It has been known since the studies by Dodds and Lawson in the thirties that some industrial chemicals can act on the hormonal system and cause endocrine disruption. Today several hundred chemicals are known to have endocrine effects. The majority of these are phenolic or aromatic compounds.

But it was not until a few years ago that a link was established and discussed publicly between the reproductive damage which has been known for some time to occur in animals in the wild, epidemiological findings on an increase in testicular cancer, genital anomalies and changes in the human sperm count and quality, and a possible impact of "environmental hormones". The endocrine effects of even low concentrations of chemicals in the environment at sensitive development stages of organisms and the compound effect of different substances, which remains to be clarified, seem to pose particular problems in this connection.

To verify the current hypotheses linking chemicals in the environment with disturbances in the endocrine system, we must investigate not only the mechanisms by which the substances take effect and their potency, address not only their occurrence in the environment and compile a more systematic record of their effects on animals and humans, but we also require information on the material flows of compounds which have been found to have an endocrine effect in experiments.

The present study addresses the material flows for 1995 of three such (groups of) substances which are used as industrial chemicals and produced in volumes between 10 and 200 kilotons per annum in the Federal Republic of Germany. The substances are bisphenol A, which is used mainly as a precursor for polycarbonates and epoxy resins, dibutyl phthalate (DBP) and benzyl butyl phthalate (BBP) , two plasticisers of which two thirds are used to soften PVC, and nonylphenola starting substance for phenol resins but mainly for nonylphenol ethoxylates, emulsifiers with a surfactant effect.

The study investigates for what purposes and in what volumes the compounds are produced, what products contain them, where they are likely to cause environmental emissions and what are their disposal routes.

Official production figures based on information supplied by companies provide only an incomplete picture of the volumes of the industrial chemicals produced, the structure of use and disposal paths. These data were collected and examined for the material flow analysis in direct searches at the Federal Statistical Office . However, the main body of data used in the study refers to production, processing and application of these chemicals by companies which produce them, trade in them (or products containing them) or consume the products as "end users". The study also draws on information from industrial associations.

The main problem was to break down the major fields of application by industrial sector and product groups. However, relatively consistent data were obtained by combining figures given by producers with those from companies which supply the various product markets. It is essential to break down the material flows by product groups if we are to localise the sources of emissions. All the substances studied have in common, though to very different degrees, that it is the products containing them which are actually the main source of emissions.

The emissions data are likewise based on figures supplied by producers or on estimates from these figures, and on information from experts and the literature. Empirically substantiated studies of emission flows for the compounds considered here have been undertaken only for specific applications. No information was available for many applications. The figures therefore refer only to identifiable emissions and must be considered lower estimates.

The quantitative information on disposal paths refers to annual consumption of the respective substances in the manufacture of products consumed within Germany, not to the actual number of products in use which, as far as durable plastic products are concerned, considerably exceeds annual output. The disposal figures are very rough estimates of the quantities involved and are likely to change as the Technical Instructions on Household Waste take effect.

The study of material flows for all chemicals was based on a five-stage approach. and considers (1) domestic production by volume, (2) domestic consumption for production, i.e. the quantities processed within Germany, (3) domestic product consumption, i.e. the quantities contained in final products consumed in Germany, (4) the amounts for disposal, and (5) the identifiable emissions from the entire production process (manufacture, processing, transport) and from products.

All three groups of substances are imported and exported on a large scale, whether as feedstocks or as processed (semi-finished and final) products. Compared with the volumes produced in Germany, the result is a net outflow at all stages. The study also considers and estimates these imports and exports as far as possible.

Like most environmental chemicals with endocrine effects, all three of the substances studied for their material flows are aromatic compounds. Bisphenol A and nonylphenol are phenol derivatives, while the two phthalates are esters of 1,2-benzene dicarboxylic acid.

Clear evidence of slight oestrogenous impacts has been found in in vitro and in vivo experiments with bisphenol A and nonyl and octylphenol. Similar findings for the two phthalates are known only from in vitro experiments which are considered contradictory and could not be reproduced in vivo. (Multi-generation toxic effects of BBP and DBP found in in vivo experiments are not considered specific proof of an oestrogenous effect, but they are consistent with this hypothesis.) The oestrogenous potencies of the three substances established in vitro are of the order of 10-4 - 10-6 times that of 17b-oestradiol.

Nonylphenol has a toxic effect on all the organisms tested. DBP is classified by manufacturers as toxic to reproduction, and BBP as possibly toxic to reproduction. Bisphenol seems to be hardly toxic at all to microorganisms but toxic to higher organisms in many cases. These are the reasons why nonylphenol, bisphenol A, DBP and BBP were included in the priority list pursuant to Regulation 793/93/EEC on the evaluation and control of environmental risks emanating from existing chemical substances.

The three substances vary considerably in terms of degradability. Bisphenol is readily biodegradable but poorly degradable abiotically. Dibutyl phthalate and benzyl butyl phthalate, being comparatively short-chained phthalates, are broken down relatively quickly both in the presence and the absence of air. Nonylphenol is persistent. It may be released in the biodegradation of nonylphenol ethoxylates.

Turning to the three substances' distribution, all of them are poorly or only moderately soluble in water. Since the polar properties of alkylphenol ethoxylates make them readily soluble in water, they give rise to waterborne emissions of nonylphenol. All the compounds studied have a high affinity with organic matter (octanol-water distribution). Where emissions occur, all the compounds are likely to bond to water and sediment, in ascending order from bisphenol A to nonylphenol and the phthalates.

Studies in Germany indicate that the phthalates and nonylphenol are aquatically significant. There are virtually no findings available for bisphenol.

The mass flows for all three substances studied are given in Table 1. The figures refer to all life cycle phases - production, processing into products, consumption in final products, emissions and disposal paths - for Germany in 1995.

"Production" is defined here as domestic output from the feedstock. "Processing" refers to the domestic input of feedstock in semi-finished and finished products, i.e. productive consumption. The difference between the two is accounted for by the balance of import and export. "Final products" refers to the consumption of the converted substances in finished products for use within Germany (product consumption). Here again, foreign trade gives rise to a difference in the volumes encountered in the previous stages. "Processing" and "final products" include not only products manufactured in Germany but also, in some cases, imports of feedstocks for processing and imports of final products for consumption in Germany. Since no estimates of emissions were possible for a number of products or for the disposal stage in general, we refer to "identifiable emissions".

Bisphenol A was produced on a scale of 210,000 tonnes in Germany in 1995. The only two producers are Bayer AG (Krefeld-Uerdingen) and Dow Deutschland Inc. (Rheinmünster). German output amounts to approx. 50 per cent of the west European total. Only small quantities of bisphenol A are imported, and only about 10 per cent is exported for use as a feedstock. The greater part of the feedstock is used in Germany. However, exports of processed products are higher. Allowing for exports and imports, only 57 per cent of processed bisphenol remains in products consumed in Germany.

In 1995, dibutyl phthalate and benzyl butyl phthalate were produced in amounts of 21,600 and 9,000 tonnes, respectively, or just less than 31,000 tonnes together. In 1995, DBP was supplied by three companies (BASF AG, Buna GmbH and Hüls AG), while BAYER AG was the sole producer of BBP in Germany. Taking both phthalates together, domestic consumption in final products amounts to approx. 70 per cent of the volume produced.

NonylphenolNonylphenol was produced only by Hüls AG. Imports of the feedstock are very low, exports considerable at 40 per cent of output. Of the nonylphenol processed in final products in Germany, the same proportion is exported in these products, so that less than one quarter of the feedstock remains in Germany.

A greater proportion of nonylphenol is exported than of any other substance in the present study.

Table 2 gives a breakdown of the main semi-finished and finished products made with the feedstocks in Germany. The figures refer to the input of feedstock.

Bisphenol A (BPA) is a polymerisable starting product used in the production of plastics. Almost all the BPA produced in Germany is used in polycarbonate (approx. 70 per cent) and epoxy resin (approx. 30 per cent). Non-polymer BPA is used as an additive for a number of special purposes (colour development component in thermal paper, antioxidant in high-temperature cables and tyres, reactant in the production of tetrabromide bisphenol A, a flameproofing agent).

The bisphenol A contained in polycarbonate and epoxy resin is chemically bonded. The residual monomer content is said to be low (ppm range). Polycarbonates are very stable plastics used for many different structural purposes (disks, sheeting, tubular products, food containers, medical purposes, etc.). Epoxy resins are liquid resins which cure to form hard, insoluble, chemical-resistant plastics when hardening agents are added. They are used particularly as adhesive, varnishing and casting resins. The main applications are surface coating, including the inside coating of metal packaging (food and beverage cans). BPA is also the starting product for the manufacture of dental composites resembling epoxy resin (filling and sealing agents).

Used as an additive in thermal paper, in high-temperature cables and rubber tyres, BPA does not form a chemical bond with the host material and is therefore released more easily.

Dibutyl phthalate and benzyl butyl phthalate are two so-called extraneous plasticisers which are admixed with plastics to modify their flexibility and viscosity. They, likewise, do not form a chemical, but only a physical, bond with the host material.

DBP is a relatively volatile, short-chain phthalate used mainly (65 per cent) as the fast fusing plasticiser component in PVC, i.e. the gelling agent which reacts fastest with the PVC. Its volatility has led to DBP being substituted by other phthalates in some cases. Paints, dispersions and adhesives account for 30 per cent of DBP applications. DBP improves the temperature-dependent and mechanical properties and the flowability of paint. In dispersions (coating material, paints, casein glue) it is used, among other things, to improve film formation. DBP raises the elasticity and flowability of adhesives, notably those based on polyvinyl acetate.

BBP is a relatively expensive special-purpose plasticiser which is likewise suitable for use as a PVC gelling aid and increases plastics' resistance to ageing. In floor coverings, the domain where BBP is used in the PVC sector (60 per cent of applications), it is also the preferred phthalate because it repels dirt. Polysulphide sealants are the second main use of DBP and account for approx. 30 per cent of the total. These sealants are used in double glazing, concrete buildings and elsewhere in the construction industry where soils and groundwater must be sealed against chemicals. BBP is the standard plasticiser used here. Its "bulk" means that it is bonded relatively firmly in the polymer matrix.

Both phthalates are also added in small amounts to produce plasticiser mixtures for numerous other plastics applications.

Nonylphenol accounts for approx. 70 per cent of alkylphenols produced in Germany. A small proportion (13 per cent) is processed to produce nonylphenol resin. Such phenol resins are used as tackifying agents in the rubber industry, to increase the tenacity, bonding strength and heat resistance of adhesives, particularly as a curing agent for epoxy resins used mainly for the inner coating of metal packaging.

The largest other application of nonylphenol by far is the production of the commercially and technically important trisnonylphenyl phosphite (TNPP) which has been cleared for food contact applications and serves as a co-stabiliser in PVC and other plastics.

More than 80 per cent of nonylphenol is used in ethoxylation. Alkylphenol ethoxylates (APEOs) are nonylphenol compounds almost without exception. APEOs are non-ionic surfactants with an emulsifying and dispersing action, which makes them suitable for a very large variety of applications.

At the processing stage, approx. 50 per cent of APEOs are used as emulsifiers for emulsion polymers based on styrol butadiene, styrol acrylate, pure acrylate or PVC systems. The range of products includes plastic coatings (paper coatings, textile coatings/carpet backs), dispersion paints and varnishes, adhesives, sealants and similar products. The emulsifier content of such dispersions varies greatly and averages approx.1.5 per cent.

The other half of APEOs are used in a wide range of product groups, although here, too, the emphasis is on their emulsifying effect. In pesticides they serve as formulation aid (spray concentrates), as they do in construction chemicals, such as air-entraining and foaming agents added to concrete, in mould release agents, wax and bitumen emulsions. In the petrochemicals industry, they are used as additives, emulsifiers and dispersing agents in cooling lubricants, lubricating and hydraulic oils and in engine oils. Other uses include offshore chemicals, textile and leather treatment, industrial cleaners, auxiliaries in the manufacture of plastics, paints and varnishes, auxiliaries in the paper and pulp industries, coagulation agents in sewage treatment plants and in medical and veterinary products.

Table 3 presents a survey of uses based on final products sold in Germany in 1995. The APEOs are listed and broken down as a separate group. Quantitative differences between the figures in Tables 2 and 3 are attributable to imports and exports.

While bisphenol A and the phthalates reveal no substantial qualitative differences in their breakdown of use in the processing (productive consumption) and final product (product consumption) stages (at least as far as we can assume, given that no detailed picture could be obtained of the use structure of phthalate in final products), clear differences do arise with regard to nonylphenol and the APEOs.

While bisphenol A and the phthalates reveal no substantial qualitative differences in their breakdown of use in the processing (productive consumption) and final product (product consumption) stages (at least as far as we can assume, given that no detailed picture could be obtained of the use structure of phthalate in final products), clear differences do arise with regard to nonylphenol and the APEOs.

More than 80 per cent of nonylphenol processed was used in ethoxylates. However, APEOs accounted for less than 50 per cent of the nonylphenol in final products sold in Germany in 1995. In other words, APEOs, specifically nonylphenol ethoxylates, produced in Germany are largely exported (approx. 75 per cent) and not used within the country. This is clearly linked to the debate about the aquatic toxicity of APEOs and their breakdown products, notably nonylphenol. In 1986, this debate prompted the industry associations to make a voluntary commitment to discontinue use of APEOs in household washing powders and cleaners which fall under the Detergents and Cleansers Act. Subsequently, APEOs were substituted in many applications, so that their consumption declined sharply. Long-term substitutes were to be found for approx. 9,500 tonnes (56 per cent) out of a total of 17,000 tonnes consumed in 1985. APEO consumption in 1995 was about the same as in 1985 minus the amount to be substituted, or that contained in products significant to wastewater under the Detergents and Cleansers Act.

The emissions given in Table 4 from material flows of the substances studied are in many ways only an approximation to the actual amounts involved. They refer to the manufacturing process for the respective substances (production, processing, transport) and to the products sold and those in use.

It was indicated above that this study addresses only identifiable emissions, since no indicators are available to estimate emissions for a number of important product groups. This is true in particular for epoxy resins in the case of bisphenol A, and for nonylphenol resins and emulsion polymers in the case of nonylphenol. For all three substance groups, no quantifiable information can be given on emissions from the disposal phase of products withdrawn from the production and consumption process, since the empirical foundations are insufficient.

The fall in nonylphenol concentrations found in surface waters in 1995 by comparison with the latter half of the eighties points to the impact of APEO substitution, but it also shows that (alongside possible remobilisation of river sediment) substantial sources of emissions still exist. Efforts to find substitutes for APEOs in detergents and cleaning agents have led to substitution in other fields of application as well, so that the increase in consumption for applications not covered by the reduction is only slight.

The reference volumes pose a second methodological problem for emission estimates. An estimate of actual emissions for durable products such as plastics must be based not only on the amounts produced and processed but also on the volume of existing products in use. This exceeds the annual volume of new products and the annual volume of products discarded and withdrawn from use. It was not considered for bisphenol A, since no data were available on existing products and the emissions resulting from them. For the phthalates, emissions were estimated using empirically based emission factors for the various stages of the material flow. The emission factors were derived by the method employed in a study by the European Chemical Society (CEFIC), which took di(2-ethylhexyl) phthalate as its model substance and based its assessment of emissions from products on those currently in use in the consumer phase. Assuming that Western European does not differ substantially from Germany in terms of the relationship between the volume of products in use and the new output of phthalates, it is safe to say that the products in the consumer phase have been considered in the estimated emission factors (see below for details). No statements are made on emissions from durable products (phenol resins, emulsion polymers) containing nonylphenol, and annual emissions of alkylphenol ethoxylates are regarded as directly proportionate to consumption in new products.

Summing up, the estimate of actual bisphenol A emissions from products is likely to be on the low side, since it does not take sufficient account of products in use. The same applies to emissions from products containing nonylphenol, since the figures do not include emissions from durable plastics applications. However, the magnitudes of the respective emission flows are not likely to be greatly distorted. Emissions of bisphenol A from products are not significant in absolute terms. They originate mainly from one specific use (recycling of thermal paper) for which the available figures are relatively accurate. In the case of nonylphenol, APEOs are the dominant source of emissions, while those from emulsion polymers and phenol resins are presumably around ten times lower. However, these qualifications demonstrate that a great deal remains to be clarified, and it is not possible to make any definitive statements.

The emissions given in Table 4 are environmentally relevant emissions following exhaust air and effluent treatment in normal operation. Whatever the uncertainty about the estimates, the different magnitudes are conspicuous. Emissions are in the tens of tonnes for bisphenol A, approx. 650 tonnes (mean value) for both phthalates together and approx. 200 tonnes for nonylphenol. Table 4 also reveals that the use of final products is a far greater source of emissions than production and processing. This is particularly pronounced for nonylphenol and less so for the phthalates, but essentially applies to all three substances. The first question which arises concerns the main factors which influence the emission patterns of the different substances.

In this context , a distinction must be drawn between (1) material properties of the substances and products containing them, (2) factors associated with processing and disposal methods and (3) factors attributable to use.

In material terms there are major differences between the products in the way they are converted or integrated into products. With the exception of a few special-purpose applications, bisphenol A reacts and forms a chemical bond with the polymer and can be released from the plastic only in very small amounts. The phthalates form only a physical bond with the polymer matrix of the various plastics and may therefore migrate, evaporate and wash out depending on volatility, temperature and ambient conditions. Although most nonylphenol is chemically transformed (phenol resins, alkylphenol ethoxylates), it may be released from APEOs in sewage and is itself relatively stable.

As far as processing and disposal methods are concerned, the crucial factors are the standard of pollution abatement technologies employed in the production, processing and transport of the substances and products containing them, as well as effluent treatment. This also includes what is done with the contaminated sewage sludge (agricultural use or incineration/landfill).

Emissions in use result largely from the intended handling and purposes of the substances or products containing them. This applies, for example, when APEOs are used as formulation aids in pesticides and in other APEO applications which give rise to significant effluent emissions. While these can be reduced by appropriate technical precautions (treatment plants), it is not possible to avoid all waterborne emissions of nonylphenol.

The differences in emission patterns for the material flows studied can be essentially summed up as follows.

The main source of bisphenol A emissions is the recycling of thermal paper containing "free" BPA as the colour development component, a co-reactant. These emissions occur in the consumer phase. Waterborne emissions totalled between 1 and 1.6 tonnes in 1995. Emissions in sewage sludge used in agriculture are estimated at 9 to 17 tonnes, although in this connection we need to determine to what extent BPA is broken down in sewage sludge. Emissions from other products were estimated only for polycarbonate (max. 1.5 tonnes from annual production).

In the case of the phthalates higher (airborne) emissions are to be expected from processing and products. Table 4 gives emission ranges based on estimates of emission factors at the various stages of production and product use.

These emission factors draw on studies of DEHP. They consider material properties (vapour pressure, volatility), processing and disposal (different emission rates resulting from different processing methods for PVC and other polymers, e.g. calendering and injection moulding), and exhaust air purification. They also take into account differences in products containing phthalates (applications involving open and sealed surfaces, e.g. floor coverings, paints and double glazing). Absolute emissions are estimated by relating the emission factors to domestic product consumption. The emission factors are in the region of 0.03 per cent (mean values) in production and transport, 0.8 per cent for DBP and 0.25 per cent for BBP in processing. Emission rates from products are put at 3.4 per cent for DBP and 1.5 per cent for BBP.

These differences indicate that products are the dominant source of emissions for phthalates, too.

According to producers, nonylphenol emissions, like those of bisphenol A, are low for the entire production process. This is because nonylphenol is produced by large-scale operations with their own wastewater treatment and exhaust air purification facilities. It was mentioned above that no information is available on emissions from nonylphenol resins and emulsion polymers. Due to the uses involved, the main sources of APEO - and thus also nonylphenol - emissions are pesticides (APEOs as formulation aids) and APEO applications giving rise to waterborne emissions calculated at 800 tonnes annually, as well as construction chemicals. Total nonylphenol emissions are estimated to be 210 tonnes per annum, of which at least 60 tonnes are direct waterborne emissions.

The figures given in Table 1 for the annual volumes of the substances dumped in landfills or incinerated are merely illustrative and represent rough estimates. They indicate that landfills contain large volumes of the three substances. Information on their behaviour, including emissions in leachate, is sparse. Where appropriate plant exists, the information available suggests that incineration of the substances studied is not likely to give rise to any emissions

The material flow analysis shows that not only the volumes of the substances produced and used are environmentally relevant, but also that their bonding and the form in which they occur in products are crucially important.

The greater part of environmental emissions result from diffuse release of the substances from products in the consumer phase. This is true across the board for the two phthalates. Specific applications which are particularly prone to emissions can be identified for bisphenol A and nonylphenol. This is true of bisphenol A in (sprays) and APEO applications in various fields which give rise to waterborne emissions of nonylphenol. The substitution of APEO in washing and cleaning agents in recent years shows that it is possible to reduce environmental emissions by targeting efforts at specific applications.

There was insufficient information to make any reliable statements for a large number of applications or for waste disposal. Further clarification should be provided by figures from the EU study of waste materials, which were not available at the time of the study.