Organic matter removal in bio-filters to achieve biologically stable water downstream of the process

Regrowth in distribution systems, especially pathogens, can have a significant impact on public health. Non-pathogens can also lead to biocorrosion and taste issues, therefore, a focus on the biostability of water and the associated controls are being adopted as they are very useful tools for today's distribution systems (Pinpin et al. 2014). Organic matter removal has also become one of the biggest challenges for pre-treatments of membranes systems (Shanmuganathan et al. 2014) as it causes severe flux decline and affects the quality of the water produced when organic fouling increases. Selection of the treatment to optimize organic matter removal is the key to ensure the best performance of the downstream process units and distribution system.

Biological treatment processes are typically robust systems that are simple to construct, have low energy requirements and are appropriate to communities with limited access to technological resources. The biological processes currently employed for water treatment are almost exclusively media filtration processes where microorganisms growing on the filter media are responsible for biological degradation of organic compounds in water. Some of the most common biological water treatment technologies are biologically active carbon filters, slow sand filters and dual media filters of sand and anthracite although nowadays lightweight expanded clay aggregates can be used successfully as a substitute for anthracite in dual media filters. Biological treatment of water is widely accepted as one of the first engineered potable water treatment technologies and the bulk of the world's sewage treatment plants rely on biological treatment (Keller et al. 2010).

Biofiltration is often used in conjunction with a strong oxidising agent such as ozone. This is an efficient method of treating drinking water as oxidation converts stable organics into compounds that the microbes can remove through assimilation and metabolism. Water produced in this way has reduced Dissolved Organic Carbon (DOC) and hence chlorine demand is low, water taste is improved, biofilm growth is reduced and operating costs are reduced significantly. There is some concern about the use of ozone treatment because it allows some assimilable and biodegradable carbon to pass through the filter allowing downstream biofilm formation.

The experiments in this study were carried out using a laboratory-scale and a full scale plant. The full scale plant is located in Western Australia and has a capacity of 167.000 m³ /day. The assessed process units in both plants are the following: 1-Enhanced coagulation, 2-Expanded clay and sand filters and 3-biological activated carbon (BAC) filters. Water has been treated through the filters, individually and combined in different orders. No oxidant has been dosed in any of the treatments. Flows, empty bed contact time (EBCT) and head losses have been monitored thought the filters to achieve the best performance. The filters have been running continuously for more than a year.

It is a dual media filter comprising sand and expanded clay in a single process unit. The expanded clay is known to provide long filtration times between backwashing, lower backwashing rates, high resistance to blinding by algal blooms and it could potentially lead to extra DOC removal, due primarily to its greater surface area. Under normal operation the filtration rate is 7.5 m/h with 15 minutes of EBCT 8.5 m/h with one filter under backwash.

BAC filtration is known to capture naturally occurring biology in a controlled way to remove organics through bio-assimilation. The BAC design is based on an EBCT of 15 minutes at full flow. This process leads to very long bed life (typically 10–15 years) and the selective removal of those contaminants that can cause taste, odour, chlorine decay, trihalomethane (THM) formation and biofilm regrowth in the distribution network.

Retention time, head loss and pressure have been adjusted to achieve the best performance of the filters.

Enhanced coagulation (C), of the raw water is carried out by optimising the pH of the inlet water with CO2. Aluminium sulphate is used as the coagulant with a polyelectrolyte to maximise coagulant effectiveness.

PH and Turbidity measurements were monitored and undertaken using on line analysers and lab equipment from Thermofisher and Hach respectively. Among the general parameters used to quantify the natural organic matter (NOM), in this study the conventional analytical techniques for measuring the DOC, UV abs 254, Biodegradable Organic Carbon (BDOC), and Assimilable Organic Carbon (AOC) were used in the raw water and downstream of each process units to assess the effectiveness of the removal of NOM. DOC was measured using a Shimadzu and Sievers 5310.

For the BDOC the samples were analysed using the Joret method. Samples were quenched with sodium thiosulfate prior to placing in contact with the biological substrate in the method ratio of 3:1. Samples were placed in duplicate bioreaction vessels. Samples were left in the presence of humidified air for 9 days. They were taken from the reaction vessels periodically and filtered through 0.45 μm membranes before analysis using a Shimadzu TOC (Total Organic Carbon) Analyser.

For AOC the samples were also analysed using the Standard Method for AOC (AWWA Standard Methods for the Examination of Water and Wastewater 21st Edition 9217B). Triplicate samples were dosed with sodium thiosulfate and mineral salts, pasteurised and then inoculated with pseudomonas fluorescens (P17) and aquaspirillum NOX. They were then sealed and incubated at 15 °C. On days 7, 8 and 9 samples were plated in triplicate at dilutions 10⁻², 10⁻³, and 10⁻⁴. Plates were incubated at 15 °C and growth on the plates was counted 5 days after incubation. Average of the growth of both P17 and NOX detected over the 3 days was used to calculate the AOC concentration expressed as μg acetate C equivalents.

Devices such DR6000 from Hach were also used to measure the UVabs at 254. Dissolved oxygen (DO) tests were also undertaken to observe any reduction of DO through the process due to the biological activity, at the same time NH⁴⁺-N was measured to identify nitrifying bacteria. The specific UV adsorption (SUVA) was also monitored and it is defined as the UV absorbance of a sample, measured at a given wavelength (λ), divided by the DOC concentration, providing the average molar absorptivity of all molecules that comprise the DOC in the sample under test. SUVA, therefore, is a parameter that indicates the nature or quality of DOC.

The benefit of sand and activated carbon in biofiltration is well known and recent studies have reported similar biodegradation performance in sand filters and activated carbon filters (Brunet Carreras et al. 2012). Studies show also the benefit of the expanded clay due to its great surface area and porosity which makes the filter able to entrap and develop biological activity on the surface (Eikebrokk & Saltne 2002), and it has been found that similar bacteria (Proteobacteria class, Bacteroidetes, Planctomycetes, Verrucomicrobia and others) to that in granular activated carbon particles can be attached to expanded clay grains (Simon et al. 2013) suggesting that these type of media could act as accumulators of biofilm promoters in equal conditions.

The NOM found in the raw water samples of this study are represented by hydrophobic components with medium molecular weight of organic matter compounds indicated by low SUVA 254 values. The average values were below 2 L mg⁻¹m⁻¹, therefore the aromatic within the NOM was not high, and hence a high biodegradability of raw water can be expected through the biofilters.

In this study, the S&EC filter was monitored in parallel and combined with the BAC filter, samples from the raw water and the outlet of each filter were collected on a daily bases and also monitored via on-line analysers. The behaviour of each filter was different along the operation time as shown in Figure 1.

The S&EC filter showed a low and uniform organic removal along the operation time considering that the maximum removal was achieved after the optimization of the system. On the other hand, the BAC performance since the beginning of the trail until the end showed a different behaviour. Three known stages were monitored in the BAC, the few first months the BAC filters were in ‘adsorption stage’ where most of the organics were removed by adsorption, then a decrease in organics removal was observed (the intermediate stage) due to saturation of the carbon surface with retained slow biodegradable components. At this stage organics were removed by adsorption and biological activity and the removal was lower than when in adsorption stage. Once the carbon was completely saturated the removal of organic carbon became constant, meaning the biological stage of the carbon was achieved. At this stage of the operation when both filters were in biological mode, aluminium sulphate was added in the inlet water after the adjustment of the pH to increase the removal of organics by enhanced coagulation.

The coagulation provides removal of high molecular weight compounds of the organic matter. It has been reported that the removal of DOC via coagulation processes is generally expected for source water with SUVA values greater than 4 L mg⁻¹ m⁻¹. Studies by Hu et al. (2013), confirmed that the contents of AOC increased significantly after coagulation. The removal of those high molecular weight organic matters was the main reason for the increase of microbial growth. Polysaccharides and/or proteins in secondary effluents were easily removed by coagulation and were thought to be possible key of organics substances affecting microbial growth potential during coagulation so it was suggested that post treatments are needed after coagulation to maintain the biological stability of reclaimed water. At the same time recent experiments show that the combination of biotreatment with coagulation could enhance process efficiency, however the same studies suggested that further experiments are needed to determine the best support medium for biofilm growth and the best combination of coagulation with biotreatment in different situations (Zhang et al. 2008). In this study the optimum dose and pH were set up to the limits where the maximum removal of organics and turbidity were observed downstream of the processes, achieving a turbidity target of 0.08 NTU. At the same coagulant dose and pH correction, values of UV 254 at the outlet of the S&EC filter were most of the time higher than in the water treated by the BAC without enhanced coagulation working in biological mode as shown in Figure 2.

The maximum removal of UV abs and DOC was found in the process unit which includes enhanced coagulation and BAC. The Figure 2 shows the UV abs 254 removal along the operation time and it can be also observed the three mentioned stages of the BAC. Once the biological activity was achieved (EBCT = 4000) the addition of coagulant with the correction of pH increased by ‘B’ the removal of organics without affecting the biological removal. As soon as the dose stopped the removal by biodegradation went back to ‘A’ levels as shown in Figure 3.

The regrowth of microorganisms, especially pathogens, can have a great impact on public health, non-pathogens can also lead to bio-corrosion and off-flavor therefore in developed countries high priority has been given to the control of bacterial regrowth in drinking water distribution systems (McGuire 2006). Generally, there are two approaches available to control bacterial regrowth in distribution systems. First, maintain an adequate disinfectant concentration for better inactivation efficiency. Chlorination and Chloramination have been trialed after each process units observing minimum decay and no presence of DBPs (disinfection by products) in the process units with higher organics removal. Second, lower the substrate concentration to cut down the food supply of microbes (Pinpin et al. 2014). At this time of the operation the BDOC and AOC reduction was measured in each of the process units considering different flow paths. AOC values were below measure range (<20 μg/L). The maximum removal of BDOC was obtained in the process lines which include BAC as shown in Table 1.

From the analysis of the BDOC the BAC without previous treatment achieved average values of 0.05 mg/l in the outlet of the filter. The BAC attached bacteria during the biological phase of the filtration, without any oxidant, is able to reduce organic matter to levels where regrowth of biofilm would be not possible as observed by other studies (Page & Dillon 2007) when BDOC in water is less than 0.15 mg/l. Opposite to the S&EC where values were higher than 0.25 mg/l. Figure 4 shows the status of the tanks downstream the S&EC and BAC filters.

The reduction of DO in the outlet of the filters was also monitored by on-line instruments and verified by daily lab tests. At the beginning of the trial there was no reduction of DO along the filtration of any of the filters but once the BAC achieved the biological activity the reduction across the BAC increased to values up to 23% from the inlet to the outlet due to the activity of the attached bacteria. Reduction in S&EC of DO was negligible. In addition, values of NH³-N (mg/l) were analyzed in Table 2 before and after the filters during the biological stage. Low values of NH³-N in the inlet do not provide accurate results in the outlet of the filters as the method is not valid for values less than 0.015 mg/L. Nevertheless the results show a small reduction along the filter media which shows that the decrease in DO of the BAC is due to biodegradation in addition to nitrification. Or – the decrease in DO of the BAC is not only due to biodegradation but also to nitrification.

This study showed that biologically activated carbon together with enhanced coagulation is the best cost/effective biofilter in removing organic matter from water, without affecting the biodegradation that occurs in the activated carbon, achieving the turbidity target of 0.08 NTU after filtration, with the biostability of the water also providing satisfactory advantages to the process. BDOC values in the treated water showed that the treatment produces biologically stable water, reducing biofilm formation which represents an important challenge as a pretreatment for conventional water treatment plants, membrane systems and in the distribution system downstream of the plants.

The surface of the carbon seems to have the most suitable porosity to contribute to the attached bacteria achieving the highest BDOC removal against the sand and expanded clay filters.

In this case having low SUVA values ozone application was not necessary for biodegradation avoiding disinfection by products.