OR 41/2010

CO2 Technology Centre Mongstad updated air dispersion calculations

OR 41/2010

CO2 Technology Centre Mongstad updated air dispersion calculations

Contents
Page Summary ................................................................................................................ 3 Sammendrag p norsk .......................................................................................... 7 1 2 Introduction .................................................................................................. 11 Air quality and water quality guidelines .................................................... 15 2.1 Air quality criteria ................................................................................... 15 2.2 Administrative norms .............................................................................. 16 2.3 Air quality recommendations from Norwegian Institute of Public Health (FHI) ............................................................................................ 17 2.4 Water quality thresholds and guidelines ................................................. 17 2.5 Maximum tolerable amount of nitrosamines .......................................... 18 2.6 Critical loads for nitrogen ....................................................................... 19 Emission data ................................................................................................ 22 3.1 Alstom using chilled ammonia technology ............................................. 22 3.2 Aker Clean Carbon using amines ............................................................ 24 3.2.1 Mobile Test Unit ......................................................................... 29 3.2.2 Emissions of CO2 ........................................................................ 29 Maximum hourly mean concentrations of main pollutants ..................... 30 4.1 Unit emission........................................................................................... 30 4.2 Emissions from Alstom Chilled ammonia technology ........................... 34 4.3 Emissions from Aker Clean Carbon amine technology .......................... 36 4.4 Summary of all calculated maximum hourly mean concentrations ........ 45 Annual mean and 8-hours mean maximum concentrations and annual deposition of main pollutants ......................................................... 47 5.1 Unit emissions ......................................................................................... 47 5.2 Emissions from Alstom facility............................................................... 49 5.3 Results of amines and degradation products using emissions from ACC facility ............................................................................................ 52 5.4 Worst case scenario nitrosamines in air ............................................... 58 5.4.1 Formation of nitrosamines from the MEA emission scenarii ..... 59 5.4.2 Formation of nitrosamines from scenario 5 (improved amines) ........................................................................................ 62 5.4.3 Maximum possible emissions of nitrosamines ............................ 65 5.4.4 Discussion on worst case results ................................................. 66 5.5 Worst case scenario nitrosamines in drinking water ............................ 67 5.5.1 If no degradation of nitrosamines in water .................................. 68 5.5.2 If degradation of nitrosamines in water occur ............................. 68 5.5.3 Discussion on worst case results ................................................. 69 Conclusions ................................................................................................... 71 References ..................................................................................................... 72

6 7

NILU OR 41/2010

Summary
The CO2 Technology Centre Mongstad has given the Norwegian Institute for Air Research (NILU) the task of calculating and assessing the dispersion of emissions to air of carbon dioxide (CO2), nitrogen oxides (NOX), ammonia (NH3), amines and amine degradation products from activities at CO2 Technology Centre Mongstad (TCM). This is an update of the report Test Centre Mongstad, Dispersion calculations for emissions to air from Test Centre Mongstad (TCM) (Berglen et al., 2008).The calculated maximum hourly mean concentrations and yearly averages are well below norms and guidelines for NH3, CO2, MEA, formaldehyde, acetaldehyde, and piperazine. Worst case calculations for nitrosamines in air show that there may be a problem, but further research must be done to reduce the uncertainties. Based on the worst case calculations for water it is not possible to conclude that there will be a problem, but the uncertainties have to be reduced, especially concerning nitrosamine degradation.

Since the previous report two technology vendors have been contracted to test their capture technology at TCM; Alstom using Chilled ammonia and Aker Clean Carbon (ACC) using amines. Both vendors have provided new emission data and NILU has made dispersion calculations based on these emission data. The results have been compared with existing threshold values and air quality guidelines. Maximum hourly mean concentrations have been estimated using CONCX, these are valid inside the Mongstad industrial area and up to approximately 5 km from the emission points. For regional dispersion The Air Pollution Model (TAPM) have been applied. These results are important concerning long term exposure and deposition. TCM will use flue gas from existing facilities at Mongstad, like the Residual Fluid Catalytic Cracker (RFCC) and the Combined Heat and Power plant (CHP). In this study, the additional emissions from the CO2 capture plant is investigated. There will be emissions of NH3 from the Alstom plant, and emissions of amines and amine degradation products from the ACC plant. ACC will test the use of monoethanolamine (MEA), as well as other proprietary blends for carbon capture. Both plants will emit the captured CO2 to air separately. For the Alstom plant, emitting NH3, the impacts of these emissions on air quality and eutrophication of the ecosystems were investigated. For the ACC plant the most important issues are the emissions to air of amines and amine degradation products formed in the capture process and their subsequent fate in the atmosphere. Several amine degradation products have severe health effects. In addition, the chemical degradation of amines is not well known. The content in this report is based on existing knowledge at this stage and will be upgraded when new knowledge is available. Concerning the ACC amine plant, the emission data contain some welldocumented substances, but also substances where large knowledge gaps exist. The emissions of well-known substances are of the same order of magnitude as the emissions of substances where the atmospheric degradation is poorly known.

NILU OR 41/2010

4 Model calculations have been performed on a very local scale (~up to 8 km using CONCX) to investigate dispersion, and on a regional scale (~up to 30 km using TAPM) to investigate dispersion and deposition. CONCX and TAPM are wellknown models used in the scientific community. Weather data for 2007 are used for regional calculations. Concerning chemistry, atmospheric degradation of amines and amine chemistry is very complex with several hundred compounds involved and reactions taking place both in gas phase, aqueous phase and on particles. At present, no full atmospheric model chemistry scheme exist for amines. To account for chemical reactions, a scaling method is applied using chemical degradation rates found in the literature. In addition there are large uncertainties. Some processes, like gas to aerosol (particle) conversion or chemistry on aerosols have not been included. The simplifications made in the scaling method imply that the calculated model concentrations represent an upper limit of the concentrations to be expected in the real atmosphere. The calculated maximum hourly mean concentrations from the Alstom facility are 55 g/m3 for NH3 and 72 mg/m3 for CO2. From the ACC plant the maximum calculated hourly concentration of MEA is about 2 g/m3 and the maximum calculated hourly mean NH3 concentration is just below 10 g/m3. The maximum calculated hourly concentrations of formaldehyde, acetaldehyde, and alkylamines are 1,4 g/m3, 2,1 g/m3, and 0.8 g/m3 respectively. Maximum CO2 concentration from the export gas stack is 71 mg/m3 for ACC. All these calculated hourly maximum concentrations for Alstom and ACC are much lower than the Administrative norm and lower than long term exposure guidelines for CO2, MEA, formaldehyde and acetaldehyde. Concerning regional dispersion, the model calculations estimate a maximum 8hours mean concentration of NH3 of 6.1 g/m3, and a maximum deposition of 3.1 mg N/(m2 year), using Alstom emission data. This deposition is a small increase of 0.3 % compared to the current background deposition, but in an area where the vegetation critical loads are already exceeded. Concerning the ACC plant, the calculated maximum annual mean concentration of MEA is 1,3 ng/m3, while the maximum annual mean concentrations of formaldehyde and acetaldehyde are 1,6 ng/m3 and 2,4 ng/m3, respectively. The maximum annual mean concentration of alkyl amines (methylamine, ethylamine, dimethylamine, diethylamine) is just below 1,0 ng/m3. The maximum concentrations for both plants and all emission scenarios are found about 10 km to the S-E of Mongstad in the area around Sundsb and Kols. A Worst case estimate has been calculated to provide a worst possible scenario estimate of concentrations of nitrosamines in air and water. This method is used when there are large uncertainties to estimate the highest possible (but not necessarily plausible) concentrations of compounds (nitrosamines in this case). The calculated nitrosamines concentrations in ambient air are compared to the 0.07 ng/m3 threshold value from EPA/IRIS. A formation range of nitrosamines between 2 and 10 % is assumed for components with nitrosamines formation potential (e.g. Pitts et al., 1978, Karl et. al 2008, Brten et. al 2008). This wide range reflects the uncertainties in these calculations. Yearly average maximum value of nitrosamines have been calculated for all the scenarii apart from the upset scenario. Considering these results it must be remembered that the emission

NILU OR 41/2010

5 conditions must be valid for the duration of the year for the concentration value to be valid. For the MEA scenarii, only one scenario exceed the limit value. For the improved solvents scenarii, all exceed the limit except scenario 5c (if 10% formation of nitrosamines occur). The main conclusion from the worst case study of nitrosamines in air is that there may be a potential problem concerning the EPA/IRIS threshold value of 0.07 ng/m3. Further research, including more detailed modeling, and monitoring is needed to reduce the present uncertainties associated with nitrosamines in air from CCS. Calculations of the maximum amount of components with nitrosamine formation potential in the emissions show that the concentration of nitrosamines in air can be below the threshold value, if the sum of component emission is no larger than about 0.02 g/s. This figure is dependent on the amount of directly emitted nitrosamines. Concerning nitrosamines in drinking water, the degradation of nitrosamines in water is a key question. Some studies indicate that nitrosamines are persistent. If this is the case, then all worst case scenarii exceed the 0.7 ng/l threshold value. Other studies, like e.g. Drewes et al. (2006) report a degradation under anoxic conditions with a lifetime of the order of days. If these results are applied, the worst case scenarii equilibrium concentrations are below the threshold value. Based on these calculations, it is not possible to conclude that there will be a problem with nitrosamines in drinking water, but the uncertainties have to be reduced. Scientific research should be conducted, especially to investigate the degradation of nitrosamines in water.

NILU OR 41/2010

NILU OR 41/2010

Sammendrag p norsk
CO2 Technology Centre Mongstad har gitt Norsk institutt for luftforskning (NILU) i oppdrag beregne og vurdere spredning og utslipp til luft av karbondioksid (CO2), nitrogenoksider (NOX), ammonikk (NH3), aminer og nedbrytningsprodukter fra aminer som skal slippes ut fra testsenteret som er under bygging ved Mongstad (CO2 Technology Centre Mongstad, TCM). Dette er en oppdatering av en tidligere NILU-rapport Test Centre Mongstad, Spredningsberegninger for utslipp til luft fra Test Centre Mongstad (TCM) (Berglen et al., 2008). De beregnede maksimalkonsentrasjonene (time- og rlig middel) er alle under akseptkriteriene for NH3, CO2, MEA, formaldehyd, acetaldehyd og piperazine. Worst case-beregninger har blitt gjort for nitrosaminer i luft. Disse viser at nitrosaminer kan vre et problem, men mer kunnskap behves for redusere usikkerhetene. Basert p Worst case beregninger for drikkevann er det ikke mulig konkludere om nitrosaminer vil vre et problem, men usikkerhetene m reduseres, spesielt med tanke p nedbrytning av nitrosaminer i vann.

Siden forrige rapport har to teknologileverandrer ftt i oppdrag teste sin fangsteknologi ved TCM; Alstom som benytter Chilled ammonia og Aker Clean Carbon (ACC) som benytter aminer. Begge selskaper har gitt nye oppdaterte utslippstall til NILU og NILU har utfrt spredningsberegninger basert p disse tallene. Resultatene har blitt sammenlignet med gjeldende grenseverdier og luftkvalitetskriterier. Maksimal timemiddelkonsentrasjon er beregnet ved hjelp av CONCX-modellen. Disse resultatene er gyldige innenfor industriomrdet p Mongstad og ut til omlag 5 km. For beregninger p regional skala er The Air Pollution Model (TAPM) brukt. Disse resultatene er relevante for langtidskonsentrasjoner (rlig middel) og avsetning. TCM vil benytte rykgass fra eksisterende anlegg p Mongstad; fra Residual Fluid Catalytic Cracker (RFCC, dvs. fra selve raffineriet) og fra Combined Heat and Power plant (CHP, dvs. fra gasskraftverket). I denne studien beregnes og vurderes de ekstra utslippene fra CO2-fangstanlegget. Det vil vre utslipp av NH3 fra Alstoms anlegg, samt utslipp av aminer og aminers degraderingsprodukter fra ACCs anlegg. ACC planlegger teste bruk av monoethanolamin (MEA) i tillegg til to proprietre blandinger. Begge anleggene vil slippe ut CO2 i en egen pipe. Angende Alstoms teknologi er utslipp av NH3 og effekten av disse utslippene vurdert med tanke p luftkvalitet og overgjdsling (eutrofiering). Angende ACCs teknologi er utslipp av aminer og aminers degraderingsprodukter og spredning og omdanning av disse i atmosfren det viktigste sprsmlet. Flere av degraderingsproduktene fra aminer har alvorlige helseeffekter. I tillegg er det store kunnskapshull med tanke p den kjemiske nedbrytningen i atmosfren. Denne rapporten bygger p eksisterende kunnskap pr dags dato og nye studier vil bli gjort nr ny kunnskap foreligger. ACCs anlegg vil slippe ut komponenter som er velkjente, men ogs komponenter hvor kunnskapsnivet er lavt. Utslippene av velkjente komponenter er av samme strrelsesorden som utslipp av komponenter hvor nedbrytningsveiene er lite kjent.

NILU OR 41/2010

Det er gjort modellberegninger p lokal skala for beregne spredning (opp til 8 km ved hjelp av CONCX) og p regional skala for studere spredning og avsetning (opp til 30 km ved hjelp av TAPM). Bde CONCX og TAPM er velkjente modeller brukt av forskere over hele verden. For regionale modellkjringer er vrdata for 2007 benyttet. Kjemisk nedbrytning av aminer og aminkjemien er meget komplisert med hundrevis av forskjellige komponenter og reaksjoner. Aminer brytes ned bde ved gassfasekjemi, vskefasekjemi (i drper) og p partikler. Pr dags dato finnes det intet modellkjemiskjema for aminer For ta hyde for kjemiske reaksjoner er det brukt en skaleringsmetode. Denne metoden er basert p kjemiske reaksjonsrater og laboratorieeksperimenter funnet i den vitenskapelige litteraturen. Det er dog store usikkerheter. Visse typer prosesser, som gass-til-partikkelovergang, er ikke tatt med. Forenklingene som er gjort i skaleringsmetoden innebrer samtidig at de beregnede konsentrasjonene er i det vre skikt av hva man kan forvente finne i den virkelige atmosfren. Maksimalt beregnet timemiddel for Alstoms anlegg er 55 g/m3 for NH3 og 72 mg/m3 for CO2. For ACCs anlegg er maksimalt beregnet timemiddel av MEA omlag 2 g/m3 og maksimalt timemiddel for NH3 er litt under 10 g/m3. Maksimalt beregnet timemiddel for formaldehyd, acetaldehyd, and alkylaminer er hhv. 1,4 g/m3, 2,1 g/m3 og 0,8 g/m3. Maksimal CO2-konsentrasjon fra eksportgasspipen er 71 mg/m3 for ACCs anlegg. Alle de beregnede timemiddelverdier (Alstom og ACC) er lavere enn Administrativ norm og lavere enn gjeldende luftkvalitetskriterier for CO2, MEA, formaldehyd og acetaldehyd. Nr det gjelder regional skala viser beregningene at Alstoms utslipp gir et 8timers maksimum lik 6,1 g/m3for NH3 og maksimal avsetning p of 3,1 mg N/(m2 r). Maksimal avsetning utgjr 0.3 % av allerede eksisterende Navsetning. Samtidig er dette i et omrde hvor grenseverdiene for overgjdsling allerede er overskredet. For ACCs anlegg er maksimal rlig konsentrasjon av MEA 1,3 ng/m3, mens maksimumskonsentrasjonene av formaldehyd og acetaldehyd er hhv. 1,6 ng/m3 og 2,4 ng/m3. Maksimalt rsmiddel av alkylaminer (metylamin, etylamin, dimetylamin, dietylamin) er like under 1,0 ng/m3. Beregnet maksimumskonsentrasjon forekommer om lag 10 km sr-st for Mongstad i omrdene rundt Sundsb and Kols. En worst case-beregning er gjort for finne det verst tenkelige scenario nr det gjelder konsentrasjoner av nitrosaminer i luft og drikkevann. Dette er en metode som er brukt nr det er store usikkerheter i prosessene og mlet er finne hyest mulige (men ikke ndvendigvis sannsynlige) konsentrasjon av de underskte komponentene (nitrosaminer i dette tilfellet). De beregnede konsentrasjonene er sammenlignet med grenseverdien fra EPA/IRIS (0,07 ng/m3). For de komponentene som har potensiale for nitrosamindannelse er det brukt et spenn fra 2 til 10% for nedbrytning til nitrosaminer (basert p for eksempel Pitts et al., 1978, Karl et. al, 2008, Brten et. al 2008). Dette vide spennet reflekterer de store usikkerhetene i disse beregningene. Maksimalt rlig middelkonsentrasjon av nitrosaminer er beregnet for alle scenarioer bortsett fra Upset. Disse resultatene m sees i lys av at utslippene m vre konstante over et r for at de beregnede rsmiddelkonsentrasjonene skal vre direkte sammenlignbare. For MEA-

NILU OR 41/2010

9 scenarioene er det bare et som overskrider grenseverdien. For scenarioene med Improved amines gr alle over grenseverdien bortsett fra scenario 5c (forutsatt 10% nitrosamindannelse). Hovedkonklusjonen fra Worst case-beregningen er at det kan vre et problem med tanke p nitrosamindannelse og grenseverdien p 0,07 ng/m3. Ytterligere kunnskap (inkludert mer detaljert modellering) og overvkning er ndvendig for redusere usikkerhetene i forbindelse med nitrosaminer i luft fra CO2-fangst. Beregninger av strst mulige utslipp av komponenter med potensiale for danne nitrosaminer viser utslippene kan hyst vre om lag 0,02 g/s for overholde grenseverdien. Men det er viktig at dette tallet sees i sammenheng med de direkte utslippene av nitrosaminer. Nr det gjelder nitrosaminer i drikkevann er nedbrytning av nitrosaminer i vannfase et meget viktig punkt. Noen studier indikerer at nitrosaminer er bestandige (vanskelig nedbrytbare). Hvis dette er tilfelle vil alle worst case scenarioer gi konsentrasjoner over grenseverdien p 0,7 ng/l. Noen studier, som eksempelvis Drewes et al. (2006) har funnet nedbrytning av nitrosaminer under oksygenfattige forhold med levetider for nitrosaminer i strrelsesorden ~dager. Om disse levetidene anvendes p forholdene rundt TCM s vil de verste av worst case scenarioene vre under grenseverdien. Konklusjonen fra disse beregningene er at det ikke er mulig slutte at det vil vre et problem med nitrosaminer i drikkevann, men at usikkerhetene m reduseres. Mer forskning behves, spesielt for kvantifisere nedbrytningen av nitrosaminer i vann.

NILU OR 41/2010

10

NILU OR 41/2010

11

CO2 Technology Centre Mongstad updated air dispersion calculations

Introduction

Carbon capture and storage (CCS) is an important topic and measure on the political agenda. The Norwegian government puts a great effort into research and development concerning CCS. To facilitate the development of new technologies the CO2 Technology Center Mongstad (TCM) has been built. The establishment of TCM was part of the agreement between the Norwegian government and Statoil signed in Octobre 2006 (Gjennomfringsavtale) in connection to the emission permit for the Combined Heat and Power plant (CHP) located at Mongstad. CO2 Technology Centre Mongstad will test various post-combustion techniques to capture CO2. In the short term (2011) a test plant will be constructed, capturing ~100000 tonnes CO2 per year. Later a full scale plant is planned at Mongstad and, when in operation, it will capture and store between 1.2 and 2.0 million tonnes of CO2 per year (CO2 Master Plan Mongstad, 2009). There will inevitably be emissions to air from these facilities. The dispersion, degradation and deposition of these emissions have to be quantified, and the concentrations in air of the compounds emitted from TCM have to be calculated. In 2008 Norwegian Institute for Air Research (NILU) wrote a report for the CO2 Technology Centre Mongstad, see Berglen et al. (2008). In this report NILU made dispersion calculations based on preliminary emission data. Maximum hourly mean concentrations of CO2, NH3 and NOx were presented as well as annual mean concentrations and simple deposition calculations of soluble species (e.g. MEA). NILU also recommended a minimum stack height to avoid downdraft of the emissions. Since then the consortium behind TCM (Gassnova, Statoil, Shell, and Sasol) has chosen two different technology vendors for the Technology Centre, Aker Clean Carbon (ACC) using amines and Alstom using chilled ammonia to capture CO2. These are two possible post-combustion techniques to be tested in Norway and TCM is a test facility for the two chosen technologies. The chilled ammonia technology will have emissions of well known substances and the level of knowledge is sufficient to give a good estimation of impacts from emissions to air. The emissions from capture of CO2 by amines are not so well known and the emissions vary according to the solvents used for capture. The amines also react with other substances in the atmosphere after emission. Theoretical studies have shown that toxic and carcinogenic substances can form both inside the plant and in the atmosphere. It is therefore necessary to know the toxicity and amount of these substances that are formed. Currently there are large knowledge gaps. The

NILU OR 41/2010

12 level of knowledge depends strongly on the amine or amine solvent used for capture. This lack of knowledge causes some concern. The photochemical reactions after emission are complex and partly unknown. It is common knowledge that the emissions from the amine technology consist not only of the mother amine used in the carbon capture process, but also of degradation products. This adds uncertainty to the emissions from the amine technology. To abate this uncertainty a system for measurements of emissions should in our view be implemented for the test plant. This surveillance system will need to be optimized to look at substances yet not known to be emitted. Methodology and standardization of measurement techniques need to be developed. The measurement systems must have a low detection. This monitoring system will serve two purposes; 1) to develop new analysis methods to measure new components, and 2) to monitor the substances emitted from TCM so that human and environmental exposure levels can be assessed. These emissions measurements should be made available to the scientific community and to the public. Level of scientific knowledge Some work have been initiated to investigate the processes connected to emissions to air of amines. The work have concentrated on the emissions of MEA (e.g. the Atmospheric Degradation of Amines (ADA) projects in 2009 and 2010). Some of the work is referred to below. In 2008 and 2009 a screening project lead by NILU summarized the current level of knowledge concerning amine emissions and transformation, as well as the impact of amines on the environment and human health. This was a broad project with contributions from the Norwegian Institute for Water Research (NIVA), the Norwegian Institute for Nature Research (NINA), the Norwegian Institute for Public Health (FHI), the Centre for Theoretical and Computational Chemistry at the University of Oslo (CTCC, UiO), in addition to NILU. The project was funded by Gassnova, Statoil and Shell Technologies Norway. The screening project summarized the current level of understanding and a number of key issues were identified. Ten reports were written in the project, they are listed in Table 1. All of the reports can be downloaded from the co2.nilu.no project web site . For this study, the report on degradation products of amines (Brten et al., 2008), the report on analytical methods (Dye et al., 2008), the worst case report (Karl et al., 2008), the reports on health effects of amines and the degradation products (Lg et al, 2009a, 2009b), and the report on toxicity on aquatic organisms (Brooks, 2008) have been most useful.

NILU OR 41/2010

13

Table 1:

Reports published during the Phase 1 CO2 and amines screening project coordinated by NILU.
Report Title Final report on a theoretical study on the atmospheric degradation of selected amines. Evaluation of analytical methods for amine related emissions and degradation products in emission and ambient air Report on models, model needs and requirements Amines and rainfall. Impact of amines on rainfall from plume clouds Worst Case Studies on Amine Emissions from CO2 Capture Plants Health effects of different amines relevant for CO2 capture Health effects of possible degradation products of different amines relevant for CO2 capture Effects on terrestrial vegetation, soil and fauna of amines and possible degradation products relevant for CO2 capture The toxicity of selected primary amines and secondary products to aquatic organisms: A review Summary Report: Amine Emissions to Air During Carbon Capture Phase I: CO2 and Amines Screening Study for Effects to the Environment Author CTCC NILU NILU NILU NILU FHI FHI NINA NIVA NILU

Task # Task3 Task4 Task 5.1/5.2 Task 5.3 Task 6 Task 7.1 Task 7.2 Task 8 Task 9 Task 10

Based on the theoretical study done in the Phase 1 project, UiO-CTCC and partners initiated the ADA projects (Atmospheric Degradation of Amines, see Nielsen et al., 2010). These are chamber experiments performed in the EUPHORE chamber, where gas phase oxidation of MEA is investigated. Current study This report is an update of our previous work (Berglen et al., 2008). New model calculations have been made based on more reliable emissions estimates and new knowledge has been taken into account. The aim of this study is to calculate dispersion and deposition of air pollutants emitted from TCM and assess the concentrations in relation to air quality threshold values and the deposition of nitrogen in relation to the vegetation critical loads. The main components that are studied are CO2, NOx, NH3, and MEA and some reaction- and degradation products, including, but not limited to acetaldehyde, formaldehyde, nitrosamines and alkyl amines. Maximum hourly mean concentrations of compounds emitted are calculated on a very local scale (a few kilometres from the stack), and on a semi-regional scale (up to 30 km from Mongstad). This is done to assess the concentrations inside the

NILU OR 41/2010

14 Mongstad plant (exposure on the people working at Mongstad), and to assess the concentrations for people living in the communities around the Mongstad refinery. Section 2 presents the current air quality and water quality criteria, as well as the vegetation critical loads for nitrogen deposition. The results of the calculations in this study were compared to these quality criteria and critical loads. Section 3 presents the emission data from the Alstom and ACC plants for different emission scenarios. For the concentrations of pollutants inside the Mongstad industrial area the CONCX model has been used to calculate maximum hourly mean and the results are presented in section 4. For the calculation of the dispersion at a regional scale, the model TAPM have been applied and model results showing annual mean concentrations, maximum short-term concentrations and annual nitrogen wet deposition are presented in sections 5.1-5.3. Finally, the worst case calculations/estimates of nitrosamine formation and nitrosamine concentrations are presented, both concerning air (section 5.4) and drinking water (section 5.5).

NILU OR 41/2010

15

2
2.1

Air quality and water quality guidelines

When evaluating the air quality in an area, it is normal to compare measured concentrations and calculated concentrations with air quality criteria or air quality limit values. In 1997, the Norwegian government stipulated registration and action thresholds in the Pollution Control Regulation. If a registration threshold is exceeded, potential measures must be evaluated to bring the air pollution level below the limit value. If an action threshold is violated, this must be followed up by measures to reduce air pollution. The EU has recently defined new limit values for air quality in the European Union (including the EEA area). These are primarily based on the World Health Organisation's recommended guidelines (WHO, 1999). The EUs limit values for means over 1 hour, 8 hours and 24 hours may be exceeded a certain number of times per year. These values are also valid in Norway through the EEA agreement. In some ways, these limits are significantly stricter than the applicable provisions of the Norwegian Pollution Control Act. In autumn 1998, the Norwegian government passed national air quality targets for cities and communities that must have been complied with since 1 January 2005 (PM10, SO2) and 1 January 2010 (PM10, NO2, benzene). These requirements are based on the EU requirements, but the Norwegian target values are somewhat stricter. All official data and reporting relating to the progress of environmental efforts, developments in the condition of the environment etc. must be undertaken in accordance with these targets. The Norwegian Climate and Pollution Agency (Klif) has stipulated so-called recommended air quality standards based on the requirement that exposure levels must be two to five times higher than the criteria before it can be safely concluded that harmful effects exist. Violating these standards can therefore not be interpreted as necessarily harmful to human health, but it cannot be ruled out that particularly vulnerable people may feel the effects at levels below the criteria. These criteria are significantly lower than the registration and action thresholds in the Pollution Control Regulation, and they are also lower than the EUs limit values and the Norwegian national targets. As opposed to the requirements set out in the Pollution Control Regulation and the EUs limit values, the Klifs recommended air quality standards are not legally binding. The Klifs recommended air quality standards contain the lowest values; when the air quality satisfies these values, other requirements are also satisfied. Table 2 provides a summary of the different limit values and standards.

NILU OR 41/2010

16

Table 2:

Klifs recommended air quality standards, national targets, the Pollution Control Acts action thresholds and registration thresholds, and the EUs new limit values for air quality relating to effects on human health. The limit values are in g/m3. The figures brackets show how many times it is permitted to exceed the limit value.
Mean time Klifs recommended air quality standards National targets (and the number of violations permitted) The Pollution Control Acts action threshold The Pollution Control Acts registration threshold The EUs new limit values (and number of times they may be exceeded) 1 hour 100 150 (8 per year) 300
1) 1)

Substance NO2

8 hours

24 hours 75

6 months 50

1 year 30

200 200 (18 per year)

40

1)

1)

must be complied with by 1 January 2010

2.2

Administrative norms

The Norwegian Labour Inspection Authority has defined maximum concentrations that must not be exceeded in the working environment (see Veiledning om administrative normer for forurensning i arbeidsatmosfre in http://www.arbeidstilsynet.no/binfil/download2.php?tid=77907, Guidelines relating to administrative norms for pollution in the workplace - in Norwegian only, see Norwegian Labour Inspection Authority in the reference list). It is these limit values that apply within the industrial area at TCM and they have been used to evaluate the results (see Table 3). Table 3: Administrative norms for 8-hours mean concentrations for the different components discussed in this report.
Component Administrative norm 3 Ammonia (NH3) 18 mg/m 3 CO2 9000 mg/m 3 Nitrogen dioxide (NO2) 1.1 mg/m 3 MEA (CAS 141-43-5) 2.5 mg/m 3 Piperazine (CAS 110-85-0) 0.3 mg/m 3 Formaldehyde* 0.6 / 1.2 mg/m 3 Acetaldehyde 45 mg/m 3 Methylamine 12 mg/m 3 Ethylamine 4 mg/m 3 Dimethylamine 4 mg/m 3 Diethylamine 15 mg/m 3 * 1.2 mg/m is maximum concentration that cannot be exceeded under any circumstances.

The figures show that the limit for CO2 is very high, it is permitted to have almost

NILU OR 41/2010

17 0.5% CO2 in the air without breaching the norm. In evaluating ambient air concentrations other than in the work atmosphere, it is common to use a factor of 1/30 to 1/100 to estimate ambient air concentrations guidelines, compared to the administrative norms for working atmosphere . This factor will vary according to how much is known about the substance and how toxic it is. 2.3 Air quality recommendations from Norwegian Institute of Public Health (FHI)

During the phase 1 of the screening project coordinated by NILU in 2008-2009 (see Table 1), the Norwegian Institute of Public Health (FHI) did a study to investigate health effects of different amines relevant for CO2 capture and their degradation products (Lg et al, 2009a, 2009b). The aim was also to establish exposure guidelines for the general population. The four most relevant amines were investigated (MEA, Methyl diethanolamine (MDEA), AMP and piperazine). Concerning degradation products special attention was given to nitrosamines, nitramines, aldehydes and amides. Human health effects with regard to acute toxicity (short term exposure), chronic toxicity (long term exposure), mutagenicity (affect DNA), carcinogenicity (cause cancer), teratogenicity (ability to cause birth defects), irritation (in e.g. eyes or respiratorial airways), and sensitisation (the act or process of inducing an acquired sensitivity or allergy) were investigated. The guideline values established by FHI are listed in Table 4. Concerning acetaldehyde the value from FHI is higher than the WHO value. WHO regards 0.3 mg/m3 as a limit for tolerable lifetime cancer risk. Table 4: Exposure guidelines for the general population over time, concentrations in air. Numbers from Lg et al. (2009a and 2009b)
Exposure guidelines 3 10 g/m 3 120 g/m 3 6 g/m 3 5 g/m Reduce human exposure to an absolute minimum. No values concluded 3 1.2 g/m 3 2 mg/m

Component MEA (CAS 141-43-5) MDEA (CAS 105-59-9) AMP (CAS 124-68-5) Piperazine (CAS 110-85-0) Nitrosamines Nitramines Formaldehyde (CAS 50-00-0) Acetaldehyde (CAS 75-07-0)

Table 4 only lists the well known compounds. There will also be emissions from TCM of compounds that are not well known and for which threshold values do not exist. The possible health effects and toxicity of these unknown compounds are not established. These compounds and their possible environmental effects have to be investigated further. E.g. nitrosamines is a group of compounds where some are very toxic and carcinogenic, while others cause less concern. 2.4 Water quality thresholds and guidelines

During the screening project Norwegian Institute for Water Research (NIVA) conducted a literature survey to find ecotoxicity data of amines and their

NILU OR 41/2010

18 degradation products (Brooks and Wright, 2008). A summary of the data found is listed in Table 5. Please note that these values refer to effects on aquatic organisms due to concentrations in water, and that they cannot be compared directly to concentrations in air. Table 5: Critical values for the determination of risk for each of the main chemical groups based on the most sensitive environmental effect measured. Taken from Brooks and Wright (2008).
LOEC (mg/L) Assessment Factor 100 1000 100 100 50 1000 1000 Calculated PNEC (g/L) 7.5 20 5 100 24 0.025 0.2 Required PEC for an RQ >1 (g/L) >7.5 >20 >5 >100 >24 >0.025 >0.2

Compound

MEA AMP MDEA PIPA Amides Nitrosamines Nitramines

0.75 20 0.5 10 1.2 0.025 0.2

LOEC: Lowest observable effect concentration PNEC: Predicted no effect concentration PEC: Predicted environmental concentration RQ:

2.5

Maximum tolerable amount of nitrosamines

One important task in this project is to calculate maximum tolerable amount of nitrosamines from TCM (worst case). The calculations are based on normal operational flow rates for the ACC plant and the concentrations for acceptable risk levels in drinking water and air. In the calculation of the maximum tolerable emission calculation from TCM, the value corresponding to a cancer risk of one in a million (10-6) is used. The nitrosamine concentrations for acceptable risk levels used in the calculation of the maximum tolerable emission are: 0.07 ng/m3 in air 0.7 ng/l in drinking water Originally the nitrosamine concentration in air for the acceptable risk level was 4 ng/m3, and it has been used in the worst case study by Karl et al. (2008). The state of California has determined that 0.04 and 0.004 g/m3 of NDMA (a nitrosamine) per day were equivalent to a 10-5 and 10-6 risk of cancer, respectively 1. But recently the Environmental Protection Agency (EPA) in the US reported that threshold value for nitrosamines for 10-6 risk of cancer is 0.07 ng/m3 in ambient air (see Integrated Risk Information System [IRIS],
1

0,004 g/m3 equals 4 ng/m3

NILU OR 41/2010

19
http://www.epa.gov/ncea/iris/subst/0045.htm ,

section II.C.1. Summary of Risk Estimates). This threshold value of 0.07 ng/m3 is here used for the worst case scenario calculation presented in section 5.4. It should be noted though that the 0.07 ng/m3 value refer to a lifetime cancer risk (60 years) whereas TCM will be in operation for a shorter period of time (5 years). The acceptable concentration of nitrosamines in water used in the worst case study by Karl et al. (2008) is 7 ng/l, which corresponds to a cancer risk of 1 in 100 000 (10-5). Reference is given to the Reach guidance on information requirements and chemical safety assessments (chapter R.8, p. 47. see Reach 2008), which states that: "Although there is no EU legislation setting the 'tolerable' risk level for carcinogens in the society, cancer risk levels have been set and used in different contexts. Based on these experiences, cancer risk levels of 10-5 and 10-6 could be seen as indicative tolerable risks levels when setting DMELs (Derived Minimal Effect Level) for workers and the general population, respectively. The amines worst case studies by Karl et al. (2008) also refers to a 10-6 cancer risk level of 1 ng/l for N-Nitrosodiethylamine (NDEA). The EPA/IRIS limit of 0.7 ng/l is chosen as the concentration to be used for nitrosamines as group for the maximum tolerable emission calculations. EPA/IRIS refer to NDMA in this case. 2.6 Critical loads for nitrogen

Nitrogen acts as a fertiliser for vegetation, but adding too much nitrogen can have undesirable effects. Vegetation and soil naturally receive nitrogen in the form of wet and dry deposition. In addition, plants can absorb and bind nitrogen from the atmosphere through so-called biological nitrogen fixation. In Norway, nitrogen compounds will largely be absorbed by the vegetation since many of the vegetation types have sub-optimal access to nitrogen (cf. Stuanes and Abrahamsen, 1996). The capacity to utilise the increased access to nitrogen is different in different types of vegetation. An increase in nitrogen can therefore lead to plant species with higher tolerance limits suppressing species with lower tolerance limits and thus affect bio-diversity (Tamm, 1991). Figure 1 shows total deposition of nitrogen (N) in Norway 1997-2001 and 20022006 (Aas et al., 2008). Around Mongstad the N deposition is of the order of 8001000 mg N/year. Figure 2 shows the vegetation critical loads for eutrophication. The critical load is defined as A quantitative estimate of an exposure to one or more pollutants below which significant harmful effects on specified sensitive elements of the environment do not occur according to present knowledge (Nilsson and Grennfelt, 1988). Figure 3 shows the estimated exceedence of the vegetation critical loads for eutrophication (Larsen et al., 2008). The figures shows that in the southern parts of Norway the nitrogen deposition is already near or above the critical loads. This includes the areas around Mongstad. Any additional nitrogen in areas where the critical load has been exceeded will have a negative impact on vegetation and on the environment.

NILU OR 41/2010

20

Figure 1: Total deposition of nitrogen (oxi+red) for 1997-2001 (left panel) and 2002-2006 (right panel) taken from Aas et al. (2008). Unit: mg N/m2 year.

NILU OR 41/2010

21

Figure 2: Map showing the vegetation critical loads for eutrophication based on the lowermost critical load for each vegetation type. Unit: mg N/m2 year.

NILU OR 41/2010

22

Figure 3: Estimated exceedence of the critical loads for eutrophication of vegetation (empirical values), taken from Larsen et al. (2008).

Emission data

Two different companies have been contracted to test their CO2 capture technology at TCM; Aker Clean Carbon (ACC) and Alstom. ACC will test their use of amines while Alstom will test their chilled ammonia technique. There will be in all four different emissions points at TCM, two for each technology vendor. One of these stacks will emit the clean flue gas, and the other will emit CO2 that is captured in the process. In the full scale plant the captured CO2 will be sent in pipelines to be stored below the seabed in the North Sea. The flue gas used at TCM comes from two different sources, the Residual Fluid Catalytic Cracker (RFCC) at the Mongstad refinery and the Combined Heat-andPower plant (CHP), i.e. the power plant which is constructed close to the refinery at Mongstad. One important criteria for the location of TCM at Mongstad is the agreement between Statoil and Norwegian Government concerning the emission permit for the CHP plant. In addition to the components listed here there will also be emissions of e.g. NO x, SOx and particles. These compounds are already part of the flue gas, while NH3, amines and amine degradation products come from the capture process. The compounds from the capture process are the most important topic in this study. However, emissions of NOx (NO2) will be investigated since a reduction in the NOx level has occured due to installation of SNCR (Selective Non-Catalytic Reduction) in the RFCC flue gas system. 3.1 Alstom using chilled ammonia technology

The emission data provided by Alstom is listed in Table 6 and Table 7. In addition to ammonia (NH3) there are also emissions of CO2 and water vapour (H2O). N2, O2 and Argon are also listed, this is just normal air and these components are

NILU OR 41/2010

23 not part of the emission data. Base case is the normal emission rate expected at TCM whereas worst case is the maximum emissions that can occur. For some purposes the base case emission data is used, giving the average load or concentration, while for other purposes the worst case emission data is used giving the maximum values. Table 6: Emission data for the Alstom plant using flue gas from the RFCC. To the left the original emission data provided by Alstom, to the right the emission rates per second.
Scenario 1 Base case Stack height (m) Stack diameter (inch) Height on adjacent buildings Gas temperature C Gas volume (m3/h) Velocity out of stack (m/s) 48 40 * 17.7 44287 15.7 Scenario 2 Worstcase 48 40 * 17.7 44293 15.7 Amount of gas emitted (kg/s) NH3 (g/s) CO2 (g/s) H2O (g/s) N2 (g/s) O2 (g/s) Ar (g/s) CO2 (g/s) NH3 (g/s) Gas volume (m3/s) 12.30 12.30 Scenario 1 Base case Scenario 2 Worstcase

100% RFCC gas

Amount of gas emitted (kg/h) 53144 NH3 (kg/h) 0.3 CO2 (kg/h) 1799 H2O (kg/h) 697 N2 (kg/h) 47057 O2 (kg/h) 2840 Ar (kg/h) 751 Export gas stack CO2 (kg/h) 10196 NH3 (kg/h) 0.09 * Flue gas stack highest point in CAP area

53152 8 1799 697 47057 2840 751 10196 0.5

14.76 0.08 499.7 194 13071 788.9 209 2832 0.025

14.76 2.2 499.7 194 13071 788.9 209 2832 0.14

NILU OR 41/2010

24

Table 7:

Emission data for the Alstom plant using flue gas from the CHP. To the left the original emission data provided by Alstom, to the right the emission rates per second.
Scenario 3 Base case Scenario 4 Worstcase 48 40 * 18.3 48979 17.3 18.3 48987 17.3 Amount of gas emitted (kg/s) NH3 (g/s) CO2 (g/s) H2O (g/s) N2 (g/s) O2 (g/s) Ar (g/s) CO2 (g/s) NH3 (g/s) Gas volume (m3/s) 13.61 13.61 Scenario 3 Base case Scenario 4 Worstcase

100% CHP gas

Stack height (m) Stack diameter (inch) Height on adjacent buildings Gas temperature C Gas volume (m3/h) Velocity out of stack (m/s)

48 40 *

Amount of gas emitted (kg/h) 58285 NH3 (kg/h) 0.3 CO2 (kg/h) 496 H2O (kg/h) 732 N2 (kg/h) 46505 O2 (kg/h) 9759 Ar (kg/h) 793 Export gas stack CO2 (kg/h) 2813 NH3 (kg/h) 0.02 * Flue gas stack highest point in CAP area

58294 8.7 496 732 46505 9759 793 2813 0.25

16.19 0.08 138 203 12918 2711 220 781 0.006

16.19 2.4 138 203 12918 2711 220 781 0.07

3.2

Aker Clean Carbon using amines

ACC has provided 6 emission scenarios. Five of these are fairly similar, the only difference is the scaling of the various components. These five emission scenarios are named Expected (scenario 1), Design (scenario 2), Worst Case (scenario 3), Upset (scenario 4), and Minimum case (scenario 6). The physical parameters are given in Table 8 and components emitted are listed in Table 9 for flue gas from CHP and in Table 10 for flue gas from RFCC. Scenario 1 Expected and scenario 2 Design are listed. Scenario 4 Upset is considered to be the maximum possible emission scenario from TCM. These large amounts of MEA (6 ppm) will not be emitted over long periods of time, but rather as events or episodes. Scenario 3 Worst case emitting 3 ppm of MEA (not listed here) is considered to be the largest long term emission. In the literature expected emissions from a post-combustion CO2 plant using amines are between 1 and 4 ppm amines (Karl et al., 2010). There is continuous scientific research and efforts going on to reduce these emissions. By introducing new equipment (e.g. water wash) these emissions may be reduced considerably. It is a challenge to find the best possible technical solutions both from an engineering point of view and from an environmental point of view at the lowest possible cost. Scenario 1 (Expected) will emit 0.5 ppm MEA, this reflects the latest achievements in emission reducing technologies.

NILU OR 41/2010

25 These emissions contain a mother amine (e.g. MEA) in addition to e.g. NH3, formaldehyde, acetaldehyde, acetone and formamide. The emissions of these components are about the same order of magnitude (in g/s). These components are relatively well known and threshold values exist (Table 3 - Table 4). In addition there will be emissions of a large number of other products. These are formed chemically inside the ACC plant. Some of these compounds are not well known, and there are no threshold values. The sum of the lesser known compounds is of the same order of magnitude as the emissions of the mother amine. Hence the emissions of these less known compounds are not negligible. Another particular aspect concerning the Technology Centre is that the CO2 removed from the flue gas will be emitted to air separately. Also other components will be emitted along with the extracted CO2. For example Scenario 2 (Design) estimate 0.03 g/s of acetic acid using flue gas from RFCC. In a full scale plant this CO2 and the other products will not be emitted to air, but will be sent by pipeline to storage under the seabed in the North Sea. Table 8: Physical parameters concerning emissions from Absorber stack and CO2 stack, using flue gas from both CHP and RFCC.
CHP Absorber out 1.0 45.9 101.0 2418.6 66882.7 63496.6 1.0 22.5 1.05 0.0051 0.0991 0.1377 0.7581 RFCC CO2 out 1.0 27.8 151.0 71.5 3101.4 1175.2 0.156 2.64 0.9754 0.0242 0.0001 0.0003 Absorber out 1.0 48.0 101.0 1837.0 50148.2 48535.7 1.0 17.2 1.03 0.0133 0.1102 0.0435 0.8329 CO2 out 1.0 27.8 151.0 226.6 9830.3 3723.8 0.156 2.64 0.9757 0.0241 0.0000 0.0002

Vapour Fraction Temperature [C] Pressure [kPaa] Molar Flow [kgmole/h] Mass Flow [kg/h] Actual Volume Flow [Am3/h] Stack diameter [m] Velocity out of stack [m/s] Mass Density [kg/m3] Master Comp Mole Frac (CO2) Master Comp Mole Frac (H2O) Master Comp Mole Frac (Oxygen) Master Comp Mole Frac (Nitrogen)

NILU OR 41/2010

26 Table 9: Emission data given by ACC using flue gas from the CHP. Two scenarii are given, scenario 1-Expected and scenario 2-Design, quantities given in mixing ratio and in g/s.
Scenario 1 - Expected Absorber out Absorber out (g/s) (ppmv) 0.5 0.020517418 2.6 0.029746897 1.3 0.026227226 1.3 0.038471838 0.13 0.005072518 0.26 0.007867294 0.0026 0.00010318 0.26 0.005425359 0.026 0.000787428 0.26 0.007874281 0.0026 0.000127756 0.0026 0.000127756 0.0026 0.000225748 0.0026 0.0026 0.00010325 0.000152245 Scenario 2 - Design Absorber out Absorber out (g/s) (ppmv) 1 0.041034835 5.1 0.058349682 2.55 0.051445712 2.55 0.075463989 0.255 0.00994994 0.51 0.015432 0.0051 0.000202391 0.51 0.01064205 0.051 0.001544571 0.51 0.015445706 0.0051 0.000250599 0.0051 0.000250599 0.0051 0.000442813 0.0051 0.0051 0.000202528 0.000298635

CHP

MEA NH3 Formaldehyde Acetaldehyde Acetone Formamide Acetamide Methylamine Ethylamine Dimethylamine Diethylamine 1-Butanamine Dibutylamine Nmethylethanamine N-methyl 1butanamine N-ethyl 1butanamine 1-Propanamine Dipropylamine Formic acid Acetic acid Butyric acid Propinoic acid DEA HEI HEF Nitrosamines

MEA

0.0026 0.000176752 0.0013 5.16248E-05 0.00026 1.76752E-05 0.026 0.000804022 0.026 0.001048914 0.026 0.001538963 0.026 0.001293981 0.0005 3.53176E-05 0.026 0.001958614 0.026 0.001556166 0.0015 9.06959E-05 Scenario 1 - Expected CO2 stack CO2 stack (g/s) (ppmv) 0.166666667 0.000202187

0.0051 0.000346706 0.00255 0.000101264 0.00051 3.46706E-05 0.051 0.00157712 0.051 0.002057486 0.051 0.003018734 0.051 0.002538194 0.001 7.06353E-05 0.051 0.003841897 0.051 0.00305248 0.003 0.000181 Scenario 2 - Design CO2 stack CO2 stack (g/s) (ppmv) 0.333333333 0.000404375

NILU OR 41/2010

27 Table 10: Emission data given by ACC using flue gas from the RFCC. Two scenarii are given, scenario 1-Expected and scenario 2-Design, quantities given in mixing ratio and in g/s.
RFCC Scenario 1 - Expected Absorber out Absorber out (g/s) (ppmv) 0.5 0.015583844 16.5 0.143385138 0.825 0.012641972 0.825 0.018544085 0.825 0.024450408 1.65 0.037921707 0.0165 0.000497344 0.165 0.002615116 0.165 0.003795538 0.0165 0.000379554 0.0165 0.000615807 0.0165 0.000615807 0.0165 0.001088144 0.0165 0.0165 0.00049768 0.000733849 Scenario 2 - Design Absorber out (ppmv) 1 33 1.65 1.65 1.65 3.3 0.033 0.33 0.33 0.033 0.033 0.033 0.033 0.033 0.033 Absorber out (g/s) 0.031167687 0.286770277 0.025283944 0.03708817 0.048900815 0.075843413 0.000994687 0.005230232 0.007591077 0.000759108 0.001231614 0.001231614 0.002176288 0.000995361 0.001467698

MEA NH3 Formaldehyde Acetaldehyde Acetone Formamide Acetamide Methylamine Ethylamine Dimethylamine Diethylamine 1-Butanamine Dibutylamine Nmethylethanamine N-methyl 1butanamine N-ethyl 1butanamine 1-Propanamine Dipropylamine Formic acid Acetic acid Butyric acid Propinoic acid DEA HEI HEF Nitrosamines

MEA NH3 Formaldehyde Acetaldehyde Acetone Formic acid Acetic acid

0.0165 0.000851975 0.00825 0.00024884 0.00165 8.51975E-05 0.165 0.003875524 0.165 0.005055947 0.165 0.007418063 0.165 0.006237211 0.0005 2.68252E-05 0.165 0.009440855 0.165 0.007500988 0.0015 6.88873E-05 Scenario 1 - Expected CO2 stack CO2 stack (g/s) (ppmv) 0.166666667 0.000640866 0.230325895 0.000246932 0.279895105 0.00052914 3.816231555 0.010582793 0.102267734 0.000373925 0.730414947 0.002116559 3.172672773 0.011993832

0.033 0.001703951 0.0165 0.00049768 0.0033 0.000170395 0.33 0.007751049 0.33 0.010111894 0.33 0.014836127 0.33 0.012474423 0.001 5.36505E-05 0.33 0.018881709 0.33 0.015001975 0.003 0.000138 Scenario 2 - Design CO2 stack (ppmv) CO2 stack (g/s) 0.333333333 0.001281733 0.460651791 0.000493864 0.727727273 0.001375763 9.922202043 0.027515263 0.265896109 0.000972206 1.899078862 0.005503053 8.248949209 0.031183964

In addition ACC will test two different Improved solvents (scenario 5). These are listed in Table 11 and Table 12, values given in ppmv only. Improved solvent 1 Design (5a) and Worst case (5b) are both based on S3. Improved solvent 2 (5c and 5d) is based on S1, S4 and S9. The emission values from the two flue gas

NILU OR 41/2010

28 sources CHP and RFCC are identical in mixing ratio (ppmv), but the values are different in mass (g/s) due to different flue gas strength between the two sources (CHP has higher actual volume flow, see Table 8). For this scenario 5 the emissions of less known degradation products are often as high or higher than the mother amine. This makes it complicated to evaluate the environmental effect of the emissions. Table 11: Emission data given by ACC using flue gas from the CHP and RFCC. Scenario 5 Improved solvent 1, quantities given in mixing ratio (ppmv).
Scenario 5 - Improved solvent 1
Tertiary amine Ammonia Formaldehyde Acetaldehyde Acetone Formamide Acetamide Methylamine Ethylamine Dimethylamine Diethylamine 1-Butanamine Dibutylamine N-methylethanamine N-methyl 1-butanamine N-ethyl 1-butanamine 1-Propanamine Dipropylamine

5a Design (S3)
0.002 5 1 2 0.5 0.1 0.001 0.255 0.165 0.255 0.0165 0.0165 0.0165 0.0165 0.0165 0.0165 0.0083 0.0017

5b - Worst case(S3)
0.02 5 1 2 0.5 0.1 0.001 0.515 0.329 0.515 0.0329 0.0329 0.0329 0.0329 0.0329 0.0329 0.01645 0.00329

2,2'-[[2-[(2-hydroxyethyl)methylamino]ethyl]imino]bisEthanol Ethyleneglycol N,N.dimethylethanolamine MMEA 2-methylaminoethanol 1-hydroxyethyl-3-methyl imidazolidone Hydroxyethyl oxazolidone Formic acid Acetic acid Butyric acid Propinoic acid DEA Nitrosamines

0.1 0.1 0.002 0.002 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.0002 0.003

0.1 0.1 0.02 0.02 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.002 0.003

NILU OR 41/2010

29

Table 12: Emission data given by ACC using flue gas from the CHP and RFCC. Scenario 5 Improved solvent 2, quantities given in mixing ratio (ppmv).
Scenario 5 - Improved solvent 2 (S1, S4, S9) Primary amine Secondary amine Tertiary amine Ammonia Formaldehyde Acetaldehyde Acetone Formamide Acetamide Piperazine 1,2 ethylenediamine 4,4-dimethyl-2-oxazolidinone 2-methyl-2-(methylamino)-1-propanol NO Nitrosamines 5c - Design 0,88 0.115 0.005 1 0.2 0.4 0.1 0.02 0.0002 0.1 0.1 0.1 0.1 0.1 0.003 5d - Worst case 2.24 0.7 0.06 1 0.2 0.4 0.1 0.02 0.0002 0.1 0.1 0.1 0.1 0.1 0.003

3.2.1 Mobile Test Unit In the early phase of the project, before the construction of TCM is completed, there will be tests run on a Mobile Test Unit (MTU). The ACC emission data depicted in Table 9 - Table 12 will also be used. However the quantities emitted will be very small. The actual volume flow from MTU will be maximum 1112,7 m3/h compared to 50000-60000 m3/h from the ACC plant. This means that the emissions from the MTU in mass (g/s) are a factor 50 less than emissions from the plant even though the emissions in ppmv are similar. Maximum emissions of MEA will for instance be 0.0042 g/s. For this reason the emissions from MTU and the subsequent concentrations are not investigated in detail. 3.2.2 Emissions of CO2 The CO2 that is captured will be emitted through a separate stack. The density of CO2 is about 1,5 the density of air. Some concern have been raised that pure CO2 will sink to the ground and cause high concentrations. This situation has been observed elsewhere during very stable and calm atmospheric conditions. Mongstad is located at the west coast of Norway. During the NILU monitoring project at Kollsnes (20 km from Mongstad) 2008-2009 there were wind 100% of the time (wind speed > 0.4 m/s). This means that and there is hardly never calm and stable conditions in the Mongstad area, and the CO2 emitted is expected to be well mixed.

NILU OR 41/2010

30

Maximum hourly mean concentrations of main pollutants

This section contains calculated maximum hourly mean concentrations in the vicinity of TCM (up to a few kilometres from Mongstad). For these calculations the CONCX model is used (Bhler, 1987). CONCX is a simple and robust Gaussian distribution model that calculates concentrations downwind of an emission source at various wind speeds and under various atmospheric stability conditions. There is no chemical degradation included, nor loss by deposition. This means that the compounds are assumed to chemically inert. Such an assumption is valid if the components have a lifetime longer than minutes to hours. The model requires the following input: Stack height, height of adjacent buildings (due to the turbulence zone around these buildings), emission velocity and flue gas temperature (in order to calculate thermal lift). This information is listed in section 3. Then the CONCX model calculates concentrations at ground level as a function of the distance from the source. For Gaussian models in general, there is 95 % probability that the model result is within a factor of 2 of the reality, However, due to the conservative approach in the parameter setting in CONCX a better uncertainty estimate is that it is 95 % probability for the real concentrations to be within a range of 0.25 to 1.25 of the predicted results. The results are compared with the Norwegian Labour Inspection Authoritys administrative norms and national and international air quality criteria (see section 2). The maximum short term concentrations of pollutants emitted from TCM will occur between a few hundred meters up to a few kilometres from TCM, depending on the atmospheric conditions and wind. These maxima are well within Statoil Mongstads industrial area, where the Norwegian Labour Inspection Authoritys norms apply, while other air quality criteria apply outside the area. 4.1 Unit emission

The calculations were first carried out using emissions of 1 g/s, a so-called unit emission. If the emission component is inert, i.e. not chemically degraded within the time scale of local dispersion (up to a few hours) and the emission parameters remain the same, these results can be linearly scaled in order to get concentrations for other quantities emitted. This must be regarded as a first approach to estimating concentrations. The next step in model development would be to include a chemistry scheme into the model, but for the components investigated here and their chemical lifetime ( > minutes/hours) it is valid to assume that they remain inert at this short geographical scale (kilometres).

NILU OR 41/2010

31

30 25 20 g/m3 15 10 5 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8

Alstom RFCC Emissions 1 g/s

Ustable 1 m/s Neutral 3 m/s Light stable 1 m/s

km downwind 25 20 15 Unstable 1 m/s 10 5 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Neutral 1 m/s Light stable 1 m/s

Alstom CHP Emissions 1 g/s

g/m3

km downwind

Figure 4: Results for unit emission of 1 g/s, Alstom emission data, flue gas from RFCC (top panel) and CHP (lower panel For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3.

NILU OR 41/2010

32
14 12 10 g/m3 8 6 4 2 0 0,1 10 9 8 7 g/m3 6 5 4 3 2 1 0 0,1 0,2 0,3 0,4 0,5 1 2 km downwind 3 4 8 Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s 0,2 0,3 0,4 0,5 1 2 km downwind 3 4 8
Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s

ACC RFCC Emissions 1 g/s

ACC CHP Emissions 1 g/s

Figure 5: Results for unit emission of 1 g/s, ACC emission data, flue gas from RFCC (top panel) and CHP (lower panel). For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3. Figure 4 and Figure 5 show unit emission results for the two technology vendors (Alstom and ACC). The wind speed that yield the highest concentrations for each stability class are shown for each source. The concentration profile shows zero concentration near the stack, then a maximum at a certain distance (between 400 and 3000 m), and then the concentration decreases. This concentration pattern shows that the stack height is sufficient for dispersion. Neutral, unstable and light stable describe the atmospheric stability, i.e. the vertical mixing. Neutral conditions, when temperature decreases with height, occur during cloudy weather, rain and during strong wind. Neutral conditions give fairly good vertical mixing. Unstable conditions, when temperature decreases

NILU OR 41/2010

33 rapidly with height, occur during strong incoming solar radiation that heats the ground. Unstable conditions give very good vertical mixing of the emissions, but for emissions from high stacks this may be a disadvantage since the emissions are brought rapidly to the ground close to the stack. Light stable and stable conditions, when temperature increases with height, occur during nighttime and during wintertime when there is strong cooling at the ground by outgoing longwave radiation. Stable conditions are ideal for emissions from high stacks, since the emissons will reach the ground at a long distance from the stack. The Alstom calculations show higher concentrations closer to the stack than the ACC calculations. E.g. for the RFCC flue gas source the calculated maximum concentrations are 25g/m3 400 m from the stack for Alstom while ACC have maximum 11,5g/m3 500 m from the stack. This is due to difference in the outlet temperature. The temperature from the ACC facility is higher (45C vs. 18 C for Alstom), hence there will be a thermal lifting. The plume is lifted higher before horizontal dispersion occurs and maximum concentration using the ACC emission data is found further downwind and is consequently lower. The results presented here are calculated using outflow temperature of 18 C. Lower temperatures may occur (~10 C). However this will not alter the results significantly regarding to compliance with the air quality guidelines. The ACC emissions using flue gas from the CHP (Figure 5 lower panel) will give somewhat lower hourly mean. This is due to higher velocity out of the stack, 22,5 m/s for CHP vs. 17,2 m/s for RFCC. Due to this the CHP plume will rise 10 m higher than the RFCC plume before horizontal dispersion. Hence the plume will reach the ground at a longer distance from the stack, and hence show lower concentrations. This is also valid for the Alstom emission data (Figure 4), the CHP case gives higher outlet temperature and higher velocity out of the stack. Alstom may run their process with outlet temperature down to 10C. In this case there will be hardly thermal lifting, vertical dispersion will depend on velocity out of the stack only. In addition to the emissions from the absorber units discussed above there will be emissions from the export gas stack. The gas will consist mainly of CO2, with the addition of small quantities of NH3 for Alstom, and the components listed in the bottom part of Table 10 for ACC. Model runs have been made for CO2, the maximum concentration value of the additional components have been calculated by scaling.

NILU OR 41/2010

34

4.2

Emissions from Alstom Chilled ammonia technology

To obtain maximum hourly concentration from TCM the results using unit emission (1 g/s) are scaled according to the emission data described in section 3.1 (Table 6). For NH3, the results from both the RFCC and CHP emissions are shown, emitting 2,2 g/s and 2,4 g/s respectively in worst case. For CO2, only the RFCC emission data is shown. The reason for this is that the RFCC emission data is three times larger than the CHP (499,7 g/s vs. 138 g/s). The results are shown in Figure 6 and Figure 7.
60 50 40 g/m3 30 20 10 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 1 m/s Neutral 3 m/s Light stable 1 m/s

Alstom RFCC Worst Case 2,2 g NH3/s

km downwind 60 50 40 g/m3 30 20 10 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 1 m/s Neutral 1 m/s Light stable 1 m/s

Alstom CHP Worst Case 2,4 g NH3/s

km downwind

Figure 6: Results for worst case emission scenario, Alstom emission data, RFCC emitting 2,2 g NH3/s (top panel) and CHP emitting 2,4 g NH3/s (lower panel). For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3.

NILU OR 41/2010

35

14000 12000 10000 g/m3 8000 6000 4000 2000 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8

Alstom RFCC Worst Case 499,7 g CO2/s

Unstable 1 m/s Neutral 3 m/s Light stable 1 m/s

km downwind

Figure 7: Results for worst case emission scenario, Alstom emission data, RFCC emitting 499,7 g CO2/s. For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3. For NH3 the calculated maximum hourly concentration is 55 g/m3 (0.055 mg/m3), for CO2 the maximum is 12400 g/m3 (12.4 mg/m3). Administrative norms for NH3 and CO2 that applies at the TCM plant are 18 mg/m3 and 9000 mg/m3 respectively (Table 3). This means that the concentrations are more than a factor 300 and 700 lower than the norm. Export gas stack The concentrations of CO2 for emissions from the export gas stack are shown in Figure 8. The maximum ground level CO2 concentration is calculated to be 71 mg/m3 using export gas from RFCC cleaning. This is less than 1 % of norm from the Labour Authorities. For NH3 the maximum concentration is calculated to be 3.5 g/m3 (worst case emissions with RFCC cleaning). This is less than 1 of the norm.

NILU OR 41/2010

36

Alstom CO2 export RFCC

Alstom CO2 export CHP

Downwind distance (m)

Figure 8: Concentrations of CO2 from emissions through the export gas stack, both scenarios. Upper pane RFCC gas, lower pane CHP gas.

4.3

Emissions from Aker Clean Carbon amine technology

In the same manner, to obtain maximum hourly concentration from TCM the results using unit emission (1 g/s) are scaled according to the emission data described in section 3.2. In scenarii 1-4 and 6 (Table 9 and Table 10) there are 25 components emitted, for scenarii 5a-5d Improved solvents (Table 11 and Table 12) there are 30 and 14 components respectively. MEA, NH3, formaldehyde, acetaldehyde, and alkylamines (methylamine, ethylamine, dimethylamine, diethylamine) are shown here. These are the main components emitted from the ACC facility. For the other components the quantities emitted are fairly small, i.e.

NILU OR 41/2010

37 mg/s or g/s scale. And since emissions of 1 g/s gives a maximum concentration of 25 g/m3 (Figure 4), then emissions of 1 mg/s will give 25 ng/m3 maximum concentration and so on. These are much lower concentrations than any of the air quality guidelines or threshold values listed in Table 3 - Table 5. For MEA, scenario 4 Upset emitting 0.187 g MEA/s (RFCC) and 0.246 g MEA/s (CHP) respectively are shown. Concerning NH3, emitted at a rate of 0.85 g NH3/s (equals 98,7 ppmv, see Table 10) only the RFCC emission data is displayed. Formaldehyde, acetaldehyde, and alkylamines are emitted at a rate 0.15 g/s, 0.22 g/s and 0.08 g/s respectively using flue gas from CHP.
2,5 2 1,5
Unstable 1 m/s

ACC MEA RFCC 0,187 g/s

g/m3

Unstable 2 m/s Neutral 1 m/s

0,5 0 0,1 0,2 0,3 0,4 0,5 1 2 km downwind 2,5 2 1,5 Unstable 1 m/s 1 Unstable 2 m/s Neutral 1 m/s 0,5 0 0,1 0,2 0,3 0,4 0,5 1 2 km downwind 3 4 8 3 4 8

ACC MEA CHP 0,246 g/s

Figure 9: Results using ACC emission data, flue gas from RFCC emitting 0.187 g MEA/s (top panel) and flue gas from CHP emitting 0.246 g MEA/s (lower panel). For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3.

g/m3

NILU OR 41/2010

38

12 10 8 g/m3 6 4 2 0 0,1 0,2 0,3 0,4 0,5 1 2 km downwind 3 4 8

ACC NH3 RFCC 0,85 g/s

Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s

Figure 10: Results using ACC emission data, flue gas from RFCC emitting 0.85 g NH3/s. For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3. The maximum calculated hourly concentration of MEA is about 2 g/m3. This is much lower than the Administrative norm (Table 3) and also lower than long term exposure guidelines (Table 4). Maximum calculated NH3 concentration is just below 10 g/m3. This is a factor 1800 lower than administrative norm (18 mg/m3).

NILU OR 41/2010

39
1,6 1,4 1,2 g/m3 1 0,8 0,6 0,4 0,2 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s

ACC Formaldehyde CHP 0,15 g/s

km downwind

Figure 11: Results of formaldehyde, acetaldehyde, and alkylamines using ACC emission data, flue gas from CHP emitting 0.15 g/s formaldehyde (upper panel), 0.22 g/s acetaldehyde (mid panel) and 0.08 g/s alkylamines (lower panel). For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3.

NILU OR 41/2010

40

2,5 2 1,5

ACC Acetaldehyde CHP 0,22 g/s

g/m3

Unstable 1 m/s 1 0,5 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 2 m/s Neutral 1 m/s

km downwind 0,9 0,8 0,7 0,6 g/m3 0,5 0,4 0,3 0,2 0,1 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s

ACC Alkyl amines CHP 0,08 g/s

km downwind

Figure 11: Contd. The maximum calculated hourly concentrations of formaldehyde, acetaldehyde, and alkylamines are 1.4 g/m3, 2.1 g/m3, and 0.8 g/m3 respectively. The administrative norms for all these components range from 0.6 mg/m3 (formaldehyde) to 45 mg/m3 (acetaldehyde, see Table 3), and all calculated maximum concentrations are well below these values.

NILU OR 41/2010

41
0,35 0,3 0,25 g/m3 0,2 0,15 0,1 0,05 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s

ACC Methylamine CHP 0,03 g/s

km downwind

Figure 12: Results of methylamine using ACC emission data, flue gas from CHP emitting 0.03 g/s For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3. Methylamine has a strong smell (rotten fish), and therefore it is investigated separately. Maximum emissions of methylamine is 0.03 g/s (scenario 4 Upset, CHP) and the maximum hourly mean concentration is 0.3 g/m3. This is the mean concentration for one hour. However the short term concentrations may vary considerably within one hour depending on the atmospheric conditions (turbulence etc.). For smell the short term maximum concentration is perhaps more important than the long term mean. As a general rule the maximum short term concentration within one hour is typically a factor 10 higher than the hourly mean. This means that for short periods of time methylamine concentrations of up to 3 g/m3 may occur. The threshold value for smell (luktterskel in Norwegian) for methylamine is 5 ppmv. This corresponds to 6 mg/m3, i.e. that the value for smell is half the administrative norm. Then the estimated short term concentrations outlined here will be much lower than the detection limit for smell.

For Scenario 5d Improved solvent 2 Worst case (S1, S4, S9) the maximum hourly concentrations of piperazine, primary amine, secondary amine, and tertiary amine are calculated. These will be emitted at a rate of 0.1ppmv/0.005 g/s for piperazine, and 2.24 ppmv for primary amine, 0.7 ppmv for secondary amine, and 0.06 ppmv for tertiary amine. The exact composition of this solvent is not known to NILU and the molar mass is based on a given estimate. This corresponds to approximately 0.09 g/s for primary amine (assuming molar mass about 60), and 0.045 g/s for secondary and 0.004 g/s for tertiary amine (assuming molar mass of about 100).

NILU OR 41/2010

42 The results are shown in Figure 13. Maximum concentrations of piperazine, primary amine, secondary amine, and tertiary amine are 0.055 g/m3, 0.85 g/m3, about 0,4 g/m3, and slightly less than 0.04 g/m3 respectively. Threshold value for working environment for piperazine is 0.3 mg/m3, so the maximum hourly concentration is well below the threshold value. Since the exact composition of these primary, secondary and tertiary amine emissions is not given, it is not possible to evaluate the environmental and health effects of these emissions.

0,06 0,05 0,04

ACC Piperazine CHP 0,0058 g/s

0,03 0,02 0,01 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8

Unstable 1 m/s
Unstable 2 m/s Neutral 1 m/s

km downwind

Figure 13: Results of piperazine, primary amine, secondary amine, and tertiary amine using ACC emission data 5d Improved solvent 2 Worst case (S1, S4, S9), flue gas from CHP. For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3.

NILU OR 41/2010

43
0,9 0,8

0,7
0,6

ACC Primary amine CHP 0,09 g/s

g/m3

0,5

0,4
0,3 0,2

Unstable 1 m/s
Unstable 2 m/s Neutral 1 m/s

0,1
0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8

km downwind 0,45 0,4

0,35
0,3
g/m3

ACC Secondary amine CHP 0,045 g/s

0,25

0,2
0,15 0,1

Unstable 1 m/s
Unstable 2 m/s Neutral 1 m/s

0,05
0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8

km downwind 0,04 0,035

0,03
0,025

ACC Tertiary amine CHP 0,004 g/s

g/m3

0,02

Unstable 1 m/s
Unstable 2 m/s Neutral 1 m/s

0,015
0,01

0,005
0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8

km downwind

Figure 13: Contd.

NILU OR 41/2010

44 The total emissions of NOx will be between 148 and 208 ppm using gas from RFCC (TCM design basis). Flue gas from CHP will have NOx levels below 5 ppm. Maximum emissions of NO2 from RFCC is 6 ppm. This is equal to an emission rate of 0,14 g/s footnote 2. The maximum hourly concentrations of NO2 are showed in Figure 14.
1,8 1,6 1,4 1,2 g/m3 1 0,8 0,6 0,4 0,2 0 0,1 0,2 0,3 0,4 0,5 1 2 3 4 8 Unstable 1 m/s Unstable 2 m/s Neutral 1 m/s

ACC NO2 RFCC 0,14 g/s

km downwind

Figure 14: Results of NO2 using ACC emission data, flue gas from RFCC emitting 6 ppm/0.14 g/s. For each source, the three meteorological conditions that provide the highest concentrations were selected. For information on the emission conditions see section 3. The horizontal axis shows the distance from the source in kilometres and the vertical axis shows concentration as an hourly mean in g/m3. Maximum calculated hourly concentration from TCM is then 1.6 g/m3. The NO2 contribution from TCM has to be added to the background concentration. In 20082009 NILU performed a monitoring program at Kollsnes (report in preparation). The average background concentration of NO2 at Kollsnes is below 4 g/m3. In sum the NO2 concentration at Mongstad will be below all air quality standards outlined in Table 2. The maximum downwind NO2 concentration from the emissions will depend upon the upwind ozone concentration rather than the small NO2 emissions (the reaction NO+O3 -> NO2+O2 will titrate O3 and give enhanced NO2 concentrations). NO2 is photolyzed back to NO (NO2+hv -> NO+O). This photolysis depend on the amount of sunlight and is not investigated further here. It should be noted that recently it has been a reduction in the NOx level due to installation of SNCR (Selective Non-Catalytic Reduction) in the RFCC flue gas system.

Assume total flow rate of 1837 kmole pr hour and molar mass 46.01 for NO2.

NILU OR 41/2010

45 Emissions through the export gas stack The primary component of the gas is obviously CO2. The ground level concentrations will be highest for RFCC gas cleaning. The concentrations of CO2 are shown in Figure 15. The maximum concentration is lust above 70 mg/m3, less than 1 % of the CO2 norm.

ACC CO2 export RFCC

mg/m3

Unstable 1 m/s Neutral 1 m/s Light Stable 1 m/s

Downwind distance (m)

Figure 15: Ground level CO2 concentrations from the export gas stack, cleaning of RFCC gas. The concentrations of the additional components given in the emission scenarios will be highest for scenario 4 upset. The calculated maximum ground level concentrations are shown in Table 13. These concentrations are small compared to the contribution from the absorber, except for acetaldehyde which is marginally smaller than the contribution from the absorber. Table 13: Maximum ground level concentrations from additional components emitted through the export gas stack. Component
MEA NH3 Formaldehyde Acetaldehyde Acetone Formic acid Acetic acid

Concentration (g/m3)
0.201 0.039 0.083 1.656 0.059 0.331 1.877

4.4

Summary of all calculated maximum hourly mean concentrations

To give a clearer overview all results from this chapter are summarized in Table 14. These are not new results, but taken from the text. These values are lower than the Administrative norms for the components listed in Table 3 and lower than the guidelines from FHI for the components listed in Table 4.

NILU OR 41/2010

46 Table 14: Maximum hourly mean concentrations of the components investigated in this section. Unit: g/m3. Technology vendor / component Alstom NH3 CO2 (from export gas stack) ACC MEA NH3 Formaldehyde Acetaldehyde Alkyl amines Methylamine Piperazine Primary amine Secondary amine Tertiary amine NO2 (from RFCC flue gas) CO2 (from export gas stack) Maximum hourly concentration 55 g/m3 71 900 g/m3 2 g/m3 10 g/m3 1.4 g/m3 2.1 g/m3 0.8 g/m3 0.3 g/m3 0.055 g/m3 0.85 g/m3 ~0.4 g/m3 slightly less than 0.04 g/m3 1.6 g/m3 70 500 g/m3

NILU OR 41/2010

47

5
5.1

Annual mean and 8-hours mean maximum concentrations and annual deposition of main pollutants
Unit emissions

To calculate annual mean and 8-hours mean concentrations, annual deposition and concentration in drinking water The Air Pollution Model (TAPM) developed by CSIRO, Australia, is applied. A detailed description of the model and its application can be found in Karl et al. (2010) or at TAPM (2009). TAPM is also used for dispersion calculations from Krst (Attalla and Azzi, 2010). For convenience a short description is given here. TAPM is an integrated model consisting of a prognostic meteorological module and a set of air quality modules. For this study the meteorological module is nested three times, from an initial domain of 600 600 km2 (grid resolution of 15 km) down to a domain of 80 x 80 km2 (2 km resolution) centered on TCM. Meteorological data represents year 2007. Initial and boundary conditions for the outermost grid are taken from the LAPS and GASP models from the Australian Bureau of Meteorology. Physical surface data, such as topography, land use and sea surface temperature are taken from various sources (the US Geological Survey, Earth Resources Observation Systems (EROS) Data Center Distributed Active Archive Center (EDC DAAC) and the US National Center for Atmospheric Research NCAR). The air quality modules used in the simulations are the Lagrangian Particle Model (LPM) and the Plume Rise Module (PRM). The LPM is based on the PARTPUFF model (Hurley, 1994) whereby mass is represented as a puff in the horizontal direction, and as a particle in the vertical direction. The PRM, used for point source emission, determines plume rise based on stack configuration and emission dynamics. In this study the emitted amines are assumed to be chemically inert but undergo both wet and dry deposition processes. The deposition of the amines are treated in the same way as for sulfur dioxide. The amines are assumed to be highly soluble and totally removed by wet deposition. Amines are deposited directly to the ground. Further degradation in soil/vegetation surfaces is not considered. The first results for unity emission (1 g/s) are shown. Assuming that the emitted amines are chemically inert (see section 5.3 for discussion on this assumption), these results can be scaled according to the quantity emitted. This is the same approach used to calculate maximum hourly concentration in section 4. Both annual mean concentration, maximum 8-hours mean concentration and annual wet deposition are calculated. Results from a model run emitting 1 g/s (unity emission) are shown in Figure 16. Annual mean concentrations are most important for the compounds where long term exposure, carcinogenic effects or chronic toxicity are the key parameters. Maximum 8-hours mean concentrations are important for the compounds where short term exposure or acute toxicity are the key parameters, whereas annual mean deposition calculation is necessary to determine possible effects on eutrophication and drinking water concentrations.

NILU OR 41/2010

48

Figure 16: Annual mean concentration (upper panel), maximum short term concentration (8 hrs., mid panel) and annual wet deposition (lower panel) for a model run emitting 1 g/s. Horizontal scale is in km. Units: g/m3 for concentrations and mg/(m2 year) for deposition.

NILU OR 41/2010

49 For the annual mean concentration the prevailing wind directions are clearly seen, from the south and from the north-west. The maximum annual mean concentration (0.03 g/m3) is found about 10 km downwind of TCM in S-E direction. Maximum monthly mean concentrations vary between 0.03 and 0.15 g/m3 (for the 12 months January-December, results not shown). Note that the location of the monthly maxima differ throughout the year. Hence the maximum annual mean value is not equal to the average of the maximum monthly values. The highest short term concentration, 2.4 g/m3 for a 8-hr period is found in June, again S-E of the CO2 Technology Centre. This is considerably lower than the hourly mean concentrations showed in Figure 4 and Figure 5. This is due to differences in spatial and temporal resolution in the model results. CONCX gives a detailed description of the maximum hourly concentration close to the stack. In Figure 16 TAPM gives 8-hour average values representative for 22 km2 grid squares. The wet deposition is a combination of synoptic scale weather patterns modified by topography. Wet deposition has maximum east of TCM. This is due to orographic rainfall, a well-known phenomenon in the western part of Norway. Annual mean rainfall at Mongstad is 1890 mm/year (met.no, eKlima.no), but there are considerable local differences. Maximum annual wet deposition is calculated to be 39 mg/(m2 year) emitting unit emission of 1 g/s. TAPM calculates the atmospheric meteorological parameters (advection/wind, temperature, humidity, rainfall) based on the laws of physics that governs the atmosphere. There are uncertainties in these calculations. The uncertainty or error for TAPM is less than 30 % for predictions of observed mean concentrations of passive agents released from stacks. It is also important to emphasize that the meteorological data (weather) represents the year 2007. In the real atmosphere there are inter-annual variations in weather. Hence the dispersion from TCM will change accordingly. This means that deposition, location and magnitude of maxima etc. may vary from year to year. However the results presented here are considered representative for the situation that will occur in the vicinity of Mongstad when TCM will be in operation. 5.2 Emissions from Alstom facility

The emission data from Alstom is listed in Table 6 and Table 7. It does not contain any toxic substances, only NH3, CO2 and water vapour + normal air (N2, O2 and Ar). CO2 is not toxic. It is a green house gas, but that topic is beyond the scope of this report. Concerning NH3, the current analysis has covered concentrations in working atmosphere and ambient air and possible eutrophication effects on vegetation. The calculations of maximum hourly concentrations (section 4.2) reveal that calculated concentrations of NH3 are much lower than the administrative norms for working environment (18 000 g/m3) and the estimated air quality guideline of 180 g/m3 (1/100 of the norms). NH3 will, however,

NILU OR 41/2010

50 contribute to eutrophication, i.e. excess of nitrogen in soil and water. Hence, the annual mean concentration in air and annual mean deposition of nitrogen must be calculated. The maximum emission rate from the Alstom facility is 2.4 g NH3/s (Worst case using flue gas from CHP, see Table 7) and a total of 2.5 g/s when adding the emissions from the export stack. For short term (8 hours mean) maximum concentration, this is the correct scenario to use. Hence the results from the unity emissions are scaled by a factor 2.5. On the other hand, the average emission rate of NH3 from Alstom is 0.08 g NH3/s (both base case scenarii, RFCC and CHP, see Table 6 and Table 7) and a total of 0.105 g/s adding the amissions from the export stack. For calculation of annual mean deposition, the base case scenario is most correct, and 0.105 g NH3/s equals 0.086 g N/s footnote 3. Both maximum short term concentration and annual mean wet deposition are shown in Figure 17. Evidently the results are very similar to the results in Figure 16 using unity emission since it is a question of scaling (for the 8-hr maximum with a factor of 2.5 and for the annual deposition with a factor of 0.086). The maximum 8-hrs mean concentration (6.1 g/m3) is seen S-E of Mongstad whereas wet deposition shows two maxima, one to the north of Mongstad and one to the east. Calculated maximum deposition is 3.2 mg N/(m2 year). Concerning wet deposition the value of 3.2 mg N/(m2 year) is relatively low, around 0.3 %, compared to the annual nitrogen deposition of 800-1000 mg N/(m2 year) near Mongstad (Figure 1). However if the worst case scenario applies the deposition will be nearly a factor 30 higher and wet deposition increases to over 70 mg N/(m2 year) (results not shown). Given that nitrogen deposition at Mongstad is already over the threshold values for vegetation any extra nitrogen will contribute to eutrophication.

Molar mass 14 for N and 1 for H, thus multiply by 14/17 to get total N instead of NH 3.

NILU OR 41/2010

51

Figure 17: Maximum short term concentration of NH3 (upper panel) and annual mean wet deposition of N (lower panel) from the Alstom facility assuming an emission rate of 2,4 g NH3/s and 0.08 g NH3/s respectively. Horizontal scale is in km. Units: g/m3 for concentrations and mg/(m2 year) for deposition.

NILU OR 41/2010

52

To summarize this section, the maximum short term concentration from Alstom chilled ammonia plant is calculated to be 6.1 g NH3/m3 (June) whereas the maximum wet deposition is 3.2 mg N/(m2 year). 5.3 Results of amines and degradation products using emissions from ACC facility

In this section calculated concentrations of amines and degradation products are presented. The results are based on the unit emission calculations outlined in section 5.1, scaled according to the emission data given by ACC. This method is valid if the compounds are assumed to be inert or to have a lifetime longer than hours/days. For example MEA has an atmospheric lifetime of about a day, formaldehyde has a lifetime of typically 8-10 hrs, acetaldehyde about half to one day. This means that most of the mass emitted from TCM will not be degraded within the 30 km model grid. To give an example; if the wind speed is 2 m/s, or 7,2 km/hr, the flue gas emitted from TCM will be transported 30 km away in about 4 hours. So if the lifetime of a component is about half a day (12 hours) or more, most of the molecules of this component will not be degraded within 30 km from TCM, and it may be considered inert in these calculations. Of course there will be some chemical loss, even if the lifetime is long, but the loss will be small compared to the total mass/concentration of the component. The calculated concentrations will be compared to air quality limit values (see section 2). Concentrations of MEA, acetaldehyde, formaldehyde, alkyl amines and piperazine will be shown. Concerning nitrosamines, please see section 5.4. (worst case calculations). MEA: Scenario 2 (Design,) estimates 1 ppmv of MEA in the flue gas. This is equal to 0.04 g MEA/s using flue gas from CHP. Hence the results using unit emissions are scaled by a factor 0.04. The calculated annual mean concentrations are shown in Figure 18. Maximum annual concentration of MEA is below 1.3 ng/m3 (0.0013 g/m3) downwind of TCM in S-E direction. This is several orders of magnitude lower than the exposure guideline for MEA (10 g/m3) given by FHI (Table 4).

NILU OR 41/2010

53

Figure 18: Annual mean concentration of MEA for a model run emitting 0.04 g/s. Horizontal scale is in km. Units: ng/m3.

NILU OR 41/2010

54

Formaldehyde: Scenario 2 (Design) estimates 0.05 g/s of formaldehyde using flue gas from CHP. Hence the results using unit emissions are scaled by a factor 0.05. The calculated annual mean concentrations are shown in Figure 19. Maximum annual concentration of formaldehyde is just above 1.6 ng/m3 downwind of TCM in S-E direction. This is several orders of magnitude lower than the exposure guideline (1.2 mg/m3) given by FHI (Table 4).

Figure 19: Annual mean concentration of formaldehyde for a model run emitting 0.05 g/s. Units: ng/m3.

NILU OR 41/2010

55

Acetaldehyde: Scenario 2 (Design) estimates 0.07 g/s of acetaldehyde using flue gas from CHP. Hence the results using unit emissions are scaled by a factor 0.07. The calculated annual mean concentrations are shown in Figure 20. Maximum annual concentration of acetaldehyde is 2.4 ng/m3 (0.0024 g/m3) downwind of TCM in S-E direction. This is much lower than the exposure guideline (2 g/m3) given by FHI (Table 4).

Figure 20: Annual mean concentration of acetaldehyde for a model run emitting 0.15 g/s. Horizontal scale is in km. Units: ng/m3 .

NILU OR 41/2010

56

Alkyl amines: Scenario 2 (Design) lists several alkyl amines, among others methylamine, ethylamine, dimethylamine, diethylamine. Methylamine (0.011 g/s) and dimethylamine (0.015) are the most important ones in terms of emission strength. In all the total emissions of alkyl amines are about 0.03 g/s using flue gas from CHP. Hence the results using unit emissions are scaled by a factor 0.03. The calculated annual mean concentrations are shown in Figure 21. Maximum annual concentration of alkyl amines is just below 1.0 ng/m3 downwind of TCM in S-E direction. Secondary and tertiary alkyl amines have a potential to form nitrosamines, primary alkyl amines do not. However this is not investigated in further detail here.

Figure 21: Annual mean concentration of alkyl amines for a model run emitting 0.03 g/s. Horizontal scale is in km. Units: ng/m3.

Piperazine Scenario 5c (Improved solvent 2 Design (S1, S4, S9)) estimates 0.1 ppmv/0.0057 g/s of piperazine using flue gas from CHP. Hence the results using unit emissions are scaled by a factor 0.0057. The calculated annual mean concentrations are shown in Figure 22. Maximum annual concentration of piperazine is 0.18 ng/m3. downwind of TCM in S-E direction. This is much lower than the exposure guideline (5 g/m3) given by FHI (Table 4).

NILU OR 41/2010

57

Figure 22: Annual mean concentration of piperazine for a model run emitting 0.0057 g/s (scenario 5c Improved solvent 2 Design, S1, S4, S9). Horizontal scale is in km. Units: ng/m3.

Summary all calculated maximum annual mean concentrations To give a clearer overview of all results from this chapter, the results from ACC facility are summarized in Table 15. These are not new results, but taken from the text above. These maximum annual mean concentrations were compared to the guidelines from FHI (Table 4) for MEA, formaldehyde and acetaldehyde and are well under the guidelines. There are no guidelines for long term exposure to alkyl amines nor for piperazines. Table 15: Annual mean concentrations of the components investigated in section 5.3. Unit: g/m3. Component MEA Formaldehyde Acetaldehyde Alkyl amines Piperazine Emission scenario 2 (Design) 2 (Design) 2 (Design) 2 (Design) 5c Improved solvent 2 Design Emissions 0.04 g/s 0.05 g/s 0.07 g/s 0.03 g/s 0.0057 g/s Maximum annual concentration 1.3 ng/m3 1.6 ng/m3 2.4 ng/m3 1.0 ng/m3 0.18 ng/m3

NILU OR 41/2010

58 5.4 Worst case scenario nitrosamines in air

The worst case scenario is an estimate that maximizes the risk of the amine emissions to air. This is a model estimate that represents the worst possible scenario. The model assumptions made for these calculations are a simplification of the actual processes involved in amine dispersion and degradation in air. In the model calculations there is no specific amine chemistry scheme included, i.e. the model does not have a chemistry scheme to account for degradation of the compounds emitted from TCM. The chemical degradation of amines is complicated and there are considerable knowledge gaps. For instance, gas phase degradation of MEA involves more than 100 possible degradation products, whereas degradation of piperazine involves over 500 degradation products (Brten et al., 2008). The problem is that these degradation pathways are not well known. Simplifications have therefore to be made. Several ongoing research projects are aimed at studying amine degradation, like the ADA-2009 and ADA2010 projects (Nielsen et al., 2010). Thus making an amine chemistry scheme is a task for future research, e.g. ExSIRA project under the KMB program (CLIMIT, Norwegian Research Council and industrial partners). However, in order to get useful results without a chemistry scheme included, a possible approach is to apply scaling. This means multiplying the results by a factor equal to the formation rate of a certain degradation product. To explain in more detail; if for example a chemical component A is emitted at a rate of 1 g/s (unit emission, equals about 31.5 tonnes per year4), then the model is first run for a year emitting 1 g/s. Atmospheric processes like emissions, transport and dry and wet deposition processes are included (see chapter 5.1 for model description). The output from the model gives average concentration of component A, maximum concentration of component A, and annual deposition of component A. Now if we assume that a certain degradation product B is formed with a yield of 1%, when A is degraded/oxidized, 1% of A is degraded into B, then the results of component A may be scaled by 1% (i.e. multiplied by 0.01) to get average concentration, maximum concentration, and annual deposition of this specific degradation product B. This is a simple method to get results for degradation products without including a complex amine chemistry scheme in the model. In section 5.3 the results were scaled according to emission strength, assuming that the components had an atmospheric lifetime of more than about a day, so that they may be considered chemically inert within the model grid. In the calculations discussed in this section, the components are assumed to have a short atmospheric lifetime (less than a few hours) so that chemical degradation occur close to TCM, and well within the model domain (30 km scale). According to the maximizing principle, contribution of directly emitted nitrosamines have been added to the potential formation. In general, primary amines have little potential to form nitrosamines, tertiary amines have higher potential and secondary amines have the highest potential to form nitrosamines. Also cyclic amines like piperazine have large potential to form nitrosamines. The potential to form nitrosamines of the different amines is not easy to quantify. However, based on common knowledge in the literature it is
4

1 g/s 3600 s/hr 24 hrs/day 365 days/year = 31536000 g/year

NILU OR 41/2010

59 most likely between 2 and 10% (see e.g. Pitts et al., 1978, Karl et. al 2008, Brten et. al 2008). It may be lower (e.g. 1%), or it may be higher than 10% (Grosjean, 1991 refer to a 30% potential), but 2-10% is considered as a plausible range. The various amines have different chemical properties, and their atmospheric lifetimes vary. In the same manner nitrosamines are a group of compounds and their chemical properties and toxicity differ. In this worst case analysis we have chosen to give a range of nitrosamine formation potential. We emphasize the 10% value according to the maximizing principle as it is appropriate to carry out in a worst case scenario analysis. It is important to emphasize though, that while applying the worst case method, all uncertain parameters are scaled to the worst possible value. This is according to the precautionary principle. If the calculated worst case concentrations are below the threshold values, this means that concentrations in the real atmosphere will most probably be lower than these threshold values. On the other hand, if the calculated worst case concentrations are above the threshold values, this does not necessarily mean that concentrations in the real atmosphere will be higher than these threshold values. It means that further more detailed investigations are needed to be able to establish a better estimate of the risk of amines and nitrosamines in air. This should be kept in mind when analysing the results in sections 5.4 and 5.5. Concerning nitrosamines, there will be direct emissions of nitrosamines, but also emissions of precursors. These precursors are assumed to form nitrosamines with a 2-10% yield. 5.4.1 Formation of nitrosamines from the MEA emission scenarii Current information for the gas phase potential to form nitrosamines is found in Pitts et al. (1978), and from the MEA oxidation which did not form nitrosamines by atmospheric reactions in the ADA-2009 project (Nielsen et al., 2010). There will be an update to methyl/dimethyl/trimethyl amines in ADA-2010. In the liquid phase all secondary and tertiary amines, but also amides, carbamates, and guanidines can form nitrosamines. Heterocyclic secondary amines (e.g. piperazine) are more easily nitrosated in the liquid phase than the secondary amines. Primary amines form no nitrosamines in the liquid phase. There may be chemical reactions during the amine capture processes that can form nitrosamines. With high NOx concentrations (25-50 ppmv) nitrosamines may form in a plant run with MEA (Steinar Pedersen, IEAGHG seminar Oslo 16. February 2010). With low NOx, no nitrosamines will form. Therefore it is difficult to quantify the exact formation potential from MEA. It is considered to be relatively low, but some nitrosamines may occur, although in small quantities. However the potential of MEA to form nitrosamines under various NOx conditions have to be investigated further. This will e.g. be done in MEA-Test, a project coordinated and funded by NILU. In Table 16 the compounds with potential to form nitrosamines for the MEA emission scenarii are listed.

NILU OR 41/2010

60 Table 16: Components with potential to form nitrosamines in the MEA emission scenarii. In addition there are direct emissions of nitrosamines.
Component Dimethylamine Diethylamine Dibutylamine N-methylethanamine N-methyl 1-butanamine N-ethyl 1-butanamine Dipropylamine DEA

In Table 17, the sum of the potential nitrosamine forming components (assuming 2% and 10% nitrosamine formation rate) and the direct nitrosamine emissions are given for both flue gas treatments for the five MEA scenarii. Table 17: Sum of the potential nitrosamine forming components + direct nitrosamine emissions in the MEA scenarii. For the nitrosamine forming components a 2% yield and a 10% yield is applied. Unit g/s.
Scenario CHP 2% formation yield + direct emissions 10% formation yield + direct emissions RFCC 2% formation yield + direct emissions 10% formation yield + direct emissions Expected Design Worst case Upset Minimum

0.000265

0.000523

0.001055

0.001581

0.000108

0.000962

0.001891

0.003822

0.005726

0.000178

0.000154

0.000309

0.000618

0.000928

7.75E-05

0.000497

0.000994

0.001987

0.002989

0.000112

The sum of direct emissions and nitrosamine precursor is higher for the CHP feed gas than for the RFCC feed gas. Now the sum of estimated sum of direct emissions and precursors may be used to scale the unitary emission results. This line of thought would build upon the inherent assumption that the emission scenario is the dominant condition over a year. For the upset scenario, this assumption is obviously incorrect. For the rest of the scenarii the range from lowest to highest possible emission is from 0.00011 g/s to 0.0038 g/s (for 10% yield). Two typical concentration distributions (for two emission strengths within the range above) are shown in Figure 23. The calculated maximum annual mean concentration values associated with each of the emission strengths are shown in Table 18 , omitting the upset scenario.

NILU OR 41/2010

61

Table 18: Maximum annual mean concentration value for the different MEA scenarii, assuming a 2% and 10% nitrosamine formation yield (+ direct emissions). Unit: ng/m3.
Scenario CHP 1-Expected 2-Design 3-Worst Case 6-Minimum RFCC 1-Expected 2-Design 3-Worst Case 6-minimum Max. concentration assuming 2% nitrosamine formation yield 0.008 0.015 0.032 0.003 Max. concentration assuming 10% nitrosamine formation yield 0.029 0.057 0.11 0.005

0.004 0.009 0.019 0.002

0.015 0.030 0.060 0.003

Worst case emission using feed gas from CHP (over a year) may exceed the EPA/IRIS limit value. All other combinations of scenario and feed gas are calculated to be below the limit value.

Figure 23: Annual mean concentration of nitrosamines for a model run with emissions (direct and precursors) equivalent of 0.00034 g/s (upper panel), and 0.0017g/s (lower panel). Horizontal scale is in km. Units: ng/m3

NILU OR 41/2010

62

Figure 23: Contd.

5.4.2 Formation of nitrosamines from scenario 5 (improved amines) Scenario 5 basically consist of two blends, solvent 3 (S3) forming the basis for scenario 5a (design) and 5b (worst case), and a mix of solvents 1, 4 and 9 (S1,S4,S9) forming the basis for scenario 5c (design) and 5d (worst case). The lists of emitted components are different for the two solvents, but the same for design and worst case sub scenarii. Components that have a potential to form nitrosamines are listed in Table 19.

NILU OR 41/2010

63

Table 19: Components with potential to form nitrosamines in the scenario 5 emission inventory. In addition there are direct emissions of nitrosamines.
Scenario 5a and 5b Tertiary amine Dimethylamine Diethylamine Dibutylamine N-methylethanamine N-methyl 1-butanamine N-ethyl 1-butanamine Dipropylamine 2,2'-[[2-[(2-hydroxyethyl)methylamino]ethyl]imino]bisEthanol N,N.dimethylethanolamine MMEA 2-methylaminoethanol 1-hydroxyethyl-3-methyl imidazolidone Hydroxyethyl oxazolidone DEA Scenario 5c and 5d Secondary amine Tertiary amine Piperazine 4,4-dimethyl-2-oxazolidinone

As with the MEA scenarii, the sum of the potential nitrosamine forming components (with estimated 2% and 10% nitrosamine formation rate) and the direct nitrosamine emissions are given for both flue gas treatments for the scenario 5 inventories. The results are shown in Table 20. Conversion from mixing ratio (ppm) to mass (g/s) of secondary and tertiary amines is based on a molecular weight of 100. Table 20: Sum of the potential nitrosamine forming components + direct nitrosamine emissions in the improved amines scenarii (scenario 5a5d). For the nitrosamine forming components a 2% yield and a 10% yield is applied. Unit g/s.
Scenario CHP 2% formation yield + direct emissions 10% formation yield + direct emissions RFCC 2% formation yield + direct emissions 10% formation yield + direct emissions 5a (S3) Design 5b (S3) Worst case 5c (S1,S4,S9) Design 5d (S1,S4,S9) Worst case

0.0012

0.0015

0.00061

0.00147

0.0063

0.0069

0.00233

0.00664

0.0009

0.0012

0.00047

0.00112

0.0047

0.0052

0.00177

0.00504

NILU OR 41/2010

64 The sum of direct emissions and nitrosamine precursor is higher for the CHP feed gas than for the RFCC feed gas. Now the sum of estimated sum of direct emissions and precursors may be used to scale the unitary emission results. This line of thought would build upon the inherent assumption that the emission scenario is the dominant condition over a year. This assumption would be more valid for design than worst case scenarii. The sum of direct emissions and nitrosamine precursors (assuming 2% and 10% yield) range from 0.0005 g/s to 0.0069 g/s. Two concentration distributions are shown in Figure 24, one for emission near the lowest figure in the range and one near the highest number in the range. The calculated maximum annual mean concentration values associated with each of the emission strengths are shown in Table 21. Table 21: Maximum annual mean concentration value for the different improved amines scenarii (scenario 5a-5d), assuming a 2% and 10% nitrosamine formation yield (+ direct emissions). Unit: ng/m3.
Scenario CHP 5a (S3) Design 5b (S3) Worst Case 5c (S1,S4,S9) Design 5d (S1,S4,S9) Worst Case RFCC 5a (S3) Design 5b (S3) Worst Case 5c (S1,S4,S9) Design 5d (S1,S4,S9) Worst Case Max. concentration assuming 2% nitrosamine formation yield Max. concentration assuming 10% nitrosamine formation yield

0.036 0.045 0.018 0.044

0.19 0.21 0.07 0.20

0.027 0.036 0.014 0.034

0.14 0.16 0.05 0.15

If 2% nitrosamine formation occur, the maximum annual mean values are below the EPA/IRIS limit value of 0.07 ng/m3. For a 10% formation rate, only scenario 5 c Design has a maximum value below the limit.

NILU OR 41/2010

65

Figure 24: Annual mean concentration of nitrosamines for a model run with emissions (direct and precursors) equivalent of 0.0007 g/s (upper panel), and 0.007 g/s (lower panel). Horizontal scale is in km. Units: ng/m3.

5.4.3 Maximum possible emissions of nitrosamines Based on these calculations, to comply with the EPA/IRIS 0.07 ng/m3 value for nitrosamine concentration in air, the maximum possible emission is calculated to

NILU OR 41/2010

66 be 0.0023 g/s given as the sum of directly emitted nitrosamines and 10 % of the sum of components with a potential to form them through reactions after emissions. One example of maximum possible emissions can be found in scenario 5 c Improved Solvent 2 (S1,S4,S9) Design. This in turn corresponds to an emission rate of about 0.02 g/s of components with potential to form nitrosamines. This number is dependent though on the direct emissions of nitrosamines. 5.4.4 Discussion on worst case results The calculated maximum concentrations presented in sections 5.4.1 and 5.4.2. are above the EPA/IRIS threshold value of 0.07 ng/m3. As pointed out in the introduction to this chapter; if the calculated worst case concentrations are above the threshold values, this does not necessarily mean that concentrations in the real atmosphere will be higher than these threshold values. The worst case method reflects the uncertainties in the processes included and further investigations are needed to reduce these uncertainties to provide more accurate calculations of the concentration of amines and nitrosamines in air. The following recommendations are given: 1) Knowledge gaps to be elaborated Formation rate of nitrosamines. Now a 2-10% range is used. The nitrosamine formation potential vary considerably, from practically zero up to 30% (Grosjean, 1991). More scientific knowledge will reduce the uncertainties, and hence refine the formation rate (more compound specific). The type of nitrosamines formed (speciation). Nitrosamines is a large group of compounds where chemical properties and toxicity vary. It is hence crucial to determine which nitrosamines are formed in the atmosphere. The toxicity of the different nitrosamines. Knowledge exist on e.g. tobacco relevant nitrosamines, but less is known concerning the nitrosamines related to CCS. If more toxicity data on these nitrosamines exist there will be possible to identify threshold values for each specific nitrosamine, not for all nitrosamines as a group. Gas phase/particle phase/aqueous phase. The most recent research on amine degradation focus on gas phase chemistry (e.g. ADA-2009, ADA-2010). In the real atmosphere, and especially in the plume from TCM, there will be both gas phase, aqueous phase and particle phase taking place. More knowledge on particle phase and aqueous phase chemistry is needed to understand this chemistry both qualitatively and quantitatively. More information on the actual emissions and their speciation is also needed. In wintertime there is less sunlight in Norway and hence slower photochemistry. Nitrosamines are degraded by photolysis. More information on amine and nitrosamine chemistry in the absence of sunlight is important. 2) Elements that will modify the maximum values in the real atmosphere The TAPM calculations are done for one year giving annual mean concentrations. However there are seasonal variations in the weather pattern. According to the test plan, ACC will run MEA scenarii for 6 months, improved solvent 1 (S3) for 6 months, and improved solvent 2 (S1, S4, S9) for 6 months. Both Design and Worst case scenarii will be tested. No scenario will be run for more than 3 months. Many of these inventories give maximum annual mean values below the threshold value (even in worst case), especially those based on

NILU OR 41/2010

67 MEA. Due to seasonal variations at Mongstad, the future observed annual mean concentrations to occur at Mongstad will be a combination of the various emission inventories tested. The TAPM model runs are performed for the year 2007. Inter-annual variations in weather exist, and hence location and magnitude of the maxima may vary. Gas to particle conversion. Gas molecules behave differently than particles. The calculations presented here assume that everything behaves like molecules. In the real atmosphere particles will be formed. The main conclusion from the worst case study of nitrosamines in air is that there may be a potential problem concerning the EPA/IRIS threshold value of 0.07 ng/m3. Further research, including more detailed modeling, and monitoring is needed to reduce the present uncertainties associated with nitrosamines in air from CCS. 5.5 Worst case scenario nitrosamines in drinking water

There are large knowledge gaps concerning nitrosamine chemistry and the degradation of nitrosamines in water (see e.g. ATSDR, 1989 for more information). This lack of knowledge introduce uncertainties in the calculations. In this study, the worst case method used by Karl et al. (2010) is applied with some modifications. The fundamental principle using the worst case method is that when there are uncertainties, the worst possible option or the worst possible value is applied. In these calculations, the following assumptions have been made: no inter-annual variation in rain pattern. This means that maximum rainfall and maximum rain water concentration will occur at the same location from year to year. 10% nitrosamines formation is applied. In the previous sections a 2-10% range is given. For the worst case, only the maximum value is used. the part of the drinking water catchment that receives the maximum deposition of nitrosamines by rain constitutes 20% of the total catchment. Model calculated maximum deposition occur over a very small area. In these model results maximum deposition would cover an area of approximately 8 km2. It would be placed over two catchment areas, Nordgulelva and Dingja. These catchment areas have sizes of 21 km2 and 28 km2. The peak wet deposition would therefore be diluted by a factor of approximately 5 due to the gradients in the deposition field over these two catchment areas.

One fundamental question concerning nitrosamines in drinking water is whether there is degradation of nitrosamines in water. Some nitrosamines are very persistent (OME, 1991), while others degrade easily (Drewes et al., 2006). In this study these two cases are analyzed separately.

NILU OR 41/2010

68 5.5.1 If no degradation of nitrosamines in water This assumption is in accordance with the worst case study by Karl et al. (2008). If there is no degradation of nitrosamines in water, then at equilibrium, after some years of operation, the drinking water concentration will be 1/5 of the rain water concentration. This is valid if the rain pattern remains constant from year to year, if there is no changes in the emission pattern, and if the peak wet deposition would be diluted by a factor of approximately 5. Then, from the results for unity emissions (section 5.1), emissions of 1 g/s give a maximum deposition of 39 mg/(m2 year). The annual rainfall in the Mongstad region is 1890 mm/year. This is equal to 1890 litre/(m2 year). So, if this mass of 39 mg/(m2 year) is deposited by 1890 litre/(m2 year) of rain, this gives an average rain water concentration of 0.02 mg/litre, or 20 g/litre footnote 5. If this is in turn diluted by a factor 5 from rain water to drinking water, the drinking water concentration will in worst case be 4 g/litre. These results may be used to scale the different emission quantities used in this study. A summary of the results is shown in Table 22. Table 22: Scenario related nitrosamines in drinking water assuming no degradation (persistent nitrosamines). Scenario Emissions* Maximum Max. rain Max. drinking g/s deposition water water mg/(m2 concentration concentration year) Unity emissions 1 39 20 g/litre 4.0 g/litre 1-Expected 0.00096 0.0374 19 ng/litre 3.8 ng/litre 2-Design 0.00189 0.0718 38 ng/litre 7.6 ng/litre 3-Worst case 0.00382 0.149 76 ng/litre 15 ng/litre 6-Minimum 0.00018 0.007 3.6 ng/litre 0.7 ng/litre 5a Design 0.0063 0.0246 126 ng/litre 25 ng/litre 5b Worst case 0.0069 0.269 138 ng/litre 28 ng/litre 5c Design 0.00233 0.0909 47 ng/litre 9.3 ng/litre 5d Worst case 0.00664 0.259 133 ng/litre 26 ng/litre Scenarii 1, 2, 3, and 6 are based on MEA, scenarii 5a-5d are based on improved solvents. * Based on the highest emission from the two possible feed gas flows Assume that the part of the drinking water catchment that receives the maximum deposition of nitrosamines by rain constitutes 20% of the total catchment

All these results show a drinking water concentration above the threshold value of 0.7 ng/litre, except for scenario 6-Minimum based on MEA. 5.5.2 If degradation of nitrosamines in water occur Some studies report that nitrosamines do degrade in water. E.g. Drewes et al., (2006) investigated 7 different nitrosamines (NDMA, NDEA, NMEA, NDPA,
5

39 mg/(m2 year) / 1890 litre/(m2 year) = 0.02 mg/litre

NILU OR 41/2010

69 NDBA, NPIP, and NPYR). Their estimated degradation rate under anoxic conditions vary from k = 0.0976 day-1 to k = 0.5485 day-1. This corresponds to a lifetime (e-folding time) between 1.8 and 10.2 days. This study refers to degradation in soil and soil water. In Norway typically 7-10% of the water in lakes fall directly into lakes. The rest falls onto soil/land. In some extreme cases as much as 80% of lake water may come from direct rainfall (Richard Wright, NIVA, personal communication). In that respect the results from Drewes et al. (2006) may be applicable for this study. If the simplest formula of the mass balance equation is applied, the equilibrium concentration is defined as C = P / k where C is equilibrium concentration, P is production of nitrosamines in the reservoir, and k is chemical decay. Assuming rainfall of 1890 mm/(m2 year), rain water concentration of 138 ng/l (5b-Worst case Table 22), and a ratio of 1:5 for lake surface:lake volume, concentrations reduced by a factor 5 due to catchment size, and k=0.0976 day-1 (Drewes et al., 2006), then the equilibrium concentration C = 0.3 ng/litre. Equilibrium concentrations for the different scenarii are shown in Table 23. These simple calculations show that even with degradation in soil water and water, there may be nitrosamine concentrations close to but below the threshold value. Table 23: Scenario related nitrosamines in drinking water assuming degradation with chemical lifetime equal to 10,2 days (k = 0,0976 day-1). Scenario Emissions* Maximum Max. rain Max. drinking g/s deposition water water 2 mg/(m concentration concentration year) Unity emissions 1 39 20 g/litre 42.4 g/litre 1-Expected 0.00096 0.0374 19 ng/litre 0.04 ng/litre 2-Design 0.00189 0.0718 38 ng/litre 0.08 ng/litre 3-Worst case 0.00382 0.149 76 ng/litre 0.16 ng/litre 6-Minimum 0.00018 0.007 3.6 ng/litre 0.007 ng/litre 5a Design 0.0063 0.0246 126 ng/litre 0.3 ng/litre 5b Worst case 0.0069 0.269 138 ng/litre 0.3 ng/litre 5c Design 0.00233 0.0909 47 ng/litre 0.1 ng/litre 5d Worst case 0.00664 0.259 133 ng/litre 0.3 ng/litre Scenarii 1, 2, 3, and 6 are based on MEA, scenarii 5a-5d are based on improved solvents. * Based on the highest emission from the two possible feed gas flows Assume that the part of the drinking water catchment that receives the maximum deposition of nitrosamines by rain constitutes 20% of the total catchment

5.5.3 Discussion on worst case results As already pointed out the fundamental question concerning nitrosamines in water is whether degradation occur or not. Tate and Alexander (1975) report that

NILU OR 41/2010

70 oxidation, hydrolysis, biotransformation, and biodegradation are not significant factors affecting the fate of NDMA in lake water. Hence photo-degradation is the main process for removing NDMA from the aquatic environment. In Norway, there is a long winter with less sunlight, and hence slower degradation. The lakes and reservoirs are also covered by ice, this prevents the already reduced amount of sunlight to penetrate into water. In addition there is very little sunlight at the bottom of lakes where the water intake is located. This line of thought is also supported by the studies of groundwater compartment, where, in the absence of light, NDMA has the potential to persist (OME, 1991). Nitrosamines (NDMA) is also formed as a byproduct of chlorination , and is frequently found in municipal waste water effluents and surface waters receiving waste water discharges (Drewes et al., 2006). Chlorination is also used for treatment of drinking water. If nitrosamines are formed during chlorination, this means that there will be an existing background concentration of nitrosamines in the drinking water. If so, the possible nitrosamines from CCS will add to this background level and enhance the concentrations even more. This possible background level is not taken into account in this study, but should be kept in mind when investigating and monitoring nitrosamines in water.

These calculations show that if there is little degradation of nitrosamines in water, i.e. if nitrosamines are practically persistent, nitrosamines in drinking water may be a problem. However it is not possible to conclude that there will be a problem with nitrosamines in drinking water. The uncertainties in the present estimates have to be reduced, and scientific research should be conducted, especially to investigate degradation of nitrosamines in water. In addition the concentrations of nitrosamines should be monitored and analyzed when TCM will be in operation.

NILU OR 41/2010

71

Conclusions

In this study, the Norwegian Institute for Air Research has calculated and assessed the dispersion of emissions to air of carbon dioxide (CO2), nitrogen oxides (NOX), ammonia (NH3), amines and amine degradation products from activities at European CO2 Technology Centre Mongstad (TCM). This is an update of the report Test Centre Mongstad, Dispersion calculations for emissions to air from Test Centre Mongstad (TCM) (Berglen et al., 2008). Two technology vendors have been contracted to test their capture technology at TCM; Alstom using Chilled ammonia and Aker Clean Carbon (ACC) using amines. Flue gas from the Residual catalytic cracker (RFCC) and the Combined Heat and Power plant (CHP) will be used to test CO2 capture techniques. Alstom provided two emission scenarios for the calculations (Base case and Worst case). ACC provided five emission scenarios based on MEA (Expected, Design, Worst case, Upset, and Minimum), as well as 4 emission scenarios based on two proprietary solvents (Design and Worst case for Improved solvent 1 and Improved solvent 2 respectively). The calculated maximum hourly mean concentrations from the Alstom facility are 55 g/m3 for NH3 and 71.9 mg/m3 for CO2. From the ACC plant the maximum calculated hourly concentration of MEA is about 2 g/m3 and the maximum calculated hourly mean NH3 concentration is just below 10 g/m3 and the maximum hourly concentration of CO2 is 70.5 mg/m3. The maximum calculated hourly concentrations of formaldehyde, acetaldehyde, and alkylamines are 1,4 g/m3, 2,1 g/m3, and 0.8 g/m3 respectively. All these calculated hourly maximum concentrations for Alstom and ACC are much lower than the Administrative norm and lower than long term exposure guidelines for NH3, CO2, MEA, formaldehyde and acetaldehyde. Concerning regional dispersion, the model calculations estimate a maximum 8hours mean concentration of NH3 of 6.1 g/m3, and a maximum deposition of 3.1 mg N/(m2 year), using Alstom emission data. This deposition is a small increase compared to the current background deposition, but in an area where the vegetation critical loads are already exceeded. Concerning the ACC plant, the calculated maximum annual mean concentration of MEA is 1.3 ng/m3, while the maximum annual mean concentrations of formaldehyde and acetaldehyde are 1.6 ng/m3 and 2.4 ng/m3, respectively. The maximum annual mean concentration of alkyl amines (methylamine, ethylamine, dimethylamine, diethylamine) is just below 1.0 ng/m3. The maximum concentrations for both plants and for all emission scenarios are found about 10 km to the S-E of Mongstad in the area around Sundsb and Kols. All calculated annual mean values are lower than the air quality guidelines for the components where guidelines exist. The maximum tolerable amount of nitrosamines (Worst case) have also been investigated. The calculated nitrosamines concentrations in ambient air are compared to the 0.07 ng/m3 threshold value from EPA/IRIS. A formation range of nitrosamines between 2 and 10 % is assumed for components with nitrosamines formation potential (e.g. Pitts et al., 1978, Karl et. al 2008, Brten et. al 2008).

NILU OR 41/2010

72 Yearly average maximum value of nitrosamines have been calculated for all the scenarii apart from the upset scenario. Considering these results it must be remembered that the emission conditions must be valid for the duration of the year for the concentration value to be valid. For the MEA scenarii, only the worst case emissions exceed the limit value. For the improved solvents scenarii, all exceed the limit except scenario 5c. Calculations of the maximum amount of components with nitrosamine formation potential in the emissions show that the concentration of nitrosamines in air can be below the threshold value, if the sum of component emission is no larger than about 0.02 g/s. This figure is dependent on the amount of directly emitted nitrosamines. Concerning nitrosamines in drinking water, the degradation of nitrosamines in water is a key question. Some studies indicate that nitrosamines are persistent. If this is the case, then all worst case scenarii exceed the 0.7 ng/l threshold value. But even if degradation of nitrosamines in soil water and water occur (as pointed out by Drewes et al., 2006) then the highest worst case scenarii equilibrium concentrations are close to but below the threshold value. Based on these calculations, it is not possible to conclude that there will be a problem with nitrosamines in drinking water, but the uncertainties have to be reduced. Scientific research should be conducted, especially to investigate the degradation of nitrosamines in water.

References

ATSDR (1989) Toxicological profile for N-nitrosodimethylamine. Atlanta, GA, Agency for Toxic Substances and Disease Registry (ATSDR), U.S. Public Health Service in collaboration with U.S. Environmental Protection Agency (EPA). Available at: http://www.atsdr.cdc.gov/toxprofiles/tp141.pdf Aas, W., Hjellbrekke, A., Hole, L.R. and Trseth, K. (2008) Deposition of major inorganic compounds in Norway 2002-2006. Kjeller (NILU OR 72/2008). Achermann, B. and Bobbink, R. (ed.) (2003) Empirical critical loads for nitrogen. Expert workshop, Berne, 11-13 November 2002. Proceedings. Bern, Swiss Agency for the Environment, Forests and Landscape, SAEFL (Environmental documentation 164). Attalla, M. and Azzi, M. (2010) Environmental impacts of emissions from CO2 capture Oral presentation at IEAGHG conference, Oslo 16th February 2010. Berglen, T.F., Gjerstad, K.I., Knudsen, S., Mc Innes, H. and Tnnesen, D.A. (2008) Dispersion calculations of emissions to air from Test Center Mongstad.. Kjeller (NILU OR 12/2008) (In Norwegian). Brooks, S. (2008) The toxicity of selected primary amines and secondary products to aquatic organisms: A review. Oslo, Norwegian Institute for Water Research (NIVA report, 5698-2008).

NILU OR 41/2010

73

Brten, H.B., Bunkan, A.J., Bache-Andreassen, L., Solimannejad, M. and Nielsen, C.J. (2008) Final report on a theoretical study on the atmospheric degradation of selected amines. Kjeller (NILU OR 77/2008). Drewes, J.E., Hoppe C. and Jennings, T. (2006) Fate and transport of Nnitrosamines under conditions simulating full-scale groundwater recharge operations. Water Environ. Res., 78, 2466-2473. Grosjean, D. (1991) Atmospheric chemistry of toxic contaminants. 6. Nitrosamines: Dialkyl nitrosamines and nitrosomorpholine. J. Air Waste Manage. Assoc., 41, 306-311. Karl, M., Brooks, S., Wright, R. and Knudsen, S. (2008) Amines worst case studies. Worst case studies on amine emissions from CO2 capture plants (Task 6). Kjeller (NILU OR 78/2008). Karl, M., Denby, B., Wright, R., Knudsen, S. and Berglen, T.F. (2010) Worst case study method to assess the environmental impact of amine emissions from a CO2 capture plant. Manuscript in preparation. Larsen, T., Lund, E. and Hgsen, T. (2008) Overskridelser av tlegrenser for forsuring og nitrogen for Norge. Oppdatering med perioden 2002-2006. Oslo, Norsk institutt for vannforskning (NIVA Rapport, 5697-2008). Nielsen, C.J., DAnna, B., Dye, C., George, C., Graus, M., Hansel, A., Karl, M., King, S., Musabila, M., Mller, M., Schmidbauer, N., Stenstrm, Y. and Wisthaler, A. (2010) Atmospheric Degradation of Amines (ADA) Summary Report: Gas phase photo-oxidation of 2-aminoethanol (MEA) CLIMIT Project no. 193438. Kjeller (NILU OR 8/2010). Nilsson, J. and Grennfelt, P. (Eds.) (1988) Critical loads for sulphur and nitrogen. UNECE/Nordic Council workshop report, Skokloster, Sweden, March 1988. Copenhagen, Nordic Council of Ministers (NORD 1988:15). Lg, M., Andreassen, ., Instanes, C. and Lindemann, B. (2009a) Health effects of different amines relevant for CO2 capture. Report prepared by Norwegian Institute of Public Health. Oslo/Kjeller (NILU OR 5/2009). Lg, M., Instanes, C., Lindemann, B. and Andreassen, . (2009b) Health effects of possible degradation products of different amines relevant for CO2 capture. Final report. Report prepared by Norwegian Institute of Public Health. Oslo/Kjeller (NILU OR 7/2009). Norwegian Labour Inspection Authority (2010) Veiledning om administrative normer for forurensning i arbeidsatmosfre. URL: http://www.arbeidstilsynet.no/veiledning.html?tid=78034

NILU OR 41/2010

74 OME (1991) N-Nitrosodimethylamine. Toronto, Ontario, Ontario Ministry of the Environment, Hazardous Contaminants Coordination Branch (Scientific Criteria Document for Multimedia Standard Development No. 01-90). REACH (2008) Guidance on information requirements and chemical safety assessments, dated May 2008. URL: http://guidance.echa.europa.eu/docs/guidance_document/information_requirem ents_en.htm Stuanes, A.O. and Abrahamsen, G. (1996) Tlegrenser for nitrogen i skog. En vurdering av kunnskapsgrunnlaget. s, Skogforsk (Aktuelt fra Skogforsk, 7/96). Tamm, C.O. (1991) Nitrogen in terrestrial ecosystems. Questions of productivity, vegetational changes and ecosystem stability. Berlin, Springer Verlag (Ecological Studies, 81). TAPM (2009) The Air Pollution Model (TAPM). Hobart, CSIRO. URL: http://www.cmar.csiro.au/research/tapm Tate, R.L. III and Alexander, M. (1975) Stability of nitrosamines in samples of lake water, soil, and sewage. J. Natl. Cancer Inst., 54, 327-330. U.S. EPA (2010) N-Nitrosodimethylamine (CASRN 62-75-9), II.C.1. Summary of risk estimates. Washington, U.S. Environmental Protection Agency. Available from: Integrated Risk Information System (IRIS), URL: http://www.epa.gov/ncea/iris/subst/0045.htm [2010-08-17]. WHO (2006) N-Nitrosodimethylamine in drinking-water. Background document for development of WHO guidelines for drinking-water quality. Geneva, World Health Organization, Geneva. URL: http://www.who.int/water_sanitation_health/dwq/chemicals/ndma2ndadd.pdf [2010-08-17].

NILU OR 41/2010

Norwegian Institute for Air Research P.O. Box 100, N-2027 Kjeller, Norway Associated with CIENS and the Environmental Research Alliance of Norway ISO certified according to NS-EN ISO 9001
ISBN: 978-82-425-2284-9 (print) ISSN: 0807-7207 SIGN. NO. OF PAGES 74 TITLE CO2 Technology Centre Mongstad updated air dispersion calculations Update of OR 12/2008 PROJECT LEADER Tore Flatlandsmo Berglen NILU PROJECT NO. O-109125 AUTHOR(S) Tore Flatlandsmo Berglen, Dag Tnnesen, Christian Dye, Matthias Karl, Svein Knudsen, and Leonor Tarrasn CONTRACT REF. 4501850931 Contact persons: Vibeke Randgaard Talstad and Elise Otterlei REPORT PREPARED FOR Statoil / TCM ABSTRACT Norwegian Institute for Air Research has performed new air dispersion calculations for European CO2 Technology Centre Mongstad (TCM) based on emission data from the two technology vendors Alstom (using Chilled ammonia) and Aker Clean Carbon, ACC (using amines). Flue gas from the Residual Fluid Catalytic Cracker (RFCC) and Combined Heat and Power plant 3 3 (CHP) will be used. Maximum hourly mean concentrations using Alstom emission data are 55 g/m for NH3 and 12,4 mg/m 3 3 3 for CO2. For ACC the maximum hourly mean concentrations are 2 g/m for MEA, just below 10 g/m for NH3, and 1,4 g/m , 3 3 2,1 g/m , and 0.8 g/m for formaldehyde, acetaldehyde, and alkylamines respectively. Worst case calculations for nitrosamines have also been performed. These calculations show that emission from solvents based on MEA is below the 3 EPA/IRIS threshold value of 0.07 ng/m , except scenario 3. For improved solvents, most scenarii give maximum concentrations above the threshold values. For nitrosamines in drinking water, the degradation of nitrosamines in water is a crucial question. If nitrosamines are persistent, then all scenarii give worst case concentrations above the threshold value. With degradation, concentrations are below the threshold values. However, it is not possible to conclude that there will be a problem, the uncertainties have to be reduced NORWEGIAN TITLE CO2 Technology Centre Mongstad- Oppdatering av spredningsberegninger til luft KEYWORDS Carbon Capture and Storage Amine technology Amine degradation products CLASSIFICATION * B PRICE NOK 150.-

REPORT SERIES SCIENTIFIC REPORT DATE

REPORT NO. OR 41/2010

ABSTRACT (in Norwegian) Norsk institutt for luftforskning har foretatt nye spredningsberegniger for utslipp til luft fra CO2 Technology Centre Mongstad. Utslippstall fra teknologileverandrene Alstom (Chilled ammonia) og Aker Clean Carbon (aminer) er brukt. Timemiddelverdier av NH3, MEA, formaldehyd, acetaldehyd, alkylaminer og CO2 er alle under gjeldende luftkvalitetskriterier. Det samme gjelder for rsmiddel p regional skala. En worst case studie er gjort for nitrosaminer i luft og drikkevann. Resultatene viser at nitrosaminer kan vre et problem bde nr det gjelder luft og drikkevann, men det er store usikkerheter. Mer kunnskap behves, spesielt ang. dannelse av nitroasaminer i luft og ang. nedbrytning av nitrosaminer i drikkevann. * Classification A Unclassified (can be ordered from NILU) B Restricted distribution C Classified (not to be distributed)

REFERENCE: DATE: ISBN

O-109125 AUGUST 2010 978-82-425-2284-9

NILU is an independent, nonprofit institution established in 1969. Through its research NILU increases the understanding of climate change, of the composition of the atmosphere, of air quality and of hazardous substances. Based on its research, NILU markets integrated services and products within analyzing, monitoring and consulting. NILU is concerned with increasing public awareness about climate change and environmental pollution.

REFERENCE: DATE: ISBN

O-109125 AUGUST 2010 978-82-425-2284-9

NILU is an independent, nonprofit institution established in 1969. Through its research NILU increases the understanding of climate change, of the composition of the atmosphere, of air quality and of hazardous substances. Based on its research, NILU markets integrated services and products within analyzing, monitoring and consulting. NILU is concerned with increasing public awareness about climate change and environmental pollution.