Comparison and predesign cost assessment of ozonation, membrane ltration and activated carbon for the treatment of recalcitrant organics, a conceptual study

The presence of micropollutants in the environment is today of major concern. These pollutants could have long-term impacts on the environment and on population health. Biological treatment of wastewater is generally insucient to allow their complete elimination. The establishment of ecient treatments is then needed to degrade the refractory organic matter. Activated carbon adsorption, membrane ltration and oxidation processes are common suitable solutions. All of them have advantages and are effective to treat wastewaters but drawbacks are well known such as waste production, energy consumption or by-products formation. This study aims at dening a strategy to choose the best option according to the nature of the wastewater and the treatment objectives. A methodology was designed for the rating of theses processes to choose the best strategy regarding environmental, technical and economic criteria. A simulation of three wastewater treatment scenarios was carried out to compare the costs of ozonation, adsorption and reverse osmosis. According to the result obtained, a decision tree is proposed to dene the best option for a tertiary treatment to reach reuse or discharge objectives.


Introduction
The presence of micropollutants in today's environment is of major concern. Research has permit to create new chemical compounds for medicine, chemistry, cosmetics and agriculture (phytosanitary) uses. These compounds belong to the family of surfactants, ame retardants, pharmaceuticals, cosmetics, petrol additives, biocides, pesticides and all of their degradation products. The fate of these new compounds in the environment need to be measured as their behavior, their fate and their (eco)toxicological effects are not very well known (Norman, 2018).
They can have toxic effects (carcinogenic, mutagenic or reprotoxic) or even interfere with the hormonal system of living beings (endocrine disruptors). They could also have health and environmental effects in short or long terms. Furthermore, the potential risk of a "cocktail" effect due to a mixt of these compounds must be considered. Over 1,000 substances have been referenced as emerging environmental substances (Norman, 2019). Since 2000, the European Directive 2000/60/EC (European commission, 2000) sets the reduction target of hazardous substance emission in water. A list of substances or group of substances concerned by the emission reduction (de ned as priority substances) or elimination (considered as hazardous With the tightening of regulatory constraints, releases of some pollutants are decreasing, such as pesticides, however other substance releases are increasing, like pharmaceutical substances (Metz & Ingold, 2014). The organic pollutants including emerging pollutants in urban or industrial e uents are not completely biodegradable and a part of these compounds are refractory to biological treatments. Consequently, pollutants are measured at the output of activated sludge treatments ( , which are the most conventional treatment for urban wastewaters. The main mechanisms involved in the elimination of micropollutants in biological treatments are mainly biodegradation but also sorption onto sludge, air stripping and phototransformation (Gusmaroli et al., 2020). The implementation of tertiary treatments may be required to allow the elimination of these refractory pollutants and to increase the water quality discharged to surface waters.
Adsorption on activated carbon, Reverse Osmosis (RO) and oxidation are the common treatment solutions implemented as tertiary treatment. Activated carbon is widely employed to remove organic compounds in industrial or drinking waters. This treatment presents various advantages: easily to implement and operate, no chemical needs and no generation of by-products. However, the replacement of the spent activated carbon is necessary and the cost involved limits its use on low COD concentration e uents. RO presents also very high treatment performances but requires e cient pre-treatment and high opertational costs due to energy consumption, elimination of the retained pollutants (retentate) and a quali ed staff to operate and control the process. Only the oxidation processes make it possible to really degrade polluants but formation of unknown and potential toxic by-products can occur due to partial oxidation of the compouds (Hamdi El Najjar et al., 2014;Wu et al., 2019).
Studies comparing the costs of tertiary treatments to reduce the discharge of micropollutants are generaly available for urban wastewater treatment plants (Bui et al., 2016;Wahlberg et al., 2010). Few works exist on the eld of tertiary treatments employed for industrial wastewaters. This study aims to achieve a comparison and a predesign cost assessment of ozonation, reverse osmosis and activated carbon for the treatment of an industrial e uent with a COD concentration higher than values commonly found in urban discharges. Attempts have been made to estimate the capital and annual operating and maintenance cost (O&M cost) for a 2,000 m 3 d -1 capacity treatment plant with a COD concentration of 500 mg L -1 . Secondly, a multi-criteria approach is proposed to compare the performance of the processes. Finally, a decision tree is designed to select the most appropriate treatment for the elimination of dissolved organics.

Granular Activated Carbon (GAC) cost
GAC is implemented in a lter bed for the adsorption of organics on granular carbon. Process energy requirements are low for GAC and include both supply and backwash pumping (Hansen et al., 1979). The pollutants are eliminated by adsorption due to their a nity with the activated carbon and its high speci c surface area of this adsorbent. The consumption of activated carbon can be rst estimated from the COD load treated, typically in the range of 250 to 500 g COD kg -1 AC or higher (Truc, 2007). In rst approach, a consumption of 250 to 300 g COD kg -1 AC is generally considered. Performances of this treatment are dependant of the organics compounds to be adsorbed and in particular the polarity, molecular weight, solubility and concentration. This can be evaluated in laboratory by adsorption isotherms. After the saturation, activated carbon must be replaced and reactivated in high temperature ovens. In France, the reactivation of coal is done in specialized centers, it is too expensive to be carried out on user sites (Bui et al., 2016).
The reactivation yield is dependent on the type of carbon and the nature of the molecule adsorbed. For charcoal made from softwood (pine) this yield is relatively low (70-90%), while for charcoal made from coconut it can reach 98% (information obtained from Chemviron 2019). Treatment of the spent GAC in a reactivation center will require a prior acceptance certi cate with limits to be respected for certain parameters such as sulfur, chlorine and uorine.
The cost of GAC is generally between 1 and 4 € kg -1 and the cost for the reactivation is 0.6-0.7 € kg -1 (excluding transport). The reactivation cost is slightly higher than the elimination cost (0.4-0.5 € kg -1 ), but leads to savings on the purchase of new GAC until it cannot be reactivated. Activated carbon treatment should not be used when treated uxes have too high COD due to the costs associated with the carbon reprocessing. The investment costs were identi ed and estimated by the company IRH as part of a study carried out for the Rhône Mediterranean Corsica Water Agency. This preliminary design approach is exclusive of taxes and fees and doesn't include supply (cost and mankind), the contracting authority staff and the project management This equation was mainly de ned using a simulation tool created by the Water Research Foundation and the USEPA.
Altogether, the investment costs depend on the ow treated and on its composition in uencing the nature of the material used, the kinetic of ltration end the number of lters used.

Membrane ltration cost assessment
Reverse osmosis and Nano ltration (NF) are implemented for water reuse or very strict constraints on discharges (very low threshold or low water ow rate authorized for discharge). They require more e cient pre-treatment than adsorption and ozonation. Energy consumption is higher and the elimination of retentate (10 to 30% of the initial volume of water treated) remains problematic impacting the operating fees. NF and RO investment costs, identi ed by the company IRH, are estimated . Only few works exist on these tertiary treatments on industrial wastewaters. The economic considerations drive the selection of the process implementation. Here, a simulation of different scenarios for is proposed to treat a conceptual e uent with a COD concentration at 500 mg L -1 (higher than urban waste concentration). A comparison of the 3 different treatment processes (ozonation, adsorption and membrane ltration) is carried out on a cost basis. Treatment channels considered are presented in Fig. 1. The channel with membrane ltration (ultra ltration / reverse osmosis) is considered to study a scenario with water reuse.
In the different scenarios tested, the three post processes are described as follow: Membrane channels: wastewater is rst ltered by ultra ltration and reserve osmosis and retentate is post-treated by ozonation. Two scenarios are proposed here with permeate directly discharged (Channel 1) or reused (Channel 1 bis).
Activated carbon channels: wastewater is treated by AC post-treatment, the obtained water can directly be discharged.
Two scenarios here are considered using either reactivated AC (Channel 2) or new AC (Channel 2 bis).
Finally, wastewater can be processed directly by ozonation treatment allowing direct discharge.

Hypothesis used in the simulation
General assumptions used for simulation are: A ow rate at 2 000 m 3 d -1 , 24h/24, 365 days per year; An electricity cost at 0.1 € kWh -1 ; A staff cost xed at 50 € h -1 .
The capital cost is amortized over 20 years (n) ( Speci c hypotheses of each scenario are listed in Table 1.  This simulation provides a comparison of the costs for the different tertiary treatment plants and shows the impact of the various assumptions on the overall result. However, estimations are carried out on the basis of average treatment performance assumed for a ctitious e uent and not from available data for a real e uent. The treatment rates necessary for ozonation, the consumption of activated carbon as well as the performances of membrane ltration are speci c to each e uent and could differ signi cantly from the assumptions made for this conceptual study. The results showed that (thermal) reactivation allows to avoid impacts linked to mining and transformation of raw coal, to natural gas consumption used (reactivation represents 40% of the gas consumption needed for the AC production), to toxicity risk (by 25%) and to CO 2 emissions (by almost 90%) (Bayer et al., 2005). The implementation of a tertiary treatment is generally effective in reducing the residual toxicity of an e uent if the potential generation of by-products generated by oxidation processes (ex. ozonation) is controlled. However, the tertiary treatment requires additional consumption (chemicals, electricity, construction materials) which may affect the overall environmental impact. This is the conclusion reached by Rahman et al. (2018) for the treatment of emerging pollutants leaving urban wastewater treatment plants.

Comparison Of Processes And Strategy
Membrane ltration is the most impacting process due to the high energy consumption implemented unlike reactivated carbon ltration. In the future, an optimization of tertiary treatments is necessary to reduce their environmental impact and in particular their energy consumption (Margot, 2014).

Page 8/19
A comparison of the different processes studied is complex because the mechanisms for removing pollutants are very different. Membrane ltration and adsorption are separative processes while oxidation processes are destructive processes. Indeed, pollutants retained by RO or GAC are effectively eliminated from the treated water while the oxidation processes generally produce oxidation by-products leading to potential incomplete mineralization of pollutants. Moreover, performances are dependent on the wastewater quality (salinity, type and pollutant concentrations), the technology used and the operation conditions (type of membrane, pressure applied, nature of the activated carbon, kinetics, etc.).
Despite these points of attention, a summary of process performances according to category of pollutants is proposed in Table  2. Reference Membrane ltration reach the better performances for the elimination of priority and emerging pollutants. Granular activated carbon is e cient on a large part of the pollutants studies but organometallics and some pesticides can't be eliminated. In the same way, ozonation is not o good option for organometallics and per uorinated compounds but it can reach good performances on some pesticides, drugs and Polycyclic Aromatic Hydrocarbon.
Performances of ozonation and activated carbon were compared in the framework of the MicroPoll project and showed that ozonation followed by a sand lter and Powder Activated Carbon (PAC) is effective in removing the majority of micropollutants with similar average removal rates. Ozone is very effective for some types of pollutants while PAC (Powder Activated Carbon) acts on a wider range of substances (but with lower yields) (Margot et al., 2013). Ozone and PAC signi cantly reduce the e uent toxicity with comparable costs (PAC separated by sand lter) (Margot et al., 2013) or higher (ultra ltration used for PAC recovery). Combination of ozone and activated carbon allow to avoid the risk of toxic oxidation by-product rejection, adsorbed by AC. This coupling is implemented on urban wastewater treatment plants in Switzerland (Grelot et al., 2017). The French project Ampère has shown that RO has very high micropollutants abatement performance. Molecules hardly eliminated by activated carbon and ozonation were retained (Ruel et al., 2011).

Comparison of the processes
The three processes studied have speci c characteristics with their own performances and degradation or separation e ciency. A comparison of the tertiary processes is proposed in Table 3 considering the main criteria that impact the choice of the treatment. To complete this rst comparison, a multi-criteria approach is proposed to assess and compare the processes. Adapted from the methodology reported by Fast et al. (2017), this approach is based on the evaluation of process performance according to technical, environmental and economic criteria. Different key points have been de ned with a rating base proposed allowing the reproducibility of this approach presented. The rating is carried out by applying a score ranging from 1 to 5. The highest scores correspond to the best performances. Technical and environmental criteria rating are based on the data listed previously and regrouped in topics detailed by method in Supplementary material. This ranking is not an absolute indication indeed, all the criteria were considered without weighting according to the method used by Fast et al. (2017). This could be discussed, in the same way other criteria could have been considered. The scoring results obtained are shown in Table 4 and Fig. 2.   First, all the process analyzed have a ranking between 3.3 and 4.3 showing globally good performances. Activated carbon adsorption obtains the best results when reactivation is applied on the spent carbon. On the contrary the RO presents the lower score due to the complexity of operation with the risk of fouling and scaling. Moreover the production of wastes (retentate) is a major drawback that increases the O&M cost. On contrary the possibility of reusing water is a great advantage that can lead to choose this solution. It can solve problems of availability of freshwater or impossibility to discharge wastewater.
The results of this ranking have to be considered only for the scenarios studied. Other results could have been found for other wastewaters with lower owrate or organic concentration. The choice of the process will be made according to the e uent to be treated ( ow rate, load, nature of pollutants) and the local context. Moreover, the quotation applied may differ depending on the person carrying out this quotation with a subjective part on the weighting parameters for example. However, the proposed methodology makes it possible to de ne technical, environmental and economic criteria and to balance it according to country legislation.

Strategy de ned for the estimation of the tertiary treatment best option
These different processes can be complementary and must be chosen according to the e uent nature to be treated, its ow rate and the treatment objectives to be achieved.
Biological processes not evaluated in this study have the lowest costs and are preferred when pollution is biodegradable. They are widely used for the treatment of urban and industrial wastewater. Partial oxidation can be necessary to increase the biodegradability of an e uent before a bio-treatment in presence of recalcitrant pollutants. This option is generally less expensive than total oxidation when the residual COD is greater than 350 mg/L. Membrane ltration will be useful for water reuse projects and/or when the e uent is refractory to oxidation. Adsorption will be implemented for the treatment of low ows of organics or pollutants refractory to oxidation. The adsorption can also make it possible to eliminate oxidation by-products. A simpli ed decision tree has been designed to choose the best option according to the e uent composition for the treatment of organic pollutants from industrial or urban e uents (Fig. 3).
As described previously, the choice of the best wastewater treatment scheme will be a combination of processes on a case-bycase basis depending on the e uent to be treated and the treatment objectives. However, by coupling this decision tree and the ranking methodology presented, an impartial and customizable choice can be made.

Conclusion
This study aims at establishing the better course for e uent remediation with dissolved organic matter load with refractory compounds. Different wastewater treatment strategies can be studied depending on the nature of the wastewater, the water ow rate and the treatment objectives.
A methodology was designed for the rating of tertiary processes to choose the best strategy on environmental, technical and economic criteria.
From this tool, we have compared the most effective tertiary treatments for recalcitrant COD. The best result was obtained by the activated carbon when reactivation is applied on the spent carbon which allowed to reduce the O&M cost and the environmental footprint. On the contrary the reverse osmosis presents the lower score due to the complexity of operation and the production of retentate which is a major drawback that increases the O&M cost and the environmental footprint. However, the possibility of reusing water is a great advantage that can lead to reduce the O&M cost and solve problems of availability of freshwater or impossibility of discharge of wastewater. Ozonation obtained intermediate result. It is very effective but costs can be high when the organics concentration of the wastewater is high. Despite their performance, oxidation processes generally do not completely mineralize organic matter and the formation of potentially toxic oxidation by-products can be problematic. A post-treatment (sand lter, adsorption on activated carbon or bio-treatment) may be necessary to eliminate these by-products. Generally, ozonation is very effective, however, it can't oxidize all the molecules. Advanced Oxidation Processes (AOPs) could be a solution in this case. They are able to generate non-selective and highly reactive hydroxyl radicals, which are effective on the majority of organic compounds. They can be chemical, photochemical, catalytic, electrochemical, sonochemical and physical. The AOPs will be reserved for the treatment of the most problematic e uents with di cult compounds. Generally, adsorption on activated carbon and membrane ltration (NF/RO) can deal with e uents refractory to oxidation processes. When the refractory COD concentrations are high, other methods will be more suitable. Evapo-concentration or thermal oxidation processes (wet oxidation, supercritical oxidation) can manage e uent concentrations up to 100 g L − 1 of COD, above these concentrations incineration is generally the only treatment solution. Wastewater treatment scenarios used for the cost evaluation. The COD concentration of the water discharge is less than 50 mg L-1. Figure 2