Commercial Application of Air Phase Biofilters

ABSTRACT

Biofilter design criteria and performance data have been collected over the past 5 1/2 years from pilot unit research on actual commercial exhaust and industrial biofilter installations on air flows ranging from 600 cfm to 140,000 cfm. Reaction and diffusion kinetics have been analyzed under actual operating conditions for a host of chemical compounds including alcohols, ketones, alphatic, and aromatic hydrocarbons. The biofilter system design can create economic advantages or disadvantages and must encompass the permitting requirements of the regulatory authorities.

Specific data may be protected by confidentiality agreement. However, the discussions outlines below are derived from PPC Biofilter pilot unit research or commercial installations and operating experience with companies such as Weyerhaeuser, Amoco, Union Camp, Dow, General Tire in the following industry groups:

Introduction:

The two primary decision parameters used to evaluate biofiltration for commercial applications are 1) economics and 2) regulatory permit requirements. Both parameters can fluctuate independently based on the gas streams to be treated in regards to chemical composition, concentration flow rate, temperatures and relative humidity.

The capital cost of a biofiltration system has a direct relation to the size of the reactor, which is dependent on flow rate, chemical composition, and concentration.

Pre-treatment equipment requirements for humidification and temperature control are a direct function of the flow rate, moisture percent, dry bulb operating temperatures and particulate loading. These parameters can effect the quantity of water discharged and, coupled with the organic loading of the biofilter, the ability to discharge water from the system to the local POTW.

The biofiltration system's ability to consistently demonstrate removal efficiencies in compliance with a regulatory permit will be a direct function of concentration, variable loading from batch operation, and fluctuating chemical composition.

The operating cost of a biofiltration system can include electricity consumption, natural gas consumption, steam, maintenance cost, filter media replacement cost, along with water consumption and media disposal cost. These costs are directly related to the design and size of the biofilter.

These and other issues are discussed below using actual industry examples.

DISCUSSION

1. PRE-TREATMENT - TEMPERATURE CONTROL AND HUMIDIFICATION

The air phase biofilters in operation today, predominantly operate in a mesophilic bacteria temperature range of 60F to 105F. Most commercial gas streams do not automatically fit this requirement and must be conditioned. Furthermore, temperatures can affect the removal efficiency depending on the compound to be degraded.

The removal efficiency of the reactor has a direct relationship to the bacteria count within the media. The bacteria count has direct relationship to moisture content of the support media. To prevent wide fluctuations in the moisture content of the support media and therefore the removal efficiencies, it is an established practice to saturate the gas stream before entering the reactor.

Reviewing both of the parameters before implementing a pilot unit study can save considerable time and energy.

A. In the wood products industry, controlling plywood veneer dryer emissions is required in certain regions. In this example, the total air flow can exceed 250.000 acfm at 325F. The 325 degree does not rule out biofiltration. However, the exhaust is approximately 31% moisture, creating a saturation temperature of 160F. Adiabatic cooling from 325 to 160 degrees can be accomplished cheaply and efficiently by evaporating water from the flue gas. The saturated gas stream must be cooled from 160 degrees to 100 degrees for mesophilic bacteria operation. Force cooling can be engineered but creates a major problem for the wood industry.

Force cooling requires a direct contact cooler, heat exchanger, and cooling tower. Total cooling requirement on a 100,000 acfm gas stream, at the above conditions, is roughly 80 MMBtu. While the capital and operating cost for this equipment skew the economic benefits of biofiltration towards incineration it is not the overriding decision to eliminate biofiltration technology.

Force cooling would generate over 130 gallons of condensate water per minute. The condensate will have VOCs in the water phase. Most wood industry plants are "zero discharge" facilities. Engineering the system to adapt to biofiltration has taken an air pollution problem and created water pollution problem with no economic way to deal with roughly 200.000 gallons of discharge water per day.

Force cooling the gas stream lowers the temperature below the melting point of some of the condensable compounds previously in the vapor state such as pine tars, linoleic acid and hydrocarbon wax. These condensed tars and acids can foul the heat exchange system and clog the biofilter media.

In summary, detailed review of the gas stream specifications in advance, eliminated any economic advantages biofiltration held over regenerative thermal oxidation.

B. In a similar case, the corn processing industry injects SO₂ into the corn hull to loosen the gluten and the germ. Gluten dryers and germ dryers represent approximately 14,000 acfm at 220⁰F. The saturation temperature after adiabatic cooling was 115 degrees F. Emissions to be controlled included methanol, ethanol, and SO₂.

While the SO₂ was in a gaseous state it is highly soluble and it's acidic nature can affect the pH of the filter media. Therefore, a caustic scrubber was used as pre-treatment. Using the same scrubber as a direct contact cooler, the recirculation water passes through a plate and frame heat exchanger, which is cooled by plant cooling water. This equipment force cools the gas stream to the required mesophilic temperature of 100 degrees F. This force cooling does generate approximately 17 gpm of condensate which has ethanol and methanol concentrations. Applying the Henry's gas law constants at the operating temperature provides a good estimate of the methanol and ethanol concentrations in the water phase.

The difference in this case, as opposed to the wood industry dryers, was the ability to dispose of condensate water through a waste treatment system and discharge to the POTW. The biofilter system demonstrates greater than 95% removal of SO₂, methanol and ethanol.

In this case, biofiltration provided a superior technology as scrubbing required a 70 gpm one time pass through, and there was minimal fuel value for incineration.

Both of these cases are representative of the condensate issues associated with dryer exhausts. A thorough understanding of the physical state of the dryer exhaust allows for the evaluation of condensate issues that must be addressed for a specific facility.

C. Meeting the design saturation requirement can also cool the gas stream below a design standard for mesophilic operation. On wood industry press vents the exhaust temperature is approximately 120-140 degrees F. However, the plant draws in considerable amounts of ambient air and therefore the wet bulb temperature after humidification can vary seasonally from 60 degrees F to 100 degrees F, depending upon ambient temperatures and humidity. The metabolic rate of mesophilic bacteria can double every 12 degrees F. We have noticed the removal efficiencies can vary on these systems depending on ambient conditions. Removal efficiencies of greater than 98% removal are documented at biofilter temperatures of 95 degrees and dropping to 70% removal at temperatures below 65 degrees F.

In one commercial application, excess plant steam is injected into the inlet duct work during the winter time to raise the wet bulb temperature of the gas stream and maintain higher removal efficiencies. In a second case, natural gas burners are employed to maintain optimum operating temperatures.

D. Research provided by Weyerhaeuser ES&T demonstrates that part of the temperature/removal efficiency relationship is compound specific. This has also been demonstrated with compounds such as styrene and other aromatic hydrocarbons.

In the printing industry, pilot unit research was conducted on a gas stream with 22 separate compounds. The pilot unit research resulted in a commercial application to process 45,000 cfm. In this case, lowering the operating temperature from 80 degrees to 70 degrees showed no impact on the removal efficiency. It is not clear whether an increase in removal efficiency would result when the temperature rises from 80 degrees to 95 degrees.

Temperature effects in a biofilter are complex phenomena because of the number of different physical properties which come into play. While the temperature effects on the physical properties inside the biofilm will tend to enhance performance (solubility, molecular diffusitivity, cell metabolism), temperatures adversely affects the physical properties in the gas phase and at the boundary between the biofilm (vapor pressure, Henry's constants). Determination of the effect of temperature will depend on the system and is beyond the scope of this paper.

Summary

A careful review of the gas stream specifications for the exhaust to be treated is recommended prior to biofiltration project. The pertinent data to collect includes ACFM, dry bulb temperature, moisture % or relative humidity, front and back half particulate loading, along with a list of all the chemical compounds found and their respective concentrations.

2. CHEMICAL COMPOUNDS AND CONCENTRATION.

A great deal of bench scale pilot research has been accomplished to develop the elimination capacities for individual compounds. However, most commercial exhausts contain mixtures of compounds which vary on an hourly and daily basis with respect to each compound and its respective concentration.

A. A 3M process uses ethyl acetate, toluene, acetone and MEK on Monday and Wednesday. On Thursday and Friday methanol and ethanol appear in the production process. Concentration of each individual compound can vary from 10 ppm to 60 ppm.

In this examples, preferential degradation of the alcohols was observed. On Monday through Wednesday, the total removal efficiency would stabilize at 94% and acetone removal was relatively consistent at 94%. On Thursday and Friday when ethanol and methanol appeared in the production process, the acetone removal efficiency declined to 75% while the alcohols experienced 96% removal and the total removal declined to 92%.

This six month research project demonstrated the ability of biofiltration to obtain consistent removal efficiencies of total hydrocarbons between 88-94% on a gas stream which is variable by chemical composition and concentration. However, many regulatory permits would require consistent 95% removal or even 99% removal (under MACT).

Biofiltration, in this examples represents a superior technology with lower capital cost, and operating cost when compared to incineration or activated carbon alternatives. The annual energy savings on this application of biofiltration compared to incineration amount to $250.000/year. The biofilter does not produce NO_x emissions, factor in fuel oil or natural gas consumption and secondary NO_x production from electricity generation and it appears that the regulatory agencies have not seen the total picture. Couple the energy savings and the reduction of secondary contaminants with the intrinsic safety of a biofilter, and the economic advantages of a biofilter become much more apparent.

However, biofiltration on this type of fluctuating gas stream delivers a fluctuating performance between 88-96% removal. Therefore, similar applications will need to insure that their permit allows fluctuation.

B. Serigraph is a silk screen printer which produces printed dashboard components such as speedometers and air conditioning controls for the automotive industry. New plant construction in Wisconsin required a reduction in the total emissions from the existing facilities. The solvents used in their process include over 30 separate compounds of which 5-8% are alcohols, 45-55% are esters or ethers, 13-15% are ketones, 8-11% are aliphatic and 12-15% are aromatics.

This smorgasbord of chemical compounds changes in total hydrocarbon concentration on an hourly and daily basis. Serigraph and PPC combined efforts to research the application of biofiltration to a potential 45,000 acfm gas stream. A commercial pilot unit was installed on an existing dryer to determine the removal efficiency at varying concentrations for a four month period.

The pilot research demonstrated removal efficiencies of 95% at concentrations of 70-80 ppm. The research also showed removal efficiency dropped to 75% at concentrations of 50 ppm and less, operating at the same retention time. This phenomenon is consistent with past research which shows that mass transfer limitations control the biofilter at low concentrations. While the outlet concentration was very low at low inlet concentration, the percentage removal declines. This can be explained by the physical properties of the VOCs in the gas stream. For volatile, slightly soluble hydrocarbons, the low outlet concentration may be approaching thermodynamic limits which would basically fix the outlet concentration, regardless of the inlet concentration. For compounds that have a higher affinity for water in an air water system, such as alcohols, this does not necessarily apply.

D. Bush Boake Allen (BBA) is a flavor manufacturer located in the heart of Chicago. As their production level increased so did the odor complaints from surrounding neighborhoods. While they were not "permit restricted" based on pounds of annual emissions, they were restricted as many U.S. organizations are, under the rule of "no odors past the property line." In 1994 BBA experienced 13 odor complaints and began researching odor control alternatives. The odors from strawberry, grape, lemon, butter etc. flavor production are created from volatile organic compounds such as esters, ethers, and anthranilates.

The odors in this application are created from volatile organic compounds. Each flavor represents specific organic compounds. If production of a certain flavor such as "strawberry" occurs only once every two weeks, then the bacteria in the biofilter are not highly adapted to the organic compounds and concentrations found in strawberry production. Therefore, during the initial production of this flavor, the removal efficiency may slightly decline until the bacteria adapt to that condition. Most of the organic compounds in this application are similar in chemical groups and therefore the same bacteria can provide high oxidation efficiencies between different flavor productions.

Removal efficiencies in these types of applications can decline when a certain flavor has different chemical structure and an high concentration. However, after one year of operation the odor complaints dropped from 13 to 1. The one complaint could not be traced back to production. Odor test using dynamic olfactometry provided the following results:

Summary

As a general rule, the bacteria in a biofilter, will degrade the simple, short chain compounds first and then work in successive order to the more difficult compounds such as C-8 aliphatics and finally the difficult aromatic compounds such as benzene. Usually, the more complex the chemical structure the longer the residence time required for complete oxidation. A thorough review of all the compounds in the gas stream is recommended.

REACTOR DESIGN AND FILTER MEDIA

Biofiltration systems have been constructed in many different designs ranging from bio-trickling filters employing totally inert media, to up-flow open top reactors using media ranging from peat moss to dirt, to totally enclosed insulated reactors using engineered media. Each system design has unique characteristics that affect removal efficiencies and operating cost.

A. The majority of the operating costs associated with a biofilter are from electrical cost for running the fan to move the treated air stream through the biofilter. Fan requirements are determined by the pressure drop through the system. The pressure drop through the system is a function of gas velocity, which is a function of reactor design, and the biofilter media used. Pressure drop through a biofilter can be modeled using the Ergun equation:

This equation takes into account fluid properties, physical characteristics of the media, such as particle size and void fraction, and the velocity of the gas through the media. To reduce the pressure drop through the media, it is necessary to reduce the velocity through the biofilter and/or provide a media with a large enough void fraction to allow reasonable flow velocities at a reduced pressure drop.

Biofilters employing "soil" as a filter media have an extremely high pressure drop (15-20" w.c.) and are not economically feasible on large gas streams. For instance, a 100,000 cfm gas stream such as found in the wood industry, paint booths, or boat manufacturers would require a 500 hp fan to overcome the pressure drop through a soil filter. At $.06 per kW/hr this represents an electrical consumption cost of $22.00/hr which deters the economic advantages of biofilters.

Different wood mulches, compost and peat moss have been employed independently or mixed with inert ingredients to reduce the pressure drop through the filter media. However, velocity is a critical design issue as these media tend to be relatively dense and therefore require a substantial foot print. In this 100,000 cfm example, a reactor designed for a media height of 5'. At a velocity of 11 feet per minute the foot print is 9,000 square feet. A dense media such as 100% compost can have a pressure drop in excess of 9" w.c. and an operating cost of $9.75/hr. An engineered media with a dense structure but inert ingredients added to decrease compaction can have a pressure drop of 4" w.c. and an operating cost of $4.33/hr. A media which provides a large void fraction can reduce pressure drop over the media to 0.5" w.c. with a resulting operating cost of $0.54/hr.

Many industrial locations simply do not have the "real estate" for a traditional biofilter and require a smaller footprint. This can be accomplished by stacking the filter media to greater heights (i.e. 15-20 feet). Compaction problems are solved by intermediate bed supports. However, the increase in velocity by using a reactor with a smaller surface area will result in a higher pressure drop over the media and therefore higher electrical consumption. In these situations, a media with a high void fraction is preferred and can potentially offer savings when compared to thermal systems.

Dense media such as compost and peat moss prohibit a stacked bed design as velocities above 10 feet per minute are usually the maximum. Engineered media can traditionally handle velocities of 20 feet per minute. Media which employ large void fractions can obtain velocities of 40 feet per minute at reasonable pressure drops but require increased water circulation for mass transfer of the chemical compound in the gas phase to the biofilm in the water phase.

B. The filter media employed in a biofilter can impact performance. In most viable biofiltration applications, the performance of the system is limited by the mass transfer associated with moving the VOC's from the gas phase into the liquid biofilm surrounding the media particles. Because of this limitation, the selected media and design velocity can have a significant impact on the performance of the system. Intuitively, a filter media with a small particle size and hence a small void fraction would appear to perform better than a media with a large particle size and void fraction. While this may be true at low VOC loadings and modest velocities, a more open media can offer design flexibility, significant operational advantages and may even provide improved improved performance in some cases.

A more open media, with a reduced pressure drop can be more easily "stacked", providing a biofilter with a significantly reduced foot print. This foot print reduction increases the design velocity. The increased velocity will increase pressure drop, but it will also improve the mass transfer in the system. An open media can also offer improved performance resulting from improved moisture control, which is directly related to the biological activity in the bed.

While there is some flexibility regarding the nature of the media used in the biofilter, not anything will work as a biofilter media. Different materials may have inherent problems associated with them that only become apparent after evaluation. Wood based media, for example, an have significant amounts of fungal growth which will significantly increase pressure drop, even to the point of rendering the system inoperable. This type of evidence can only be gathered from actual testing of the media and takes a great deal of time and money to gather.

A pilot study was conducted at whiskey warehousing facility to evaluate biofiltration as means of controlling ethanol emissions. During the course of the study, two separate media,and engineered media and a totally organic media, were tested. Ethanol concentrations ranged from 200 to 700 ppm and gas residence times ranged from 17 to 30 seconds.

It was evident from the data collected that performance depended primarily on the moisture content of the media. The engineered media provide good performance, but at high ethanol loadings, biological activity tended to dry out the media, which caused performance to decline. The "dense" and hydrophobic nature of the engineered media made it difficult to maintain a uniform moisture content in the media and prohibited excessive application of water to the media. Removal efficiencies fluctuated between 88 and 96% removal depending on the moisture content.

After six months of operation, the engineered media was replaced with a proprietary media and operated through the spring an summer. This media was much bulkier than the engineered media and had a significantly larger void fraction. The large void fraction provided better drainage and allowed for more uniformity in moisture throughout the bed. The ability to add relatively large volumes of water to media kept the media from drying out at high organic loadings. This resulted in a more stable operating system with consistent VOC removal efficiencies of greater than 95%.

Summary

Although there are numerous "general" guidelines to follow when designing a biofiltration system, a standard design can not be applied to each specific application. To successfully implement biofiltration as VOC control strategy, a complete, experienced based, engineering evaluation is required. The biofilter should be custom engineered for a particular application, taking care to address all of the issues mentioned above such as temperature and pretreatment/humidification. To effectively address the pertinent design issues for a particular facility, a thorough understanding of the process and the characteristics of the gas stream is essential. This information, coupled with knowledge of the mechanical characteristics of the biofilter media, and an understanding of the biology involved allows the design and implementation of a biofiltration system to successfully meet the control requirements of a particular facility.

Back to Menu