On January 29, 2003, a massive explosion rocked the West Pharmaceutical facility in Kinston, N.C. The blast killed six plant workers, injured dozens more, cost the economically depressed town more than two hundred jobs, and destroyed millions of dollars worth of property. Such was the devastation that few clues, and sadly fewer close eyewitnesses, survived.
What is known for sure is that the Kinston facility used moistened polyethylene powder to manufacture plastic, medical-grade stoppers and syringes. Armed with that information and knowledge of the plant's construction, investigators pieced together a plausible theory about what went wrong. Stephen Selk, an engineer with the U.S. Chemical Safety and Hazard Investigation Board (CSB), Washington, D.C. and CSB's lead investigator at Kinston, theorized that when the polyethylene powder dried, plastic dust as fine as talcum powder probably migrated upward and settled, over many weeks or months, in the area above ceiling panels. A spark of some sort " from electrical equipment, a static discharge, or some heat source " probably ignited a small quantity of the dust. The shock wave from this initial blast may have dispersed the remaining reservoir of powder, which was ignited by the flame front that followed, creating a much larger, more deadly secondary blast. Selk and others familiar with the West disaster believe that a secondary blast was probably responsible for obliterating the Kinston site.
Dust explosions are more insidious, and potentially more dangerous, than vapor blow-ups. As fluids, flammable gases mix and diffuse through the atmosphere more or less homogeneously, through predictable concentration gradients. As a result, vapor explosions tend to involve a single, relatively localized blast. Because they are solids, dusts and powders settle heterogeneously onto equipment, in passageways, and in nooks and crannies of the plant superstructure. When airborne particles ignite, the blast front propels settled dust into the air. The flame front from the initial fire then ignites the newly airborne material to cause successive explosions. The primary explosion typically damages a limited area; secondary blasts tend to be much more destructive.
The Kinston facility was technically a rubber compounding and forming plant, not a pharmaceutical manufacturing operation. However, the sheer devastation of January 29 should cause every process industry that works with powders to take notice: Any finely divided organic material is a potential dust explosion hazard.
The National Fire Protection Association (NFPA) defines any organic powder with an average particle size of 420 microns or less as a potential hazard. according to Amy Spencer, a spokeswoman for the organization. "The smaller the particle, and the higher its concentration, the greater the hazard," she says.
Recipe for Disaster
With dust, what you can't see is sometimes more dangerous than what you can. Obvious accumulations of dust or powder are routinely vacuumed away at most manufacturing facilities. But according to NFPA, a nearly invisible one-thirty-second of an inch layer of dust can be hazardous. Even if dust present is not sufficient to blow the roof off a building, it can serve as a trigger for other explosions. Dust or powder need not even precipitate the "big bang."
"Dust explosions are usually secondary events," Selk explains. "A small explosion that is not dust-related can send dust airborne, where it gets ignited by the fire front from the first explosion or from an independent ignition source."
Fuel, oxidant, and ignition source are the three ingredients required by all explosions. Almost any organic material and many metals serve as fuel which is why so many different industries need to become aware of the dangers associated with errant powder and dust. Fire safety experts recommend understanding your powders as much as possible: How flammable are they, in what concentrations, and under which conditions? Pharmaceutical processors should take special care in any area where dust and powder are stored, processed, or handled, including process filters, conveying systems, transfer lines and tubes, vacuum systems, mixing and blending operations, filling equipment, and even bulk storage containers.
Know the Risks
No single fire/explosion prevention strategy fits all process situations. Often, engineers adopt a combination of approaches that minimize risk factors (fuel, oxidant, and/or ignition) at appropriate times and locations within a process.
As one safety engineer who preferred anonymity (due to his involvement with the West explosion) notes, engineers first must identify danger zones and the risks each of the three critical explosion factors pose in those areas. "By calculating the risks and costs associated with various mitigation strategies, it's possible to arrive at an explosion prevention strategy that mitigates potential risks intelligently and economically."
Risk assessment can go a long way towards implementing a smart, economic explosion strategy. Bill Stevenson, general manager at safety consultant CV Technology, mentions a Midwestern manufacturer who spent $25,000 per year maintaining an active explosion suppression system on a spray drier, unaware that dust concentrations barely surpassed 20% of the material's minimum explosive concentration, or MEC. "They hadn't even considered that their risk didn't pose any hazard whatsoever," Stevenson comments.
No Red Flags
The most important preventive measure, say safety experts, is awareness of the hazards of handling solids, powders, and dusts they generate. Excipients, actives, sugars, and other organic powders don't scream 'flammable' at the end-user in the same way as ether or hydrogen do. For example, flammable gases and liquids carry U.S. Department of Transportation (DOT) symbols on their packaging indicating hazards, but many bulk solids and powders used in process industries are not classified as flammable.
Similarly, Material Safety Data Sheets (MSDSs) list relevant vapor flammability information for liquids and gases, but those data are frequently missing from MSDSs for solids. What engineers see is usually a cryptic or explanation that the material presents a vague dust explosion hazard, without reference to the conditions under which the hazard exists. Relevant powder/dust properties to look for include MEC as well as minimum ignition energy (MIE), minimum ignition temperature (MIT), maximum explosion pressure (Pmax) and the explosion severity index (Kst). For powder mixtures, the NFPA suggests using values for the finest and driest components, just to be safe. MIE and MIT are particularly important because they determine a powdered material's sensitivity to various ignition sources such as hot surfaces, electrical or frictional sparks. The lower the MIE and MIT values, the more dangerous the powder.
Because pharmaceutical products and intermediates tend to be unique materials, the only way to determine a powder's potential hazard is to test it for flammability and explosivity. For example Fike (Blue Springs, Mo.) is one of a handful of companies that can test, re-create, and simulate explosion conditions using a manufacturer's powdered material. The company generates relevant explosivity and flammability parameters which define a compound's explosion hazard over a range of concentrations and conditions. With that information in hand, process engineers may use a package like Dust Expert from London, U.K.-based DNV Software to design explosion prevention and mitigation systems.
Three Steps to Safety
Perhaps the easiest way to prevent explosions is also the most obvious: minimize fuel concentration. For process industries that generate dust, NFPA recommends keeping dust concentration below 25% of the minimum explosive concentration through use of local exhaust ventilation. Although airborne dust is less of a problem within punctiliously clean manufacturing settings, localized dust accumulation can occur within smallish processing areas. Good housekeeping and local exhaust ventilation goes a long way towards reducing fuel in these locations.
Eliminating the oxidation source " oxygen in air " is problematic but possible. In some processes, the fraction of oxygen in air, 21%, may be reduced by the addition of nitrogen, carbon dioxide, or other inert gases. NFPA 69 recommends oxygen concentrations that are two percentage points lower than the limiting oxygen concentration (LOC) " the minimum oxygen concentration required for a material to combust. For many organic materials, that means oxygen levels in the 8-12% by volume range. This standard only applies when processes are continuously monitored. With occasional monitoring, NFPA suggests going down to 60% of the LOC, and even lower when the LOC is less than 5% by volume.
Inerting can be a risky business due to the potential for asphyxiation, especially in situations where workers stick their heads into pipes and equipment for extended periods. According to NFPA eighty nitrogen asphyxiations have occurred in the United States over the past decade.
Expensive options like inerting are generally reserved for high-risk operations. Low-risk, low-value operations might benefit from less costly solutions such as explosion protection, of which three major types exist: explosion containment, explosion suppression, and relief venting. Explosion isolation, a separate strategy, may be used with all three forms protection schemes.
Light my Fire
Ignition is a tricky condition to control for powder processes. Even after eliminating obvious conditions such as open flames, smoking, and unattended heating equipment, manufacturing environments abound in ignition sources. These include sparks from machinery, static electricity, foreign objects that drop into powder containment or processing systems. Even when all these bases are covered, under the right conditions adventitious sparks resulting from impact or friction between two hard surfaces (e.g. piping, metal equipment, cement floors, fixtures) can be just as dangerous as an open flame.
Controlling ignition sources often involves strategically shielding electronics, switches, and other electrical sources of sparking. In some situations electronics can be installed in a safe area, or within purged or explosion-proof boxes.
In the case of instrumentation, an increasingly popular option to the heavy explosion-proof box is the intrinsically safe circuit. In this approach, a barrier, installed in a non-classified area, limits the circuit current such that even an errant screwdriver cannot generate a spark.
Bulk bags, also known as flexible intermediate bulk containers (FIBCs), have become popular in pharmaceutical processing, as they have been for low-value materials for many years. When most woven plastic bags are filled or discharged, static charges build up which can ignite the dust clouds generated during material transfer. Processors therefore turn to conductive bags whenever an explosion potential exists.
Type C conductive bags are made from interwoven carbon and polypropylene fibers linked together "like a Faraday cage," according to Malcolm Ranson, director of international sales at powder handling system manufacturer Spiroflow in Monroe, N.C.
When the bags are properly grounded, static charges are dissipated to earth. Type D bags are made from a conductive fabric which dissipates static continuously through a mild corona discharge. "If you don't use an antistatic bag and earth it properly, you've wasted all the money you've invested in zoning your electrical equipment," Ranson observes.
Keep a Lid on It
Because pharmaceutical processors have long recognized the potency and toxicology of their powders, modern drug-making facilities are significantly freer of explosive dust than, say, grain or confection process plants. Containment strategies mitigate most pharmaceutical dust and powder mishaps.
"Pharmaceutical processors don't typically have dust from active ingredients floating around," Spiroflow's Ranson notes. However, local powder accumulation and dispersal during transfer, storage, mixing, and general handling operations represent potential trouble spots, even in the most scrupulously clean facilities. In certain situations a fire or explosion that begins within a contained processing area can spread due to the combined effects of a pressure wave, which spreads combustible material beyond the confines of a container or processing suite, and the subsequent flame front which ignites everything in its path.
Dust explosions can pack a wallop, with pressures in the 8-10 bar range, which is often considerably higher than the strength of powder handling machinery. When they blow in unprotected areas, the initial explosion may multiply several-fold in strength and destructiveness by propagation through space or within ductwork.
Every process container has some built-in explosion containment capability. High-risk operations might therefore incorporate equipment built on the estimate that an explosion will generate between eight and ten atmospheres of pressure. Other equipment might employ relief venting " panels that "blow" during an explosion, venting explosion pressure and product to a safe location and thereby protecting both equipment and personnel.
Suppression systems use pressure detectors that pick up explosions at the earliest stages, before the "big bang." Pressure changes trigger a canister to release fire suppression agents which extinguish the fire/explosion before damage occurs.
Explosion isolation has a specialized meaning in the world of explosion prevention. Explosion isolation keeps shock waves and flame fronts from propagating up- or downstream from the source to connected equipment. Isolator systems use explosive-triggered (actually pressure-triggered) gate valves.
Myth or Reality?
Given drug-makers' fetish for cleanliness, how much of a threat do dust explosions actually pose, and how frequently do they occur in our industry? Estimates for non-pharmaceutical industries vary, from one per week to one per day in Europe alone, with similar frequency in the United States.
FM Global, which insures one-third of all Fortune 1000 companies, says that, between1992 and 2002, FM Global's pharmaceutical and chemical clients reported a combined total of nine dust explosions, resulting in $32 million in losses. What percentage of those events occurred in pharmaceutical plants vs. chemical facilities? Stephen Zenofsky, who provided the numbers, said he could not break them down by industry. But citing the industry's penchant for cleanliness, he said he wouldn't be surprised if pharmaceutical incidents were at the very low end of the single-digits during that time period, or even zero.
However, just because no major drug manufacturing plant has suffered the same frightful fate as West's rubber compounding facility doesn't mean explosions don't occur, comments Bill Stevenson of CV Technology. Industry-wide attention to cleanliness and safety, combined with lower material use compared, say, to grain or cocoa processing, means explosions are smaller and usually confined to single vessels. "Minor explosions often go unreported, and even for significant events it may not be clear from the evening news that they're reporting a dust explosion," he notes.
Given the devastating consequences of dust explosions, and how common these accidents are in other industries, drug makers should be proud of their safety record. When it comes to explosion prevention, erring on the side of caution is no mistake.