Introduction

The U. S. Environmental Protection Agency (EPA) promulgated effluent guidelines for organic chemicals, plastics, and synthetic fibers (OCPSF) industry category in late 1987. (Federal Register, November 5, 1987) Pretreatment of OCPSF wastewaters prior to discharge into a public-owned treatment works is now a requirement.

The pretreatment standards are very stringent, and the compliance deadline, extremely tight. Many of the standards promulgated for 47 pollutants are below 500 micrograms/L. For some pollutants, the standards are as low as 19 micrograms/L. The deadline for compliance is November, 1990.

BCM Engineers, Inc. (HQd) has been retained by a large-size chemical company to assist in its compliance with fine regulation for a plant in the east coast.

The project is the preliminary engineering phase, which includes the following tasks:

• Wastewater charcterizaticn
• Assessment of alternative ocmpliance strategies
• Source control evaluation
• Treatability studies, including bench-scale aryl pilot-scaletestings
• Process design

The work is in progress. This paper highlights the complexity of the project and the results of the work completed as to date. Some of the complexity stems from the heterogeneous product mix and intermittent wastewater flow generation patterns. 

Products and Manufacturing Processes 

The plant manufactures more than 900 organic chemical products. Moreover, the plant also operates a large number of pilot-scale process development units for making new products. Functionally, the plant is divided to six production areas, more or less along the product lines.

A significant number of production processes employ batch operation. Typically, the production process involves: charging the reactor with raw materials: reaction under controlled conditions: separation of products from other materials: and reactor wash and cleanout before product charqeaver. The reactor sizes range from 100 gallons to 6,000 gallons. Some products share the same reactors and separation equipment.

The product schedule is highly dependent on market demand for each product, which is not totally predictable. The high product variability and episodic nature present a unique challenge to wastewater characterization and abatement.

Wastewater Stream and Major Contaminants 

The plant currently generates about 3.1 million gallons per day (MD) of wastewater from aver 200 streams. After solvent separation and PH adjustment with lime, the wastewater is being discharged into a public sewer for further treatment. About 75 percent of the waste streams fall in the OCPSF category.

The major sauces of wastewater are reaction vent gas scrubbing, reactor aryl tank washes and cleanarts, floor washes, vacuum jets and tank waga-/rail tank car cleaning. Another group of wastewaters is storuarater runoff on process areas, material transfer stations, and tank farms.

Table 1 shows the key OCPSF contaminants and expected concentrations in the combined plant effluent.

Most contaminants are organic solvents. In addition to these contaminants, other impurities, such as surfactants, monaneric organics are present. These impurities can interfere treatment processes.

Pretreatment Effluent Standards

Table 1 also shows the concentration standards for the contaminants. The standards apply to OCPSF-regulated wastes only, prior to mixing with any other stream.

When treated OCPSF wastes are mixed with aryl dilution water, such as sanitary waste and non-cooling water, the values in Table 1 must be proportionally lowered in accordance to the EPA's Combined Wastestream Formula. (40 CER 403 EPA Pretreatment Standards) Therefore, depending on the quantity of the dilution water, the actual concentration limitations at the plant outfall can be significantly lower than what Table 1 shows.

Alternative Compliance Strategies  

Three strategies being considered are: control or treatment of wastewaters at sources (source control); treatment of entire plant combined effluent at a central facility (end-of-pipe); and combination of these two strategies.

Philosophically, plant management prefers source control over other strategies. Management believes that the best solution to pollution is to avoid pollution in the first place. Each production area most be responsible for its pollution. This strategy assures the accountability and the fair allocation of abatement costs to the production area that contributes to the waste generation. This strategy encourages waste minimization.

End-of-pipe treatment is an opposite strategy. The potential advantage of this strategy is the ewes of scale. Since many production areas generate the same contaminants, a centralized treatment facility world be able to remove them together at a lea costs. Also, a central facility wand be amenable to a full-time staff whose primary responsibility is to treat the wastewater.

While the evaluation of the alternative strategies is still in progress, it appears that a combination of scarce control and end-of-pipe treatment would be most attractive. source control will be implemented selectively on the abet concentrated streams or wastes that bear unique contaminants. End-of-pipe treatment will control lightly contaminated streams that share ccammn contaminants.

Waste Characterization

Because of the high product and waste variability acrd a large number of waste streams, accurate wastewater characterization is a major challenge.

The wastewater characterization is based on a oambination of three techniques: (1) Wastewater survey; (2) sampling: and (3) cuter simulation.

Wastewater Survey

This involves a thorough review of production processes to identify all wastestreams. For each waste, hydraulic flow, frequency and duration of corresponding production rates, contaminants and their mass flow rates are estimated. The compilation of the wastewater data has been tended a waste inventory.

The wastewater survey, however, cannot account for all contaminants present. This is especially true when a unapt is a trace impurity in the raw materials, or a by-product of the reaction. Therefore process review alone cannot provide sufficient Information for design. Therefore, these technique were supplemented with wastewater sampling.

Wastewater Sampling

As of March, 1989, over 75 samples have been collected at sources, junction roaW oles in the sewer, and at the plant effluent. The analyses include the organic priority pollutants (EPA GC/N5 Method 624/625) and a wide range of convention parameters. As required by FAA protocols, all volatile organics samples were grab samples.

While sampling generates useful design data, it has significant limitations.

First, complex matrix and interferences encountered in the samples have caused elevated Method Detection Limits (MDLs). Often, the MDLS are several order-of-magnitude higher than the OCPSF concentration standards. Consequently, it is not discernible as whether a contaminant is present in a waste stream, and if present, whether the concentration is higher than the standard.

One cause of the problem is an extremely wide range of concentrations of the OCPSF contaminants present in a given sample. To protect the instrument, and to lower the peak concentration within the acceptable range of the instrument, dilution of the sample is needed. Hut the dilution obscures the oCPSF contaminants with lower, nevertheless significant, concentrations. A second cause is coelution: contaminants of low or moderate concentrations elute from the column at the same time aryl the GC/MS peaks for each contaminant coincide.

To overcome these problems, a site-specific "two-dimensional" GC/MS method is being developed by a research-oriented laboratory. The two-dimensional GC/MS uses a purge and trap interface on a GC equipped with a nondestructive thermal conductivity detector (TCD). To detect alai quantitate minor components when present with large quantities of other components, the GC effluent will be split into manageable fractions. The TCD will be used as an on-line screening tool to monitor GC efficient levels. Sections of the column effluent containing large, unresolved peaks will be refocused by one adsorbent trap while low level analytes will be refocused on a separate trap. This will allow improved separation of coeluting species under different chromatographic conditions and will facilitate identification of species formerly obscured by high level components. The second chromatographic separation will be performed on a GC/MS system equipped with ester library search.

The second limitation is the uncertain representativeness of the sampling. Because of time constraints aryl cast considerations, only limited amount of sampling is practical. Moreover, EPA protocols require collection of grab samples, rather than 24-hour composites, for volatile organics. Sampling therefore only provides a "snap-shot" of the highly variable wastewater characteristics. Also, the sampling does not fully reflect the effect of full-scale source control measures that yet to be implemented. The Monte Carlo ccuprter simulation intends to partially mitigate these limitations. 

Computer Simulation 

The Monte Carlo model is used for the simulation. Figure 1 depicts a simplified flow chart of the model. For specified sources and contaminants, the model randomly predicts whether a scarce is discharging in an increment time period (e.g., an hour). For that time period, waste hydraulic flows and contaminants mass loads of all discharging sources are added to yield total emissions. The simulation continues for all time periods (e.g., over one year). The flow and load data are then analyzed statistically to produce design information, such as 90 percentile values.

The model can stochastically estimate the wastewater characteristics for individual buildings, combination of buildings, aryl at the plant outfall. The simulation can generate hourly, daily, or monthly wastewater characteristics.

The simulation will be used to performance "what if" analyses. As source control/treatment projects are defined, the simulation will show the impact of these projects on the wastewater characteristics.

The data in the waste inventory is the input for the simulation. The information generated from sampling, and historic plant wastewater data will be used for the model calibration.

 Source Control 

The objective of source control is to reduce the volume and/or strength of the wastewater. Another objective is to eliminate contaminants that can interfere with downstream wastewater treatment processes. The current goal of swine control to reduce the plant wastewater flow by more 50 percent, appears achievable. Far certain OCPSF contaminants, a reduction of mass loading by more so percent is also possible.

A variety of source control techniques have been or are being evaluated by the plant personnel and BCM. Techniques being considered include, but are not limited to, the following:

• Material recovery and recycle
• Chemical substitutions
• Employment of dry processes
• Containment of material transfer stations
• Segregation of wastes
• Good Housekeeping

Material Recovery and Recycle

 Over recent years, the plant has been evaluating ways to reduce or eliminate the discharge of OCPSF chemicals. For example, one production area has devised a technique to drastically reduce the discharge of 1,2-dichloroethane and 1,2-dichloropropane. These solvents are recovered by stripping with nitrogen girl subsequent condensation.

Evaluation of recovery and reuse of toluene from one major source is underway. If successful, this source control measure can reduce the toluene discharge by a considerable percentage. Bench scale testing is required to firmly define the performance.

 Chemical Substitution 

Potentially, the toluene discharge can be further reduced by chi substitution. One production area is evaluating whether toluene can be replaced with methanol, which is a non-OCPFS chemical and highly biodegradable.

Employment of Dry Processes

Replacement of vacuum jets with vacuum pumps will significantly lower contaminant discharge. Vacuum jets are widely used to evacuate gases from reactors. Stem, which is used by the vacuum jets, is in direct contact with the gases. Therefore, the steam condensate is laden with contaminants, such as phenol. In contrast, vacuum pumps, which are dry operations, will not create wastewater.

Contaiment of Material Transfer Stations

Since a major source of OCPSF contaminants is stormwater from the material transfer stations, contaiment and diversion of the runoff for treatment will significantly reduce the discharge.

Segregation of Wastes

Segregation of heavily contaminated and lightly contaminated wastewaters will aid in source control. While this technique may not eliminate wastewater discharge, it will reduce the size and complexity of the end-of-pipe treatment.

An engineering evaluation of a dual sewer system is in progress. (See Figure 2) One sewer will receive OCPSF wastes, and the other sewer will carry non-OCPSF wastes. The OCPSF wastes will be treated by a central treatment system at the plant: the non-OCPSF wastes will be treated by the municipal wastewater treatment system.

The dual-sewer system offers potentially attractive features. First, the size of the implant treatment system for OCPSF wastes alone will be much smaller than an end-of-pips treatment system for the combined OCPSF and non-OCPSF plant effluent. Second, each production area will have options of: (1) sending the wastewater to the OCPSF sewer and paying for the treatment: (2) eliminating the wastewater completely through source control; or (3) treating the wastewater at the scarce and discharging it to the non-oCPSF sewer.

Good Housekeeping

Equipment modifications, operation procedure charges, and training will farther reduce the discharge of contaminants fran each production area. A comprehensive monitoring program will be implemented to enforce the good housekeeping.

Treatment Technologies

Parallel to source control investigations, evaluation of treatment technologies, which can be used at sources or at the end-of-pipe, one in progress. The technologies being evaluated include the following unit processes:

• Stripping
• solvent extraction
• Foam fractionation
• UV/peroxide oxidation
• Carbon adsorption
• Resin adsorption
• Biological oxidation

Stripping

Many of the OCPSF contaminants present in the wastewaters are volatile, therefore, amenable to stripping. Both air stripping and steam stripping are being evaluated. Air stripping will remove chlorobenzene, ethylbenzene, methylene chloride, methyl chloride, 1,1,1-trichloroethane, toluene, tetrachloroethylene, and trans-l,2-dichloroethylene. Since air stripping merely transfers the contaminants from the liquid to the gas phase, an air pollution control device will probably be needed.

Air stripping will not effectively remove less volatile species such as 1,2-dichloroethane and 1,2-dichlorapropane; steam stripping would be necessary for then. Steam stripping should also remove some phenol and naphthalene. However, additional polishing steps following steam stripping may be required to meet the OCPSF standards.

Solvent Extraction

Extraction of phenol and other materials of low volatility from concentrated streams at sources is being considered. Candidate solvents are methyl isobutyl ketone (MIBK and hexane. Although MIBK is an effective extractant, its solubility in water is relatively high. Recovery of MIBK( from the extracted wastewater will probably be necessary.

Foam Fractionation

Foam fractionation is based on the selective adsorption or attachment of material on the surfaces of gas bubbles passing through a solution or suspension. This process could potentially remove the surfactants. If not removed, the surfactant may impede chemical coagulation and stripping effliciency.

UV/Peroxide Oxidation

UV/peroxidation creates the hydroxy radical, which has an high oxidation power to degrade many of the organics present in the wastewaters. However, this process is very energy-intensive. Therefore, it is more suitable for polishing.

Carbon Adsorption

Many of the organics present in the wastewater are adsorbable by activated carbon. These contaminants include chlorobenzene, ethylbenzene, naphthalene, and toluene. Methylene chloride, 1,2-dichloroethane, 1,1,1-trichlomethane, tetrachloroethylene and trans-1,2-dichloroethylene are also adsorbable, but to a lesser degree.

Like UV/peroxide oxidation, activated carbon will most likely be a polishing process. Using activated carbon as the primary means of organic removal will not be economical because of exorbitant carbon consumption. The organic contaminant concentrations are very high, and activated carbon cannot selectively remove OCPSF species.

Resin Adsorption

Resin adsorption is more selective than activated . Resins, which are organic polymeric adsorbents, are commercially available. The resins are reportedly effective for removing phenolics, chlorinated pesticides and other non-polar organics from aqueous wastes. The binding affinities between the resin and the organics are weaker than that between activated carbon and the organics. Therefore, the resin can be readily regenerated with water-miscible organic solvent, such as methanol. Resin adsorption, too, is likely suitable for polishing only.

Another potential application of resin adsorption is treatment of the air stripper off-gas. As cared to activated carbon, the resin is less sensitive to the humidity. Therefore, heating or dehumidication of the off-gas before adsorption may not be necessary.

Biological Oxidation

Biological oxidation using the activated sludge process or the powdered activated carbon-augmented activated sludge process were considered. However, among other reasons, this process requires a large space, which the plant does not have, and generates significant quantity of sludge, which is not consistent with the plant's waste minimization policy. Therefore, this process has been dismissed.

Conceptual End-Of-Pipe Treatment Scheme

A pilot plant has been constructed to test the end-of-pipe treatment schemes far OCFSF wastes as depicted in Figure 2. (See Figure 3) Other unit processes are being evaluated at the bends-scale.

The pilot plant consists of pretreatment, primary treatment, and polishing step, using the treatment technologies discussed above.

Pretreatment

The pretreatment step is to remove suspended solids. Removal of the solids is critical to protect the integrity aryl performance of the downsstream processes. The processes are chemical flocculation using a ferric salt, gravity settling and filtration. Preliminary indications are that the sludge generation rates are excessive and that the sludge may be a hazardous waste, thus raking offset disposal expensive. Alternative processes are being tested on a bench-scale basis include dissolved sir flotation and microfiltration.

The alternative primary treatment processes steps being tested are air stripping and steam stripping. preliminary results indicate that air stripping does not remove enough of the total mass of organic contaminants to render the polishing step cost-effective. Steam stripping is more capable, however, there are significant treatment interferences in this wastewater that could make steam stripping ineffective. In particular, there is a high level of surface active materials in the wastewater (in the range of 50 mg/1 to 500 mg/1). This can result in excessive foaming to the point where the steam stripper may not work. In that case air stripping may have to utilized after all if stripping is to be used.

Polishing Step

Polishing treatment is seen generally in relation to stripping in this project. Compounds of concern can be classified into two major categories: adsorbable and nonadsorbable since a major portion of the adsorbable compounds are more difficult to strip, a natural division of roles between the two unit processes is envisioned.

Activated carbon has been modelled mathematically using the HSDM developed at Michigan Technological University (Crittenden, J. and David Hand). This information in turn was used to refine the design of the pilot study and to set priorities on source control based on the relative difficulty of treatment.

Discussion and Conclusion

This chemical manufacturer's corporate policy is to be responsive to environmental regulations, both promulgated and potential. In addition, the company has specific policies encouraging waste minimization and source control. This project is an example how innovative approaches to complex problems can be achieved when a manufacturer has done the necessary groundwork and is omaaitted to an appropriate solution.