An official website of the United States government.

This is not the current EPA website. To navigate to the current EPA website, please go to This website is historical material reflecting the EPA website as it existed on January 19, 2021. This website is no longer updated and links to external websites and some internal pages may not work. More information »

Significant New Alternatives Policy (SNAP)

Carbon Dioxide as a Fire Suppressant: Examining the Risks

A PDF version of this report is also available.


This document has been reviewed in accordance with U.S. Environmental Protection Agency policy and approved for publication and distribution. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.


Under the Clean Air Act Amendments of 1990, the U.S. Environmental Protection Agency (EPA) has the statutory authority to set phase-out dates for ozone-depleting substances (ODS) and to evaluate potential risks posed by proposed ODS substitutes. Under the terms of the Montreal Protocol on Substances that Deplete the Ozone Layer, EPA promulgated regulations to phase out the production of Halon 1301. In response to the halon phase-out effective January 1, 1994, the fire protection industry has been searching for alternatives. A number of alternative technologies have been proposed, including carbon dioxide (CO2 ) systems. This report was written to provide users of total flooding halon systems, who may be unfamiliar with total flooding carbon dioxide systems, with information regarding the potential dangers associated with carbon dioxide systems. Appropriate precautions must be taken before switching to carbon dioxide systems and with this report EPA attempts to raise awareness and promote the responsible use of carbon dioxide fire suppression systems. The authors of this report consulted with experts in the industry during the information-gathering stage for development of the report. An early draft of the document was read by members of the United Nations Environment Programme (UNEP) Halons Technical Options Committee (HTOC). Many experts within the fire protection industry provided data on incidents. The penultimate document was peer reviewed in September 1999 for its technical content by a distinguished group of experts, including:

  • Rich Hansen (Test Director), United States Coast Guard - R&D Center
  • Matsuo Ishiyama, member of HTOC, Corporate Advisor and Auditor, Halon Recycling and Banking Support Committee, Japan
  • Joseph A. Senecal, Ph.D., Director of Suppression Engineering, Kidde-Fenwal, Inc.
  • Charles F. Willms, P.E., Technical Director, Fire Suppression Systems Association
  • Thomas Wysocki, P.E., President and Senior Consultant, Guardian Services, Inc.
  • Roy Young, HTOC member, United Kingdom

Comments were received from all peer reviewers. Some reviewers expressed concern that the document be written clearly enough to lay out the associated risks in a way that neither promoted nor unduly discouraged the use of carbon dioxide-based fire extinguishing systems, and changes were made in the introduction to address this concern. A reviewer described the document as "a very valuable contribution to the safety subject and. . .should be used by carbon dioxide systems providers as a positive tool to promote training, maintenance, and adherence to proven standards." All reviewers were pleased that a report on the risks associated with carbon dioxide systems had been prepared.

One reviewer found the report to accurately reflect current "land-based" requirements, but added information related to the importance of training both new crew and contracted maintenance workers in marine applications. The conclusions of the report were changed to reflect this comment. One reviewer commented that a statement in the report was overly speculative. The report language was edited to clearly indicate that the statement is speculative. Specific technical definitions and information related to an accident event were contributed by one reviewer who also provided consistency between language of the report and correct technical terminology as used in standard National Fire Protection Association (NFPA) documentation. Extensive changes were made to the sections Extinguishing Mechanisms of Carbon Dioxide and Life Safety Considerations of Carbon Dioxide on the advice of one reviewer. Most other comments were minor editorial remarks generally for clarification. All comments were addressed in the final document.

EPA wishes to acknowledge everyone involved in this report and thanks all reviewers for their extensive time, effort, and expert guidance. EPA believes the peer reviewers provided information necessary to make this document technically stronger. Without the involvement of peer reviewers and industry contacts this report would not be possible. EPA accepts responsibility for all information presented and any errors contained in this document.


This paper provides information on the use and effectiveness of carbon dioxide in fire protection systems and describes incidents involving inadvertent exposure of personnel to the gas. Because carbon dioxide fire extinguishing systems will likely be used in place of those based on halon in some applications, this paper attempts to provide an increased awareness of the potential dangers associated with the use of carbon dioxide. EPA recognizes the environmental benefits of using carbon dioxide, but is concerned that personnel accustomed to the use of halon fire suppression systems may not be properly alerted to the special hazards of carbon dioxide. Governmental, military, civilian, and industrial sources were researched to obtain information on deaths and injuries associated with the use of carbon dioxide as a fire extinguishing agent. An examination of the risks associated with carbon dioxide extinguishing systems is also presented.

Carbon Dioxide as an Extinguishing Agent

Fire protection applications generally can be divided into two basic categories: 1) applications that allow the use of water-based sprinklers and 2) special hazards that require the use of some other fire extinguishing agent such as carbon dioxide, halon, halon replacements, dry chemicals, wet chemicals, or foams. According to industry consensus, special hazard applications comprise approximately 20 percent of total fire protection applications. Of the special hazard applications, approximately 20 percent of the market (based on dollars) is protected by carbon dioxide extinguishing agents. Carbon dioxide has been used extensively for many years in the special hazard fire protection industry worldwide. Between the 1920s and 1960s, carbon dioxide was the only gaseous fire suppression agent used to any degree, but halon-based systems were used extensively beginning in the 1960s. Carbon dioxide continues to be used in numerous applications around the world for the extinguishment of flammable liquid fires, gas fires, electrically energized fires and, to a lesser degree, fires involving ordinary cellulosic materials such as paper and cloth. Carbon dioxide can effectively suppress fires of most materials with the exception of active metals, metal hydrides, and materials containing their own oxygen source, such as cellulose nitrate (Wysocki 1992). The use of carbon dioxide is limited primarily by the factors influencing its method of application and its intrinsic health hazards.

Carbon dioxide is used internationally in marine applications in engine rooms, paint lockers, vehicle transport areas on cargo vessels, and in flammable liquid storage areas (Willms 1998). Large marine engine room systems may require as much as 20,000 lb of carbon dioxide per system. Carbon dioxide fire suppression systems are currently being used by the U.S. Navy and in commercial shipping applications.

The steel and aluminum industries also rely heavily on carbon dioxide fire protection. In the aluminum industry, for example, the rolling mill process requires the use of kerosene-like lubricants and coolants. Fires are prevalent in this application, occurring on the average of 1 per week in the typical aluminum plant (Wysocki 1998, Bischoff 1999). One particular aluminum processing company averages about 600 system discharges per year worldwide in all their fire protection applications using carbon dioxide, such as rolling mills, control rooms, and aluminum sheet printing (Stronach 1999). Many carbon dioxide systems in the metal processing industry are rapid discharge local application systems. In these applications, the carbon dioxide storage containers are located close to the outlet nozzles such that liquid carbon dioxide starts to discharge from the nozzle(s) in under 5 seconds (Wysocki 1998, Stronach 1999). These local application carbon dioxide systems range in size from 800 to 10,000 lb of compressed carbon dioxide (Bischoff 1999, Stronach 1999).

Carbon dioxide systems also are used in computer rooms (subfloor), wet chemistry benches, particle board chippers, equipment dust collectors, printing presses, cable trays, electrical rooms, motor control centers, switch gear locations, paint spray booths, hooded industrial fryers, high-voltage transformers, nuclear power facilities, waste storage facilities, aircraft cargo areas, and vehicle parking areas (Willms 1998, Wysocki 1998). Small carbon dioxide systems, such as those protecting paint lockers or fryers, use approximately 50 lb of carbon dioxide. Other systems use an average of about 300 to 500 lb of carbon dioxide (Willms 1998), but can use as much as 2,500 lb (Ishiyama 1998).

Several properties of carbon dioxide make it an attractive fire suppressant. It is not combustible and thus does not produce its own products of decomposition. Carbon dioxide provides its own pressurization for discharge from a storage container, eliminating the need for superpressurization. It leaves no residue, and hence precludes the need for agent clean up. (Clean up of fire-released debris would, of course, still be necessary in the case of a fire event.) Carbon dioxide is relatively nonreactive with most other materials. It provides three-dimensional protection because it is a gas under ambient conditions. It is electrically nonconductive and can be used in the presence of energized electrical equipment.

Extinguishing Mechanism of Carbon Dioxide

Flame extinguishment by carbon dioxide is predominantly by a thermophysical mechanism in which reacting gases are prevented from achieving a temperature high enough to maintain the free radical population necessary for sustaining the flame chemistry. For inert gases presently used as fire suppression agents (argon, nitrogen, carbon dioxide, and mixtures of these), the extinguishing concentration (As measured by the cup burner method (NFPA 2001)) is observed to be linearly related to the heat capacity of the agent-air mixture (Senecal 1999).

Although of minor importance in accomplishing fire suppression, carbon dioxide also dilutes the concentration of the reacting species in the flame, thereby reducing collision frequency of the reacting molecular species and slowing the rate of heat release (Senecal 1999).

Extinguishing Effectiveness of Carbon Dioxide

Carbon dioxide is the most commonly used "inert" gas extinguishing agent, followed by nitrogen (Friedman 1992). On a volume basis, carbon dioxide is approximately twice as effective as nitrogen (e.g., for ethanol fires, the minimum required volume ratios of carbon dioxide and nitrogen to air are 0.48 and 0.86, respectively). However, because carbon dioxide is 1.57 times heavier than nitrogen [44 and 28 molecular weight (MW), respectively] for a given volume, the two gases have nearly equivalent effectiveness on a weight basis.

Gas Volume Equivalent (GVEq) = vol. ratio for N2 / vol. ratio for CO2 =1.8
Weight Equivalent = GVEq x MWN 2 / MWCO2 = 1.1

The amount of carbon dioxide needed to reduce the oxygen level to a point at which various fuels are prevented from burning is relatively high and is also at a level where humans will suffer undesirable health effects. Table 1 presents the minimum required ratios of carbon dioxide to air (v/v), the corresponding oxygen concentration that will prevent burning of various vapor fuels at 25 degrees C, the theoretical minimum carbon dioxide concentration, and the minimum design concentration of carbon dioxide for various fuels.

Table 1 refers only to gases or vapors; however, the data are also relevant to liquids or solids because they burn by vaporizing or pyrolyzing. Generally, with a few exceptions such as hydrogen or carbon disulfide, a reduction of oxygen to 10 percent by volume would make fires and explosions impossible.

Use of Carbon Dioxide Extinguishing Systems

Carbon dioxide fire extinguishing systems are useful in protecting against fire hazards when an inert, electrically nonconductive, three-dimensional gas is essential or desirable and where clean up from the agent must be minimal. According to the NFPA, some of the types of hazards and equipment that carbon dioxide systems protect are "flammable liquid materials; electrical hazards, such as transformers, switches, circuit breakers, rotating equipment, and electronic equipment; engines utilizing gasoline and other flammable liquid fuels; ordinary combustibles such as paper, wood, and textiles; and hazardous solids" (NFPA 12).

Table 1.  Required Ratios (v/v) and Minimum Carbon Dioxide Concentrations to Prevent Combustion

Vapor Fuels CO2/aira (v/v) O2 Concentration (%) Theoretical Minimum CO2 Concentrationb (%) Minimum Design CO2 Concentration (%)
Carbon Disulfide 1.59 8.1 60 72
Hydrogen 1.54 8.2 62 75
Ethylene 0.68 12.5 41 49
Ethyl Ether 0.51 13.9 38 46
Ethanol 0.48 14.2 36 43
Propane 0.41 14.9 30 36
Acetone 0.41 14.9 27 34
Hexane 0.40 15.0 29 35
Benzene 0.40 15.0 31 37
Methane 0.33 15.7 25 34

a   Friedman 1989.
b   Coward and Jones 1952.

Life Safety Considerations of Carbon Dioxide

Health Effects

The health effects associated with exposure to carbon dioxide are paradoxical. At the minimum design concentration (34 percent) for its use as a total flooding fire suppressant, carbon dioxide is lethal. But because carbon dioxide is a physiologically active gas and is a normal component of blood gases at low concentrations, its effects at lower concentrations (under 4 percent) may be beneficial under certain exposure conditions. ( Appendix B discusses the lethal effects of carbon dioxide at high exposure levels (Part I) and the potentially beneficial effects of carbon dioxide at low exposure concentrations, as well as the use of added carbon dioxide in specialized flooding systems using inert gases (Part II))

At concentrations greater than 17 percent, such as those encountered during carbon dioxide fire suppressant use, loss of controlled and purposeful activity, unconsciousness, convulsions, coma, and death occur within 1 minute of initial inhalation of carbon dioxide (OSHA 1989, CCOHS 1990, Dalgaard et al. 1972, CATAMA 1953, Lambertsen 1971). At exposures between 10 and 15 percent, carbon dioxide has been shown to cause unconsciousness, drowsiness, severe muscle twitching, and dizziness within several minutes (Wong 1992, CATAMA 1953, Sechzer et al. 1960). Within a few minutes to an hour after exposure to concentrations between 7 and 10 percent, unconsciousness, dizziness, headache, visual and hearing dysfunction, mental depression, shortness of breath, and sweating have been observed (Schulte 1964, CATAMA 1953, Dripps and Comroe 1947, Wong 1992, Sechzer et al. 1960, OSHA 1989). Exposures to 4 to 7 percent carbon dioxide can result in headache; hearing and visual disturbances; increased blood pressure; dyspnea, or difficulty breathing; mental depression; and tremors (Schulte 1964; Consolazio et al. 1947; White et al. 1952; Wong 1992; Kety and Schmidt 1948; Gellhorn 1936; Gellhorn and Spiesman 1934, 1935; Schulte 1964). Part I of Appendix B discusses human health effects of high-concentration exposure to carbon dioxide in greater detail.

In human subjects exposed to low concentrations (less than 4 percent) of carbon dioxide for up to 30 minutes, dilation of cerebral blood vessels, increased pulmonary ventilation, and increased oxygen delivery to the tissues were observed (Gibbs et al. 1943, Patterson et al. 1955). These data suggest that carbon dioxide exposure can aid in counteracting effects (i.e., impaired brain function) of exposure to an oxygen-deficient atmosphere (Gibbs et al. 1943). These results were used by the United Kingdom regulatory community to differentiate between inert gas systems for fire suppression that contain carbon dioxide and those that do not (HAG 1995). During similar low-concentration exposure scenarios in humans, however, other researchers have recorded slight increases in blood pressure, hearing loss, sweating, headache, and dyspnea (Gellhorn and Speisman 1934, 1935; Schneider and Truesdale 1922; Schulte 1964). Part II of Appendix B discusses these results in greater detail.

Safety Measures

As with other fire protection systems, a number of regulatory agencies or authorities having jurisdiction (AHJ) administer the design, installation, testing, maintenance, and use of carbon dioxide systems. The authority that regulates the system depends on where the system is located, the intended scenario, and the type of system. Many AHJs that regulate industrial, commercial, and nonmarine applications utilize the NFPA consensus standard covering carbon dioxide extinguishing systems (NFPA 12). Although the standard itself does not hold the force of law, governments and local authorities adopt the standard as their governing fire code. Marine applications are regulated depending on whether the vessels navigate domestic or international waters. U.S. Coast Guard (USCG) regulations pertain to ships in domestic waters and are published in the Code of Federal Regulations (46 CFR Part 76.15). Internationally registered vessels are covered under the International Maritime Organization's Safety of Life at Sea (SOLAS) (IMO 1992). In workplaces that are land-based, the Occupational Safety and Health Administration (OSHA) regulates the exposure to carbon dioxide in order to ensure worker safety.

Design, Specification, and Component Approval

Generally, the process of acquiring fire suppression system approval starts with the manufacturer "listing" its components through organizations such as Underwriters Laboratory or Factory Mutual in the United States. Part of the listing process is the development of an instruction and maintenance manual that includes a description of the full operation of the system along with system drawings. Specifications or plans for the carbon dioxide system are prepared under the supervision of an experienced and qualified person knowledgeable in the design of carbon dioxide systems and with the advice of the AHJ. The designs are then submitted to the AHJ before installation begins.

Installation and Testing

Installation of the carbon dioxide system is usually performed by manufacturers' representatives or distributors. Although the installers are not given a formal accreditation or certification, they are trained by the manufacturer regarding proper installation of system components. The completed system is inspected and tested by appropriate personnel to meet the approval requirements of the AHJ. Often these requirements include:

(A) Performance of a full discharge test of the entire design quantity through the piping and into the intended hazard area, for each hazard area, if the system protects more than one. A check to verify that the design concentration is achieved and maintained for the specified hold time applies to total flooding type systems only.
(B) Operational checks of all devices necessary for proper functioning of the system, including detection, alarm, and actuation.
(C) Checks for proper labeling of devices and protected areas warning occupants of the possible discharge of carbon dioxide. In addition, signage must be present to warn personnel to vacate the area when the alarm sounds. (No foreign language requirements (e.g., Spanish) for signage are specified by U.S. AHJs. Ideally all labels and warning signs should be printed both in English and in the predominant language of non-English-reading workers (NIOSH 1976))
(D) Complete inspections of the system and the hazard area to ensure that the system meets the specifications and that it is appropriate for the type of fire hazard.

Use Controls

Despite the use of carbon dioxide in fire-fighting applications above its lethal concentration, NFPA 12 does not limit its use in occupied areas. The standard calls for safeguards such as pre-discharge alarms and time delays to ensure prompt evacuation prior to discharge, prevent entry into areas where carbon dioxide has been discharged, and provide means for prompt rescue of any trapped personnel.

The standard also requires that personnel be warned of the hazards involved as well as be provided with training regarding the alarm signal and safe evacuation procedures. In addition, NFPA 12 requires that a supervised "lock-out" be provided to prevent accidental or deliberate discharge of a system when persons not familiar with the system and its operation are present in a protected space (NFPA 12).4 The Appendix to NFPA 12 lists the following steps and safeguards that may be used to prevent injury or death to personnel in areas where carbon dioxide is discharged: (The degree of compliance with the suggestions provided in NFPA 12 varies across different facilities. The 2000 edition of NFPA 12 will include an additional provision for mandatory evacuation of the protected area prior to conducting any testing, servicing, or maintenance on the carbon dioxide system (Willms 1999))

(A) Provision of adequate aisle ways and routes of exit. These areas should be kept clear at all times.
(B) Provision of the necessary additional or emergency lighting, or both, and directional signs to ensure quick, safe evacuation.
(C) Provision of alarms within such areas that will operate immediately upon activation of the system on detection of the fire, with the discharge of carbon dioxide and the activation of automatic door closures delayed for sufficient time to evacuate the area before discharge begins. (In the next edition of the NFPA 12 Standard this provision will be revised to state that time delays and predischarge alarms that operate prior to discharge should be used (Willms 1999))
(D) Provision of only outward swinging, self-closing doors at exits from hazardous areas, and, where such doors are latched, provision of panic hardware.
(E) Provision of continuous alarms at entrances to such areas until atmosphere has been restored to normal.
(F) Provision for adding an odor to the carbon dioxide so that hazardous atmospheres in such areas may be recognized.
(G) Provision of warning and instruction signs at entrances to and inside such areas.
(H) Provision for prompt discovery and rescue of personnel that may be rendered unconscious or physically impaired in such areas. This may be accomplished by having such areas searched immediately after carbon dioxide discharge stops by trained personnel equipped with proper breathing equipment. Those rendered unconscious by carbon dioxide can be restored without permanent injury by artificial respiration, if removed quickly from the hazardous atmosphere. Self-contained breathing equipment and personnel trained in its use, and in rescue practices including artificial respiration, should be readily available.
(I) Provision of instructions and drills of all personnel in the vicinity of such areas, including maintenance or construction people who may be brought into the area to ensure their correct action when carbon dioxide protective equipment operates.
(J) Provision of means for prompt ventilation of such areas. Forced ventilation will often be necessary. Care should be taken to really dissipate hazardous atmospheres and not merely move them to another location. Carbon dioxide is heavier than air.
(K) Provision of such other steps and safeguards necessary to prevent injury or death as indicated by a careful study of each particular situation.
(L) Provision for mandatory evacuation of the protected area prior to conducting any testing, service, or maintenance on the CO2 system.

Industrial Risk Insurers (IRI), one of the insurance companies that provides property and business interruption insurance to large Fortune 500 companies such as Ford, General Motors, and Chrysler (IRI 1994), uses NFPA 12 as a basis for their insurance process and has prepared an interpretative guideline to the NFPA 12 Standard (IM 13.3.1). IM 13.3.1 interprets NFPA 12 and also specifies the use of a "system lock-out." A system lock-out is a device that either mechanically or electrically prevents the system from discharging. Examples of system lock-outs include manually operated valves that block the flow of an agent through downstream pipe work. Similarly, IRI also suggests that for normally unoccupied areas where fast growth fires may occur, a "supervised intermittent time delay" may be desired. Such devices function only when personnel are in the protected area and allow the system to discharge gas only after an extended time delay, thus allowing personnel to egress the area prior to discharge.

International maritime use of carbon dioxide extinguishing systems is extensive. Fire protection in these applications is covered under the regulations and requirements set forth in the International Maritime Organization's SOLAS (IMO 1992). As with NFPA 12, SOLAS does not prevent the use of carbon dioxide in normally occupied areas. Also similar to NFPA, SOLAS requires that "means be provided for automatically giving audible warning of the release of fire-extinguishing medium into a space in which personnel normally work or to which they have access." The alarm must operate for a suitable amount of time prior to the gas being released. Similar to NFPA 12, SOLAS requires that access doors to the areas where fire-extinguishing medium is stored shall have doors that open outwards. These requirements are not differentiated for carbon dioxide or halogenated hydrocarbon or inert gas agent systems. Unlike NFPA, SOLAS mandates that "automatic release of gaseous fire-extinguishing medium shall not be permitted" except with respect to local application systems.

USCG regulations for carbon dioxide systems in passenger vessels are documented in 46 CFR Part 76.15. Separate subparts describe different types of vessels. Similar to SOLAS, 46 CFR Part 76.15 stipulates manual control of cylinder activation. (It should be noted that 46 CFR Part 76.15-20 stipulates that "Systems...consisting of not more than 300 lb of carbon dioxide, may have the cylinders located within the space protected. If the cylinder stowage is within the space protected, the system shall be arranged in an approved manner to be automatically operated by a heat actuator within the space in addition to the regular remote and local controls.") 46 CFR Part 76.15 also requires that systems using more than 300 lb of carbon dioxide must be fitted with an "approved delayed discharge" arranged in such a way that when the alarm sounds the carbon dioxide is not released for at least 20 seconds. This requirement also may pertain to systems of less than 300 lb depending on the number of protected levels and the egress pathway configurations. To minimize the possibility of inadvertent actuations, USCG specifies that two separate manual controls be operated for release of carbon dioxide, thereby requiring two independent actuations to occur before carbon dioxide discharges into the protected space. In addition, all personnel must be evacuated from the protected space prior to performing any testing or maintenance on the carbon dioxide system (Willms 1999). (The 2000 edition of the NFPA 12 Standard includes a chapter on marine applications mandating evacuation of a space prior to testing and other activities (Willms 1999))

In land-based workplace environments, OSHA regulates the use of carbon dioxide. These regulations are provided in 29 CFR Parts 1910.160 and 1910.162, which outline the requirements for general and gaseous fixed extinguishing systems, respectively. Despite the fact that the concentration of carbon dioxide needed to extinguish fires is above the lethal level, OSHA does not prevent the use of carbon dioxide in normally occupied areas. (However, OSHA does explicitly limit the use of chlorobromomethane and carbon tetrachloride as extinguishing agents where employees may be exposed (29 CFR Part 1910.160 (b) (11).) For carbon dioxide systems, OSHA requires a predischarge alarm for alerting employees of the impending release of carbon dioxide when the design concentration is greater than 4 percent (which is essentially true for all carbon dioxide systems, see Table 1). This predischarge alarm must allow sufficient time delay for personnel to safely exit the area prior to discharge. Although it is speculative, it is likely that these regulations would confer adequate protection only in the event of planned discharge, not accidental discharge. Accidental discharges have occurred, however, in which adherence to regulations has provided personnel protection, whereas some planned discharges have resulted in injury to personnel.

The purpose of the predischarge alarm required by OSHA, NFPA, and SOLAS is to allow occupants time to evacuate an area into which carbon dioxide will be discharged. However, ensuring egress from spaces that are either very large or that have obstacles or complicated passageways has proven to be difficult. Evacuation is particularly difficult once discharge begins because of reduced visibility, the loud noise of discharge, and the disorientation resulting from the physiological effects of carbon dioxide.

In a number of the regulations, concern is given to the possibility of carbon dioxide leaking or flowing into adjacent, low-lying spaces such as pits, tunnels, and passageways. In these cases, carbon dioxide can inadvertently create suffocating atmospheres that are neither visible nor detectable.

Two examples of the ideal fire scenario and how the carbon dioxide systems/safeguards are expected to work are described below for two applications (car parks in Japan and a marine engine room). Carbon dioxide systems are used in Japan in car parks (known in the United States as parking garages) such as tower parking or floor machinery parking, but not in normally occupied car parking facilities, where clean agents are generally used. The enclosed volume of the typical garage facility ranges from 1,000 m 3 to 1,500 m 3 [roughly 35,000 ft 3 to 53,000 ft 3 ], where 800 kg to 1,125 kg [1,764 lb to 2,480 lb] of carbon dioxide are used. The system operates through automatic discharge with a manual override option. The typical fire scenario for a carbon dioxide system in a tower parking or floor machinery parking facility is shown in Figure 1 (Ishiyama 1998).

Marine applications, such as engine rooms, are areas where carbon dioxide systems are often used. The typical fire scenario for a carbon dioxide system in a large marine engine room is shown in Figure 2. Most of these systems function through manual activation (except systems containing less than 300 lb [136 kg] of carbon dioxide, which correspond to enclosure volumes less than 6,000 ft 3 [170 m 3 ]). A typical engine room will be on the order of 250,000 ft 3 [7,079 m 3 ] and use 10,000 lb [4,536 kg] of carbon dioxide (Gustafson 1998). Despite the safeguards that are required by regulation and meant to guard against injuries associated with carbon dioxide fire extinguishing systems, accidents resulting in injuries and deaths have occurred, primarily caused by not following established safety procedures.

Figures 1 and 2

Review of Incidents (Accidents/Deaths) Involving Carbon Dioxide as a Fire Extinguishing Agent

A comprehensive review of carbon dioxide incidents in fire protection was undertaken by searching governmental, military, public, and private document archives. The variability in record-keeping practices of various organizations has impacted the success of the data collection effort.

Incident Record Search

Library/Internet Searches Completed
Literature Searches

Two literature searches were conducted. The first literature search (1975-present) was conducted to collect information on incident reports on injuries/deaths associated with carbon dioxide as a fire protection agent. Key words used in the searches included: death(s), incident(s), injury(ies), accident(s), carbon dioxide (or CO2 ), fire extinguishing agent(s), fire suppressant(s), maritime, marine, shipping industry, military, civilian, industry(ies), company(ies), firm(s), human, men, worker(s), employee(s), laborer(s). All relevant articles were retrieved. The following databases were searched:

  • OSHA 1973-1997
  • MEDLINE 1966-1997
  • Toxline 1965-1997
  • Energy SciTec 1974-1997
  • NTIS 1964-1997
  • GPO Publications Reference File
  • IAC Trade and Industry Database 1976-1997
  • Life Sciences Collection 1982-1997
  • Ei Compendex 1970-1977
  • Wilson Applied Science and Technology Abstracts 1983-1997
  • Chemical Safety NewsBase 1981-1997
  • GPO Monthly Catalog 1997

National Institute for Occupational Safety and Health (NIOSH) Library Search: A search of the NIOSH database at their library in Cincinnati, Ohio, was conducted.

Internet Search: An Internet search using the same key words used in the library search also was conducted within the following electronic databases:

  • Government Printing Office
  • FireDoc
  • NFPA Online Database

Professional Contacts

Contacts were asked to provide information on incidents concerning human deaths and/or injuries associated with the accidental or intentional discharge of carbon dioxide fire protection systems. (Accidental discharges include those occurring during maintenance operations on or near the carbon dioxide system, testing exercises, or those resulting from operator error or a faulty system component. Intentional discharges are generally those occurring in fire situations; however, they also include some discharges during testing exercises or due to a false alarm.) Details of the incident (e.g., date, site name, and location of the incident) were requested, as well as a description of the cause of the incident and the number of people injured or killed. Although this information was requested, the amount of information available varied by incident.

Associations/Private Companies/Government Organizations/Research Laboratories

All relevant information was retrieved directly from the following sites and/or from contacts that were identified therein:

  • The Society of Fire Protection Engineers
  • National Association of Fire Equipment Distributors
  • Fire Suppression Systems Association
  • Hughes Associates, Inc.
  • Kidde International
  • Ansul Fire Protection
  • Fike Corporation
  • Insurance companies that specialize in high-performance risk protection
  • National Defense Canada
  • U.S. Department of the Navy
  • U.S. Department of Energy (DOE)
  • USCG
  • NIOSH - Division of Safety Research
  • Center for Global Environmental Technologies, New Mexico Engineering Research Institute (NMERI)
  • National Fire Laboratory, Canadian Research Council
  • Ship Support Agency, United Kingdom Ministry of Defense
  • Association of German Safety Engineers
  • Bavarian Land Institute for Labor Protection
  • Bavarian Land Institute for Medicine
  • Coordinating Office for Labor Protection
  • Directorate of Fire Brigade Affairs
  • Environmental Department (Umweltbundesamt)
  • Federal Labor Association
  • Federal Union of Fire Extinguishers and Installation
  • Federal Union of Professional Safety Engineers
  • Federal Institute for Occupational Safety and Health
  • Fire Shelter Industries
  • German Society of Occupational Health and Hazard
  • German Fire Union
  • Home Office of the Federal State of Baden Württemberg
  • Hygiene Institute
  • Institute of Research for Fire Safety (Universitaet Karlsruhe)
  • Labor Protection and Technical Safety
  • Ministry of Internal Affairs
  • Office of Damage Prevention
  • Union of Safety (Insurance)
  • Australian Maritime Safety Authority
  • Richard Bromberg, HTOC representative from Brazil (A more detailed library search was performed to collect corroborating information on the incident provided by this source.)
  • Matsuo Ishiyama, HTOC representative from Japan
  • Syncrude Canada Ltd.
  • Loss Prevention Council, U.K.

Search Results

The results of this comprehensive data review are presented in Appendix A. From 1975 to the present, a total of 51 carbon dioxide incident records were located that reported a total of 72 deaths and 145 injuries resulting from accidents involving the discharge of carbon dioxide fire extinguishing systems. (Information was requested on any incidents of death or injury resulting from the use of carbon dioxide fire extinguishing systems. Data were requested on both fire- and nonfire-related incidents; however, it was significantly more difficult to gather information on fire-related incidents. Injuries and fatalities from fire situations are generally classified only as fire-related and are not broken down by the fire suppression agent that was used. Therefore, carbon dioxide deaths and injuries from fire-related situations may not be adequately represented. In addition, it should be noted that any discharge of carbon dioxide which resulted in no injuries and/or deaths was not included in the analysis.) All the deaths that were attributed to carbon dioxide were the result of asphyxiation. Details about the injuries were generally not provided in the incident reports, although some OSHA inspections listed asphyxia as the nature of the injury.

Prior to 1975, a total of 11 incident records were located that reported a total of 47 deaths and 7 injuries involving carbon dioxide. Twenty of the 47 deaths occurred in England prior to 1963; however, the cause of these deaths is unknown. Table 2 presents a categorical breakdown of the carbon dioxide incident reports and the deaths/injuries identified.

Although a comprehensive review was performed, it should be noted that data developed through this process may be incomplete because: 1) additional sources of data may be difficult to uncover (e.g., international incidents), 2) records are incomplete, 3) agencies are not required to report, 4)anecdotal information is sketchy and difficult to verify, and 5) fire-related deaths due to CO2 are generally not well documented.

Table 2.  Search Results

Use Category Number of Incidents Deaths Injuries
United States and Canada
1975- Present Military 9 10 15
Nonmilitary 20 19 73
Before 1975 Military 3 11 0
Nonmilitary 5 3 3
Total 37 43 91
1975- Present Military 1 4 5
Nonmilitary 21 39 52
Before 1975 Military 0 0 0
Nonmilitarya 3 33 4
Total 25 76 61
Total 62 119 152

a Included in the total international nonmilitary incidents, deaths, and injuries before 1975 are the 20 deaths resulting from the use of carbon dioxide as a fire suppressant in England from 1945 to the mid 1960s, for which the cause is unknown.

All of the 13 military incidents reported since around 1948 were marine-related. Only 11 of the 49 civilian (commercial, industrial, or state) incidents reported during the same time period were marine-related. The remaining incidents occurred in data processing centers, nuclear power plants, pilot training centers, airplanes, bus garages, emergency unit communication centers, waste storage facilities, underground parking garages, steel rolling mills, motor vehicle assembly lines, and other facilities.

Results presented in Appendix A show that accidental exposure to carbon dioxide during maintenance or testing was found to be the largest cause of death or injury. In some cases, personnel did not follow required safety procedures that may have prevented the injury or death and perhaps even the exposure itself. In several instances, new procedures have been introduced as a result of the incident. The causes of the injuries and/or deaths are summarized in Table 3.

In some cases, maintenance on items other than the fire extinguishing system itself was the cause of the accidental discharge. The most recent reported case occurred at the Test Reactor Area, Idaho National Engineering and Environment Lab (a major DOE site) where carbon dioxide was accidentally released into an electrical switchgear building during routine preventative maintenance on electrical breakers. In another recent incident on a Brazilian oil tanker docked in harbor, a cleaning crew accidentally discharged the carbon dioxide system while working below deck. Similarly, at the Murray Ohio Manufacturing Company, workers discharged the carbon dioxide system while performing an installation near a detector that actuated the system. On the Navy Replenishment Oiler, a maintenance worker lost his footing and stepped on the activation valve while performing maintenance on an overhead light. In these incidents, it was not noted whether preliminary precautionary measures were followed as stated in OSHA, SOLAS, or NFPA guidance. However, in certain other instances, the required precautionary measures were not followed. For example, in the USS Sumter incident, sailors were performing planned maintenance on a carbon dioxide system in a paint locker when the system discharged. Later it was determined that these personnel skipped three of the four preliminary steps on the Maintenance Requirement Card.

In testing and training situations, discharges causing death and injuries were not always accidental. In two reported incidents, the carbon dioxide system was intentionally discharged for testing purposes and the gas escaped into an adjacent area (University of Iowa Hazardous Waste Storage Facility, A.O. Smith Automotive Products Company). In a 1993 incident in Japan, CO2 was intentionally discharged into an outdoor pit as part of a training exercise. Personnel subsequently entered the pit, unaware of the discharge. Two deaths occurred during a "puff" test of the carbon dioxide system onboard the Cape Diamond cargo vessel. Subsequent investigations indicated that shipboard personnel were not evacuated from the engine room during the test, as should have occurred in accordance with established safety procedures. Furthermore, the main discharge valve was not closed completely, releasing more carbon dioxide than anticipated.

Table 3. Causes of Injuries and/or Death Associated with Carbon Dioxide Discharges After 1975. a

Cause of Injuries/Death Incident Referenceb
Accidental Discharge During Maintenance/Repairs to the Carbon Dioxide System Navy Aircraft Carrier (1993) USS Sumter
Turbo Generator
Little Creek Naval
Navy Aircraft Carrier (1980) Cartercliffe Hall Cargo Vessel Carolina Fire Protection Automated Fire Suppression Systems
Autoridad Energia Electrica-Planta
Darwin 1997
Heath 1993
Allen 1997
Heath 1993
Darwin 1997
Warner 1991
Allen 1997
OSHA 1999 OSHA 1999
Accidental Discharge During Maintenance in the Vicinity of Carbon Dioxide System Brazilian Oil Tanker Murray Manufacturing Co. Navy Replenishment Oiler Oiler Kalamazoo
Navy Submarine Tender
SS Lash Atlantico
Stevens Technical Services Inc. Test Reactor Area, Idaho National Engineering and Environment Lab
Bromberg 1998
McDonald 1996
Darwin 1997
Heath 1993
Darwin 1997
Hager 1981
OSHA 1999
Caves 1998
Accidental Discharge During
Cape Diamond Marine Casualty Investigation
Report 1996
Accidental Discharge During Fire
LNG Carrier
Surry Nuclear Power Station
Paci 1996
Warnick 1986
Accidental Discharge from Faulty
Installation or System Component
Dresden Sempergalerie
Hope Creek
Drescher and Beez 1993
Caves 1998
Accidental Discharge from
Operator Error
French Data Center
Car Park (Japan)
Gros et al. 1987
Ishiyama 1998
Accidental Discharge - False Alarm Consolidated Edison Co. Barge
Meredith/Burda Corporation
OSHA 1998
OSHA 1999
Intentional Discharge During
U. of Iowa Hazardous Waste
Storage Facility
Japanese Outdoor Pit
A.O. Smith Automotive Products
Bullard 1994 Ishiyama 1998
OSHA 1999
Intentional Discharge During Fire
Navy Aircraft Carrier (1966) Australian Naval Ship Westralia Airline Constellation Ravenswood Aluminum Corporation
Muscle Shoals Construction Site
Darwin 1997
Webb 1998
Gibbons 1997
OSHA 1999 OSHA 1999
Intentional Discharge - False Alarm Japan Ishiyama 1998

a Incidents where the cause of discharge was uncertain are not included in the table.
b References from Table 3 are listed in Appendix A.

Examining the Risks Associated with Carbon Dioxide Extinguishing Systems

The risk involved with the use of carbon dioxide systems is based on the fact that the level of carbon dioxide needed to extinguish fires (and, thus, to protect an enclosure) is many times greater than the lethal concentration. For instance, the minimum design concentration to suppress a propane fire is 36 percent. This concentration of carbon dioxide can produce convulsions, unconsciousness, and death within several seconds. Since carbon dioxide cylinder store rooms are often relatively small compared to the protected areas, inadvertent discharges into these store rooms will also produce levels much higher than the lethal level. Because the consequences of exposure happen quickly and without warning, there is little or no margin for error.

It is intended that total flooding carbon dioxide systems be designed such that human exposure does not occur during fire-fighting scenarios. Predischarge alarms and time delays are prescribed in NFPA 12, OSHA, and SOLAS guidelines to prevent such exposure. Hence, relatively few accidents involving carbon dioxide systems occur during fire events; rather, accidents most often occur during maintenance of the carbon dioxide system itself, during maintenance around the carbon dioxide system, or to a more limited extent, during testing of the fire suppression system. Of the accidental discharges that occurred during maintenance, results of the survey indicated that the deaths and/or injuries from carbon dioxide exposure were caused by: 1) inadvertently actuating the system because there was a lack of adequate safety procedures to prevent such discharges, 2) failure to adhere to safety procedures, or 3) low technical proficiency of personnel in the vicinity of the carbon dioxide system.

Although the risk associated with the use of carbon dioxide for fire protection in protected enclosures is fairly well understood by regulators, standard-setting bodies, and insurers, the risk of carbon dioxide may not be well understood by the maintenance workers who perform functions on or around carbon dioxide systems. The failure to adhere to prescribed safety measures is a demonstration of this lack of understanding and appreciation of the dangers associated with carbon dioxide. Precautionary measures must be mandated to ensure that personnel follow strict guidelines, even if those personnel are simply entering the storage areas where the carbon dioxide system cylinders and components are being housed.

This point is exemplified by the German experience with the use of carbon dioxide in fire protection. In Germany, a large number of carbon dioxide systems are used to protect facilities and installations. Most of these are equipped with automatic release of carbon dioxide, even in occupied spaces. Despite the relative abundance of carbon dioxide systems in Germany and an exhaustive search of German records for accidents involving carbon dioxide, only one reported nonfire event was found. Personal communication with a number of sources (Brunner 1998, Schlosser 1997, Lechtenberg-Autfarth 1998) supports the finding that relatively few accidents during nonfire events have occurred with carbon dioxide in Germany. (It should be noted, however, that accidents during fire events were more difficult to locate because German data sources did not distinguish between fatalities and injuries caused by the fire and fatalities and injuries caused by the use of carbon dioxide.) The good safety record of the German experience may be attributed to their approach in installing and operating carbon dioxide systems.

In Germany (and much of Europe), unlike in the United States, only certified carbon dioxide-specialized installers are allowed to install the carbon dioxide systems. Once the system is installed, it is checked and approved by VdS Schadenverhütung (VdS), an approval body much like Factory Mutual. Regulations on system operations are strictly enforced and ensure that time delays are adequate to allow egress, that the alarms are functioning properly, and that rules and warnings are posted in the vicinity of the carbon dioxide system. Approval is granted to use the system only if it meets all the standards and requirements. In addition, according to the Comité Européen des Assurances (CEA) (The CEA is the federation of national insurance company associations in European market economy countries), the carbon dioxide installation and protected risk are required to be inspected at least once a year by an expert of the AHJ (CEA 1997).

In addition to the system of double and triple checks imposed by the German authorities, the prevalence of carbon dioxide use in Germany may have provided increased awareness and education of the agent's risks and dangers.

Because of the widespread use of Halon 1301 in the United States, which is safer than carbon dioxide at fire-fighting concentrations, there may be a lower awareness of the hazards surrounding carbon dioxide use. Experience has shown that a relatively higher margin of safety has been experienced with the use of Halon 1301 compared to carbon dioxide. This high safety margin may add to the lack of awareness of the dangers involved with using carbon dioxide systems.

Conclusion and Recommendations

A review of accidental deaths or injuries related to carbon dioxide use in fire protection indicates that the majority of reported incidents occurred during maintenance on or around the carbon dioxide fire protection system. In many of the situations where carbon dioxide exposure led to death or injury during maintenance operations, the discharge resulted from personnel inadvertently touching, hitting, or depressing a component of the system. In some cases, personnel did not adhere to the precautionary measures prescribed. In other cases, the safety measures were followed, but other accidental discharge mechanisms occurred.

Examination of the accident records shows that a disproportionately large number of accidents involving carbon dioxide have occurred on marine vessels. A number of factors may play a part in these occurrences. First, a limited number of personnel on the ship's crew have training and authority to activate the carbon dioxide system (Gustafson 1998). These few crew members are very well trained regarding the system's operation, however, the remaining personnel would not have the same level of sophisticated knowledge. In particular, new crew members and contracted maintenance workers may be unfamiliar with a ship's particular installation, even if they are aware of the potential dangers of carbon dioxide systems in general. This unfamiliarity could result in an inadvertent actuation, and it is therefore important that ship operators provide instruction on and require adherence to ship-specific procedures (Hansen 1999). The lack of training may cause certain personnel to touch, tamper, or hit system components, which then trigger activation. In addition, untrained personnel may disregard warning signs or alarms because they have not been adequately informed of the hazards. In addition, because of the design of many shipboard systems, the manual activation mechanism is sometimes a cable connected from a lever to the actuation device. In some designs the cable is not enclosed in a protective casing where it attaches to the pilot cylinders. The exposed nature of this device makes accidental deployment easier. In most system designs, however, the cable runs in conduit with pulleys to provide for turns and bends in the cable run. Furthermore, two separate controls are necessary to activate USCG-approved shipboard systems over 300 lb, thereby reducing the risk of accidental discharge resulting from exposed cables (Wysocki 1999).

Another factor influencing the safety record of marine applications is the nature of the regulatory requirements governing use of carbon dioxide systems. Maritime regulations (46 CFR Part 76.15 and SOLAS) do not provide detailed requirements to ensure safety of personnel. These maritime regulations can be contrasted with the NFPA standard that has more specific suggestions to protect personnel against the adverse effect of carbon dioxide. Improvement of maritime regulations would at least provide specific requirements that would presumably help reduce the accidental exposures that occur in marine applications.

Additionally, in certain instances language barriers may present a source of additional risk. For example, if signage and training manuals are available only in English, non-English-reading personnel may not receive adequate or timely warning. Hence, making these materials available in the predominant language of non-English-reading workers may help to educate personnel and thereby reduce risks.


Bischoff, Bernie. 1999. Chemetron Fire Systems, Matteson, IL, personal communication.

Brunner, Dr. Walter. 1998. envico AG, Gasometer Strasse 9, Ch 8031 Zurich, Switzerland, personal communication.

CATAMA. 1953. Aviation Toxicology-an Introduction to the Subject and a Handbook of Data.

Committee on Aviation Toxicology, Aero Medical Association. The Blakiston Co.: New York, NY. pp. 6-9, 31-39, 52-55, 74-79, 110-115.

CCOHS. 1990. Carbon Dioxide Chemical Infogram. Canadian Center for Occupational Health and Safety, Hamilton, Ontario. October.

CEA. 1997. CO2 Systems Planning and Installation. Comité Européen des Assurances: Paris, France.

29 CFR Part 1910.160 (b) (11). 1994. Fixed Extinguishing Systems. Code of Federal Regulations, September 1.

29 CFR 1910.162. 1994. Fixed Extinguishing Systems, Gaseous Agent. Code of Federal Regulations, September.

46 CFR Part 76.15. 1997. Ch. I - Carbon Dioxide Extinguishing System, Details. Code of Federal Regulations, October 1.

Consolazio, W.V.; Fisher, M.B.; Pace, N.; Pecora, L.J.; Pitts, G.C.; Behnke, A.R. 1947. Effects on man of high concentrations of carbon dioxide in relation to various oxygen pressures during exposures as long as 72 hours. Am. J. Physiol. 51:479-503.

Coward, H.W.; Jones, G.W. 1952. "Limits of Flammability of Gases and Vapors." Bulletin 503, USDI Bureau of Mines: Pittsburgh, PA.

Dalgaard, J.B.; Dencker, G.; Fallentin, B.; Hansen, P.; Kaempe, B.; Steensberger, J.; Wilhardt, P. 1972. Fatal poisoning and other health hazards connected with industrial fishing. Br. J. Ind. Med. 29:307-316.

Dripps, R.D.; Comroe, J.H.. 1947. The respiratory and circulatory response of normal man to inhalation of 7.6 and 10.4 percent carbon dioxide with a comparison of the maximal ventilation produced by severe muscular exercise, inhalation of carbon dioxide and maximal voluntary hyperventilation. Am. J. Physiol. 149:43-51.

Friedman, R. 1989. Principles of Fire Protection Chemistry, 2nd Edition. National Fire Protection Agency: Quincy, MA.

Friedman, R. 1992. Theory of Fire Extinguishment. Fire Protection Handbook, 17th Edition, Ed. A. Cote. National Fire Protection Agency: Quincy, MA.

Gellhorn, E. 1936. The effect of O2-lack, variations in the carbon dioxide-content of the inspired air, and hyperpnea on visual intensity discrimination. Am. J. Physiol. 115:679-684.

Gellhorn, E.; Spiesman, I.. 1934. Influence of variations of O2 and carbon dioxide tension in inspired air upon hearing. Proc. Soc. Exp. Biol. Med. 32:46-47.

Gellhorn, E.; Spiesman, I.. 1935. Influence of hyperpnea and of variations of O2- and CO2-tension in the inspired air upon hearing. Am. J. Physiol. 112:519-528.

Gibbs, F.A.; Gibbs E.L.; Lennox, W.G.; Nims, L.F. 1943. The value of carbon dioxide in counteracting the effects of low oxygen. J. Aviat. Med. 14:250-261.

Gustafson, Matthew. 1998. U.S. Coast Guard Headquarters, Washington, DC, personal communication.

HAG. 1995. "A Review of the Toxic and Asphyxiating Hazards of Clean Agent Replacements for Halon 1301," prepared by the Halon Alternatives Group (HAG) in the UK, February 1995. As cited in a letter dated May 9, 1995 from J.S. Nicholas, Ansul Inc. to Karen Metchis, EPA.

Hansen, Richard. 1999. Fire Program Manager/Project Manager, USCG Research and Development Center, Groton, CT, personal communication.

IMO. 1992. SOLAS Consolidated Edition, 1992, Consolidated Test of the International Convention for the Safety of Life at Sea, 1974, and its Protocol of 1978: Articles, Annex, and Certificates. International Maritime Organization: London, England.

IRI. 1994. Information Manual 13.3.1-Carbon Dioxide System. June 1994. Industrial Risk Insurers: Chicago, IL.

Ishiyama, M. 1998. Nohmi Bosai, Ltd., HTOC representative from Japan, personal communication.

Kety, S.S. and Schmidt, C.G. 1948. The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J. Clin. Invest. 27:484-492.

Lambertsen, C.J. 1971. "Therapeutic Gases -- Oxygen, Carbon Dioxide, and Helium." Drill's Pharmacology in Medicine. Chapter 55, Ed. By J.R. DiPalma. McGraw-Hill Book Co.: New York, NY.

Lechtenberg-Autfarth. 1998. Bundesanstalt Fur Arbeitsschutz und Arbeitsmedizin. (Federal Institute for Occupational Safety and Health), Dortmund, Germany, personal communication. NFPA 12. Standard on Carbon Dioxide Extinguishing Systems. 1998 Edition. National Fire Protection Association: Quincy, MA.

NFPA 2001. Standard on Clean Agent Fire Extinguishing Systems. 1996 Edition. National Fire Protection Association: Quincy, MA. Appendix A, Sec. A-3-4.2.2.

NIOSH. 1976. Criteria for a Recommended Standard: Occupational Exposure to Carbon Dioxide. HEW Publication No. 76-194, National Institute for Occupational Safety and Health, August.

OSHA. 1989. Carbon Dioxide, Industrial Exposure and Control Technologies for OSHA Regulated Hazardous Substances, Volume I of II, Substance A - I. Occupational Safety and Health Administration. Washington, DC: U.S. Department of Labor, March.

Patterson, J.L.; Heyman, H.; Battery, L.L.; Ferguson, R.W. 1955. Threshold of response of the cerebral vessels of man to increases in blood carbon dioxide. J. Clin. Invest. 34:1857-1864.

Schlosser, Ingeborg. 1997. VdS Schadenverhütung GmbH. Koln, Germany, personal communication.

Schneider, E.C.; Truesdale, E. 1922. The effects on the circulation and respiration of an increase in the carbon dioxide content of the blood in man. Am. J. Physiol. 63:155-175.

Schulte, J.H. 1964. Sealed environments in relation to health and disease. Arch. Environ. Health 8:438-452.

Sechzer, P.H.; Egbert, L.D.; Linde, H.W.; Cooper, D.Y.; Dripps, R.D.; Price, H.L. 1960. Effect of CO2 inhalation on arterial pressure, ECG and plasma catecholamines and 17-OH corticosteroids in normal man. J. Appl. Physiol. 15(3):454-458.

Senecal, Joseph. 1999. Kidde-Fenwal, Inc., Ashland, MA, personal communication.

Stronach, Ian. 1999. ALCAN Aluminum LTD, Montreal, Quebec, personal communication.

White, C.S.; Humm, J.H.; Armstrong, E.D.; Lundgren, N.P.V.. 1952. Human tolerance to acute exposure to carbon dioxide. Report No. 1: Six percent carbon dioxide in air and in oxygen. Aviation Med. pp. 439-455.

Willms, C. 1998. FSSA Technical Director, Baltimore, MD, personal communication.

Willms, C. 1999. FSSA Technical Director, Baltimore, MD, personal communication.

Wong, KL. 1992. Carbon Dioxide. Internal Report, Johnson Space Center Toxicology Group. National Aeronautics and Space Administration: Houston, TX.

Wysocki, T. J. 1992. Carbon Dioxide and Application Systems. Fire Protection Handbook. 17th Edition. Ed. A. Cote. National Fire Protection Agency: Quincy, MA.

Wysocki, T. J. 1998. Guardian Services, Inc., personal communication.

Wysocki, T. J. 1999. Guardian Services, Inc., personal communication.