Annex 18 Complementary information to source category 2c Iron and Steel Production and Foundries

Overview of recent revisions

PCDD/PCDF emission factors for residues were revised for class 3 (iron and steel making). In particular, the previous definition of classes 2 and 3 (iron and steel making) only included clean scrap; this was revised to cover both clean and dirty scrap. As for iron foundries, the revised emission factors confirm in general the previous values. Regarding hot-dip galvanizing plants, air emission factors were confirmed while residues emission factors were significantly modified. Emission factors for other unintentional POPs are listed below:

Table III.18.1 PCB emission factors for source category 2c Iron and Steel Production and Foundries

2c Iron and Steel Production and Foundries		Emission Factors (µg TEQ/t LS)
Classification		Air	Water	Land	Product	Residue
Iron and Steel Making
1	Dirty scrap (cutting oils, general contamination), scrap preheating, limited controls
2	Clean scrap/virgin iron or dirty scrap, afterburner and fabric filter
3	a. Clean scrap/virgin iron or dirty scrap, EAF equipped with APC designed for low PCDD/PCDF emission, b. BOF furnaces	0.001
4	Blast furnaces with APCS	0.001
Iron foundries
1	Cold air cupola [or hot air cupola] or rotary drum with no gas cleaning
2	Rotary Drum - fabric filter	0.5
3	Cold air cupola – fabric filter [or wet scrubber]	0.5				0.1
4	Hot air cupola, or induction furnace, fabric filter	0.02				0.01

Table III.18.2 HCB emission factors for source category 2c Iron and Steel Production and Foundries

2c Iron and Steel Production and Foundries		Emission Factors (µg TEQ/t LS)
Classification		Air	Water	Land	Product	Residue
Iron and Steel Making
1	Dirty scrap (cutting oils, general contamination), scrap preheating, limited controls	2500
2	Clean scrap/virgin iron or dirty scrap, afterburner and fabric filter	2500
3	a. Clean scrap/virgin iron or dirty scrap, EAF equipped with APC designed for low PCDD/PCDF emission, b. BOF furnaces	2 BOF 2500 EAF
4	Blast furnaces with APCS	1

For iron and steel making, emission factors for PCB/HCB are provided:

• With a medium level of confidence for HCB emissions related to classes 2 and 4, as emission factors are not based on expert judgment but are not derived from a broad geographical coverage;
• With a low level of confidence for HCB emissions related to classes 1 and 3 and for PCB emissions (all classes), as emission factors are based on extrapolations and expert judgment.

For iron foundries, emission factors for PCB are provided:

• With a medium level of confidence for class 3 and 4 air emissions and for class 4 residue releases, as emission factors are not based on expert judgment but are not derived from a broad geographical coverage;
• With a low level of confidence for class 2 air emissions and for class 3 residue releases, as emission factors are based on extrapolations and expert judgment.

Derivation of emission factors

Release to Air

I. Iron and steel making

PCDD/PCDF will be released into gases from furnaces. It can be difficult to capture all the gases from the process, and a large fraction of the gas and PCDD/PCDF may be present in fugitive emissions rather than in stack gases. Emissions seem to increase greatly by poor quality mixed scrap feeds, in particular where metal working residues, including cutting oils, are fed. The preheating of scrap to improve energy efficiency can lead to increased emissions as well; concentrations up to 9.2 ng TEQ/Nm³ have been measured (Germany, LUA 1997). In Europe, PCDD/PCDF measurements gave emission factors that ranged 0.07-9 μg I-TEQ/t LS (liquid steel); based on European data, a conversion factor of 940 kg pig iron/t LS was used.

Regarding blast furnaces (class 4), flue gas volumes from hot stoves are between 100,000 and 600,000 Nm³/h per blast furnace. Emission factors determined from measurements from four EU member States were from <0.001 to 0.004 μg I-TEQ/t LS. Such data are confirmed by more recent measurements from the European Union, as the iron and steel BREF document reports an emission factor of 0.003 µg TEQ/t (BREF 2012).

In BOFs during oxygen blowing, converter gas is released, which contains small amounts of PCDD/PCDF. Basic oxygen steel making plants in Europe generally have quite low emission factors, slightly higher than blast furnaces (with an upper end of 0.06 μg I-TEQ/t LS based on measured data). Polish data (Grochowalski et al. 2007) are consistent with this figure as an emission factor of 0.02 µg TEQ/t was derived from measurement at a plant scale. As emission factors of BOF are similar to those of class 3 EAF (see below), BOF are incorporated in class 3.

For electric arc furnaces, most measured emission data relate to plants using relatively clean scrap and virgin iron and which are fitted with some after-burners and fabric filters for gas cleaning. Emission factors derived from plants in Sweden, Germany, and Denmark gave emission factors between 0.07 and 9 μg I-TEQ/t LS. For the Toolkit, an emission factor of 3 μg TEQ/t LS is applied (class 2) (Bremmer et al. 1994, SCEP 1994, Napier 1998). This emission factor is supported by data from several publications issued between 2003 and 2009. For instance, the following emission factors were derived: 1.33-7.6 µg TEQ/t in Taiwan (Wang et al. 2009), 1.7 µg TEQ/t in Japan (Sakai et al. 2009), 4.8 µg TEQ/t in Italy (ENEA 2003).

Emissions from EAF plants using dirty scrap containing cutting oils or plastic materials as well as plants with scrap preheating and relatively poor controls were found to have higher concentrations of PCDD/PCDF in stack gases as found in Germany (SCEP 1994). In such cases, an emission factor of 10 μg TEQ/t LS is used (poor plants could emit more) for class 1. This emission factor is supported by data from many publications issued between 2004 and 2009. For instance, the following emission factors were derived: 6.3 µg TEQ/t in Taiwan (Hwang et al. 2006), 11-90 µg TEQ/t in Sweden (Oberg 2004).

Where careful controls are placed on the scrap used (excluding cutting oils and heavily contaminated scrap) and efficient gas cleaning is used with secondary combustion and fabric filters (sometimes in combination with a rapid water quench) emissions below 0.1 ng TEQ/Nm³ can be achieved. For these plants an emission factor of 0.1 μg TEQ/t should be used (class 3). The same low concentrations were measured in the flue gases from basic oxygen furnaces; e.g. a median concentration of 0.028 ng I-TEQ/Nm³ (LUA 1997). This emission factor is supported by data from many publications issued between 2003 and 2010. For instance, the following emission factors were derived: 0.3-0.92 µg TEQ/t in New Zealand (Merz 2004, Graham and Bingham 2010), 0.03 µg TEQ/t in Sweden (Oberg 2004), 0.26 µg TEQ/t in Italy (ENEA 2003).

Regarding PCBs, based on results got on industrial sites from Poland (Grochowalski et al. 2007), an emission factor of 0.001 µg TEQ/t is proposed for classes 3 and 4.

Regarding HCB, Japan measurements on EAFs have been published for the last years in several publications (Ota et al. 2005, Sakai et al. 2009). EAF emission factors range between 2,100 and 2,900 µg/t. As a consequence, the same emission factor is proposed for EAFs related to classes 1 to 3. Measurements at a BOF and a blast furnace in Poland correspond to an emission factor of 2 µg/t and 1 µg/t respectively (Grochowalski et al. 2007).

II. Iron foundries

For foundries, there are hardly any data available: testing in Germany (SCEP 1994) showed that hot air cupolas and induction furnaces fitted with fabric filters had low emissions to air, an emission factor of 0.03 μg TEQ/t of product should be used (class 4). This initial emission factor has been confirmed by recent measurements carried out in France (0.0087 µg TEQ/t, Duquet and Fiani 2006), in Poland (0.02-0.06 µg TEQ/t, Grochowalski et al. 2007) and in South Korea (0.1 µg TEQ/t, Yu et al. 2006).

Cold air cupolas showed higher emissions and a factor of 1 μg TEQ/t is used for plants with fabric filters (class 3). This initial emission factor has been confirmed by recent Chinese measurements (Lv et al. 2011a).

Limited testing on rotary drum furnaces showed higher levels again and a factor of 4.3 μg TEQ/t is applied to plants with fabric filters for gas cleaning (class 2).

Where cold air cupolas or rotary drum furnaces are used which do not have fabric filters or equivalent for gas cleaning a higher emission factor of 10 μg TEQ/t should be used (class 1).

If poor quality scrap (high contamination) or poorly controlled furnaces with gas cleaning other than effective fabric filters is found this should be noted.

Regarding PCBs, new data have been published since 2005. A Chinese team (Lv et al. 2011a) reported an emission factor of 0.5 µg TEQ/t for cold air cupolas (class 3) while a South Korean team (Yu et al. 2006) reported an emission factor range of 0.01-0.03 µg TEQ/t for hot air cupolas (class 4). As class 2 and 3 emission factors for PCDD/PCDFs are very similar, it is proposed to use the same PCB emission factor for theses two classes.

III. Hot-dip galvanizing plants

For hot-dip galvanizing plants, as for any other thermal plant, the presence or absence of flue gas cleaning equipment, will be a determining factor as to the magnitude of the PCDD/PCDF air emissions. Some plants do not have flue gas cleaning devices, others have bagfilters. From Germany and without further specification, PCDD/PCDF concentrations between 0.007 and 0.132 ng I-TEQ/Nm³ were measured in the flue gases from four installations (LUA 1997); the median was 0.016 ng I-TEQ/Nm³. The Danish inventory utilized the German data and a stack volume of 33,000 Nm³ emitted per ton of iron/steel galvanized to estimate its national emissions.

A Spanish study (Fabrellas et al. 2003), investigated hot-dip galvanizing plants equipped with bagfilters. The concentrations – 0.003-0.014 ng I-TEQ/Nm³ - were much lower than those reported by Germany (LUA 1997). Nevertheless, the emission factor to air reported in these two studies were very similar: Spanish study =0.041-0.061 µg I-TEQ/t of galvanized steel and 0.007-0.027 µg I-TEQ/t of galvanized steel for plants without and with degreasing step, respectively; whereas the German study resulted in an air emission factor of 0.061 ng I-TEQ/t of galvanized steel. The Spanish team has carried out further measurements on six plants (Martinez et al. 2008): emission factors ranging from 0.007 to 0.061 were derived.

Release to Water

Releases to water could occur where wet scrubbers or quenches are used. No data were available to provide an emission factor. Where an effluent is released this should be noted and information reported.

Release to Land

No release to land is expected.

Release in Products

No significant release is expected with the product steel from this process, it has been subject to high temperatures and PCDD/PCDF is likely to have been driven off or destroyed.

Release in Residues

The principal residues of interest are slag and dust collected in flue gas treatment systems. Other dust deposited from fugitive emissions may also contain PCDD/PCDF.

I. Iron and steel making

From blast furnaces, 9-15 kg of dust and sludge per ton of LS are generated from the gas purification system. 280 kg of slag are produced per ton of LS.

In BOF steel making, 12-27 kg of dusts and slags are generated per ton of LS from BOF gas treatment. Converter slag is 99 kg per ton of LS. Electric arc furnaces produce more slags, e.g. 129 kg/t LS for carbon steels and 161 kg/t LS for high alloyed and stainless steels.

An average emission factor for PCDD/PCDF in residues can only be given for EAFs: from gas cleaning operations (fabric filter) an emission factor of 15 μg TEQ/t is based on an average of UK data (Dyke et al. 1997). This factor assumes similar gas cleaning equipment; the release may be different with other systems. This factor is used for the poorly controlled and average plants (classes 1 and 2). Such emission factor has been confirmed by recent publications. The following emission factors have been derived recently: 22 µg TEQ/t (Chang et al. 2006), 32 µg TEQ/t (Du et al. 2009).

A lower emission factor of 0.15 μg TEQ/t is used for the best plants (Bremmer et al. 1994): class 3. The fate or use of the residues should be noted (PCDD/PCDF can be introduced into other processes if these residues are used as feedstock in recycling processes).

II. Iron foundries

From foundries, cupolas and EAFs emit particulate matter, which is likely to contain PCDD/PCDF. Induction furnaces emit much less particulates. Initial emission factors were derived from data obtained in Germany (SCEP 1994). Slag can be generated as well as sand casting technologies will generate substantial volumes of sand, which may be reused in the plant or be sent off for use as construction material (USEPA 1998b). These emission factors have been confirmed by some recently published data from other parts of the world. For instance, measurements carried out in 14 Chinese foundries which can be allocated to classes 3 and 4 showed an average emission factor of 0.365 µg TEQ/t.

First PCB data with respect to residues were published recently (Lv et al. 2011a). From these results, emission factors of 0.1 µg TEQ/t and 0.01 µg TEQ/t are proposed for classes 3 and 4 respectively.

III. Hot-dip galvanizing plants

PCDD/PCDF concentrations in fly ashes from hot-dip galvanizing plants were measured to be 2.15-9.6 ng I-TEQ/kg ash with a geometric mean of 3.9 ng I-TEQ/kg fly ash (German data in LUA 1997); no emission factor can be derived from these data. The Spanish study did derive emission factor ranges of 487-8,075 pg I-TEQ/g of filter dust for plants without degreasing step and of 127-1,804 pg I-TEQ/g of filter dust for plants with degreasing step, respectively (Fabrellas et al. 2003). The mean value for both cases is 2,000 pg TEQ:g of filter dust. Based on Spanish and Danish publications (Martinez et al. 2008, Hansen 2001), it is assumed that 1 kg of filter dust is generated by the air pollution control for each ton of galvanized steel produced. As a consequence and assuming that higher emission factors are expected in the case of facilities without degreasing step, emission factors for classes 2 and 3 are proposed to be set at 2 µg TEQ/t and 1 µg TEQ/t respectively.

With respect to class 1, emission factors are based on ash residues concentrations measured in Spain (Martinez et al. 2008) and in China (Lv et al. 2011b). The Spanish team measured concentrations of 0.7-107 pg TEQ/g of ashes (n = 9), giving a mean value of 20 pg TEQ/g of ashes. The Chinese team obtained a similar result (12-38 pg TEQ/g of ashes). From this figure, the Chinese team derived an emission factor of 0.00065-0.0167 µg TEQ/t of galvanized steel. Therefore an emission factor of 0.01 µg TEQ/t galvanized steel is proposed for class 1.