ERI Home  IRR Home           Last updated 8/10/04

Commodity Chemicals

Impacts, Risks and Regulations

[Note:  Two asterisks indicate a provisional conclusion that will be expanded into more detailed narrative.  Three asterisks indicate an additional topic area, not yet boiled down to a conclusion, that is being prepared for incorporation into the document.]


Mom and Pop are not in the commodity chemicals business.  This subsector is differentiated from the rest of the chemical sector primarily by bulk.  The object is to carry out chemical transformations on a massive scale, producing relatively simple molecules at the lowest possible cost.  Lowering environmental impacts in commodity chemical production involves large capital expenditures and long time horizons.

If the challenges to introducing change into this subsector are great, the potential rewards are correspondingly substantial.  The atmospheric emissions of volatile organic compounds (VOCs) from its operations are second only to petroleum refining in total volume, and its emissions of nitrogen oxides (NOx) are in the top ten.  The chemical sector as a whole is the leading manufacturing sector (at the two digit SIC level) on many TRI emissions categories, including releases to air, transfers to treatment and POTWs, and treatment on and off site, and the commodity chemical subsector is a major contributor to those totals.  For impact reduction, commodity chemicals are the biggest, if one of the most challenging, targets of all the manufacturing sectors.

The chemical industry, accustomed to being in the line of fire, has developed voluntary programs on its own initiative, and has responded enthusiastically (for an industry sector) to certain voluntary programs devised by others, in an attempt to channel public pressure into directions less onerous, and more mutually beneficial, than traditional regulation.  The "Responsible Care®" program, developed in the wake of the Bhopal disaster, has been a model for other sectors, and this industry's level of participation in the EPA's "33/50" program of the early 90s was, on the whole, exemplary.

However, by its very nature, the chemical industry, and the commodity chemical subsector in particular, is likely to remain the long pole in the environmental improvement tent.  It has recently been the subject of criticism from some veteran industry watchers for resting on its earlier laurels, and for having made little progress since its significant gains in the earlier part of the last decade.  As the public's expectations grow, and the floor level for what constitutes acceptable baseline environmental performance continues to rise, the chemical industry will continue to be challenged to find new ways to improve.  And, if recent trends are any indication, it will increasingly be held accountable, not only for its processes, but for the fate of its products. 


Industry profile

Environmental impacts and risks

   Issues List

  Quantitative impact data

Effects of existing and future regulations on impacts

Information sources

Industry profile

The chemical industry is the sector that produces most of the vast array of materials found in modern commerce.  Materials technology has, particularly over the past two centuries, developed into a hugely complex web of capabilities and practices, and the structure of chemical industry is a fair measure of that complexity. 

The first task in analyzing the environmental impact of the industry is to break that overwhelming complexity down into more manageable subcategories.

Types of producers

The chemical sector as a whole can be subcategorized according to several alternative schemes.  Three alternatives are particularly relevant for the subject matter analyzed here:

  1. Categories based on the physical nature of the materials being manufactured
  2. Categories based on the engineering processes used to make the materials
  3. Categories based on the economic constraints associated with the markets in which the materials are being sold

A materials-based scheme is the most appropriate for understanding impacts, and particularly for converting the impacts to risks.  A process-based scheme is probably more appropriate for devising broadly applicable ways to ameliorate the impacts.  Both of these physically oriented schemes are enlightening, but the available data are typically broken down by materials, so much of the following discussion will deal with a materials-based approach.

There are also several applications in which a market-based scheme is useful, such as when considering the economic consequences of introducing regulations, or when designing voluntary environmental improvement programs based on finding common interests.  Some discussion of a market-based segmentation of the sector is offered below.

Materials-based categories of chemical manufacturers: 

Chemical manufacturers can be divided according to what they make.  The processes that they employ to make those materials follow from the laws of chemistry and physics (and the degree to which the state-of-the-art has progressed to take cleverest advantage of them).  Assuming that producers are operating in a regulatory environment that stimulates them to minimize their impacts on the physical environment, the impacts are largely conditioned by the fundamental requirements of the processes.  Our discussion here will focus on those distinctions with the most pronounced environmental consequences (and thus may segment the industry a bit differently from other, perhaps more familiar schemes that are more concerned with other aspects of analysis, such as economics or energy usage).

The first and most fundamental distinction, common to virtually all schemes (and to elementary, undergraduate, and graduate education, for that matter) is the distinction between organic and inorganic chemicals.  The former relates to compounds containing covalently bonded carbon, the latter to all substances encompassing the other 91 naturally occurring elements, as well as those containing carbon that doesn't happen to be covalently bonded to itself or hydrogen.  There would seem to be a lot more scope in the latter category, consisting as it does of materials comprising everything that exists with the exception of those containing one element in a particular molecular configuration, but paradoxically the former category comprises the vast majority of distinct materials in commerce (not to mention all life as we know it).  Carbon is protean stuff.

Although organic chemicals dominate the picture in terms of the variety of compounds, inorganics dominate the very top of the list in terms of sheer tonnage.  In a list of the top fifty chemicals produced in the U. S., ordered by quantity produced (1997 data), eight of the top ten chemicals are simple inorganics.  Among the top twenty, the split is half inorganic, half organic.  But by the time one looks at the entire list of fifty, over thirty are organic compounds.

Among the organic compounds, the most common distinction is between the aliphatic, or straight-chain compounds, and the aromatic, or ring compounds.  This distinction is made because of the peculiar chemical stability of a particular configuration, a ring of six carbon atoms, each contributing electrons to a swirl of negative charge that is free to circulate around the whole ring, forming a unified entity that is broken apart only with considerable difficulty.  This ring, once formed, persists as a motif even when additional groups and chains are added. 

***  [more detail on environmental impacts associated with specific types of materials]

Process-based categories of chemical manufacturers:

Chemical engineers have devised a scheme of "unit operations" to classify the types of production processes used throughout the chemical industry.  While there are tens of thousands of materials to contend with, the number of unit operations is, depending on granularity of the specific scheme chosen, generally more like one or two dozen.

In many cases, the general types of impact associated with specific materials can be predicted from the unit operations used in their manufacture.  For example, one can assume that separative operations involving the gas phase, such as distillation, will involve fugitive emissions to the atmosphere, while separations involving liquid-liquid extractions, particularly those involving the partitioning of the desired material between a water and a solvent phase, will result in wastewater containing solvent residues. 

*** [more detail on environmental impacts associated with specific unit operations]

Market-based categories of chemical manufacturers: 

Chemical manufacturers can also be divided according to how they make and sell their products.  Producers of large-volume chemicals, especially chemicals that are used as feedstocks by other manufacturers, are concerned with producing materials of acceptable quality at the lowest possible cost.  Markets for any given material are large enough to entice other manufacturers into competition, leading to a constant battle to increase yields and efficiency, and to drive down costs.  Continuous processes are preferred over batch processes whenever possible.  Equipment size grows, driving up capital costs.  This subsector of the industry is called the commodity chemical sector.

But if buyers of chemicals are always looking for the lowest possible cost, why has the entire industry not developed in this direction?  Producers of chemicals for which the demand is more limited, or for which the market is fragmented into many different kinds of customer, each with a different set of criteria, are less likely to feel the pressure, or to have the ability to attract the capital, to rationalize its production methods to the highest possible degree.  Manufacturers in the specialty chemical sector are more likely to develop niches in which they excel in meeting specific customer requirements.  Their production methods may be somewhat more amenable to modification to meet environmental performance goals (in those cases when environmental goals conflict with cost minimization), since their profit margins may not be so dependent on the differential between cost of production and market price.

A note on standard industry classification:  Neither the SIC nor the NAIC systems call out "commodity chemicals" as a specifically defined subdivision.  Thus impact data derived from data sources that sort data according to one or the other of these systems will depend on which specific codes are included in the working definition of "commodity chemicals".  The NAIC system explicitly groups together industries that use similar processes.  Although definitions of SIC organizing principles do not seem to be spelled out as explicitly in available sources, it appears, at least for the chemical sector, that the SIC system is more focused on grouping according to similarity of product.  Accordingly, the NAIC system's focus on process would be best suited to differentiating companies according to technical considerations, whereas the SIC system's focus on products would be more aligned with market considerations.  However, we do not always have the luxury of choice.  Some of the available data (such as greenhouse gas emissions data) have been segmented by the information provider into NAICS categories, while other data (such as for criteria pollutants and hazardous air pollutants) have been, at least historically, segmented according to SIC codes.  In what follows, we can only do the best with what we've got.

The market-based definition used in this analysis is as follows:

The SIC category 2869 unavoidably includes specialty manufacturing, but to exclude it from consideration would exclude most of what would normally be considered "commodity" chemical production.  The working definition excludes the NAICS category 325199, All Other Basic Organic Chemical Manufacturing, on the grounds that much of the commodity subsector is covered in the other NAICS categories that are included.  More detailed descriptions of these categories are accessible from the U.S. Census website NAICS page for the 325 code.

[wet scrubbers discussion -- not futile, since pollutant stays put longer and is more amenable to treatment;   inverse of stripping]

Trade and research organizations

Environmental impacts and risks

Issues list

Quantitative impact data

Air emissions data for certain key criteria pollutants (ozone precursors) are available from the National Emission Trends (NET) database (1999), and hazardous air pollutant emissions data are available from the National Toxics Inventory (NTI) database (1996 is the most recent year for which final data are available).  For the commodity chemical sector's SIC codes (2865 and 2869), the total emissions for volatile organic compounds (VOC), nitrogen oxides (NOx) and hazardous air pollutants (HAPs) are as follows (in tons per year):

SIC Subsector VOC NOx HAP
2865 Cyclic Crudes and Intermediates 5,837 5,903 6,329
2869 Industrial Organic Chemicals, not elsewhere classified 125,333 156,338 42,118
2873 Nitrogenous Fertilizers 8,530 53,862  
2874 Phosphatic Fertilizers 201 3,588  
  Total, commodity chemical subsectors (as defined here) 139,901 219,691 48,447
  Total for entire chemical sector (SIC 28xx) 376,812 423,975 452,050

These levels are substantial.  The commodity subsector alone dwarfs the total VOC emissions of every other industry sector at the two digit SIC code level, except for petroleum and paper.  Adding the commodity subsector to the rest of the chemical sector, we find the chemical sector as a whole in a class by itself for VOC emissions, as well as for HAPs.

In terms of greenhouse gas emissions, the chemical sector as a whole (NAICS code 325) is responsible for 491.8 teragrams (million metric tons, Tg) of carbon dioxide equivalent emissions, according to 2000 data.  This is second only to petroleum products (NAICS 324, at 533.8 Tg CO2 equivalent) among industry sectors at the three digit NAICS level as a contributor to global warming.  (These figures include manufacturing processes only, fuel and non-fuel emissions.  They do not include subsequent emissions of greenhouse gases due to products released or burned by customers outside the respective sectors.)  The following table lists greenhouse gas emissions in Tg CO2 equivalent for selected NAICS sectors.

NAICS Subsector GHG, fuel GHG, non-fuel Total GHG
32511 Petrochemical Manufacturing 51.0 1.7 52.7
32512 Industrial Gas Manufacturing 14.8   14.8
325192 Cyclic Crude and Intermediate Manufacturing 6.1   6.1
325199 Other Basic Organic Chemicals 105.3   105.3
325311 Nitrogenous Fertilizers 31.8 31.8* 63.6
325312 Phosphatic Fertilizers 6.7   6.7
  Total, commodity chemical sectors (as defined here) 215.7 33.5 249.2
  Total, all chemical sectors (NAICS 325) 395.2 136.3 531.5

* 16.0 Tg CO2 equivalent from ammonia production (as CO2), plus 15.8 Tg CO2 equivalent from nitric acid production (as N2O);  the equivalence between fuel and non-fuel quantities for this subsector is coincidental.

More information on the calculation of greenhouse gas emissions is provided in another document in this series, Greenhouse Gas Estimates for Selected Industry Sectors.  (Note that the totals given for the chemical sector are defined differently in that document, since for example it breaks out plastics and agricultural chemicals as separate categories.)

]The chemical industry is a major contributor to releases to wastewater, including both direct releases to the environment and indirect releases to publicly owned treatment works (POTWs).  ***

***  Aggregate water data hard to find.  [Use TRI data as a relative indication.]

***  [Solid waste data]


A quantitative assessment of the risks associated with a given material involves combining a measure of the probability of exposure to the material with a measure of the toxicity of the material when exposed.  Like everything else relating to the chemical industry, risk assessment is complicated by the sheer variety of materials produced.

One way to approach the calculation of the probability of exposure would be to considered releases on a material-by-material basis.  There are a number of data sources that could provide input for such estimates.  TRI  environmental release data are reported in terms of specific materials, and would be a good place to start for those materials covered in the TRI list of reportable substances, although it would cover only large sources (facilities whose scale of operations exceeds TRI reporting thresholds) and would not be as valuable for materials that are produced or processed by numerous small sources.  Another potential data source would be estimates of emission factors (quantity of pollutant emitted on average per specific quantity of a given material undergoing a given production step).  Data for many processes are covered in an EPA data source, AP-42, Compilation of Air Pollutant Emission Factors.  Unfortunately, emission factors for many processes have not yet been determined.  But for those that have, one could combine emission factor data with data on the production of a given material to find typical releases.  Such data would be most useful if it were geographically specific.

Detailed consideration of materials is essential in evaluating toxicity, since compounds that are closely-related in molecular structure can differ greatly in their physiological effects.  A summary of the relative toxicity of the products of the chemical industry could (and does) fill libraries as a topic unto itself.  A useful starting point in evaluating human health risks for specific materials in the EPA's Integrated Risk Information System (IRIS).  A corresponding database for environmental toxicity can be found through the EPA's Ecotox database. 

Effects of existing and future regulations on impacts


There are numerous finalized and proposed NESHAP rules that affect either processes or products of the commodity chemical sector.  They include:

A number of organic rules directed specifically toward polymer production are listed in the document in this series on the plastics sector.


Existing Effluent Guidelines relating to the chemical sector are found in the Code of Federal Regulations, Title 40, Chapter 1, Subchapter N, and include:

Solid and hazardous wastes

***  [list of RCRA categorical wastes pertaining to the chemical industry]

Non-regulatory incentives for environmental improvement

***  [Effects of TRI on motivating "beyond compliance" behavior in the chemical industry.]

***  [Voluntary approach -- description and evaluation of Responsible Care® program, sponsored by the American Chemistry Council]

The Future

Turning now toward potential regulatory trends which are just starting to appear over the horizon, one of the most potentially portentous for the commodity chemical sector, but which is probably not yet very prominent on anyone's radar screen, is a logical extension of "take-back" legislation.  These laws, which have been the subject of much debate and some implementation in Europe, essentially shift the onus of responsibility for the environmental impacts of consumer products from the consumer (or society in general) to the producer.  A somewhat related trend, which has made more inroads in this country, has been the development of product liability.  This concept, generally pursued more avidly and successfully through the courts than through legislation, holds producers legally and financially responsible for the consequences stemming from the uses or abuses of their products (whether or not they are used as intended or instructed).

Up to now, both take-back and product liability have applied mainly to discrete parts manufacturers (such as those making automobiles, computers, and the like).  The question is whether the pushing-back of responsibility that is at the heart of both of these concepts will begin to pass back up the supply chain.  That it has not happened to any great extent already is probably due to the fact that the original equipment manufacturers (OEMs) who tend to bear front-line responsibility for the areas in which take-back and product liability have been most active up to now are sufficiently deep-pocketed that they have been adequate targets for pressure groups and plaintiffs.  (This has not been the case in yet another arena in which concepts of legal responsibility have recently been in play:  assigning responsibility in remediation situation according to the "polluter pays" principle.  In that case, it is not uncommon for the relatively small company who has sent waste to a landfill or has had a spill or release to attempt to turn around and pass responsibility back to its suppliers.)

The reason that chemical manufacturers may ultimately find themselves in the line of fire in upcoming producer responsibility conflicts is because they are, in a sense, the "molecular OEMs" -- they are the entities immediately responsible for "building toxicity into the molecule".  And commodity chemical manufacturers especially tend to be the large, deep-pocketed companies that make such tempting targets.

To round out this discussion, we will consider

  1. the risk to chemical manufacturers posed by the extension of producer responsibility up the supply chain
  2. a mechanism whereby the efforts of suppliers to defend themselves against that risk might be channeled in ways that would provide more benefit both to suppliers and to the public interest than would be the case if matters proceed along primarily adversarial lines

Risk:  (Note:  One man's risk is another man's protection.   We are discussing this "risk" in the "Regulations" rather than the "Risk" section of the document because the risk under consideration is to corporate business practices, not the environment.  Furthermore, the effect of this risk is to drive companies in the same direction as regulation would.)

From a manufacturer's point of view, potential liability costs associated with the production of a particular chemical have to be factored into calculations of the product's profit potential.  Any attempt to quantify those costs involves a risk assessment.  In this case, the assessment hinges on the extent to which a producer can be held liable for potential consequences built into a product.  The producer can (at least currently) find some measure of protection by providing, along with the product, detailed instructions for proper use, and cautions against misuse.  But the producer, and no doubt plaintiffs' counsel, must be aware that historically some portion of their output has been, for whatever reason, used improperly, and has caused harm to persons or to the environment.  Should manufacturers be permitted to continue to create these materials with no responsibility for the consequences, in the face of that knowledge?

An analogy may be useful.  At present, gun manufacturers have been largely immune from liability for misuse of their products, on the grounds that consumers are presumably aware that the gun's purpose is to shoot, and the product is functioning as intended.  The responsibility for shooting the wrong target has, at least until now, generally fallen solely upon the shooter.  There is a vigorous effort underway to change this, but the details of this unfolding story are not the center of interest here.  Instead, let us assume for purposes of discussion that we do in fact want to extend responsibility up the supply chain, and ask where that might lead.  Surely the gun manufacturer falls within the scope of the extension.  But what about the bullet manufacturer?  The producer of the powder?  The miner and refiner of the lead?

This analogy actually serves two purposes here.  The first is the demonstration that, just as with life-cycle analysis, one is inevitably faced with the question of how far up the chain to go, and there is generally no obvious, logically defensible way to determine a stopping point.  The second is the observation that the chain sooner or later involves a chemical manufacturer.  This is no accident.  Very few situations in modern life involve only natural, unaltered materials.

Resolution:  There is at present no mechanism to insure the public against the risks posed by specific materials.  Even if a far-sighted public had, at the dawning of the age of modern chemistry, sought to provide such a mechanism, the actuarial data would not have existed to make an insurance-type approach feasible.  Nobody could reasonably have foreseen all of the consequences now familiar today, including such surprises as bioconcentration in food webs, stratospheric ozone depletion, and other complex effects. 

But we are now well into the second century of that age, and, gradually, the data are becoming available.  This is not to say that we have discovered all the surprises that may be in store for us.  (Genetic engineering, for example, is likely to prove unpredictable enough to disabuse us of that notion.)  But perhaps we know enough now to make some reasonable provisions. 

We might imagine the existence of companies or agencies whose business it is to indemnify producers of materials against all claims which might arise from damage done by the material.  Claims might be filed by injured parties, for example, or by government agencies acting on the public's behalf in the case of environmental damage.  An indemnifying organization would defend itself against the claims, or arrive at settlements with the claimants, in much the same way that individual producers are doing now.  But producers would no longer be trying to evade responsibility as their primary legal strategy.  Instead, producers would be formulating strategies that would lower the rates that indemnifying organizations would be charging them for assuming the risks.

As for the matter of how to apportion liability along a supply chain, the workings of the market might provide the means for a practical solution. The overall cost of insuring any given material would be a function of the damage it has the potential to cause, modulated by the actual history of its use and misuse.  As the material is manufactured from starting materials, and as it passes through subsequent companies (being refined, mixed into formulations, etc.) on its way to an end consumer, it gains in value.  Added value calculations are determined by transaction prices, which is one area (perhaps the only one) in which quantified data are virtually always available.  Companies will presumably be willing to buy indemnification against claims arising from their use or sale of that material in proportion to the value that they add to it (a measure of how much that material is worth to the company).  That will establish a baseline apportionment, which would then be adjusted up or down by the indemnifying organization establishing its rate on the basis of the record any particular company has established in handling and selling the material responsibly (or otherwise).

Ideally, this situation would not be established by government fiat.  Life insurance, automobile insurance, and the like are competitive markets, and they work.  Government insurance schemes (such as deposit insurance, for example) have a more checkered history.  Government's best course of action, if such a system as described here is deemed a good way to lessen the negative consequences (particularly including the negative environmental consequences) associated with the creation of new materials, would be to develop reliable data relevant to impacts and risks (quantities released, fate and transport, human and environmental toxicity, etc.) and to make it publicly available.  To its credit, the EPA spends much of its resources doing just that.

Information sources

The EPA Office of Enforcement and Compliance Assurance (OECA) devotes several documents in its Sector Notebooks series to the chemical industry.  Two with particular relevance to this analysis are:

The chemical industry is one of the "Industries of the Future" of the program run by the Department of Energy's Office of Industrial Technologies.  Information on documents produced under that program is available on the web:

One of the OECA Compliance Assistance Centers, "Chemalliance", is devoted to the chemical sector, found at