Chlorine has been used in water treatment for more than 100 years, and of the 63,000 water treatment plants in the United States and Canada, more than 98 per cent Chlorine or its metabolites are used on plants (Whitfield and Gray, 2009, para. 5; para. 9). The simplicity of using chlorine and its derivatives for this purpose is one of the wonders of modern chemistry: it’s cheap, it’s safe, and it works. For larger applications, gaseous chlorine and sodium hypochlorite are used, and sometimes-dry calcium hypochlorite briquettes find their way into small operations outside the normal use of pools and wellheads (Brennan, 2006).
Chlorine has a Variety of Uses in a Water Treatment Plant
It is used on water intake structures for the removal of aquatic organisms; as pre-filtration to oxidize metals such as iron and manganese for removal, and to kill algae and bacteria. (Please refer to the Appendix for a more comprehensive list of pathogens to which chlorine is applied.) Most importantly, when used in the disinfection of drinking water, its residual effects continue through the system preventing further contamination once the water leaves the treatment plant.
Volbeda (2006) presents a background discussion of chlorine chemistry, but, when chlorine gas is dissolved in water, it creates hypochlorous acid. Combining sodium hypochlorite (bleach) with water produces hypochlorous acid and the hypochlorite ion. The volatile nature of the chlorine molecule constitutes both its disinfection power and its danger as a chemical. Chlorine doesn’t care where it gets its eighth electron and will rip it out of any constituent in the water, bacteria and viruses included. Chlorine combines with anything and even in combination can be toxic to bacteria and viruses. While the disinfection process is slowed when using a combined chlorine product or byproduct for disinfection, disinfection still occurs due to the toxic nature of chlorine.
Chlorine is a greenish yellowish gas that is 2.5 times heavier than air. In its gaseous form is extremely toxic and dangerous. Kerri succinctly presents the dangers of chlorine gas:
It has a very high coefficient of expansion. If there is a temperature increase of 50 degrees, the volume will increase from 84-89 per cent. This expansion could easily rupture a cylinder or line full of liquid chlorine. For this reason, all chlorine containers must not be filled to more than 85 per cent of their volume. One litre of liquid chlorine can evaporate and produce 450 litres of chlorine gas (Kerri, 1989, p. 261).
With this vivid picture in mind, process engineers must carefully assess the risks of using chlorine in its various forms near municipalities. Industrial accidents at water treatment plants are fairly rare and usually quickly contained; the dangers of chlorine are most manifested in railway accidents, which put thousands of people at risk.
Hazards and Risks
Guidelines set forth for water treatment professionals illustrate the process of using chlorine gas in a treatment plant (Kerri, 1989). Under vacuum supplied by different forms of induction machines, chlorine gas is fed into water treatment device. The entire chlorination piping system should be under vacuum to ensure safety. Chlorine tanks have a system of regulators to reduce pressure from the tank, feed equipment to measure feed rate accurately, and vacuum educators to deliver the chemical into the receiving water. Piping connections must be sealed with proper pipe thread compound to ensure pipe joints are not subject to chlorine attack, and compression fittings must be sealed with a new lead washer. When the system is first started, and each time an empty chlorine tank is switched for a new one, a simple ammonia check should be performed. Ammonia and chlorine combine to make a white smoke, indicating a leak. At each point in the process, automatic and manual overrides exist if a leak is detected at any point.
Chlorine gas scrubbers should be installed in any facility that uses chlorine gas. These scrubbers are designed to handle the volume of one tank in the event of a rupture. A leak detection system turns the scrubber on when a leak is detected, and the system scrubs one tank-volume worth of chlorine gas into a sodium hydroxide solution. The resulting combination of chlorine and sodium hydroxide is a 10 per cent bleach solution which can then be used as a liquid form of chlorine.
The Environmental Protection Agency requires wastewater plants which store 2,500 pounds or more of chlorine gas to prepare a risk management plan (Stephenson, 2007). Risk reduction begins with using the smallest cylinders possible of chlorine gas for the application. Using 150-pound cylinders minimizes the risk of tank rupture. Keeping in mind that one litre of chlorine gas under pressure will expand to 450 litres, a rupture will make a very large cloud of toxic gas but is the smallest in terms of the container used. As demand and facility size increase, cylinders are manifolded together; a 150-pound cylinder is not appropriate for most municipal water treatment applications. A majority of water treatment plants use ton cylinders manifolded together.
Extreme care is taken at every point in the system to ensure that a tank rupture or leak does not happen. Omaha Metropolitan Utilities District recently redesigned the municipal water treatment system to provide for the efficient use of liquid withdrawal methods from ton cylinders. Rather than expanding the scrubber system, they installed actuators on each of the 24-ton tanks in the facility (Koenig and Slaydon, 2008). Using systems based on this model, water treatment plants can manifold as many ton containers as necessary while controlling for leaks at each container and throughout the entire system. Daily, weekly, and monthly maintenance checks and safety drills of manifold systems are necessary for conjunction with using technology to control for accidents automatically.
Minimizing the Risks of Using Chlorine for Water Treatment
When using gas chlorination, the location of the water treatment plant becomes important. The water plant should be located as far out of the city as possible, downwind of the prevailing winds so that should a rupture occur and the scrubbing system fails, the prevailing winds will dissipate the killing cloud of mustard gas away from population centres. Planners must not only assess the current risk to the population but must also account for planned growth in a city. It is possible with proper planning and safety protocols to have a gas chlorination water treatment facility in the middle of a population centre, though this situation is not ideal.
In addition to sizing and locating a municipal water treatment plant correctly, booster systems at strategic locations may also help reduce risks (Tryby, Boccelli, Uber, and Rossman, 2002). Booster systems are a common practice, but protocols regarding their safety and efficiency are not always rigorous. A network beginning at the main treatment plant pipes treated water to various parts of the system, where population demands determine the need for booster nodes. A model such as this spreads out the risk of chemical spills, reduces the need for onsite chemical storage at any one location, and allows for the more economical use of alternative technologies. On the downside, booster stations require more person-hours for maintenance and inspections, and more equipment overall to operate a large system of nodes.
Risk assessment software such as Phast (developed by the Det Norske Veritas Corporation in Oslo, Norway) provides planners and retrofitters with a tool to determine various levels of risk. According to the DNV North America website (2009), users input data into the tool, and it calculates models showing various effect distances based on hazardous event parameters. The software contains more than 1,600 chemicals and covers the results of leaks, ruptures, and equipment failure; it is used in a variety of industries, including maritime applications, petroleum, and water treatment. Risk assessment professionals can utilize a wide range of options available through this tool to determine the worst-case and more reasonable risks.
Worst Case Scenarios vs Real World Accidents
Recently, rail companies have petitioned the Obama administration to refuse to transport toxic chemicals of all types, including chlorine, through population centres (Frank, 2009). More than three million tons of chlorine are shipped by rail annually in the United States. The railroad cites “remote but deadly risks” in its petition, and are more motivated by liability protection than the actual risks, which would indeed be deadly if a rail car carrying chlorine gas ruptured (para. 7). The rupture of a ton cylinder could potentially produce a cloud one mile high by a half-mile wide by 1 mile long of toxic mustard gas that will kill everything in its wake. There are indeed catastrophic risks involved with producing, storing and transporting large amounts of chlorine, but fictitious worst-case scenarios created while planning for emergencies do not accurately reflect reality (Logomasini, 2003). As far back as 1987, risk assessors called for more thorough inspections of rail cars to prevent derailments and the use of technologically advanced steel and double-layered hazmat protection when transporting chlorine gas (Swoveland, 1986). In the incident which sparked Swoveland article, a train in Ontario derailed and a tank car of chlorine gas ruptured; if there had not been a large propane fire funnelling the heavier-than-air mustard gas upwards into the atmosphere, many thousands of people in the city of Mississauga might have died or been injured.
Despite the almost-catastrophic outcome of the Ontario incident, real-world accidents are rare and usually contained within moments of occurrence. In a 2002 incident in southern California, plant operators had just switched two empty ton containers with two full ones, following all plant protocols. An automatic leak detection system sounded an alarm, the system was automatically closed, and operators also manually closed the system using backup protocols. The valves were closed within three seconds of leak detection, and the entire incident was avoided in less than ten seconds. Approximately one pound of chlorine gas was released into the container room (Connell, 2002). Operators used SCBA units with full chemical suits and an emergency leak kit to find the source of the leak and fix the problem.
The Agency for Toxic Substances and Disease Registry has maintained a Hazardous Substances Emergency Events Surveillance system for more than 20 years. Still, this state-based system is not universal across the United States (Hall, Haugh, Price-Green, Dhara, and Kaye,1996). A single state agency coordinates emergency personnel such as firefighters, first responders, hospitals, and municipal planning agents. A total of 3,125 hazardous substance releases were recorded from 1990 to 1992; of these, 128 involved chlorine as a single substance accidentally released, and approximately 30 per cent of those incidents required evacuations of any sort. None were considered catastrophic, and none resulted in a death directly related to chlorine gas release.
Currently, water treatment plants have a variety of means to disinfect water: chlorine, heat, ultrasonic waves, iodine, bromine, sodium hydroxide, and ozone. Chlorine is by far the cheapest form to use, most widely accepted and has the fewest risks to public health. It also leaves a “residual” in the water that can accurately be tested to ensure adequate disinfection according to good water treatment practices. Additional technological alternatives to using chlorine and its derivatives are under consideration, such as UV treatments, nanofiltration, and modular ozone generation. At this point, these technologies are novelties and are incredibly expensive to retrofit or to develop from the ground up (Whitfield and Brown, 2009). The Government Accountability Office estimates that converting from chlorine gas to alternative methods would cost wastewater plants anywhere from $650,000 to $13 million, depending on the size of the plant, and major municipalities such as Washington, D.C. and Cincinnati, Ohio, could face costs upwards of $2 billion (Stephenson, 2007).
This is not to say that process engineers should not explore these options—developing new technologies could reveal a method that works better than current methods or that could be used more cheaply in the long run. For now, despite its risks, chlorine is the chemical of choice for water treatment, for very good reason. Other technologies should be developed and implemented as costs go down.
It seems that the main risk in using chlorine in water treatment does not arise at the plant itself. Leak detection systems, automatic and manual shutoffs, safety protocols, inspections, and so on contribute to minimizing the risks of a gas leak at a treatment plant. Rail transportation of chlorine gas presents a much bigger worst-case scenario and real-world hazard. Risk reduction should begin in this area, while water treatment plants continue to refine their operations now and in the future.
Appendix: Pathogenic Diseases Removed Through the Use of Chlorine and Derivatives
Disinfection of public drinking water is vitally important for public safety. The following are some examples of pathogenic diseases transmitted by water (Kerri, 1989, p. 257).
- Salmonella (salmonellosis)
- Shigella (bacillary dysentery)
- Bacillus typhous (typhoid fever)
- Salmonella paratype (paratyphoid)
- Vibrio cholerae (cholera)
- Coxsackie Virus
- Echo Virus
- Infectious Hepatitis
- Entamoeba Histolytica (amoebic dysentery)
- Giardia Lamblia (giardiasis)
- Ascaris Lumbricoides (Giant Roundworm)
- Brennan, J. (2005, March). Utilities taking a new look at dry chlorine. WaterWorld 21(3): 14.Connell, G. (2002, December). Emergency preparedness pays off during chlorine leak. WaterWorld 18(11): 18 (2 pp.)
- DNV (2009). Products: Safeti and Phast software description. http://www.dnv.com/services/software/products/safeti/index.asp
- Frank, T., (2009, May 20). Rail industry petitions to reduce toxic cargos. USA Today, News section, p. 3a.
- Hall, H., Haugh, g., Price-Green, P., Dhara, V., and Kaye, W. (1996, June). Risk factors for hazardous substance releases that result in injuries and evacuations: data from nine states. American Journal of Public Health 86(6): 855-857.
- Kerri, K.D. (1989). Water Treatment Plant Operation: A Field Study Training Program (Vol. 1). California State University, Sacramento: Hornet Foundation.
- Koenig, M. and Slaydon, G. (2008, May). System offers secure control of chlorine tone containers. WaterWorld 24(5): 44-46.
- Logomasini, A. (2003, May 21). Supporting a risky water policy. Washington Times, Commentary, p. A14.
- Stephenson, J. (2007, March). Securing wastewater facilities: Costs of vulnerability assessments, risk management plans, and alternative disinfection methods vary widely. GAO Reports, GAO-07-480.
- Swoveland, C. (1987, July). Risk analysis of regulatory options for transport of dangerous commodities by rail. Interfaces 17(4): 90-107.
- Tryby, M., Boccelli, D., Uber, J., and Rossman, L. (2002, September). Facility location model for booster disinfection of water supply networks. Journal of Water Resources Planning and Management 128(5): 322-333.
- Volbeda, J. (2006, November). Continuous monitoring of chlorine helps improve plant performance. WaterWorld 22(11): 28 (3 pp.).
- Whitfield, R., and Brown, F. (2009, July 27). The bottom line. ICIS Chemical Business 276(3): 32-33.