The trickling filter process has exhibited certain practical characteristics which have made it a popular secondary waste treatment method. Simplicity, operational economy, ability to give good effluents under wide variations of load, and quick recovery from pH variations and shock loads of toxic substances are factors which frequently result in choice of trickling filtration over the activated sludge process.

Solid surfaces are of practical importance in processes depending upon bacterial action on dilute solutions of organic substances. Zobell(1) in 1937 suggested that enhancement of the physiological activities of bacteria when allowed to grow in dilute nutrient solutions in the presence of solid surfaces could be attributed to:

1. The possible concentration of nutrients and enzymes by adsorption to the

surface.

2. The interstices between bacterial cells and surfaces acting as concentration

points retarding the diffusion of exoenzymes and metabolites away from

the cell thereby favoring both digestion and adsorption of foodstuffs.

3. Interstices between surfaces and cells serving to produce optimum

oxidation-reduction or other physico-chemical conditions.

4. Surfaces functioning as attachment points for organisms which are

Zobell worked with surfaces submerged in aqueous solutions containing minute concentrations of biologically oxidizable compounds, reminiscent of the procedure for determination of the biochemical oxygen demand of wastes. Both the trickling filter and activated sludge processes have provided fertile areas for speculation as to the mechanism or mechanisms involved in the rather profound changes that occur in the characteristics of a waste which has undergone treatment. Of the two, the trickling filter has proven to be the more elusive and intriguing. One reason for its ability to excite the observer is the relatively short time of passage through the filter during which the liquid fraction of a waste changes its character so markedly.

Contact time, which is partly a function of hydraulic application rate, may be as short as a minute. During this apparently brief contact time a degree of oxygen demand satisfaction may be observed which is only obtained after hours of processing in aeration tanks or after many days of natural "self-purification."

A trickling filter may be defined as a device which concentrates within a relatively compact structural arrangement a large amount of surface upon which bacterial films can develop in contact with thin films of the waste undergoing treatment. A typical stone bed trickling filter six feet deep may provide 90 to 120 square feet of oxygen transfer-bacterial attachment surface for every square foot of land occupied by the unit. It would appear that further concentration of surface area could be achieved by packing filters in greater depths; however, a head loss or pumping penalty is associated with greater packed depths. Within limits an economic analysis can be used to select the best configuration of a unit. The economic balance is partly affected by the nature of the biological reaction which has been observed to progress at a rate proportional to the concentration of organic matter remaining to be removed. Velz(2) expressed a basic law for the performance of trickling filters,

He pointed out the similarity between this and the relationship expressed by Phelps (3) in his development of an expression for the self-purification of streams,

With "t" expressed in days, the value cited by Phelps for "k" was 0.1. Velz cited values for "k" of 0.175 and 0.1505 for low and high rate filters with corresponding values of L of 0.90 and 0.784 respectively.

In developing a theory relative to the relationship of contact time to satisfaction of oxygen demand, Howland(4) arrived at an expression for inclined surfaces assuming laminar flow conditions as follows:

T = (3v)/(g * sin Q ) * L/(q^2/3)

  where L = Length of plane

Q = Angle of inclination

g = Gravitational constant

v = Kinematic viscosity of water

T = Contact time (time of flow, distance, L)

q = Liquid application rate

Bloodgood, Teletzke, and Pohland(5) verified Howland's prediction of contact time varying inversely with the 2/3 power of liquid application rate. Howland(4) stated that an analysis of published data indicates that fractional removal of oxidizable organic matter accomplished depends primarily upon the time of flow. The implications of their work were that a deep filter containing the smallest practical media is to be preferred to obtain an optimum contact time and maximum efficiency.

Selection of the smallest practical media is largely a resolution of conflicting basic principles. The proportion of surface area to occupied volume increases as the particle size of filter media decreases, but porosity and permeability to air and liquid flow decreases. In order to satisfy the oxygen demand in waste waters, (usually greatly in excess of the saturation level of oxygen in water) filter media must possess favorable free-vent properties for simultaneous passage of air and waste. This imposes a practical minimum on particle size and for any given particle size, a practical maximum upon the depth of the unit. Trickling filter literature contains conflicting opinions as to the effect of depth on trickling filters. Buswell(6) found the quality of effluent at the ten-foot depth in experimental filters to be no better than that at the six-foot depth.

Ponding and slogging became limiting factors at higher rates of filtration. He concluded that shallow filters at lower rates of application were better than deep filters at higher rates because there was no need to pump as high and the lower reaches were ineffective. Clark(7), however, concluded that ten-foot filters could be operated successfully with four to five times the load of that on a five-foot filter. The Committee on Development of Uniform Standards for Sewage Works(8) has limited the choice of bed depth from a minimum of five feet to a maximum of seven feet for standard rate filters while not specifying depth limitations for high rate filters.

It may be concluded that conventional filter media impose a restriction on the designer who may find ample economic justification for using substantially deeper oxygen transfer beds than permissible with stone media.

Ingram(9) introduced an approach to trickling filter design in which substantially greater media depths were suggested possible through provision for application and removal of waste at intermediate positions and insuring aerobic conditions throughout the filter by providing forced ventilation. The experimental unit for which data were reported contained an overall pacing depth of 18 feet. Although offering a substantial challenge to the engineer in extending this concept to large scale installations using stone media, it indicates progress toward giving the designer greater freedom of choice in configuration of the media.

Shallower beds became progressively more important as more effective usage was made of filter media by higher hydraulic rates of application and various methods of recirculation. A portion of the pumping cost advantage of shallow filters over deep filters was lost, however, by increasing the number of passes through the filter.

Many early difficulties with ponding and clogging of deep filters were probably attributable to lack of care in the selection, preparation and placement of filter stone. The Committee of the Sanitary Engineering Division on Filtering Materials for Water and Sewage Works of the American Society of Civil Engineers in 1937(10) cautioned that "Too much emphasis cannot be placed upon the necessity of excluding fines and in so placing the material that the sizes will not be segregated or the underdrains obstructed".

Escritt(11) in 1953 reiterated Buswell's(6) opinion that an optimum filter depth exists; however, Escritt attributed the existance of an optimum depth to his observation that for a given configuration a maximum retention of liquid occurred at the optimum application rate. He reasoned and substantiated with data that under the condition of maximum retention, maximum work rate was attained. He suggested that recirculation, alternating double filtration and construction of deep filters have accomplished this end without conscious effort to achieve conditions favoring maximum retention.

Numerous advances have been made in the design and operation of the basic trickling filter process over the years since its inception. Prefabrication underdrains were developed, methods of liquid distribution improved, and the high rate trickling filter systems with various recirculation schemes were developed.

Many attempts at improvement of the basic component of the system--the media--have met with little practical success. Stone continues to be the most common material used despite problems attributable to its physical properties. These deficiencies include:

1. Limited ventilation caused by a low percentage of relatively poorly

connected voids.

2. Surfaces at media contact positions which are not available for slime

adherence.

3. Plugging, ponding and channeling caused by

Materials tried as trickling filter media with limited success included: corn cobs, wood chips and slabs, coke, cinders, broken tile, broken brick, etc.

In attempting to improve treatment efficiency and eliminate plugging problems associated with the relatively uncontrolled size, shape and gradation of crushed rock or slag media, prefabricated media of uniform size and shape have been developed. Noteable among these was the prefabricated tile media developed by A.A. Page of the Red Wing Sewer Pipe Corporation in 1933, trade-named Aero-Block, providing 20 to 22 square feet of surface area per cubic foot of volume and a porosity of approximately 53%. The physical structure consists of on inch round holes extending from top to bottom of a block separated by 1/4 inch walls. According to Leibee(12) trickling filter installations utilizing Aero-Block media in the Minnesota and northern Wisconsin area have functioned successfully for many years.

Levine et.al.(13) reported on trickling filters packed with Raschig rings of various sizes and a prefabricated media called Straight's block. These materials worked satisfactorily as filter media but showed little improvement in performance over rock or slag media although all had the advantage of uniformity.

Ideally, a trickling filter packing would have the following properties:

1. High surface area to volume ratio for effective oxygen transfer from air to

2. Low unit weight for structural simplicity and economy.

3. Low resistance to simultaneous passage of liquid and air.

4. Large clear openings to minimize clogging.

5. Uniformity

6. Availability

7. Durability

8. Low cost
9. Design flexibility to conform to space availability and required degree of

In order to gain ultimate acceptance, the packing would be required to compete on an overall cost basis with media which do not completely meet these criteria but which may be available at lower unit cost.

The trend in more widespread use of plastics for industrial applications logically led to their consideration as materials from which a closer approach to the ideal packing could be fabricated.

One such material, Dowpac (Registered trade mark of The Dow Chemical Company), a product of The Dow Chemical Company has been under development since 1953. Table 1 lists properties of Dowpac and several conventional and synthetic filter media.

Dowpac consists of a series of corrugated sheets of thermoplastic bonded together to form a structural packing module. The sheets of Dowpac are shipped nested and then expanded to their packed configuration at the job site. (See Figure 1). The sheets can be formed in a continuous extrusion-thermal forming operation from any thermoplastic having suitable chemical resistance and structural properties.

The ribs on the sheets of Dowpac space the sheets uniformly one inch apart when assembled thus freeing 94% of the packed volume for the build-up of biological slimes and the unrestricted flow of air and liquid waste. The honeycomb formed at the top and bottom of each assembled module of the expanded packing is believed to facilitate uniform distribution throughout the packing so that all surfaces become wetted and support bacterial slimes. The available surface area does not appear to decrease appreciably as slimes build up because nearly all of the packing surfaces are widely spaced and parallel with little possibility for bridging or plugging with excessive slimes. The gradual corrugations prevent free-fall of the liquid waste through the packing but allow the packing to be self-cleaning. Approximately 25 square feet of surface area is provided in each cubic foot of packed volume. Dowpac is commercially available made from two thermoplastic materials: high impact polystyrene (Dowpac 10) and saran, a copolymer of vinyl and vinylidene chloride (Dowpac 20).

Bryan (13) in 1955 presented data obtained in an evaluation of the comparative effectiveness of Dowpac 10 (formerly Dowpac HCS) and conventional trickling filter stone (blast furnace slag).

Dowpac was developed by The Dow Chemical Company with the objective of as closely as practicable meeting the requirements of an ideal filter medium.

A number of incidental observations made during this initial evaluation are particularly significant. A tendency for rather prolific slime accumulation seemed to refute the frequently reported need for using media with a rough surface, an advantage frequently claimed for blast furnace slag. The unit was fed only pure phenol which was the only organic carbon source and biologically oxidizable substance present. There was never any evidence of the presence of phenol in the slime or sloughed solids. For phenolic substrates then, the theory regarding biological adsorption, storage and subsequent oxidation appeared invalid. A materials balance indicated approximately one-half of the phenol fed was fixed into bacterial protoplasm. Cursory bacteriological examination indicated a mixed nondescript bacterial culture predominantly Pseudomonas and Achromobacter species.

Figure 2 indicates the comparative performance with no recirculation. The slag packed portion gave a constant phenol removal efficiency of 62% up to a loading of 85 lbs. of oxygen demand per 1,000 cubic feet of packing per day. The section packed with Dowpac 10 gave a constant removal efficiency of 82% up to the same organic loading. Influent phenol concentrations at this loading were approximately 180 ppm. The phenol oxidation efficiency dropped off above loading of 85 lbs. Of oxygen demand per 1,000 cubic feet per day; however, both the slag packed section and the section containing Dowpac 10 removed increasing amounts of phenol as the loading increased.

Figure 3 illustrates the comparative effects of recirculation. At all application and recirculation combinations the Dowpac performed consistently better than the slag packed portion.

As a result of the interest generated in the plastic packing by the report of this study and subsequent publication, The Dow Chemical Company initiated a pilot plant program to field test the packing. Approximately twenty industrial and municipal pilot plant studies were run on various types of waste. These studies were generally made by the industry concerned, with the plastic packing and technical assistance provided by The Dow Chemical Company.

The studies have covered a wide variety of wastes. Those on which extensive data were obtained included:

• Domestic sewage
• General organic chemical plant wastes
• Domestic sewage with food processing waste
• By-product coke plant phenolic wastes
• Kraft process paper mill waste
• Ragmill waste
• Limited amounts of data and experience have been obtained on:
• Fiberboard reprocessing wastes
• Cannery wastes
• Dairy waste
• De-inking and white water wastes
• Domestic sewage and fruit canning waste
• Cellulose waste (similar to bleached Kraft waste)
• Chlorinated phenol waste
• Vinyl cyanide waste
• Glycol waste

The pilot plants were generally of the same 3 ft. by 3 ft. square configuration and varied in packed depth from 21 to 42 ft. The sidewalls were invariably enclosed to prevent excessive cooling. Distribution of the liquid wastes were accomplished by means of a 3ft. diameter rotary distributor or a perforated pan distributor.

Natural draft aeration was sufficient to maintain aerobic conditions in the pilot tower at all times, even at packed depths as great as 42 feet. Observations of natural draft were made in a 42 ft. deep tower in one instance to check on the presence or absence of air flow. Figure 4 shows the relationship between the volume of air flow versus the temperature differential between influent air and the waste water. Sewage flow to the tower during this study ranged from 0.3 to 2.0 gpm/sq.ft. or from 430 to 2,880 gallons per day per sq. ft. During the period of relative stagnation, aerobic conditions apparently still prevailed in the tower.

Short term observations of liquid contact time on unslimed Dowpac have also been made. Hydraulic load cells calibrated to read directly in pounds were placed beneath a 10-1/2 foot deep bottom section in a 21 ft. Dowpac 20 pilot tower. After taring the weight of the packing, direct readings of retained waste water plus packing weight were obtained at various hydraulic loading rates. These data are shown on Figure 5 where calculated contact times in seconds are plotted against hydraulic loadings. Run No. 2 was made one day later than Run No. 1 which had been made with clean, unslimed packing. It is believed Run No. 2 is the more correct since it is likely that complete wetting of the packing was not achieved during Run No. 1 because of the initial hydrophobic properties of the clean plastic.

Dowpac is being utilized in full scale treatment installations to take advantage of its properties which are unique for a filter media.

Table II summarizes operating data from daily spot samples obtained furing operation from December 1957 to September 1959.

Raw waste flows from the sour water stripper average between 80 and 100 gpm. The total hydraulic loading to the Dowpac tower varied between 120 and 160 gpm or approximately 300 to 400 gpd per square foot (within the normal range for high rate trickling filters). The unit was originally designed for a hydraulic application rate of 400 gpm or 1000 gpd/sq. ft. but recirculation has been purposely reduced. Anderegg states, "Maximum phenol removal across the trickling filter is not desired because of the need to have some organic feed for the activated sludge unit." He believes if design hydraulic loads were applied to the unit, the phenol poundage removals and percentage removals would be increased because the additional recycle would improve waste-organism contact and smooth out and dilute radical changes in waste strength and pH.

Removal of phenolics from April 1957 through September 1959 averaged 74%. Phenolic concentrations in the feed during this period averaged 228 ppm with extremes of 2 ppm and 2200 ppm. The average of the monthly median of H₂S concentration was 23 ppm with extremes of 0 and 720 ppm. The average daily temperature variation of the mixed feed to the filter unit was 8.5^oF. with feed temperature variation as great as 52^oF. occurring in one day. The average temperature of the raw feed and recycle is approximately 130^oF.

Frame(17) described the modern waste treatment plant at the Cities Service Oil Company Trafalgar Refinery near Oakville, Ontario. After conventional treatment of oily waste waters in API separators, neutralization, chemical coagulation and sedimentation, the waste waters are then pumped to primary and secondary aeration tanks. The mixed liquors from these aeration tanks pass directly to two 21 foot deep Dowpac 20 oxidation towers operated in parallel. Although operating data are very limited, these towers apparently perform a polishing operation and smooth out the peaks of organic loading when the waste stream becomes strong or the bacterial culture in the aeration tanks become upset. The trickling filter units provide a stable source of seed if the aeration tank cultures become upset. Hydraulic application rates to the towers are approximately 730 U.S. gallons per day per sq. ft. including recycle which varies inversely with raw flow. Subsequent treatment includes sedimentation, ozonation, application of activated carbon, and rapid sand filtration.

Reynolds and Mauterer(18) reported briefly on the combined cooling tower - biological oxidation unit at the A. E. Staley Manufacturing Company which has been in operation for two years. This installation consists of 9,000 cubic feet of Dowpac 10 placed in one bay of a standard redwood induced draft cooling tower. This unit performs a dual function of cooling 6 MGD of process waters and removing organic contaminants from vegetable oil refining. Some plugging of the Dowpac was experienced due to a combination of high organic content (up to 5,000 ppm total solids) and a pH as low as 4. These two conditions favored the development of a very prolific growth of filamentous fungi which would not slough off the packing as did the bacterial slimes adhering to the packing in the pilot study previously made. Automatic pH control of the cooling waters, together with mechanical flushing and caustic treatment has eliminated this plugging problem. The cooling performance in the Dowpac packed section is on a par with the conventional wood slat packed sections:

Air rate: 1155 lbs./hr./sq. ft.

Water rate: 1815 lbs./hr./sq. ft.

Inlet water temp.: 110^oF.

Outlet water temp.: 80^oF.

Data on wet bulb temperatures are not available; however, the above listed average data are for the months of August, September, and October. B.O.D. removals in the tower overall are approximately 1800 lbs. per day.

The performance of Dowpac in full scale trickling filter installations and pilot studies indicates that prefabricated plastic surface media for biological oxidation have practical application in waste treatment. As development work continues on improved shape modifications and manufacturing technique, it is believed biological treatment utilizing plastic media will take an increasingly prominent position as a practical tool for sanitary engineers in water pollution control.

 REFERENCES

1. Zobell, C. E., "The Influence of Solid Surface Upon the Physiological Activities of Bacterial in Seawater", J. Bact. 33:86 (1937)
2. Velz, C. J., "A Basic Law For the Performance of Biological Filter, Sewage Works Journal, 20:607-617 (1948)
3. Phelps, E. B., "Stream Sanitation", John Wiley and Sons, Inc., New York, New York (1944)
4. Howland, W. E., "Flow Over Porous Media As In A Trickling Filter", Proc. 12 Industrial Wastes Conference, Purdue Univ., 94: 435 - (1958)
5. Bloodgood, D. E., Teletzke, G. H., and Pohland, F. G., "Fundamental Hydraulic Principles of Trickling Filers", Sewage and Industrial Wastes, 31: 243-253 (1959)
6. Buswell, A. M., Strickhouser, et. al. Illinois Water Survey Bulletin 26 (1928) Abs. Sewage Works Journal 1:511-512 (1928)
7. Clark, H.W., "Past and Present Developments in Sewage Disposal and Purification", Sewage Works Journal, 2:561-571 (1930)
8. Upper Mississippi River Board of Public Health Engineers and Great Lakes Board of Public Health Engineers "Standards for Sewage Works" (1952)
9. Ingram, W. T., "A New Approach to Trickling Filter Design " Proc. Am. Soc.

Civil Engr. Paper No. 999, Vol. 82 (1956)

10. Committee of the Sanitary Engineering Division, ASCE Manual No. 13 "Filtering Materials for Sewage Treatment Plants" (1937)
11. Escritt, L. B., "Depth: A Possible Major Factor in Percolating Filter Design" Surveyor, 111:143 (1952)
12. Aero-Block, Bulletin 135, Hugh C. Leibee, Consultant, Minneapolis, Minnesota
13. Levine, et. al. "Observations on Ceramic Filter Media and High Rates of Filtration" Sewage Works Journal, 8, No. 5, 701 (1936)
14. Bryan, E. H., "Molded Polystyrene Media for Trickling Filters", Proc. Purdue Industrial Waste Conference (1955)
15. Report of Sub-Committee on Sewage Treatment in Military Installation of the Committee on Sanitary Engineering National Research Council, Sewage Works Journal, 18:791 (1946)
16. Anderegg, Fred C., "Biological Disposal of Refinery Wastes" Proceedings, 14^th Purdue Industrial Waste Conference (1959)
17. Frame, John D. "Refining" Refinery Wastes", Wastes Engineering, 30, 380 (1959)
18. Reynolds, J. C. and Mauterer, R. J. , "Staley Saves Water, Cuts Process Waste", Power Engineering, (February 1959)