View all publications

6 Comparison with International Data

6.1 Data limitations

The secondary metal sector is characterised by a wide variety of furnace types operated in different modes and with a range of air pollution controls. Furthermore, the input materials being processed will vary considerably. Taken together, these factors mean that emissions can be expected to vary widely from plant to plant, over time, and with changes to the operation of an individual plant.

Emissions of PCDD and PCDF from the secondary metal sector have not been comprehensively studied internationally. Much of the data that is available is derived from tests in Germany, where standards of pollution control on plants and controls over input materials are likely to be among the most stringent in the world. Consequently, the relatively low emission factors from these studies may not be representative of other countries. Conversely, some data is available for older plants with practices that are no longer carried out, or where air pollution control has improved.

In some cases systems are in place to collect particulate matter from flue gases. The filter or cyclone dusts may contain PCDD and PCDF at high levels and these become a further discharge to the environment upon disposal to land.

6.2 Ferrous foundries

The US EPA Draft Dioxin Reassessment (US EPA, 2000) contains emission factors for ferrous foundries. The emission factors used were from data reported from Germany for a variety of furnaces. The mean of the data was 1.26 µg I-TEQ tonne-1 of metal feed, although the factors spanned four orders of magnitude.

Testing by the California Air Resources Board [AB-2588 facility test report.] (CARB) had been undertaken but was not considered to be representative of the US industry. The CARB testing was based on batch-operated cupola furnaces charged with pig iron, scrap iron, steel scrap, coke and limestone, with an oil-fired afterburner and a fabric filter (US EPA, 2000). The CARB data gave an emission factor of 0.37 µg I-TEQ tonne-1 of metal feed (0.42 µg WHO-TEQ tonne-1).

German testing of electric arc furnaces showed a wide range of emissions factors (Working Group, 1996). Flue gas treatment using fabric filters appeared to capture about 90% of the PCDD and PCDF. Emissions ranged from about 0.025 to 5 µg I-TEQ tonne-1 (product). It should be noted that runs using a plant with scrap pre-heat gave emission concentrations about 10 times higher than the other data. (No emission factor could be calculated from the data available, but it would likely be of the order of 20-50 µg I-TEQ tonne-1.)

Tysklind et al (1989) tested an EAF continuously charging with a mixed scrap containing a variable chlorine content, and measured 0.8-30 µg Nordic-TEQ tonne-1 dry without fabric filtration and 0.2-7.7 µg Nordic-TEQ tonne-1 with fabric filtration. The worst case was batch charging of an open furnace with materials containing cutting oils, which resulted in 190 µg Nordic-TEQ tonne-1. The reported emission factors were 13 µg Nordic-TEQ tonne-1 for batchwise charging and 7.7 µg Nordic-TEQ tonne-1 for continuous charging.

A study of nine Swedish steel plants remelting scrap in an EAF (Öberg and Allhammar, 1989) gave dioxin emissions in the range <0.23-9.0 µg TEQ (Eadon) tonne-1 product.

The European Dioxin Inventory Stage II (Quass et al, 2000) emission factor for 1985 was 3 µg I-TEQ tonne-1. This was three times higher than the 'typical' factor presented in the Stage I report, but within the Stage I range of 0.3-5.7 µg I-TEQ tonne-1. Emission factors have been lowered over time to reflect a decrease in installations using scrap pre-heating devices and improved flue-gas cleaning. The reference also reported tests carried out on "cold blast cupolas" for producing cast-iron and steel. Six tests gave emission concentrations of between 0.003 and 0.184 ng I-TEQ/m3, yielding an average emission factor of 0.35 µg I-TEQ tonne-1 of product with a maximum of 1.45 µg I-TEQ tonne-1 of product.

Table 13 summarises emission factors for liquid steelfrom the Standardized Toolkit for Identification and Quantification of Dioxin and Furan Releases (UNEP, 2001a). These are also mainly drawn from testing in Germany.

UNEP recommends that an emission factor of 3 µg TEQ tonne-1 of liquid steel is used for EAFs based on tests from plants using clean scrap and virgin iron with some afterburners and fabric filters.

Emissions from EAFs using dirty scrap and containing cutting oils or plastic, and plants with scrap pre-heating and with poor controls, have been found to have higher emissions of PCDD and PCDF. In such cases, UNEP recommends an emission factor of 10 µg TEQ tonne-1 of liquid steel.

Where controls are placed on scrap to exclude cutting oils and heavily contaminated scrap, and gas cleaning with secondary combustion and fabric filtration are used, emissions below 0.1 µg TEQ tonne-1 of liquid steel can be achieved.

Table 13: UNEP iron foundry emission factors


Emission factor
(µg I-TEQ tonne-1 of liquid steel)

Cold air cupola or rotary drum with no gas cleaning


Rotary drum with fabric filter


Cold air cupola with fabric filter


Hot air cupola or induction furnace with fabric filter


EAFs designed for low emissions using clean scrap or virgin iron


6.3 Secondary copper

The European Stage I inventory gives a typical emission factor of 50 µg I-TEQ tonne-1 of refined copper, with a minimum and maximum value of 5 and 500 µg I-TEQ tonne-1 respectively. The factor was assigned high uncertainty because it was based on a poor data pool. Measurements from the French inventory programme suggested that emission factors were lower, but the Stage I figure was maintained for the European Stage II inventory because there was insufficient evidence to show that the French data were representative of European secondary copper installations.

Emission factors reported for German shaft furnaces or converters and reverberatory furnaces ranged from 5.6 to 110 µg I-TEQ tonne-1 and from 0.005 to 1.56 µg I-TEQ tonne-1 respectively. Emission factors reported for two smelter and casting furnaces in Sweden, in which relatively clean scrap was used as input, were 0.024 and 0.04 µg I-TEQ tonne-1. A smelter in Austria was reported to have an emission factor of 4 µg I-TEQ tonne-1.

A study was conducted of a plant with electrically heated melters fed with copper and zinc to produce brass ingots in Portugal (Coutinho et al, 2001). The study yielded emission factors ranging from 0.09 to 0.4 µg I-TEQ tonne-1. The exhaust system had two gas treatment lines including cooling (to prevent de novo synthesis), cyclones and fabric filtration.

UNEP's published emission factors for the secondary copper industry are summarised in Table 14.

Table 14: UNEP secondary copper and brass emission factors


Emission factor
(µg TEQ tonne-1 of brass or copper)


Basic technology - mixed materials with simple fabric filtration


Well controlled - scrap copper with afterburners and fabric filters


Optimised for PCDD/PCDF control (i.e. rapid water quench or activated carbon)



Simple melting furnace


Induction furnace with air pollution control


The US EPA dioxin inventory (US EPA, 2000) discusses releases from secondary copper production. The earliest US data relates to a batch-fed cupola-type blast furnace controlled by a gas-fired afterburner and a fabric filter processing a mixture of scrap copper-containing materials and considerable plastic contamination. Emissions were determined to be 779 µg I-TEQ tonne-1 of scrap (810 µg WHO-TEQ tonne-1).

Further testing was carried out on another cupola blast furnace system producing blister copper and fitted with afterburner, cooling tower and fabric filter. Emissions from this plant were 16,618 µg I-TEQ tonne-1 of scrap (16,917 µg WHO-TEQ tonne-1 ). This plant was operated under reducing conditions.

A third plant was tested which operated under oxidising conditions. The feed was relatively pure and clean, and the rotary furnace was equipped with a primary quench and venturi scrubber. Emissions were reported at 3.6 µg I-TEQ tonne-1 of scrap(3.66 µg WHO-TEQ tonne-1).

6.4 Secondary aluminium

A Portuguese study of secondary aluminium production gave emission factors ranging from 47 to 200 µg I-TEQ tonne-1 (Coutinho et al, 2001). Salts fluxes such as sodium chloride, potassium chloride and fluoride were added to the process. There was no treatment of air emissions from these furnaces.

In the US, four facilities were tested in 1995 and two facilities were tested in 1992. The results of testing are summarised in Table 15 (US EPA, 2000).

Table 15: US factors from test data for secondary aluminium


Emission factor
(µg I-TEQ tonne-1)

Top-charge melt furnace; clean scrap


Sweat furnace pre-cleaner, and reverberatory furnace with afterburner


Roasting dryer and reverberatory furnace with afterburner and fabric filter (lime addition on the furnace fabric filter)


Roasting dryer and reverberatory furnace with afterburner and fabric filter (lime addition on the furnace fabric filter)




Venturi scrubber


a Charge material.

b Aluminium product.

The average value in the US tests from the six facilities was 21.1 µg I-TEQ tonne-1 of charge material, including the value with very clean scrap.

For the European Stage I inventory, default emission factors ranged from 5 to 100 µg I-TEQ tonne-1 with a typical value of 22 µg I-TEQ tonne-1 , which was used for the estimation in Stage II. Data from 11 facilities tested in Germany had emission factors calculated at 0.01-167 µg I-TEQ tonne-1 of charge material. The mean emission factor for the facilities was 42 µg I-TEQ tonne-1 of charge material (Working Group, 1996).

UNEP (2001a) has published emission factors for the secondary aluminium industry per tonne of aluminium, as summarised in Table 16.

Table 16: UNEP secondary aluminium emission factors


Emission factor
(µg TEQ tonne-1 of aluminium)

Thermal processing with simple or no dust removal


Thermal processing; well-controlled fabric filters with lime injection/afterburners


Drying (pre-treatment) of shavings


Optimised for PCDD/PCDF control, scrap-cleaning lime injection, afterburners, fabric filters and activated carbon


6.5 New Zealand measurements compared to international data

Figure 6 provides a comparison of the range of New Zealand measurement data compared to data measured in overseas studies for aluminium (Al), iron and steel (Fe) and copper (Cu). The range of measurements reflects the range of process variables; in particular, scrap quality and air pollution control. Apart from the results for two sites, emissions were well below overseas data, indicating the New Zealand industry generally has low emissions.

Figure 6: PCDD and PCDF emissions per tonne of metal compared to international data