Ultrasound-facilitated electro-oxidation for treating cyan ink effluent.

Chua, Chee-Yong ; Loh, Kai-Chee

INTRODUCTION

The majority of ink-jet printers utilize aqueous-based inks, which contain co-solvents, such as ethylene glycol-glycerol and a mixture of water-soluble dyes and/or pigments as colouring agents. The dyes and pigments are highly non-biodegradable and contribute significantly to the dissolved organic matter in ink waste waters. Much of the waste effluents from ink manufacturing facilities are therefore rich in non-biodegradable compounds although the volumes of such effluents are usually low. In a recent ink waste water characterization study, Chua and Loh (2004) revealed that the biodegradability index, represented by the ratio of Biochemical Oxygen Demand at 5 days ([BOD.sub.5]) to Chemical Oxygen Demand (COD) of untreated ink effluent ranged from only 0.2 to 0.3 in a scale where only ratios of [BOD.sub.5]:COD > 0.4 may be considered thoroughly biodegradable (Bekbolet et al., 1996).

For the ink industry, Fenton's reaction remains a widely used method. However, because ink-jet inks waste waters contain high strength soluble organic matters, this process requires relatively large quantities of [H.sub.2][O.sub.2] and acid (for subsequent neutralization) and consequently produces a large amount of sludge, leading to a secondary problem in the final sludge disposal. Thus, the development of an advanced oxidation process for ink effluent treatment, leading to smaller chemical usage and sludge production, as well as being a cleaner and more energy efficient technology, is a worthy approach.

Leshem et al. (2006) recently reported the colour and COD removal for a waste water effluent containing acid-reactive, natural dye and pigment using electrochemical oxidation. The result was a greater than 80% colour removal but less than 50% COD reduction. They also found that COD removal occurred at a slower overall rate than colour reduction. Gotsi et al. (2005) have also conducted a study on the electrochemical oxidation of olive oil mill waste waters over a titanium-tantalum-platinum-iridium anode. Olive oil mill waste waters with COD values ranging from 1475 to 6575 mg/L were used in their investigation. It was reported that nearly complete degradation of phenols and decolourization were achieved within short treatment times of up to 60 min. This was, however, accompanied by a relatively low COD removal that never exceeded 40% even after prolonged (up to 4 h) durations.

The effectiveness of electrochemically generated iron (II) catalyst in the presence of hydrogen peroxide for reducing the COD in aqueous solutions containing H-acid and Reactive Black 5 was recently reported (Rao et al., 2006). The higher efficiency of the method was attributed to the incremental addition of [Fe.sub.2+] and the accompanying higher [H.sub.2][O.sub.2]/[Fe.sub.2+] molar ratio. A more unusual technique is the use of [H.sub.2][O.sub.2] with ultrasonic power. This technique has the advantage of producing hydroxyl radicals, along with high local temperatures and pressures in the cavitation bubbles formed. The hydroxyl radicals are an extremely powerful oxidant, of which the rate coefficients with organic molecules are generally in the range of [10.sup.8] to [10.sup.10] [M.sup.-1] [s.sup.-1] (Glaze and Kang, 1989). Lorimer et al. (2000) conducted some studies on the decolourization of acidic dye effluent (Sandolan Yellow--an azo group in association with two aromatic systems and auxochromes) with applied ultrasound, electro-oxidation and the combined process. No decolourization was observed for using ultrasound alone (20 and 40 kHz). The selected dye (50 mg/L) was resistant to decolourization by using 0.5 mol/L of hydrogen peroxide. However, the addition of sodium hypochlorite solution (2.5 x [10.sup.-4] mol/L) was able to effectively decolourize a solution of Sandolan Yellow ([10.sup.-4] mol/L). This has resulted in electro-oxidation of dye effluent with the addition of aqueous sodium chloride. Lorimer et al. (2000) also found that when inert electrodes were used for the electro-oxidation of Sandolan Yellow, the discolourization effectiveness was enhanced by the applied ultrasound. However, due to the low allowable waste water discharge limit of chloride ion, it is unwise to add sodium chloride into the ink effluent.

An assessment of the untreated inks waste waters for values that exceeded or contributed significantly to those listed in the regulatory discharge limit was conducted by Chua and Loh (2004). The COD of all the untreated ink waste water tested exceeded the allowable discharge limit by 9-22 times. A baseline Fenton's reaction experiment performed on cyan ink waste water concluded that COD reduction was a more stringent criterion for Fenton's treatment in the ink waste water than [BOD.sub.5] or true colour value reduction. In addition, for the case of cyan ink waste water, the conventional Fenton's reaction could not mitigate copper for proper discharge, exposing a potential source of toxicity in the sewer and a sewage treatment plant's biological treatment units.

A series of combination treatment processes that involved sonolysis and Fenton's reaction was conducted for treating cyan ink waste water, which has been found to be the toughest ink to treat. It was found that the treatment processes based on sonolysis alone gave no reduction in COD value, true colour unit or copper content. This process was simply not effective for treating cyan ink waste water within the experimental duration of 3 h, the duration based on the industrially accepted Fenton's treatment. On the other hand, by combining sonolysis with Fenton's reaction, a COD reduction of 45 % more than that achievable by the Fenton's reaction alone was obtained. However, a comparable amount of the sludge generation was noted, and the copper discharge limit was still not met.

In this research, a modification to Lorimer et al.'s (2000) ultrasound electro-oxidation process was attempted for treating cyan ink effluent. Cyan ink effluent was investigated because magenta, yellow, and black inks did not contain copper higher than the detection limit of 0.1 mg/L. In this modification, the inert electrodes were replaced by iron electrodes for both the anode and the cathode to generate and regenerate the catalytic iron needed for Fenton's reaction. The electro-oxidation process was further facilitated by ultrasonication to generate hydrogen peroxide. With the resulting iron (II) and [H.sub.2][O.sub.2] generated, we anticipated that an enhanced Fenton's reaction with lower chemicals usage and lower sludge generation would emerge.

MATERIALS AND METHODS

Ink Waste Water

Synthetic cyan ink effluent was made up from inks obtained from Hewlett Packard HP51649A (colour). Cyan inks from the ink cartridges were diluted 50 times, indicative of the typical strength of inks effluent generated from cleaning production lines and equipment. The high purity water for the dilution was obtained from a Milli-Q system (Milli-Pore Corporation, New South Wales, Australia) with a resistivity of 18.2 M(cm and less than 50 [micro]g/L of organic carbon content.

Chemicals/Reagents

The chemicals/reagents used in this study included stabilized extra pure 35 wt% hydrogen peroxide (Riedel-de Haen, Seelze, Germany) and iron (II) sulphate FeS[O.sub.4]. 7[H.sub.2]O (Merck, Darmstadt, Germany). The quenching reagent used in the kinetics study was bovine liver catalase, EC1.11.1.6 C-40, at 24 640 units/mg solid; 1 unit decomposes 1.0 [micro]mol [H.sub.2][O.sub.2] per minute at pH 7.0 and 25[degrees]C (Sigma Aldrich, Darmstadt, Germany).

Water Analyses

Unless otherwise indicated, the various water quality parameters outlined in the trade effluent discharge limits of Singapore (ENV, 1997) were determined using standard methods (American Public Health Association, 1998). These water quality parameters included [BOD.sub.5], COD, total suspended solids (TSS), total dissolved solids (TDS), conductivity, pH, UV-vis (ultraviolet-visible) spectra and the regulated chemical constituents. In particular, the COD measurement was conducted using the MN Filter Photometer PF-11 (Macherey-Nagel, Duren, Germany) and VELP ECO 16 Thermoreactor. The dichromate method was used in the colourimetric COD measurement via MN reagents. The detection limits were 10 mg/L for single samples and 6 mg/L for triplicate samples with a coefficient of variation of less than 6%. The turbidity reading in formazine attenuation units (FAU) was also measured by the PF-11. The Shimadzu UV-1601 with 10 mm path length quartz cuvette was used to obtain the UV-visible spectral scan and absorbance reading of the ink sample. An Orion 720A meter with a pH probe and Thermo Orion oxidation/reduction potential (ORP) 9179 probe was used for pH and ORP measurements. Another Orion 115 meter equipped with Microelectrodes M1-915 conductivity electrode (K = 1.0) was used to measure sample conductivity. A Testo Digital Thermometer (for in situ temperature measurement) was immersed in the synthetic waste water during the chemical oxidation process.

[FIGURE 1 OMITTED]

For determination of the trace metal concentrations in the ink waste water, a pre-treatment of acid digestion was conducted to dissolve the metal ions in complexes with pigment or dye components in the inks. The acid digestion was carried out with 65 concentrated nitric acid. The trace metal concentrations were then determined by inductively coupled plasma (ICP) atomic emission spectrometry (ICP, Perkin Elmer, Waltham, MA). The standard method protocol (APHA, 1998) was followed for this acid digestion and ICP measurement.

To assess the colour of the ink waste water, an ISO method by measuring light absorbance at the three visible wavelengths of 436, 525, and 620 nm was employed (International Organization for Standardization, 1994). The sum of the absorbances at these three wavelengths (STCV-3[lambda]) was then used as an indication of the strength of the true colour present in the ink waste water.

Sludge Quantitation

During the treatment process, samples of the reacted mixtures were taken to measure the amount of sludge generated. The sludge in the ink waste water refers to the suspended and non-filterable residue left in the treatment process and which requires subsequent disposal. A sludge quantification method defines residue, non-filterable as those solids retained by a glass fibre filter and dried to constant weight at 103-105[degrees]C. Samples of known volume were first filtered through a prepared glass fibres filter, and the residue retained on the filter was dried to constant weight in the oven (105[degrees]C) for at least 1 h. The mass of the residue was then determined.

Fenton's Reaction

The industrially accepted Fenton-based COD reduction method was used as the baseline technique for comparison. Five hundred millilitres of the ink waste water was mixed in a 3 L magnetic-bar stirred beaker, and the computed amount of iron sulphate solution (based on a molar ratio of 1:10 for FeS[O.sub.4]:[H O.sub.2][O.sub.2]) was added. A stoichiometric amount of 35% w/w [H O.sub.2][O.sub.2] reagent based on the initial COD was then introduced (Eckenfelder, 2000). The pH of the reaction mixture was recorded, and the reacted solution was neutralized with 0.1 M NaOH when gas evolution had stopped. The final treated waste water was then filtered for the corresponding water parameter analyses.

During treatment, the temperature and ORP were monitored after hydrogen peroxide was introduced. Conductivity of the mixture was also monitored to ensure uniform mixing through an adjustment of the magnetic stirrer speed. Ten millilitres of the reaction mixture was sampled at 10, 20, 30, 45, and 60 min intervals. Unreacted hydrogen peroxide was destroyed by the catalase enzyme to quench Fenton's reaction and to prevent its interference with the analytical measurements.

Proposed Two-Step Method

The proposed cyan ink effluent treatment scheme featured two steps as shown in Figure 1. The experimental set-up for step 1, which is an ultrasound-assisted electro-oxidation process, is shown in Figure 2. Five hundred millilitres of ink effluent in a 1 L Hysil beaker was placed in an ultrasound bath (TRU-SWEEP model 575STAG, Crest Ultrasonic, Trenton, NJ) with tank size of 29.5 cm x 15 cm x 15 cm. The ultrasound bath had been filled with water and the ultrasound switched on for about 10 min to allow the system to stabilize before suspending the beaker in it, 2 cm above the base of the bath. Mild steel (Grade SS41) 15 cm length and 3 mm thickness was used as electrodes. The width of the anode measured 24 mm while that of the cathode was 37 mm. The electrodes were polished with sandpaper to ensure that their surfaces were clean prior to being connected to the DC power supply (Electro-Automatik, Viersen, Germany) rated DC 32 V and 2.5 A output with constant V and constant A mode. The electrodes were suspended 1.0-1.5 cm apart in the beaker. The various monitoring meters, probes or electrodes were then connected and suspended in the beaker. These included the temperature, conductivity, ORP, and pH probes.

The voltage and current across the electrodes were adjusted to 18 V and 1.00 A, respectively. Care was taken to maintain the current at about 1.00-1.02 A throughout the experiment. Every 10 min, 2 mL samples were withdrawn for sampling. Step 1 lasted 50 min after which the ultrasound and power supply were switched off. The electrodes were removed, and the beaker was transferred to a magnetic stirrer. Transfer between steps 1 and 2 took at most 10 min.

For step 2, 1.8 g of solid iron (II) sulphate was dissolved in 8 mL of ultra-pure water and the solution was added to the ink solution. Thirty-nine millilitres of 35 wt% hydrogen peroxide was then added. The stopwatch was started immediately, and a 2 mL sample was withdrawn for initial time sampling. Subsequently, 2 mL samples were taken every 10 min. All the samples from step 2 were quenched by pipetting them into glass vials filled with 2 mL of quenching reagent immediately after they were withdrawn. Step 2 was conducted for a total of 60 min.

[FIGURE 2 OMITTED]

The reaction mixture was then left for another 60 min to make up the total treatment time of 3 h, comparable to that commonly practised in the industrial process. It was then stirred well, after which two 50 mL samples were taken. One sample was filtered to analyze the sludge content, and the other sample was neutralized to pH7 using 0.1 M NaOH. The neutralized sample was then filtered for analysis of sludge content as well. The neutralized solution was also analyzed for COD.

All the quenched samples were left for 12 h to ensure that the hydrogen peroxide was removed by the quenching reagent and would not contribute to the COD value. They were then analyzed for COD. For COD analysis, 0.2 ml, of the sample was added to a tube of Hach COD reagent. The tubes were then heated for 2 h in the COD reactor-heating block. Upon cooling, they were analyzed in the HACH colourimeter. Selected samples from step 1 were diluted 100 times and analyzed in the UV-spectrophotometer.

Ultrasound Power Density and Cavitations Intensity in Reaction Vessel

The energy density (E) and intensity (I) are often used to characterize ultrasound devices. I = P/A and E = P/V are commonly used to describe the power input P from the sound source into the liquid. The intensity is normalized by the radiating surface A, whereas the energy density is normalized by the sonicated liquid volume V.

For power density, calorimetric power measurement methods were carried out using different sizes of ultrasonic baths. Both the beaker and bath were filled with distilled water. When the ultrasonic bath was switched on, the temperatures of water inside the beaker (if there were any) and that surrounding the beaker were monitored at regular time intervals using the thermometer model 1303 (TES Electronic, Hanover, Germany) with type K thermocouple SS type which offered T1 and T2 differential measurement. The total ultrasonic power received by the reaction volume in the beaker was calculated based on:

Power (W) = dT/dt x [C.sub.p] x m

where [C.sub.p] is the specific heat capacity of water (4.184 J/g [degrees]C), m the mass of reaction volume (g), dT/dt the steady-state temperature gradient ([degrees]C/s).

Cavitations occurred when high intensity ultrasonic waves were directed into the liquid. In order to establish the evidence of cavitations present within the reaction volume, a cavitations intensity meter was used. The cavitations intensity meter model CM-3-100 (Alexy, Bethel, NY) with the standard 45 cm long probe was placed at the location within the ultrasonic field in the liquid to measure the cavitations intensity in "CAVIN." The meter was calibrated to read from 0 to 1000 CAVIN with one CAVIN representing 1/1000 of the peak cavitations observed in the universal peak value established by the manufacturer.

RESULTS AND DISCUSSION

Proposed Two-Step Integrated Treatment Scheme

A preliminary treatment was conducted using ultrasound-facilitated electro-oxidation alone (step 1 only). It was found that although decolourization of the cyan ink effluent was possible, the whole treatment demanded a reaction duration of more than 3 h. An evaluation of the treated ink effluent at the end of 3 h revealed only a 65% reduction of copper content (140-49 mg/L), a 20% COD reduction (13 247-10 600 mg/L) and a little sludge generated. A two-step treatment scheme consisting of a first step which involved the release of [Fe.sup.2+] using ultrasound-assisted electro-oxidation and the second step which involved the addition of FeS[O.sub.4] and [H.sub.2][O.sub.2] (modified industrial Fenton's) was attempted.

This treatment scheme required much less FeS[O.sub.4] demanded by Fenton's reaction resulting in a treated cyan ink effluent with little to no sludge generated. These met our expected criteria of needing less chemicals and generating little sludge. The possibility of executing the two proposed steps in reverse order was also explored. It was found that in this case, the same amount of sludge as Fenton's reaction alone was generated. Therefore, it was concluded that the sequence of the treatment steps was important if the objective of having less sludge was of paramount importance, in addition to the reduction in COD and true colour. With regards to sludge formation, the two-step integrated treatment scheme resulted in little to no sludge produced as compared to about 6.7 g/COD of sludge generated from the Fenton's reaction. This is because the amount of FeS[O.sub.4] used was less than the conventional Fenton's reaction, and the final pH of the treated cyan ink effluent at the end of the two-step scheme was as low as 1.8, resulting in a higher solubility product for the iron sludge.

The developed two-step treatment was optimized in subsequent experiments based on the amount of chemicals needed, the sonication current supplied, and the distribution of the total reaction time between the two steps. The result of these was an optimized treatment requiring, for treating 500 mL of cyan ink effluent, 1.8 g of FeS[O.sub.4] and 39 mL of [H.sub.2][O.sub.2] (50% less than Fenton's reaction) and a maximum sonication current supply, which was restricted by the positioning of the electrodes, of 1.02 A.

The operating conditions for step 1 were also quantified. The measured ultrasound power density and cavitations intensity for the submerged reaction volume were found to be 11.64 W/L and 200 CAVIN, respectively. The true input power for the ultrasonic bath with rated input electric power of 240 W was measured as 69 W The bath ultrasound power density and cavitations intensity were also determined as 20.43 W/L and 250 CAVIN, respectively. The frequency of ultrasound irradiated in this bath was 35 kHz. The electro-oxidation as the cooperating part of the step 1 process was characterized by using a pair of iron electrodes--Fe(II)/Fe(III) reversible redox reagents system operating at 18 V, 1 A and anodic current density of 28.2 mA/[m.sup.2].

During the experiment, several process monitoring parameters, namely temperature, conductivity, pH, and ORP were monitored. These parameters were useful in providing the progress indication of the reaction mixture during the course of treatment. The temperature rise was linear for step 1 from 26 to 43[degrees]C which was from the heat generated during the ultrasonic operation. This was followed by a 10[degrees]C rise in step 2 mainly contributed by the chemical oxidation reaction. The conductivity of the mixture generally increased from 1.07 to 1.54 mS in step 1 resulting from the iron dissolution from the anode. In step 2, the conductivity of the solution further rose to 3.91 mS due to the addition of the Fe(II) from the iron salts. This also indicated an increased in the concentration of charged particles in the reaction mixtures. The production of organic acids in the ink degradation process by Fenton's reagent was another source for such a rise in the conductivity in step 2.

The pH of the solution dropped in the first 5 min of step 1 (from 6.9 to 3.4) and then rose to a steady value of 4.3 at the end of step 1. In step 2, this dropped sharply in the first 5 min from 4.3 to 2.5 and decreased steadily after that. The initial drop in step 1 could be due to the introduction of additional iron (II) ions from the anode. In step 2, the initial drop was the result of rapid oxidation and degradation of the complex organic molecules in the ink into organic acids. This continued throughout the rest of the reaction, causing the pH to drop steadily to 1.8 at the end of step 2. Lastly, the ORP also dropped in the first 5 min of step 1 which then increased at a decreasing rate for the rest of the step before reaching 430 mV In step 2, the ORP rose rapidly in the first 10 min, which then remained at around 590 mV The initial drop in step 1 could be due to the introduction of Fe(II) ions which had acted as a reducing agent. The subsequent rise showed that electro-oxidation was indeed taking place. In step 2, the initial rise was due to the introduction of hydrogen peroxide. The maintenance of the ORP at 590 mV also meant that overall the solution was in a more oxidizing state than its original state of 350 mV.

Outcome of Two-Step Treatment

The earlier ink waste water characterization study (Chug and Loh, 2004) had identified COD, copper, iron, and sulphate contents as the high-risk non-compliance parameters. Table 1 indicates the changes of these water quality parameters or component concentrations at the various stages of the treatment process (refer to Figure 1) evaluated against the regulatory discharge limits (ENV, 1997). As indicated in Table 1, the treated cyan ink effluent from the two-step treatment method met entirely with the listed discharge limits.

In the two-step treatment process, the H202 usage was found to be 50% less than Fenton's reaction. Hence, the corresponding FeS[O.sub.4] usage in step 2 (modified Fenton's reaction) was much reduced to 16-30% of the conventional Fenton's usage. Furthermore, the total treatment scheme could be affected in less than 2 h with 50 min needed in step 1. In fact, it was 45% faster than Fenton's reaction (1 h 40 min vs. 3 h). Although both the Fenton's reaction and the two-step integrated treatment process were able to achieve more than 80% reduction in the COD, our proposed treatment scheme was superior because in addition to the shorter treatment time, there was also reduced chemicals usage and little sludge production. More importantly, the copper content was reduced significantly to meet the discharge limit, a clear advantage over the industrial Fenton's reaction.

Sludge Quantitation

The use of the two-step treatment scheme resulted in a 98 % reduction in the amount of sludge generated, which was quantified as 1.4 g/L in an 18 h settling period as compared to 71 g/L of sludge generated from the conventional Fenton's reaction during the same period. The sludge that was commonly found in Fenton's type of treatment schemes consisted mostly of Fe[(OH).sub.3], In order for precipitation of Fe[(OH).sub.3], its solubility product must be greater than 4 x [10.sup.-38] M (at 298 K). The final pH of the treated cyan ink effluent at the end of the two-step scheme, as well as the Fenton's reaction, both dropped to a low of about pH 1.8. Given the acidic condition of the treated ink effluent, [Fe.sup.2+] would be more stable than [Fe.sup.3+]. Also, with an overall reduced addition of [Fe.sup.2+] in the two-step scheme, lesser [Fe.sup.3+] could be formed.

[FIGURE 3 OMITTED]

Sludge formed in both processes consisted mainly of iron (III) hydroxide, formed during the terminating step of the Fenton's reaction. The decrease in sludge production was attributed to the reduction in iron added during step 2 since the contribution of [Fe.sup.2+] to the reaction mixture in step 1 was minimal. That this was the case was confirmed by varying the amount of iron (II) sulphate added in step 2. The amount of iron (II) sulphate added was varied from 8% to 32%, with 16% (0.016 mol FeS[O.sub.4]/mol of [H.sub.2][O.sub.2]) being the standard concentration used in the two-step treatment scheme. The amount of sludge produced ranged from 1.38 to 1.73 g/L treated ink effluent.

Although sludge quantitation by the weighing method provided a good primary reference, the analysis was time consuming. It was worthy to develop a field monitoring parameter that would allow the amount of sludge generated to be estimated rapidly. We anticipated that a rapid estimation of the sludge concentration could be determined using refractive index (RI) measurements. The principle behind this is based on the material balance for the amount of Fe(II) added during treatment. The RI provided an indication of the amount of Fe(II) remaining in solution that had not precipitated as sludge. The estimated sludge generation could then be read from an experimentally obtained calibration curve. The RI of the supernatant in the treated water samples was determined using the AR200 refractometer (Reichert, Seefeld, Germany). The AR200 refractometer used has a measuring range of 1.33-1.56 nD, where nD refers to the measurement when the light emerges from the sample detected at the wavelength of the sodium D line.

After the cyan ink effluent had been treated, the sludge was allowed to settle over a 1 h period. The RI of the supernatant was measured and correlated against the amount of sludge produced. Figure 3 shows the calibration curve for RI versus the sludge dry weight per treated L of water. The data obtained were from experiments performed in triplicates. It can be seen that the correlation between RI and sludge produced was linear and excellent ([R.sup.2] = 0.99). Instead of determining the amount of sludge through drying and weighing, which is time consuming and tedious, the RI of the supernatant of the settled sample can therefore be used as an indication of the amount of sludge generated.

Physical Observations of Electrode Surfaces During Step 1

During the step 1 operation, dye deposition was observed on the cathode. In addition, this deposition process was found to be selective. Most deposited dye was found on the side of the cathode not facing the anode. The closely spaced electrodes in the step 1 arrangement have created a strong anodic and cathodic potential field interaction between the electrodes. This overlap of the potential field led to a strong induction motion of the charged ions toward the cathode side not facing the anode. The cavitations promoted a coagulation of dyes enhancing their attachment to the cathode. This was confirmed through an experiment in the absence of ultrasound performed on the cyan ink effluent. It was noted that the amount deposited on the cathode was reduced to a thin and lightly adhered layer. It was thus concluded that cavitations played a significant role in aiding dye adsorption onto the cathode.

The effect of ultrasound on the ink solution at this stage might seem to have promoted the movement of the coagulated molecules toward the specific cathodic site. Ensminger (1998) found that ultrasound has no effect on the electrolysis cell voltage or the cathode current efficiency in a series of studies on the effect of ultrasound energy during electrolysis and electroplating operation. However, he found that ultrasound was able to reduce the cathodic and anodic polarization during eletrodeposition and permitted an increase in the rate of metal deposition. According to these studies, movement of gas bubbles and cavitations in an ultrasonic field could affect the acceleration of the process of electrocrystallization. He also concluded that the key factor in accelerating electrocrystallization was the steady-state micro-eddies that arose in the electrolyte at the surface of the cathode in the ultrasonic field. Our result on dye deposition is in accord with his conclusions. This phenomenon of dye deposition is thought to be a physical separation between the dye and the solvent. Since the colour of the cyan ink effluent was mainly contributed by the dye, the deposition process aided in the reduction in the true colour strength of the cyan ink effluent. A significant consequence of this was the reduction of the copper concentration in the cyan ink effluent during the step 1 process, as depicted in Table 1, further providing strong evidence of the synergistic effect between ultrasound and electrochemical energies during the step 1 operation.

Proposed Mechanism for Proposed Treatment Scheme

The two-step integrated treatment scheme is valued for its reduction in chemicals requirement. Step 2 resembles Fenton's reaction in its chemical reaction. However, the fact that it needed less FeS[O.sub.4] suggests that step 1 had altered the composition of the ink waste water such that the waste water did not demand as much chemicals for degradation in step 2. It was also noted that significant COD reduction was not achieved during step 1. Furthermore, the strength of the colour of the cyan ink effluent was not lowered greatly. It was therefore hypothesized that step 1 had merely simplified the complex structures of the organic compounds in the cyan ink effluent. A high COD value was maintained throughout step 1. It was thus suggested that the re-structuring of bonds could have occurred, resulting in the formation of intermediates with simpler bond structures. Since UV-vis spectrophotometry is able to detect chromophores (Dean, 1999), such intermediates might be detected in a UV spectrum in the 250-400 nm range. A scan of the samples collected during step 1 showed that the area under the absorption curve between 275 and 400 nm was significantly reduced over the 50 min reaction time, as illustrated in Figure 4.

[FIGURE 4 OMITTED]

Multiple bonds in some organic compounds absorb strongly between 250 and 400 nm, giving rise to the peak in the UV spectrum (Ewing and Wood, 1985; Harwood and Claridge, 1997). Although a minute amount of [H.sub.2][O.sub.2] was produced in step 1, its contribution to the absorbance value within the wavelength range of interest was negligible. This was ascertained by performing an analysis on pure [H.sub.2][O.sub.2] alone. This reduction in area, thus, could possibly correspond to a decrease in the concentration of multiple bonds of organic compounds in the partially treated waste water. A check was further conducted on the UV spectrum belonging to cyan ink effluent treated with the conventional Fenton's reaction. A similar but a greater amount of absorbance reduction was found in the 275-400 nm window (data not shown).

This could mean that the formation of such intermediates was a necessary step in both treatment methods. We therefore propose that the treatment of cyan ink effluent consisted of two reactions in series: the formation of intermediates followed by the degradation of these intermediates. It was thus deduced that in our two-step treatment scheme, step 1 enhanced the first reaction by forming the intermediates while step 2 dealt with the degradation of these intermediates, which were more amenable to peroxide oxidation with much less chemicals needed.

Reaction Kinetics for Two-Step Treatment Process

In our ink effluent treatment, since the reaction mixture comprised multiple reacting components and not all of the mixture components were known with regards to their structure and compositions, treatment efficiency was evaluated based on a lumped parameter such as COD. In defining the reaction kinetics, we have therefore used for step 1 the STUV-3[lambda] measurements and for step 2 the COD measurements, as representations of the reactant concentrations. The difference in treatment efficiency attributed to temperature change was investigated over the range of 20-50[degrees]C in the two-step treatment process. For step 1, it was further assumed that the rate of mass transfer of [Fe.sup.2+] into the liquid phase was not limiting due to the agitation effect supplied by the ultrasonic bath.

The data obtained for both steps fitted second-order kinetics best, as shown by the plots of 1/[C.sub.i] versus t (Figures 5 and 6). The rate constants for the intermediates reduction reaction in step 1 was determined as [k.sub.1,Int] = 3.02 x [10.sup.-4] [AU.sup.-1] [min.sup.-1] (298 K) while that for the COD reduction reaction in step 2 was [k.sub.2,COD] = 4.17 x [10.sup.-6] L/[mg.sup.-1] [O.sub.2] [min.sup.-1] (298 K). Figures 5 and 6 also show the temperature effect on the kinetics of the reactions. Based on these, the activation energies and the pre-exponential Arrhenius constants were extracted. The pre-exponential Arrhenius constants were found to be 0.471 1/(AU min) and 0.0156 1/(COD min) for steps 1 and 2, respectively. The corresponding activation energies were 18.2 and 20.4 kJ/mol.

[FIGURE 5 OMITTED]

Due to the paucity of data available in the literature for activation energies involved in the decomposition of inks and that the inks are multi-component in dye and pigment composition, it may be instructive to note that the activation energies obtained are within the upper bounds of typical bond strengths in organic molecules, the highest being about 300 kJ/mol (Morrison and Boyd, 1983). Denisov et al. (2003) also reported that the activation energies for the thermal decomposition of azo compounds typically ranged from 85 to 223 kJ/mol. In comparison, the activation energies determined for both steps of the two-step treatment scheme were much smaller, indicating the strong catalytic effects of the ultrasound-facilitated electro-oxidation and Fenton's reactions.

Cost Analysis

The cost effectiveness of the proposed two-step treatment system vis-a-vis the conventional Fenton's treatment system can be assessed based solely on operating cost. The major operating cost includes chemicals usage cost and final disposal cost. The latter can be further subdivided into the cost of sludge disposal and the cost of water needed for diluting the treated effluent to meet the discharge limit of copper in the case of cyan ink effluents. In all these cost aspects, the proposed two-step treatment system is advantageous. For example, in terms of chemicals usage, the current system uses 50% less [H.sub.2][O.sub.2] and 16-30% less [Fe.sub.2]S[O.sub.4] usage. Sludge generation from the proposed treatment system showed a 98% reduction as compared to the conventional method. This translates into significant savings in the final sludge disposal cost, particularly in countries where sustainable land development is vital. On site solid waste disposal is usually not allowed for in such countries, and additional transportation and handling cost would be incurred as a major portion of disposal cost. Hence, an effective on-site waste water treatment process is highly desirable. Finally, since copper discharge limit can be met with the present approach, there is no additional dilution requirement.

[FIGURE 6 OMITTED]

CONCLUSIONS

COD reduction and copper removal have been found to be the stringent treatment criteria for cyan ink effluent. A study involving sonolysis, electro-oxidation and the use of hydrogen peroxide in various arrangements has been performed that successfully led to the development of the two-step treatment scheme. For this proposed treatment methodology, a list of beneficial factors has been observed in the treatment of cyan ink effluent. In step 1, an ultrasound-assisted electro-oxidation process, significant copper removal, formation of reaction intermediates and iron generation through electrodeposition were noted. Both COD and copper removals and faster reaction as well as significant sludge reduction were observed in the second step, the modified Fenton's reaction. In addition, the kinetics study for this proposed scheme concluded second order reaction kinetics for both steps, with much lower activation energies.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the financial support provided by the National University of Singapore under the research grant R-279-000-082-112. We also would like to acknowledge Ms Yap Mei Xia, Ms Hannah Lee Chang En, and Ms Lee Lingzhi for their technical assistance and contribution in the course of this research.

Manuscript received July 18, 2006; revised manuscript received April 2, 2007; accepted for publication June 8, 2007.

REFERENCES

American Public Health Association (APHA), American Water Works Association (AWWA), and Water Pollution Control Federation (WPCF), "Standard Methods for the Examination of Water and Wastewater," 20th ed., American Public Health Association, Washington, D.C (1998).

Bekbolet, M., F. Cecen and G. Ozkosemen, "Photocatalytic Oxidation and Subsequent Adsorption Characteristics of Humic Acids," Water Res. 34, 65-72 (1996).

Chua, C. Y. and K. C. Loh, "Systematic Approach for Quick Quality Assessment of Ink Effluents for Treatment and Discharge," J. Environ. Eng. 130, 417-424 (2004).

Dean, J. A. "Langer's Handbook of Chemistry," 15th ed., McGraw Hill (1999).

Denisov, E. T., T. G. Denisova and T. S. Pokidova, "Handbook of Free Radical Initiators," John Wiley & Sons, New Jersey, U.S.A. (2003), pp. 308-315.

Eckenfelder, W. W., "Industrial Water Pollution," McGraw-Hill, Singapore (2000), pp. 462-477.

Ensminger, D., "Ultrasonic: Fundamentals, Technology, Applications," Marcel Dekker, New York, U.S.A. (1998), pp. 505-506.

ENV, Ministry of the Environment, Republic of Singapore, "Code of Practice on Pollution Control," Appendix 9, Pollution Control Report, ENV, Singapore (1997).

Ewing, G. W. and G. Wood, "Instrumental Methods of Chemical Analysis," McGraw-Hill, Singapore (1985), pp. 63-64.

Glaze, W. H. and J. W. Kang, "Advanced Oxidation Process. Description of a Kinetic Model for the Oxidation of Hazardous Materials in Aqueous Media with Ozone and Hydrogen Peroxide," J. Ind. Eng. Chem. Res. 28, 1537-1580 (1989).

Gotsi, M., N. Dalogerakis, E. Psillakis, P. Samaras and D. Mantzavinos, "Electrochemical Oxidation of Olive Oil Mill Wastewaters," Water Res. 39, 4177-4187 (2005).

Harwood, L. M. and D. W. Claridge, "Introduction to Organic Spectroscopy," Oxford University Press, UK (1997), pp. 13-16.

International Organization for Standardization, The (ISO), "Water Quality--Examination of Color," ISO 7887:1994, 2nd ed. ISO, Switzerland (1994).

Leshem, E. N., D. S. Pines, S. J. Ergas and D. A. Reckhow, "Electrochemical Oxidation and Ozonation for Textile Wastewater Reuse," J. Environ. Eng. 132, 324-330 (2006).

Lorimer, J. P., T. J. Mason, M. Plattes and S. S. Phull, "Dye Effluent Decolourisation Using Ultrasonically Assisted Electro-Oxidation," Ultrason. Sonochem. 7, 237-242 (2000).

Morrison, R. T. and R. N. Boyd, "Organic Chemistry," 4th ed., Allyn and Bacon, U.S.A. (1983).

Rao, N. N., B. Gaurav, K. Pradnya and S. N. Kaul, "Fenton and Electro-Fenton Methods for Oxidation of H-Acid and Reactive Black 5," J. Environ. Eng. 132, 367-376 (2006).

Chee-Yong Chua (1) and Kai-Chee Loh (2) *

(1.) Eliddell Engineering Technology, PSA Building, P.O. Box 637, Singapore, Singapore

(2.) Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, S117576 Singapore, Singapore

* Author to whom correspondence may be addressed. E-mail address: [email protected]

Table 1. Values of high risk non-compliance parameters at various
points in the two-step treatment (refer to Figure 1 for sampling
points)

Parameter A B C
 (mg/L) (mg/L) (mg/L)

COD 10500 10 360 1740
Iron 34 40 660
Sulphate <1 <1 1140
Copper 140 50 17

Parameter D Regulatory limits for
 (mg/L) discharge (mg/L)

COD 499 600
Iron 0.08 50
Sulphate 181 1000
Copper 1.92 5