Biofilter as a VOC Control Device for
an Industrial Screen Printing Operation

ABSTRACT

Emissions of Volatile Organic Compounds (VOCs) are generated from inks, thinners, and cleanup solvents used at Serigraph’s four printing facilities in West Bend, Wisconsin. Since southeastern Wisconsin is designated as a severe non-attainment area for Ozone, the construction of a new 166,000 square feet screen printing facility included the installation of a commercial biofilter as a VOC control device to provide for Serigraph’s growth and future expansion.

The main design features of this biofilter are the hybrid filter-media moisture control, two identical but independent filter chambers, and sophisticated control system to protect the biofilter. Also, the system includes an innovative air-cascading system to maximize the off-gas VOC concentration and minimize the energy consumption, while providing the required flexibility for manufacturing, and incorporates EPA’s method 204 for Permanent Total Enclosure (PTE).

The following paper will provide detailed information on the first commercial biofilter application for an industrial screen printer, including regulatory issues, the industry perspective, important features of the system design, and the operational performance data.

INTRODUCTION

Serigraph is a screen printer and sheet fed offset printer located in West Bend, Wisconsin, about 30 miles northwest of Milwaukee. The company produces OEM products for the automotive, appliance, computer, electronics, and recreational industries, as well as displays for the point-of-purchase market. Ninety five percent of its printed substrate is plastic. Serigraph operates five facilities in West Bend, and one in Mexico. The company employs 1,200 people in its West Bend facilities.

West Bend is part of the greater Milwaukee metropolitan area. The region is classified as a severe non-attainment area for ozone. Companies located in the area are required to limit their growth to a net emissions increase of 25 tons of VOCs in any five consecutive year period, or subjected to New Source Review (NSR) process. The NSR requirements for a severe non-attainment area are the implementation of Lowest Achievable Emission Rate (LAER) technology and the emission offsetting reduction ratio of 1.3. The alternative is to achieve reduction in emissions prior to be subjected to the NSR requirements. This scenario is known as netting-out.

Serigraph was successful in utilizing the netting-out scenario by reducing its VOC emissions for several years through material substitution and implementation of better management practices. However, an average annual growth of ten percent required additional measures to reduce VOC emissions, and then, it felt that those measures included VOC control technology. The Air Pollution Control (APC) permit requirements, for Serigraph, are based on total tons per year of VOCs emission, and not on a percentage removal basis. Serigraph, therefore, had the opportunity to establish the amount of reduction in its VOC emissions for its growth and future expansion. Thus, in order to achieve desired reductions and to gain a reasonable payback on investment, it was calculated that the chosen APC technology should achieve a minimum 80% VOC removal efficiency.

PROCESS AND EMISSION CHARACTERISTICS

In 1995, Serigraph made the decision to build a new screen printing facility to house its existing equipment. The new facility included clean rooms where all the printing operations would take place. The challenge that Serigraph faced was that the VOC emission from screen printing operations is of low concentration, and the concentration varies constantly depending upon the number of printing presses and dryers in operation, and the amount of materials being used. Cleaning operations associated with the printing account for about fifty-percent of the VOC emissions. Of the remaining emissions, emanating from inks during printing, only thirty percent are discharged through stacks. The balance, about seventy-percent, evaporate at the press or prior to the products enter the curing ovens or dryers. Thus, up to eighty-five percent of VOC emissions from screen-printing are fugitive, and therefore considered non-controllable.

The total exhaust air from the new facility was projected to vary between 20,000 and 45,000 cubic feet per minute (CFM), which combines the fugitive exhaust from the clean room, and the exhaust from ten independent gas fired dryers. An average concentration of 0.1g/m3 of total hydrocarbon (THC), as the molecular weight of the compound, was projected for the new facility. The curing temperatures for plastic substrate are between 140 F - 180 F, with little moisture content.

Over twenty different VOCs are found in inks and solvents used at Serigraph. These VOCs include 5-8% alcohols, 10-15% ethers, 35-40% esters, 13-15% ketones, 8-11% aliphatic and 12-15% aromatic hydrocarbons. The solvents include ethanol, methanol, isopropanol, MEK, butyl cello-solve, isophorone, butyrolactone, ethyl acetate, C6-C7 alkanes, EB acetate, PM acetate, 1,2-dimethoxy 2-propanol acetate, 2-ethoxy ethanol acetate, toluene, glycol ethers, hydrofuranone, di-acetone alcohol, cyclohexanone, pentanone, benzene, etc.

ECONOMIC ANALYSIS OF VOC CONTROL TECHNOLOGIES

Serigraph reviewed all the potential VOC emission control technologies for their applicability, performance, capital cost, and operating costs. Following is a summary of the findings.

Scrubbing efficiencies were difficult to obtain with the low solubility organic compounds. An activated carbon system for 45,000 CFM presented excessive carbon replacement and regeneration cost. Employing the use of adsorption technology appeared to be equally undesirable for couple of reasons. First, the system would be relatively large for 45,000 CFM with a low VOC concentration of about 40 PPM(volume). Second, it would need to convert up to 99 tons of VOCs annually, with a varying exhaust flow-rate and concentration. Other traditional means of destroying VOC emissions include the use of thermal oxidizers. The low concentration of VOCs, combined with the low curing temperatures of the exhaust added a very little fuel value to decrease the fuel consumption required for incineration. Oxidizers with concentrators were high in capitol cost. While without concentrators, the operating cost would be exorbitant. Thus, economically, thermal treatment options were not feasible for Serigraph’s application. The ideal device would be required to accept a high volume of air, low VOC concentrations, saturation temperatures between 70 F– 90 F, minimal operating cost, and achieve a minimum of 80% removal efficiency. Biofiltration seemed to be the answer. Table 1 shows the cost comparison between a few VOC control technologies for Serigraph’s application.

RESEARCH AND PILOT UNIT STUDIES

During 1994, from July through November, Serigraph conducted initial pilot study to determine the viability of the technology for its specific solvents. Though this pilot unit did not have any sophisticated controls for weather conditioning or system maintenance, average removal efficiency of 72% was achieved. In January 1996, encouraged by the result of the initial pilot test, Serigraph contracted PPC Biofilter to provide a commercial biofiltration system to control the VOC emissions from their new printing facility.

Prior to installing the commercial system, PPC Biofilter installed a pilot unit at Serigraph. The purpose of the pilot study was to confirm the theoretical size for the commercial system, operating protocols, and to insure that the desired control efficiencies would be obtained. It was first connected to an exhaust from a gas-fired dryer with sheet-fed presses, and then to an exhaust from an electric dryer with a web press. The total duration of the study was from March through August of 1996.

The pilot unit included a pre-treatment humidifier with the capability to control the inlet air temperature and flow rate. Throughout the course of the study, the pilot unit was operated at a flow rate ranging from 100 to 170 ACFM, representing a retention time of 18-32 seconds, and a temperature range between 40 to 95 degree F. Figure 1 shows the temperature and flow rates during the operation of the pilot unit. The pilot unit was controlled with a PLC unit to monitor and data log the necessary operating parameters in real time. Moisture content of the filter media was monitored by a load cell, and maintained by the control logic. The filter media in the reactor represented a cross-sectional flow surface area of 10.76 square feet with a depth of 5 feet.

The pilot unit performed to the expectations, providing a stable operation at different flow rates. Average efficiency removal of more than 80% was achieved for gas fired dryer’s exhaust with moderate concentration levels. Inlet and outlet VOC concentrations were monitored with a portable Flame Ionization Detector (FID). Figure 2 depicts the VOC loading and the removal efficiency of the pilot unit during the course of the study.

Validity of the FID Data: Although, the flame ionization detectors (FIDs) are an acceptable and, most times, required to quantify and/or to analyze total hydrocarbons for monitoring the performance of VOC control equipment, there are shortcomings with these devices. Particularly, when monitoring the biofiltration system where inlet stream contains mixed organic compounds. It is important to be aware of FID’s lower response factors for oxygenated hydrocarbons and relatively higher response factors for non-oxygenated ones. Accuracy of the FID data, therefore, raises questions where the inlet air-stream to the biofilter contains oxygenated compounds, such as ethanol, isopropanol, methanol, di-acetone alcohol, etc., along with alkanes and aromatics, whose FID response factors are different.

Typically, for Serigraph’s exhaust, the FID understates the inlet VOC concentration because of the presence of large portion of oxygenated compounds. Also, the FID overstates the exhaust concentration since the oxygenated compounds are typically removed in the biofilter and do not exist in the exhaust air stream. Therefore, the FID readings show an overall decreased removal efficiency compared to the actual removal efficiency. The only way to qualify this result is through GC/MS analysis, which becomes very expensive for more than 20 different organic compounds.

It is of great importance to understand the source of the VOC emissions for monitoring a biofilter with a FID. In case of Serigraph, gas fired dryers are used where combustion is not complete, and the incomplete combustion will result in the emissions of methane into the dryers’ exhaust gas stream. This methane will not be abated in the biofilter, and hence, the appropriate adjustments should be made for the FID readings to account for the methane. Repeated analysis and laboratory tests suggest that, on an average, more than two thirds of the outlet concentration from Serigraph’s biofilter is methane. There is, therefore, a substantial difference between the VOC removal efficiency of the biofilter with and without adjusting for methane. Also, at Serigraph, the methane concentration in the exhaust air stream can vary substantially depending on the number of dryers operating.

In summary, the removal efficiencies which are unadjusted for inlet methane concentrations, shown in Figure 2, demonstrate the ability of the biofilter to achieve much greater than 80% VOC removal, adjusted for methane, at a media retention time of about 30 seconds.

Theoretical Modeling of the Biofilter: A thorough discussion of the theoretical model applied to biofiltration is beyond the scope of this paper. However, in general, there are two distinct operating ranges for a biofiltration system. The first range is usually referred to as diffusion limited, where the system exhibits first order kinetics, and the mass transfer drives the filter size. The second scenario is called reaction limited, where the system exhibits zero order kinetics, and the system biology is the limiting factor. In practice, most biofiltration applications fall within the diffusion limited regime, which is the case for Serigraph.

The standard model applied to biofiltration proposed by Ottengraf, makes the assumption of zero order kinetics, with diffusion limitation up to a certain boundary concentration, which is compound specific. In reality, the system exhibits Monod type kinetics, which becomes apparent at low concentrations. In the diffusion-limited regime, this model shows the elimination capacity to be linear or first order with respect to the contaminant concentration in the gas phase. And, in the reaction-limited regime, the elimination capacity is independent of the contaminant concentration in the gas phase or of zero order.

Figure 3 shows a plot of the elimination capacity of the system against the square root of the inlet concentration. From this, the slope of the best-fit line is proportional to the fundamental sizing parameter for a diffusion-limited biofilter. It appears from the collected data that, at very low organic concentrations, the assumption of zero order kinetics breaks down as expected. This can be seen from the change in the slope of the characteristic graph shown in Figure 3. Typically, this type of theoretical modeling is applied to analyze the elimination capacity for a single compound. In the case of Serigraph where more than 20 different solvents exist with varying solubilities, Henry’s gas law constants, vapor pressures, etc., the model becomes quite complex. Biologically, it is assumed that the simpler chemical structures, such as, ethanol, IPA, methanol, etc. are degraded first, and then the bacteria will work in successive order, degrading the most difficult compound at the end. In effect, all the compounds are treated as one imaginary compound.

While the data presented here is specifically applicable to Serigraph’s emissions, similar trends should be expected where several different gas streams exist or a single gas stream with mixed chemical composition. The fundamental difference would be in defining the "very low" concentration range for specific gas streams with mixed chemical components. It is very important to note that there is a possibility of reduced efficiencies at "very low" concentration range. In general, if a system achieves 95% removal efficiency at 110 ppm (below the boundary concentration), one may not expect the system to achieve similar removal efficiencies at 10 ppm Because, for the biofilter, at a given residence time, depending on the physical properties of the VOCs of interest, there are limitations on the lowest achievable effluent VOC concentration.

The above modeling was studied in detail, and it played an important role in determining whether the technology could maintain the emissions below the required levels on a consistent and continuous basis. However, during the final analysis, the retention time or the filter volume, and other operating variables, such as, temperature, moisture control, etc., were thoroughly reviewed to design a reliable system to achieve the desired removal efficiency.

Commercial Biofiltration System

The nature of Serigraph’s manufacturing process causes the individual organic components, the combined VOC concentrations, operating temperatures, etc. to vary dramatically. A single dryer may be operating in a given period of time, and in the next hour several dryers may be in production. Exhaust airflow to the biofilter, therefore, could range between 20,000 CFM to 65,000 CFM, with inlet dry bulb temperature before the primary humidification ranging between 70 and 180 degrees F.

Air Cascading & Collection System: To control the above mentioned varying airflow, VOC concentration, and temperature, Serigraph engineered a sophisticated air collection and cascading system. Figure 4 shows the schematic of the general layout of the air collection system, including the controls and their operating ranges. During normal operation, the return air from the clean room is directed to a common supply duct as a major portion of the supplied air to the dryers. When all the dryers are operating, the amount of return air from the clean room may not be sufficient, and hence, fresh air is added through a set of outside air intake ducts with modulating dampers (AD-1) to make up for the total supply air demand. Combusted hot air in the dryers is recycled twice before being exhausted out, and collected into a common exhaust duct. During weekends, return air from the clean room is short-circuited to the common exhaust duct through a bypass damper (AD-2) to maintain the clean room’s indoor air quality, and to insure that the biofilter receives oxygen/air on a continuous basis. The air handling system is operated by an independent PLC to work in conjunction with four blowers; two each located at the inlet and at the outlet of the biofilter. Variable frequency drives are employed on all four blowers to maintain the room pressure, and to balance the varying airflow from various dryers.

This system has aided in maintaining a more consistent VOC concentration in the gas stream, has reduced the overall exhaust to below 40,000 CFM, has increased the VOC concentration by more than 25%, and has maintained the inlet temperature within the operating limitations of the biofilter. Additional temperature control is provided by utilizing the dryers, which may or may not be in actual production.

Above-mentioned air collection system is also designed to achieve Permanent Total Enclosure (PTE) in accordance with EPA’s Method 204. The clean room is set to maintain -0.004" WC with respect to the outside manufacturing area. When all the dryers are running, bypass damper (AD-2) is closed, forcing the return air from the clean room to go through the dryers, and exhausted into the common exhaust duct. Modulating dampers in the outside air intake ducts (AD-1) will open only if the clean room pressure goes to a negative to the set point, and will make up for the additional supply air to the dryers. Inlet fans to the biofilter are always set to achieve the required negative pressure in the clean room. When no dryer is running, return air is short-circuited from the common supply duct to the common exhaust duct through a bypass damper (AD-2), and maintains the clean room at required negative pressure. Also, the bypass damper AD-2 will open if the room pressure is positive to the set point, allowing the air in the clean room to be pulled out by the inlet fans, and thereby forcing the room to be at desired negative pressure.

The above system, therefore, has helped to achieve several important operational aspects. It has maintained a more consistent VOC concentration and temperature in the combined exhaust gas stream. It has increased the VOC concentration by more than 25%. It has helped in reducing the energy consumption of the dryers by more than 15% in taking the warm room air as a supply air to the burners instead of outside cold air, and helped in keeping the biofilter oxygenated at all times. The most important feature of this system is that, it has reduced the overall size of the control device by about 20,000 CFM and there by has reduced the capitol cost of the control device by more than 20%.

Biofilter Equipment: PPC Biofilter designed, fabricated and installed a state of the art biofiltration system employing a fully automated PLC controls and a computer MMI / data logging. Figure 5 shows the Process and Instrumentation Diagram (PID) for the full-scale biofilter. The biofilter is composed of two independent, totally enclosed, insulated reactors working in parallel. Each biofilter has its own pre-treatment humidifier with an independent recirculation system and its own inlet fan with a variable frequency drive for flow control. A single reactor may be in operation while the second reactor may be off-line for maintenance and/or media replacement so that the production can run continuously. PPC Biofilter cultured a mesophilic bacteria inoculant from the solvents found in the exhaust and applied to the filter media during loading. The biofilter vessels are constructed of 8" thick, poured in-place concrete walls, with integral pretreatment humidifiers. The roof is constructed of pre-stressed hollow core concrete and is insulated to prevent condensation inside the biofilter.

Controlling the moisture of the filter media is the single most important variable affecting the performance of a biofiltration system. To insure stable and uniform moisture content in the filter media, PPC designed two independent pre-treatment counter current humidifiers, which are integral to the biofilter vessel. Each humidifier contains 600 cubic feet of structured packing and recirculates 1000 GPM of water from a common sump. Saturated exhaust stream from the humidifier passes through a 90 square feet of demister to prevent water droplets carryover on to the media and to distribute evenly across the reactor. These humidifiers provide >95% relative humidity and cool the 102-160 F exhaust to 70-90 F for an optimum biological activity. Load cells were installed to monitor the weight and the moisture content of the media. To insure adequate moisture of the media, a hybrid spray system has been installed, which allows the operator to spray on a timer basis through the solenoid activated over bed spray systems. Additional water will be sprayed if the load-cell controls call for the same. Thermocouples monitor the inlet and above the media temperature and the flow switches on the humidifier recirculation pipes monitor the flow rate. Water meters on the secondary humidification lines monitor the water usage of the over bed spray system to insure uniform moisture content in the filter media. Differential pressure transmitters monitor the pressure drop over the pre-treatment humidifiers and the biofilter beds. All instruments are data logged through the PLC/MMI. In the event of a process upset, such as high temperatures, loss of sump water, etc., system alarms are activated and an automatic by-pass opens protecting the biofilter. An automatic emergency paging system is then activated.

Performance Data of the Commercial Biofilter

The commercial system was designed to handle 45,000 ACFM of exhaust at 140 degrees F with an average inlet loading of 0.1g/m3. The system was placed into operation on May 23, 1997. A great deal of performance data has been collected, and the full-scale system has performed at or above design expectations.

Figure 6 shows the results of four laboratory tests conducted to confirm the amount of methane and total hydrocarbons present in the inlet gas stream to the biofilter, and Figure 7 shows the removal efficiency of the commercial system as measured with the portable FID. The system was designed to provide an 80% removal efficiency, corrected for inlet-methane, and the system has provided an average removal efficiencies of 78% for total VOCs, and >85% for non-methane hydrocarbons. Figure 6 shows a remarkable difference between the removal efficiencies of the biofilter for non-methane organics and total hydrocarbons. Figure 7 demonstrates that as the inlet concentration decreases, the removal efficiency declines. However, the exhaust emissions less methane remain fairly constant thus far. Current emissions going into the biofilter represent an annual emission rate of 54 tons of VOCs, while the exhaust from the biofilter represents about 8 tons per year. The data from both the pilot unit and the commercial installation show that the Serigraph’s system can take increased concentrations and mass loading while providing higher removal efficiencies.

CONCLUSIONS

Biofiltration is a cost effective control technology, and if used to treat relatively low concentration VOCs it can provide a significantly lower cost per ton treated than any other conventional APC technologies. One has to invest a significant amount of time in maintenance and "baby-sitting" of the biofilter, especially during the first 3-4 months of operation. But, it is worth the trouble because a considerable savings in the operating cost is one of the biggest benefits of this technology. Though it may be difficult to achieve >95% removal efficiencies on a continuous basis, by providing the necessary design features the technology can provide removal efficiencies of about 85% on a consistent basis. It is evident that, for a successful application of the biofiltration, off-gas components should thoroughly be analyzed for solubility, biodegradability, oxidation byproducts, acidification or alkalization, the need for the pretreatment of off-gas, etc. Until further technological advancements are made, pilot tests are highly recommended to avoid any unpleasant surprises.

ACKNOWLEDGMENTS

Serigraph acknowledges Wisconsin Department of Commerce (previously, Wisconsin Department of Development) for the financial grant towards the construction of the full-scale unit under the Waste Technology and Production Development (WTPD) fund. The authors and Serigraph also acknowledge Dr. Gero Leson, Leson Environmental Consulting, California, for his expert consulting, valuable inputs, and his involvement throughout the project.

REFERENCES

Ottengraf, S.P.P.; Van den Oever, A.H.C. Biotechnology and Bioengineering, 1982, Vol. XXV; pp. 3089-3102

This manuscript contains graphics that must be viewed and printed by the Adobe Acrobat (TM) plug-in included with this program. Please click here to activate Acrobat and view the graphics.

Back to Menu