (ELS-8000, Otsuka Electronics Co., Ltd., Osaka, Japan). An electrophoresis quartz cell was used for measuring the electrophoretic mobility (EPM) of polystyrene latex reference particles (500 nm in average diameter) that were coated with hydroxypropyl cellulose. Fibrous supports were fixed between the open sides of the cell. The cell was filled with an appropriate amount of 10 mM NaCl (pH 7). An electric potential in the range of 40−80 V was applied between two Pt electrodes mounted at the both ends of the cell. Electroosmotic flow exhibiting a parabolic velocity profile was observed in the case of a closed electrophoresis quartz cell. This profile was converted to electrophoretic mobility by Mori and Okamoto’s equation,16 and then further converted to the ζ-potential by Smoluchowski’s equation. The ζ-potential of the electrically conductive fiber, CF, was determined by the streaming potential measurement using Mü tek PCD-04 (BTG Instruments GmbH, Herrsching, Germany). The fibrous supports were cut into pieces (ca. 5 mm in length and 0.5 g in weight), suspended in 100 mL of deionized water, and poured into the measurement cell of the instrument. A hydraulic pressure was applied to the measurement cell and the resulting electric potential difference (i.e., streaming potential) created by the pressure drop through the measurement cell was measured with a pair of electrodes. To accurately calculate the streaming potential, the solution was titrated with 0.001 N poly(diallyldimethylammonium chloride) (Sigma-Aldrich, St. Louis, MO) to reach an end point of zero streaming potential. The ζ-potential can be calculated by using the Helmholtz−Smoluchowski equation.17
■
MATERIALS AND METHODS Microorganisms and Cultivation. Seed sludge was obtained from an aerobic basin of a municipal wastewater treatment plant. The seed sludge was cultivated in either organic synthetic wastewater (peptone, 0.48 g L−1; meet extract, 0.32 g L−1; urea, 0.08 g L−1; NaCl 0.024 g L−1; NaHPO4 0.08 g L−1; KCl 0.011 g L−1; CaCl2 0.011 g L−1; MgSO4 0.008 g L−1) or inorganic synthetic wastewater ((NH4)2SO4, 1.146 g L−1; CaCl2, 1.387 g L−1; FeSO4·7H2O, 0.0005 g L−1; and KH2PO4, 0.004 g L−1), for more than one month before the initiation of experiments. The sludge cultivated by the organic or inorganic synthetic wastewater was used as activated sludge or nitrifying bacterial sludge, respectively. Population of nitrifying bacteria in the sludge was quantified by fluorescent in situ hybridization (FISH) with probes of Fluorescein isothiocyanate-labbeld EUB338mix for total bacteria and Cy3-labbeled Nso190 for ammonia-oxidizing β-proteobacteria. Detailed procedure of FISH and information of the probes are found in elsewhere.15 FISH confirmed that ammonia-oxidizing β-proteobacteria dominated more than 80% in total bacteria when the sludge was applied to the experiments. Pure cultures of the following bacterial strains were used: C. hutchinsonii (NBRC15051), A. faecalis (NBRC13111), B. subtilis (NBRC13719), and E. coli (NBRC3301). Before adhesion tests, these strains were aerobically cultured for 1 day at 30 °C in a liquid medium containing 10 g L−1 tryptone, 5 g L−1 yeast extract, and 5 g L−1 NaCl. Cells were harvested in the exponential growth phase by centrifugation (8000g, 10 min) and suspended in phosphate buffered saline (pH 7.2). Fibrous Supports. The four types of fibrous supports used in this study were produced and supplied by Teijin Ltd. (Osaka, Japan). CF (HTA-12k; tensile strength, 400 kgf/mm2; tensile modulus, 24 × 103 kgf/mm2) used in this study was PAN-based pure carbon fiber, which was bundled with 12 000 threads of carbon fiber. CF was washed in deionized water several times before use for experiments to completely remove a hydrophilic sizing agent (mainly hydrophilic surfactant and polyvinyl alcohol) coated on CF. Diameters and surface areas per unit weight of the fibrous supports, which are cylindrical in shape, are summarized in Table 1. Observation by scanning electron
ζ = Uηκ /(pε)
(1)
where ζ is the ζ -potential, U is the electroosmotic velocity, η is the viscosity of the solution, κ is the conductivity of the solids, p is the hydraulic pressure, and ε is the dielectric constant of the solution. Because the ζ-potentials of the other three fibrous supports (i.e., AP, PAN, and PE) were obtained as described above and because they were correlated with the streaming potential according to eq 1, the ζ-potential of CF was calculated based on the calibrated equation of ζ-potential and streaming potential of the other three fibrous supports. Sludge Immobilization Tests. The sludge was suspended in a rectangular tank with a working volume of 10 L (30 cm wide, 20 cm high, and 25 cm long). The sludge concentration was adjusted to give mixed liquor suspended solids of 2000 mg/L. Fibrous materials were cut into segments of 20 cm in 10176
dx.doi.org/10.1021/es3020502 | Environ. Sci. Technol. 2012, 46, 10175−10181
Environmental Science & Technology
Article
length and were grouped into 10 g bundles. Six bundles of a fibrous material were fixed on one side to a stick (30 cm long) and hung from the top of the tank such that they were immersed in the sludge. Four sticks (one type of support material on each stick) were placed at the top of the tank in a single experiment. Aeration was conducted with two ballshaped air diffusers (2 cm in diameter), which were set in opposite corners on the bottom of the tank. Care was taken to prevent bubbles from shearing off the adhering sludge directly or from disturbing interaction of the sludge with the fibrous supports. At every sampling time point, a 5 cm length of the fiber bundle was cut off for sampling. The adhering sludge was detached from the fiber sample by sonication, suspended in distilled water, dried at 105 °C for 1 h,18 and weighed to measure the amount of suspended solid (SS) that adhered to the fibrous support during immersion in the sludge suspension. The dry weight of the fibrous supports after removal of the adhering sludge was also measured. Finally, the amount of adhered SS per unit surface area of the fibrous supports was calculated. Cell Adsorption Tests. Bacterial cells in nitrifying bacterial sludge and pure bacterial cells were harvested at the exponential growth phase by centrifugation (8000g, 10 min) and suspended in phosphate buffered saline (pH 7.2). When the effect of ionic strength (I) on bacterial adhesion was investigated, the harvested cells were suspended in solutions prepared by 10fold serial dilution of phosphate buffer (I = 202 mM).19 The prepared cell suspension (50 mL) was placed in a 100 mL beaker. The concentration of the cell suspension was adjusted to OD660 of 0.1. Each fibrous support was cut into segments of 3 cm in length. The pieces of each fibrous support were weighed to have a surface area of 5.0 × 10−2 m2, and then immersed in the cell suspension, which was agitated at 200 rpm at 30 °C. The adhesion rate of the pure bacterial cells to each support was calculated from the decrease in OD660 of each cell suspension. The adhesion rate constant k was defined by cell concentration C of the suspension (OD660) as follows.20
V (dC /dt ) = −kAC
total interaction energy = repulsion energy + attraction energy =πεrε0a[(Ψ1 + Ψ2)2 ln(1 + e−κ h) + (Ψ1 − Ψ2)2 ln(1 − e−κ h)] − Aa /6h ,
where εr is the relative permittivity of the suspension medium, ε0 is the vacuum permittivity, a is the radius of the bacterial cell, Ψ1 is the surface potential of the fibrous support, Ψ2 is the surface potential of the bacterial cell, κ is the inverse Debye− Hückel length, A is the Hamaker constant, and h is the distance of closest approach between the fibrous support and the bacterial cell. The Hamaker constant for two materials separated by water can be calculated as a function of the Hamaker constants of the individual materials in their condensed state. For a system where a bacterium is interacting with a solid surface in water, the Hamaker constant (Abws) can be written as23 Abws = (Abb1/2 − A ww1/2 )(A ss1/2 − A ww1/2 )
where Aij is the Hamaker constant of between materials i and j (j = b, w, s) and subscripts b, w, and s represent bacteria, water, and solid surface, respectively.
■
RESULTS AND DISCUSSION Ability of Fibrous Supports to Immobilize Bacterial Sludge. The time courses of the amount of adhered activated sludge to the fibrous supports are shown in Figure 1A. It was revealed that CF has an extremely high capacity for immobilizing activated sludge in comparison with the other three fibrous supports. The amount of adhered activated sludge to CF increased with incubation time and reached more than 22.5 g/m2 at 24 h. With regard to AP and PAN, the amounts of adhered sludge reached respective maximum values at 12 h, which were near half that of CF at 24 h, and remained constant thereafter. To PE, the activated sludge adhered only slightly. Also, nitrifying sludge, which is difficult to be retained in a reactor due to the low growth rate,24 was evaluated in terms of adhesion to these fibrous supports. As shown in Figure 1B, the nitrifying bacterial sludge also adhered most to CF among the fibers tested. The maximum capacity for supporting the nitrifying bacterial sludge was in the order of CF, AP, PAN, and PE, and that of CF was 3.0 g/m2, which was more than twice that of AP. To date, the highest level of immobilizing capacity for nitrifying bacterial sludge has been reported to be 2.5−3.0 g/m3 on ferro-nickel slag,25 and on polypropylene and polyethylene sheets modified with polyethylene glycol (PEG) chains with two different functional groups (−PEG-NH2 and −PEG-CH3).26 Thus, CF was shown to be a superior support that can efficiently immobilize nitrifying bacterial sludge, as well as activated sludge, and to be a promising material for efficient nitrification in wastewater treatment. In Figure 1B, initial adhesion of nitrifying bacterial sludge was not higher for CF than for PAN (at 1 h) and AP (at 1 to 3 h) although CF reversed the capacity for immobilizing the sludge with PAN and AP at 6 h. In addition, the amount of the sludge adhering to PAN decreased in the initial of the immobilization test. In these sludge-immobilization tests, a high concentration of the sludge samples (2000 mg/L) was used. Therefore, flocks and aggregates rather than single cells adhered to the fibrous supports and data could fluctuate because the effect of adhesion and detachment of relatively
(2)
Then, k = −(V /A)(1/t )ln(Ct /C0)
(3)
where V, A, t, Ct, and C0 are the volume of the cell suspension, the surface area of the fibrous support, the contact time, OD660 at time t, and initial OD660, respectively. Surface Charge of Bacterial Cells. EPM of the bacterial cells was measured on an electrophoretic light-scattering spectrophotometer (Sysmex, Hyogo, Japan). EPM measurements were conducted over a broad range of ionic strengths, from 5 to 200 mM KCl, at pH 7 and room temperature. Following the last rinse, the bacterial cell pellet was suspended in KCl solution of each ionic strength to obtain OD660 of 0.01. The measured bacterial EPM at each ionic strength was calculated as the average of three replicates. Between EPM measurements, the electrode set was thoroughly rinsed, first with 75% ethanol and then with deionized water. The surface charge of bacterial cells was calculated according to the soft particle theory.21 Calculation of Interaction Energy Profile. Interaction profiles are calculated from the following equation based on the theory.22 10177
dx.doi.org/10.1021/es3020502 | Environ. Sci. Technol. 2012, 46, 10175−10181
Environmental Science & Technology
Article
Surface Potentials of Fibrous Supports. To elucidate the underlying factors in the high capacity of CF for supporting microbial cells, the ζ-potentials of the fibrous supports were measured. Fibrous materials cannot be subjected to direct EPM measurement. Particles obtained by grinding the fibrous materials might have different surface properties and might show different electrokinetics from those of the original fibers. Therefore, their ζ-potentials were determined from the measurement of the EPM of reference particles, whose ζpotentials were zero but were moved by electroosmotic flow in the electric field within a cell which had fibers affixed to its side surface. Furthermore, because CF is a conductive material, its electrokinetics cannot be accurately measured from the electroosmotic flow. Therefore, the ζ-potential of CF was determined by measuring the streaming potential induced by a hydraulic pressure through a cell containing a suspension of cut CF fibers. Although the ζ-potentials of all four of the fibrous supports used in this study were negative, CF had the least negative ζ-potential among them (Table 1). Typical nitrifying bacteria, such as Nitrosomonas europaea and Nitrobacter winogradskyi, have cell surfaces with a negative charge at near neutral pH,25,26 causing repulsive force between the bacterial cells and most solid surfaces, which usually are negatively charged in water. Therefore, the nitrifying bacteria are considered able to more easily adhere to CF that has a less negatively charged surface than to the other fibrous materials. Surface Potentials of Pure Bacterial Strains. To clarify the details of the mechanism underlying the high capacity of CF to immobilize sludge containing microbial cells, we analyzed the interactions between the four fibrous supports and pure bacterial strains, which were subjected to EPM measurement to determine cell surface potentials. In Figure 3, EPMs of the
Figure 1. Time-courses of amount of adhered (A) activated sludge and (B) nitrifying bacterial sludge to fibrous supports. Mean value and standard deviation out of triplicate independent experiments are shown. Note that three samples were taken at a single sampling point of a single experiment.
large flocks on the experimental data would be large. Therefore, adhesion of bacterial cells in the nitrifying bacterial sludge, in which ammonia-oxidizing β-proteobacteria dominated for more than 80% of total bacteria, was examined by the adsorption tests in an agitated flask using a low concentration of the sludge (OD600 = 0.1,
