Incidence of an ozonation stage on the treatment of cherry stillage by activated sludge

This article presents a laboratory study of the ozonation of diluted cherry stillage, a high-strength wastewater. Influence of variables, kinetics, and the effects of an ozonation stage coupled with the biological treatment by activated sludge are addressed. Single activated sludge processing was shown effective to remove biological oxygen demand (BOD) and chemical oxygen demand (COD) but polyphenols were reduced to a lesser extent. On the other hand, direct wastewater ozonation did not reduce COD and total organic carbon (TOC) appreciably, and foaming problems were experienced when a high gas flow rate was applied. However, polyphenols and UV₂₅₄ absorbance decreased substantially by means of ozonation. To best achieve complete cherry stillage purification, two ways of coupling ozonation with activated sludge are proposed. Ozonation prior to activated sludge is advised for high-concentration wastewater to reduce polyphenol concentration, thus removing inhibiting effects. For wastewater with low polyphenol concentration the sequence activated sludge–ozonation–activated sludge is preferred to enhance the overall process performance in terms of oxidation efficiency and sludge settling.

Keywords: Activated Sludge;; Cherry Stillage;; Wastewater Treatment;; Polyphenol; Ozone;

Production of spirits by distillation of fruit and vegetables is an industrial activity of increasing interest. However, as in many other food processing industries, large amounts of solid and liquid wastes are produced. Liquid wastes, namely stillage, usually constitute a wastewater with high concentration of organic acids and phenols in addition to varying amounts of suspended solids and inorganic compounds. As a result, high values of chemical oxygen demand (COD) are typically found in these wastes ([26]). To avoid a negative environmental impact, an effective treatment of this wastewater must be carried out before its discharge to rivers and lakes.

Activated sludge processes and anaerobic digestion are usually considered the most cost-effective methods for the treatment of high-strength wastewater from food-related manufacturers ([21]; [22]). Anaerobic digestion is normally applied on site since the methane-containing biogas produced is sometimes used as fuel in the same processing plant. However, anaerobic digestion can hardly achieve an effluent with high quality and, therefore, tertiary treatment is usually required. When the factory is situated near a municipal wastewater treatment plant (MWTP), the most economic option to deal with this wastewater is to release to the MWTP where various combined wastewaters are treated in an activated sludge process and pay the appropriate charges for it ([16]). However, in many cases, the presence of recalcitrant or inhibitory substances (such as polyphenols and fatty acids in cherry stillage) makes conventional biological treatment methods inappropriate to achieve an effluent complying with the required standards ([15]). [25] reported four cases in which the combination of chemical and biological treatment processes was potentially advantageous to obtain a complete abatement of wastewater pollutants. Wastes containing recalcitrant and inhibitory substances together with biodegradable compounds require chemical pretreatment. According to previous investigations, ozone seems a good choice in combination with activated sludge ([1]; [6]). Major benefits of ozonation on assisting biological oxidation of complex wastewater derive from its high and selective oxidation potential toward olefins and aromatics. Thus, theoretically in the case study, it is expected that ozone would not degrade sugars of cherry stillage while attacking double bonds of unsaturated fatty acids and phenols, which would be a key factor in possible biodegradation inhibition.

A previous work showed the ability of ozonation to effectively degrade certain phenol-like compounds that have been demonstrated to be present in distillery wastewater ([8]). However, there is lack of data on how ozonation works when polyphenols are accompanied by other substrates, as in real wastewater, and how the ozonation step may induce biodegradability to these compounds in the wastewater to facilitate subsequent biological removal.

The specific purposes of this work were to study the ozonation of wastewater containing cherry stillage and to explain the effects of such ozonation on the subsequent biodegradation by activated sludge. The overall performance of the integrated process ozone–activated sludge was compared to that obtained by single biodegradation.

Raw wastewater was obtained from the spent wash generated from a distillery unit of a cherry-based spirits manufacture plant located at Valdastillas (Cáceres Province, Spain). Spent wash contained a fraction of solids that were settled down and separated from the liquid phase. After solids separation, wastewater was characterized to have the values in Table I.

TABLE I Main Features of the Wastewater Used in This Work

Sludge from a full-scale MWTP was used as inoculum to start-up the lab-scale activated sludge process. An acclimated microbial population was prepared. Microorganisms of the seed failed to acclimate to concentrated wastewater likely due to toxicity of organic compounds or metals to the aerobic population. Therefore, raw wastewater was diluted 1 to 80 by volume with tap water for acclimation purposes. Main features of diluted wastewater are also given in Table I. Acclimation was carried out in a batch reactor operating in fill and draw mode as described elsewhere ([10]). Ozone was generated from oxygen by means of a Sander 307 ozone generator able to produce 2.5% ozone by volume at the most.

Ozonation was performed in a glass bubble column (i.d., 9 cm; length, 45 cm) operated as an isothermal semibatch reactor. Typically, it was loaded with 1.5 L of the wastewater and the oxygen-ozone mixture was continuously supplied through a porous glass plate of 16–40 μm pore diameter. Recycle flow was used to provide good mixing within the reactor. Temperature control was performed by pumping water from a thermoregulated bath to the reactor jacket. Wastewater samples were periodically withdrawn to be analyzed. Ozone in the gas phase, both entering and leaving the system, was continuously monitored by a GM19 analyzer.

Aerobic wastewater treatment was carried out in a 3 L working volume glass digester operating as a sequencing batch reactor (SBR), each cycle comprising filling, reaction, settling, drawing, and sludge wasting stages. Agitation was provided by a mechanical stirrer, set at 500 rpm, thus achieving good mixing within the reaction mixture. Isothermal reaction conditions (20°C) were maintained by circulating water through a reactor jacket. Air was supplied through a diffuser to attain a dissolved oxygen concentration (DO) above 2.5 mg L−1 during the reaction stage of each SBR cycle. Sodium hydroxide was added to the feed wastewater up to pH 7 to avoid cell death because of the acidity of raw wastewater. Filling (i.e., feeding), settling, and sludge wasting were automated by a biocontroller ADI1030. Also, pH, DO, and agitation were controlled. Samples of the reactor were withdrawn periodically during the reaction period, centrifuged, filtered, and analyzed.

The COD, total Kjeldahl nitrogen (TKN), ammonia (N-NH3), volatile suspended solids (VSS), and sludge volume index (SVI) were measured according to Standard Methods (APHA, 1985). BOD was analyzed by the respirometric method ([19]). A Dohrmann DC-190 analyzer was used to determine total and inorganic carbon (TC and IC, respectively). Total organic carbon (TOC) was calculated as the difference between TC and IC values. UV absorbance was measured at 254 nm with a Hitachi 2000 spectrophotometer. Total phenols were analyzed by the Folin Ciocalteau method using gallic acid as the standard as described elsewhere (García-García et al., 1997). For wastewater characterization, the concentration of metals was carried out by atomic emission following the induced coupled plasma (ICP) technique. Ozone was analyzed in water using the method proposed by [4], while in the gas phase it was measured photometrically with an Anseros GM19 analyzer.

Preliminary experiments of aerobic biodegradation were first carried out in the sequencing batch reactor. The aim of this series was to ascertain whether bacterial growth inhibition took place or not because of high concentration of biorefractory or toxics in the feed. Aerobic biodegradation was effected in wastewater with different initial COD in the 1–5 g L−1 range but with the same initial VSS concentration. Raw wastewater dilution was always performed with tap water. Figures 1 and 2 show the effect of the substrate concentration on the evolutions of the VSS concentration and residual normalized COD, respectively.

Graph: FIGURE 1 VSS profiles during the reaction stage of a cycle of the SBR. Reaction conditions: T =20°C; pH0 =7; COD0 symbols: ▪, 1.08 g L−1; ▴, 1.58 g L−1; •, 2.14 g L−1; ▾, 4.40 gL−1.

Graph: FIGURE 2COD/COD0 profiles during the reaction stage of a cycle of the SBR. Reaction conditions: T =20°C; pH0 =7; VSS0 =1.6 gL−1; COD0 Symbols: ▪, 1.08 gL−1; ▴, 1.58 gL−1; •, 2.14 gL−1; ▾, 4.40 gL−1.

Figure 1 shows the typical VSS profile for a substrate limiting growth system: lag stage, exponential growth, stationary, and decay phases. For experiments performed with low organic load (initial COD of about 1 g L−1), lag phase was not observed at the beginning of the experiment likely due to the use of previously acclimated microorganisms. However, for experiments carried out with more concentrated wastewater (initial COD>1.5 g L−1) despite using an acclimated culture, the lag phase was pronounced, perhaps because of a combination of factors (i.e., a decrease in the rate of oxygenation and the presence of higher concentration of inhibitors) which lowered the respiration rate. For wastewater with more than 1.5 g L−1 initial COD, the exponential growth of microorganism finished at about 15 h of reaction; from then the endogenous metabolism would became a major factor, thus leading to microorganism death.

Profiles of Figure 2 corroborate the inhibition effects when treating wastewater with COD higher than 1 g L−1. As a consequence, COD depletion rate was decreased when increasing the initial COD from the critical value. Figure 2 leads to the conclusion that, regardless of the substrate utilization rate, activated sludge process was able to remove most of the organic matter of diluted cherry stillage, the overall decrease of COD being about 75% with 24 h reaction. Moreover, BOD and TOC reductions reached 95% and 75%, respectively. However, an overgrowth of filamentous microorganisms was observed for such a long reaction time. Thereby, activated sludge settling problems were experienced within the system. Reasons for the appearance of bulking sludge could be the low BOD after the removal of most of the biodegradable organic substances and, more likely, a shortage in available nutrients. In this sense, it is important to note that the COD/TKN ratio in raw wastewater, over 80/1 (see Table I) is very much higher than the ratio commonly accepted as necessary to sustain healthy biological growth.

To quantify sludge settling capacity SBR was operated at different reaction time, and oxidation efficiency and SVI were measured thereafter. Figure 3 presents the results obtained in this series of experiments. Figure 3 shows that activated sludge could not achieve complete purification of wastewater since polyphenols were not properly removed. Also at optimum retention time for good sludge settling properties (i.e., 8–10 h), removals of COD and TOC were lower than 60%.

Graph: FIGURE 3 Oxidation efficiency and sludge settling ability of activated sludge treating diluted cherry stillage. Reaction conditions: T=20°C; pH0 =7; VSS0 =1.6gL−1; COD0 =2.4gL−1; symbols: •, COD; ▴, TOC; □, polyphenols; ×, SVI.

Semibatch ozonations of about 2.5 g L−1 initial COD wastewater (see Table I for other wastewater characteristics) were carried out under various ozone dose in the 50–600 mg O3 per L wastewater range to examine the oxidation and mass transfer efficiencies. Gas flow rate was set to a maximum of 20 L h−1 because higher flow rates produced foaming problems. Oxidation efficiency was expressed as the percentages of COD, TOC, polyphenol, and UV254 absorbance removal. On the other hand, since ozone was not found in solution, ozone mass transfer efficiency (OMT) was calculated from the ratio between utilized and applied ozone, according to Equation [1]. In this equation, and stand for the ozone concentration in the inlet gas and in the off-gas, respectively. Results obtained are summarized in Table II.

TABLE II Oxidation and Ozone Mass Transfer Efficiencies

Graph

What is first observed in Table II is the small significance of specific pollution removal by means of ozonation in terms of COD and TOC, even at high ozone doses. Ratios of COD and TOC removal to utilized ozone were determined taking into account OMT. Thus, an average value of 1.07 g COD/g O3 was calculated, which is close to that reported for the ozonation of wine distillery wastewater ([7]). Accordingly, excessive ozone demand would be required to obtain significant organic matter removal which would incur such an expensive wastewater treatment process. Therefore, the goal of ozonation in an integrated ozone–activated sludge system treating cherry stillage should be the partial degradation of given compounds resistant to biological oxidation rather than the complete cleaning up of the wastewater. In this sense polyphenols, recognized as major inhibiting compounds in conventional activated sludge processing, can be reduced to acceptable concentration levels without excessive ozone consumption. As an average figure, from data of Table II and features of raw wastewater (see Table I), 0.15 g of polyphenol (measured as gallic acid) was calculated to be removed with 1 g of utilized ozone. UV254 absorbance was also reduced to a large extent. In fact, as shown in Figure 4, a linear relationship between polyphenol concentration and UV254 absorbance was observed on different ozonated wastewater samples. This provided an acceptable method for the quick estimation of polyphenol concentration in ozonated cherry stillage by a rapid and low sample-consuming measure of UV254 absorbance. Here, it is important to stress that such a procedure must not be applied without worry to other wastewaters, since a linear relationship between polyphenol concentration and UV254 absorbance is hardly found in other wastewaters.

Graph: FIGURE 4 Relationship between polyphenol concentration and UV254 absorbance during ozonation of diluted cherry stillage. Ozonation conditions: T =20°C; pH0 =4.1; COD0 =2.4 g L−1; ozone dose =50–600 mg L−1.

To model ozonation kinetics of complex wastewater, the COD usually has been taken as an indicator of organic matter concentration of the solution side. Then, the concentration of all the reactants with ozone can be lumped together under this variable ([11]; [20]). However, when the goal of ozonation is only the removal of specific compounds as polyphenols to assist subsequent biological oxidation, the objective of the kinetic study should be to establish the competitive reaction rates between these substances and other ozone-reacting compounds. The most simple attempt is to consider two parallel chemical reactions as follows:

Graph

Graph

where P refers to polyphenols and B to other ozone-reacting compounds.

As in any gas-liquid reaction system, the ozone absorption rate is known to depend on a mass transfer coefficient (kLa) and the enhancement factor (E) ([12]):

Graph

is the concentration of ozone at the gas-liquid interface that can be related to the ozone concentration in the gas by the Henry's law:

Graph

where H is the dimensionless Henry constant. For the case study, this was taken as 3.12 mg of gaseous O3/mg of dissolved O3 according to previous calculated values for the ozone–distillery wastewater system ([5]).

Fluid dynamics of the gas and liquid phases flowing through the reactor must be stated before going further with kinetic determinations. Thus, step-response experiments were carried out in order to establish the residence time distribution (RTD) and assess the flow behavior through the actual ozonation reactor. Ozone and potassium hydrogen phthalate were used as tracers in the carrier gas and liquid, respectively. Procedure for determining RTD is described elsewhere (Beltrán et al., 2000a; [13]). Perfect mixing in the liquid phase was obtained, as expected from the agitation induced by the gas flow and the liquid recycling ratio. The gas phase flow behavior was found to be very close to that in a plug flow reactor. Under these circumstances, such a procedure based on the mass balance of gaseous ozone has been described to calculate E·kLa (Beltrán et al., 1995; [27]):

Graph

where Q g is the gas flow rate and V the reaction volume.

Before going further with kinetic modeling, it is useful to establish the kinetic regime of ozone absorption. As can be observed in Table II OMT efficiency became lower as far as the ozone dose increased. In addition, the concentration of dissolved ozone was found to be negligible throughout the course of experiments referred to in Table II. From these experimental facts, conclusions about dynamics of ozone absorption could be drawn. Thus, the kinetic regime of ozone absorption was fast, the reaction between ozone and dissolved compounds taking place at the gas-liquid interface or the liquid side film. For ozone doses higher than those used in this work, wastewater likely would became exhausted in most reactive compounds toward ozone, thereby decreasing ozone efficiency and the absorption kinetic regime could change from fast to moderate or even slow (Beltrán et al., 1995).

If a fast pseudo-first-order kinetic regime is assumed for both reactions [R1] and [R2], and the film model is used, the enhancement factor for the reaction system can be related to the dimensionless Hatta numbers (HaP and HaB, for polyphenols and other ozone-reacting substances, respectively) according to Equation [5] ([23]):

Graph

Besides, Condition [6] is to be fulfilled for fast pseudo-first-order reactions ([12]):

Graph

In Equation [6], Ei stands for the instantaneous enhancement factor. Definition of Ha and Ei must be considered as follows:

Graph

Graph

The subscript j in Equations [7] and [8] may be replaced by either P or B for polyphenols or other reactants, respectively. Concentration of polyphenol in water, CP, could be estimated after applying equation of the straight line of Figure 4 to UV254 measures. CB was taken as the COD of dissolved organic matter other than polyphenols. Therefore, to calculate CB, the oxygen demand of polyphenols should be subtracted from the actual COD. In order to know the contribution of polyphenols to overall COD, the COD of an equimolar mixture of three phenolic acids (i.e., gallic acid, p-hydroxybenzoic acid, and syringic acid) was determined to be 0.73 mg O2/mg polyphenol. Then, CB could be estimated from Equation [9]:

Graph

To apply the kinetic model, values of diffusivities (DO3, DP, and DB), mass transfer coefficients (kL and kLa), and stoichiometric coefficients (zP and zB) must also be previously established. Ozone diffusivity was taken from literature as 1.76·10−9 m2·s−1 ([18]); polyphenol diffusivity was assumed to be that of gallic acid since this compound was chosen as a reference phenol-like compound. Gallic acid diffusivity was calculated using the equation of Wilke-Chang to be 7.1·10−10 m2 s−1 (Beltrán et al., 2000b). Diffusivity of other ozone reactants was chosen as 5·10−10 m2 s−1 since it is an average value of those corresponding to the diffusion of a number of organics in water ([24]). The individual liquid phase mass transfer coefficient kL was estimated from the equation of Calderbank valid for gas-liquid contactors similar to the column used in this work (Froment and Bischoff, 1979), while kLa was obtained from oxygen absorption experiments in the specific wastewater ([17]) followed by correction according to the film model. The specific interfacial area, a, was then obtained from the ratio kLa/kL as 30.7 m−1. Value of zP was also considered to be that of gallic acid, which has been determined previously to be 3.54 g gallic acid/g O3 (Beltrán et al., 2000b). Finally, zB was taken as 1.05 g O2/g O3 according to a value reported for a similar distillery wastewater (Beltrán et al., 1999b).

By combining Equations [4], [5], [7], and [9], the following equation can be obtained:

Graph

A series of semibatch wastewater ozonations was performed to provide enough data to enable one to correlate variables of Equation [10] and, therefore, to calculate rate constants. Figure 5 shows a three-dimensional plot where it can be seen that experimental data fit Equation [10] well. From the multivariable regression analysis based on the Levenberg-Marquardt algorithm, the unknowns kP and kB were determined to be 1.122 ± 0.057 L·mg−1·s−1 and 2.74 10−3 ± 0.68·10−3 L (mg O2)−1 s−1, respectively.

Graph: FIGURE 5 Relationship between ozone uptake and COD and polyphenol concentration depletions during wastewater ozonation according to Equation [10].

Finally, once the rate constants were known, Condition [6] of the fast pseudo-first-order kinetic regime was satisfactorily checked for data used to model the system of Reactions [R1] and [R2].

Integrated treatment systems were performed on diluted cherry stillage in two ways: ozonation followed by activated sludge on initial COD wastewater of 7.2 g L−1 and the sequence of activated sludge–ozonation–activated sludge treatments on about 2.2 g L−1 initial COD.

The objective of the first procedure was to study the ability of ozonation to partially degrade compounds in wastewater responsible for the inhibition of aerobic microorganisms. Ozone dose of 0.1 g L−1 was applied achieving a polyphenols removal rate of about 63%. Activated sludge treatment with initial biomass of 1.75 g L−1 then followed with a residence time of 45 h. Figure 6 shows the evolution of VSS and COD during this reaction period. A lag period could not be observed from the VSS profile. Also, the exponential decay of COD concentration suggests no inhibition of substrate degradation. SVI of the population developed after an operational time of 45 h was higher than 200 mL g−1, which is exceeds the acceptable value for proper activated sludge operation as settling would become a limiting factor.

Graph: FIGURE 6 COD and VSS profiles during the reaction stage of a cycle of the SBR operating on preozonated wastewater. Reaction conditions: T =20 °C; pH0 =7; VSS0 =1.75 g L−1; COD0 =7.2 g L−1.

The activated sludge–ozonation–activated sludge integrated system was applied to wastewater with relatively low initial COD (i.e., 2.2 g L−1); that is, for which biological inhibition occurred to a low extension. The aim of this method was to stop the first biological oxidation with hydraulic retention time to achieve good sludge settleability in addition to significant organic matter removal (about 12 h as deduced from Figure 4). The effluent thus obtained was then subjected to ozonation (ozone dose applied 65 mg L−1) to remove most of the polyphenols and induce biodegradability. Finally, a second activated sludge treatment of 8 h was used to complete purification of wastewater. Table III shows the results obtained with this procedure and they are compared to those from one-step activated sludge completed for 24 h of reaction time.

TABLE III Comparative Performance of Single Activated Sludge and Integrated Activated Sludge–Ozone–Activated Sludge Treating Diluted Cherry Stillage

As observed from Table III, a combined treatment system enables one to obtain better overall process performance as a result of higher COD, TOC, and polyphenol removals. In addition, the sludge from any of the activated sludge steps of the integrated system showed better settling characteristics than those of single activated sludge. This is an important finding because of the relevant cost of settling devices in activated sludge MWTPs ([1]).

One of the choices for the treatment of cherry stillage is dilution and release to an MWTP where an activated sludge process is applied. However, this practice has been shown to present some constraints if high organic load in the wastewater concentration of inhibitors, such polyphenols, is high enough to substantially decrease the specific biodegradation rate. On the other hand if low organic load is present, inhibition effects are not noticed, but even so long reaction time is required to achieve proper wastewater purification, thereby leading to sludge settling problems. From this investigation, we conclude that ozonation is a suitable technology to be combined with activated sludge for the treatment of cherry stillage. Since the ozonation effectiveness to oxidize polyphenols depends on the amount of ozone uptake, ozone absorption rate should play an important role in the efficient use of ozone to assist biological oxidation. The goal of ozonation is then the selective oxidation of polyphenols which develops in a fast pseudo-first-order kinetic regime. Therefore, the kinetic study presented here is relevant to design and scale-up of ozonation systems. Beneficial effects of ozonation when combining with activated sludge are the removal of inhibitory effects and the improvement of sludge settling at the biodegradation stage. In addition, purification of wastewater by two-stage activated sludge with ozonation in between is good enough to use effluent for dilution of the incoming wastewater.

The authors thank the CICYT of Spain and the European Commission for financial support of this investigation under the Grant 1FD97-0087.