Terrorism and the Chemical Infrastructure: Protecting People and Reducing Vulnerabilities (2006)

The products of the chemical industry are ubiquitous in modern life. Plastics, fibers, drugs—these are a few of the products we encounter everyday in our workplaces and homes that are direct products of the chemical industry. Many more of the products we use daily—paper, fabrics, cosmetics, and electronics—are produced using the products of the chemical industry. The chemical sector is a key part of the national economy: while its products represent only 2 percent of the U.S. gross domestic product,¹ they underpin most other manufactured goods and enable our way of life.

The chemical sector is diverse and wide-ranging. It includes firms that manufacture huge volumes of chemicals intended for many uses—such as major refineries processing thousands of tons of petrochemical feedstocks daily—and firms that produce small quantities of materials with highly specific uses, such as small pharmaceutical companies producing products in gram or kilogram quantities after many days of processing and purifying. The facilities in which chemicals are produced are similarly varied—from refineries covering square miles of land with many high-volume chemicals on-site, to startup companies occupying thousands of square feet in light

industrial parks. Products can be transported to their final place of use by truck, rail, pipeline, or other means in both large and small quantities.

To identify the vulnerabilities posed to the nation to terrorist attack on or other catastrophic loss in the nation’s chemical infrastructure, it is necessary first to somehow succinctly characterize this large and varied sector. This is done below in two steps: (1) a scheme for categorizing the vast number of chemicals produced by the sector and defining these categories and (2) a general model describing the sector’s supply chain.

Virtually all chemical manufacturing, storage, and use in the United States fits into one of the following categories:

• Petrochemicals and fossil fuels

• Inorganic chemicals including fertilizers

• Industrial gases

• Specialty chemicals

• Pharmaceuticals

• Consumer products.²

A description of each of these categories, including a generalized discussion of the manufacture, transport, and use of the chemicals within that category, is given below:

This category entails chemicals produced from hydrocarbon feedstocks, such as crude oil products and natural gas. It includes such chemicals as hydrocarbons and industrial chemicals (e.g., alcohols, acrylates, acetates), aromatics (e.g., benzene, toluene, xylenes), and olefins (e.g., ethylene, propylene, butadiene, methanol).

Manufacture and Use. Most of these chemicals are produced and sold in large volumes—so-called commodity chemicals—and most can be pro-

duced through several different chemical routes or processes. They are used as building blocks in many manufactured products. Many of these chemicals can be replaced by another, with only minor modification required to the user’s manufacturing process and with only minor changes in the performance of the final manufactured product.

Hazard. Most of the products, intermediates, and by-products in this category are highly flammable, and some are toxic (e.g., hydrogen cyanide, hydrogen sulfide, phosgene). Some can form explosive vapor clouds upon release.

Locations of Production. A large percentage of manufacturing facilities for petrochemicals and fossil fuels are located along the Texas and Louisiana Gulf Coast, but significant installations are also found in the industrial areas of the East Coast, Midwest, and California.

Location of Storage and Usage. Nationwide.

Distribution. Refinery products and a number of petrochemicals (e.g., ethylene, naphtha, ethylene oxide, benzene) are transported to other plants via pipeline, barge, rail, or truck for further processing.

This category includes acids (e.g., sulfuric, nitric) and alkalis (e.g., caustic soda, soda ash), chlorine, ammonia, and ammonia-derived fertilizers. It also includes fluorine derivatives (e.g., hydrogen fluoride), phosphates, potash, pigments (e.g., titanium dioxide), and certain metals such as mercury.

Manufacture and Use. Many of these chemicals, such as chlorine, ammonia, and ammonia-derived fertilizers, are produced and purchased in large volumes as commodity chemicals and may rely on natural gas or crude oil as a feedstock. They are used both as building blocks for other manufactured goods and as end products in themselves (e.g., chlorine, ammonia-derived fertilizers). The chemicals in this category can be substitutable, although not always readily; for example, substances other than chlorine gas can be used to purify drinking water, but at a cost of time and money to effect the substitution that may not be acceptable to all communities.

Hazards. While many of products in this category are nontoxic and relatively unreactive such as potash and pigments, hazards found in this category include corrosives such as acids and fluorine derivatives and toxics such as chlorine, alkalis, ammonias, and heavy metals.

Location of Production. Inorganic chemicals are produced in many parts of the United States, but largely in the South and Midwest.

Location of Storage and Use. Fertilizers are stored and used throughout agricultural areas. Other products tend to be stored and used near manufacturing sites. Chlorine is often found stored in or near petrochemical plants where it is produced and stored in large quantities. Chlorine can also be stored in or near water treatment facilities. Other inorganic chemicals are used in mining and many other industries throughout the country.

Distribution. Chemicals in this category are transported to other plants or to end users via pipeline, barge, rail, or truck. Some facilities are able to generate some of these chemicals on-site on an as-needed basis in order to avoid transport and storage.

This category encompasses two general classes: (1) gases used primarily in large quantities as auxiliaries in other manufacturing processes (e.g., refining, petrochemical or steel manufacture), including nitrogen, oxygen, hydrogen, and carbon monoxide, and (2) specialty gases that are produced in smaller quantities to serve the electronics, food, and other industries.

Hazards. Hazards are dependent on the chemical under consideration and the quantity in which it is being used. For instance, nitrogen is a chemical asphyxiant that displaces oxygen at high concentrations, oxygen promotes combustion, hydrogen is flammable, and carbon monoxide is a simple asphyxiant that binds more strongly than oxygen to hemoglobin and is also flammable. Hazards from specialty gases are due typically to toxic, irritant, or asphyxiant properties.

Manufacture and Use. These gases are produced by multiple firms in both large and small facilities.

Location of Production. The first class of industrial gases is produced in large plants, often adjacent to large users, such as refineries on the Gulf Coast. Specialty gas manufacture is more distributed and occurs in smaller facilities.

Location of Storage and Use. High-volume gases can be manufactured at or adjacent to the point of use and are usually stored as liquids in specially designed cryogenic tanks. Both classes are stored in smaller quantities in gas cylinders.

Distribution. High-volume chemicals can be distributed via pipeline or as cryogenic liquids in railcars or tank trucks. Cylinder-sized volumes are transported by rail or truck.

This category comprises a large number of chemicals that are used as aids to the manufacture of other major products (e.g., in paper milling, plastics production, water treatment, mining), are used as end products (e.g., pesticides in farming), or are components of consumer products (personal care products, paints and coatings, adhesives and sealants, photographic chemicals).

Manufacture and Use. Most chemicals in this category are produced and sold in relatively small quantities (grams to hundreds of kilograms). Unlike the categories discussed above, specialty chemicals often have a single manufacturer. Their use is highly specific, but if some deviation in final product performance can be tolerated, most are substitutable.

Hazards. Some specialty chemicals are toxic.

Location of Production. Diversified.

Location of Storage and Use. Diversified.

Distribution. Chemicals in this category are produced in either continuous or batch (smaller, noncontinuous) plants and are shipped to users in 55-gallon drums or smaller containers.

This category includes prescription and over-the-counter drugs, diagnostic substances, vaccines, vitamins, and preparations for both human and veterinary uses.

Manufacture and Use. The large majority of pharmaceuticals have multiple manufacturers; most have available substitutions. A notable exception, however, is vaccines, which tend to have one or two suppliers and may have no substitutions.

Hazards. The production processes used are similar to those in large chemical plants, but pharmaceuticals are produced in much smaller quantities and with much more stringent quality control. The production of pharmaceuticals can entail the generation of combustible dust as well as use of toxic industrial chemicals in relatively small quantities.

Location of Production. Diversified.

Location of Storage and Use. Pharmaceuticals are warehoused in prepackaged units until distributed to end users.

Distribution. Pharmaceutical manufacturing includes the production of pills or other end-products that are then packaged and sent to distributors and pharmacies.

This category entails formulated products, such as soaps, detergents, bleaches, paints, solvents, glues, toothpaste, shampoos, cosmetics, skin care products, perfumes, and colognes intended for direct consumer use.

Manufacture and Use. Although each product within this category is somewhat differentiated from its competitors, all are readily substitutable with products from multiple manufacturers.

Hazards. The contents of consumer products may be toxic, corrosive, or flammable; many have quick skin-bonding characteristics; or their packaging may be pressurized.³ Consumer products can be particularly vulnerable to tampering since their distribution is widespread and under relatively loose control at the retail level.

Location of Production. Nationwide.

Location of Storage and Use. These products are used widely and are commonly found in households and retail outlets nationwide.

Distribution. All of these items are available to the general public and are sold in packaged form in retail stores, by mail order, and on-line.

The description of the chemical categories above includes some specific characteristics of feedstocks, manufacturing, storage, and use for each. From these specifics it is possible and instructive to draw some generalizations about the chemical supply chain. Here that supply chain is presented as a network using a model described by the characteristics of the materials and infrastructure involved; the pathways, links, and nodes between manufacturers and users; and the ownership and control of the components. The model as described applies to most of the chemical industry. Exceptions to the model are noted where relevant. The main characteristics of the network follow.

The chemical industry is materials intensive. Most chemical products can be produced from a variety of starting materials, although many of

those starting materials are dependent upon crude oil or natural gas as a feedstock. The supply network for these starting materials tends to be diversified: in general, a single supplier does not provide all of the starting materials for a final manufactured chemical product. Suppliers of materials may be domestic or international.

The chemical industry is highly capital intensive, requiring substantial facility and equipment investment. Many chemical facilities contain specialized equipment that, if destroyed, would be difficult to replace quickly (e.g., cracking facilities). Most chemical manufacture requires a small number of personnel per square foot of facility. For example, a large petrochemical refinery with a footprint of several square miles may have only a few hundred employees on-site under normal circumstances.

The supply chain for any chemical is characterized by multiple nodes, links, and pathways. A node is a facility at which the chemical is produced, stored, or consumed; a link is the means (road, rail, barge, or pipeline) by which the chemical is transported from one node to another; and a pathway is the sequence of nodes and links by which the chemical is produced, transported, and transformed from its initial source to its ultimate consumer. Dominant nodes are geographic locations in which a substantial proportion of a chemical is concentrated—possibly because of a small number of facilities with large capacities or a large number of facilities with small capacities. Dominant links are similarly defined as links over which a substantial proportion of a chemical is concentrated during its passage through the supply chain. Pathways and links exist in the transfer of materials between facilities, companies, and sectors. Transfer of materials requires transport, usually via rail, road, ship, or pipeline between plants.

The chemical supply chain also has dominant nodes—small geographic areas where large concentrations of products or intermediaries exist, (e.g., the Gulf Coast) or connecting points (dominant links) where many diverse streams in the supply chain converge, (e.g., a pipeline, large natural gas or ethylene storage terminals). Although the industry as a whole is geographically dispersed, some sectors of the chemical industry are geographically

clustered (e.g., petrochemicals in the Gulf Coast), thereby creating a dominant node.

A highly complex and distributed supply chain makes it difficult to shut down the entire system by interruption at a single point. However, determining if an attack at a dominant node or a dominant link of the supply chain of a critical chemical would bring production to a standstill is a high priority in determining overall vulnerability. In order to achieve a widespread interruption in production for an extended period, multiple, well-placed interruptions would have to occur at one or more dominant nodes or dominant links (or both)—in most cases, at multiple geographic locations.

Most of the chemical infrastructure is under private ownership with the exception of certain feedstocks (e.g., oil reserves). Although most chemical companies in the United States are domestically owned, a number of companies are based in other countries with sites in the United States. Most large chemical companies have multiple sites or locations domestically and globally, which allows for highly decentralized ownership and control of the supply chain. Removing a single node or link is not sufficient to disrupt the entire supply chain for a given chemical. However, lack of centralized control may hinder a timely response to a terrorist incident or to a series of terrorist incidents.

At the plant level, many control systems are automated. Because of safety concerns, manufacturers give high priority to ensuring that automation does not exacerbate an emergency. The supervisory control and data acquisition (SCADA) system software is designed to shut down a process when it exceeds safe operating parameters, but it is not the primary safety system in a well-designed chemical production facility. Good process design in the chemical industry also includes safety shutdowns that are independent of the SCADA system. Plants conforming to standards ISA 584.01, International Electrotechnical Commission (IEC) 61508, and IEC 61511 are less vulnerable to catastrophic release due to attack through the SCADA system than those that are not compliant, unless these independent safety shutdown systems—either hard-wired or independent automated systems with no remote access to the outside world—have also been altered. Thus, effecting a release or other disruption through the automated

control system requires not only reprogramming the SCADA software, but also physically infiltrating the site and systematically disabling these other fail-safe mechanisms.⁴ However, sabotage of an automated process remains a possibility, and the controls in place to mitigate the effects of an accident may not be adequate when intentional disruption or destruction occurs if the standards cited above have not been conformed to or if fail-safe mechanisms have not been properly maintained.⁵

Since the products of the chemical sector play such an important, underpinning role in our modern life and the national economy, the economic effects of a disruption to this supply chain must be considered seriously.

Unlike many other sectors that are considered part of the nation’s critical infrastructure, such as utilities, the chemical sector has more than 200 years of experience operating in the free market. Strong competition has ensured that manufacturers have contingency plans in place to meet customer demands in the event of a disruption to their manufacturing capacity. These include product stockpiles, plans to shift manufacturing to other locations, plans to assist customers in shifting their processes temporarily to similar but alternative products, and in the case of large-scale interruption of manufacturing capacity, cooperative agreements between competitors during emergencies to ensure that critical needs for a given chemical can be met. In many cases, customers can simply acquire supplies from domestic and international competitors of their suppliers to fill their need. In cases where a specialty chemical with one or a few suppliers is disrupted, customers can be shifted temporarily to alternative chemicals, which may come at the price of slightly higher production cost or slightly altered product characteristics, but are generally within acceptable standards. To have a cata-

strophic national economic impact or result in catastrophic casualties, the chemical or chemicals whose supply is disrupted would have to (1) be critical to a large portion of the economy or to public health; (2) have a production process that is not readily restarted in a short period of time; (3) be unsubstitutable; and (4) be intended for an end use for which there is no other substitute. In other words, both the chemical itself and the product in which it is ultimately used would have to be unsubstitutable by some other chemical, product, or method of achieving the final end within the period required to drain whatever stockpile of the chemical or product exists. The competitive marketplace in which the chemical sector exists works against such a possibility, at least in the case of civilian products and uses.⁶

Examination of the historical record backs up these assertions. For example, in 1999 two of the three plants worldwide capable of producing hydroxylamine, a key component of specialty chemicals used in semiconductor fabrication, suffered catastrophic explosions and ceased production. In response, chemical firms doing significant business with electronics manufacturers directed their customers to hydroxylamine-free products and processes until hydroxylamine capacity was restored. Although the substitution raised costs for semiconductor manufacturers slightly, fabrication did not suffer a significant disruption. Hurricanes, most recently Hurricane Katrina, have shut down significant portions of the nation’s petroleum and natural gas supply (key feedstocks to the chemical industry), with subsequent economic loss but without catastrophic economic consequences. By the end of 2005, some four months after Hurricane Katrina hit, approximately 25 percent of petroleum and 20 percent of natural gas capacity from the Gulf of Mexico remained “shut in,” down from roughly 50 percent in the weeks after the storm.⁷ The full extent of economic consequences from the disruption to supply remains to be seen as of this writing and will require evaluation in the years to come; the nation has seen higher gasoline and heating costs, and costs of consumer goods are expected to reflect these increased energy prices. Even with this, however, economic consequences from the disruption to the chemical infrastructure (though not from the hurricane overall) have been notable but not catastrophic to date. Katrina pro-

vides other examples of how the chemical infrastructure responds to disruption. For example, cooperative agreements between liquid hydrogen producers were activated to ensure that strategic needs for this cryogenic material were met. Even the worst accidents on record (Bhopal, Toulouse, Texas City) did not result in a situation in which the supply of the chemical or fertilizer in question could not be made available in a short time.

Agricultural chemicals and pharmaceuticals are the areas most likely to be impacted by single suppliers of specialty chemicals. However, major herbicides and fungicides have alternatives that perform as well or almost as well, so that normal agricultural practice can continue should the supply of preferred product be disrupted.⁸ Likewise, pharmaceutical manufacturers typically stockpile two to three months’ supply of their products as a contingency against disruption of their manufacturing capacity. Should a disruption occur that cannot be rectified in that period, doctors have two to three months to migrate their patients to alternative drugs and treatment, which exist for every major category of pharmaceutical on the market. The exception is organism-specific pharmaceuticals, such as vaccines, for which a substitute may not exist.