Possibility of treating rubber factory wastewater in biological reactors using media with low specific surface area

Four Covered Activated Ditch (CAD) type test reactors set with 25, 50, 75 and 100 m'/m specific surface areas (SSA) of Bio-brush were tested under 0.5, 1.0, 2.5 and 3.5 CODkg/rn/d organic loading rates (OLR). Chemical Oxygen Demand (COD) of the treated effluent of each reactor were monitored to see the efficiency of treatment. In order to minimize the cost of a treatment system the main aim of this study is to investigate the possibility of treating rubber factory waste in CAD system with low SSA (<100 m/m) Bio-brush media. Although the cost of Bio-brush media with low SSA is less, the possibility of treating effluent with media below 100 m/m SSA in completely anaerobic treatment systems is not economical in terms of treatment efficiency. However, a CAD system with low SSA of Bio-brush media may be effectively used to remove COD (about 60-75%) with a final aerobic treatment part for removing remaining pollutants.


Introduction
In Sri Lanka natural rubber industry is one of the most water polluting industries in the island. From an economical point of view biological treatment is considered the most suitable for rubber waste treatment when compared to extensive physical and chemical treatment methods.
In addition, the treatment facility must be cost effective as raw rubber processing industries, are highly vulnerable to price fluctuations. Anaerobic treatment of waste is a simple and cost effective method of treatment, which can contribute significantly to the improvement of environmental standards in developing countries.

K V V S Kudaligama et al.
As the loading capacity of an anaerobic reactor is not limited by the supply of any required reagent, such as oxygen in aerobic systems, in principle the loading potentials of anaerobic treatment systems are dictated by,

Media design
The bristle fibre coir used in preparation of Bio-brushes was selected from a same batch in order to minimize variations.
Mean surface area/unit weight of coir fibre used was 0.015 m 2 /g coir. Coir fibre was cut in to 15 cm lengths for preparation of Bio-brushes with 15 cm diameters. Different amounts of coir fibre (weighed to get the correct SSA), was spread evenly in between two HDPE ropes (diameter 1.5mm) and using a special hand operated machine, twisted to get a bottle brush like flexible brush as explained by the Sri Lankan patent, No. 10951 (Warnakula, 1993). Different SSA tested with Bio-brush media were 25, 50,75,100m 2 /m 3 .

Reactor design
Four identical ditch type test reactors ( Fig. 1) were constructed with cement blocks, supported with concrete reinforcement. Length, width and depth of each reactor were 9.09 m, 0.3 m and 0.69 m respectively. Ditches were lined with a UV stabilized polyethylene sheet for waterproofing and were set with different SSAs of media. Bio-brushes were fixed in a vertical position and anchored to the bottom to prevent uplifting due to gas formation. The final effective volume of each ditch was 1.62 3 m .

Feed characteristics and organic loading
Fresh rubber serum was used as the feed. The reactors were tested under four different organic loads; 0.5, 1.0, 2.5 and 3.5 CODkg/m 3 /d. The retention time set was 2.66 days. Each organic load was introduced until the test reactors reached a steady state.

Inoculum
Anaerobic biomass collected from a local crepe rubber factory wastewater treatment system was used for preparation of anaerobic seed sludge for test reactors.

Start-up of reactors
Reactors were filled with water and inoculated with 50 1 of seed sludge mixture with 5000 mg/1 of suspended solids content.
Dilute rubber factory wastewater was introduced as the feed for a period of 2 weeks. COD of the feed increased gradually in 50 mg/1 up to reach 750 mg/1 weekly.

Analytical procedure
Samples were collected daily from each test reactor and tested for COD.
Closed reflux colourimetric method (APHA, 5220D) was used to measure COD. pH of samples was measured daily using Jenway 3305 pH meter calibrated at pH 4 and pH 7. A photometric method (HACH 8006) was used to determine the suspended solids content. 25 ml, homogenized samples, collected daily were measured at 810nm using a DR/2010 HACH spectrophotometer. Daily COD values of the effluent were measured in each reactor at each organic load until the reactor seems to coming to a steady state and the daily COD removal% of the reactors were evaluated associated with 3-day moving averages.

At
the beginning of 0.5 CODkg/m 3 /d OLR, the reactors with 25 and 50 m 2 /m 3 SSA showed a COD removal% of below 50, while other two reactors showed an COD removal% above 50 while other two reactors showed a comparatively high COD removal%.
In final 10 days of test period, all four test reactors showed their highest COD removal % and the reactors having higher SSA of media showed higher COD removal% at all the time (Fig. 2-a).
During the first five days run with 1.0 CODkg/m 3 /d OLR the COD removal % in all four reactors did not show an increase in COD removal % and the COD removal % of all 4 reactors was above 50 % from the commencement (Fig. 2-b). From day 31 to 40 the increase of COD removal % was low.
From the beginning of 2.5 CODkg/m 3 /d OLR, all four reactors showed a COD removal% above 55 (Fig. 2-c). Reactor with 100 m 2 /m 3 SSA showed the highest COD removal % during the first half of the test period. In the second half of the test period reactors with 75 and 100 m 2 /m 3 SSA showed a similarly higher COD removal % (Fig. 2-c).
At the beginning of 3.5 CODkg/m 3 /d feeding rate, the COD removal % of all four reactors were below 50 and for about seven days, the increase of COD removal % in all four reactors were also very low (Fig. 2-d). Reactor with 100 m 2 /m 3 SSA showed the highest COD removal% during the first thirty days of the test period. From day 30 to 45, reactors with 75 and 100 m 2 /m 3 SSA showed a similar pattern in COD removal% and did not differ significantly. After day forty five the COD removal% of reactor with 100 m 2 /m 3 SSA increased, achieving a removal of 75%, which was the highest. During the seventy days of test period, the COD removal % of reactor with 25 m 2 /m 3 SSA was the lowest (Fig. 2-d).
In all four reactors the time taken to reach to a steady state increased with the OLR introduced. The analysis of variance carried out for the dependent variable COD removal efficiency, using the model with factors such as SSA and OLR was found to be significant at the probability level 0.001. The interaction effect; OLR x SSA and both the main effects were found to be significant at the probability level, 0.001.