E

Ta

b

U

a

ARRAA

KCICPAI

1

cneraeaIht1

(

g

1h

International Journal of Greenhouse Gas Control 10 (2012) 310–328

Contents lists available at SciVerse ScienceDirect

International Journal of Greenhouse Gas Control

j our na l ho me p age: www.elsev ier .com/ locate / i jggc

ffect of CO2 capture on the emissions of air pollutants from industrial processes

akeshi Kuramochia,∗, Andrea Ramírezb, Wim Turkenburgb, André Faaijb

Climate Change Group, Institute for Global Environmental Strategies (IGES), 2108-11 Kamiyamaguchi, Hayama, Miura-gun, Kanagawa 240-0115, JapanCopernicus Institute of Sustainable Development, Department of Innovation, Environmental and Energy Sciences, Faculty of Geosciences, Utrecht University, Budapestlaan 6, 3584CDtrecht, The Netherlands

r t i c l e i n f o

rticle history:eceived 7 March 2011eceived in revised form 26 May 2012ccepted 29 May 2012vailable online 20 July 2012

eywords:O2 capture

ron and steelementetroleum refineriesir pollutants

a b s t r a c t

This study assesses whether the deployment of CO2 capture technologies in the European industrialsector would result in significant changes in the emissions of air pollutants (NOx, SO2, PM, and NH3) inthe short term. The industrial sectors investigated were: cement, petroleum refineries, and iron and steel.The analysis included onsite emissions and changes associated with grid electricity consumption due toCO2 capture. Post-combustion capture using monoethanolamine (MEA) was considered for the cementsector and petroleum refineries, and Top Gas Recycling Blast Furnace (TGRBF) with vacuum-pressureswing adsorption (VPSA) for the iron and steel sector.

The results show that when all three industrial sectors in the EU-27 are fully equipped with CO2 capture,industrial SO2 emissions in the EU-27 may decrease by 40–70% whereas NH3 emissions may increase by120–520% (equivalent to 2–8% of total European emissions). The large increase in NH3 emissions is due tothe degradation of MEA. Cement and petroleum refineries account for nearly all these changes. The results

ndustry also show limited impact (within ±10% of EU-27 industrial emissions) on NOx and PM emissions. Emissionchanges due to electricity import/export are found to be equally important as onsite emission changes.For the iron and steel sector, the changes in National Emissions Ceilings Directive (NECD) emissions arefound to be limited for the selected CO2 capture technique under conservative assumptions. However,the changes in the NECD emissions could vary largely depending on how the steel mill will adapt andoperate their coke oven batteries that supply the coke to the blast furnace (BF).

© 2012 Published by Elsevier B.V.

. Introduction

Industry and petroleum refineries are among the largestontributors to anthropogenic CO2 emissions, accounting forearly 40%1 of these emissions globally (IEA, 2010). Together withnergy efficiency improvement and deploying nuclear energy andenewables, carbon capture and storage (CCS) is gaining attentions a promising option to achieve significant reduction of CO2missions in the atmosphere. The potential of CCS in the industrialnd petroleum refining sectors is considered to be significant. Thenternational Energy Agency (IEA) estimates that, in a scenario toalve global greenhouse gas (GHG) emissions in 2050 compared

o 2007 level, nearly half of all CCS deployed (up to more than0 Gt/yr) would be in industrial processes (cement, iron and steel,

∗ Corresponding author. Tel.: +81 46 826 9613; fax: +81 46 855 3809.E-mail addresses: kuramochi@iges.or.jp, takeshi.kuramochi@gmail.com

T. Kuramochi).1 Includes coke ovens and blast furnaces and also CO2 emissions from power

eneration and process emissions (IEA, 2007).

750-5836/$ – see front matter © 2012 Published by Elsevier B.V.ttp://dx.doi.org/10.1016/j.ijggc.2012.05.022

and chemicals) and the fuel transformation sector (petroleumrefineries and liquefied natural gas production) (IEA, 2008).

However, CO2 is not the only substance emitted from theindustrial sector. Since the industrial revolution, power plants andindustries have been a major source of environmental pollutionsuch as acidification, eutrophication, and toxification of waters.Large efforts have been conducted by governments and industriesto mitigate these effects. The deployment of (new) technologiesat large scale, such as CCS, should therefore take into accountthe possibility of side-effects on other industrial plant emissions.Therefore, it is important to assess the possible impact of CO2capture technologies on the emissions of air pollutants such asNOx, SO2, ammonia, non-methane volatile organic compounds(NMVOC), and particulate matter (PM), which contribute to theformation of ground-level ozone, and damage ecosystems by acid-ification and eutrophication. In the European Union (EU), currentemission limits for the first four pollutants are set under the EUNational Emissions Ceilings Directive (NECD) for 2010 (European

Commission, 2001), and all four pollutants are covered by theGothenburg Protocol of the United Nations Economic Commissionfor Europe to abate acidification, eutrophication, and ground-levelozone (UNECE, 1999).

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 311

Table 1Contribution of iron and steel, cement and oil refining sectors to total CO2 and NECD pollutant emissions in the EU-27 in 2007. For comparison, data for public electricity andheat production, the total industrial sector, and the total EU-27 emission values are also presented.

Substance/sector CO2a NOx SO2 PM10 NH3

[Mt (%-total)] [kt NO2-equivalent (%-total)] [kt (%-total)] [kt (%-total)] [kt (%-total)]

Public electricity and heat production 1.4 × 103 (33) 1.9 × 103 (17) 3.9 × 103 (54) 1.3 × 102 (6) 4 (0.1)Industryb 9.3 × 102 (22) 1.8 × 103 (16) 1.3 × 103 (18) 5.6 × 102 (27) 55 (1.5)Iron and steelc 2.1 × 102 (5) 2.7 × 102 (2.4) 2.6 × 102 (3.6) 1.2 × 102 (5.5) <1Cementd 1.7 × 102 (3) 3.2 × 102 (2.9) 1.4 × 102 (2.0) 15 (<1) >1Petroleum refininge 1.3 × 102 (3) 1.7 × 102 (1.5) 5.2 × 102 (7.1) 17 (<1) 1EU-27 total 4.2 × 103 (100) 1.1 × 104 (100) 7.3 × 103 (100) 2.1 × 103 (100) 3.9 × 103 (100)

Data source: EEA (2011a), unless otherwise stated.a The figure excludes net CO2 removals from land use, land-use change and forestry (LULUCF) and emissions from international aviation and international maritime

transport (EEA, 2011a).b Includes the following NFR (New Format for Reporting) sectors: 1A2 (Combustion: Manufacturing industries and construction) and 2 (Industrial processes) (EEA, 2010).c Includes NFR sectors 1A1c (Manufacture of solid fuels and other energy industries), 1A2A (Combustion: Iron and steel), 1B1b (Fugitive emission from solid fuel: Solid

fuel transformation), and 2C1 (Process: Iron and steel production) (EEA, 2010). Sectors 1A1c and 1B1b are entirely or mostly due to coke manufacture. This study assumesthat all coke production is attributable to iron and steel industry.

d Because the EEA does not publish combustion-related emission figures specific for the cement sector, figures were taken from various sources. CO2 emissions data (netemissions excluding emissions from electric power) are taken from WBCSD (2011). Net emissions are 10 Mt lower than the gross emissions because the emissions fromalternative fuels are excluded from net emission figures. Emissions data for NOx , SO2 and PM are interpolated from the emissions projection for years 2005 and 2010 inthe NEC 2010 Baseline scenario of the GAINS model (IIASA, 2010). PM emissions data was obtained from EEA (2011a). Although the data does not include emissions fromcombustion, we did not consider this as an issue because PM emissions from combustion are limited (EEA, 2009). The minimum value for NH3 emissions was obtained fromEEA (2011a) for process emissions. Data does not include emissions from combustion. EEA (2009) suggests that NH3 emissions mainly originate in fuel combustion. GAINSm

ing/st

tpaCfmk

CsW2milpinotottfatINcct

2

psasT

odel also does not consider NH3 emissions from these processes.e Includes NFR sectors 1A1b (Combustion: Petroleum refining) and 1B2a iv (Refin

To date, in the scientific literature there are only a few studieshat include possible effects of CO2 capture on the emissions of airollutants. These studies have mainly taken the form of life cyclessessment (LCA) studies, particularly those on post-combustionO2 capture from power plants. An overview of literature can beound in Koornneef et al. (2010). To our knowledge, such an assess-

ent has not been performed systematically for CO2 capture fromey industrial emission processes.

This study aims to quantitatively assess the possible impacts ofO2 capture on the emissions of air pollutants from key industrialectors in Europe at the plant level and the European sectoral level.

e based the current study on our previous work (Kuramochi et al.,012), where we present a comparative techno-economic assess-ent of various CO2 capture technologies for carbon-intensive

ndustrial processes. In the current study we investigated the fol-owing industrial sectors: cement, iron, and steel (integrated steellants), and petroleum refining (furnaces and boilers). The focus

s on four key air pollutants: NOx, SO2, PM, and NH3. NMVOC isot considered because there is a lack of quantitative informationn the effect of CO2 capture on NMVOC emissions in the litera-ure (Koornneef et al., 2010). Taking into account the timeframef the revised Gothenburg Protocol, we investigated CO2 captureechnologies that are likely to become commercially available inhe short term, i.e. around 2020. This study focuses on the effectsrom CO2 capture and compression only: CO2 transport and storagere outside the scope of the research. Finally, this study quantifieshe net effect of CO2 capture on the emissions of NECD substances.n cases where CO2 capture leads to an increase in emissions of aECD substance, additional abatement measures (e.g., wash tower)an be applied using current available technologies at additionalost. These abatement measures were, however, not included inhis study.

. General description of the sectors studied

Table 1 shows the contribution of iron and steel, cement, andetroleum refining sectors to total CO2 and NECD substance emis-

ions in the EU-27 in 2007 (EEA, 2010, 2011a; IIASA, 2011). The tablelso presents figures for the public electricity and heat productionector as a reference. Apart from direct CO2 emissions presented inable 1, industry is accountable for about 570 MtCO2 emissions for

orage) (EEA, 2010).

the purchase of electricity from the power sector (EEA, 2011b). Thismakes industry responsible for about 1.5 GtCO2 per year, or 36% oftotal EU-27 emissions. Regarding NECD substance emissions, thecement sector shows large NOx emissions and the petroleum refin-ing sector shows large SO2 emissions relative to CO2 emissions. NH3emissions are negligibly small for all three sectors.

In 2007 the annual cement production in the EU-27 was 270 Mt,accounting for about 11% of world production (Cembureau, 2010).The cement sector is one of the most CO2 intensive industrial sec-tors, accounting for 130 MtCO2 emissions in the EU-27 in 2007 (IEA,2010). High CO2 emission intensity is not only caused by its largeenergy requirement but also by the emissions from raw materials,which is mainly comprised of calcium carbonate (main componentof limestone). Depending on the clinker/cement ratio, around 60%of the CO2 emissions originates in the calcination process whereclinker is produced (IPPC, 2009). The rest of the emissions areattributable to combustion of fossil fuels which are mostly coal(IPPC, 2009).

Petroleum refineries account for nearly 1 GtCO2/yr or around 4%of global total CO2 emissions worldwide (van Straelen et al., 2010).There are 116 petroleum refineries in the EU-27 as of 2008 andthe total capacity is a little below 1 Gt crude oil refined per year,accounting for about 20% the global total capacity (IPPC, 2010a).At the refineries, CO2 is emitted from various sources such as fur-naces and heaters, onsite heat and power plants (including CHPplants), and catalytic crackers. Onsite generation of electricity andheat is responsible for the bulk of the CO2 emissions in petroleumrefineries (Wilkinson et al., 2003). In particular, furnaces and boil-ers account for 65% of CO2 emissions from refineries worldwide(IEA GHG, 1999).

The iron and steel sector alone emitted 2.3 GtCO2 worldwidein 2007, accounting for 30% of total direct CO2 emissions fromthe industry (IEA, 2010). Crude steel production in the EU-27 was210 Mt in 2007, accounting for about 15% of global total productionof 1.35 Gt/yr (World Steel Association, 2010). 60% of the total crudesteel production in the EU-27 (126 Mt/yr) is from integrated steel-making plants, which consist of mainly five sections: coking, iron

ore agglomeration, blast furnace (BF), basic oxygen furnace (BOF),and final product manufacturing (World Steel Association, 2010).Most of the rest is produced from recycled steel scrap using electricarc furnace (World Steel Association, 2010). Significant amount of

312 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

for th

cbtbc

3

3

sfCeCtpe

pttfesipci

3u

NtlistpNfle

Fig. 1. System layout and system boundary

oal is consumed at an iron and steel plant not only to generate heatut also to reduce iron ore in the form of coke. Around 60% of world’sotal steel production is derived from a general description of car-on flows in an integrated iron and steel production process thatan be found in de Beer et al. (1998), Daniëls (2002), and elsewhere.

. Methodology

.1. System boundary

Fig. 1 shows the system boundary defined for this study. Theystem investigated accounts for all onsite emissions, i.e., emissionsrom the industrial process, onsite power and/or steam generation,O2 capture, purification, and compression to 110 bar, as well as themissions associated with electricity purchased from the grid. TheO2 capture process includes onsite energy plants specifically builto supply energy for CO2 capture. Therefore, if the onsite energylant generates excess electricity, then it replaces the centralizedlectricity generation.

There may be cases where CO2 capture changes the amount ofrocess gas export. For example, converting a BF from air-blowno oxygen-blown is an attractive option for CO2 capture becausehe fuel-quality process gas from BF becomes nearly nitrogen-ree. Moreover, this conversion improves the energy and reductionfficiency of the BF, requiring less coal and coke. The reduced con-umption of coal and coke also reduces the amount of BF gas whichs often used for power generation. In such cases, the emissionrofile at a centralized power plant may change because fuel mixhanges. Such indirect changes in emissions were also consideredn this study.

.2. Research steps, performance indicators, and assumptionssed throughout the study

For each sector, the impact of CO2 capture on the emissions ofECD substances was first assessed at the plant level. Subsequently,

he impact of CO2 capture at the European sectoral level was calcu-ated based on the plant-specific results. The methodology followedn this study is summarized in seven main steps. First, the referencepecific energy and material flows for the production of indus-rial products were defined based on a literature survey on typical

lants in Europe. Second, plant-specific data on the emissions ofECD substances (e.g. concentrations in flue gas and process gasows) as well as the annual industrial production and total CO2missions for existing plants in the EU-27 were gathered from the

e calculation of the emission of pollutants.

literature. Third, specific emissions of substance i per unit of prod-uct from industrial process j (cement: Cem, refineries: Rf, and ironand steel: IS) were calculated for the plants reported in the secondstep based on the reference energy and material flows defined inthe first step. For plant k, specific emission factor (EF) without CO2capture (reference case: Emi,j,k,ref, g/t industrial product) is definedas follows:

Emi,j,k,ref = Monsite,i,j,k,ref

Mind,j,k, (1)

where Monsite,i,j,k,ref is the reference annual onsite emissions fromthe industrial process j in the plant k (g/yr), Mind,j,k is the annual pro-duction from the industrial process j in the plant k (tonne industrialproduct/year) and Monsite,i,j,k,CC is the annual onsite emissions fromthe industrial process j in the plant k equipped with CO2 capture(g/yr).

Fourth, short-term CO2 capture technologies were selectedbased on our recently performed comparative assessment of CO2capture technologies for the industry and petroleum refineries per-formed for the short–mid term (within 10–15 yr) and the long term(20 yr or more) (Kuramochi et al., 2012). The changes in materialand energy flows due to CO2 capture were identified because someCO2 capture systems are not of an end-of-pipe nature and may altermass and energy flows in the industrial process, leading to changesin the generation levels of CO2 and the NECD substances. Fifth, thegas specifications (possible need for pre-treatment of CO2-rich gasprior to CO2 capture) and pollutant removal performance for dif-ferent CO2 capture technologies were assessed with the fourth steptaken into account. Sixth, specific NECD substance emissions withCO2 capture were calculated for the reference industrial productionfor a range of specific NECD substance emission levels identified inthe third step. Specific emissions (Emi,j,k,CC) for plant k with CO2capture are defined as follows:

Emi,j,k,CC = Monsite,i,j,k,CC + EFi,el�Eel,j,k + �EFi,fswitch�Egas,j,k

Mind,j,k, (2)

where Mind,j,k is the annual production from the industrial pro-cess j in the plant k (tonne industrial product/year), Monsite,i,j,k,CCis the annual onsite emissions from the industrial process j in theplant k equipped with CO2 capture (g/yr), �Eel,j,k is the change inannual electricity purchase from the grid for the industrial process j

in the plant k due to CO2 capture (MJe/yr), EFi,el is the EF of substancei for grid electricity (g/MJe), �EFi,fswitch is the EF of substance i forswitching centralized power plant fuel from natural gas to indus-trial process gas (g/MJLHV), and �Egas,j,k is the change in annual net

of Greenhouse Gas Control 10 (2012) 310–328 313

pp

te(N(

M

sWtsp

(ab2t2

eyciso2t

cocrsarmitttadE(ttat

fc

am

Table 2Emission factors (EF) used in this study.

Parameter Unit Value

Nominal Alternative

CO2 emission factor of fuelsa

Coal g/MJLHV 95 –Natural gas g/MJLHV 56 –

Grid electricity emission factors (EFi,Elec)CO2

b g/MJe 130 39NOx

b g/MJe 0.16 0.08–0.2SO2

b g/MJe 0.22 0.044PMc g/MJe 9.9 × 10−3 5.0 × 10−3

NH3c g/MJe 5.3 × 10−4 0.013

Emission factor for switching fuel from natural gas to steelmaking process gas(�EFi,fswitch)

SO2d g/MJLHV 0.01 –

PMd g/MJLHV 7 × 10−3 –NOx and NH3

d g/MJLHV Negligible –

High and low values represent ±30% of the nominal values.a See (IPCC, 2006).b The emission intensities of CO2, NOx , SO2 in public electricity and heat pro-

duction in the EEA32 (27 EU member states plus Iceland, Liechtenstein, Norway,Switzerland and Turkey) in 2007 were 620 g/kWh, 0.95 g/kWh (NO2 equivalent),and 2.2 g/kWh, respectively (EEA, 2011c). In the past decades, significant reduc-tions in the emission intensities of have been observed the three substances and weexpect a similar trend between 2007 and 2020. The average annual emission inten-sity reduction was 1.5%/yr between 1990 and 2007 and 2.3%/yr between 1990 and2007 for CO2; 4.4%/yr between 1990 and 2007 and 3.8%/yr between 2000 and 2007for NOx; and 7.0%/yr between 1990 and 2007 and 8.1%/yr between 2000 and 2007 forSO2. We therefore assumed annual reduction rates of 2% in CO2 emission intensity,4% in NOx emission intensity, and 7.5% in SO2 emission intensity, respectively (EEA,2011c). For alternative emission factors, oxyfuel CO2 capture reduces NOx emissionswhile post-combustion and pre-combustion capture shows unchanged or slightlyhigher emissions (Koornneef et al., 2010). Therefore an alternative NOx emissionfactor was assumed to range between 50% and 125% of the nominal emission level.The alternative CO2 and SO2 emission factor was assumed to be 30% and 20% of thenominal emission factor.

c The PM10 and NH3 emission intensities of public electricity and heat produc-tion in the EU-27 in 2007 were calculated to be 6.1 × 10−2 g/kWh and 1.9 × 10−3

g/kWh, respectively (data on the emissions from public electricity and heat produc-tion from (EEA, 2011a), data on the total outputs from public thermal power stationsfrom (EEA, 2011c)). Although there is no data available on the trend of PM emissionintensity in the previous years, PM emission intensity was assumed to reduce 4%annually, taking into account the annual emission intensity reduction rate for CO2,NOx and SO2. NH3 emission intensity was assumed to remain unchanged. The alter-native PM emission factor was assumed to be half of the nominal value. We assumethat post-combustion capture and oxyfuel capture, which reduces PM emissionssignificantly, are the more popular technologies compared to pre-combustion cap-ture. The alternative NH3 emission factor value was assumed to be a factor 25 higherthan the nominal value, which agrees with the range presented in Koornneef et al.(2010).

d

T. Kuramochi et al. / International Journal

rocess gas export from the industrial process j in the plant k toower plants due to CO2 capture (MJLHV/yr).

Lastly, the total changes in NECD substance emissions forhe entire EU-27 was calculated by extrapolating the aggregatedmissions for the existing European plants with emissions datacollected in the second step). The EU-27 sectoral emissions ofECD substances for the industrial process j with CO2 capture

MEU,i,j,CC: kt/yr) were calculated as follows:

EU,i,j,CC =MEU,i,j,ref

∑k

(Emi,j,k,CCMind,j,k)

∑k

(Emi,j,k,ref Mind,j,k), (3)

where, MEU,i,j,ref is the reference total emissions of NECD sub-tance i for the industrial process j in the EU-27 in 2020 (kt/yr).

hen there is no plant-specific industrial production data (Mind,j,k)hat corresponds with the emissions data collected in the secondtep, we assumed a typical annual production level for all Europeanlants.

The reference European sectoral emissions in year 2020MEU,i,j,ref) were assumed to be equivalent to those in 2007 forll NECD substances and industrial sectors. This assumption cane justified for two reasons. Firstly, a baseline scenario (“National010 Baseline”) in the Greenhouse Gas and Air Pollution Interac-ions and Synergies (GAINS) model developed by the IIASA (2011)indicates that both the industrial activity levels and the related

missions for the three sectors are projected to remain similar forears between 2000 and 2020 (projections for the entire industryan be found in, e.g., Amann et al. (2007); projections for individualndustrial sectors can be found in e.g., IIASA (2011)). Secondly, thecope of our research is to provide order-of-magnitude estimatesf the effect of CO2 capture, not exact predictions of emissions in020. Therefore, the use of 2007 emissions data would be valid forhis study.

Table 2 presents the parameters common for all industrial pro-esses to calculate the impact of CO2 capture on the emissionsf NECD substances. Parameters that are specific to certain CO2apture technologies or industrial processes are presented in theirespective sections. The nominal grid electricity EFs for NECD sub-tances in 2020 are calculated based on the EFs observed in 2007nd the assumption that the EFs will reduce at rates observed inecent years until 2020. This assumption implies that there is noajor introduction of CCS in the power sector. On the other hand,

f there is a large-scale deployment of CCS in the European indus-rial sector as assumed in this study, it is also natural to think thathere is a large-scale deployment of CCS also in the power sec-or. Therefore, the alternative grid electricity EFs which take intoccount the effect of CCS are defined for sensitivity analysis. Theefined emissions factors are based on the survey results on theFs available in the public literature presented in Koornneef et al.2010). The alternative EFs do not represent specific CO2 captureechnology and are only representative figures. Another impor-ant assumption is that no additional pollutant control measuresre included in cases the NECD substance emissions increase dueo CO2 capture.

In the following sections, key assumptions, research methods

or specific industrial processes, and results are presented and dis-ussed per sector.

2 The GAINS model deals with costs and potentials of air pollution controlnd greenhouse gas mitigation and assesses interactions between policies in theedium-term (until 2030).

For SO2 and PM, the mean emission values for gas-fired boilers and turbinesfrom more than 20 power plants using process gases from iron and steel works(IPPC, 2011).

4. Cement production

For the cement sector, cement kiln is the most CO2-emittingprocess and where CO2 capture process will likely be fitted to.Therefore, we considered the emission changes in cement plantswith cement kiln.

4.1. Reference production process and plant specific emissionsdata

Cement production can be distinguished by the moisture con-tent of the feed going into the kiln. Today’s best available technique(BAT) 3 is based on the dry process (IEA GHG, 2008). Dry process

3 The term “best available techniques” is defined by the EU Directive (EuropeanCommission, 2008) as “the most effective and advanced stage in the development

314 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

Table 3Reference European cement plant specification and the range of NECD substance concentrations in the clinker kiln flue gas investigated in this study.

Parameter Unit Symbol Value

Reference process specificationClinker/cement ratioa kg/kg 0.8Flue gas CO2 concentrationb vol% 22.5Flue gas O2 concentrationc vol% 10Flue gas NO2/NOx ratiod mol/mol 0.05Specific flue gas volume 103 Nm3/t clinker VSp,Cem 2.3Specific CO2 emissione tCO2/t clinker 0.81

Range of NECD substance concentrations in the flue gas investigated (typical values observed among European plants)NOx (as NO2)f mg/Nm3 CNOX,Cem 150–1200SO2

g mg/Nm3 CSO2,Cem 50–400vPMh mg/Nm3 CPM,Cem 0.27–30

NH3i mg/Nm3 CNH3,Cem 10–35

Data source: Reference: IPPC (2009), unless otherwise stated.a IPPC (IPPC, 2009) uses the ratio of 0.8 for their calculations. This values agrees with the arithmetic average for 2007–2008 (0.779) calculated based on the “Getting the

Numbers Right” database developed by World Business Council for Sustainable Development and PriceWaterhouseCoopers (Holcim, 2010).b Flue gas CO2 concentration ranges around 15–30% (IEA GHG, 2008).c Based on IEA GHG (2008).d This value has been reported for the rotary kiln flue gas from a number of existing plants in Europe. The literature values range between 2% (Leibacher et al., 2006) and

10% (IEA GHG, 2008).e The value is calculated using the presented assumptions and it agrees with the range (0.73–0.93 tCO2/t cement) indicated in Mahasenan et al. (2005).f NOx emissions are largely attributable to fuel combustion in the clinker furnace. The NOx formation level is around 1 g/Nm3 and the generated NOx is reduced by selective

non-catalytic reduction (SNCR) in most cement plants. With SNCR, NOx is reduced to N2 by injecting ammonia water (up to 25% NH3), ammonia precursor compounds orurea solution into the combustion gas at about 830–1050 ◦C. About 150 of 258 reported installations in Europe show emissions levels below 0.8 g/Nm3.

g SO2 is the main sulfur compound (99%) emitted from cement plants and its emission levels are primarily determined by the content of the volatile sulfur in the rawmaterials and possibly by the fuels. SO2 emissions originate in various outputs of the kiln system: exhaust gases, CaSO4 and other combined compounds in the clinker, andPM. Around two-thirds of the reported installations in Europe show emission levels below 0.2 g/Nm3.

hievedp

ia pro

cppaTase

fssso

cakaoar

4

Hmnc

ooda

h Continuously measured emissions. Emission levels of below 10 mg/Nm3 are aclants were less than 30 mg/Nm3.

i NH3 is emitted via ammonia slip value from de-NOx process (SNCR) as well as v

ement plants account for about 90% of European total cementroduction (IEA GHG, 2008). In this study, a dry process cementlant was assumed. Table 3 shows the data on typical productionnd the emission levels of the reference European cement plant.he emissions of NECD substances from cement plants are largelyttributable to fuel combustion. Note that cement plants exten-ively use waste as a fuel, which largely affects the pollutants’mission profiles (IPPC, 2009).

For NOx, SO2, and PM, flue gas concentration data are availableor most European cement plants. Large differences in NECD sub-tance concentrations among plants can be seen. Therefore in thistudy, the effect of CO2 capture on the emissions of these sub-tances was assessed for a range of concentration levels insteadf an average value.

To calculate EU-27 sectoral NECD substance emissions with CO2apture (MEU,Cem,CC), the data on the concentrations4 of NOx, SO2,nd PM in cement process flue gas for 257 out of 268 cementilns in EU-27 (IPPC, 2009) was used in this study. The data avail-bility for NH3 emissions is limited, thus we used an average EFf 5.7 gNH3/t clinker based on 2007 statistics and a number ofssumptions (270 Mt/yr cement as in Section 2, clinker-to-cementatio of 0.8 as in Table 3, and 1.2 kt NH3/yr emissions as in Table 1).

.2. Short-term CO2 capture technology

Based on an assessment made by the IEA GHG (2008) andegerland et al. (2006), post-combustion CO2 capture using

onoethanolamine (MEA) is considered to be the only tech-

ology that is feasible in the short term. Post-combustion CO2apture is conceptually an end-of-pipe technology that does not

f activities and their methods of operation which indicate the practical suitabilityf particular techniques for providing in principle the basis for emission limit valuesesigned to prevent and, where that is not practicable, generally to reduce emissionsnd the impact on the environment as a whole”.4 Continuous measurement results.

in 37% of the 253 reported installations in Europe, the dust emissions from most

cessing of raw materials.

require fundamental changes in the clinker production process(Mott MacDonald, 2010). In this study, it was assumed that thecement production process itself is unaffected by chemical absorp-tion CO2 capture. This assumption is also used in previous studies(Hegerland et al., 2006; IEA GHG, 2008).

Table 4 shows the flue gas specifications and the pollutantremoval performance for this technology. It was assumed thatCO2 is captured from all process gas flows (IEA GHG, 2008). Thesolvent was assumed to be a 30 wt% MEA solution. Flue gas spec-ifications for contaminants such as NO2, SO2, and PM are definedbecause these contaminants have significant negative impacts onthe performance of CO2 capture. Therefore, flue gas may need pre-treatment before it enters the CO2 scrubber unit to assure goodtechnical and economical operation. For NO2 control, techniquessuch as selective catalytic/non-catalytic reduction (SCR/SNCR) areoften used. It was assumed that the flue gas NO2/NO ratio is unaf-fected by the de-NOx process. For SO2 control, the following threetechniques can be used: absorbent addition, wet scrubbing, andactivated carbon filter (IPPC, 2010b). Among them, the wet scrub-bing process using CaCO3 is likely to be used to meet the CO2scrubber specifications (Table 4). Wet scrubbing also removes otherair pollutants, notably particulates. For PM control, techniques suchas electrostatic precipitation (ESP) and bag filters are used. In thisstudy, the conservative assumption was made that for additionaldesulfurization prior to the CO2 scrubber, co-removal of other NECDsubstances is not taking place.

4.3. Energy requirement and energy supply options forpost-combustion CO2 capture in cement plants

Chemical absorption CO2 capture using MEA requires signifi-cant amounts of heat to regenerate the solvent. The reboiler duty

of chemical absorption capture process is affected by the flue gasCO2 concentration. A modeling study by Kothandaraman et al.(2009) reports 4.5 GJth/tCO2 for natural gas combustion gas and4.2 GJth/tCO2 for coal combustion gas. The difference is due to the

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 315

Table 4Flue gas specifications and pollutant removal performance for post-combustion CO2 capture using monoethanolamine (MEA).

Parameter Unit Symbol Value

Flue gas specificationNO2 contenta mg/Nm3 CNOX,lim

20SO2 contentb mg/Nm3 CSO2,lim

29PM contentc mg/Nm3 CPM,lim 15

Post-combustion CO2 capture performanceCO2 removal efficiencyd % 85NOx removal efficiencyd % rNOX NO2: 25, NO: 0SO2 removal efficiencyd % rSO2 90PM removal efficiencyd % rPM 50

NH3 emissions due to additional de-NOx

NH3 consumption ratee mol NH3/mol NO reduced 1.5Ammonia slip rated %-total NH3 consumption 1

a Based on IEA GHG (2004).b The nominal value has been suggested by several studies for 30 wt% MEA solvents (Chapel et al., 1999; IEA GHG, 2004). For Kerr McGee/ABB Lummus Crest solvent, which

is a 15–20 wt% MEA solution, it has been suggested that SO2 removal is economically unjustifiable for concentrations below 50 ppm (142 mg/Nm3) (Barchas and Davis, 1992).c Based on Chapel et al. (1999).

NOx

w O ratio

ric

H

xtctEedeirat

G

P

nnaohutvwuCmpspr

se

d Based on Koornneef et al. (2008).e For SNCR, the ratio depends on the initial NOx concentration and the degree ofhich is consistent with the post-combustion CO2 capture requirements, at NH3/N

ich loading that can be achieved for coal combustion gas. Assum-ng 3% CO2 for natural gas combustion gas and 13% CO2 for coalombustion gas, the following linear equation was derived:

Sp = −3x + 4.59 (0.03 ≤ x ≤ 0.13), (4)

where HSp is the specific reboiler duty (GJth/tCO2 captured) and is the CO2 concentration (v/v). The assumption of a linear rela-ion between specific reboiler duty and flue gas CO2 concentrationan be justified for the range of CO2 concentration considered inhis study (3–13 vol%) according to literature (e.g., Alie et al., 2005;gberts et al., 2003)5. The range for reboiler duty is not the low-st reported in the literature for existing solvents. Specific reboileruty for new plants will likely be much lower. For retrofits, how-ver, reboiler duty would be somewhat higher due to limited heatntegration. Specific reboiler duty is indicated to be 30% higher foretrofits compared to new plants (Egberts et al., 2003). Taking intoccount that our study assumes retrofit CO2 capture, we considerhe assumed range for specific reboiler duty to be valid.

The equation for calculating specific power requirement (PSp:Je/tCO2 captured) is taken from Chapel et al. (1999):

Sp = 0.014/x + 0.035. (5)

Cement plants often do not have a usable heat source of sig-ificant scale in their vicinity. Consequently, solvent regenerationeeds heat either generated onsite by a boiler or a combined heatnd power (CHP) plant, or imported from external sources. Becausef its high energy efficiency, the use of CHP plants for CO2 captureeat generation was investigated. In this study, we considered nat-ral gas and coal as possible CHP fuels. Moreover, we consideredwo possibilities with regard to CO2 generated by the CHP plant: (1)ented (“stand-alone CHP” configuration), or (2) captured togetherith the CO2 from the industrial process (“integrated CHP” config-ration”. CHP plants that would be installed to supply energy forO2 capture are scaled on the basis of CO2 capture heat require-ent and do not cover other heat demands within the industrial

lant. In “integrated CHP” configurations, the CHP plant supplies

ufficient heat to capture CO2 from both cement flue gas and CHPlant flue gas. Table 5 presents three heat supply options and theirespective emissions profiles. In this study, we assumed that CHP

5 The linear relation is not valid for CO2 concentrations above 15 vol%, wherepecific reboiler duty is reported to become nearly constant (Alie et al., 2005; Egbertst al., 2003).

reduction. Two Swedish plants reduce NOx emission levels to below 200 mg/Nm3,s of 1.2–1.8 (IPPC, 2009).

plants installed for post-combustion CO2 capture were consideredto be equipped with BAT.

4.4. Calculation of Specific NOx, SO2 and PM emissions fromcement production with post-combustion CO2 capture

Fig. 2 shows the scheme for the calculation of NOx, SO2 and PMemissions from an industrial plant with post-combustion CO2 cap-ture for different CO2 capture heat supply options. For NO2, SO2and PM, specific emissions for plant k with post-combustion CO2capture (Emi,j,k,CC) with integrated CHP are calculated as follows:

Emi,j,k,CC = min{Ci,lim(VSp,j + VSp,CHP), Ci,jVSp,j

+Ci,CHPVSp,CHP}(1 − ri), (6)

where VSp,CHP is the specific flue gas volume from the CHP plant(Nm3/t industrial product).

For cases with stand-alone CHP, specific emissions are calcu-lated as follows:

Emi,j,k,CC = min{Ci,lim, Ci,j}VSp,j(1 − ri) + Ci,CHPVSp,CHP. (7)

Specific total NOx emissions are calculated as the sum of specificemissions of NO2 and NO, which act differently in the post-combustion CO2 capture process. For cases with integrated CHP,the emissions are calculated as follows:

EmNOX,j,k,CC = EmNO2,j,k,CC

{1 + CNO,j

CNO2,j(1 − rNO2 )

}. (8)

For cases with stand-alone CHP, the emissions are calculated asfollows:

EmNOX,j,k,CC = min{CNO2,lim, CNO2,j}VSp,j

{CNO,j

CNO2,j+ (1 − rNO2 )

}

+(CNO2,CHP + CNO2,CHP)VSp,CHP. (9)

4.5. Calculation of specific NH3 emissions from cementproduction with post-combustion capture

NH3 emissions are calculated differently than other NECDsubstances because NH3 does not negatively affect the post-combustion capture process and is formed in the CO2 captureprocess. Besides the slip from de-NOx installation, NH3 can be

316 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

Table 5Three heat supply options and their emissions profile.

Heat supply configuration I-PC I-NG S-NG

Configuration Integrated PC-CHP Integrated NG-CHP Stand-alone NG-CHPTreatment of CO2 from CHP plant Capture together with CO2

from industrial processCapture together with CO2

from industrial processVented

CHP plant type Pulverized coal combustionboiler + steam turbine

Combined cycle Combined cycle

Steam generation efficiencya 50% 50% 50%Electricity generation efficiencya 30% 40% 40%

Concentration of pollutants in the CHP flue gasNOx (as NO2; mg/Nm3)b 100 50 50NO2/NOx (vol%)c 5 10 10SO2 (mg/N m3)d 200 0 0NH3 (mg/N m3)e 5 0 0PM (mg/N m3)f 10 0 0Flue gas O2 concentration (vol%) 6 15 15

Data source: IPPC (2006), unless otherwise stated.a For PC-CHP plants, an exergetic efficiency of 45–55% and an overall energy efficiency of 75–90% have been suggested for BAT coal fired CHP plants (IPPC, 2006).b For coal-fired PC-CHP plants, emission levels of 90–300 mg/Nm3 for 50–200 MWth plants, 90–200 mg/Nm3 for 100–300 MWth plants, and 50–200 mg/Nm3 for >300MWth

plants have been suggested. For NGCC-CHP plants, emission levels of 20–50 mg/Nm3 have been suggested for new plants. For both coal and gas-fired plants, we made theconservative assumption of taking high-end values (IPPC, 2006).

c Based on Tzimas et al. (2007).d For PC-CHP plants, emission levels of 150–400 mg/Nm3 for 50–100 MWth plants, 100–200 mg/Nm3 for 100–200 MWth plants, and 20–200 mg/Nm3 for >300 MWth plants

have been suggested (IPPC, 2006). No SO2 emissions were assumed for NGCC-CHP plants (IPPC, 2006).e No NH3 emissions are expected for NGCC-CHP plants as a dry low-NOx burner was assumed to be used.f For PC-CHP plants, emission levels of 5–20 mg/Nm3 for 50–300 MWth plants and 5–10 mg/Nm3 for >300 MWth plants have been suggested. No PM emissions are expected

for NGCC-CHP plants (IPPC, 2006).

nd th

fitV

[

Fig. 2. Flue gas flows to the post-combustion CO2 capture unit a

ormed as a result of oxidative degradation of MEA by flue gasmpurities such as SO2, NO2 and O2. In this study, the oxida-

ive degradation rate of MEA is defined as follows (adapted fromeltman et al., 2010):

NH3] = [MEA]ox-deg = 2fSO2 [SO2] + 2fNO2 [NO2] + fO2 [O2] , (10)

e need for additional flue gas treatment prior to CO2 scrubbing.

where [NH3] is the NH3 formation rate (mol/s), [MEA]ox-deg isthe oxidative degradation rate of MEA (mol/s), fSO2 is the reaction

efficiency of SO2 with MEA (=SO2 removal efficiency: 90%), fNO2 isthe reaction efficiency of NO2 with MEA (=NO2 removal efficiency:25%), [SO2] is the SO2 flow rate (mol/s), [NO2] is the NO2 flow rate(mol/s), fO2 is the reaction rate of MEA with O2 (1.8 × ×10−4 mol

of Gre

Mao

bdg9MaMvtdRtuR

4

4

trtt

4c

pttsese4b(optsop

aOCMeCS

aIca(Sp

A(

In this study, CO2 capture from petroleum refinery furnaces andboilers were investigated. CO2 capture from existing onsite CHP andgas turbine plants was assumed to be performed separately from

6 The GAINS model does not provide any data on NH3 emissions from the cementsector. We assumed an emission factor of 20 mg/Nm3 flue gas, which is within therange observed in European plants (1–26 mg/Nm3) (IPPC, 2010b). For PM emissions,

T. Kuramochi et al. / International Journal

EA/mol O2) and [O2] is the O2 flow rate (mol/s). The underlyingssumption is that each mole of MEA degraded produces one molef NH3 (Rao et al., 2004; Veltman et al., 2010).

The fO2 value was calculated based on an experimental studyy Goff and Rochelle (2004), which reports the O2-induced degra-ation rate of MEA to be 0.29–0.73 kg/t CO2 captured for a flueas containing 3% CO2 and 5% O2, and 30 wt% MEA solution at0% capture efficiency. Using the geometric mean value of 0.46 kgEA/t CO2 captured as suggested by Veltman et al. (2010), we

ssumed the O2–induced oxidative degradation rate to be 0.34 kgEA per tonne of O2 in the flue gas, which corresponds to the fO2

alue presented above. Note that the MEA degradation rate dueo O2 is mass transfer dependent rather than kinetic rate depen-ent for solvents with MEA concentration above 30 wt% (Goff andochelle, 2004). It is also reported that the NH3 formation due tohe degradation of MEA is linear with flue gas O2 concentrationp to 17 vol% and increases at larger O2 concentrations (Goff andochelle, 2004).

.6. Results

.6.1. Energy and CO2 balanceTable 6 shows the CO2 concentration of the process flue gas at

he CO2 scrubber inlet, the CO2 capture rate and the specific heatequirement for CO2 capture solvent regeneration. CO2 concentra-ions are significantly lower for the integrated CHP option becausehe CHP flue gas dilutes the cement kiln flue gas.

.6.2. Changes in NECD substance emissions by flue gasoncentration level

Fig. 3 shows specific emissions of NECD substances from cementlants (Emi,Cem: g/t clinker) with and without CO2 capture as a func-ion of cement kiln flue gas concentration (Ci,Cem), and Fig. 4 showshe breakdown of specific emissions by onsite emissions and emis-ion credits for avoided centralized electricity production whenquipped with CO2 capture. It is shown that the selection of heatupply configuration affects the results significantly. Specific NOx

missions (EmNOX,Cem) remain constant for S-NG for CNOX,Cem above00 mg/Nm3, at which NO2 concentration reaches the CO2 scrub-er limit (CNOX,lim: 20 mg/Nm3). For integrated heat supply optionsI-PC and I-NG), in contrast, EmNOX,Cem does not curb for the rangef CNOX,Cem investigated because CHP flue gas dilutes the cementlant flue gas with regard to NO2, thus additional de-NOx prior tohe CO2 scrubber is not required. Both I-PC and I-NG shows emis-ion levels similar to the reference case because the increase innsite emissions is offset by the avoided emissions in centralizedower plants.

Specific SO2 emissions (EmSO2,Cem) were found to become neg-tive for all heat supply options and CSO2,Cem levels investigated.nsite SO2 emissions are reduced by about 90% or more by theO2 capture process together with possible flue gas pre-treatment.oreover, emissions reduction is credited for avoided centralized

lectricity production due to excess electricity production by theHP plant. EmSO2,Cem is therefore sensitive to the grid electricityO2 EF (Fig. 4).

Specific PM emissions (EmPM,Cem) are reduced significantly forll heat supply options. For integrated heat supply options (I-PC and-NG), cement plant flue gas is diluted by the CHP flue gas. The PMoncentration of the combined flue gas therefore reaches CPM,limt much higher CPM,Cem than for stand-alone heat supply optionsoutside of the figure). Negative EmPM,Cem are observed for I-NG and-NG because of emission credits for avoided centralized electricity

roduction (Fig. 4).

Finally, NH3 emissions increase for all heat supply options.voided NH3 emissions due to electricity export are negligible

Fig. 4). The emission levels for S-NG is a factor 3–4 lower than

enhouse Gas Control 10 (2012) 310–328 317

the other two options mainly because of the lower O2-inducedMEA degradation as CO2 is captured only from cement processflue gas. In order to examine the approach to calculate NH3 emis-sions in this study, we compared the NH3 formation rate pertonne of CO2 captured (without water wash) obtained in thisstudy with values reported in the literature. For higher flue gasCO2 concentrations, for instance, Koornneef et al. (2008) report0.21 kgNH3/tCO2 captured for PC power plants (about 12–14 vol%CO2). In this study, the I-PC case showed a NH3 formation rate of0.32 kgNH3/tCO2 captured for 14.3 vol% flue gas CO2 concentration,which is similar to that of PC power plants. For lower CO2 concen-trations, Statoil (Statoil, 2005) reports a NH3 formation rate of about0.4 kgNH3/tCO2 captured for NGCC plants (generally 3–4 vol% CO2).For the I-NG case in this study with zero SO2 content in the cementflue gas (to enable a fair comparison with natural gas combus-tion flue gas which generally contains negligible amount of SO2),the NH3 formation rate was calculated to be 0.28 kgNH3/tCO2 cap-tured for 5.7 vol% CO2 concentration. These results strongly indicatethat the approach applied for the calculation of NH3 formationis valid.

4.6.3. Changes in European sectoral emissionsTwo assumptions were made for the calculation of total sectoral

emissions from cement sector with CO2 capture (MEU,NH3,Cem,ref ).First, it was assumed that all the NECD substances are emittedthrough the cement kiln flue gas, from which CO2 is captured. Sec-ond, an identical clinker production scale per plant (Mind,Cem,k) wasassumed for all plants because there was no plant-specific clinkerproduction data that corresponds with the emissions data collectedin this study.

We therefore refer to Table 1 to estimate the total emissions forNOx, SO2 and NH3; MEU,NOx,Cem,ref , MEU,SO2,Cem,ref and MEU,PM,Cem,ref

are 3200 kt/yr, 1400 kt/yr and 15 kt/yr, respectively6.The results are presented in Fig. 5. The changes in net NOx

emissions due to CO2 capture ranged between a 50% decreaseand a 60% increase compared to the reference situation. The rangelargely depends on the CO2 capture heat supply plant option. NetSO2 emissions were found to decrease by 120–340%. The gridelectricity EF is shown to have a significant impact on the totalemission changes. As with SO2, PM emissions were also found toreduce significantly due to both, onsite emissions reduction andavoided emissions from centralized power plants. Regarding NH3emissions, our bottom–up estimate on MEU,NH3,Cem,ref was about5 kt/yr, which is considered to be in the right order of magni-tude when compared to process-related emissions reported in EEA(2011a,b) (Table 1). When equipped with post-combustion CO2capture, the EU-27 sector-wide NH3 emissions were calculatedto increase by 40–170 kt/yr as a consequence of the degrada-tion of MEA. The incorporation of CO2 capture from centralizedpower plants in the grid electricity EF has limited impact on theresults.

5. Refinery process furnaces and boilers

the share of cement kiln gas in total emissions is limited. Using the PM concentrationdata in IPPC (IPPC, 2010b), cement kiln gas only accounts for 10 kt/yr, which is lessthan 20% of sectoral total emissions. We assumed that CO2 capture only affectsthe emissions through cement kiln gas, and emissions from other cement plantprocesses are unaffected.

318 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

Table 6Technical results of post-combustion CO2 capture from the reference cement plant for three different CO2 capture heat supply options.

CHPconfiguration

Flue gas CO2

concentrationCO2 capture rate CO2 capture heat requirement Cogenerated

powerExcess powergeneration

CO2 avoidance rate

v/v t/t clinker GJ/tCO2 captured GJ/t clinker GJe/t clinker GJe/t clinker Onsite only Incl. electricityexport credits

I-PC 14.3% 2.6 4.2 11.0 6.6 5.2 54% 120%I-NG 5.8% 1.49 4.4 6.6 5.3 4.3 74% 140%S-NG 22.5% 0.86 3.9 3.4 2.7 2.3 48% 68%

-15 00

-10 00

-500

0

500

1000

50 100 150 200 250 300 350 400

EmSO

2,Ce

m(g

/t c

linke

r)

CSO2,Cem (mg/Nm3)

(b) SO2

-40

-20

0

20

40

60

80

0 5 10 15 20 25 30

EmPM

,Cem

(g/t

clin

ker)

CPM,Cem (mg/Nm3)

(c) PM

I-PC I-NG S -NG Referenc e

15 mg/Nm3 cli nker kilnflue gas (CPM,li m)

0

500

1000

1500

2000

2500

3000

150 300 450 600 750 900 1050 1200

EmN

Ox,

Cem

(g/t

clin

ker)

CNOX,Cem (mg/Nm3)

(a) NOX: NO2-equivalent

20 mg/Nm3 cli nker kilnflue g as (CNO2,li m)

0

100

200

300

400

500

600

700

800

900

1000

10 15 20 25 30 35

EmNH

3,Ce

m(g

/t c

linke

r)

CNH3,Cem (mg/Nm3)

(d) NH3

Fig. 3. Specific emissions of NECD substances for cement production (g/t clinker) with post-combustion CO2 capture as a function of pollutant concentration of each NECDsubstance cement flue gas. The flue gas specifications for the CO2 scrubber are 15 mg/Nm3 for PM, 20 mg/Nm3 for NO2, and 29 mg/Nm3 for SO2. NH3 emissions are calculatedfor 100 mg/Nm3 SO2 and 700 mg-NO2 eq./Nm3 NOx .

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 319

-150 0

-100 0

-50 0

0

500

1000

1500

2000

2500

3000

Reference I-PC I-NG S-NG Reference I-PC I-NG S-NG Refere nce I-PC I-NG S-NG Reference I-PC I-NG S-NG

NOx: 700mg NO2-eq/Nm3 SO2: 100 mg/Nm3 PM: 10 mg/Nm3 (x0.1t ) NH3: 20 mg/Nm3

Emi,C

em(g

/t cl

inke

r)

Onsite emissio ns on ly Ons ite + Electricity credits

NOX: 700 mgN O2-eq/Nm3 SO2: 100 mg/Nm3 PM: 10 mg /Nm3 (x 0.1) NH3: 20 mg/ Nm3

Fig. 4. Breakdown of specific emissions of NECD substances for cement production (g/t clinker) with post-combustion CO2 capture. The error bars show the results for casesCO2 is captured from grid electricity generation.

-10

0

10

20

30

Refe rence I-PC I-NG S-NG

Kt N

O2-e

q/yr

(b) PM

Onsi te emissi ons only Onsi te + El ectrici ty credits

0

200

400

600

800

Refe rence I-PC I-NG S-NG

kt N

O2-

eq/y

r

(a) NOX

0

50

100

150

200

Refe rence I-PC I-NG S-NG

kt N

O2- e

q/yr

(d) NH3

-400

-300

-200

-100

0

100

200

Reference I-PC I-NG S-NG

kt N

O2-

eq/y

r

(b) SO2

Fig. 5. Annual total emissions of NECD substances from the cement sector in the EU-27 countries with and without post-combustion CO2 capture using MEA. The error barsshow the results for the case CO2 is captured from grid electricity generation.

320 T. Kuramochi et al. / International Journal of Gre

Table 7Reference refinery combined stack flue gas specification and the range of NECDsubstance concentrations in the flue gas investigated in this study.

Parameters Unit Symbol Value

Reference plant specificationGas pressure Bar 1.01CO2 concentrationa vol% 10O2 concentration vol% 3Specific flue gas volumeb 103 Nm3/tCO2 VSp,RF 5.06

Range of NECD substance concentrations in the flue gasNOx (as NO2)c,d mg/Nm3 CNOX,Rf

100–500SO2

e mg/Nm3 CSO2,Rf20–1900

PMf mg/Nm3 CPM,Rf 20–700NH3

g mg/Nm3 CNH3,Rf 0.1–22

Data source: IPPC (2010a), unless otherwise stated.a Based on van Straelen et al. (2010), which indicates 8–12%.b Based on the CO2 density of 1.977 kg/Nm3.c Based on data reported by 31 European refineries (IPPC, 2010a).d The fraction of NO2 in total NOx was assumed to be 5%. NO contributes over 90%

of the total NOx in most combustion processes (IPPC, 2010a).e Based on data reported by 34 European refineries (IPPC, 2010a). A survey on 67

European refineries in 2006 suggests that about 4% of total sulfur intake in theserefineries has been emitted to the air in the form of SO2. Note that sulfur is some-times added to furnaces feed to minimize coke formation in the furnace so that theoperating cycles of the unit can be extended.

f Based on the data reported by 43 European refineries (IPPC, 2010a). Includes allparticulate sizes. The reported data in g/t of crude oil processed was converted byapplying specific flue gas volume of 1.0 × 103 Nm3/t of crude oil processed, whichwas taken from IPPC (2003).

g Based on data reported by 17 European refineries (IPPC, 2010a). NH3 may risefrom various sour water and sour gas streams or from gasifiers. The reported datain g/t of crude oil processed was converted by applying specific flue gas volume of1.0 × 103 Nm3/t of crude oil processed, which was taken from IPPC (2003).

fbfnc

5d

stbecdiaocioc

sesf

wa

2

urnaces and heaters because CHP plants and gas turbine plants wille large enough in scale to have its own CO2 capture unit. There-ore, flue gases from these existing CHP and gas turbine plants wereot combined with the flue gases from furnaces and boilers andonsequently not taken into account in this study.

.1. Reference production process and plant specific emissionsata

Table 7 shows the reference refinery combined stack flue gaspecification and the ranges of NECD substance concentrations inhe flue gas investigated in this study. For refinery furnaces andoilers, specific emissions (Emi,RF) were calculated for 1 t of CO2mitted from the reference plant (no CO2 capture) instead of 1 t ofrude oil processed because the fuel combustion rate in a refineryepends strongly on the degree of energy integration and complex-

ty of the refinery (IPPC, 2010a). The combustion plant capacity in refinery is reported to range widely from 1.7 to 5.4 GJ/t crudeil processed (IPPC, 2010a). The ranges of flue gas NECD substanceoncentrations represent the 5th–95th percentile values reportedn IPPC (2010a). As with the cement sector, we assessed the effectf CO2 capture on the NECD substance emissions for a range ofoncentration levels.

The range of NECD substance concentrations in the refinerytack gas investigated in this study (Table 7) is representative of

xisting European refineries. To calculate EU-27 sectoral NECD sub-tance emissions with CO2 capture (MEU,Rf,CC), the data inventoryor more than 30 out of 116 European refineries (IPPC, 2010a)7

7 There are no data for total PM emissions from furnaces and boilers. Therefore,e estimated the emissions based on the CO2 emissions from furnaces and boilers,

ssuming typical CO2 concentration of 10% (v/v) and an average PM concentration

enhouse Gas Control 10 (2012) 310–328

on NOx and SO2 emissions8 was used in this study. For PM andNH3, no plant-specific data on flue gas concentration levels wereavailable. These emissions were therefore estimated by assumingfor all refineries a typical CO2 concentration of 10 vol% an averagePM10 concentration of 15 mg/Nm3 (the 50th percentile value for the25 European refineries reported in (IPPC, 2010a)), and a NH3 con-centration of 10 mg/Nm3 (50th percentile value for the 17 Europeanrefineries reported in (IPPC, 2010a)).

The share of furnaces and boilers in EU-27 sectoral emissionswere assumed to be 65%, which is equivalent to the share for CO2emissions (IEA GHG, 1999) for all NECD substances. EU-27 sectoralemissions presented in Table 1 were multiplied by 65% to calcu-late MEU,NOx,Rf,ref (109 kt/yr), MEU,SO2,Rf,ref (337 kt/yr), MEU,PM,Rf,ref(11 kt/yr) and MEU,NH3,Rf,ref (0.6 kt/yr).

5.2. Short-term CO2 capture technology

The CO2 concentration of flue gases from petroleum refinerycombined stacks is about 8–12 vol% (van Straelen et al., 2010).Because of the low CO2 partial pressure, chemical absorption usingMEA is considered to be the only feasible CO2 capture option forflue gases from petroleum refinery combined stack gas in the shortterm (Kuramochi et al., 2012). As with the cement sector (clinkerproduction process), it was assumed that the refining process itselfis unaffected by chemical absorption capture. The flue gas specifi-cations and the technical performance of post-combustion captureare presented in Table 4. The need for additional flue gas pre-treatment prior to CO2 capture was assessed using the schemepresented in Fig. 2. With regard to CO2 capture heat supply, thefollowing two options were considered: (1) integrated NG-CHP (I-NG); and (2) stand-alone NG-CHP (S-NG). These CHP plants wereassumed to be newly built plants and not the existing ones inthe refinery because it was expected that the CO2 capture energydemand would be larger than the existing plants could supply.

5.3. Results

5.3.1. Energy and CO2 balanceTable 8 shows the CO2 concentration of the process flue gas at

the CO2 scrubber inlet, the volume of CO2 captured and the specificheat requirement for CO2 capture. CO2 concentrations are signifi-cantly lower for the integrated CHP options because the CHP fluegas dilutes the refinery flue gas. Note that CO2 from the CHP plantsnearly doubles the amount of CO2 captured.

5.3.2. Changes in NECD substance emissions by flue gasconcentration level

Fig. 6 shows specific emissions (g/tCO2 reference emissions) ofNECD substances from petroleum refinery combined stacks withand without chemical absorption CO2 capture as a function ofstack flue gas concentration. Fig. 7 shows the breakdown of specificemissions by onsite emissions and emission credits for avoided cen-tralized electricity production. The results are similar to those forthe cement sector because similar CO2 capture process is applied.The exception is PM emissions, which resulted in constant emissionlevel for the range of CPM,Rf assessed because additional controlmeasures prior to CO capture was required at all concentration

levels. However, the impact of credits for avoided centralized elec-tricity production is relatively smaller for SO2 and PM, and largerfor NOx compared to the case for cement plants because of the

of 180 mg/Nm3, which is a median value observed for European refineries (IPPC,2010a).

8 NOX emissions data were available for 31 refineries and SO2 emissions data for34 refineries.

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 321

Table 8Technical results of post-combustion CO2 capture from the refinery combined stack with two natural gas heat supply options.

CHP configuration Flue gas CO2

concentrationCO2 capturerate

CO2 captureheatrequirement

Cogeneratedpower

Excess powergeneration

CO2 avoidance rate

[v/v] [t/t referenceplantemissions]

[GJ/tCO2

captured][GJ/tCO2

referenceemissions]

[GJ/tCO2

referenceemissions]

Onsite only Incl. excesselectricity credits

I-NG 4.8% 147% 4.4 5.2 4.2 74% 128%S-NG 10.0% 85% 4.3 2.9 2.4 44% 76%

0

500

1000

1500

2000

2500

3000

100 150 200 250 300 350 40 0 450 500

EmN

Ox,

Rf(g

/tCO

2ge

nera

ted)

CNOX,RF ( mg NO2-eq/Nm3)

(a) NOX (NO2-equivalent)

20 mg NO2/Nm3 refinery stack ga s

-2000

-1000

0

1000

2000

3000

4000

5000

0 300 600 900 1200 150 0 180 0

EmSO

2,Rf

(g/t

CO2

gene

rate

d)

CSO2,RF (mg/Nm3)

(b) SO2

-100

0

100

200

300

400

500

0 50 100 150 20 0

EmPM

,Rf(

g/tC

O2

gene

rate

d)

CPM, RF (mg/Nm3)

(c) PM

I-NG S-NG Re fere nce

0

500

100 0

150 0

200 0

0 5 10 15 20

EmN

H3,

Rf(g

/tCO

2ge

nera

ted)

CNH3,RF (mg/Nm3)

(d) NH3

Fig. 6. Specific emissions of NECD substances for petroleum refinery furnaces and boilers (g/tCO2 reference plant emissions) with post-combustion CO2 capture as a functionof pollutant concentration of each NECD substance in the combined stack flue gas. NH3 emissions were calculated for 100 mg/Nm3 SO2 and 700 mg-NO2 eq./Nm3 NOx . Thehigh and low range bars show the sensitivity of the results to the solvent reboiler duty.

322 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

-100 0

-500

0

500

1000

1500

2000

2500

3000

Reference I-NG S-NG Reference I-NG S-NG Refere nce I-NG S-NG Reference I-NG S-NG

Nox: 250 mg NO2-eq/Nm3 SO2: 490 mg/Nm3 PM: 180 mg/Nm3 NH3: 0.1 mg/Nm3

Emi,R

f(g/

tCO

2re

fere

nce

plan

t em

issio

ns)

Onsite emission s onl y Onsit e + Electricit y credit s

NOX: 250 mg NO2-eq/Nm3 SO2: 490 mg/Nm3 PM: 180 mg/Nm3 NH3: 0.1 mg/N m3

Fig. 7. Breakdown of specific emissions of NECD substances for petroleum refinery furnaces and boilers (g/t stack gas CO ) with post-combustion CO capture. The error barss

dr

5

awr1b

6

6d

Eecateia

pGrtofi

sw

how the results for the case CO2 is captured from grid electricity generation.

ifferences in pollutant concentrations observed for cement andefinery flue gases.

.3.3. Changes in European sectoral emissionsFig. 8 presents the changes in NECD substance emissions when

ll the refinery furnaces and boilers in the EU-27 are equippedith post-combustion CO2 capture. It is shown that NOx emissions

educe by around 5% and SO2 emissions decrease by more than00% compared to the reference situation. PM emissions reduce,ut the degree of reduction differs largely by the heat supply option.

. Iron and steel production

.1. Reference production process and plant-specific emissionsata

An inventory of plant-specific data on air pollutant emissions foruropean plants is not available in the public literature. Moreover,nergy and material flows in integrated iron and steel plants areomplex, making bottom-up estimates on non-CO2 emissions for

reference iron and steel plant more difficult than for other sec-ors. Therefore, the rough assumption is made that the made thatmission data used in this paper is representative of the Europeanntegrated iron and steel plants. The potential consequences of thisssumption are discussed in Section 8.

The reference iron and steel production process is taken fromrevious work (Kuramochi et al., 2012), which is based on a typicalerman BF presented in Schmöle and Lüngen (2004). The average

eductant consumption of German BFs is similar to the plants inhe EU-15 (Schmöle and Lüngen, 2004), which accounts for mostf the integrated steelmaking plant capacity in the EU-27. There-ore, it is considered this reference to be representative of European

ntegrated steelmaking plants.

The emissions of NECD substances from European integratedteel plants were calculated by subtracting EAF-related emissions,hich are derived based on the EFs for European EAF-based steel

2 2

plants presented in IPPC (2012), from EU-27 total emissions for theiron and steel sector presented in Table 1. The breakdown of emis-sions by process is also derived from the information presentedin the IPPC (2001) . Fig. 9 shows the breakdown of annual NECDsubstance emissions of the reference integrated steel plant by pro-cess. It can be seen that the pelletizing and sintering processesare the main emitters in integrated steel plants. Table 9 describesthe detailed NECD substance emissions profile in typical integratedsteelmaking plants.

6.2. Short-term CO2 capture technology

In a conventional integrated steelmaking plant (Fig. 10a), gener-ally around 85% of the carbon introduced into the process is presentin three gas flows: approximately 70% in the BF gas, 9% in the cokeoven gas (COG) and 7% in the BOF gas (Farla et al., 1995). Thereforemuch research has been performed on CO2 capture from the BF orfrom the BF gas (see, for example, Kuramochi et al. (2012) for anoverview of technologies).

In the short–mid term, Top Gas Recycling Blast Furnace (TGRBF,presented in Fig. 10b), which is being developed by the EuropeanULCOS programme (Birat, 2010), is considered to be a promisingoption for cost-effective CO2 capture (Birat and Lorrain, 2009; Torp,2005). The techno-economic assessment by Kuramochi et al. (2012)also concludes that TGRBF is one of the short–mid technologiesthat may enable relatively low CO2 capture cost at relatively highCO2 removal rate. Therefore, this study selected the TGRBF tech-nology for further analysis. Note, however, that the TGRBF is notthe only CO2 capture technology that can be considered in theshort–mid term. There are many CO2 capture technologies applica-ble to integrated steelmaking plants (details of possible CO2 capturetechnologies are described and discussed in e.g. Birat and Maizière-

lès-Metz, 2010 and Kuramochi et al. 2012) . The best suited CO2capture technology will depend on the individual steel plant andthus, it is likely that there will be no dominant CO2 capture tech-nology either in the short-term or the long-term future.

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 323

-200

-100

0

100

200

300

400

Refere nce I-NG S-NG

kt/y

r

(b) SO2

0

50

100

150

200

Refe rence I-NG S-NG

kt /y

r(a) NOX

0

5

10

15

20

Refe rence I-NG S-NG

kt/y

r

(c) PM

Onsit e emission s Onsite + Electricity credits

0

50

100

150

Refere nce I-NG S-NGkt

/yr

(d) NH3

F EU-2s

iaCvgTpabi

yfr2cA

F(

S

ig. 8. Annual total emissions of NECD substances from the oil refining sector in thehow the results for the case CO2 is captured from grid electricity generation.

The main difference of TGRBF from conventional BF is that its oxygen-blown. Its process gas therefore contains little nitrogennd has CO2 and CO concentrations, which are reported to be 38%O2 and 47–48% CO in a modeling study (Danloy et al., 2008). Thesealues are significantly higher than the values for conventional BFas (17–25% CO2 and 20–28% CO (IPPC, 2001)). The CO2 contained inGRBF gas can be removed by various technologies such as vacuum-ressure swing adsorption (VPSA), physical absorption, or chemicalbsorption. TGRBF gas after CO2 removal is CO-rich and is recycledack to the furnace as a reducing agent, leading to an improvement

n BF performance.In this study, TGRBF using VPSA was considered for the anal-

sis. Among a number of CO2 removal technologies applicableor TGRBF, VPSA is technically the most feasible technology and

eported to be cheaper than other removal technologies (Torp,005). Moreover, this combination of technologies was tested suc-essfully at an iron production level of 1.5 t/h (Danloy et al., 2009).lthough there are competing technologies (e.g. PSA, chemical

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Others

Bas ic ox yge nfurna ce

Blast furnace

Pelle�zing andsinterin g

Coke ove n

ig. 9. Breakdown of specific emissions of NECD substances per tonne of crude steeltcs).

ource: Own calculations based on EEA (2011a) and IPPC (2001).

7 countries with and without CO2 capture (post-combustion: MEA). The error bars

absorption, etc.), VPSA can be considered as one of the more proba-ble technologies that will be combined with TGRBF for commercialCO2 capture in the short–mid term. Table 10 presents the keymaterial inputs and outputs of a conventional BF (air-blown) anda TGRBF. The major differences in energy and material flows arethe reduction in coke input and the consequent reduction in theamount of BF gas that can be used in or exported to other processes.For other processes, the volumes of intermediate products, i.e. ironore sinter and pellets, and BOF and hot roll mill, are unaffected byTGRBF CO2 capture.

TGRBF gas needs to be de-dusted for operating VPSA, but noadditional equipment is necessary because BF gas is usually wet-cleaned anyway before it can be redistributed (de Beer et al., 1998).BF gas contains pollutants such as PM, cyanides, NH3 and H2S, andis usually cleaned up in two stages before being used in other pro-cesses (IPPC, 2012)9. VPSA only uses electricity for CO2 removal, sothere are no onsite emissions of the NECD substances from this pro-cess. Moreover, VPSA does not remove any of the NECD substancesnor H2S.

Regarding the emissions from BF gas combustion, it wasassumed that PM and NH3 emissions from the preheaters areproportional to the firing capacity (in the TGRBF, pebble heatersare used instead of hot stoves). The assumption implies thatPM and NH3 concentrations (in mg/MJLHV heat content) are thesame for the conventional BF gas and the TGRBF gas. Moreover,it was assumed that the mix of process fuel gases for internalconsumption is unchanged because the process fuel gas produc-tion is barely sufficient for internal gas consumption (shown inTable 10). Table 11 summarizes the assumptions made for the

calculation of non-CO2 emissions from the iron and steel plantwith CO2 capture. Note that our assumptions on the emissions,especially SO2, from combustion processes may be conservative.

9 Coarse PM is removed in the first stage, and PM including zinc oxide and car-bon, cyanide and NH3 are removed in the second stage by wet scrubbing or wetelectrostatic precipitation (IPPC, 2011).

324 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

Table 9Description of emissions of NECD substances from integrated steel making plants.

SO2 NOx

Sulfur is predominantly emitted as SO2 and the rest as H2S. Nearly half of totalplant SO2 emissions. SO2 emissions is attributable to pelletizing andsintering processes, which are primarily from fuel combustion (coal and oilfor pelletizing, coke breeze for sintering) and to a lesser extent from iron ore.For BF, about half of the SO2 emissions is attributable to hot stove fuelcombustion and the other half to casting process and slag granulation (IPPC,2001). Large amount of sulfur is input to BF through coal and coke, but slagabsorbs the vast majority of it and most of the rest is contained in the hotmetal. Gases containing H2S and SO2 are generated when hot metal and slagare removed from the BF hearth and exposed to air (slag pit process) orwater vapor (granulating process) (IPPC, 2011; Pelton et al., 1974). The SO2

emissions from a hot rolling plant are due to the combustion of coke ovengas (COG) and BF gas.

More than half of total NOx emissions originate in fuel combustion in pelletizingand sintering processes. Emissions from blast furnace (BF) are mostly attributableto BF stove. Various mix of fuels are used in these processes.

NH3

Most of the NH3 emissions from iron and steel plants are from the coke oven andhot roll mill. In coking process, NH3 is emitted from coke quenching as well asduring coal charging. The formed NH3 appears in both COG and the condensatefrom COG (weak liquor), with typically 20–30% of the NH3 being found in the weakliquor. The emissions from hot roll mill are due to NH3 slip from the de-NOx unit.PMLarge part of PM is emitted through handling, crushing, screening, and conveyingraw materials or output products. Pelletizing and sintering processes are thelargest emitters, followed by coke oven and basic oxygen furnace.

Data source: IPPC (2012), unless otherwise stated.

Elec tricity CO2Fuels Materials Stea m

Coke oven

Blast furna ce

Coa l

Iron

ore

Sintering/

pelletizing

Coke

Air

Coke

oven gas

Pig

iron

ASU

O2

(a) Convention al blast furnac e

Powe r plant

(on site/offsite) CO2 vented

Melt

sho p

Finish-

ing millHot

rolled

coil

2-3 bar,

100 ºC

Air

ASU

(b) Top Gas Recyc ling Blast Furnace

Coke oven

Blast furnace

Coal

Iron

ore

Sintering/

pelletizing

Coke

Coke oven gas

Pig

iron

O2

CO2

cap tureCO2 for storag e

Power plant

(on site/offsite) CO2 vented

Melt

sho p

Finish-

ing mil lHot

rolled

coil

2-3 bar,

100 ºC

Electricity

from the grid

Fig. 10. Schematics of (a) conventional BF-based steelmaking process and (b) Top Gas Recycling Blast Furnace (TGRBF) steelmaking process.

Table 10Key material and energy flows per tonne of crude steel (tcs) for an integrated steel making process, with CO2 capture (TGRBF with VPSA) and without CO2 capture (air-blownBF). Note that the TGRBF requires a CO2 capture unit for the operation.

Input Unit Air-blown BF TGRBF + VPSA

Coke consumption t/tcs 0.34 0.24Coal injection to BF t/tcs 0.18 0.18Gas consumption in BF stove GJ/tcs 1.5 0.72Net gas consumptiona GJ/tcs −3.2 0.02Electricity consumption GJ/tcs 0.85 1.7CO2 capture rate t/tcs 0 0.72CO2 avoidance rate %-total carbon input to the reference integrated steel plant – 45

Source: Kuramochi et al. (2012).a Includes BF gas, coke oven gas and basic oxygen furnace gas.

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 325

Table 11Summary of assumptions for quantifying the effect of CO2 capture (TGRBF with VPSA) on the emissions of NECD substances from an iron and steel plant.

NECD substance/process NOx SO2 PMe NH3

Coke oven Proportional to coke consumptionBlast furnace Proportional to hot

stove/top gaspreheater fuelconsumptiona

Hot stove/top gaspreheater:Proportional to heatingfuel consumptionb

Unaffected Proportional to hotstove/top gas preheaterfuel consumptiond

Casting process andslag granulation:Proportional to total Sinput to BFc

Sinter/pellet plant Proportional to sinter and pellet consumptione

BOF Proportional to crude steel productionCombustion of exported BF gas Unaffected Proportional to the

amount of exported BFgas (in heat content)f

Unaffected Unaffected

Others Unaffected

a NOx formation depends largely on the flame temperature. Because the fuel mix for hot stove and top gas preheater can vary significantly by the plant, the conservativeassumption was made that the NOx formation rate per unit firing capacity is unaffected.

b The contributions of preheater fuel combustion (hot stove for conventional BF, pebble heaters for TGRBF) and other processes in SO2 emissions from BF were assumed tobe 50% each. For preheaters, a mix of various fuels (BFG, coke oven gas (COG), BOF gas and natural gas) is used and most of the SO2 emissions are attributable to BFG and COGwhich contain H2S. BFG and COG are equally important because BFG is low in H2S concentration but high in the volume consumed while COG is high in H2S concentrationbut low in the volume consumed. Note, however, that the contribution of COG becomes dominant if the COG is not desulfurized, which is not common in the EU but stillpracticed in some plants (IPPC, 2011). It is not known how the preheating fuel mix as well as the sulfur concentration in BFG may change when the conversion to TGRBF takesplace. Therefore, a conservative assumption was made that the sulfur concentration of TGRBF gas per unit heat content is unchanged between conventional BF and TGRBF.

c Since the vast majority of the sulfur input is absorbed by either slag or hot metal, this is a reasonable assumption. There is no information in the literature indicating achange in sulfur concentration in BF gas due to a nitrogen-free operation of the BF. Sulfur content (wt%) in coke is estimated to be 85% of that in coking coal, based on theinformation that generally 60–70% of total sulfur in coking coal remains in coke (Hu et al., 2008), and that 1 t of coking coal is converted into 0.8 t of coke (Diemer et al., 2004).Moreover, sulfur contents (wt%) in coking coal and non-coking coal were assumed to be the same.

d NH3 emissions were assumed to be entirely from de-NOx units.e The assumptions for PM emissions are conservative when the open source emissions are concerned. Regarding BF and coke oven, coal and coke consumption reduces

and it is therefore likely that some fugitive emissions related to coal coke handling from BF and coke oven will be reduced. However, there is no literature that suggests towhat extent these emissions will reduce. For the emissions from BF, half of them originate in cast house, and most of the rest in raw material charging and coal preparation( able.

n in Bc ional B

Aptrggp

Fo

IPPC, 2001). Since sinter and pellet inputs are unchanged, the assumption is reasonf There is no information in the literature indicating a change in H2S concentratio

oncentration of the TGRBF gas per unit heat content is identical to that of convent

lthough it is assumed that the NECD substance emissions are pro-ortional to the firing capacity, the possible reduction of COG dueo the reduced coke demand may lead to significant emissions

eduction. While most of the combustion processes in the inte-rated steel plant uses a mix of COG, BFG, BOF gas and naturalas, COG is considered to be more emissive than other fuel gases,articularly SO2.

ig. 11. NEC emissions from an iron and steel plant with top gas recycling and CO2 captuf the results to grid electricity emission factors.

F gas due to a nitrogen-free operation of the BF. We therefore assumed that the H2SF gas.

6.3. Results

Fig. 11 shows the relative emission level of NECD substances

from an iron and steel plant when CO2 is captured from the BF.The reduction in onsite SO2 emissions (6%) is mainly due to thereduction of coke consumption and the substitution of reduced BFgas combustion by natural gas, and these reductions are largely

re using VPSA compared with a reference plant. The error bars show the sensitivity

326 T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328

-80%

-70%

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

Cement Refinery Stee l Total Cement Refinery Steel Total Cement Refinery Stee l Total

NOX SO2 PM

Rela

�ve

chan

ge c

ompa

red

to to

tal E

U-2

7 in

dust

rial

emiss

ions

(0 =

refe

renc

e le

vel)

Onsite emiss ions +electricit y credit sOnsite emiss ions

NOX SO2

0%

100%

200%

300%

400%

500%

600%

Cement Refinery Steel Total

NH3NH3

Fig. 12. Relative changes in NECD substance emissions from key industrial processes due to CO2 capture compared to total industrial emissions in EU-27 in 2020. The dotsr ent Cr plus e

cirbw(ttC

oNaw

7

owCviFToew

tSet2cettp

epresent onsite emissions under nominal parameter sets (average values for differefining sectors). The gray bars represent total emission changes (onsite emissions

ancelled by the additional emissions due to increased electric-ty consumption. Onsite NOx emissions also reduce (10%) due toeduced coke consumption; the reduction is also partly cancelledy the grid electricity-associated emissions. PM and NH3 emissionsere found to reduce when the plant is equipped with CO2 capture

5% and 21%). The figure also shows that the indirect emissions dueo grid electricity consumption does not have large enough impacto change the ranking of integrated steel plants with and withoutO2 capture on NECD substance emissions.

With regard to the changes in sector-wide emissions, we extrap-lated the calculated specific emissions to the EU-27 sectoral level.Ox and SO2 emissions were found to decrease by about 25 kt/yrnd 16 kt/yr, respectively. The changes in PM and NH3 emissionsere found to be negligible.

. Potential impact on European industrial emissions

In the previous sections, the potential changes in the emissionsf NECD substances due to CO2 capture in three industrial sectorsere calculated. In order to understand these potential impacts ofO2 capture in a broader context, the results obtained in the pre-ious sections were compared with the total industrial emissionsn the EU-27 (the last row of Table 1). The results are presented inig. 12 for both individual sectors and the three sectors combined.he diamond-shaped plots represent the results onsite emissionsnly and the gray bars represent the sum of onsite emissions andlectricity emission credits using EFs for grid electricity with andithout CO2 capture (the range is presented in Table 4).

When all three industrial sectors in the EU-27 investigated inhis study were fully equipped with CO2 capture, total industrialO2 emissions may decrease by 40–70% while total industrial NH3missions would increase by 120–520%, which is about 2–8% ofotal European NH3 emissions. Industrial NOx emissions in the EU-7 may increase by up to 10% due to the cement sector whenoal-fired CHP plants are used to supply CO2 capture energy. PM

missions are also found to reduce by up to 10% of European indus-ry total. The changes in industrial NECD substance emissions dueo CO2 capture will be attributable mainly to the cement and theetroleum refinery sectors.

O2 capture energy supply options are presented for the cement and the petroleumlectricity credits) with and without CO2 capture from grid electricity generation.

8. Limitations of the study

One of the main limitations of this study concerns the avail-ability of the plant-specific emissions data inventory for existingEuropean plants. For the petroleum refineries, we assumed that thedistribution of flue gas emission levels observed for the surveyedplants can be applied to the entire European petroleum refiningsector. The uncertainty due to this assumption may affect the SO2emission results.

Data availability was particularly an issue for the iron and steelsector. Firstly, the lack of plant-specific emissions data resulted inmaking the rough assumption that the emissions from the refer-ence integrated steel plant were representative of all integratedsteel plants in the EU-27. This assumption itself is unlikely to affectthe sectoral results for the CO2 capture technique studied in thisarticle (TGRBF with VPSA) because the changes in NECD emissionswere found to be limited compared to the total EU-27 emissionsand the selected CO2 capture process does not have specificationson NECD substances for the gas flows. However, it may cause largeuncertainties when other CO2 capture technologies are considered.Our conclusions are therefore limited to the impact of TGRBF inNECD emissions. It should also be stressed that an (open) inven-tory of plant-specific emissions data will be necessary for futurein-depth assessment of the iron and steel sector.

The second issue regarding the iron and steel sector relates tothe assumptions on the operation of coke oven batteries in caseof CO2 capture. This study assumed that the emissions from thecoke oven are proportional to the coke input to the BF. However,the adaptation of coke oven batteries due to the change in cokedemand at the BF may be plant-dependent. The adaptation of cokeoven batteries affects not only the direct emissions from coke ovenbatteries but also the production of COG, which has EFs for NOx,SO2 and NH3. Because coke oven accounts for significant fraction ofvarious NECD substance emissions from the integrated steel plant,more focus should be on the coke oven operation and the relevantmaterial flows in the future research.

Thirdly, due to the specificity of process integration and the lackof plant-specific emissions profile for European steel plants, infor-mation from the Tata Steel IJmuiden was often referred to in thestudy. Since Tata Steel IJmuiden plant is one of the most efficient

T. Kuramochi et al. / International Journal of Greenhouse Gas Control 10 (2012) 310–328 327

Table 12Summary of the potential changes in emissions (onsite emissions plus electricity credits) due to CO2 capture from three key industrial processes relative to reference emissions(0% = reference level). The increased emissions are marked ↑ and reduced emissions are marked ↓.

Sector CO2 NOx SO2 PM NH3

Cement (vs EU-27 sectoral) 50–140% ↓ 50 ↓–60% ↑ 120–340% ↓ ∼150% ↓ Factor 30–140 ↑10 ↓–10%

10 ↑–

stisdfout

patstTt

Nptgsotfd(stswcte

iwis

9

psiswbtttwnCs

Refinery (vs EU-27 sectoral) 50–130% ↓

Steel (vs EU-27 sectoral) ∼50% ↓

Three sectors combined (vs total EU-27 industrial) 25–50% ↓

teel plants in Europe, its emissions profile may not be represen-ative of typical European integrated steel plants. Therefore, morenformation on plant-specific emissions profile for other Europeanteel plants is necessary to conduct more robust analysis. Third, aetailed process analysis is necessary to obtain more reliable resultsor CO2 capture technologies that require modification at the coref the industrial process like the TGRBF technology. Although thencertainty of the results due to the lack of data would be limited athe European sectoral level, it may be significant at the plant level.

For the cement sector, there were sufficient emissions data forlants with cement kiln (data for 257 plants of total 268). It wasssumed that all European cement plants are of identical scale, buthis does not affect the results on SO2 and NH3 emissions becausepecific emissions for these two substances are nearly constant forhe range of flue gas concentrations observed for European plants.he results for NOx and PM, however, may be more sensitive thanhose for SO2 and NH3.

Regarding system boundaries, this study took into account theECD substance emissions associated with changes in centralizedower generation due to CO2 capture in the industrial sectors inves-igated. The results have shown that the emissions associated withrid electricity have a significant impact on total NOx and SO2 emis-ions. However, literature indicates that other indirect emissionsf NECD substances, which are outside the system boundary ofhis study, may also be significant. Indirect emissions from coaluel preparation and transport is found to be as important as theirect emissions for the case of coal-fired plants with CO2 captureKoornneef et al., 2008; van Harmelen et al., 2008a). These emis-ions should therefore be taken into account to fully understandhe consequences of CO2 capture on the NECD substance emis-ions. From the results presented in van Harmelen et al. (2008b),e expect that considering secondary and tertiary emissions from

oal use would result in increased emissions for chemical absorp-ion CO2 capture process using coal-fired CHP plants and reducedmissions for TGRBF steel making process.

With regard to NH3 emissions results obtained in this study, its important to note that a water-wash of CO2 scrubber outlet gas

as not considered as discuss previously in the introduction. Fornstance, a water wash can remove 95% of the NH3 as well as otherubstances such as evaporated MEA (Veltman et al., 2010).

. Conclusions

The objective of this study was to quantitatively assess theotential impact of CO2 capture on the emission of NECD sub-tances (PM, SO2, NOx and NH3) in Europe (EU-27) from keyndustrial sectors when applying CO2 capture technology in thehort term (timeframe: 2020). The following industrial processesere investigated: cement, petroleum refineries (furnaces and

oilers), and iron and steel (integrated steel plants). We analyzedhe impact on onsite emissions as well as emissions associated withhe changes in centralized electricity production due to CO2 cap-ure. We investigated the changes in emissions at the plant level as

ell as at the EU-27 sector level. For the cement sector and for fur-aces and heaters used in petroleum refineries, post-combustionO2 capture using MEA was considered. For the iron and steelector, oxygen-blown TGRBF with CO2 capture using VPSA was

15% ↑ 70–80% ↓ 5–60% ↓ factor 30–130 ↑↓ 5–15% ↓ 5% ↓ No significant change

10% ↓ 40–70% ↓ 0–10% ↓ 120–520% ↑

selected in our analysis. For post-combustion CO2 capture, differ-ent solvent regeneration heat supply options were investigated.The main differences were: (1) fuel type (natural gas or coal) and(2) the treatment of CO2 from the heat supply plant (vented orcaptured together with CO2 from the industrial process).

Table 12 shows the summary of the results obtained in the cur-rent study. It is shown that CO2 capture short-term CO2 capturemay have considerable impact on SO2 and NH3 emissions. When allthree industrial processes in the EU-27 are fully equipped with CO2capture, total industrial SO2 emissions in EU-27 may decrease by40–70%. About half of the decrease in SO2 emissions is attributableto onsite emissions reduction and the rest to avoided emissionsfrom centralized power plants by exporting electricity to the grid.Total industrial NH3 emissions may increase by 120–520% (about2–8% of total European NH3 emissions). NOx emissions rangedbetween a 10% increase and a 10% decrease. The increase is mainlyattributable to the cement sector when coal-fired CHP plants areused to supply CO2 capture energy. PM emissions are also foundto reduce by up to 10% of total European industrial emissions. Thetotal changes in NECD substance emissions due to CO2 capture inthe short term were found to be attributable to the cement and thepetroleum refinery sectors; contribution of the iron and steel sectorwas negligible.

Regarding individual industrial processes, in the cement sectorthe changes in NOx emissions ranges between a 10% decrease and10% increase compared to total EU-27 industrial emissions, largelydepending on the CO2 capture heat supply plant option. SO2 emis-sions were found to decrease by 10–40% compared to total EU-27industrial emissions. The relatively large range of the change inemissions is largely due to the selection of CO2 capture energy sup-ply configuration. For the petroleum refineries, the reduction inSO2 emissions amounts up to about 30% of the total EU-27 indus-trial SO2 emissions. The changes in PM emissions were found tobe marginal compared to total EU-27 industrial emissions for bothsectors. The study also found that in the case of post-combustioncapture for the cement sector and the petroleum refineries, avoidedcentralized power plant emissions, in particular SO2 and PM, due toelectricity export from the CHP plant are as important as changesin onsite emissions.

For the iron and steel sector, the effect of CO2 capture was foundto be limited for the selected CO2 capture technique (TGRBF withVPSA) under the assumptions that the material and energy flow andemissions data used in this paper is representative of the Europeanintegrated iron and steel plants. SO2 emissions may reduce by about5% of total EU-27 industrial emissions, but the changes in otherNECD substance emission levels compared to total EU-27 industrialemissions were marginal. CO2 capture from BF does not largelyaffect the most emitting processes within the integrated steel plantsuch as pelletizing, sintering and coking processes. Note, however,that the changes in the NECD emissions could range from limitedreduction to significant reduction, depending on how the steel millwill adapt and operate their coke oven batteries that supply thecoke to the BF.

It should be emphasized that the results obtained in the cur-rent study are for specific CO2 capture technologies and thereforedo not necessarily apply to other CO2 capture technologies. There-fore, the results presented in this paper should be handled carefully.

3 of Gre

RmopCpaC

A

aclZlr

R

A

A

A

B

C

D

D

D

d

E

E

EE

E

E

F

G

H

28 T. Kuramochi et al. / International Journal

ecommendations for future research are: (1) detailed flowsheetodeling for CO2 capture technologies that integrated into the core

f industrial processes for a more accurate estimation of changes inollutant emission levels, (2) similar assessment for applying otherO2 capture technologies such as oxyfuel combustion capture,ost-combustion capture using advanced solvents and membranes,nd (3) assessment of the economic impact of the changes in non-O2 emissions due to CO2 capture.

cknowledgments

This research is part of the Dutch policy research programme onir and climate (BOLK) and the Dutch research programme on CO2apture technology development (CAPTECH). The authors wouldike to thank Joris Koornneef (Ecofys), Gerard Jägers, Christiaaneilstra, Chris Treadgold (Tata Steel IJmuiden), Toon van Harme-en, Arjan van Horssen, Magdalena Jozwicka (TNO), and anonymouseferees for their contributions to the study.

eferences

lie, C., Backham, L., Croiset, E., Douglas, P.L., 2005. Simulation of CO2 capture usingMEA scrubbing: a flowsheet decomposition method. Energy Conversion andManagement 46, 475–487.

mann, M., Asman, W., Bertok, I., Cofala, J., Heyes, C., Klimont, Z., Rafaj, P., Schöpp, W.,Wagner, F., 2007. Updated baseline projections for the revision of the NationalEmission Ceilings Directive. In: NEC Scenario Analysis Report Nr.4. InternationalInstitute for Applied Systems Analysis, Laxenburg, Austria.

mann, M., Asman, W., Bertok, I., Cofala, J., Heyes, C., Klimont, Z., Rafaj, P., Schöpp, W.,Wagner, F., 2007. Updated baseline projections for the revision of the NationalEmission Ceilings Directive, NEC Scenario Analysis Report Nr. 4. InternationalInstitute for Applied Systems Analysis, Laxenburg, Austria.

irat, J.-P., Maizière-lès-Metz, D., 2010. Steel sectoral report. Contribution to theUNIDO roadmap on CCS—fifth draft, G. In: lobal Technology Roadmap for CCS inIndustry. United Nations Industrial Development Organization, Vienna, Austria.

hapel, D.G., Mariz, C.L., Ernest, J., 1999. Recovery of CO2 from flue gases: commercialtrends. In: Canadian Society of Chemical Engineers Annual Meeting, Saskatoon,Saskatchewan, Canada.

aniëls, B.W., 2002. Transition Paths towards CO2 Emission Reduction in the SteelIndustry. Rijksuniversiteit Groningen, Groningen, the Netherlands.

anloy, G., Berthelemot, A., Grant, M., Borlée, J., Sert, D., van der Stel, J., Jak, H.,Dimastromatteo, V., Hallin, M., Eklund, N., Edberg, N., Sundqvist, L., Sköld, B.-E., Lin, R., Feiterna, A., Korthas, B., Muller, F., Feilmayr, C., Habermann, A., 2009.ULCOS-Pilot testing of the low-CO2 Blast Furnace process at the experimentalBF in Luleå. La Revue de Metallurgie, January.

anloy, G., van der Stel, J., Schmöle, P., 2008. Heat and mass balances in the ULCOSblast furnace. In: 4th ULCOS Seminar.

e Beer, J., Worrell, E., Blok, K., 1998. Future technologies for energy-efficient ironand steel making. Annual Review of Energy and the Environment 23, 123–205.

EA, 2010. European Community Emission Inventory Report 1990–2008 under theUNECE Convention on Long-Range Transboundary Air Pollution (LRTAP). Euro-pean Environmental Agency, Copenhangen, Denmark.

EA, 2011a. Annual European Union Greenhouse Gas Inventory 1990–2008 andInventory Report 2010.

EA, 2011b. CO2 Emissions in Industry—EU Countries (1990, 2007).gberts, P., Keppel, F., Wildenborg, T., Hendriks, C., van der Waart, A.-S., 2003.

GESTCO-DSS: A Decision Support System for Underground Carbon DioxideSequestration. TNO and Ecofys, Utrecht, the Netherlands.

uropean Commission, 2001. Directive 2001/81/EC of the European Parliament andof the Council of 23 October 2001 on National Emission Ceilings for CertainAtmospheric Pollutants.

uropean Commission, 2008. Directive 2008/1/EC of the European Parliament andof the Council of 15 January 2008 concerning integrated pollution preventionand control.

arla, J.C.M., Hendriks, C.A., Blok, K., 1995. Carbon dioxide recovery from industrialprocesses. Climatic Change 29, 439–461.

off, G.S., Rochelle, G.T., 2004. Monoethanolamine degradation: O2 mass trans-fer effects under CO2 capture conditions. Industrial & Engineering Chemistry

Research 43, 6400–6408.

egerland, G., Pande, J.O., Haugen, H.A., Eldrup, N., Tokheim, L.-A., Hatlevik, L.-M.,2006. Capture of CO2 from a cement plant—technical possibilities and eco-nomical estimates. In: 8th International Conference on Greenhouse Gas ControlTechnologies, Trondheim, Norway.

enhouse Gas Control 10 (2012) 310–328

IEA, 2007. Tracking Industrial Energy Efficiency and CO2 Emissions. InternationalEnergy Agency, Paris, France.

IEA, 2008. CO2 Capture and Storage: A key carbon abatement option. InternationalEnergy Agency, Paris, France.

IEA, 2010. Energy Technology Perspectives 2010. Scenarios & Strategies to 2050.International Energy Agency, Paris, France.

IEA GHG, 1999. The reduction of greenhouse gas emission from the oil refining andpetrochemical industry. IEA Greenhouse Gas R&D Programme, Cheltenham, UK.

IEA GHG, 2008. CO2 Capture in the Cement Industry. International Energy AgencyGreenhouse Gas R&D Programme, Cheltenham, UK, pp. 221.

IIASA, 2011. Greenhouse Gas and Air Pollution Interactions and Synergies (GAINS)Europe. International Institute for Applied System Analysis, Laxemburg, Austria.

IPPC, 2006. Draft Reference Document on Best Available Techniques for Large Com-bustion Plants. European Integrated Pollution Prevention and Control Bureau,Sevilla, Spain.

IPPC, 2009. Draft Reference Document on Best Available Techniques in the Cement,Lime and Magnesium Oxide Manufacturing Industries. European Integrated Pol-lution Prevention and Control Bureau, Sevilla, Spain.

IPPC, 2010a. Draft Reference Document on Best Available Techniques for Mineral Oiland Gas Refineries. Draft 1—July 2010. European Integrated Pollution Preventionand Control Bureau, Sevilla, Spain.

IPPC, 2010b. Reference Document on Best Available Techniques in the Cement, Limeand Magnesium Oxide Manufacturing Industries. European Integrated PollutionPrevention and Control Bureau, Sevilla, Spain.

IPPC, 2012. Best Available Techniques (BAT) Reference Document for Iron and SteelProduction. Industrial Emissions Directive 2010/75/EU (Integrated PollutionPrevention and Control). European Integrated Pollution Prevention and ControlBureau, Sevilla, Spain.

Koornneef, J., van Keulen, T., Faaij, A., Turkenburg, W., 2008. Life cycle assess-ment of a pulverized coal power plant with post-combustion capture, transportand storage of CO2. International Journal of Greenhouse Gas Control 2,448–467.

Koornneef, J., Ramirez, A., van Harmelen, T., van Horssen, A., Turkenburg, W., Faaij,A., 2010. The impact of CO2 capture in the power and heat sector on the emis-sion of SO2. NOx , particulate matter, volatile organic compounds and NH3 in theEuropean Union. Atmospheric Environment 44, 1369–1385.

Kothandaraman, A., Nord, L., Bolland, O., Herzog, H.J., McRae, G.J., 2009. Comparisonof solvents for post-combustion capture of CO2 by chemical absorption. EnergyProcedia 1, 1373–1380.

Kuramochi, T., Ramírez, A., Faaij, A., Turkenburg, W., 2012. Comparative assessmentof CO2 capture technologies for carbon-intensive industrial processes. Progressin Energy and Combustion Science 38, 87–112.

Mott MacDonald, 2010. Sectoral Assessment: Cement, Global Technology Roadmapfor CCS in Industry. United Nations Industrial Development Organization,Vienna, Austria.

Rao, M., Rubin, E.S., Berkenpas, M.B., 2004. An integrated modeling framework forcarbon management technologies-Technical documentation: Amine-based CO2

capture and storage systems for fossil fuel power plant, vol. 1. Carnegie MellonUniversity, Pittsburgh, PA, USA.

Schmöle, P., Lüngen, H.B., 2004. Hot metal production in the blast furnace gas froman ecological point of view. In: 2nd International Meeting on Ironmaking and1st International Symposium on Iron Ore, Vitoria, Brazil.

Statoil, 2005. Study and estimate for CO2 capture facilities for the proposed 800 MWcombined cycle power plant - Tjeldbergodden, Norway. Fluor and Statoil, Oslo,Norway.

Torp, T.A., 2005. Drastik reduksjon av drivhusgasser fra stalproduksjon med CO2

Fangst & Lagring (CCS)—ULCOS prosjektet. Technical Committee, Rome, Italy.UNECE, 1999. The 1999 Protocol to Abate Acidification. Eutrophication and Ground-

level Ozone.van Harmelen, T., Koornneef, J., van Horssen, A., Ramírez Ramírez, A., van Gijlswijk,

R., 2008a. The impacts of CO2 capture technologies on transboundary air pollu-tion in the Netherlands. TNO Built Environment and Geosciences, Utrecht, theNetherlands, p. 150.

van Harmelen, T., Koornneef, J., van Horssen, A., Ramirez, A., van Gijlswijk,R., 2008b. The impacts of CO2 capture technologies on transboundary airpollution in the Netherlands. Utrecht Centre for Energy research (UCE),p. 150.

van Straelen, J., Geuzebroek, F., Goodchild, N., Protopapas, G., Mahony, L., 2010. CO2

capture for refineries, a practical approach. International Journal of GreenhouseGas Control 4, 316–320.

Veltman, K., Singh, B., Hertwich, E.G., 2010. Human and environmental impactassessment of postcombustion CO2 capture focusing on emissions from amine-based scrubbing solvents to air. Environmental Science and Technology 44,1496–1502.

Wilkinson, M.B., Boden, J.C., Gilmartin, T., Ward, C., Cross, D.A., Allam, R.J., Ivens, N.W.,2003. CO2 capture from oil refinery process heaters through oxyfuel combus-tion. In: 6th International Conference on Greenhouse Gas Control Technologies,Kyoto, Japan.

World Steel Association, 2010. Crude Steel Statistics Total 2009.