Figure 1. BW agents shown in relative size and superimposed over the performance curve for a representative MERV 13 filter.
The anthrax mailing incidents following the events of September 11 heightened public awareness of the danger posed by terrorist use of biological weapons (BW). One of the primary concerns raised across the nation was whether or not buildings could be protected against intentional releases of biological agents. Governmental agencies and private industry expressed considerable interest in technologies such as HEPA filtration, ultraviolet germicidal irradiation (UVGI), and biodetection technologies but were frustrated to learn that retrofitting HEPA filters could be prohibitively expensive, that biodetection technology was still in an embryonic state, and that some experts claim UVGI is ineffective.

Such conclusions reflect misperceptions and misdirection in the approaches taken and do not preclude the possibility of protecting buildings in a cost-effective manner. This article offers engineers new perspectives by reviewing how effective existing technologies can be in biodefense applications.

Figure 2. Performance curve for a combined MERV 13/URV 13 system. The UVGI system has a 0.75 second exposure time.

Biological Weapon Agents

Thanks to the industriousness and ingenuity of military researchers around the globe, several dozen deadly pathogens exist today that have potential as weapons of mass destruction. As if in fulfillment of the old maxim that any weapon built will eventually be used, BW technology has trickled down into the hands of those with no inhibitions about using them in support of their political, religious, or personal agendas.

Several incidents in the United States alone have highlighted the potential threat of bioweapons. In 1984 a religious extremist group in Oregon disseminated Salmonella in supermarkets using hand-held sprayers. In 1991, right-wing extremists in Minnesota developed the deadly toxin ricin. In 1998, a white supremacist associated with the group Christian Identity was caught ordering anthrax, after successfully obtaining plague bacilli. The perpetrator of the 2001 anthrax mailings has yet to be apprehended.

Perhaps the greatest fear among building owners today is that a BW agent might be released into the air intakes of their building. The casualties and economic devastation resulting from such a bioterrorist attack might be as intolerable as any building collapse from a bomb blast. The possibility that numerous buildings might be simultaneously subject to bioterrorist releases could put the problem on the scale of a national disaster. What practical measures can be taken, if any, to mitigate this threat? The answer to this question requires a detailed look at how effective existing technologies can be at removing microorganisms.

Figure 3. Results of simulated releases of three BW agents in a 20-story building air intake for increasing levels of air cleaning (MERV/URV).

BW Agent Removal Technologies

BW agents consist of both microorganisms and toxins, which are biological poisons. They possess a definable size range and have some predictable filterability. Figure 1 shows approximately 50 known or suspected BW agents, including toxins, superimposed over a performance curve for a MERV 13 filter. Toxins are typically ground to a size somewhere between 1 and 6 microns, and as a result they are represented in Figure 1 by the minimum, the maximum, and the logmean size of 2.28 microns. It can be observed in Figure 1 that most agents will be removed at over 50% efficiency.

The retrofit of new filters in any ventilation system should include an engineering evaluation of the fan capacity since system pressure losses may be increased. Such a retrofit may either increase power consumption or decrease the total airflow volume, depending on the fan type. The cost of retrofitting higher-efficiency filters may also include installing new filter frames, sealing or welding of the filters and frames to limit bypass, instrumentation, maintenance costs, and possible expansion of duct size to accommodate design air velocities.

The dip in Figure 1 between 0.1 and 0.3 microns represents the most penetrating particle size range. It just so happens that microbes in this range tend to be susceptible to UVGI. Figure 2 shows a hypothetical 'performance curve' for a MERV 13 filter combined with an URV 13 UVGI system with 0.75 second exposure time (see the sidebar for an explanation of the URV standard). Note how the effect of combining UVGI with the filter is to attenuate the most penetrating particle size range and to level out the performance curve in comparison with Figure 1. Although the spores are not greatly affected by this level of UVGI and the toxins are not affected at all, there is a significant impact on the other bacteria and viruses. This improvement, furthermore, occurs without the pressure loss that would occur by increasing filter efficiency alone.

Retrofitting UVGI systems can be relatively simple where space is available. UV lamps are available today that can be mounted externally after drilling holes through ductwork. Pressure losses across such lamps are often negligible, but attention must be paid to maintaining design operating conditions (e.g., air velocity and temperature) of the UV lamps, otherwise UV power output could suffer drastically. Other factors that may need to be addressed in UVGI system design include rh, entry of visible light (which may reduce kill rates through photoreactivation), escape of UV into inhabited areas, and design of reflective panels to enhance UV exposure.

Figure 4. Particle counter capable of detecting micron-sized airborne particles. Reprinted with permission from Pacific Scientific Instruments.

Immune Building Design Criteria

"We want a system that removes 99.9% of biological agents" is a request often heard in the filter and UVGI industries lately. But what is the basis for this criterion? The fact is that this criterion has no basis and simply reflects the desire to remove as much of the agent as possible. In some cases concerned clients have specifically requested the installation of HEPA filters, without fully realizing the operating expense that such filters may incur and the possible impact on the ventilation system fans.

No criteria have yet been established for the performance of systems designed for protecting building occupants against biological agents. Any such criteria must have some analytical or empirical basis. Perhaps the most practical approach, and one being undertaken today in both military circles and the private sector, is the simulation of bioterrorist attack scenarios on computer models of buildings.

Figure 5. Self-powered air sampler. Reprinted with permission from Biochem Technologies, LLC.

Simulation Of Bioterrorist Attacks

In the Architectural Engineering Department of the Pennsylvania State University, research into immune building technologies is leading to new ideas and possible future standards. Multizone models of various types of buildings and ventilation systems are used to study the spread of contaminants over time. These results provide a basis for calculating the inhaled doses of pathogens and estimating casualties. This approach allows investigators to test the effectiveness of different filters and UVGI systems at reducing casualties.

Figure 3 shows one example of a series of simulations in which various combinations of MERV/URV systems were modeled in a 20-story building subject to releases of anthrax spores, smallpox, and botulinum in the outside air intakes. The average results for all three agents are shown in terms of the number of occupants protected from inhalation of fatal doses. The model building consists of a constant-volume system with 15% outside air and typical commercial office building volume and occupancy.

In the example shown in Figure 3, considerable protection is offered by the MERV 11/URV 11 air cleaning system and no significant benefits are provided by increasing the air cleaning level beyond MERV 13/URV 13. Similar results have been obtained for other attack scenarios, including internal releases in the AHU, other BW agents, and other building sizes. Such simulations can be performed for almost any building to determine the most cost-effective solutions. The previous example is idealized, but buildings can be modeled to any desired level of detail, including modeling of door and window leakage, filter bypass, plate-out effects on internal surfaces, and wind or stack effects.

In the example of Figure 3, the use of HEPA filters for air cleaning would appear to be overkill, which may be good news for engineers seeking economic solutions. In fact, this same conclusion about HEPA filters was arrived at previously for health care facilities by researchers over 30 years ago. The "99.9%" removal rate criterion often requested by clients seems to have no analytical justification. In the case of anthrax, for example, a removal rate of about 60% would seem to be adequate to prevent most, if not all, fatal infections in any recirculating constant-volume system with at least 15% outside air.

Ideally, air cleaning and disinfection systems would operate full-time and provide round-the-clock protection to building inhabitants. However, not all buildings can be easily retrofitted with such systems, and in some cases the retrofit may not be cost effective. There may also be a need to provide a higher degree of protection to certain areas of some buildings, such as control rooms, communication facilities, and sheltering zones. For such situations, options are limited to either separate ventilation systems or automatic control systems. The latter requires the use of detection technology.

Biodetection Technology

The holy grail of biodetection science is a sensor that will instantly identify any airborne threat and then initiate emergency system operation. Unfortunately, such technology does not exist yet. The biosensors and biodetectors available today can identify no more than a dozen BW agents and their response time leaves much to be desired. Furthermore, the expense of such systems can be prohibitive. Automatic detection technology holds great promise for the future but little practical value for the present.

Two alternatives exist that offer more cost-effective ways of dealing with the problem: particle detectors and air sampling. Particle detectors, or particle counters, can identify the presence of concentrations of microbial-sized particles in the air. They cannot distinguish dust or droplets from microbes, but this may not be critical or necessary. If a particle detector is located downstream of a filter and registers the presence of micron-sized particles, then either there is a major release upstream or the filter has broken through. In either case it would be prudent to sound an alarm and take any necessary action, which may include alarming to evacuate the building or isolating the system to prevent further spread of potential contaminants. The former control architecture is known as "detect-to-alarm" while the latter is known as "detect-to-isolate."

Many particle detectors are available today as scientific instruments, and these units can cost $15,000 to $50,000 or more. Newer portable or compact particle counters and particulate monitors are also becoming available today with prices as low as $3,000.

Air Sampling

Air sampling has been available for some time and is in common use for identifying microbial contamination problems in buildings. It is a relatively inexpensive technology but requires manual operation. Air drawn from a duct or room is impinged on a petri dish, and this plate is then cultured and evaluated in a laboratory. It can take 1 to 2 days to identify the presence of microbial agents but this is not necessarily a problem since almost all microbial agents require at least 3 days of incubation before causing symptoms. Detecting an airborne pathogen within 1 to 2 days of exposure allows sufficient time for medical treatment of building occupants unless the inhaled dosage is extraordinary. This approach is known as "detect-to-treat" control architecture.

A side benefit of air sampling is that building owners will have a record of the microbial air quality, which may assist in identifying the more mundane microbial hazards of indoor air. The average annual cost for such a detect-to-treat system can be as low as a few thousand dollars, depending on frequency of sampling and what lab arrangements can be made.


New computer simulations of BW agent releases in buildings suggest that high levels of protection against bioterrorism may be possible by combining the existing technologies of filtration and UVGI. The example results presented here do not necessarily apply to all buildings since considerable variations in airflow characteristics and system operation are possible, but the method of simulation can be used to provide a customized solution for any building.

Biodetection systems are not yet advanced enough or sufficiently all-encompassing to provide automatic detection and identification of microorganisms, but alternatives such as particle counters and air sampling offer practical and cost-effective solutions for applications where they are needed.

The subject of bioterrorism defense of buildings covers far more scope, and requires far more detailed attention, than can be addressed in this single article, but readers should be assured that the problem is a manageable one and that affordable solutions are possible. The use of immune building technologies for biodefense may even produce collateral benefits by reducing the incidence of respiratory disease. In fact, the widespread use of immune building technologies may even lead to a future in which many common respiratory diseases will cease to be a problem. ES

Some images associated with this article do not transfer to the Internet. To view them, please refer to the print version.

URV: A Proposed Standard for Sizing UVGI Systems

Sizing of UVGI systems can be a complicated matter that often requires the use of computer software. A proposed standard for UVGI systems being developed at Penn State University consists of a UVGI Rating Value (URV). The URV standard has been designed to be analogous to the MERV standard for filters developed by ASHRAE. Table 1 lists some of the proposed URV ratings and the values for average intensity that they represent. For any given average intensity, the dose can be computed by multiplying by the exposure time, as shown in the following table. The dose can then be used to compute the kill rate for any microbe for which a UVGI rate constant is known, as shown in the examples for smallpox and influenza at an exposure time of 0.75 seconds.

The URV standard is designed to represent the typical range of UVGI system sizes such that any UVGI system can be mated to any MERV filter to provide a roughly balanced removal rate across the full range of microbes. A MERV 11 filter, for example, can be matched with an URV 11 UVGI system if the exposure time is in the range of about 0.5 to 1.5 seconds. Exposure times much lower or higher than these may require scaling the URV system up or down, respectively, to obtain a similar dose rate.